Journal of Clinical Endocrinology and Metabolism

Tamura, Ai;Ogasawara, Tatsuki;Fujii, Yoichi;Kaneko, Hiyori;Nakayama, Akitoshi;Higuchi, Seiichiro;Hashimoto, Naoko;Miyabayashi, Yui;Fujimoto, Masanori;Komai, Eri;Kono, Takashi;Sakuma, Ikki;Nagano, Hidekazu;Suzuki, Sawako;Koide, Hisashi;Yokote, Koutaro;Iseki, Kozue;Oguma, Rena;Matsue, Hiroyuki;Nojima, Hiroyuki;Sugiura, Kensuke;Yoshitomi, Hideyuki;Ohtsuka, Masayuki;Rahmutulla, Bahityar;Kaneda, Atsushi;Inoshita, Naoko;Ogawa, Seishi;Tanaka, Tomoaki

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02646pmid: 29688432

Abstract Context Necrolytic migratory erythema (NME) occurs in approximately 70% of patients with glucagonoma syndrome. Excessive stimulation of metabolic pathways by hyperglucagonemia, which leads to hypoaminoacidemia, contributes to NME pathogenesis. However, the molecular pathogenesis of glucagonoma and relationships between metabolic abnormalities and clinical symptoms remain unclear. Patient A 53-year-old woman was referred to our hospital with a generalized rash and weight loss. NME was diagnosed by histopathological examination of skin biopsy tissue. Laboratory tests revealed diabetes, hyperglucagonemia, marked insulin resistance, severe hypoaminoacidemia, ketosis, and anemia. Enhanced computed tomography scans detected a 29-mm pancreatic hypervascular tumor, which was eventually diagnosed as glucagonoma. Preoperative treatment with octreotide long-acting release reduced the glucagon level, improved the amino acid profile, and produced NME remission. Surgical tumor excision normalized the metabolic status and led to remission of symptoms, including NME. Interventions Whole-exome sequencing (WES) and subsequent targeted capture sequencing, followed by Sanger sequencing and pyrosequencing, identified biallelic alteration of death-domain associated protein (DAXX) with a combination of loss of heterozygosity and frameshift mutations (c.553_554del:p.R185fs and c.1884dupC:p.C629fs) in the glucagonoma. Consistently, immunohistochemistry confirmed near-absence of DAXX staining in the tumor cells. Tumor expression of glucagon and somatostatin receptor subtype 2 and 3 messenger RNA was markedly upregulated. Conclusions This is a report of glucagonoma with biallelic inactivation of DAXX determined by WES. The tumor manifested as glucagonoma syndrome with generalized NME. This case showed the relationship between hypoaminoacidemia and NME status. Further investigations are required to elucidate the underlying mechanisms of NME onset and glucagonoma tumorigenesis. Glucagonoma is a rare pancreatic neuroendocrine tumor (PanNET) originating from islets of Langerhans α-cells. Glucagonoma occurs in 3% to 7% of patients with PanNETs, and 8% to 10% of glucagonomas are associated with multiple endocrine neoplasia type 1 (1–4). Glucagonoma syndrome includes weight loss, cheilosis or stomatitis, diarrhea, necrolytic migratory erythema (NME), and diabetes mellitus. NME occurs in 50% to 70% of patients with the syndrome (1, 3, 5). It has been attributed to hypoaminoacidemia and zinc or essential fatty acid deficiency (6); however, details of NME pathogenesis are still unclear. Recent genome-wide analysis identified death-domain associated protein (DAXX) and alpha thalassemia/mental retardation syndrome X-linked (ATRX) as candidate pathogenetic genes for PanNET (7, 8), but the genetics of glucagonoma remain unknown. We describe metabolic and NME changes in a woman with glucagonoma during her clinical course. Genetic analyses including whole-exome sequencing (WES) and subsequent targeted capture sequencing identified somatic mutations at two DAXX sites along with chromosomal loss in broad genomic regions associated with the tumor. Case Description A 53-year-old woman presented to our hospital with anorexia, weight loss, and generalized rash. The rash began on the scalp and demonstrated irregular erythema with erosions and crusts (Fig. 1A). It waxed and waned, without obvious triggers. Lesion biopsy showed parakeratosis, necrosis, and separation of the upper epidermis with vacuolization of keratinocytes (Supplemental Fig. 1), consistent with NME. Figure 1. View largeDownload slide Clinical presentation of glucagonoma syndrome. (A, B) Skin lesions of NME. (A) At patient’s initial visit to our hospital, she had severe erythema with erosions and crusts involving mainly the back, hips, and genital area. (B) At 3 months postoperatively, the NME was completely resolved. (C-E) Imaging studies. (C) Contrast-enhanced CT scan showing a 29 × 25 × 19 mm hypervascular pancreatic tumor (red arrowheads). (D) Gadolinium-enhanced MRI of the tumor demonstrating high signal intensity on T1-weighted imaging (red arrowheads). (E) 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/CT showing mild 18F-FDG uptake in the tumor lesion (maximum standardized uptake value, 2.32). (F, G) Pathological findings of the glucagonoma. (F) Gross appearance of the tumor, which was 30 × 30 mm and contained hemorrhage and necrosis. Scale bar = 1 cm. (G) Hematoxylin and eosin staining showing tumor cells forming a rosette arrangement. (H-K) Immunostaining of the tumor for glucagon (H, in brown), insulin (I, in brown), Ki-67 (J, in brown), and SSTR2 (K, in brown). (L) Chronologic changes in amino acids during the clinical course. Glucogenic, ketogenic, and both amino acids are shown in red, green, and purple characters, respectively. Numbers 1 and 2 represent fold-induction determined by the lower limit of the reference range for each amino acid. On admission, all amino acids (glucogenic and ketogenic) except aspartic acid and glutamic acid were remarkably low (blue line). After Oct-LAR treatment, both essential and nonessential amino acids increased, with some reaching the lower limit of the reference range (yellow line). After surgical treatment, almost all amino acids levels normalized immediately (red line). Amino acids in the pyruvic acid pathway changed more than those in the acetoacetate pathway. Figure 1. View largeDownload slide Clinical presentation of glucagonoma syndrome. (A, B) Skin lesions of NME. (A) At patient’s initial visit to our hospital, she had severe erythema with erosions and crusts involving mainly the back, hips, and genital area. (B) At 3 months postoperatively, the NME was completely resolved. (C-E) Imaging studies. (C) Contrast-enhanced CT scan showing a 29 × 25 × 19 mm hypervascular pancreatic tumor (red arrowheads). (D) Gadolinium-enhanced MRI of the tumor demonstrating high signal intensity on T1-weighted imaging (red arrowheads). (E) 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/CT showing mild 18F-FDG uptake in the tumor lesion (maximum standardized uptake value, 2.32). (F, G) Pathological findings of the glucagonoma. (F) Gross appearance of the tumor, which was 30 × 30 mm and contained hemorrhage and necrosis. Scale bar = 1 cm. (G) Hematoxylin and eosin staining showing tumor cells forming a rosette arrangement. (H-K) Immunostaining of the tumor for glucagon (H, in brown), insulin (I, in brown), Ki-67 (J, in brown), and SSTR2 (K, in brown). (L) Chronologic changes in amino acids during the clinical course. Glucogenic, ketogenic, and both amino acids are shown in red, green, and purple characters, respectively. Numbers 1 and 2 represent fold-induction determined by the lower limit of the reference range for each amino acid. On admission, all amino acids (glucogenic and ketogenic) except aspartic acid and glutamic acid were remarkably low (blue line). After Oct-LAR treatment, both essential and nonessential amino acids increased, with some reaching the lower limit of the reference range (yellow line). After surgical treatment, almost all amino acids levels normalized immediately (red line). Amino acids in the pyruvic acid pathway changed more than those in the acetoacetate pathway. Laboratory tests revealed normochromic normocytic anemia, hypoalbuminemia, and hyperglucagonemia (hemoglobin, 10.4 g/dL; mean corpuscular volume, 86.3 fL; albumin, 3.5 g/dL; plasma glucagon, 402 pg/mL) (Table 1; Supplemental Table 1). Severe hypoaminoacidemia along with marked ketosis suggested a hypermetabolic state (total amino acids, 820.2 nmol/mL; total ketone body, 1263 μmol/L; 3-hydroxybutyric acid, 986 μmol/L) (Table 1; Supplemental Table 1). Glycated hemoglobin was 6.5%, and a 75-g oral glucose tolerance test revealed diabetes with marked hyperinsulinemia, possibly because of hepatic insulin resistance (Table 1). Plasma zinc and other pancreatic hormone levels were normal (Supplemental Table 1). Table 1. NME and Metabolic Profile Changes During the Clinical Course On Admission After Oct-LAR After Surgery NME Waxing and Waning; Generalized Remission; No New Lesions Complete Remission Reference Range Metabolic profile  Glucagon (pg/mL) 402 228 167 7-174  Total ketone body (μmol/L) 1263 184 106 ≧130  Total amino acids (nmol/mL) 820.2 1337.1 2787.5 2068.2–3510.3  EAAs (nmol/mL) 333.0 509.1 792.8 660.0–1222.3  NEAAs (nmol/mL) 469.2 828.0 1994.7 1381.6–2379.4  α-Linolenic acid (μg/mL) 23.3 ND 21.1 11.5–45.8  Linoleic acid (μg/mL) 865.5 ND 1094.5 708.1–1286.0  Arachidonic acid (μg/mL) 160.2 194.5 245.2 135.7–335.3 On Admission After Oct-LAR After Surgery NME Waxing and Waning; Generalized Remission; No New Lesions Complete Remission Reference Range Metabolic profile  Glucagon (pg/mL) 402 228 167 7-174  Total ketone body (μmol/L) 1263 184 106 ≧130  Total amino acids (nmol/mL) 820.2 1337.1 2787.5 2068.2–3510.3  EAAs (nmol/mL) 333.0 509.1 792.8 660.0–1222.3  NEAAs (nmol/mL) 469.2 828.0 1994.7 1381.6–2379.4  α-Linolenic acid (μg/mL) 23.3 ND 21.1 11.5–45.8  Linoleic acid (μg/mL) 865.5 ND 1094.5 708.1–1286.0  Arachidonic acid (μg/mL) 160.2 194.5 245.2 135.7–335.3 On Admission, Time (min) After Surgery, Time (min) 75-g OGTT 0 15 30 60 90 120 0 15 30 60 90 120 Plasma glucose (mmol/L) 6.6 9.3 10.3 10.4 10.5 12.8 4.8 6.2 8.1 8.5 7.3 6.8 IRI (μIU/mL) 12 129 124 120 159 362 3 29 35 48 73 50 On Admission, Time (min) After Surgery, Time (min) 75-g OGTT 0 15 30 60 90 120 0 15 30 60 90 120 Plasma glucose (mmol/L) 6.6 9.3 10.3 10.4 10.5 12.8 4.8 6.2 8.1 8.5 7.3 6.8 IRI (μIU/mL) 12 129 124 120 159 362 3 29 35 48 73 50 Abbreviations: EAA, essential amino acid; IRI, immunoreactive insulin; ND, no data; NEAA, nonessential amino acid; OGTT, oral glucose tolerance test. View Large Table 1. NME and Metabolic Profile Changes During the Clinical Course On Admission After Oct-LAR After Surgery NME Waxing and Waning; Generalized Remission; No New Lesions Complete Remission Reference Range Metabolic profile  Glucagon (pg/mL) 402 228 167 7-174  Total ketone body (μmol/L) 1263 184 106 ≧130  Total amino acids (nmol/mL) 820.2 1337.1 2787.5 2068.2–3510.3  EAAs (nmol/mL) 333.0 509.1 792.8 660.0–1222.3  NEAAs (nmol/mL) 469.2 828.0 1994.7 1381.6–2379.4  α-Linolenic acid (μg/mL) 23.3 ND 21.1 11.5–45.8  Linoleic acid (μg/mL) 865.5 ND 1094.5 708.1–1286.0  Arachidonic acid (μg/mL) 160.2 194.5 245.2 135.7–335.3 On Admission After Oct-LAR After Surgery NME Waxing and Waning; Generalized Remission; No New Lesions Complete Remission Reference Range Metabolic profile  Glucagon (pg/mL) 402 228 167 7-174  Total ketone body (μmol/L) 1263 184 106 ≧130  Total amino acids (nmol/mL) 820.2 1337.1 2787.5 2068.2–3510.3  EAAs (nmol/mL) 333.0 509.1 792.8 660.0–1222.3  NEAAs (nmol/mL) 469.2 828.0 1994.7 1381.6–2379.4  α-Linolenic acid (μg/mL) 23.3 ND 21.1 11.5–45.8  Linoleic acid (μg/mL) 865.5 ND 1094.5 708.1–1286.0  Arachidonic acid (μg/mL) 160.2 194.5 245.2 135.7–335.3 On Admission, Time (min) After Surgery, Time (min) 75-g OGTT 0 15 30 60 90 120 0 15 30 60 90 120 Plasma glucose (mmol/L) 6.6 9.3 10.3 10.4 10.5 12.8 4.8 6.2 8.1 8.5 7.3 6.8 IRI (μIU/mL) 12 129 124 120 159 362 3 29 35 48 73 50 On Admission, Time (min) After Surgery, Time (min) 75-g OGTT 0 15 30 60 90 120 0 15 30 60 90 120 Plasma glucose (mmol/L) 6.6 9.3 10.3 10.4 10.5 12.8 4.8 6.2 8.1 8.5 7.3 6.8 IRI (μIU/mL) 12 129 124 120 159 362 3 29 35 48 73 50 Abbreviations: EAA, essential amino acid; IRI, immunoreactive insulin; ND, no data; NEAA, nonessential amino acid; OGTT, oral glucose tolerance test. View Large Enhanced CT and MRI scans revealed a 29-mm pancreatic tumor (Fig. 1C and1D). Positron emission tomography/CT showed mild uptake in the tumor and normal uptake elsewhere (Fig. 1E). The patient was diagnosed with glucagonoma, and preoperative treatment with 20 mg octreotide long-acting release (Oct-LAR) was begun because she had severe malnutrition and NME; octreotide treatment has been reported to improve NME symptoms in glucagonoma (2, 9–12). Indeed, this treatment improved her glucagon levels, hypoaminoacidemia, and ketosis, and produced NME remission within a month (Table 1). She underwent excision of the central pancreatic lesion. The tumor was 30 × 30 mm and exhibited typical NET pathological findings (Fig. 1F and 1G; Supplemental Fig. 2A). Immunohistochemistry was positive for glucagon (Fig. 1H; Supplemental Fig. 2B), synaptophysin, and chromogranin A, and negligible for insulin (Fig. 1I; Supplemental Fig. 2C). Ki-67 labeling index was approximately 3% (Fig. 1J). The tumor was diagnosed histologically as grade 2 (G2) PanNET (World Health Organization 2010 criteria). Postoperatively, hypoaminoacidemia, ketosis, and diabetes parameters, especially severe insulin resistance, returned to near normal almost immediately (Table 1; Supplemental Table 2). NME resolved completely within one month after surgery (Fig. 1B; Table 1). Materials and Methods The Committee on Ethics in Human Research of Chiba University approved this study. The patient provided written informed consent for publication. Tumor and attached nontumor tissues were carefully dissected from formalin-fixed, paraffin-embedded (FFPE) tissue samples (Supplemental Fig. 3A). Genomic DNA and total RNA were extracted from these tissues using QIAamp DNA FFPE Tissue Kit (Qiagen) and RNeasy FFPE Kit (Qiagen). Genomic DNA was extracted from peripheral blood using MagneSil Blood Genomic, Max Yield System (Promega). WES, targeted capture sequencing, Sanger sequencing, and pyrosequencing are described in the Supplemental Material and Methods. We performed reverse-transcription quantitative PCR as previously described (13, 14). All mRNA expression values were normalized to β-actin. Primers and experimental conditions are described in Supplemental Material and Methods. Detailed information about immunohistochemistry and confocal microscopy analysis is also described in the Supplemental Materials and Methods. Genetic and Gene Expression Analyses To assess the pathophysiology of our patient’s glucagonoma, we performed genetic and gene expression analyses using genomic DNA from peripheral blood and genomic DNA and total RNA from the tumor and attached nontumor tissue. We analyzed the WES data as described in the Supplemental Materials and Methods for somatic mutations and chromosomal aberrations using our in-house pipeline, as previously reported (15). WES analyses based on paired tumor-normal DNA samples identified 31 genes with single-nucleotide variants and/or insertions-deletions as potential candidate genes of somatic mutations in this glucagonoma; a number of altered genes were reasonably distributed by their variant allele frequencies (VAFs) (Supplemental Table 3; Supplemental Fig. 3B). There were two distinct DAXX alterations: a c.553_554del:p.R185fs in exon3 and a c.1884dupC:p.C629fs in exon6 (Fig. 2A), which was located at a colon cancer mutation site reported in the Catalogue of Somatic Mutations in Cancer database (mutation ID: COSM1443783) (Supplemental Materials and Methods). Given that a DAXX gene alteration has been suggested to function as a putative driver gene for PanNET (8), the validation of these mutations by targeted capture sequencing was subsequently performed. We confirmed the presence of these somatic mutations at the DAXX gene locus (Fig. 2A). In this context, it has been shown that copy number changes in PanNET are classified into four groups based on arm length copy number patterns; one of them, recurrent pattern of whole chromosomal loss, was significantly enriched in G2 PanNET (8). Therefore, we next investigated copy number variants according to our in-house pipeline using WES data (Supplemental Materials and Methods) and detected chromosomal loss patterns over broad regions, including chromosomes 1, 2, 3, 6, 8, 10, 11, 15, 16, and 22 that have been previously found in G2 PanNET (Fig. 2B) (8). Notably, chromosomal loss, which appears to be related to a loss of heterozygosity (LOH), was observed at 6p21.32 containing the entire DAXX locus (Fig. 2B, vertical line). These results, together with our findings showing VAF of 0.367 and 0.24 for c.553_554del:p.R185fs and c.1884dupC:p.C629fs, respectively (Fig. 2A), which both result in a truncated form of the DAXX gene product (Fig. 2C), suggest that chromosomal loss, including at the DAXX gene locus, and subsequent truncated mutations cause biallelic inactivation and loss of DAXX function in this case of glucagonoma and may play a pathogenic role in this disease. Because biallelic inactivation of DAXX was detected in approximately 20% of PanNET (8), we decided to confirm a putative second hit mutation of the DAXX locus, in addition to LOH determined by copy number analysis, using Sanger sequencing and pyrosequencing of paired tumor-normal tissue DNA samples. Expectedly, we detected the same DAXX somatic mutations in tumor, but not in nontumor, tissue (Fig. 2D and 2E; Supplemental Fig. 3C and 3D). To assess the DAXX expression at the protein level in tumor cells, we performed immunohistochemistry and confocal microscopy analyses with single and/or double staining for DAXX and glucagon. Consistent with its genetic status, immunohistochemistry revealed no or little DAXX protein expression in tumor cells, particularly in glucagon-positive tumor cells (Fig. 2F; Supplemental Fig. 4C). In contrast, in nontumor tissues, DAXX-positive cells were clearly observed as the major cell population, with high expression in islet cells (Supplemental Fig. 4A, 4B, and 4D). In addition, confocal microscopy analysis confirmed DAXX protein inactivation in glucagon-positive tumor cells at the single-cell level (Fig. 2G), suggesting that DAXX biallelic inactivation is present in this tumor. Figure 2. View largeDownload slide Whole-exome and targeted capture sequencing identification of biallelic inactivation of DAXX gene in the glucagonoma. (A) DAXX mutations revealed by WES and subsequent TCS. Sequencing data of the DAXX gene locus was aligned using the Integrative Genomics Viewer, which identified a frameshift deletion in exon3 (left) and a frameshift insertion in exon6 (right). Sequence depth, variant number, and VAF in WES and TCS are shown at bottom. (B) CN alterations in the patient’s tumor. Total CN is shown in the at top (blue). Hetero-SNPs and allelic ratio are shown in the center (dark green) and bottom (red and green). Chromosomal loss was found in chromosomes 1, 2, 3, 5, 6, 8, 9, 10, 11, 15, 16, 18, 21, 22, and X as indicated by the dark blue line. LOH at 6p21.32 containing the DAXX gene locus is indicated by the vertical line. (C) Domain structure of human DAXX protein. Two distinct gene alterations causing a truncate form of DAXX gene product are shown (arrowheads). (D, E) DAXX mutations were confirmed by Sanger sequencing (D, left: exon3, c.553_554del:p.R185fs; E, right: exon6, c.1884dupC:p.C629fs). (F) Immunostaining for DAXX in the tumor (brown). A magnified view is shown with representative images of DAXX-negative tumor cells and a DAXX-positive nontumor cell (inset). (G) Confocal laser scanning microscopy analysis for DAXX (red), glucagon (green), and DAPI (nuclear staining in blue). (H) Reverse-transcription quantitative PCR analysis showing high mRNA expression of glucagon, SSTR2, and SSTR3 in the tumor part (T), compared with nontumor tissue (N). By contrast, insulin and DAXX mRNA expression was lower in the tumor. Data represent mean ± SEM. CN, copy number; DAPI, 4′,6-diamidino-2-phenylindole; SNP, single nucleotide polymorphism; TCS, targeted capture sequencing. Figure 2. View largeDownload slide Whole-exome and targeted capture sequencing identification of biallelic inactivation of DAXX gene in the glucagonoma. (A) DAXX mutations revealed by WES and subsequent TCS. Sequencing data of the DAXX gene locus was aligned using the Integrative Genomics Viewer, which identified a frameshift deletion in exon3 (left) and a frameshift insertion in exon6 (right). Sequence depth, variant number, and VAF in WES and TCS are shown at bottom. (B) CN alterations in the patient’s tumor. Total CN is shown in the at top (blue). Hetero-SNPs and allelic ratio are shown in the center (dark green) and bottom (red and green). Chromosomal loss was found in chromosomes 1, 2, 3, 5, 6, 8, 9, 10, 11, 15, 16, 18, 21, 22, and X as indicated by the dark blue line. LOH at 6p21.32 containing the DAXX gene locus is indicated by the vertical line. (C) Domain structure of human DAXX protein. Two distinct gene alterations causing a truncate form of DAXX gene product are shown (arrowheads). (D, E) DAXX mutations were confirmed by Sanger sequencing (D, left: exon3, c.553_554del:p.R185fs; E, right: exon6, c.1884dupC:p.C629fs). (F) Immunostaining for DAXX in the tumor (brown). A magnified view is shown with representative images of DAXX-negative tumor cells and a DAXX-positive nontumor cell (inset). (G) Confocal laser scanning microscopy analysis for DAXX (red), glucagon (green), and DAPI (nuclear staining in blue). (H) Reverse-transcription quantitative PCR analysis showing high mRNA expression of glucagon, SSTR2, and SSTR3 in the tumor part (T), compared with nontumor tissue (N). By contrast, insulin and DAXX mRNA expression was lower in the tumor. Data represent mean ± SEM. CN, copy number; DAPI, 4′,6-diamidino-2-phenylindole; SNP, single nucleotide polymorphism; TCS, targeted capture sequencing. We also compared gene expression using reverse-transcription quantitative PCR in tumor and nontumor tissue. Accordingly, glucagon mRNA expression was drastically upregulated in the tumor, whereas insulin mRNA was downregulated (Fig. 2H). Consistent with the effectiveness of Oct-LAR in our patient, both mRNA and protein expression of somatostatin receptor subtypes SSTR2 and SSTR3 were higher in tumor than in nontumor tissues (Fig. 1K; Fig. 2H; Supplemental Fig. 2D and 2E); this reflects previous reports of elevated SSTR2 expression in pancreatic tumors (16). In accordance with our notion, the expression of DAXX mRNA was markedly downregulated in the tumor (Fig. 2H). Discussion In our patient, glucagonoma syndrome presented as diabetes with severe insulin resistance, hypoaminoacidemia, and ketosis. Glucagon directly and indirectly regulates various enzymes, stimulating glycogenolysis and gluconeogenesis and inhibiting glycogenesis and glycolysis (17). Hepatic insulin resistance pivotally influences impaired glucose tolerance in glucagonoma, which likely contributed to the marked insulin resistance in our patient. Because glucagon also regulates amino acid uptake in the liver, severe hypoaminoacidemia mediated by excessive amino acid catabolism occurs frequently with glucagonoma. Although glucagon normally functions in ketogenesis primarily through free fatty acid β-oxidation, catabolism of ketogenic amino acids could promote ketone bodies formation in a pathological context such as glucagonoma. Indeed, hypolipidemia is uncommon in glucagonoma, as in our case. Intriguingly, we found that Oct-LAR treatment produced NME remission in parallel with a large improvement in hypoaminoacidemia and only minor changes in lipids. This suggests that hypoaminoacidemia was a more important contributor to NME than essential fatty acid deficiency. Furthermore, amino acids involved in the pyruvic acid pathway were altered to a greater extent than those in the acetoacetate pathway after treatment (Fig. 1L). This suggests that glucagon-signaling effects on glucogenic amino acid catabolism occur via the liver and may have resulted in NME in this patient. Several somatic mutations and chromosomal aberrations were previously identified in PanNET, including alterations of MENIN, CDKN1B, mTOR signaling pathway, and DAXX/ATRX pathway genes (7, 8, 18). Loss of DAXX/ATRX is thought to be involved in disease pathophysiology and progression through telomerase-independent methods of telomere stabilization: alternative lengthening of telomeres and chromosomal instability (8, 19). DAXX/ATRX mutations were associated with large tumor size, late tumor stage, and poor prognosis in G2 subgroup PanNET (8, 19); thus, biallelic inactivation of ATRX and/or DAXX and LOH may be related to malignant alteration rather than tumorigenesis. Biallelic DAXX inactivation through a combination of LOH and frameshift mutations in our patient suggests that close follow-up will be necessary because malignancy cannot be excluded in terms of the genetic background of this tumor. The DAXX mutation site is located in the histone-binding domain (Fig. 2C) (20), which is functionally linked to formation of the ATRX-DAXX histone chaperone complex that is implicated in gene repression and telomere chromatin structure. Because this complex mediates histone 3.3 deposition, histone-3 lysine-9 trimethylation levels, and DNA methyltransferase 1 recruitment followed by alternative lengthening of telomere activation and/or chromosomal instability, the detected DAXX mutation may result in the loss of epigenetic regulation and thereby contribute to pathogenesis of the glucagonoma. Conclusion In our patient with glucagonoma, NME, and diabetes resolved with treatment. Hypoaminoacidemia was closely related to NME status. This article reports on biallelic DAXX inactivation in glucagonoma. Abbreviations: Abbreviations: ATRX alpha thalassemia/mental retardation syndrome X-linked DAXX death-domain associated protein G2 grade 2 LOH loss of heterozygosity NME necrolytic migratory erythema Oct-LAR octreotide long-acting release PanNET pancreatic neuroendocrine tumor VAF variant allele frequency WES whole-exome sequencing Acknowledgments We thank Erika Sugawara and Noriko Yamanaka for technical support. Financial Support: This work was supported by Grants-in Aid for the Foundation for Growth Science; Advanced Research and Development Programs for Medical Innovation, Scientific Research (B) 17H04037 (T.T.) and (C) 17K09875 (S.S.); Young Scientists (B) 17K16160 (I.S.); the Takeda Science Foundation (T.T.); Foundation for Growth Science in Japan (T.T.); SENSHINE Medical Research Foundation (T.T.); and the Cooperative Research Project Program of the Medical Institute of Bioregulation, Kyushu University (T.T.). Disclosure Summary: The authors have nothing to disclose. References 1. Kindmark H , Sundin A , Granberg D , Dunder K , Skogseid B , Janson ET , Welin S , Oberg K , Eriksson B . Endocrine pancreatic tumors with glucagon hypersecretion: a retrospective study of 23 cases during 20 years . Med Oncol . 2007 ; 24 ( 3 ): 330 – 337 . 2. Ramage JK , Ahmed A , Ardill J , Bax N , Breen DJ , Caplin ME , Corrie P , Davar J , Davies AH , Lewington V , Meyer T , Newell-Price J , Poston G , Reed N , Rockall A , Steward W , Thakker RV , Toubanakis C , Valle J , Verbeke C , Grossman AB ; UK and Ireland Neuroendocrine Tumour Society . Guidelines for the management of gastroenteropancreatic neuroendocrine (including carcinoid) tumours (NETs) . Gut . 2012 ; 61 ( 1 ): 6 – 32 . 3. Falconi M , Eriksson B , Kaltsas G , Bartsch DK , Capdevila J , Caplin M , Kos-Kudla B , Kwekkeboom D , Rindi G , Klöppel G , Reed N , Kianmanesh R , Jensen RT ; Vienna Consensus Conference participants . ENETS Consensus guidelines update for the management of functional p-NETs (F-p-NETs) and non-functional p-NETs (NF-p-NETs) . Neuroendocrinology . 2016 ; 103 :( 2 ): 153 – 171 . 4. Ito T , Igarashi H , Nakamura K , Sasano H , Okusaka T , Takano K , Komoto I , Tanaka M , Imamura M , Jensen RT , Takayanagi R , Shimatsu A . Epidemiological trends of pancreatic and gastrointestinal neuroendocrine tumors in Japan: a nationwide survey analysis . J Gastroenterol . 2015 ; 50 ( 1 ): 58 – 64 . 5. Wermers RA , Fatourechi V , Wynne AG , Kvols LK , Lloyd RV . The glucagonoma syndrome. Clinical and pathologic features in 21 patients . Medicine (Baltimore) . 1996 ; 75 ( 2 ): 53 – 63 . 6. Tierney EP , Badger J . Etiology and pathogenesis of necrolytic migratory erythema: review of the literature . MedGenMed . 2004 ; 6 ( 3 ): 4 . 7. Jiao Y , Shi C , Edil BH , de Wilde RF , Klimstra DS , Maitra A , Schulick RD , Tang LH , Wolfgang CL , Choti MA , Velculescu VE , Diaz LA Jr , Vogelstein B , Kinzler KW , Hruban RH , Papadopoulos N . DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors . Science . 2011 ; 331 ( 6021 ): 1199 – 1203 . 8. Scarpa A , Chang DK , Nones K , Corbo V , Patch AM , Bailey P , Lawlor RT , Johns AL , Miller DK , Mafficini A , Rusev B , Scardoni M , Antonello D , Barbi S , Sikora KO , Cingarlini S , Vicentini C , McKay S , Quinn MC , Bruxner TJ , Christ AN , Harliwong I , Idrisoglu S , McLean S , Nourse C , Nourbakhsh E , Wilson PJ , Anderson MJ , Fink JL , Newell F , Waddell N , Holmes O , Kazakoff SH , Leonard C , Wood S , Xu Q , Nagaraj SH , Amato E , Dalai I , Bersani S , Cataldo I , Dei Tos AP , Capelli P , Davì MV , Landoni L , Malpaga A , Miotto M , Whitehall VL , Leggett BA , Harris JL , Harris J , Jones MD , Humphris J , Chantrill LA , Chin V , Nagrial AM , Pajic M , Scarlett CJ , Pinho A , Rooman I , Toon C , Wu J , Pinese M , Cowley M , Barbour A , Mawson A , Humphrey ES , Colvin EK , Chou A , Lovell JA , Jamieson NB , Duthie F , Gingras MC , Fisher WE , Dagg RA , Lau LM , Lee M , Pickett HA , Reddel RR , Samra JS , Kench JG , Merrett ND , Epari K , Nguyen NQ , Zeps N , Falconi M , Simbolo M , Butturini G , Van Buren G , Partelli S , Fassan M , Khanna KK , Gill AJ , Wheeler DA , Gibbs RA , Musgrove EA , Bassi C , Tortora G , Pederzoli P , Pearson JV , Waddell N , Biankin AV , Grimmond SM ; Australian Pancreatic Cancer Genome Initiative . Whole-genome landscape of pancreatic neuroendocrine tumours . Nature . 2017 ; 543 ( 7643 ): 65 – 71 . 9. Kaltsas G , Caplin M , Davies P , Ferone D , Garcia-Carbonero R , Grozinsky-Glasberg S , Hörsch D , Tiensuu Janson E , Kianmanesh R , Kos-Kudla B , Pavel M , Rinke A , Falconi M , de Herder WW ; Antibes Consensus Conference participants . ENETS Consensus Guidelines for the Standards of Care in neuroendocrine tumors: pre- and perioperative therapy in patients with neuroendocrine tumors . Neuroendocrinology . 2017 ; 105 ( 3 ): 245 – 254 . 10. Kimbara S , Fujiwara Y , Toyoda M , Chayahara N , Imamura Y , Kiyota N , Mukohara T , Fukunaga A , Oka M , Nishigori C , Minami H . Rapid improvement of glucagonoma-related necrolytic migratory erythema with octreotide . Clin J Gastroenterol . 2014 ; 7 ( 3 ): 255 – 259 . 11. Lo CH , Ho CL , Shih YL . Glucagonoma with necrolytic migratory erythema exhibiting responsiveness to subcutaneous octreotide injections . QJM . 2014 ; 107 ( 2 ): 157 – 158 . 12. Wei J , Lin S , Wang C , Wu J , Qian Z , Dai C , Jiang K , Miao YI . Glucagonoma syndrome: A case report . Oncol Lett . 2015 ; 10 ( 2 ): 1113 – 1116 . 13. Sakuma I , Higuchi S , Fujimoto M , Takiguchi T , Nakayama A , Tamura A , Kohno T , Komai E , Shiga A , Nagano H , Hashimoto N , Suzuki S , Mayama T , Koide H , Ono K , Sasano H , Tatsuno I , Yokote K , Tanaka T . Cushing syndrome due to ACTH-secreting pheochromocytoma, aggravated by glucocorticoid-driven positive-feedback loop . J Clin Endocrinol Metab . 2016 ; 101 ( 3 ): 841 – 846 . 14. Takiguchi T , Koide H , Nagano H , Nakayama A , Fujimoto M , Tamura A , Komai E , Shiga A , Kono T , Higuchi S , Sakuma I , Hashimoto N , Suzuki S , Miyabayashi Y , Ishiwatari N , Horiguchi K , Nakatani Y , Yokote K , Tanaka T . Multihormonal pituitary adenoma concomitant with Pit-1 and Tpit lineage cells causing acromegaly associated with subclinical Cushing’s disease: a case report . BMC Endocr Disord . 2017 ; 17 ( 1 ): 54 . 15. Suzuki H , Aoki K , Chiba K , Sato Y , Shiozawa Y , Shiraishi Y , Shimamura T , Niida A , Motomura K , Ohka F , Yamamoto T , Tanahashi K , Ranjit M , Wakabayashi T , Yoshizato T , Kataoka K , Yoshida K , Nagata Y , Sato-Otsubo A , Tanaka H , Sanada M , Kondo Y , Nakamura H , Mizoguchi M , Abe T , Muragaki Y , Watanabe R , Ito I , Miyano S , Natsume A , Ogawa S . Mutational landscape and clonal architecture in grade II and III gliomas . Nat Genet . 2015 ; 47 ( 5 ): 458 – 468 . 16. Reubi JC , Waser B , Schaer JC , Laissue JA . Somatostatin receptor sst1-sst5 expression in normal and neoplastic human tissues using receptor autoradiography with subtype-selective ligands [published correction appears in Eur J Nucl Med. 2001;28(9):1433]. Eur J Nucl Med . 2001 ; 28 ( 7 ): 836 – 846 . 17. Quesada I , Tudurí E , Ripoll C , Nadal A . Physiology of the pancreatic alpha-cell and glucagon secretion: role in glucose homeostasis and diabetes . J Endocrinol . 2008 ; 199 ( 1 ): 5 – 19 . 18. Maxwell JE , Sherman SK , Li G , Choi AB , Bellizzi AM , O'Dorisio TM , Howe JR . Somatic alterations of CDKN1B are associated with small bowel neuroendocrine tumors [published online ahead of print September 15, 2015] . Cancer Genet . doi: 10.1016/j.cancergen.2015.08.003. 19. Di Domenico A , Wiedmer T , Marinoni I , Perren A . Genetic and epigenetic drivers of neuroendocrine tumours (NET) . Endocr Relat Cancer . 2017 ; 24 ( 9 ): R315 – R334 . 20. Hoelper D , Huang H , Jain AY , Patel DJ , Lewis PW . Structural and mechanistic insights into ATRX-dependent and -independent functions of the histone chaperone DAXX . Nat Commun . 2017 ; 8 ( 1 ): 1193 . Copyright © 2018 Endocrine Society

Salih, Mahdi;Bovée, Dominique M;van der Lubbe, Nils;Danser, Alexander H J;Zietse, Robert;Feelders, Richard A;Hoorn, Ewout J

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00065pmid: 29726949

Abstract Context Increased renal sodium reabsorption contributes to hypertension in Cushing syndrome (CS). Renal sodium transporters can be analyzed noninvasively in urinary extracellular vesicles (uEVs). Objective To analyze renal sodium transporters in uEVs of patients with CS and hypertension. Design Observational study. Setting University hospital. Participants The uEVs were isolated by ultracentrifugation and analyzed by immunoblotting in 10 patients with CS and 7 age-matched healthy participants. In 7 patients with CS, uEVs were analyzed before and after treatment. Main Outcome Measure Abundance of protein in uEVs. Results The 10 patients with CS were divided in those with suppressed and nonsuppressed renin-angiotensin-aldosterone system (RAAS; n = 5 per group). Patients with CS with suppressed RAAS had similar blood pressure but significantly lower serum potassium than patients with CS with nonsuppressed RAAS. Compared with healthy participants, only patients with suppressed RAAS had higher phosphorylated Na+-K+-Cl− cotransporter type 2 (pNKCC2) and higher total and phosphorylated Na+-Cl− cotransporter (pNCC) in uEVs. Serum potassium but not urinary free cortisol correlated with pNKCC2, pNCC, and Na+-Cl− cotransporter (NCC) in uEVs. Treatment of CS reversed the increases in pNKCC2, NCC, and pNCC. Conclusions CS increases renal sodium transporter abundance in uEVs in patients with hypertension and suppressed RAAS. Potassium has recently been identified as an important driver of NCC activity, and low serum potassium may also contribute to increased renal sodium reabsorption and hypertension in CS. These results may also be relevant for hypertension induced by exogenous glucocorticoids. A common clinical feature of Cushing syndrome (CS) is hypertension, which occurs in ~75% of patients (1, 2). Because of the pleiotropic effects of glucocorticoids, the pathogenesis of hypertension in CS is believed to be multifactorial (3–8). In the kidneys, excess glucocorticoids increase renal blood flow and increase renal tubular sodium reabsorption (3). Glucocorticoids can activate all of the major apical sodium transport proteins along the nephron (9–12). This includes the Na+/H+ exchanger type 3 (NHE3) in the proximal tubule (9), the Na+-K+-Cl− cotransporter type 2 (NKCC2) in the thick ascending limb of the loop of Henle (11), the Na+-Cl− cotransporter (NCC) in the distal convoluted tubule (10), and the epithelial sodium channel (ENaC) in the collecting duct (3, 11). In addition to these sodium transporters expressed at the apical plasma membrane, glucocorticoids can also increase activity of the basolateral Na+-K+-ATPase (13). Activation of these transport proteins may be mediated through activation of the glucocorticoid receptor, which is expressed throughout the nephron (3, 14). In addition, supraphysiological concentrations of cortisol can overwhelm the capacity of 11β-hydroxysteroid dehydrogenase type 2 to inactivate cortisol to cortisone. This enables activation of the mineralocorticoid receptor, with subsequent activation of NCC and ENaC. Because ENaC is electrochemically coupled to potassium secretion, CS may cause hypokalemic hypertension, similar to primary aldosteronism (15, 16). However, mineralocorticoid receptor antagonists do not fully prevent hypertension in CS, suggesting that this mechanism does not fully explain hypertension in CS (15, 17). The effects of glucocorticoids on renal sodium transport have mainly been studied in vitro and in experimental animals (9–11, 13, 18). Isolation of urinary extracellular vesicles (uEVs) in healthy persons and patients now allows for a noninvasive analysis of renal sodium transport. uEVs are nanosized vesicles that can be isolated from human urine and contain renal transport proteins (19). We previously confirmed in both rats and humans that aldosterone activation of sodium transporters in the kidney results in corresponding changes of these transporters in uEVs (20). In the current study, we analyze renal sodium transport proteins in uEVs isolated from patients with CS and hypertension. In addition to total abundances, we also analyze the phosphorylated forms of NCC (pNCC) and NKCC2 (pNKCC2), which are considered the active forms of these transporters (21). Because ENaC is difficult to analyze in uEVs, we instead analyzed one of its activating serine proteases, prostasin, as has been done previously (20, 22). Finally, we analyzed the pleiotropic kinase Rac1 because it has been considered a measure of mineralocorticoid receptor activity and is detectable in uEVs (23, 24). Materials and Methods Patients and healthy controls Our medical ethics committee approved this study (MEC-2007-048). All patients gave written informed consent. Consecutive patients with newly diagnosed or recurrent CS in our center were considered for inclusion during the period 2012–2015. ACTH-dependent and ACTH-independent hypercortisolism was biochemically established with measurement of 24-hour urinary cortisol excretion, late-night salivary cortisol, plasma cortisol levels (8:00 am) after 1 mg dexamethasone overnight, and ACTH concentrations (25). To analyze renal sodium transporters in uEVs, we excluded patients who were using medication that interferes with renal sodium transport (renin-angiotensin inhibitors or diuretics). Patients with CS were further subdivided into those with or without suppressed plasma renin-angiotensin-aldosterone system (RAAS), as defined by the lower levels of normal for plasma renin (10 μU/mL) and/or aldosterone of (50 pg/mL). The rationale for this was that patients with suppressed RAAS are more likely to have increased renal sodium reabsorption by other mechanisms than the RAAS itself, which may be reflected in uEV transporter status. Biochemical and clinical remission was reached in all patients. For 7 of the 10 patients with CS, urinary samples were available before and after treatment. Healthy participants were recruited and matched to patients with CS by age; they did not use any drugs and had no hypertension or history of endocrine disorders or chronic kidney disease. Measurements Blood pressure was measured in the supine position by using an automatic oscillometric device for at least 15 minutes; the average of the last three measurements was used for analysis. Hypertension was defined as systolic blood pressure ≥ 135 mm Hg and/or diastolic blood pressure ≥ 85 mm Hg (26). Both 24-hour urine and spot urine samples were collected. Plasma and urine electrolytes were measured by using an ion-selective electrode (ISE Indirect, Cobas; Roche Diagnostics, Mannheim, Germany). Creatinine was measured by using an enzymatic colorimetric method (Crep2, Cobas; Roche Diagnostics). Renin in plasma was measured by using an immunoradiometric kit (Renin III; Cisbio, Gif-sur-Yvette, France). Aldosterone was measured by solid-phase radioimmunoassay (Diagnostic Products Corp., Los Angeles, CA). Plasma renin and aldosterone were measured midmorning after the patient had been up for at least 2 hours, with blood collection after the patient had been seated for 5 to 15 minutes (27). Plasma and saliva cortisol were measured by using liquid chromatography coupled to mass spectrometry. Urinary free cortisol is expressed as the fold-elevation above the upper limit of normal. Isolation and immunoblotting of uEVs The spot urine samples from which uEVs were isolated were collected around the time of the blood collections. Spot urine samples were collected in a urinary container with a protease inhibitor (cOmplete; Roche, Woerden, The Netherlands) and centrifuged (3000g for 5 minutes) to remove cells and debris before storage at −80°C. The uEVs were isolated by using high speed and ultracentrifugation, as described previously (28, 29). Briefly, urine was first centrifuged at 17,000g for 15 minutes to pellet high-density particles, separating the supernatant (supernatant 1). Dithiothreitol was then used to disrupt the Tamm-Horsfall polymers, after which the samples were diluted in isolation buffer and centrifuged at 17,000g (supernatant 2) (29). The two supernatants were then combined and ultracentrifuged at 200,000g for 2 hours. The pellets were suspended in Laemmli buffer for immunoblot analysis and subsequently heated at 60°C for 15 minutes. The uEVs of patients with CS and healthy controls were isolated simultaneously. Urine creatinine was used for normalization of spot samples, as done previously (28, 30). SDS-PAGE was carried out on a gradient gel (4% to 20%), after which the gel was transferred to a Trans-Blot Turbo system (Bio-Rad, Hercules, CA). The membranes were blocked in 5% milk or BSA and incubated overnight at 4°C. Antibodies against the following proteins were used: NHE3 [1:1000; StressMarq (31)], NCC [1:1000; StressMarq (32)], pNCC [1:500, kindly provided by Dr. Fenton (33)], NKCC2 (1:1000, kindly provided by Dr. Knepper), pNKCC2 [1:500, kindly provided by Dr. Mutig (34)], Rac1 [1:500; Millipore, product # 05-389 (35)], prostasin (1:500; BD Biosciences), aquaporin 2 [1:1000; StressMarq (36)], Na+/K+-ATPase (1:250; Abcam; product # ab7671), and CD9 (1:500; Santa Cruz Biotechnology; product # sc-13118). Secondary antibodies were peroxidase-conjugated goat anti-rabbit or -mouse (1:3000; Sigma-Aldrich). Imaging was obtained by using an Amersham Imager 600 (GE Healthcare, Uppsala, Sweden), and the densitometry measurements were performed by using ImageQuant, version 8.1.0.0 (GE Healthcare, Chicago, IL). All immunoblots were performed twice to verify consistency and reproducibility. CD9 was used as a measure of uEV number (28). Although urine creatinine and CD9 are highly correlated (28), we did not perform a second normalization on CD9 and did not correct densitometry for CD9. The reason to do so reflects the unresolved issue of whether differences in uEV proteins between groups reflect more protein per uEV or more uEVs expressing this protein. Statistical analyses Results are expressed as mean and SD or median and range, as appropriate. Data were logarithmically transformed before analysis in case of non-normal distribution. A Student t test or ANOVA was used for group comparison. A paired t test was used to analyze effects in uEVs before and after treatment. Pearson correlation was used to analyze correlations between serum levels of potassium or aldosterone to urinary sodium transporters (NCC, pNCC, and NKCC2). A P value < 0.05 was considered to indicate a statistically significant difference. Statistical analyses were performed with SPSS software, version 21 (IBM, Armonk, NY). Results Patient characteristics Out of the 18 consecutive patients with CS, 8 were excluded because they used interfering medication. The remaining 10 patients formed the study population (Table 1). They were compared with seven healthy participants. Eight patients with CS had an ACTH-producing pituitary adenoma (one macroadenoma, seven microadenoma), whereas two patients had a cortisol-producing adrenal adenoma. In the latter patients there was no biochemical evidence for aldosterone cosecretion (low plasma aldosterone levels). The 10 patients with CS were further divided into those with suppressed RAAS and those with nonsuppressed RAAS (Fig. 1). Patients with CS with suppressed RAAS had significantly lower body mass index (BMI), higher estimated glomerular filtration rate, and lower serum potassium levels. Blood pressure did not differ between the two groups. In each of these groups, one patient used metoprolol. Table 1. Patient Characteristics Characteristic All Patients (n = 10) Suppressed RAAS (n = 5) Nonsuppressed RAAS (n = 5) Healthy Participants (n = 7) Clinical  Age, y 43.2 ± 7.8 40.8 ± 10.9 45.6 ± 2.3 43.7 ± 2.4  Men, n (%) 5 (50) 2 (20) 3 (43) 3 (43)  BMI, kg/m2 29 ± 8 24 ± 3 34 ± 8a 24 ± 3  Cause of CS, n (%)   Pituitary adenoma 8 (80) 3 (60) 5 (80) —   Adrenocortical adenoma 2 (20) 2 (40) 0 —  Systolic BP, mm Hg 144 ± 14 141 ± 18 147 ± 11 128 ± 8  Diastolic BP, mm Hg 92 ± 17 89 ± 23 96 ± 11 90 ± 8 Blood  Sodium, mmol/L 141 ± 3 142 ± 3 140 ± 2 —  Potassium, mmol/L 4.2 ± 0.4 3.9 ± 0.2a 4.4 ± 0.3 —  eGFR, mL/min per 1.73 m2 94 ± 18 107 ± 10a 80 ± 12 — Urine  Sodium, mmol/d 196 (138–273) 196 (138–273) 196 (173–247) —  Free cortisol, × ULNb 1.9 (0.3–6.1) 2.0 (0.3–6.1) 1.6 (1.0–2.3) — Saliva  Late-night cortisol, nmol/L 13.9 (3.8–70.6) 11.2 (3.8–70.6) 18.1 (7.8–30.4) — Characteristic All Patients (n = 10) Suppressed RAAS (n = 5) Nonsuppressed RAAS (n = 5) Healthy Participants (n = 7) Clinical  Age, y 43.2 ± 7.8 40.8 ± 10.9 45.6 ± 2.3 43.7 ± 2.4  Men, n (%) 5 (50) 2 (20) 3 (43) 3 (43)  BMI, kg/m2 29 ± 8 24 ± 3 34 ± 8a 24 ± 3  Cause of CS, n (%)   Pituitary adenoma 8 (80) 3 (60) 5 (80) —   Adrenocortical adenoma 2 (20) 2 (40) 0 —  Systolic BP, mm Hg 144 ± 14 141 ± 18 147 ± 11 128 ± 8  Diastolic BP, mm Hg 92 ± 17 89 ± 23 96 ± 11 90 ± 8 Blood  Sodium, mmol/L 141 ± 3 142 ± 3 140 ± 2 —  Potassium, mmol/L 4.2 ± 0.4 3.9 ± 0.2a 4.4 ± 0.3 —  eGFR, mL/min per 1.73 m2 94 ± 18 107 ± 10a 80 ± 12 — Urine  Sodium, mmol/d 196 (138–273) 196 (138–273) 196 (173–247) —  Free cortisol, × ULNb 1.9 (0.3–6.1) 2.0 (0.3–6.1) 1.6 (1.0–2.3) — Saliva  Late-night cortisol, nmol/L 13.9 (3.8–70.6) 11.2 (3.8–70.6) 18.1 (7.8–30.4) — Values expressed with a plus/minus sign are the mean ± SD. Non-normal distributed data are shown as geometric mean and range. Abbreviations: BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; NT, not tested; ULN, upper limit of normal. a P < 0.05 (suppressed vs nonsuppressed RAAS). b Average of two measurements. View Large Table 1. Patient Characteristics Characteristic All Patients (n = 10) Suppressed RAAS (n = 5) Nonsuppressed RAAS (n = 5) Healthy Participants (n = 7) Clinical  Age, y 43.2 ± 7.8 40.8 ± 10.9 45.6 ± 2.3 43.7 ± 2.4  Men, n (%) 5 (50) 2 (20) 3 (43) 3 (43)  BMI, kg/m2 29 ± 8 24 ± 3 34 ± 8a 24 ± 3  Cause of CS, n (%)   Pituitary adenoma 8 (80) 3 (60) 5 (80) —   Adrenocortical adenoma 2 (20) 2 (40) 0 —  Systolic BP, mm Hg 144 ± 14 141 ± 18 147 ± 11 128 ± 8  Diastolic BP, mm Hg 92 ± 17 89 ± 23 96 ± 11 90 ± 8 Blood  Sodium, mmol/L 141 ± 3 142 ± 3 140 ± 2 —  Potassium, mmol/L 4.2 ± 0.4 3.9 ± 0.2a 4.4 ± 0.3 —  eGFR, mL/min per 1.73 m2 94 ± 18 107 ± 10a 80 ± 12 — Urine  Sodium, mmol/d 196 (138–273) 196 (138–273) 196 (173–247) —  Free cortisol, × ULNb 1.9 (0.3–6.1) 2.0 (0.3–6.1) 1.6 (1.0–2.3) — Saliva  Late-night cortisol, nmol/L 13.9 (3.8–70.6) 11.2 (3.8–70.6) 18.1 (7.8–30.4) — Characteristic All Patients (n = 10) Suppressed RAAS (n = 5) Nonsuppressed RAAS (n = 5) Healthy Participants (n = 7) Clinical  Age, y 43.2 ± 7.8 40.8 ± 10.9 45.6 ± 2.3 43.7 ± 2.4  Men, n (%) 5 (50) 2 (20) 3 (43) 3 (43)  BMI, kg/m2 29 ± 8 24 ± 3 34 ± 8a 24 ± 3  Cause of CS, n (%)   Pituitary adenoma 8 (80) 3 (60) 5 (80) —   Adrenocortical adenoma 2 (20) 2 (40) 0 —  Systolic BP, mm Hg 144 ± 14 141 ± 18 147 ± 11 128 ± 8  Diastolic BP, mm Hg 92 ± 17 89 ± 23 96 ± 11 90 ± 8 Blood  Sodium, mmol/L 141 ± 3 142 ± 3 140 ± 2 —  Potassium, mmol/L 4.2 ± 0.4 3.9 ± 0.2a 4.4 ± 0.3 —  eGFR, mL/min per 1.73 m2 94 ± 18 107 ± 10a 80 ± 12 — Urine  Sodium, mmol/d 196 (138–273) 196 (138–273) 196 (173–247) —  Free cortisol, × ULNb 1.9 (0.3–6.1) 2.0 (0.3–6.1) 1.6 (1.0–2.3) — Saliva  Late-night cortisol, nmol/L 13.9 (3.8–70.6) 11.2 (3.8–70.6) 18.1 (7.8–30.4) — Values expressed with a plus/minus sign are the mean ± SD. Non-normal distributed data are shown as geometric mean and range. Abbreviations: BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; NT, not tested; ULN, upper limit of normal. a P < 0.05 (suppressed vs nonsuppressed RAAS). b Average of two measurements. View Large Figure 1. View largeDownload slide Plasma renin and aldosterone concentrations in patients with CS. Based on the lower limits of normal (dashed lines), these values were used to classify patients into nonsuppressed RAAS (NS-RAAS) or suppressed RAAS (S-RAAS). Two patients had low-normal plasma renin concentrations (13 and 17 μU/mL), but suppressed plasma aldosterone concentrations (17 and 24 pg/mL). Error bars represent standard deviations. Figure 1. View largeDownload slide Plasma renin and aldosterone concentrations in patients with CS. Based on the lower limits of normal (dashed lines), these values were used to classify patients into nonsuppressed RAAS (NS-RAAS) or suppressed RAAS (S-RAAS). Two patients had low-normal plasma renin concentrations (13 and 17 μU/mL), but suppressed plasma aldosterone concentrations (17 and 24 pg/mL). Error bars represent standard deviations. CS increases renal sodium transporters in uEVs Total NCC and the phosphorylated forms of NCC and NKCC2 were more abundant only in uEVs of patients with CS and a suppressed RAAS (Fig. 2). In these patients, the uEV abundances of pNCC and pNKCC2 were increased three- to fourfold. In addition, the pNKCC2-to-total NKCC2 ratio was increased only in the patients with CS and suppressed RAAS. Although the differences in NHE3 between patients with CS and healthy participants were statistically significant, this result was largely due to the strong signals in patients 1 and 6. We also observed a trend toward higher Na+-K+-ATPase abundance in uEVs from patients with a nonsuppressed RAAS, but this did not reach statistical significance. None of the other analyzed proteins in uEVs showed significant differences in abundance, including aquaporin-2, Rac1, and prostasin. Figure 2. View largeDownload slide (A) Immunoblot analysis of renal sodium transporters and related proteins in uEVs from patients with CS and healthy participants. Patients with CS were divided according to a suppressed (S) or nonsuppressed (NS) RAAS. Numbers indicate individual patients and correspond to those in Fig. 4. F, female; M, male. (B) Group comparisons of immunoblot densitometry were performed with ANOVA using log-transformed data of ODs. *P < 0.01. Error bars represent standard deviations. Figure 2. View largeDownload slide (A) Immunoblot analysis of renal sodium transporters and related proteins in uEVs from patients with CS and healthy participants. Patients with CS were divided according to a suppressed (S) or nonsuppressed (NS) RAAS. Numbers indicate individual patients and correspond to those in Fig. 4. F, female; M, male. (B) Group comparisons of immunoblot densitometry were performed with ANOVA using log-transformed data of ODs. *P < 0.01. Error bars represent standard deviations. Serum potassium determines sodium cotransporter status in uEVs Although within the normal range, serum potassium was significantly lower in patients with CS and suppressed RAAS (Table 1). Serum potassium has recently been recognized as an important driver of NCC activity (37, 38). Therefore, we analyzed the correlations between serum potassium and aldosterone, NCC, pNCC, and pNKCC2 (Fig. 3). Although serum potassium and aldosterone showed a positive correlation, it seems unlikely that the normal serum potassium concentrations suppressed aldosterone. Of interest, serum potassium correlated negatively with NCC, pNCC, and pNKCC2 abundances in uEVs. No correlation was observed between urinary cortisol levels, renin, aldosterone, and transporter abundances (data not shown). Figure 3. View largeDownload slide (A) Correlations between the serum potassium concentration and aldosterone and serum potassium, pNKCC2, pNCC, and (B) NCC in uEVs of patients with CS. Data from one patient are missing because of an hemolytic sample. AU, arbitrary units. Figure 3. View largeDownload slide (A) Correlations between the serum potassium concentration and aldosterone and serum potassium, pNKCC2, pNCC, and (B) NCC in uEVs of patients with CS. Data from one patient are missing because of an hemolytic sample. AU, arbitrary units. Effect of treatment In seven patients, uEVs were analyzed before and after successful treatment of CS (Fig. 4). NHE3 decreased in all patients with CS, although the decrease in the patients with a suppressed RAAS was of borderline significance (P = 0.08). In patients with suppressed RAAS, the abundances of pNKCC2, NCC, and pNCC decreased significantly after treatment (∼threefold to fivefold). The seven patients received the following treatments for CS. Four of five patients with newly diagnosed Cushing disease underwent transsphenoidal adenomectomy, which was successful in three patients, whereas one patient was subsequently treated with bilateral adrenalectomy because of persistent hypercortisolism. One patient without a visible adenoma was treated with ketoconazole with normalization of cortisol production. Three patients with recurrent Cushing disease were treated with steroid synthesis inhibitors resulting in biochemical remission. Finally, two patients with adrenal CS underwent an uncomplicated laparoscopic adrenalectomy. Figure 4. View largeDownload slide Immunoblot analysis of renal sodium transporters and related proteins in uEVs from patients with CS (A) after and (B) before treatment, in those with suppressed RAAS (S-RAAS) or nonsuppressed RAAS (NS-RAAS). (A) Patient numbers correspond with the numbers in Fig. 2. (B) The log-transformed ODs were analyzed by using a paired t test in which the post-treatment abundances were set at 1. *P < 0.05. Error bars represent standard deviations. OD, optical density. Figure 4. View largeDownload slide Immunoblot analysis of renal sodium transporters and related proteins in uEVs from patients with CS (A) after and (B) before treatment, in those with suppressed RAAS (S-RAAS) or nonsuppressed RAAS (NS-RAAS). (A) Patient numbers correspond with the numbers in Fig. 2. (B) The log-transformed ODs were analyzed by using a paired t test in which the post-treatment abundances were set at 1. *P < 0.05. Error bars represent standard deviations. OD, optical density. Discussion The pathogenesis of hypertension in CS is complex and not entirely understood. To further elucidate the mechanism of how hypercortisolism increases blood pressure, we isolated and analyzed uEVs in patients with CS and hypertension to analyze renal sodium transport. We hypothesized that patients with suppressed RAAS are more likely to have increased renal sodium reabsorption by other mechanisms, which may be reflected in uEV transporter status. uEVs are nanosized vesicles that contain all of the major renal sodium transport proteins and reflect their activity in the kidney (19, 20, 39, 40). Using this approach, we found that only patients with CS and a suppressed RAAS had increased abundances of NCC. Of interest, this activation of NCC appears not to reflect the mineralocorticoid effect of hypercortisolism but rather is an indirect effect induced by a reduced serum potassium concentration. Indeed, potassium has recently been identified as a major driver of NCC activity (37, 38, 41, 42). Our results agree with findings of recent experimental and clinical studies showing a linear relationship between the serum potassium concentration and NCC expression in kidney and uEVs (41, 43). This effect of potassium appears to be specific for NCC, and it is therefore unclear whether the observed relation between serum potassium and pNKCC2 can be explained by similar mechanisms (37). Together, this proof-of-principle study provides insight in renal sodium handling during CS and the pathogenesis of hypertension in CS (summarized in Fig. 5). Previous studies on hypertension in CS also showed that some patients with CS have a suppressed RAAS, whereas others do not (44–48). In our cohort, patients with nonsuppressed RAAS had a higher BMI, which is associated with higher plasma renin and aldosterone levels (49, 50). In addition to effects on renal sodium reabsorption, hypercortisolism can also increase the peripheral vascular sensitivity to adrenergic agonists (51) and the production of angiotensinogen by the liver (47). The increased uEV abundances of pNKCC2, pNCC, and NCC in CS observed in this study merit further discussion. The increase of pNKCC2, pNCC, and NCC in uEVs of patients with CS and a suppressed RAAS suggests that increased sodium reabsorption through these transporters contributed to an increase in extracellular fluid volume (and therefore suppression of the RAAS). This may also in part explain the higher estimated glomerular filtration rate. In a new steady state, increased sodium reabsorption is not detectable in 24-hour urine (which reflects dietary intake); this likely explains the similar urinary sodium excretions in patients with or without suppressed RAAS (Table 1). In adrenalectomized rats, Velázquez et al. (10) showed that both aldosterone and dexamethasone increased NCC-mediated sodium reabsorption fivefold. More recently, Ivy et al. (52) showed that glucocorticoids cause nondipping of blood pressure via NCC activation. Glucocorticoid and mineralocorticoid receptors are expressed in the distal convoluted tubule, and activation of these receptors could therefore explain NCC activation (3). However, recent data indicate that NCC activation may occur indirectly via potassium. For example, Veiras et al. (53) showed that angiotensin II stimulates sodium-potassium exchange through ENaC and the renal outer medullary potassium channel (ROMK). This results in potassium deficiency and thereby NCC activation (53). Wolley et al. (43) confirmed this mechanism clinically in patients who were undergoing confirmatory testing for primary aldosteronism using the fludrocortisone suppression test. Similar to our data, they also showed that serum potassium was strongly and negatively correlated with pNCC and NCC in uEVs (43). Similarly, glucocorticoids may have activated ENaC and ROMK with increased potassium secretion and a decrease in serum potassium (11, 18, 54). Unfortunately, ENaC is difficult to analyze in uEVs (55). We used prostasin as a surrogate marker for ENaC but found no consistent effect of glucocorticoids. However, ENaC activation is not universally accompanied by an increase of prostasin in uEVs, as recently shown by Qi et al. (56). Finally, it is unclear how to explain the increase in pNKCC2. In adrenalectomized rats, Stanton et al. (57) did not observe that glucocorticoids increased sodium reabsorption in the thick ascending limb. Frindt and Palmer (11), however, did show an increase in NKCC2 in dexamethasone-treated rats but did not analyze pNKCC2. Of interest, when transporter activity would be expressed as the ratio between the phosphorylated and total protein abundance (i.e., pNKCC2/NKCC2 and pNCC/NCC), this would favor activation of NKCC2 rather than NCC. One would have expected more severe hypercortisolism in the patients with CS and suppressed RAAS. This might in part be explained by differences in 11β-hydroxysteroid dehydrogenase 2 (11β-HSD2) activity between both patient groups. Studies suggest a positive relation between 11β-HSD2 activity and BMI (58, 59). Patients with nonsuppressed RAAS in our study had a higher BMI than did those with suppressed RAAS, which may have modulated 11β-HSD2 activity next to possible other factors. We did not measure urinary steroid metabolites to assess 11β-HSD2 activity, but this could be a subject for future studies. The glucocorticoid activation of NCC is well described. Figure 5. View largeDownload slide Proposed model for the pathogenesis of hypertension in CS, including vascular and renal effects. The contribution of activation of renal sodium transport is partly based on the results in this study. ECF, extracellular fluid. Figure 5. View largeDownload slide Proposed model for the pathogenesis of hypertension in CS, including vascular and renal effects. The contribution of activation of renal sodium transport is partly based on the results in this study. ECF, extracellular fluid. A number of limitations of this study should be mentioned. First, because this was an observational study, we could not discontinue antihypertensive drugs and therefore had to exclude patients already receiving renin-angiotensin inhibitors or diuretics. This resulted in a relatively small number of patients [although similar to numbers in previous uEV studies (43, 56)]. The patient group was too small to analyze possible sex differences in sodium transporters, as was recently reported (60). Second, although several groups have shown that transporter analysis in uEVs correlates with their activity in the kidney (19, 20, 39, 40), this still remains an indirect measure of true transporter activity. We tried to address this by analyzing the phosphoproteins of the transporters, which are generally considered the active forms (21). In conclusion, our study provides insights in the pathogenesis of hypertension in CS showing that cortisol excess increases renal sodium transporter abundance in uEVs, especially in patients with suppressed RAAS. In addition to excess glucocorticoids, low serum potassium may also contribute to increased renal sodium reabsorption and hypertension in CS. These results may also be relevant for hypertension induced by exogenous glucocorticoids. Abbreviations: Abbreviations: 11β-HSD2 11β-hydroxysteroid dehydrogenase 2 BMI body mass index CS Cushing syndrome ENaC epithelial sodium channel NCC Na+-Cl− cotransporter NHE3 Na+/H+ exchanger type 3 NKCC2 Na+-K+-Cl− cotransporter type 2 pNCC phosphorylated Na+-Cl− cotransporter pNKCC2 phosphorylated Na+-K+-Cl− cotransporter type 2 RAAS renin-angiotensin-aldosterone system ROMK renal outer medullary potassium channel uEV urinary extracellular vesicle Acknowledgments We thank Drs. Fenton, Knepper, and Mutig for providing antibodies. We thank Usha Musterd-Bhaggoe for technical support. Financial Support: Dutch Kidney Foundation (KSP-14OK19 and CP16.01) (to E.J.H.). Disclosure Summary: R.A.F. received grant support from Novartis. The remaining authors have nothing to disclose. References 1. Feelders RA , Pulgar SJ , Kempel A , Pereira AM . The burden of Cushing’s disease: clinical and health-related quality of life aspects . Eur J Endocrinol . 2012 ; 167 ( 3 ): 311 – 326 . 2. Mancini T , Kola B , Mantero F , Boscaro M , Arnaldi G . High cardiovascular risk in patients with Cushing’s syndrome according to 1999 WHO/ISH guidelines . Clin Endocrinol (Oxf) . 2004 ; 61 ( 6 ): 768 – 777 . 3. Hunter RW , Ivy JR , Bailey MA . Glucocorticoids and renal Na+ transport: implications for hypertension and salt sensitivity . J Physiol . 2014 ; 592 ( 8 ): 1731 – 1744 . 4. Ferrari P . Cortisol and the renal handling of electrolytes: role in glucocorticoid-induced hypertension and bone disease . Best Pract Res Clin Endocrinol Metab . 2003 ; 17 ( 4 ): 575 – 589 . 5. Ong SL , Whitworth JA . How do glucocorticoids cause hypertension: role of nitric oxide deficiency, oxidative stress, and eicosanoids . Endocrinol Metab Clin North Am . 2011 ; 40 ( 2 ): 393 – 407, ix . 6. Smets P , Meyer E , Maddens B , Daminet S . Cushing’s syndrome, glucocorticoids and the kidney . Gen Comp Endocrinol . 2010 ; 169 ( 1 ): 1 – 10 . 7. Cicala MV , Mantero F . Hypertension in Cushing’s syndrome: from pathogenesis to treatment . Neuroendocrinology . 2010 ; 92 ( Suppl 1 ): 44 – 49 . 8. Whitworth JA , Mangos GJ , Kelly JJ . Cushing, cortisol, and cardiovascular disease . Hypertension . 2000 ; 36 ( 5 ): 912 – 916 . 9. Bobulescu IA , Dwarakanath V , Zou L , Zhang J , Baum M , Moe OW . Glucocorticoids acutely increase cell surface Na+/H+ exchanger-3 (NHE3) by activation of NHE3 exocytosis . Am J Physiol Renal Physiol . 2005 ; 289 ( 4 ): F685 – F691 . 10. Velázquez H , Bartiss A , Bernstein P , Ellison DH . Adrenal steroids stimulate thiazide-sensitive NaCl transport by rat renal distal tubules . Am J Physiol . 1996 ; 270 ( 1 Pt 2 ): F211 – F219 . 11. Frindt G , Palmer LG . Regulation of epithelial Na+ channels by adrenal steroids: mineralocorticoid and glucocorticoid effects . Am J Physiol Renal Physiol . 2012 ; 302 ( 1 ): F20 – F26 . 12. Li C , Wang W , Summer SN , Falk S , Schrier RW . Downregulation of UT-A1/UT-A3 is associated with urinary concentrating defect in glucocorticoid-excess state . J Am Soc Nephrol . 2008 ; 19 ( 10 ): 1975 – 1981 . 13. Lorenz JN , Loreaux EL , Dostanic-Larson I , Lasko V , Schnetzer JR , Paul RJ , Lingrel JB . ACTH-induced hypertension is dependent on the ouabain-binding site of the alpha2-Na+-K+-ATPase subunit . Am J Physiol Heart Circ Physiol . 2008 ; 295 ( 1 ): H273 – H280 . 14. Uhlén M , Fagerberg L , Hallström BM , Lindskog C , Oksvold P , Mardinoglu A , Sivertsson Å , Kampf C , Sjöstedt E , Asplund A , Olsson I , Edlund K , Lundberg E , Navani S , Szigyarto CA , Odeberg J , Djureinovic D , Takanen JO , Hober S , Alm T , Edqvist PH , Berling H , Tegel H , Mulder J , Rockberg J , Nilsson P , Schwenk JM , Hamsten M , von Feilitzen K , Forsberg M , Persson L , Johansson F , Zwahlen M , von Heijne G , Nielsen J , Pontén F . Proteomics. Tissue-based map of the human proteome . Science . 2015 ; 347 ( 6220 ): 1260419 . 15. Clore JN , Estep H , Ross-Clunis H , Watlington CO . Adrenocorticotropin and cortisol-induced changes in urinary sodium and potassium excretion in man: effects of spironolactone and RU486 . J Clin Endocrinol Metab . 1988 ; 67 ( 4 ): 824 – 831 . 16. Christy NP , Laragh JH . Pathogenesis of hypokalemic alkalosis in Cushing’s syndrome . N Engl J Med . 1961 ; 265 ( 22 ): 1083 – 1088 . 17. Montrella-Waybill M , Clore JN , Schoolwerth AC , Watlington CO . Evidence that high dose cortisol-induced Na+ retention in man is not mediated by the mineralocorticoid receptor . J Clin Endocrinol Metab . 1991 ; 72 ( 5 ): 1060 – 1066 . 18. Bailey MA , Mullins JJ , Kenyon CJ . Mineralocorticoid and glucocorticoid receptors stimulate epithelial sodium channel activity in a mouse model of Cushing syndrome . Hypertension . 2009 ; 54 ( 4 ): 890 – 896 . 19. Salih M , Fenton RA , Zietse R , Hoorn EJ . Urinary extracellular vesicles as markers to assess kidney sodium transport . Curr Opin Nephrol Hypertens . 2016 ; 25 ( 2 ): 67 – 72 . 20. van der Lubbe N , Jansen PM , Salih M , Fenton RA , van den Meiracker AH , Danser AH , Zietse R , Hoorn EJ . The phosphorylated sodium chloride cotransporter in urinary exosomes is superior to prostasin as a marker for aldosteronism . Hypertension . 2012 ; 60 ( 3 ): 741 – 748 . 21. Yang SS , Fang YW , Tseng MH , Chu PY , Yu IS , Wu HC , Lin SW , Chau T , Uchida S , Sasaki S , Lin YF , Sytwu HK , Lin SH . Phosphorylation regulates NCC stability and transporter activity in vivo . J Am Soc Nephrol . 2013 ; 24 ( 10 ): 1587 – 1597 . 22. Olivieri O , Castagna A , Guarini P , Chiecchi L , Sabaini G , Pizzolo F , Corrocher R , Righetti PG . Urinary prostasin: a candidate marker of epithelial sodium channel activation in humans . Hypertension . 2005 ; 46 ( 4 ): 683 – 688 . 23. Shibata S , Mu S , Kawarazaki H , Muraoka K , Ishizawa K , Yoshida S , Kawarazaki W , Takeuchi M , Ayuzawa N , Miyoshi J , Takai Y , Ishikawa A , Shimosawa T , Ando K , Nagase M , Fujita T . Rac1 GTPase in rodent kidneys is essential for salt-sensitive hypertension via a mineralocorticoid receptor-dependent pathway . J Clin Invest . 2011 ; 121 ( 8 ): 3233 – 3243 . 24. Tapia-Castillo A , Carvajal CA , Campino C , Hill C , Allende F , Vecchiola A , Carrasco C , Bancalari R , Valdivia C , Lagos C , Martinez-Aguayo A , Garcia H , Aglony M , Baudrand RF , Kalergis AM , Michea LF , Riedel CA , Fardella CE . The expression of RAC1 and mineralocorticoid pathway-dependent genes are associated with different responses to salt intake . Am J Hypertens . 2015 ; 28 ( 6 ): 722 – 728 . 25. Lacroix A , Feelders RA , Stratakis CA , Nieman LK . Cushing’s syndrome . Lancet . 2015 ; 386 ( 9996 ): 913 – 927 . 26. Mancia G , Fagard R , Narkiewicz K , Redon J , Zanchetti A , Böhm M , Christiaens T , Cifkova R , De Backer G , Dominiczak A , Galderisi M , Grobbee DE , Jaarsma T , Kirchhof P , Kjeldsen SE , Laurent S , Manolis AJ , Nilsson PM , Ruilope LM , Schmieder RE , Sirnes PA , Sleight P , Viigimaa M , Waeber B , Zannad F ; Task Force for the Management of Arterial Hypertension of the European Society of Hypertension and the European Society of Cardiology . 2013 ESH/ESC practice guidelines for the management of arterial hypertension . Blood Press . 2014 ; 23 ( 1 ): 3 – 16 . 27. Funder JW , Carey RM , Mantero F , Murad MH , Reincke M , Shibata H , Stowasser M , Young WF Jr . The management of primary aldosteronism: case detection, diagnosis, and treatment: an Endocrine Society clinical practice guideline . J Clin Endocrinol Metab . 2016 ; 101 ( 5 ): 1889 – 1916 . 28. Salih M , Fenton RA , Knipscheer J , Janssen JW , Vredenbregt-van den Berg MS , Jenster G , Zietse R , Hoorn EJ . An immunoassay for urinary extracellular vesicles . Am J Physiol Renal Physiol . 2016 ; 310 ( 8 ): F796 – F801 . 29. Fernández-Llama P , Khositseth S , Gonzales PA , Star RA , Pisitkun T , Knepper MA . Tamm-Horsfall protein and urinary exosome isolation . Kidney Int . 2010 ; 77 ( 8 ): 736 – 742 . 30. Zhou H , Yuen PS , Pisitkun T , Gonzales PA , Yasuda H , Dear JW , Gross P , Knepper MA , Star RA . Collection, storage, preservation, and normalization of human urinary exosomes for biomarker discovery . Kidney Int . 2006 ; 69 ( 8 ): 1471 – 1476 . 31. Dynia DW , Steinmetz AG , Kocinsky HS . NHE3 function and phosphorylation are regulated by a calyculin A-sensitive phosphatase . Am J Physiol Renal Physiol . 2010 ; 298 ( 3 ): F745 – F753 . 32. Tiwari S , Li L , Riazi S , Halagappa VK , Ecelbarger CM . Sex and age result in differential regulation of the renal thiazide-sensitive NaCl cotransporter and the epithelial sodium channel in angiotensin II-infused mice . Am J Nephrol . 2009 ; 30 ( 6 ): 554 – 562 . 33. Pedersen NB , Hofmeister MV , Rosenbaek LL , Nielsen J , Fenton RA . Vasopressin induces phosphorylation of the thiazide-sensitive sodium chloride cotransporter in the distal convoluted tubule . Kidney Int . 2010 ; 78 ( 2 ): 160 – 169 . 34. Mutig K , Paliege A , Kahl T , Jöns T , Müller-Esterl W , Bachmann S . Vasopressin V2 receptor expression along rat, mouse, and human renal epithelia with focus on TAL . Am J Physiol Renal Physiol . 2007 ; 293 ( 4 ): F1166 – F1177 . 35. Mira JP , Benard V , Groffen J , Sanders LC , Knaus UG . Endogenous, hyperactive Rac3 controls proliferation of breast cancer cells by a p21-activated kinase-dependent pathway . Proc Natl Acad Sci USA . 2000 ; 97 ( 1 ): 185 – 189 . 36. DiGiovanni SR , Nielsen S , Christensen EI , Knepper MA . Regulation of collecting duct water channel expression by vasopressin in Brattleboro rat . Proc Natl Acad Sci USA . 1994 ; 91 ( 19 ): 8984 – 8988 . 37. Terker AS , Zhang C , McCormick JA , Lazelle RA , Zhang C , Meermeier NP , Siler DA , Park HJ , Fu Y , Cohen DM , Weinstein AM , Wang WH , Yang CL , Ellison DH . Potassium modulates electrolyte balance and blood pressure through effects on distal cell voltage and chloride . Cell Metab . 2015 ; 21 ( 1 ): 39 – 50 . 38. Sorensen MV , Grossmann S , Roesinger M , Gresko N , Todkar AP , Barmettler G , Ziegler U , Odermatt A , Loffing-Cueni D , Loffing J . Rapid dephosphorylation of the renal sodium chloride cotransporter in response to oral potassium intake in mice . Kidney Int . 2013 ; 83 ( 5 ): 811 – 824 . 39. Gonzales PA , Pisitkun T , Hoffert JD , Tchapyjnikov D , Star RA , Kleta R , Wang NS , Knepper MA . Large-scale proteomics and phosphoproteomics of urinary exosomes . J Am Soc Nephrol . 2009 ; 20 ( 2 ): 363 – 379 . 40. Corbetta S , Raimondo F , Tedeschi S , Syrèn ML , Rebora P , Savoia A , Baldi L , Bettinelli A , Pitto M . Urinary exosomes in the diagnosis of Gitelman and Bartter syndromes . Nephrol Dial Transplant . 2015 ; 30 ( 4 ): 621 – 630 . 41. Terker AS , Zhang C , Erspamer KJ , Gamba G , Yang CL , Ellison DH . Unique chloride-sensing properties of WNK4 permit the distal nephron to modulate potassium homeostasis . Kidney Int . 2016 ; 89 ( 1 ): 127 – 134 . 42. Ishizawa K , Xu N , Loffing J , Lifton RP , Fujita T , Uchida S , Shibata S . Potassium depletion stimulates Na-Cl cotransporter via phosphorylation and inactivation of the ubiquitin ligase Kelch-like 3 . Biochem Biophys Res Commun . 2016 ; 480 ( 4 ): 745 – 751 . 43. Wolley MJ , Wu A , Xu S , Gordon RD , Fenton RA , Stowasser M . In primary aldosteronism, mineralocorticoids influence exosomal sodium-chloride cotransporter abundance . J Am Soc Nephrol . 2017 ; 28 ( 1 ): 56 – 63 . 44. Ganguly A , Weinberger MH , Grim CE . The renin-angiotensin-aldosterone system in Cushing’s syndrome and pheochromocytoma . Horm Res . 1983 ; 17 ( 1 ): 1 – 10 . 45. Mantero F , Armanini D , Boscaro M . Plasma renin activity and urinary aldosterone in Cushing’s syndrome . Horm Metab Res . 1978 ; 10 ( 1 ): 65 – 71 . 46. Saruta T , Suzuki H , Handa M , Igarashi Y , Kondo K , Senba S . Multiple factors contribute to the pathogenesis of hypertension in Cushing’s syndrome . J Clin Endocrinol Metab . 1986 ; 62 ( 2 ): 275 – 279 . 47. van der Pas R , van Esch JH , de Bruin C , Danser AH , Pereira AM , Zelissen PM , Netea-Maier R , Sprij-Mooij DM , van den Berg-Garrelds IM , van Schaik RH , Lamberts SW , van den Meiracker AH , Hofland LJ , Feelders RA . Cushing’s disease and hypertension: in vivo and in vitro study of the role of the renin-angiotensin-aldosterone system and effects of medical therapy . Eur J Endocrinol . 2014 ; 170 ( 2 ): 181 – 191 . 48. Yasuda G , Shionoiri H , Umemura S , Takasaki I , Ishii M . Exaggerated blood pressure response to angiotensin II in patients with Cushing’s syndrome due to adrenocortical adenoma . Eur J Endocrinol . 1994 ; 131 ( 6 ): 582 – 588 . 49. Tuck ML , Sowers J , Dornfeld L , Kledzik G , Maxwell M . The effect of weight reduction on blood pressure, plasma renin activity, and plasma aldosterone levels in obese patients . N Engl J Med . 1981 ; 304 ( 16 ): 930 – 933 . 50. Dudenbostel T , Ghazi L , Liu M , Li P , Oparil S , Calhoun DA . Body mass index predicts 24-hour urinary aldosterone levels in patients with resistant hypertension . hypertension . 2016 ; 68 ( 4 ): 995 – 1003 . 51. Pirpiris M , Sudhir K , Yeung S , Jennings G , Whitworth JA . Pressor responsiveness in corticosteroid-induced hypertension in humans . Hypertension . 1992 ; 19 ( 6 Pt 1 ): 567 – 574 . 52. Ivy JR , Oosthuyzen W , Peltz TS , Howarth AR , Hunter RW , Dhaun N , Al-Dujaili EA , Webb DJ , Dear JW , Flatman PW , Bailey MA . Glucocorticoids induce nondipping blood pressure by activating the thiazide-sensitive cotransporter . Hypertension . 2016 ; 67 ( 5 ): 1029 – 1037 . 53. Veiras LC , Han J , Ralph DL , McDonough AA . Potassium supplementation prevents sodium chloride cotransporter stimulation during angiotensin II hypertension . Hypertension . 2016 ; 68 ( 4 ): 904 – 912 . 54. Gallazzini M , Attmane-Elakeb A , Mount DB , Hebert SC , Bichara M . Regulation by glucocorticoids and osmolality of expression of ROMK (Kir 1.1), the apical K channel of thick ascending limb . Am J Physiol Renal Physiol . 2003 ; 284 ( 5 ): F977 – F986 . 55. Pisitkun T , Shen RF , Knepper MA . Identification and proteomic profiling of exosomes in human urine . Proc Natl Acad Sci USA . 2004 ; 101 ( 36 ): 13368 – 13373 . 56. Qi Y , Wang X , Rose KL , MacDonald WH , Zhang B , Schey KL , Luther JM . Activation of the endogenous renin-angiotensin-aldosterone system or aldosterone administration increases urinary exosomal sodium channel excretion . J Am Soc Nephrol . 2016 ; 27 ( 2 ): 646 – 656 . 57. Stanton B , Giebisch G , Klein-Robbenhaar G , Wade J , DeFronzo RA . Effects of adrenalectomy and chronic adrenal corticosteroid replacement on potassium transport in rat kidney . J Clin Invest . 1985 ; 75 ( 4 ): 1317 – 1326 . 58. Livingstone DE , Jones GC , Smith K , Jamieson PM , Andrew R , Kenyon CJ , Walker BR . Understanding the role of glucocorticoids in obesity: tissue-specific alterations of corticosterone metabolism in obese Zucker rats . Endocrinology . 2000 ; 141 ( 2 ): 560 – 563 . 59. Müssig K , Remer T , Haupt A , Gallwitz B , Fritsche A , Häring HU , Maser-Gluth C . 11beta-hydroxysteroid dehydrogenase 2 activity is elevated in severe obesity and negatively associated with insulin sensitivity . Obesity (Silver Spring) . 2008 ; 16 ( 6 ): 1256 – 1260 . 60. Veiras LC , Girardi ACC , Curry J , Pei L , Ralph DL , Tran A , Castelo-Branco RC , Pastor-Soler N , Arranz CT , Yu ASL , McDonough AA . Sexual dimorphic pattern of renal transporters and electrolyte homeostasis . J Am Soc Nephrol . 2017 ; 28 ( 12 ): 3504 – 3517 . Copyright © 2018 Endocrine Society

Begaye, Brittany;Piaggi, Paolo;Thearle, Marie S;Haskie, Kaitlyn;Walter, Mary;Schlögl, Mathias;Bonfiglio, Susan;Krakoff, Jonathan;Vinales, Karyne L

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00387pmid: 29788444

Abstract Context In healthy adults with detectable cold-induced brown adipose tissue activation (CIBA), the relationships between sympathetic nervous system (SNS) or thyroid activity during energy balance (EBL) with CIBA and body composition change are undetermined. Objective To investigate the relationships between CIBA and thermoneutral catecholamines and thyroid hormones measured during EBL and to determine if CIBA, catecholamines, or thyroid hormones predict body composition changes. Design, Setting, Participants, and Interventions Twelve healthy volunteers (seven male and five female) with positive CIBA [>2 standardized uptake value (g/mL)] had 24-hour energy expenditure (24hEE) assessed during EBL via whole-room indirect calorimetry while residing on a clinical research unit. Positron emission tomography/computed tomography scans were performed after exposure to 16°C for 2 hours to quantify CIBA. Main Outcome Measures CIBA, 24hEE during EBL, and thermoneutrality with concomitant measurement of urinary catecholamines and plasma free T3 and free T4. Body composition at baseline and 6 months by dual-energy X-ray absorptiometry. Results Lower urinary norepinephrine and free T4 were associated with higher CIBA (r = −0.65, P = 0.03; and r = −0.75, P < 0.01, respectively), but CIBA was not associated with 24hEE at thermoneutrality (P = 0.77). Lower CIBA (β = −3.5 kg/standardized uptake value; P < 0.01) predicted fat mass gain, whereas higher urinary norepinephrine and free T4 predicted future fat mass gain at 6 months (β = 3.0 kg per twofold difference in norepinephrine, P = 0.03; and β = 1.2 kg per 0.1-ng/dL difference in free T4, P = 0.03, respectively). Conclusion Lower SNS and free thyroid measurements at baseline indicate a greater capacity for CIBA, which may be predictive against fat mass gain. Obesity is the result of sustained energy intake exceeding energy expenditure (EE). Brown adipose tissue (BAT), a thermogenic tissue that dissipates energy (1), may be a contributor of daily EE and may play a role in regulating weight or body composition changes. The sympathetic nervous system (SNS) is a primary regulator of BAT function via promotion of brown preadipocyte proliferation and activation of beta-adrenergic receptors on mature brown adipocytes (1, 2). Sympathetic activation of BAT results in nonshivering thermogenesis in response to cold exposure, a powerful stimulus of BAT, in both rodents and humans (3–5). Cold-induced BAT activation (CIBA) can be quantified through positron emission tomography (PET) scans in adult humans and is associated with increased EE (6–8). The thermogenic properties of BAT make it an intriguing target for tackling obesity (9); however, the relationship of CIBA to change in weight or body composition in humans is not well understood. CIBA is inversely associated with measures of adiposity such as body mass index and body fat percentage (3, 4, 10). We have reported that CIBA, measured in standardized uptake value (SUV), is negatively associated with fat-free mass (FFM) and that subjects with CIBA (SUV ≥2 g/mL) who have lower activation of BAT are more likely to have an increase in fat mass (FM) after 6 months in free-living conditions (11). Further, the treatment of 6 weeks with cold exposure in adults with previously undetectable CIBA (<2 g/mL SUV) activates BAT and increases FM loss in those with the greatest increase in CIBA after therapy (12). In humans, SNS activity can be assessed by urine or plasma catecholamine measurements (13). Studies investigating the association between catecholamines and BAT have predominantly examined individuals diagnosed with pheochromocytoma, a tumor in the adrenal medulla that oversecretes catecholamines and causes BAT hypertrophy (14–16). In these studies, those with higher plasma catecholamine concentrations have higher BAT activity (17). However, among healthy individuals, there are no differences in plasma norepinephrine concentrations in those with detectable vs undetectable CIBA (18). Whether SNS activity measured during thermoneutrality and energy balance (EBL), as a static marker of SNS activity, is associated with CIBA has not been examined in healthy individuals. Free T3 and free T4 are reduced in individuals with positive BAT activity compared with BAT-negative individuals (19). A case report has shown the presence of substantial BAT volume in an adolescent with severe primary hypothyroidism, a low thyroid hormone state (20), that subsequently decreased in volume after 2 months of treatment with levothyroxine. However, this study used infrared thermal imaging and MRI techniques rather than gold-standard PET/CT to quantify BAT activity within the supraclavicular region. Conversely, in a hyperthyroid state, excess thyroid hormone concentrations increase BAT activity in rodents and humans (21, 22). These studies indicate that thyroid hormones are a potential regulator of BAT in humans, and additional investigation in healthy subjects with normal thyroid function is warranted. To further understand the role of the SNS and thyroid hormones in humans with cold-induced activation of BAT, we evaluated the relationships between CIBA and the SNS and thyroid function by assessing 24-hour urinary catecholamines and plasma thyroid hormones (free T3 and free T4) measured during EBL and thermoneutrality in healthy individuals with CIBA >2 SUV (11). In so doing, we aimed to establish a means to identify subjects who have greater potential for cold-induced BAT that might allow for detection of those who might benefit from interventions that act via BAT. We also examined whether CIBA, urinary catecholamines, or thyroid hormones predicted weight or body composition changes after 6 months in free-living conditions. Methods Participants Fifty-two volunteers between the ages of 18 and 50 years recruited from the Phoenix, AZ metropolitan area between 2009 and 2012, deemed healthy by medical history, physical examination, and laboratory measures, were admitted to the clinical research unit as part of a larger ongoing study (NCT00523627). Thirty-six volunteers were excluded based on PET/CT substudy criteria, which included an age <18 or >40 years, radiation exposure to the torso within the past 12 months, or declining participation. Of the 16 volunteers who participated in this substudy, 12 participants had measurements of CIBA and thyroid hormones, and 11 had urinary catecholamines analyzed. On admission, volunteers were placed on a weight-maintaining diet (WMD) consisting of 50% carbohydrates, 30% fat, and 20% protein. A 75-g oral glucose tolerance test was performed after 3 days on the WMD, and only volunteers with normal glucose regulation (23) continued the study. Body composition was measured by a dual-energy X-ray absorptiometry scan (DPX-1; Lunar Corp, Madison, WI). Volunteers were invited to return to our unit for follow-up at 6 months to remeasure weight and body composition. All participants provided written and informed consent prior to beginning the study. This study was approved by the Institutional Review Board of the National Institute of Diabetes and Digestive and Kidney Diseases. EE Measurements Following 4 days of a WMD, a whole-room indirect calorimeter was used to measure 24-hour EE (24hEE) by an initial eucaloric acclimation session followed by a second eucaloric session for the precise determination of 24hEE, as previously described (24), in thermoneutrality (mean ± SD; ambient temperature 23.9 ± 1.4°C). Participants were allowed to move freely inside the metabolic chamber, and this activity was measured using radar sensors, expressed as a percentage of time when activity was detected. As part of this study, participants also had 24hEE measured during fasting conditions (participants allowed only water and noncaloric beverages). Diet-induced thermogenesis (DIT) was calculated in two manners, the first as the difference between 24hEE during the eucaloric session and 24hEE during fasting and a second measurement of DIT as previously described (25). PET/CT Imaging Study participants underwent an 18F-fluorodeoxglucose (18F-FDG) PET/CT scan after a day on a WMD and following an overnight fast. Prior to the scan, volunteers were exposed to mild cold (16°C) for 2 hours while wearing standardized clothing of ∼0.3 clo. All possible measures were taken to avoid shivering, and if shivering occurred, the volunteer was removed from the cold room for 5 minutes and then returned into the cold room. The images were collected 1 hour after injection of 18F-FDG into the antecubital vein (mean dose 14.7 ± 0.3 mCi). PET and CT images by the Reveal 16 High Rez (CTI Molecular Imaging, Knoxville, TN) were taken from the diaphragm to the brain, and the images were coregistered with BAT quantified using a statistical parametric mapping software (SPM12; http://www.fil.ion.ucl.ac.uk/spm/software/spm12/) package in MATLAB (The Math Works, Inc.). CIBA was defined as the average SUV activity of the collection of voxels with an SUV of ≥2.0 in the PET image, coinciding with areas in the CT image with Hounsfield units between −250 and −10, in which the mean BAT SUV and volume were calculated. PET images were reconstructed into image voxels of 1.95 mm × 1.95 mm × 4.00 mm and CT image voxels of 0.98 mm × 0.98 mm. CIBA SUV (g/mL) was calculated as the activity of BAT (Becquerel) per milliliter of tissue divided by the dose (Becquerels) per gram of body weight. CIBA SUV was further adjusted for FFM by multiplying CIBA SUV to FFM/weight ratio (26). Catecholamines and Thyroid Hormones Urinary catecholamines were collected during the 24hEE measurement, whereas volunteers were in EBL and thermoneutrality as an index of SNS tone activity. Mayo Medical Laboratories (27) measured catecholamines using the HPLC method for epinephrine and norepinephrine and liquid chromatography-tandem mass spectrometry (stable isotope dilution analysis) for metanephrine and normetanephrine. Fasting plasma free T3 and free T4 were collected the morning after EBL and stored in a freezer at −70°C. The batched samples were measured using the EIA kit from Phoenix Pharmaceuticals (Burlingame, CA) by the National Institute of Diabetes and Digestive and Kidney Diseases Laboratory Core in Bethesda, MD. The intra-assay and interassay CV were 3.0% and 4.5% for free T3 and 2.8% and 4.0% free T4, respectively. Statistical Analyses Alpha was set at 0.05. Analyses were performed using SAS software (version 9.4; Cary, NC). Data are presented as mean ± SD, unless otherwise stated. Catecholamines values were log10-transformed to meet normal distribution requirements for parametric analysis. Differences between sexes were assessed using Student t test. Pearson correlations were used to assess relationships between continuous variables. Linear models with 24hEE as a dependent variable and FFM as independent variable were used to calculate the residual 24hEE. Similar linear models were used with 24-h mean respiratory quotient (RQ) as the dependent variable and body fat percentage as an independent variable. Due to the small sample size, serial individual partial adjustments were used to verify if associations between CIBA SUV and volume and catecholamine levels were independent of age, sex, body fat percentage, FM, or FFM. To assess weight and body composition changes, linear models were used with weight or FM change as dependent variable and CIBA SUV, catecholamines, or thyroid hormones as independent variables. Results CIBA Measurements Baseline and metabolic characteristics during EBL are shown in Table 1. Twelve volunteers had positive CIBA 18F-FDG PET/CT scans with a mean CIBA SUV of 3.25 ± 0.70 g/mL (range 2.35–4.57) and mean BAT volume of 120.17 ± 91.82 cm3 (range 3.00–289.00 cm3) with no differences by sex (P = 0.23 and P = 0.15, respectively). Table 1. Baseline and Metabolic Characteristics of the Cohort Variable Total (N = 12) Male (n = 7) Female (n = 5) Race 5 AA/2 H/3 NA/2 W 1 AA/2 H/2 NA/2 W 4 AA/1 NA Age, y 31.10 ± 9.78 (19.35, 50.64) 28.89 ± 11.10 (19.35, 50.64) 34.19 ± 7.60 (21.26, 39.55) Weight, kg 78.08 ± 14.22 (56.40, 103.50) 75.40 ± 14.91 (56.40, 98.10) 75.40 ± 14.91 (60.60, 103.50) BMI, kg/m2 26.03 ± 4.81(18.29, 34.43) 25.71 ± 5.07 (18.29, 33.41) 26.48 ± 2.22 (20.74, 34.43) FM, kg 22.35 ± 14.27(4.93, 52.78) 15.90 ± 11.29 (4.93, 33.02) 31.37 ± 13.95 (13.65, 52.78) FFM, kg 55.73 ± 12.45(42.75, 79.41) 64.81 ± 10.02 (49.45, 79.41)a 44.04 ± 1.31 (42.75, 45.55) Body fat, % 27.68 ± 14.82(6.90, 53.80) 18.87 ± 9.07 (6.90, 33.0)a 39.98 ± 10.51 (24.20, 53.80) Fasting glucose, mg/dL 89.50 ± 4.77 (80.00, 95.50) 88.50 ± 6.03 (80.00, 95.00) 90.70 ± 2.84 (80.00, 95.50) Two-hour glucose, mg/dL 99.27 ± 17.57 (69.00, 124.00) 101.20 ± 21.55 (69.00, 124.00) 97.00 ± 13.38 (80.00, 116.00) 24hEE, kcal/d 2020 ± 315 (1568, 2461) 2201 ± 277 (1788, 2461)a 1767 ± 141 (1568, 1918) Intake, kcal/d 2032 ± 330 (1529, 2575) 2230 ± 275 (1779, 2575)a 1756 ± 152 (1529, 1910) EBL, kcal/d 12 ± 61 (−59, 135) 29 ± 74 (−59, 135) −11 ± 30 (−42, 31) Twenty-four–hour RQ, ratio 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.03 (0.83, 0.91) CIBA SUV, g/mL 3.25 ± 0.70 (2.35, 4.57) 3.04 ± 0.56 (2.35, 3.80) 3.55 ± 0.80 (2.75, 4.57) BAT volume, cm3 120.17 ± 91.82 (3.00, 289.00) 27.14 ± 68.80 (3.00, 185.00) 166.40 ± 107.20 (52.00, 289.00) Norepinephrine, µg/24 hb 26.39 (20.25–36.42) 30.33 (25.44–36.42) 20.25 (14.39–28.95) Normetanephrine, µg/24 hb 230.56 (193.14–315.29) 247.56 (227.17–315.29) 227.74 (159.26–240.88) Epinephrine, µg/24 hb 5.88 (4.42–7.39) 6.07 (5.88–7.39) 4.42 (2.50–5.01) Metanephrine, µg/24 hb 129.94 (123.38–160.44) 138.28 (129.22–160.42 128.44 (63.43–167.34) Free T3, pg/mL 2.86 ± 0.92 (1.32, 4.92) 3.37 ± 0.77 (2.68, 4.92)a 2.16 ± 0.63 (1.32, 2.80) Free T4, ng/dL 1.21 ± 0.18 (0.92, 1.47) 1.31 ± 0.15 (1.08, 1.47)a 1.07 ± 0.14 (0.92, 1.24) Weight change, kgc 1.80 ± 4.81 (−5.20, 8.40) 3.37 ± 1.44 (−1.60, 7.30) −0.40 ± 5.62 (−5.20, 8.40) FM change, kgc 1.18 ± 3.27 (−4.45, 5.43) 2.76 ± 1.13 (−1.96, 5.43) −0.22 ± 3.56 (−4.45, 4.73) FFM change, kgc 0.16 ± 2.16 (−2.65, 3.93) −0.45 ± 2.18 (−1.94, 3.93) 2.33 ± 1.04 (−2.65, 3.67) Variable Total (N = 12) Male (n = 7) Female (n = 5) Race 5 AA/2 H/3 NA/2 W 1 AA/2 H/2 NA/2 W 4 AA/1 NA Age, y 31.10 ± 9.78 (19.35, 50.64) 28.89 ± 11.10 (19.35, 50.64) 34.19 ± 7.60 (21.26, 39.55) Weight, kg 78.08 ± 14.22 (56.40, 103.50) 75.40 ± 14.91 (56.40, 98.10) 75.40 ± 14.91 (60.60, 103.50) BMI, kg/m2 26.03 ± 4.81(18.29, 34.43) 25.71 ± 5.07 (18.29, 33.41) 26.48 ± 2.22 (20.74, 34.43) FM, kg 22.35 ± 14.27(4.93, 52.78) 15.90 ± 11.29 (4.93, 33.02) 31.37 ± 13.95 (13.65, 52.78) FFM, kg 55.73 ± 12.45(42.75, 79.41) 64.81 ± 10.02 (49.45, 79.41)a 44.04 ± 1.31 (42.75, 45.55) Body fat, % 27.68 ± 14.82(6.90, 53.80) 18.87 ± 9.07 (6.90, 33.0)a 39.98 ± 10.51 (24.20, 53.80) Fasting glucose, mg/dL 89.50 ± 4.77 (80.00, 95.50) 88.50 ± 6.03 (80.00, 95.00) 90.70 ± 2.84 (80.00, 95.50) Two-hour glucose, mg/dL 99.27 ± 17.57 (69.00, 124.00) 101.20 ± 21.55 (69.00, 124.00) 97.00 ± 13.38 (80.00, 116.00) 24hEE, kcal/d 2020 ± 315 (1568, 2461) 2201 ± 277 (1788, 2461)a 1767 ± 141 (1568, 1918) Intake, kcal/d 2032 ± 330 (1529, 2575) 2230 ± 275 (1779, 2575)a 1756 ± 152 (1529, 1910) EBL, kcal/d 12 ± 61 (−59, 135) 29 ± 74 (−59, 135) −11 ± 30 (−42, 31) Twenty-four–hour RQ, ratio 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.03 (0.83, 0.91) CIBA SUV, g/mL 3.25 ± 0.70 (2.35, 4.57) 3.04 ± 0.56 (2.35, 3.80) 3.55 ± 0.80 (2.75, 4.57) BAT volume, cm3 120.17 ± 91.82 (3.00, 289.00) 27.14 ± 68.80 (3.00, 185.00) 166.40 ± 107.20 (52.00, 289.00) Norepinephrine, µg/24 hb 26.39 (20.25–36.42) 30.33 (25.44–36.42) 20.25 (14.39–28.95) Normetanephrine, µg/24 hb 230.56 (193.14–315.29) 247.56 (227.17–315.29) 227.74 (159.26–240.88) Epinephrine, µg/24 hb 5.88 (4.42–7.39) 6.07 (5.88–7.39) 4.42 (2.50–5.01) Metanephrine, µg/24 hb 129.94 (123.38–160.44) 138.28 (129.22–160.42 128.44 (63.43–167.34) Free T3, pg/mL 2.86 ± 0.92 (1.32, 4.92) 3.37 ± 0.77 (2.68, 4.92)a 2.16 ± 0.63 (1.32, 2.80) Free T4, ng/dL 1.21 ± 0.18 (0.92, 1.47) 1.31 ± 0.15 (1.08, 1.47)a 1.07 ± 0.14 (0.92, 1.24) Weight change, kgc 1.80 ± 4.81 (−5.20, 8.40) 3.37 ± 1.44 (−1.60, 7.30) −0.40 ± 5.62 (−5.20, 8.40) FM change, kgc 1.18 ± 3.27 (−4.45, 5.43) 2.76 ± 1.13 (−1.96, 5.43) −0.22 ± 3.56 (−4.45, 4.73) FFM change, kgc 0.16 ± 2.16 (−2.65, 3.93) −0.45 ± 2.18 (−1.94, 3.93) 2.33 ± 1.04 (−2.65, 3.67) Data are mean ± SD (lowest and highest values), except for catecholamines, which are median (interquartile range). Abbreviations: AA, African American; BMI, body mass index; H, Hispanic; NA, Native American; W, white. a P < 0.05 vs females by Student t test. b n = 11 with urinary catecholamine measurements. c n = 11 with 6-month follow-up body composition measurements. View Large Table 1. Baseline and Metabolic Characteristics of the Cohort Variable Total (N = 12) Male (n = 7) Female (n = 5) Race 5 AA/2 H/3 NA/2 W 1 AA/2 H/2 NA/2 W 4 AA/1 NA Age, y 31.10 ± 9.78 (19.35, 50.64) 28.89 ± 11.10 (19.35, 50.64) 34.19 ± 7.60 (21.26, 39.55) Weight, kg 78.08 ± 14.22 (56.40, 103.50) 75.40 ± 14.91 (56.40, 98.10) 75.40 ± 14.91 (60.60, 103.50) BMI, kg/m2 26.03 ± 4.81(18.29, 34.43) 25.71 ± 5.07 (18.29, 33.41) 26.48 ± 2.22 (20.74, 34.43) FM, kg 22.35 ± 14.27(4.93, 52.78) 15.90 ± 11.29 (4.93, 33.02) 31.37 ± 13.95 (13.65, 52.78) FFM, kg 55.73 ± 12.45(42.75, 79.41) 64.81 ± 10.02 (49.45, 79.41)a 44.04 ± 1.31 (42.75, 45.55) Body fat, % 27.68 ± 14.82(6.90, 53.80) 18.87 ± 9.07 (6.90, 33.0)a 39.98 ± 10.51 (24.20, 53.80) Fasting glucose, mg/dL 89.50 ± 4.77 (80.00, 95.50) 88.50 ± 6.03 (80.00, 95.00) 90.70 ± 2.84 (80.00, 95.50) Two-hour glucose, mg/dL 99.27 ± 17.57 (69.00, 124.00) 101.20 ± 21.55 (69.00, 124.00) 97.00 ± 13.38 (80.00, 116.00) 24hEE, kcal/d 2020 ± 315 (1568, 2461) 2201 ± 277 (1788, 2461)a 1767 ± 141 (1568, 1918) Intake, kcal/d 2032 ± 330 (1529, 2575) 2230 ± 275 (1779, 2575)a 1756 ± 152 (1529, 1910) EBL, kcal/d 12 ± 61 (−59, 135) 29 ± 74 (−59, 135) −11 ± 30 (−42, 31) Twenty-four–hour RQ, ratio 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.03 (0.83, 0.91) CIBA SUV, g/mL 3.25 ± 0.70 (2.35, 4.57) 3.04 ± 0.56 (2.35, 3.80) 3.55 ± 0.80 (2.75, 4.57) BAT volume, cm3 120.17 ± 91.82 (3.00, 289.00) 27.14 ± 68.80 (3.00, 185.00) 166.40 ± 107.20 (52.00, 289.00) Norepinephrine, µg/24 hb 26.39 (20.25–36.42) 30.33 (25.44–36.42) 20.25 (14.39–28.95) Normetanephrine, µg/24 hb 230.56 (193.14–315.29) 247.56 (227.17–315.29) 227.74 (159.26–240.88) Epinephrine, µg/24 hb 5.88 (4.42–7.39) 6.07 (5.88–7.39) 4.42 (2.50–5.01) Metanephrine, µg/24 hb 129.94 (123.38–160.44) 138.28 (129.22–160.42 128.44 (63.43–167.34) Free T3, pg/mL 2.86 ± 0.92 (1.32, 4.92) 3.37 ± 0.77 (2.68, 4.92)a 2.16 ± 0.63 (1.32, 2.80) Free T4, ng/dL 1.21 ± 0.18 (0.92, 1.47) 1.31 ± 0.15 (1.08, 1.47)a 1.07 ± 0.14 (0.92, 1.24) Weight change, kgc 1.80 ± 4.81 (−5.20, 8.40) 3.37 ± 1.44 (−1.60, 7.30) −0.40 ± 5.62 (−5.20, 8.40) FM change, kgc 1.18 ± 3.27 (−4.45, 5.43) 2.76 ± 1.13 (−1.96, 5.43) −0.22 ± 3.56 (−4.45, 4.73) FFM change, kgc 0.16 ± 2.16 (−2.65, 3.93) −0.45 ± 2.18 (−1.94, 3.93) 2.33 ± 1.04 (−2.65, 3.67) Variable Total (N = 12) Male (n = 7) Female (n = 5) Race 5 AA/2 H/3 NA/2 W 1 AA/2 H/2 NA/2 W 4 AA/1 NA Age, y 31.10 ± 9.78 (19.35, 50.64) 28.89 ± 11.10 (19.35, 50.64) 34.19 ± 7.60 (21.26, 39.55) Weight, kg 78.08 ± 14.22 (56.40, 103.50) 75.40 ± 14.91 (56.40, 98.10) 75.40 ± 14.91 (60.60, 103.50) BMI, kg/m2 26.03 ± 4.81(18.29, 34.43) 25.71 ± 5.07 (18.29, 33.41) 26.48 ± 2.22 (20.74, 34.43) FM, kg 22.35 ± 14.27(4.93, 52.78) 15.90 ± 11.29 (4.93, 33.02) 31.37 ± 13.95 (13.65, 52.78) FFM, kg 55.73 ± 12.45(42.75, 79.41) 64.81 ± 10.02 (49.45, 79.41)a 44.04 ± 1.31 (42.75, 45.55) Body fat, % 27.68 ± 14.82(6.90, 53.80) 18.87 ± 9.07 (6.90, 33.0)a 39.98 ± 10.51 (24.20, 53.80) Fasting glucose, mg/dL 89.50 ± 4.77 (80.00, 95.50) 88.50 ± 6.03 (80.00, 95.00) 90.70 ± 2.84 (80.00, 95.50) Two-hour glucose, mg/dL 99.27 ± 17.57 (69.00, 124.00) 101.20 ± 21.55 (69.00, 124.00) 97.00 ± 13.38 (80.00, 116.00) 24hEE, kcal/d 2020 ± 315 (1568, 2461) 2201 ± 277 (1788, 2461)a 1767 ± 141 (1568, 1918) Intake, kcal/d 2032 ± 330 (1529, 2575) 2230 ± 275 (1779, 2575)a 1756 ± 152 (1529, 1910) EBL, kcal/d 12 ± 61 (−59, 135) 29 ± 74 (−59, 135) −11 ± 30 (−42, 31) Twenty-four–hour RQ, ratio 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.02 (0.83, 0.91) 0.87 ± 0.03 (0.83, 0.91) CIBA SUV, g/mL 3.25 ± 0.70 (2.35, 4.57) 3.04 ± 0.56 (2.35, 3.80) 3.55 ± 0.80 (2.75, 4.57) BAT volume, cm3 120.17 ± 91.82 (3.00, 289.00) 27.14 ± 68.80 (3.00, 185.00) 166.40 ± 107.20 (52.00, 289.00) Norepinephrine, µg/24 hb 26.39 (20.25–36.42) 30.33 (25.44–36.42) 20.25 (14.39–28.95) Normetanephrine, µg/24 hb 230.56 (193.14–315.29) 247.56 (227.17–315.29) 227.74 (159.26–240.88) Epinephrine, µg/24 hb 5.88 (4.42–7.39) 6.07 (5.88–7.39) 4.42 (2.50–5.01) Metanephrine, µg/24 hb 129.94 (123.38–160.44) 138.28 (129.22–160.42 128.44 (63.43–167.34) Free T3, pg/mL 2.86 ± 0.92 (1.32, 4.92) 3.37 ± 0.77 (2.68, 4.92)a 2.16 ± 0.63 (1.32, 2.80) Free T4, ng/dL 1.21 ± 0.18 (0.92, 1.47) 1.31 ± 0.15 (1.08, 1.47)a 1.07 ± 0.14 (0.92, 1.24) Weight change, kgc 1.80 ± 4.81 (−5.20, 8.40) 3.37 ± 1.44 (−1.60, 7.30) −0.40 ± 5.62 (−5.20, 8.40) FM change, kgc 1.18 ± 3.27 (−4.45, 5.43) 2.76 ± 1.13 (−1.96, 5.43) −0.22 ± 3.56 (−4.45, 4.73) FFM change, kgc 0.16 ± 2.16 (−2.65, 3.93) −0.45 ± 2.18 (−1.94, 3.93) 2.33 ± 1.04 (−2.65, 3.67) Data are mean ± SD (lowest and highest values), except for catecholamines, which are median (interquartile range). Abbreviations: AA, African American; BMI, body mass index; H, Hispanic; NA, Native American; W, white. a P < 0.05 vs females by Student t test. b n = 11 with urinary catecholamine measurements. c n = 11 with 6-month follow-up body composition measurements. View Large Relationship of CIBA to 24hEE and body composition As previously described (11), CIBA was negatively correlated with FFM (r = −0.66; P = 0.02), but not with other baseline body composition measurements. BAT volume was positively associated with body fat percentage and FM (r = 0.76, P < 0.01; and r = 0.74, P < 0.01, respectively; Fig. 1). BAT measurements were not associated with 24hEE measured during EBL after adjustment for FFM (CIBA: r = −0.08, P = 0.80; BAT volume: r = 0.19, P = 0.56) or 24-h mean RQ after adjustment for body fat percentage (CIBA: r = 0.25, P = 0.43; BAT volume: r = 0.29, P = 0.36). CIBA was also not associated with either measurement of DIT (both P > 0.21). Figure 1. View largeDownload slide Correlation of (A) BAT volume (g) and FM (kg), (B) correlation of CIBA mean SUV (g/mL) and FFM (kg), (C) 24hEE measured during EBL and thermoneutrality, after adjustment for FFM, and (D) 24-h mean RQ measured during EBL and thermoneutrality, after adjustment for body fat percentage. Figure 1. View largeDownload slide Correlation of (A) BAT volume (g) and FM (kg), (B) correlation of CIBA mean SUV (g/mL) and FFM (kg), (C) 24hEE measured during EBL and thermoneutrality, after adjustment for FFM, and (D) 24-h mean RQ measured during EBL and thermoneutrality, after adjustment for body fat percentage. Relationship of CIBA measurements to urinary catecholamine concentrations and thyroid hormone measurements Urinary norepinephrine was not associated with baseline characteristics including age, FFM, FM, body fat percentage, or 24hEE adjusted for FFM (all P > 0.05). CIBA mean SUV was negatively associated with 24-h urinary norepinephrine and normetanephrine concentrations (r = −0.64, P = 0.03; and r = −0.66, P = 0.03, respectively; Fig. 2). After adjustment for FFM, the relationship was attenuated, but the directionality remained (r = −0.58, P = 0.07; and r = −0.62, P = 0.05). BAT volume was associated with epinephrine and metanephrine concentrations (r = −0.70, P = 0.02; and r = −0.71, P = 0.02, respectively); however, upon removal of outliers, these results were no longer significant (P > 0.70). Figure 2. View largeDownload slide Correlation between (A) CIBA mean SUV (g/mL) and urinary norepinephrine (µg/24 h) measured during EBL and thermoneutrality and (B) urinary normetanephrine (µg/24 h) measured by HPLC during EBL and thermoneutrality. Relationship between free T4 (ng/dL) and both (C) CIBA and (D) BAT volume. Lack of correlation between plasma free T3 (pg/dL) and both (E) CIBA and (F) BAT volume. Both free T4 and free T3 were measured fasting the morning after EBL diet and during thermoneutrality. Results remained similar after serial adjustment for age, sex, FM, or FFM. Figure 2. View largeDownload slide Correlation between (A) CIBA mean SUV (g/mL) and urinary norepinephrine (µg/24 h) measured during EBL and thermoneutrality and (B) urinary normetanephrine (µg/24 h) measured by HPLC during EBL and thermoneutrality. Relationship between free T4 (ng/dL) and both (C) CIBA and (D) BAT volume. Lack of correlation between plasma free T3 (pg/dL) and both (E) CIBA and (F) BAT volume. Both free T4 and free T3 were measured fasting the morning after EBL diet and during thermoneutrality. Results remained similar after serial adjustment for age, sex, FM, or FFM. Free T3 was associated with FFM (r = 0.83; P < 0.01), but not with other baseline measurements, including age, weight, and body fat percentage (all P > 0.05). Free T4 was associated with FFM (r = 0.75; P < 0.01) and inversely with body fat percentage (r = −0.69, P = 0.01), but not with other baseline measurements including age (P = 0.56), weight (P = 0.73), and FM (P = 0.07). Both CIBA mean SUV and BAT volume were negatively associated with free T4 (r = −0.75, P = 0.005; and r = −0.73, P = 0.007, respectively), but not free T3 (P = 0.27 and P = 0.37, respectively), such that individuals with higher T4 had lower CIBA. After partial adjustment for FFM, CIBA and BAT volume were still associated with free T4 (r = −0.61, P = 0.05; and r = −0.67, P = 0.02). In a model with both free T4 and urinary norepinephrine, neither were independently associated with CIBA (data not shown), but these results should be interpreted with caution given the small sample size. Pictorial comparisons of the PET/CT scans of the individual with the highest CIBA and the lowest urinary norepinephrine and free T4 vs the individual with the lowest CIBA and the highest free T4 and urinary norepinephrine are shown in Fig. 3. Sensitivity analyses using nonweighted SUV scores were used, and results were similar (data not shown). Figure 3. View largeDownload slide Comparison of the 18F-FDG PET/CT images of the subjects with highest and lowest CIBA SUV values. Volunteer with the lowest urinary norepinephrine levels and lowest plasma free T4 images in the left panel, in which PET, CT, and fused coronal PET/CT images (A) show a large area of CIBA, with the highest SUV in our cohort of 4.57 SUV (g/mL).This series of images can be compared with the PET and CT and fused coronal PET/CT images (B) of the volunteer with the highest urinary norepinephrine and free T4 levels in the right panel, in which no visualization of CIBA is seen along with a mean SUV of 2.51 (g/mL), the lowest SUV in our cohort. Figure 3. View largeDownload slide Comparison of the 18F-FDG PET/CT images of the subjects with highest and lowest CIBA SUV values. Volunteer with the lowest urinary norepinephrine levels and lowest plasma free T4 images in the left panel, in which PET, CT, and fused coronal PET/CT images (A) show a large area of CIBA, with the highest SUV in our cohort of 4.57 SUV (g/mL).This series of images can be compared with the PET and CT and fused coronal PET/CT images (B) of the volunteer with the highest urinary norepinephrine and free T4 levels in the right panel, in which no visualization of CIBA is seen along with a mean SUV of 2.51 (g/mL), the lowest SUV in our cohort. Relationship of 6-month body composition changes to CIBA and urinary catecholamine concentrations Eleven out of the initial 12 volunteers (92%) returned for follow-up assessment at 6 months. The mean weight change observed at 6 months was 1.80 ± 4.81 kg (P = 0.22 vs zero), with a mean FM change of 1.18 ± 3.27 kg (P = 0.25) and a mean FFM change of 0.16 ± 2.16 kg (P = 0.81). As previously reported (11), CIBA was associated with FM change (kg) at 6 months [r = −0.77, β = −3.5 (95% CI −5.8 to −1.3) kg per 1 SUV; P = 0.006], but not with overall weight change (P = 0.15) or FFM change (P = 0.79). BAT volume was not associated with weight (P = 0.91), FFM (P = 0.09), or FM (P = 0.40) changes at 6 months. Baseline 24-hour urinary norepinephrine concentration (r = 0.69; P = 0.03; n = 10) and free T4 (r = 0.66; P = 0.03) measured during thermoneutrality were correlated with the change in FM at 6 months (Fig. 4), such that a twofold increase in norepinephrine predicted a 3.0-kg (0.4 to 5.5) increase in FM and 0.1-µg/dL increase in free T4 predicted a 1.18-kg (0.17 to 2.19) increase in FM at 6 months. Normetanephrine (P = 0.10), metanephrine (P = 0.88), epinephrine (P = 0.55), and free T3 (P = 0.26) concentrations were not associated with FM change at 6 months. Figure 4. View largeDownload slide Relationship between FM gain at 6 months and CIBA, urinary norepinephrine, free T4, and free T3. FM change was inversely associated with (A) CIBA and positively associated with both (B) 24-h urinary norepinephrine concentration and (C) free T4, whereas it was not associated with (D) free T3. Results by Pearson correlation coefficient. Figure 4. View largeDownload slide Relationship between FM gain at 6 months and CIBA, urinary norepinephrine, free T4, and free T3. FM change was inversely associated with (A) CIBA and positively associated with both (B) 24-h urinary norepinephrine concentration and (C) free T4, whereas it was not associated with (D) free T3. Results by Pearson correlation coefficient. Discussion In 12 volunteers with CIBA, CIBA was negatively associated with markers of SNS activity and thyroid function measured during EBL and thermoneutrality, but not with 24hEE. As we previously reported, CIBA was negatively associated with baseline FFM and FM change at 6 months (11) (e.g., those with the highest CIBA had the lowest baseline FFM and lost FM at 6 months follow-up). In these same individuals, urinary catecholamine concentrations were positively associated with FM change at 6 months. Our data indicate a link between lower norepinephrine concentration and CIBA, and both were predictors of FM gain. We hypothesize that a lower baseline sympathetic activity, as reflected by 24-hour urinary norepinephrine measures, indicates a greater capacity for both increasing sympathetic, and thus BAT, activation. This increased ability to activate CIBA then drives the protection against FM gain that we observed at 6 months. We suspect the observed associations among free T4, CIBA, and FM gain indicate similar physiology. Contrary to previous studies, we observed CIBA to be negatively associated with FFM and did not find CIBA to be associated with age or adiposity measurements (3, 4, 10), possibly due to the young age of our small cohort and its ethnic diversity (3, 12). Because FFM is the most substantial determinant of EE (28), our results indicate that CIBA may not be an important predictor of EE during EBL and thermoneutrality after controlling for this confounder. Although prior studies have shown a positive association between BAT and EE during cold exposure (4, 6), these studies used a hood system that allows for the measurement of EE during a brief period of cold exposure, as short as 20 minutes, and does not capture the physiologic conditions of meals and sleep. Our methodology using an initial acclimation chamber, measured during thermoneutrality, and the gold standard whole-room indirect calorimeter allowed for very precise measurement of 24hEE with the subject in EBL (EBL 12 ± 61 kcal/d; P = 0.51 vs zero) prior to cold exposure to determine whether the baseline conditions influence BAT activation upon cold exposure. Although we did not find CIBA to be a determinant of thermoneutral and EBL EE, prior studies have assessed the contributions of BAT to overall EE. BAT has been reported to increase EE by 20% to 30% (4, 5) and involved in DIT at thermoneutrality (25); however, other studies have found minimal contribution of activated BAT to overall EE (29–31). Therefore, further larger studies are warranted to definitely determine the contribution of BAT in potentially combating obesity (32). Studies assessing catecholamine concentrations in settings in which BAT is not purposefully activated in humans have primarily been done in patients with pheochromocytomas, a medullary adrenal tumor that oversecretes catecholamines. PET/CTs performed in these patients to stage the disease have shown widespread BAT activation (17, 33, 34). However, in a study with healthy subjects, plasma norepinephrine and its metabolite dihydroxyphenylglycol increased with cold exposure, but there were no differences in those catecholamines measured before and after cold stimulation in volunteers with positive vs negative CIBA (18). In contrast, using 24-hour urinary concentrations instead of plasma measures, we demonstrated a negative association between CIBA and urinary norepinephrine measured during EBL and thermoneutrality in healthy individuals. Our results may suggest that lower sympathetic tone during EBL and thermoneutrality signify an increased ability to both activate SNS to some maximal threshold and recruit BAT upon cold exposure. One potential mechanism may be via an undersaturation of available beta-adrenergic receptors in BAT in individuals with lower sympathetic tone. Thus, those with lower SNS tone during thermoneutrality might have a greater potential to activate BAT during cold, requiring a lesser increase in norepinephrine from the SNS to do so. In this cohort of individuals with positive CIBA, we also found a negative association between free T4 and CIBA such that individuals with higher BAT activity in response to cold had lower baseline free T4 measured during EBL and thermoneutrality. This association was in the same direction as that for urinary norepinephrine, which may be not surprising given the complex synergistic interactions of thyroid hormones and the adrenergic system (35). The correlation with CIBA was observed only with free T4 and not with free T3; this could possibly be explained due to the relatively higher secretion of T4 from the thyroid and the lesser free T3 production by deiodinases during states of thermoneutrality and EBL (36). In addition, free T3 conversion is tissue specific, and peripheral measurements may not be as reflective of the local free T3 action (37) that may be observed in BAT. Higher free T4 has been reported in BAT-negative subjects (19, 38) with normal thyroid function, in line with our cohort of positive CIBA individuals, in which those individuals with the lowest activation had the highest free T4. Therefore, in our cohort of BAT-positive subjects, our free T4 results may be due to our methodology in measuring thyroid function during thermoneutrality and EBL. The similar, yet independent, results between free T4 and urinary catecholamine concentrations with CIBA and future FM change seem to reflect a similar physiology. The fact that the subjects at 6 months follow-up did not gain weight is not surprising, because our subjects were weight stable prior to admission to our unit. However, the variation in the change in FM is intriguing, and identifying intrinsic factors that predispose to fat gain in a small period of time may provide insight for identification of those prone to longer-term weight gain. Given our previous finding that greater CIBA predicted less FM gain at 6 months, we tested if body composition changes were also associated with urinary catecholamines or thyroid function measures. We found that both urinary norepinephrine and plasma free T4 measured in thermoneutrality and during EBL were positively associated with FM change at 6 months. As we noted earlier, it is possible that greater ability to activate brown fat leads to more fat utilization and therefore less fat storage. Thus, a lower urinary norepinephrine and free T4 state elicits a greater BAT response, and this could drive FM change. How BAT might affect FM change is unclear, as we did not find an association with baseline EE. However, diets with different macronutrient composition, specifically those with high carbohydrates, which are known to stimulate SNS, may activate BAT (39). Because EE was not associated with CIBA in our study, it is also possible that CIBA might be evoking an effect on food intake. Our study has several limitations. First, we have relatively small sample size; however, this cohort was carefully studied in a controlled inpatient clinical research unit, and our sample size provided enough power to assess for relationships between CIBA and both weight change and hormone measures. However, a larger sample of subjects is needed to verify if these are a conditional effect or an independent effect on future FM gain. We also did not have catecholamines or thyroid function tests collected during or after cold exposure, but these thermoneutral measurements may offer evidence of who has the potential to activate BAT and thus may have clinical and research utility. Lastly, there was no assessment of EE or dietary intake while subjects were in free-living condition, and therefore, we cannot directly assess why some individuals gained FM. In conclusion, we found in individuals with CIBA, urinary norepinephrine concentrations and plasma free T4 measured during EBL and thermoneutrality were negatively associated with CIBA. Further, lower CIBA and higher baseline urinary norepinephrine and plasma free T4 concentrations predicted FM gain at 6 months. We propose that a higher SNS tone during EBL and thermoneutrality may signify a reduced ability to recruit BAT during cold exposure, thereby limiting any potential energy dissipating properties of BAT in these individuals. The use of catecholamines and thyroid hormones as a marker for potential BAT activation may help identify those who may derive benefit from the obesity opposing functions of BAT. Abbreviations: Abbreviations: 18F-FDG 18F-fluorodeoxyglucose 24hEE 24-hour energy expenditure BAT brown adipose tissue CIBA cold-induced brown adipose tissue activation DIT diet-induced thermogenesis EBL energy balance EE energy expenditure FFM fat-free mass FM fat mass PET positron emission tomography RQ respiratory quotient SNS sympathetic nervous system SUV standardized uptake value WMD weight-maintaining diet Acknowledgments The authors thank the nursing and dietary kitchen staff of the National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases Obesity and Diabetes Clinical Research Section in Phoenix, AZ for care of the participants. Most importantly, the authors thank the volunteers for participating in this study. Financial Support: This research was supported by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases. Clinical Trial Information: ClinicalTrials.gov no. NCT00523627 (registered 31 August 2007). Disclosure Summary: The authors have nothing to disclose. References 1. Cannon B , Nedergaard J . Brown adipose tissue: function and physiological significance . Physiol Rev . 2004 ; 84 ( 1 ): 277 – 359 . 2. Virtue S , Vidal-Puig A . Assessment of brown adipose tissue function . Front Physiol . 2013 ; 4 : 128 . 3. Cypess AM , Lehman S , Williams G , Tal I , Rodman D , Goldfine AB , Kuo FC , Palmer EL , Tseng YH , Doria A , Kolodny GM , Kahn CR . Identification and importance of brown adipose tissue in adult humans . N Engl J Med . 2009 ; 360 ( 15 ): 1509 – 1517 . 4. van Marken Lichtenbelt WD , Vanhommerig JW , Smulders NM , Drossaerts JM , Kemerink GJ , Bouvy ND , Schrauwen P , Teule GJ . Cold-activated brown adipose tissue in healthy men . N Engl J Med . 2009 ; 360 ( 15 ): 1500 – 1508 . 5. Virtanen KA , Lidell ME , Orava J , Heglind M , Westergren R , Niemi T , Taittonen M , Laine J , Savisto NJ , Enerbäck S , Nuutila P . Functional brown adipose tissue in healthy adults . N Engl J Med . 2009 ; 360 ( 15 ): 1518 – 1525 . 6. Yoneshiro T , Aita S , Matsushita M , Kameya T , Nakada K , Kawai Y , Saito M . Brown adipose tissue, whole-body energy expenditure, and thermogenesis in healthy adult men . Obesity (Silver Spring) . 2011 ; 19 ( 1 ): 13 – 16 . 7. Ouellet V , Labbé SM , Blondin DP , Phoenix S , Guérin B , Haman F , Turcotte EE , Richard D , Carpentier AC . Brown adipose tissue oxidative metabolism contributes to energy expenditure during acute cold exposure in humans . J Clin Invest . 2012 ; 122 ( 2 ): 545 – 552 . 8. van der Lans AA , Hoeks J , Brans B , Vijgen GH , Visser MG , Vosselman MJ , Hansen J , Jörgensen JA , Wu J , Mottaghy FM , Schrauwen P , van Marken Lichtenbelt WD . Cold acclimation recruits human brown fat and increases nonshivering thermogenesis . J Clin Invest . 2013 ; 123 ( 8 ): 3395 – 3403 . 9. Zafrir B . Brown adipose tissue: research milestones of a potential player in human energy balance and obesity . Horm Metab Res . 2013 ; 45 ( 11 ): 774 – 785 . 10. Saito M , Okamatsu-Ogura Y , Matsushita M , Watanabe K , Yoneshiro T , Nio-Kobayashi J , Iwanaga T , Miyagawa M , Kameya T , Nakada K , Kawai Y , Tsujisaki M . High incidence of metabolically active brown adipose tissue in healthy adult humans: effects of cold exposure and adiposity . Diabetes . 2009 ; 58 ( 7 ): 1526 – 1531 . 11. Schlögl M , Piaggi P , Thiyyagura P , Reiman EM , Chen K , Lutrin C , Krakoff J , Thearle MS . Overfeeding over 24 hours does not activate brown adipose tissue in humans . J Clin Endocrinol Metab . 2013 ; 98 ( 12 ): E1956 – E1960 . 12. Yoneshiro T , Aita S , Matsushita M , Kayahara T , Kameya T , Kawai Y , Iwanaga T , Saito M . Recruited brown adipose tissue as an antiobesity agent in humans . J Clin Invest . 2013 ; 123 ( 8 ): 3404 – 3408 . 13. Grassi G , Esler M . How to assess sympathetic activity in humans . J Hypertens . 1999 ; 17 ( 6 ): 719 – 734 . 14. Lean ME , James WP , Jennings G , Trayhurn P . Brown adipose tissue in patients with phaeochromocytoma . Int J Obes . 1986 ; 10 ( 3 ): 219 – 227 . 15. Melicow MM . One hundred cases of pheochromocytoma (107 tumors) at the Columbia-Presbyterian Medical Center, 1926-1976: a clinicopathological analysis . Cancer . 1977 ; 40 ( 5 ): 1987 – 2004 . 16. Ricquier D , Nechad M , Mory G . Ultrastructural and biochemical characterization of human brown adipose tissue in pheochromocytoma . J Clin Endocrinol Metab . 1982 ; 54 ( 4 ): 803 – 807 . 17. Wang Q , Zhang M , Ning G , Gu W , Su T , Xu M , Li B , Wang W . Brown adipose tissue in humans is activated by elevated plasma catecholamines levels and is inversely related to central obesity . PLoS One . 2011 ; 6 ( 6 ): e21006 . 18. Orava J , Nuutila P , Lidell ME , Oikonen V , Noponen T , Viljanen T , Scheinin M , Taittonen M , Niemi T , Enerbäck S , Virtanen KA . Different metabolic responses of human brown adipose tissue to activation by cold and insulin . Cell Metab . 2011 ; 14 ( 2 ): 272 – 279 . 19. Zhang Q , Miao Q , Ye H , Zhang Z , Zuo C , Hua F , Guan Y , Li Y . The effects of thyroid hormones on brown adipose tissue in humans: a PET-CT study . Diabetes Metab Res Rev . 2014 ; 30 ( 6 ): 513 – 520 . 20. Kim MS , Hu HH , Aggabao PC , Geffner ME , Gilsanz V . Presence of brown adipose tissue in an adolescent with severe primary hypothyroidism . J Clin Endocrinol Metab . 2014 ; 99 ( 9 ): E1686 – E1690 . 21. Lahesmaa M , Orava J , Schalin-Jäntti C , Soinio M , Hannukainen JC , Noponen T , Kirjavainen A , Iida H , Kudomi N , Enerbäck S , Virtanen KA , Nuutila P . Hyperthyroidism increases brown fat metabolism in humans . J Clin Endocrinol Metab . 2014 ; 99 ( 1 ): E28 – E35 . 22. Martínez-Sánchez N , Moreno-Navarrete JM , Contreras C , Rial-Pensado E , Fernø J , Nogueiras R , Diéguez C , Fernández-Real JM , López M . Thyroid hormones induce browning of white fat . J Endocrinol . 2017 ; 232 ( 2 ): 351 – 362 . 23. American Diabetes Association . Diagnosis and classification of diabetes mellitus . Diabetes Care . 2014 ; 37 ( Suppl 1 ): S81 – S90 . 24. Thearle MS , Pannacciulli N , Bonfiglio S , Pacak K , Krakoff J . Extent and determinants of thermogenic responses to 24 hours of fasting, energy balance, and five different overfeeding diets in humans . J Clin Endocrinol Metab . 2013 ; 98 ( 7 ): 2791 – 2799 . 25. Hibi M , Oishi S , Matsushita M , Yoneshiro T , Yamaguchi T , Usui C , Yasunaga K , Katsuragi Y , Kubota K , Tanaka S , Saito M . Brown adipose tissue is involved in diet-induced thermogenesis and whole-body fat utilization in healthy humans . Int J Obes . 2016 ; 40 ( 11 ): 1655 – 1661 . 26. Leitner BP , Huang S , Brychta RJ , Duckworth CJ , Baskin AS , McGehee S , Tal I , Dieckmann W , Gupta G , Kolodny GM , Pacak K , Herscovitch P , Cypess AM , Chen KY . Mapping of human brown adipose tissue in lean and obese young men . Proc Natl Acad Sci USA . 2017 ; 114 ( 32 ): 8649 – 8654 . 27. Moyer TP , Jiang NS , Tyce GM , Sheps SG . Analysis for urinary catecholamines by liquid chromatography with amperometric detection: methodology and clinical interpretation of results . Clin Chem . 1979 ; 25 ( 2 ): 256 – 263 . 28. Ravussin E , Lillioja S , Anderson TE , Christin L , Bogardus C . Determinants of 24-hour energy expenditure in man. Methods and results using a respira–tory chamber . J Clin Invest . 1986 ; 78 ( 6 ): 1568 – 1578 . 29. Muzik O , Mangner TJ , Leonard WR , Kumar A , Janisse J , Granneman JG . 15O PET measurement of blood flow and oxygen consumption in cold-activated human brown fat . J Nucl Med . 2013 ; 54 ( 4 ): 523 – 531 . 30. Muzik O , Mangner TJ , Leonard WR , Kumar A , Granneman JG . Sympathetic innervation of cold-activated brown and white fat in lean young adults . J Nucl Med . 2017 ; 58 ( 5 ): 799 – 806 . 31. Peterson CM , Orooji M , Johnson DN , Naraghi-Pour M , Ravussin E . Brown adipose tissue does not seem to mediate metabolic adaptation to overfeeding in men . Obesity (Silver Spring) . 2017 ; 25 ( 3 ): 502 – 505 . 32. Marlatt KL , Ravussin E . Brown adipose tissue: an update on recent findings . Curr Obes Rep . 2017 ; 6 ( 4 ): 389 – 396 . 33. Banzo J , Ubieto MA , Berisa MF , Andrés A , Mateo ML , Tardín L , Parra A , Razola P , Prats E . Extensive hypermetabolic pattern of brown adipose tissue activation on 18F-FDG PET/CT in a patient diagnosed of catecholamine-secreting para-vesical paraganglioma . Rev Esp Med Nucl Imagen Mol . 2013 ; 32 ( 6 ): 397 – 399 . 34. Yamaga LY , Thom AF , Wagner J , Baroni RH , Hidal JT , Funari MG . The effect of catecholamines on the glucose uptake in brown adipose tissue demonstrated by (18)F-FDG PET/CT in a patient with adrenal pheochromocytoma . Eur J Nucl Med Mol Imaging . 2008 ; 35 ( 2 ): 446 – 447 . 35. Silva JE , Bianco SD . Thyroid-adrenergic interactions: physiological and clinical implications . Thyroid . 2008 ; 18 ( 2 ): 157 – 165 . 36. Bianco AC , McAninch EA . The role of thyroid hormone and brown adipose tissue in energy homoeostasis . Lancet Diabetes Endocrinol . 2013 ; 1 ( 3 ): 250 – 258 . 37. Bianco AC , Kim BW . Deiodinases: implications of the local control of thyroid hormone action . J Clin Invest . 2006 ; 116 ( 10 ): 2571 – 2579 . 38. Lapa C , Maya Y , Wagner M , Arias-Loza P , Werner RA , Herrmann K , Higuchi T . Activation of brown adipose tissue in hypothyroidism . Ann Med . 2015 ; 47 ( 7 ): 538 – 545 . 39. Vosselman MJ , Brans B , van der Lans AA , Wierts R , van Baak MA , Mottaghy FM , Schrauwen P , van Marken Lichtenbelt WD . Brown adipose tissue activity after a high-calorie meal in humans . Am J Clin Nutr . 2013 ; 98 ( 1 ): 57 – 64 .

Rawal, Shristi;Tsai, Michael Y;Hinkle, Stefanie N;Zhu, Yeyi;Bao, Wei;Lin, Yuan;Panuganti, Pranati;Albert, Paul S;Ma, Ronald C W;Zhang, Cuilin

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02442pmid: 29889229

Abstract Context T3 is the biologically active thyroid hormone involved in glucose metabolism. The free T3 (fT3)/free T4 (fT4) ratio, a marker indicating conversion of fT4 to fT3, is also implicated in glucose homeostasis. Objective To examine associations of fT3 and the fT3/fT4 ratio with gestational diabetes mellitus (GDM). Design In a case-control study, thyroid markers (fT3, fT4, TSH) were measured and the fT3/fT4 ratio was derived across four visits in pregnancy, including first (gestational weeks 10 to 14) and second (weeks 15 to 26) trimester. Conditional logistic regression adjusting for thyroid autoimmunity status and major GDM risk factors estimated trimester-specific associations of thyroid markers with subsequent GDM risk. Setting Twelve US clinical centers. Participants One hundred seven GDM cases and 214 non-GDM controls from a multiracial pregnancy cohort of 2802 women. Main Outcome Measures GDM diagnosis ascertained from medical records. Results Both fT3 and the fT3/fT4 ratio were positively associated with GDM: adjusted OR (95% CI) comparing the highest vs lowest fT3 quartile was 4.25 (1.67, 10.80) at the first trimester and 3.89 (1.50, 10.10) at the second trimester. Similarly, the corresponding risk estimates for the fT3/fT4 ratio were 8.63 (2.87, 26.00) and 13.60 (3.97, 46.30) at the first and second trimester, respectively. Neither TSH nor fT4 was significantly associated with GDM. Conclusions Higher fT3 levels, potentially resulting from de novo synthesis or increased fT4 to fT3 conversion, may be an indicator of GDM risk starting early in pregnancy. Pregnancy has a considerable physiological impact on the thyroid gland and its metabolic function (1). To meet the increased demands during pregnancy, the thyroid gland increases up to 40% in size, accompanied by an upsurge in the production of thyroid hormones T4 and T3 (1). Abnormalities in thyroid function are relatively prevalent among pregnant women and have been linked to several obstetric complications, including premature delivery and pregnancy loss, as well as adverse health outcomes in the offspring (1). However, the debate regarding the utility of routine screening and/or treatment of thyroid dysfunction during pregnancy is highly contentious and remains to be resolved. Given the important role thyroid hormones play in glucose metabolism and homeostasis, thyroid dysfunction has been suggested to play a role in the etiology of gestational diabetes mellitus (GDM), a common metabolic complication in pregnancy (2). However, the existing evidence has been conflicting and longitudinal data are sparse. Whereas a few prospective studies (3–5) report increased incidence of GDM in women who have overt or subclinical hypothyroidism, others (6–8) report no significant differences. Similarly, several (6, 9, 10) but not all (7, 11, 12) prospective studies have found that isolated hypothyroxinemia [normal TSH, low free T4 (fT4)] in pregnancy is associated with increased risk of GDM. Of the two thyroid hormones T4 and T3, T4 is considered a prohormone, serving as a substrate for the biologically active form T3 (13). The conversion of peripheral T4 to T3, by two deiodinase enzymes, accounts for 80% of all the T3 produced; the rest is produced directly by the thyroid gland (13). T3 is also the primary active hormone involved in glucose metabolism, yet most prior studies have only looked at the associations between fT4 levels and GDM. Recently, two large prospective cohort studies (10, 14) observed an inverse association between fT4 levels and GDM, but speculated that low fT4 levels in GDM women may indicate increased conversion from fT4 to free T3 (fT3) or increased deiodinase activity, which is responsible for this conversion. A commonly used method for estimating the conversion from fT4 to fT3, and potentially serving as a proxy for deiodinase activity, is the fT3/fT4 ratio (15, 16). Although studies examining its association with GDM are lacking, several cross-sectional studies have noted that the fT3/fT4 ratio is associated with higher insulin resistance and glycosylated hemoglobin as well as elevated fasting glucose, fasting insulin, and postload glucose levels (17–19). Thus, a comprehensive analysis examining the subclinical changes in thyroid hormones fT4, fT3, and their ratio with GDM risk may offer novel insights into the pathogenesis of GDM. In the current study, we prospectively investigated the associations of the fT3/fT4 ratio and related markers of thyroid function (fT3, fT4, TSH) with GDM while accounting for thyroid autoimmunity status. Because thyroid levels can change with the progression of pregnancy, we assessed these associations separately for the first and second trimester. As a secondary objective, we also examined the longitudinal trajectory of the thyroid markers across the entire pregnancy. Materials and Methods Study design This case-control study was nested within the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Fetal Growth Studies—Singleton Cohort (2009–2013), a multiracial and multicenter pregnancy cohort consisting of 2334 nonobese (20) and 468 obese women. Eligible women, 18 to 40 years of age, were enrolled between 8 and 13 weeks of gestation from 12 US clinical centers and followed throughout pregnancy. Exclusion criteria included pre-existing hypertension, diabetes, renal/autoimmune diseases, psychiatric disorders, cancer, and HIV/AIDS. Additional major exclusion criteria applicable to nonobese women (enrolled for the primary aim of developing US fetal growth standards for four self-identified US racial/ethnic groups) were smoking before pregnancy and history of pregnancy complications (e.g., GDM, severe preeclampsia). Research approval for this study was granted by the institutional review boards of all participating sites, including the NICHD. Women provided written informed consent. GDM ascertainment GDM status (n = 107) was ascertained by review of medical records. For GDM diagnosis, we applied the Carpenter and Coustan diagnostic criteria (21) to the oral glucose tolerance test (OGTT) results. The average gestational age at OGTT among GDM cases (n = 95) was 27 weeks (range, 11–36 weeks). Women without OGTT results were classified as GDM if they had an indication of medication-treated GDM on the hospital discharge diagnosis (n = 12). Non-GDM status was confirmed based on 50-g 1-hour glucose challenge test, OGTT, and hospital discharge diagnosis. We matched each case to two non-GDM controls based on age (±2 years), race/ethnicity (non-Hispanic white, African American, Hispanic, Asian/Pacific Islander), and gestational week of blood collection (±2 weeks). Matching factors were selected either because they were well-established risk factors of GDM (age, race/ethnicity) or determinants of blood biomarker levels during pregnancy (gestational age). Exposure assessment Blood specimens were collected at four visits during the course of pregnancy, targeted at gestational weeks 8 to 13, 16 to 22, 24 to 29, and 34 to 37; however, the actual ranges were as follows: 10 to 14, 15 to 26, 24 to 31, and 33 to 39 weeks, respectively. Blood specimens collected at the second visit (weeks 15 to 26) were collected after an overnight fast of 8 to 14 hours. All blood specimens were stored at −80°C and thawed immediately before assay. Utilizing the electrochemiluminescence immunoassay method, concentrations of plasma TSH ( mIU/L), fT3 (pmol/L), fT4 (ng/dL), thyroglobulin antibody (IU/mL), and thyroid peroxidase antibody (IU/mL) were measured with Roche reagents (Roche Diagnostics, Indianapolis, IN) on the Roche Cobas e411 analyzer. The fT3/fT4 ratio was derived by dividing plasma concentrations of fT3 (pg/dL) by fT4 (ng/dL). Thyroid markers were measured at all four time points of blood collection among the GDM cases and one of the two matched controls. The remaining controls only had thyroid markers measured at the two visits prior to GDM diagnosis (i.e., weeks 10 to 14 and 15 to 26). All assays were conducted blinded to case-control status and performed in a central laboratory at the University of Minnesota. The interassay coefficients of variation were ≤5% for the thyroid hormones (fT3, fT4, TSH) and ≤15% for the two thyroid antibodies. The pregnancy-specific reference range for TSH, as recommended by the 2017 American Thyroid Association guidelines (1), was applied to the study sample. Women with TSH concentrations ≤ 4 mIU/L were considered to have normal TSH levels. Isolated hypothyroxinemia was defined as having normal TSH in conjunction with low fT4 levels (<10th percentile in controls) (22). Overt or subclinical hypothyroidism was defined as having elevated TSH levels (>4 mIU/L) with low or normal fT4 concentration (≤90th percentile in controls). Women with normal TSH status and normal fT4 concentration (10th to 90th percentile in controls) were classified as euthyroid. With respect to thyroid autoimmunity status, women were considered antibody-positive if the thyroid peroxidase antibody levels were >35 IU/mL or the thyroglobulin antibody levels were >40 IU/mL (6). Covariates A structured questionnaire administered at enrollment (8 to 13 weeks) collected information on demographics and common risk factors of GDM, including maternal age (years), race/ethnicity (non-Hispanic white, African American, Hispanic, Asian/Pacific Islander), family history of diabetes (yes/no), nulliparity (yes/no), education (less than, equal to, or more than high school), smoking in the 6 months prior to pregnancy (yes/no), and alcohol consumption in the 3 month before pregnancy (yes/no). Prepregnancy body mass index (BMI; <25, 25.0 to 29.9, ≥30.0 kg/m2) was calculated from self-reported prepregnancy weight and measured height. GDM treatment (diet/lifestyle modification and/or medication) history was extracted from medical records. Gestational age at each blood collection was estimated from the reported date of the last menstrual period, which was confirmed by ultrasound measurement at the time of enrollment. Women also reported any medication use, including medications for thyroid conditions. Statistical analysis In descriptive analyses, differences in participant characteristics between cases and controls were assessed by binomial/multinomial logistic regression with generalized estimating equations for categorical variables, and generalized linear mixed effects models for continuous variables including thyroid markers, both accounting for matched case-control pairs. Conditional logistic regression was used to estimate crude and adjusted ORs (aORs) of GDM for each thyroid marker accounting for the matched case-control pairs. The thyroid markers were analyzed continuously and as quartiles. The quartiles were based on the thyroid marker distributions among the controls. ORs were calculated separately for the two visits prior to GDM diagnosis (i.e., weeks 10 to 14 and 15 to 26). For the multivariable models, a priori selected covariates included key demographic factors and conventional risk factors for GDM: education level, parity, family history of diabetes, and prepregnancy BMI (<25, 25.0 to 29.9, ≥30.0 kg/m2). As maternal age and gestational age at blood collection were only matched between cases and controls within a certain range (±2 years and 2 weeks, respectively), we further adjusted for these two variables to reduce residual confounding and derive conservative estimates. Additionally, because thyroid antibodies may influence both thyroid hormone levels and glucose homeostasis, we also included thyroid autoimmunity status (antibodies positive vs negative) in the models. Of note, we did not include smoking as a covariate, as nonobese women who smoked prior to pregnancy were not eligible for the study and only five obese women in the study reported smoking before pregnancy. Tests of linear trend were performed by using the median value for each quartile as a continuous variable in the conditional logistic regression models. To ensure temporality, we excluded one case at weeks 10 to 14 and five cases at weeks 15 to 26 from our analytical population, as their blood samples were collected after the GDM diagnosis. GDM diagnosis was made at median 15.4 weeks after blood collection at 10 to 14 weeks, and 9.9 weeks after collection at 15 to 26 weeks. In addition to examining associations with thyroid marker levels, we also estimated ORs of GDM for clinical thyroid conditions including isolated hypothyroxinemia and overt or subclinical hypothyroidism, using women who had euthyroid status as the reference group. In sensitivity analyses, we excluded women who had elevated thyroid antibodies (n = 50 at weeks 10 to 14; n = 37 at weeks 15 to 26) (6), prior history of GDM (n = 6), prior history of preeclampsia (n = 8), smoked before the current pregnancy (n = 5), or had medication-treated GDM (n = 28) (to assess whether the associations still persisted among women with less severe GDM). Of note, none of the women reported use of thyroid medications prior to GDM diagnosis. In models looking at the quartile-specific associations between thyroid markers and GDM, we also repeated the analyses limiting the sample only to women who had euthyroid status. Furthermore, we stratified our analyses by prepregnancy BMI status (BMI < 25.0 vs BMI ≥ 25.0 kg/m2), race/ethnicity (non-Hispanic white, African American, Hispanic, Asian/Pacific Islander), or family history of diabetes (yes vs no). As a part of secondary analyses, the median concentration of each thyroid marker was plotted against the four study visits to depict changes in thyroid marker levels over the course of pregnancy. Statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC). Results Table 1 shows selected participant characteristics among women with and without GDM. Compared with non-GDM controls, GDM cases were more likely to have a higher prepregnancy BMI and a family history of diabetes. Median fT4 levels were significantly lower among GDM cases, whereas median fT3 and the fT3/fT4 ratio were significantly higher among cases at the two visits in the first (weeks 10 to 14) and second (weeks 15 to 26) trimester before GDM diagnosis (Table 2). TSH levels did not differ significantly between cases and controls at either trimester. Table 1. Participant Characteristics Among Women Who Had GDM and Their Matched Controls in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) GDM Cases (n = 107) Non-GDM Controls (n = 214) Pa Age, y 30.5 ± 5.7 30.4 ± 5.4 Race/ethnicity  Non-Hispanic white 25 (23.4) 50 (23.4)  African American 15 (14.0) 30 (14.0)  Hispanic 41 (38.3) 82 (38.3)  Asian/Pacific Islander 26 (24.3) 52 (24.3) Education 0.18  Less than high school 17 (15.9) 26 (12.1)  High school graduate or equivalent 15 (14.0) 23 (10.7)  More than high school 75 (70.1) 165 (77.1) Married/living with a partner 92 (86.0) 167 (78.0) 0.12 Nulliparity 48 (44.9) 96 (44.9) 1 Family history of diabetes 40 (37.4) 48 (22.4) 0.003 Smoked before pregnancyb 4 (3.7) 1 (0.5) 0.06 Alcoholic beverage consumption before pregnancy 61 (57.0) 137 (64.0) 0.22 Prepregnancy BMI, kg/m2 <0.001  18.38–24.99 37 (34.6) 123 (57.5)  25.0–29.99 35 (32.7) 56 (26.2)  30.0–45.11 35 (32.7) 33 (15.4) GDM Cases (n = 107) Non-GDM Controls (n = 214) Pa Age, y 30.5 ± 5.7 30.4 ± 5.4 Race/ethnicity  Non-Hispanic white 25 (23.4) 50 (23.4)  African American 15 (14.0) 30 (14.0)  Hispanic 41 (38.3) 82 (38.3)  Asian/Pacific Islander 26 (24.3) 52 (24.3) Education 0.18  Less than high school 17 (15.9) 26 (12.1)  High school graduate or equivalent 15 (14.0) 23 (10.7)  More than high school 75 (70.1) 165 (77.1) Married/living with a partner 92 (86.0) 167 (78.0) 0.12 Nulliparity 48 (44.9) 96 (44.9) 1 Family history of diabetes 40 (37.4) 48 (22.4) 0.003 Smoked before pregnancyb 4 (3.7) 1 (0.5) 0.06 Alcoholic beverage consumption before pregnancy 61 (57.0) 137 (64.0) 0.22 Prepregnancy BMI, kg/m2 <0.001  18.38–24.99 37 (34.6) 123 (57.5)  25.0–29.99 35 (32.7) 56 (26.2)  30.0–45.11 35 (32.7) 33 (15.4) Data are presented as n (%) for categorical variables and mean (SD) for continuous variables. a P values for differences between case and control subjects were obtained by generalized linear mixed effects models for continuous variables and binomial/multinomial logistic regression with generalized estimating equations for binary/multilevel categorical variables, accounting for matched case-control pairs. P values are not shown for matching variables (age, race/ethnicity). b Nonobese women who smoked were not eligible for the study. View Large Table 1. Participant Characteristics Among Women Who Had GDM and Their Matched Controls in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) GDM Cases (n = 107) Non-GDM Controls (n = 214) Pa Age, y 30.5 ± 5.7 30.4 ± 5.4 Race/ethnicity  Non-Hispanic white 25 (23.4) 50 (23.4)  African American 15 (14.0) 30 (14.0)  Hispanic 41 (38.3) 82 (38.3)  Asian/Pacific Islander 26 (24.3) 52 (24.3) Education 0.18  Less than high school 17 (15.9) 26 (12.1)  High school graduate or equivalent 15 (14.0) 23 (10.7)  More than high school 75 (70.1) 165 (77.1) Married/living with a partner 92 (86.0) 167 (78.0) 0.12 Nulliparity 48 (44.9) 96 (44.9) 1 Family history of diabetes 40 (37.4) 48 (22.4) 0.003 Smoked before pregnancyb 4 (3.7) 1 (0.5) 0.06 Alcoholic beverage consumption before pregnancy 61 (57.0) 137 (64.0) 0.22 Prepregnancy BMI, kg/m2 <0.001  18.38–24.99 37 (34.6) 123 (57.5)  25.0–29.99 35 (32.7) 56 (26.2)  30.0–45.11 35 (32.7) 33 (15.4) GDM Cases (n = 107) Non-GDM Controls (n = 214) Pa Age, y 30.5 ± 5.7 30.4 ± 5.4 Race/ethnicity  Non-Hispanic white 25 (23.4) 50 (23.4)  African American 15 (14.0) 30 (14.0)  Hispanic 41 (38.3) 82 (38.3)  Asian/Pacific Islander 26 (24.3) 52 (24.3) Education 0.18  Less than high school 17 (15.9) 26 (12.1)  High school graduate or equivalent 15 (14.0) 23 (10.7)  More than high school 75 (70.1) 165 (77.1) Married/living with a partner 92 (86.0) 167 (78.0) 0.12 Nulliparity 48 (44.9) 96 (44.9) 1 Family history of diabetes 40 (37.4) 48 (22.4) 0.003 Smoked before pregnancyb 4 (3.7) 1 (0.5) 0.06 Alcoholic beverage consumption before pregnancy 61 (57.0) 137 (64.0) 0.22 Prepregnancy BMI, kg/m2 <0.001  18.38–24.99 37 (34.6) 123 (57.5)  25.0–29.99 35 (32.7) 56 (26.2)  30.0–45.11 35 (32.7) 33 (15.4) Data are presented as n (%) for categorical variables and mean (SD) for continuous variables. a P values for differences between case and control subjects were obtained by generalized linear mixed effects models for continuous variables and binomial/multinomial logistic regression with generalized estimating equations for binary/multilevel categorical variables, accounting for matched case-control pairs. P values are not shown for matching variables (age, race/ethnicity). b Nonobese women who smoked were not eligible for the study. View Large Table 2. Median Plasma Concentrations of Thyroid Markers Among Women With GDM and Their Matched Controls in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) GDM Cases Non-GDM Controls Pa Weeks 10–14b  fT3, pg/dL 300.00 (276.62, 327.27) 279.87 (263.96, 304.22) 0.002  fT4, ng/dL 1.02 (0.94, 1.13) 1.06 (0.98, 1.17) 0.04  fT3/fT4 ratioc 289.60 (263.13, 322.56) 259.74 (238.94, 288.75) 0.0001  TSH, mIU/L 1.25 (0.73, 1.83) 1.28 (0.86,1.82) 0.65  Gestational age at blood collection, wk 13.0 (12.3, 13.6) 13.0 (12.4, 13.6) Weeks 15–26d  fT3, pg/dL 292.85 (265.26, 325.00) 275.32 (255.19, 299.67) 0.0003  fT4, ng/dL 0.88 (0.81, 1.00) 0.94 (0.86, 1.03) 0.001  fT3/fT4 ratioc 334.51 (285.16, 375.76) 289.55 (257.69, 332.58) <0.0001  TSH, mIU/L 1.96 (1.26, 2.52) 1.91 (1.26, 2.76) 0.59  Gestational age at blood collection, wk 18.9 (17.3, 21.0) 19.0 (17.7, 21.1) Change from weeks 10–14 to 15–26  fT3, pg/dL 0.14 (0.08, 0.20) 0.12 (0.07, 0.20) 0.69  fT4, ng/dL 8.44 (−9.74, 24.03) 5.19 (−13.00, 24.67) 0.44  fT3/fT4 ratioc −37.87 (−63.27, −20.6) −30.22 (−47.43, −15.4) 0.0002  TSH, mIU/L −0.54 (−1.09, −0.20) −0.61 (−1.14, −0.21) 0.67 GDM Cases Non-GDM Controls Pa Weeks 10–14b  fT3, pg/dL 300.00 (276.62, 327.27) 279.87 (263.96, 304.22) 0.002  fT4, ng/dL 1.02 (0.94, 1.13) 1.06 (0.98, 1.17) 0.04  fT3/fT4 ratioc 289.60 (263.13, 322.56) 259.74 (238.94, 288.75) 0.0001  TSH, mIU/L 1.25 (0.73, 1.83) 1.28 (0.86,1.82) 0.65  Gestational age at blood collection, wk 13.0 (12.3, 13.6) 13.0 (12.4, 13.6) Weeks 15–26d  fT3, pg/dL 292.85 (265.26, 325.00) 275.32 (255.19, 299.67) 0.0003  fT4, ng/dL 0.88 (0.81, 1.00) 0.94 (0.86, 1.03) 0.001  fT3/fT4 ratioc 334.51 (285.16, 375.76) 289.55 (257.69, 332.58) <0.0001  TSH, mIU/L 1.96 (1.26, 2.52) 1.91 (1.26, 2.76) 0.59  Gestational age at blood collection, wk 18.9 (17.3, 21.0) 19.0 (17.7, 21.1) Change from weeks 10–14 to 15–26  fT3, pg/dL 0.14 (0.08, 0.20) 0.12 (0.07, 0.20) 0.69  fT4, ng/dL 8.44 (−9.74, 24.03) 5.19 (−13.00, 24.67) 0.44  fT3/fT4 ratioc −37.87 (−63.27, −20.6) −30.22 (−47.43, −15.4) 0.0002  TSH, mIU/L −0.54 (−1.09, −0.20) −0.61 (−1.14, −0.21) 0.67 Data are presented as median (25th and 75th percentile). Boldface indicates statistically significant results. a P values for differences between case and controls were obtained by generalized linear mixed effects models for continuous variables accounting for matched case-control pairs. P values are not shown for gestational age at blood collection, as it was one of the matching variables. b n = 104 and 214 for cases and controls, respectively. c The fT3/fT4 ratio was obtained by dividing plasma concentration of fT3 (pg/dL) by fT4 level (ng/dL). d n = 94 and 212 for cases and controls, respectively. View Large Table 2. Median Plasma Concentrations of Thyroid Markers Among Women With GDM and Their Matched Controls in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) GDM Cases Non-GDM Controls Pa Weeks 10–14b  fT3, pg/dL 300.00 (276.62, 327.27) 279.87 (263.96, 304.22) 0.002  fT4, ng/dL 1.02 (0.94, 1.13) 1.06 (0.98, 1.17) 0.04  fT3/fT4 ratioc 289.60 (263.13, 322.56) 259.74 (238.94, 288.75) 0.0001  TSH, mIU/L 1.25 (0.73, 1.83) 1.28 (0.86,1.82) 0.65  Gestational age at blood collection, wk 13.0 (12.3, 13.6) 13.0 (12.4, 13.6) Weeks 15–26d  fT3, pg/dL 292.85 (265.26, 325.00) 275.32 (255.19, 299.67) 0.0003  fT4, ng/dL 0.88 (0.81, 1.00) 0.94 (0.86, 1.03) 0.001  fT3/fT4 ratioc 334.51 (285.16, 375.76) 289.55 (257.69, 332.58) <0.0001  TSH, mIU/L 1.96 (1.26, 2.52) 1.91 (1.26, 2.76) 0.59  Gestational age at blood collection, wk 18.9 (17.3, 21.0) 19.0 (17.7, 21.1) Change from weeks 10–14 to 15–26  fT3, pg/dL 0.14 (0.08, 0.20) 0.12 (0.07, 0.20) 0.69  fT4, ng/dL 8.44 (−9.74, 24.03) 5.19 (−13.00, 24.67) 0.44  fT3/fT4 ratioc −37.87 (−63.27, −20.6) −30.22 (−47.43, −15.4) 0.0002  TSH, mIU/L −0.54 (−1.09, −0.20) −0.61 (−1.14, −0.21) 0.67 GDM Cases Non-GDM Controls Pa Weeks 10–14b  fT3, pg/dL 300.00 (276.62, 327.27) 279.87 (263.96, 304.22) 0.002  fT4, ng/dL 1.02 (0.94, 1.13) 1.06 (0.98, 1.17) 0.04  fT3/fT4 ratioc 289.60 (263.13, 322.56) 259.74 (238.94, 288.75) 0.0001  TSH, mIU/L 1.25 (0.73, 1.83) 1.28 (0.86,1.82) 0.65  Gestational age at blood collection, wk 13.0 (12.3, 13.6) 13.0 (12.4, 13.6) Weeks 15–26d  fT3, pg/dL 292.85 (265.26, 325.00) 275.32 (255.19, 299.67) 0.0003  fT4, ng/dL 0.88 (0.81, 1.00) 0.94 (0.86, 1.03) 0.001  fT3/fT4 ratioc 334.51 (285.16, 375.76) 289.55 (257.69, 332.58) <0.0001  TSH, mIU/L 1.96 (1.26, 2.52) 1.91 (1.26, 2.76) 0.59  Gestational age at blood collection, wk 18.9 (17.3, 21.0) 19.0 (17.7, 21.1) Change from weeks 10–14 to 15–26  fT3, pg/dL 0.14 (0.08, 0.20) 0.12 (0.07, 0.20) 0.69  fT4, ng/dL 8.44 (−9.74, 24.03) 5.19 (−13.00, 24.67) 0.44  fT3/fT4 ratioc −37.87 (−63.27, −20.6) −30.22 (−47.43, −15.4) 0.0002  TSH, mIU/L −0.54 (−1.09, −0.20) −0.61 (−1.14, −0.21) 0.67 Data are presented as median (25th and 75th percentile). Boldface indicates statistically significant results. a P values for differences between case and controls were obtained by generalized linear mixed effects models for continuous variables accounting for matched case-control pairs. P values are not shown for gestational age at blood collection, as it was one of the matching variables. b n = 104 and 214 for cases and controls, respectively. c The fT3/fT4 ratio was obtained by dividing plasma concentration of fT3 (pg/dL) by fT4 level (ng/dL). d n = 94 and 212 for cases and controls, respectively. View Large Table 3 shows the associations between thyroid markers in the first and second trimester and GDM status. fT3 and the fT3/fT4 ratio were significantly and positively associated with GDM risk at both trimesters: the aOR (95% CI) comparing the highest vs lowest quartile of fT3 was 4.25 (1.67, 10.80) at the first (Ptrend = 0.001) and 3.89 (1.50, 10.10) at the second (Ptrend = 0.007) trimester. Similarly, the corresponding risk estimates comparing the highest vs lowest quartile of the fT3/fT4 ratio were 8.63 (2.87, 26.00) and 13.60 (3.97, 46.30) at the first (Ptrend = 0.001) and second (Ptrend < 0.0001) trimester, respectively. fT4 levels were inversely associated with GDM risk at the second trimester, yet the quartile-specific associations were not significant after adjusting for potential confounders. However, women who were in the top decile of fT4 levels at the second trimester had a significantly decreased risk of GDM compared with those who were in the lowest quartile [aOR (95% CI), 0.17 (0.04, 0.76)]. TSH levels were not associated with GDM risk in either trimester. Table 3. aOR (95% CI) for GDM According to Quartiles of Thyroid Markers at Gestational Weeks 10–14 and 15–26 in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  fT3, pg/dL   Quartile 1: 1.18–4.06 13 53 1 1   Quartile 2: 4.07–4.31 18 54 1.31 (0.57, 3.00) 1.32 (0.53, 3.28)   Quartile 3: 4.32–4.68 25 52 2.05 (0.91, 4.60) 1.97 (0.78, 4.96)   Quartile 4: 4.69–7.66 43 53 4.07 (1.82, 9.11) 4.25 (1.67, 10.80)   Upper decile: 5.02–7.66 29 21 6.09 (2.47, 15.00) 6.08 (2.19, 16.87)   P for trend <0.0001 0.001   Per unit increment 1.02 (1.01, 1.03) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.70–0.98 37 55 1 1   Quartile 2: 0.99–1.06 24 52 0.69 (0.37, 1.31) 0.70 (0.34, 1.45)   Quartile 3: 1.07–1.17 23 59 0.61 (0.32, 1.16) 0.69 (0.33, 1.42)   Quartile 4: 1.18–2.26 19 48 0.58 (0.29, 1.19) 0.63 (0.26, 1.51)   Upper decile: 1.29–2.26 7 19 0.61 (0.22, 1.66) 0.83 (0.27, 2.56)   P for trend 0.11 0.30   Per unit increment 0.23 (0.04, 1.22) 0.54 (0.08, 3.75)  fT3/fT4 ratioc   Quartile 1: 0.83–3.67 9 53 1 1   Quartile 2: 3.68–4.00 12 54 1.43 (0.51, 4.02) 1.12 (0.34, 3.74)   Quartile 3: 4.01–4.44 26 52 4.03 (1.56, 10.45) 5.26 (1.70, 16.20)   Quartile 4: 4.45–6.85 51 53 8.03 (3.13, 20.61) 8.63 (2.87, 26.00)   Upper decile: 5.39–6.85 25 22 10.25 (3.62, 29.01) 9.28 (2.76, 31.26)   P for trend <0.0001 0.001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.06–0.86 29 53 1 1   Quartile 2: 0.87–1.28 22 53 0.77 (0.38, 1.56) 0.71 (0.32, 1.59)   Quartile 3: 1.29–1.82 23 51 0.85 (0.43, 1.69) 0.63 (0.28, 1.44)   Quartile 4: 1.83–30.11 25 52 0.88 (0.46, 1.71) 1.17 (0.54, 2.51)   Upper decile: 2.53–30.11 12 20 1.11 (0.48, 2.57) 1.43 (0.53, 3.84)   P for trend 0.83 0.78   Per unit increment 1.08 (0.94, 1.24) 1.11 (0.93, 1.32) Gestational weeks 15-26b  fT3, pg/dL   Quartile 1: 2.95–3.93 14 57 1 1   Quartile 2: 3.94–4.24 19 50 1.78 (0.77, 4.13) 1.25 (0.46, 3.37)   Quartile 3: 4.25–4.61 16 52 1.65 (0.70, 3.89) 1.86 (0.70, 4.92)   Quartile 4: 4.62–6.63 43 53 3.84 (1.74, 8.48) 3.89 (1.50, 10.10)   Upper decile: 4.93–6.63 27 21 6.84 (2.65, 17.65) 7.30 (2.30, 23.16)   P for trend <0.0001 0.007   Per unit increment 1.01 (1.01, 1.02) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.63–0.86 39 54 1 1   Quartile 2: 0.87–0.94 19 53 0.46 (0.24, 0.89) 0.65 (0.31, 1.24)   Quartile 3: 0.95–1.03 22 56 0.52 (0.26, 1.05) 0.57 (0.26, 1.24)   Quartile 4: 1.04–2.53 14 51 0.31 (0.14, 0.69) 0.44 (0.17, 1.14)   Upper decile: 1.11–2.53 4 26 0.14 (0.04, 0.52) 0.17 (0.04, 0.76)   P for trend 0.005 0.047   Per unit increment 0.04 (0.00, 0.39) 0.07 (0.01, 0.76)  fT3/fT4 ratioc   Quartile 1: 1.80–3.96 8 53 1 1   Quartile 2: 3.97–4.45 16 53 2.14 (0.84, 5.48) 2.46 (0.78, 7.72)   Quartile 3: 4.46–5.12 21 53 3.29 (1.31, 8.27) 4.37 (1.41, 13.50)   Quartile 4: 5.13–7.83 47 53 8.61(3.37, 21.98) 13.60 (3.97, 46.30)   Upper decile: 5.67–7.83 28 22 9.09 (3.45, 23.98) 12.73 (3.71, 43.69)   P for trend <0.0001 <0.0001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.21–1.26 24 55 1 1   Quartile 2: 1.27–1.91 21 52 1.09 (0.53, 2.23) 0.95 (0.41, 2.17)   Quartile 3: 1.92–2.76 31 54 1.42 (0.72, 2.81) 1.44 (0.66, 3.11)   Quartile 4: 2.77–8.28 18 53 0.76 (0.37, 1.57) 0.93 (0.39, 2.23)   Upper decile: 3.77–8.28 9 21 0.92 (0.35, 2.46) 1.64 (0.49, 5.44)   P for trend 0.56 0.92   Per unit increment 0.96 (0.78, 1.19) 1.07 (0.83, 1.38) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  fT3, pg/dL   Quartile 1: 1.18–4.06 13 53 1 1   Quartile 2: 4.07–4.31 18 54 1.31 (0.57, 3.00) 1.32 (0.53, 3.28)   Quartile 3: 4.32–4.68 25 52 2.05 (0.91, 4.60) 1.97 (0.78, 4.96)   Quartile 4: 4.69–7.66 43 53 4.07 (1.82, 9.11) 4.25 (1.67, 10.80)   Upper decile: 5.02–7.66 29 21 6.09 (2.47, 15.00) 6.08 (2.19, 16.87)   P for trend <0.0001 0.001   Per unit increment 1.02 (1.01, 1.03) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.70–0.98 37 55 1 1   Quartile 2: 0.99–1.06 24 52 0.69 (0.37, 1.31) 0.70 (0.34, 1.45)   Quartile 3: 1.07–1.17 23 59 0.61 (0.32, 1.16) 0.69 (0.33, 1.42)   Quartile 4: 1.18–2.26 19 48 0.58 (0.29, 1.19) 0.63 (0.26, 1.51)   Upper decile: 1.29–2.26 7 19 0.61 (0.22, 1.66) 0.83 (0.27, 2.56)   P for trend 0.11 0.30   Per unit increment 0.23 (0.04, 1.22) 0.54 (0.08, 3.75)  fT3/fT4 ratioc   Quartile 1: 0.83–3.67 9 53 1 1   Quartile 2: 3.68–4.00 12 54 1.43 (0.51, 4.02) 1.12 (0.34, 3.74)   Quartile 3: 4.01–4.44 26 52 4.03 (1.56, 10.45) 5.26 (1.70, 16.20)   Quartile 4: 4.45–6.85 51 53 8.03 (3.13, 20.61) 8.63 (2.87, 26.00)   Upper decile: 5.39–6.85 25 22 10.25 (3.62, 29.01) 9.28 (2.76, 31.26)   P for trend <0.0001 0.001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.06–0.86 29 53 1 1   Quartile 2: 0.87–1.28 22 53 0.77 (0.38, 1.56) 0.71 (0.32, 1.59)   Quartile 3: 1.29–1.82 23 51 0.85 (0.43, 1.69) 0.63 (0.28, 1.44)   Quartile 4: 1.83–30.11 25 52 0.88 (0.46, 1.71) 1.17 (0.54, 2.51)   Upper decile: 2.53–30.11 12 20 1.11 (0.48, 2.57) 1.43 (0.53, 3.84)   P for trend 0.83 0.78   Per unit increment 1.08 (0.94, 1.24) 1.11 (0.93, 1.32) Gestational weeks 15-26b  fT3, pg/dL   Quartile 1: 2.95–3.93 14 57 1 1   Quartile 2: 3.94–4.24 19 50 1.78 (0.77, 4.13) 1.25 (0.46, 3.37)   Quartile 3: 4.25–4.61 16 52 1.65 (0.70, 3.89) 1.86 (0.70, 4.92)   Quartile 4: 4.62–6.63 43 53 3.84 (1.74, 8.48) 3.89 (1.50, 10.10)   Upper decile: 4.93–6.63 27 21 6.84 (2.65, 17.65) 7.30 (2.30, 23.16)   P for trend <0.0001 0.007   Per unit increment 1.01 (1.01, 1.02) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.63–0.86 39 54 1 1   Quartile 2: 0.87–0.94 19 53 0.46 (0.24, 0.89) 0.65 (0.31, 1.24)   Quartile 3: 0.95–1.03 22 56 0.52 (0.26, 1.05) 0.57 (0.26, 1.24)   Quartile 4: 1.04–2.53 14 51 0.31 (0.14, 0.69) 0.44 (0.17, 1.14)   Upper decile: 1.11–2.53 4 26 0.14 (0.04, 0.52) 0.17 (0.04, 0.76)   P for trend 0.005 0.047   Per unit increment 0.04 (0.00, 0.39) 0.07 (0.01, 0.76)  fT3/fT4 ratioc   Quartile 1: 1.80–3.96 8 53 1 1   Quartile 2: 3.97–4.45 16 53 2.14 (0.84, 5.48) 2.46 (0.78, 7.72)   Quartile 3: 4.46–5.12 21 53 3.29 (1.31, 8.27) 4.37 (1.41, 13.50)   Quartile 4: 5.13–7.83 47 53 8.61(3.37, 21.98) 13.60 (3.97, 46.30)   Upper decile: 5.67–7.83 28 22 9.09 (3.45, 23.98) 12.73 (3.71, 43.69)   P for trend <0.0001 <0.0001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.21–1.26 24 55 1 1   Quartile 2: 1.27–1.91 21 52 1.09 (0.53, 2.23) 0.95 (0.41, 2.17)   Quartile 3: 1.92–2.76 31 54 1.42 (0.72, 2.81) 1.44 (0.66, 3.11)   Quartile 4: 2.77–8.28 18 53 0.76 (0.37, 1.57) 0.93 (0.39, 2.23)   Upper decile: 3.77–8.28 9 21 0.92 (0.35, 2.46) 1.64 (0.49, 5.44)   P for trend 0.56 0.92   Per unit increment 0.96 (0.78, 1.19) 1.07 (0.83, 1.38) Boldface indicates statistically significant results. a Adjusted for maternal age (years), gestational age at blood collection (weeks), nulliparity (yes/no), education (less than, equal to, or more than high school), family history of diabetes (yes/no), prepregnancy BMI (<25, 25.0–29.9, ≥30.0 kg/m2), and thyroid autoimmunity status (antibodies positive vs negative). b Timing of blood sample collection preceded the diagnosis of gestational diabetes in all participants. c The tT3/fT4 ratio was obtained by dividing plasma concentration of fT3 (pg/dL) by fT4 level (ng/dL). View Large Table 3. aOR (95% CI) for GDM According to Quartiles of Thyroid Markers at Gestational Weeks 10–14 and 15–26 in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  fT3, pg/dL   Quartile 1: 1.18–4.06 13 53 1 1   Quartile 2: 4.07–4.31 18 54 1.31 (0.57, 3.00) 1.32 (0.53, 3.28)   Quartile 3: 4.32–4.68 25 52 2.05 (0.91, 4.60) 1.97 (0.78, 4.96)   Quartile 4: 4.69–7.66 43 53 4.07 (1.82, 9.11) 4.25 (1.67, 10.80)   Upper decile: 5.02–7.66 29 21 6.09 (2.47, 15.00) 6.08 (2.19, 16.87)   P for trend <0.0001 0.001   Per unit increment 1.02 (1.01, 1.03) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.70–0.98 37 55 1 1   Quartile 2: 0.99–1.06 24 52 0.69 (0.37, 1.31) 0.70 (0.34, 1.45)   Quartile 3: 1.07–1.17 23 59 0.61 (0.32, 1.16) 0.69 (0.33, 1.42)   Quartile 4: 1.18–2.26 19 48 0.58 (0.29, 1.19) 0.63 (0.26, 1.51)   Upper decile: 1.29–2.26 7 19 0.61 (0.22, 1.66) 0.83 (0.27, 2.56)   P for trend 0.11 0.30   Per unit increment 0.23 (0.04, 1.22) 0.54 (0.08, 3.75)  fT3/fT4 ratioc   Quartile 1: 0.83–3.67 9 53 1 1   Quartile 2: 3.68–4.00 12 54 1.43 (0.51, 4.02) 1.12 (0.34, 3.74)   Quartile 3: 4.01–4.44 26 52 4.03 (1.56, 10.45) 5.26 (1.70, 16.20)   Quartile 4: 4.45–6.85 51 53 8.03 (3.13, 20.61) 8.63 (2.87, 26.00)   Upper decile: 5.39–6.85 25 22 10.25 (3.62, 29.01) 9.28 (2.76, 31.26)   P for trend <0.0001 0.001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.06–0.86 29 53 1 1   Quartile 2: 0.87–1.28 22 53 0.77 (0.38, 1.56) 0.71 (0.32, 1.59)   Quartile 3: 1.29–1.82 23 51 0.85 (0.43, 1.69) 0.63 (0.28, 1.44)   Quartile 4: 1.83–30.11 25 52 0.88 (0.46, 1.71) 1.17 (0.54, 2.51)   Upper decile: 2.53–30.11 12 20 1.11 (0.48, 2.57) 1.43 (0.53, 3.84)   P for trend 0.83 0.78   Per unit increment 1.08 (0.94, 1.24) 1.11 (0.93, 1.32) Gestational weeks 15-26b  fT3, pg/dL   Quartile 1: 2.95–3.93 14 57 1 1   Quartile 2: 3.94–4.24 19 50 1.78 (0.77, 4.13) 1.25 (0.46, 3.37)   Quartile 3: 4.25–4.61 16 52 1.65 (0.70, 3.89) 1.86 (0.70, 4.92)   Quartile 4: 4.62–6.63 43 53 3.84 (1.74, 8.48) 3.89 (1.50, 10.10)   Upper decile: 4.93–6.63 27 21 6.84 (2.65, 17.65) 7.30 (2.30, 23.16)   P for trend <0.0001 0.007   Per unit increment 1.01 (1.01, 1.02) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.63–0.86 39 54 1 1   Quartile 2: 0.87–0.94 19 53 0.46 (0.24, 0.89) 0.65 (0.31, 1.24)   Quartile 3: 0.95–1.03 22 56 0.52 (0.26, 1.05) 0.57 (0.26, 1.24)   Quartile 4: 1.04–2.53 14 51 0.31 (0.14, 0.69) 0.44 (0.17, 1.14)   Upper decile: 1.11–2.53 4 26 0.14 (0.04, 0.52) 0.17 (0.04, 0.76)   P for trend 0.005 0.047   Per unit increment 0.04 (0.00, 0.39) 0.07 (0.01, 0.76)  fT3/fT4 ratioc   Quartile 1: 1.80–3.96 8 53 1 1   Quartile 2: 3.97–4.45 16 53 2.14 (0.84, 5.48) 2.46 (0.78, 7.72)   Quartile 3: 4.46–5.12 21 53 3.29 (1.31, 8.27) 4.37 (1.41, 13.50)   Quartile 4: 5.13–7.83 47 53 8.61(3.37, 21.98) 13.60 (3.97, 46.30)   Upper decile: 5.67–7.83 28 22 9.09 (3.45, 23.98) 12.73 (3.71, 43.69)   P for trend <0.0001 <0.0001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.21–1.26 24 55 1 1   Quartile 2: 1.27–1.91 21 52 1.09 (0.53, 2.23) 0.95 (0.41, 2.17)   Quartile 3: 1.92–2.76 31 54 1.42 (0.72, 2.81) 1.44 (0.66, 3.11)   Quartile 4: 2.77–8.28 18 53 0.76 (0.37, 1.57) 0.93 (0.39, 2.23)   Upper decile: 3.77–8.28 9 21 0.92 (0.35, 2.46) 1.64 (0.49, 5.44)   P for trend 0.56 0.92   Per unit increment 0.96 (0.78, 1.19) 1.07 (0.83, 1.38) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  fT3, pg/dL   Quartile 1: 1.18–4.06 13 53 1 1   Quartile 2: 4.07–4.31 18 54 1.31 (0.57, 3.00) 1.32 (0.53, 3.28)   Quartile 3: 4.32–4.68 25 52 2.05 (0.91, 4.60) 1.97 (0.78, 4.96)   Quartile 4: 4.69–7.66 43 53 4.07 (1.82, 9.11) 4.25 (1.67, 10.80)   Upper decile: 5.02–7.66 29 21 6.09 (2.47, 15.00) 6.08 (2.19, 16.87)   P for trend <0.0001 0.001   Per unit increment 1.02 (1.01, 1.03) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.70–0.98 37 55 1 1   Quartile 2: 0.99–1.06 24 52 0.69 (0.37, 1.31) 0.70 (0.34, 1.45)   Quartile 3: 1.07–1.17 23 59 0.61 (0.32, 1.16) 0.69 (0.33, 1.42)   Quartile 4: 1.18–2.26 19 48 0.58 (0.29, 1.19) 0.63 (0.26, 1.51)   Upper decile: 1.29–2.26 7 19 0.61 (0.22, 1.66) 0.83 (0.27, 2.56)   P for trend 0.11 0.30   Per unit increment 0.23 (0.04, 1.22) 0.54 (0.08, 3.75)  fT3/fT4 ratioc   Quartile 1: 0.83–3.67 9 53 1 1   Quartile 2: 3.68–4.00 12 54 1.43 (0.51, 4.02) 1.12 (0.34, 3.74)   Quartile 3: 4.01–4.44 26 52 4.03 (1.56, 10.45) 5.26 (1.70, 16.20)   Quartile 4: 4.45–6.85 51 53 8.03 (3.13, 20.61) 8.63 (2.87, 26.00)   Upper decile: 5.39–6.85 25 22 10.25 (3.62, 29.01) 9.28 (2.76, 31.26)   P for trend <0.0001 0.001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.06–0.86 29 53 1 1   Quartile 2: 0.87–1.28 22 53 0.77 (0.38, 1.56) 0.71 (0.32, 1.59)   Quartile 3: 1.29–1.82 23 51 0.85 (0.43, 1.69) 0.63 (0.28, 1.44)   Quartile 4: 1.83–30.11 25 52 0.88 (0.46, 1.71) 1.17 (0.54, 2.51)   Upper decile: 2.53–30.11 12 20 1.11 (0.48, 2.57) 1.43 (0.53, 3.84)   P for trend 0.83 0.78   Per unit increment 1.08 (0.94, 1.24) 1.11 (0.93, 1.32) Gestational weeks 15-26b  fT3, pg/dL   Quartile 1: 2.95–3.93 14 57 1 1   Quartile 2: 3.94–4.24 19 50 1.78 (0.77, 4.13) 1.25 (0.46, 3.37)   Quartile 3: 4.25–4.61 16 52 1.65 (0.70, 3.89) 1.86 (0.70, 4.92)   Quartile 4: 4.62–6.63 43 53 3.84 (1.74, 8.48) 3.89 (1.50, 10.10)   Upper decile: 4.93–6.63 27 21 6.84 (2.65, 17.65) 7.30 (2.30, 23.16)   P for trend <0.0001 0.007   Per unit increment 1.01 (1.01, 1.02) 1.02 (1.01, 1.03)  fT4, ng/dL   Quartile 1: 0.63–0.86 39 54 1 1   Quartile 2: 0.87–0.94 19 53 0.46 (0.24, 0.89) 0.65 (0.31, 1.24)   Quartile 3: 0.95–1.03 22 56 0.52 (0.26, 1.05) 0.57 (0.26, 1.24)   Quartile 4: 1.04–2.53 14 51 0.31 (0.14, 0.69) 0.44 (0.17, 1.14)   Upper decile: 1.11–2.53 4 26 0.14 (0.04, 0.52) 0.17 (0.04, 0.76)   P for trend 0.005 0.047   Per unit increment 0.04 (0.00, 0.39) 0.07 (0.01, 0.76)  fT3/fT4 ratioc   Quartile 1: 1.80–3.96 8 53 1 1   Quartile 2: 3.97–4.45 16 53 2.14 (0.84, 5.48) 2.46 (0.78, 7.72)   Quartile 3: 4.46–5.12 21 53 3.29 (1.31, 8.27) 4.37 (1.41, 13.50)   Quartile 4: 5.13–7.83 47 53 8.61(3.37, 21.98) 13.60 (3.97, 46.30)   Upper decile: 5.67–7.83 28 22 9.09 (3.45, 23.98) 12.73 (3.71, 43.69)   P for trend <0.0001 <0.0001   Per unit increment 1.01 (1.01, 1.02) 1.01 (1.01, 1.02)  TSH, mIU/L   Quartile 1: 0.21–1.26 24 55 1 1   Quartile 2: 1.27–1.91 21 52 1.09 (0.53, 2.23) 0.95 (0.41, 2.17)   Quartile 3: 1.92–2.76 31 54 1.42 (0.72, 2.81) 1.44 (0.66, 3.11)   Quartile 4: 2.77–8.28 18 53 0.76 (0.37, 1.57) 0.93 (0.39, 2.23)   Upper decile: 3.77–8.28 9 21 0.92 (0.35, 2.46) 1.64 (0.49, 5.44)   P for trend 0.56 0.92   Per unit increment 0.96 (0.78, 1.19) 1.07 (0.83, 1.38) Boldface indicates statistically significant results. a Adjusted for maternal age (years), gestational age at blood collection (weeks), nulliparity (yes/no), education (less than, equal to, or more than high school), family history of diabetes (yes/no), prepregnancy BMI (<25, 25.0–29.9, ≥30.0 kg/m2), and thyroid autoimmunity status (antibodies positive vs negative). b Timing of blood sample collection preceded the diagnosis of gestational diabetes in all participants. c The tT3/fT4 ratio was obtained by dividing plasma concentration of fT3 (pg/dL) by fT4 level (ng/dL). View Large Isolated hypothyroxinemia at the second, but not first trimester, was significantly associated with increased GDM risk: the aOR (95% CI) comparing women who had hypothyroxinemia to women who had euthyroid status was 1.56 (0.63, 3.89) in the first and 2.97 (1.07, 8.24) in the second trimester (Table 4). Subclinical or overt hypothyroidism in either first [aOR (95% CI), 2.58 (0.39, 17.01)] or second trimester [aOR (95% CI), 1.78 (0.49, 6.42)] was not related to GDM risk. Table 4. Adjusted OR (95% CI) for GDM According to Hypothyroidism or Isolated Hypothyroxinemia at Gestational Weeks 10–14 and 15–26 in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  Euthyroidc 79 171 1 1  Overt/subclinical hypothyroidismd 3 3 2.09 (0.41, 10.73) 2.58 (0.39, 17.01)  Isolated hypothyroxinemiae 13 15 2.18 (0.96, 4.94) 1.56 (0.63, 3.89) Gestational weeks 15–26b  Euthyroidc 71 163 1 1  Overt/subclinical hypothyroidismd 6 14 1.00 (0.36, 2.80) 1.78 (0.49, 6.42)  Isolated hypothyroxinemiae 13 11 2.93 (1.19, 7.21) 2.97 (1.07, 8.24) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  Euthyroidc 79 171 1 1  Overt/subclinical hypothyroidismd 3 3 2.09 (0.41, 10.73) 2.58 (0.39, 17.01)  Isolated hypothyroxinemiae 13 15 2.18 (0.96, 4.94) 1.56 (0.63, 3.89) Gestational weeks 15–26b  Euthyroidc 71 163 1 1  Overt/subclinical hypothyroidismd 6 14 1.00 (0.36, 2.80) 1.78 (0.49, 6.42)  Isolated hypothyroxinemiae 13 11 2.93 (1.19, 7.21) 2.97 (1.07, 8.24) a Adjusted for maternal age (years), gestational age at blood collection (weeks), nulliparity (yes/no), education (less than, equal to, or more than high school), family history of diabetes (yes/no), prepregnancy BMI (<25, 25.0–29.9, ≥30.0 kg/m2), and thyroid autoimmunity status (antibodies positive vs negative). b Timing of blood sample collection preceded the diagnosis of gestational diabetes in all participants. c Women with normal TSH (≤4 mIU/L) and normal fT4 concentration (between 10th and 90th percentile) were classified as euthyroid. d Overt/subclinical hypothyroidism having elevated TSH levels (>4 mIU/L) with low or normal fT4 concentration (≤90th percentile in controls). e Isolated hypothyroxinemia was defined as having normal TSH (≤4 mIU/L) in conjunction with low fT4 levels (<10th percentile in controls). View Large Table 4. Adjusted OR (95% CI) for GDM According to Hypothyroidism or Isolated Hypothyroxinemia at Gestational Weeks 10–14 and 15–26 in the NICHD Fetal Growth Studies—Singleton Cohort (2009–2013) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  Euthyroidc 79 171 1 1  Overt/subclinical hypothyroidismd 3 3 2.09 (0.41, 10.73) 2.58 (0.39, 17.01)  Isolated hypothyroxinemiae 13 15 2.18 (0.96, 4.94) 1.56 (0.63, 3.89) Gestational weeks 15–26b  Euthyroidc 71 163 1 1  Overt/subclinical hypothyroidismd 6 14 1.00 (0.36, 2.80) 1.78 (0.49, 6.42)  Isolated hypothyroxinemiae 13 11 2.93 (1.19, 7.21) 2.97 (1.07, 8.24) Case (n) Control (n) Crude Model Multivariable Modela Gestational weeks 10–14b  Euthyroidc 79 171 1 1  Overt/subclinical hypothyroidismd 3 3 2.09 (0.41, 10.73) 2.58 (0.39, 17.01)  Isolated hypothyroxinemiae 13 15 2.18 (0.96, 4.94) 1.56 (0.63, 3.89) Gestational weeks 15–26b  Euthyroidc 71 163 1 1  Overt/subclinical hypothyroidismd 6 14 1.00 (0.36, 2.80) 1.78 (0.49, 6.42)  Isolated hypothyroxinemiae 13 11 2.93 (1.19, 7.21) 2.97 (1.07, 8.24) a Adjusted for maternal age (years), gestational age at blood collection (weeks), nulliparity (yes/no), education (less than, equal to, or more than high school), family history of diabetes (yes/no), prepregnancy BMI (<25, 25.0–29.9, ≥30.0 kg/m2), and thyroid autoimmunity status (antibodies positive vs negative). b Timing of blood sample collection preceded the diagnosis of gestational diabetes in all participants. c Women with normal TSH (≤4 mIU/L) and normal fT4 concentration (between 10th and 90th percentile) were classified as euthyroid. d Overt/subclinical hypothyroidism having elevated TSH levels (>4 mIU/L) with low or normal fT4 concentration (≤90th percentile in controls). e Isolated hypothyroxinemia was defined as having normal TSH (≤4 mIU/L) in conjunction with low fT4 levels (<10th percentile in controls). View Large The results were similar in sensitivity analyses excluding women who had elevated thyroid antibodies, prior history of GDM or preeclampsia, smoked before pregnancy, or had medication-treated GDM. The associations also persisted when limiting the sample to only women who had euthyroid status (n = 252 at weeks 10 to 14; n = 232 at weeks 15 to 26), or when stratifying the analyses by prepregnancy BMI status (BMI < 25.0 vs BMI ≥ 25.0 kg/m2), race/ethnicity (non-Hispanic white, African American, Hispanic, Asian/Pacific Islander), or family history of diabetes (yes vs no). In additional exploratory analyses (data not shown), we compared thyroid markers between GDM cases with 2 vs 3+ OGTT measures above the threshold, and we found that the latter group had higher fT3 and an fT3/fT4 ratio at both visits, although the differences were not statistically significant. Additionally, among GDM cases, fT3 and the fT3/fT4 ratio at both visits was found to be significantly and positively correlated with fasting glucose levels from OGTT. In secondary analyses, we examined the longitudinal changes in the level of thyroid markers during the course of pregnancy (Fig. 1). Both fT3 and fT4 levels declined with the progression of pregnancy. Overall, for most study visits, fT3 and the fT3/fT4 ratio were significantly higher, whereas fT4 was significantly lower among cases, as compared with non-GDM controls. TSH levels appeared to increase sharply from the first to second visit and then level off. The difference between cases and controls was significant only at the last visit (weeks 33 to 39), with higher TSH levels among GDM cases. Figure 1. View largeDownload slide Median plasma concentrations of (A) fT3, (B) fT4, (C) fT3/fT4 ratio, and (D) TSH at each study visit among women with GDM and their matched controls. Solid line indicates non-GDM controls; dashed line indicates GDM cases. Weeks 10 to 14, n = 104 and 214 for cases and controls, respectively; weeks 15 to 26, n = 94 and 212 for cases and controls, respectively; weeks 23 to 31, n = 102 and 106 for cases and controls, respectively; weeks 33 to 39, n = 88 and 101 for cases and controls, respectively. *P < 0.05, **P < 0.01, ***P < 0.001 for case-control comparisons at each study visit obtained by generalized linear mixed effects models accounting for matched case-control pairs. Figure 1. View largeDownload slide Median plasma concentrations of (A) fT3, (B) fT4, (C) fT3/fT4 ratio, and (D) TSH at each study visit among women with GDM and their matched controls. Solid line indicates non-GDM controls; dashed line indicates GDM cases. Weeks 10 to 14, n = 104 and 214 for cases and controls, respectively; weeks 15 to 26, n = 94 and 212 for cases and controls, respectively; weeks 23 to 31, n = 102 and 106 for cases and controls, respectively; weeks 33 to 39, n = 88 and 101 for cases and controls, respectively. *P < 0.05, **P < 0.01, ***P < 0.001 for case-control comparisons at each study visit obtained by generalized linear mixed effects models accounting for matched case-control pairs. Discussion In this longitudinal study, we provide evidence that thyroid function early in pregnancy may be an indicator for subsequent risk of GDM, a common metabolic complication in pregnancy. To our knowledge, this is the first study to identify fT3 and the fT3/fT4 ratio measured early in pregnancy as independent risk factors of GDM. Increased levels of either marker were associated with a greater GDM risk, even after adjusting for potential confounders such as prepregnancy BMI and family history of diabetes. The fT3/fT4 ratio, a marker indicating the conversion rate from T4 to T3, was most strongly associated with GDM, with women in the highest quartile in the second trimester showing an almost 14-fold increased risk compared with women in the lowest quartile. Although T3 is the primary, biologically active hormone involved in glucose homeostasis, prior studies examining thyroid biomarkers in relationship to GDM risk have been mostly focused on its precursor hormone, fT4, and their regulatory hormone, TSH. Two prospective studies (11, 23) to date have examined the association between fT3 levels and GDM, and both reported no significant differences in fT3 levels in early pregnancy between women with and without subsequent GDM. Inferences of findings from the two studies (11, 23) were hindered by the relatively small number of GDM cases, which may have limited their statistical power of detecting a significant association. Differences in demographic composition, including race/ethnicity, GDM diagnostic criteria, and population-specific reference intervals for thyroid hormones, may also have contributed to divergent findings. Notably, in our study, we identified the fT3/fT4 ratio, a commonly used proxy of peripheral deiodinase activity (15, 16), as a novel risk factor for GDM. Although the fT3/fT4 ratio has not previously been examined in relationship to GDM risk, our findings are consistent with others (17–19) showing a significant association between the fT3:fT4 ratio and measures of glucose and insulin metabolism (e.g., elevated fasting glucose). Taken together, these findings suggest that higher fT3 levels, which could result from either increased fT4 to fT3 conversion or increased T3 synthesis from the thyroid gland, could be related to the pathophysiology of GDM. Consistent with previous findings (9, 10), we observed lower fT4 levels among women with GDM, yet the associations were not significant after adjusting for potential confounders, particularly prepregnancy BMI. However, compared with women who had euthyroid status in our study, women who had isolated hypothyroxinemia (normal TSH, low fT4 levels) in the second trimester had an almost threefold greater risk of GDM. Short-term fasting is known to affect TSH and not fT4 levels (24–26), but whether the fasting status in the second trimester influenced this trimester-specific association is not clear in our study. However, other studies have also observed that isolated hypothyroxinemia in the second, but not the first trimester, is related to subsequent GDM risk (6, 9, 14). Neither hypothyroidism nor TSH levels alone were associated with GDM risk in our study, which is consistent with some studies (6–8, 11), but contrasts with others (3–5, 23) that observed an elevated GDM risk among women who had subclinical hypothyroidism or high TSH levels in pregnancy. Differences in population characteristics, study design, and sample size may account for these discrepant findings. Of note, all three studies (6–8) reporting a null association had very few women who developed both GDM and subclinical hypothyroidism. Recently, a meta-analysis of six cohort studies also showed that subclinical hypothyroidism was significantly associated with a 1.35-fold increased risk of GDM as compared with women who had euthyroid status (2). Because our sample only had 20 women who had overt or subclinical hypothyroidism before GDM diagnosis, we cannot rule out the possibility that the observed lack of a significant association could be due to inadequate statistical power. Our findings are biologically plausible. Thyroid hormones regulate hepatic gluconeogenesis, intestinal absorption of glucose, and uptake of glucose in peripheral tissues (2). Additionally, they modulate messenger RNA and protein expression of glucose transporters, promote pathways that accelerate glycogenolysis, and modify circulating insulin levels and counterregulatory hormones (27, 28). Among the thyroid hormones, T3 is the biologically active hormone responsible for stimulating endogenous glucose production, with several studies noting that fT3 levels are positively associated with insulin secretion and hyperinsulinemia (29, 30). Around 80% of circulating T3 levels are derived from monodeiodination of T4 carried out by peripheral deiodinase activity (13), supporting the notion that the fT3/fT4 ratio could serve as an important marker for glucose homeostasis. Furthermore, a missense variant (Thr92Ala) in the gene encoding type 2 deiodinase, the subtype of deiodinase specific to generating T3 from T4, has been found to be associated with insulin resistance as measured by the hyperinsulinemic–euglycemic clamp (31). As a secondary objective of our study, we profiled the longitudinal trajectory of the thyroid hormones across the entire course of pregnancy. In our sample, we observed expected changes in the levels of TSH, fT3, and fT4 hormones across pregnancy. Maternal thyroid physiology changes considerably in pregnancy owing to several mechanisms, including transient rise in human chorionic gonadotropin in early pregnancy, increased concentrations of T4 binding proteins, increased thyroid hormone metabolism by the placenta, and greater iodide excretion in the urine (1). Although women were enrolled in our study early in pregnancy between 10 and 14 weeks, we likely missed the human chorionic gonadotropin–induced decrease in TSH levels very early in gestation (1), only capturing the gradual increase in TSH levels thereafter (i.e., second and third trimesters). fT4 and fT3 levels showed a continuous decline with the progression of pregnancy, which was consistent with findings of others (32, 33). Of note, our study was unique in that we longitudinally profiled the thyroid hormone parameters in a relatively large multiracial cohort of pregnant women. There are several strengths to our study. First, we longitudinally measured several markers of thyroid function across pregnancy, allowing us to prospectively examine the trimester-specific associations between thyroid status and GDM, which is critically important given the changes in thyroid hormones across gestation (32, 33). Second, because thyroid autoimmunity status may influence both thyroid hormone levels and glucose homeostasis, an additional strength of this study was that we measured and accounted for thyroid antibody levels in our primary analyses. Third, our sample had a good representation of four major racial/ethnic groups in the United States, and GDM status in these women was well characterized based on review of medical records. Moreover, our study sample included relatively healthy women without pre-existing thyroid disease or any other chronic conditions. Thyroid medication use in pregnancy was also ascertained and accounted for. One of the limitations of this study was that due to the low frequency of clinical thyroid conditions in our study sample, we could not consider the joint effect of thyroid autoimmunity status and hypothyroidism on GDM risk. Because maternal iodine status is an important determinant of thyroid hormone levels, another limitation was the lack of iodine measurements in our study. However, it is reasonable to assume that our relatively healthy sample of US women would be iodine sufficient. Lastly, trimester-specific ranges for thyroid hormone levels were not available from our laboratory, and, as such, we used reference ranges recommended by the 2017 American Thyroid Association guidelines. In summary, findings from this longitudinal study suggest that higher fT3 levels, potentially resulting from de novo synthesis or increased deiodinase activity, may be involved in the pathophysiology of GDM. At present, the utility of routine screening for thyroid function during pregnancy is controversial. This study adds an important piece of evidence to this debate, as our findings show that women with thyroid abnormalities in early to middle pregnancy are at an increased risk for GDM and its adverse health sequelae. Our findings, in conjunction with previous evidence of thyroid-related adverse pregnancy outcomes, support the potential benefits of thyroid screening among pregnant women. Abbreviations: Abbreviations: aOR adjusted OR BMI body mass index GDM gestational diabetes mellitus fT3 free T3 fT4 free T4 NICHD National Institute of Child Health and Human Development OGTT oral glucose tolerance test Acknowledgments We thank the research teams at our study sites, including Christina Care Health Systems, University of California, Irvine, Long Beach Memorial Medical Center, Northwestern University, Medical University of South Carolina, Columbia University, New York Hospital Queens, St. Peters’ University Hospital, University of Alabama at Birmingham, Women and Infants Hospital of Rhode Island, Fountain Valley Regional Hospital and Medical Center, and Tufts University. We thank the C-TASC corporation for their assistance with data coordination. Lastly, we acknowledge the Department of Laboratory Medicine and Pathology, University of Minnesota for providing laboratory support in analyzing biospecimens and biomarkers. Financial Support: This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development intramural funding as well as by the American Recovery and Reinvestment Act funding (Contracts HHSN275200800013C, HHSN275200800002I, HHSN27500006, HHSN275200800003IC, HHSN275200800014C, HHSN275200800012C, HHSN275200800028C, HHSN275201000009C, and HHSN275201000001Z). W.B. was supported by research grants from the National Institutes of Health (Grant R21HD091458) and the Fraternal Order of Eagles Diabetes Research Center. Clinical Trial Information: ClinicalTrials.gov no. NCT00912132 (registered 3 June 2009). Author Contributions: S.R. analyzed the data and wrote the first draft of the manuscript. M.Y.T. assisted with laboratory testing, data interpretation, and reviewed the manuscript. W.B. assisted with case-control selection and coordinated biospecimen sampling from the biorepository. S.N.H., Y.Z., W.B., Y.L., P.P., and R.C.W.M. contributed to data interpretation and reviewed the manuscript. P.S.A. contributed to data analysis and interpretation and reviewed the manuscript. C.Z. obtained funding, designed and oversaw the study, and revised the manuscript. All authors contributed to the critical interpretation of the results, reviewed the manuscript for important intellectual content, approved the final version of the manuscript, and have agreed to be accountable for his/her role in this manuscript. S.R. and C.Z. are the guarantors of this work and, as such, had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Disclosure Summary: The authors have nothing to disclose. References 1. Alexander EK , Pearce EN , Brent GA , Brown RS , Chen H , Dosiou C , Grobman WA , Laurberg P , Lazarus JH , Mandel SJ , Peeters RP , Sullivan S . 2017 Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum . Thyroid . 2017 ; 27 ( 3 ): 315 – 389 . 2. Toulis KA , Stagnaro-Green A , Negro R . Maternal subclinical hypothyroidsm and gestational diabetes mellitus: a meta-analysis . Endocr Pract . 2014 ; 20 ( 7 ): 703 – 714 . 3. Sahu MT , Das V , Mittal S , Agarwal A , Sahu M . Overt and subclinical thyroid dysfunction among Indian pregnant women and its effect on maternal and fetal outcome . Arch Gynecol Obstet . 2010 ; 281 ( 2 ): 215 – 220 . 4. Tudela CM , Casey BM , McIntire DD , Cunningham FG . Relationship of subclinical thyroid disease to the incidence of gestational diabetes . Obstet Gynecol . 2012 ; 119 ( 5 ): 983 – 988 . 5. Ying H , Tang YP , Bao YR , Su XJ , Cai X , Li YH , Wang DF . Maternal TSH level and TPOAb status in early pregnancy and their relationship to the risk of gestational diabetes mellitus . Endocrine . 2016 ; 54 ( 3 ): 742 – 750 . 6. Cleary-Goldman J , Malone FD , Lambert-Messerlian G , Sullivan L , Canick J , Porter TF , Luthy D , Gross S , Bianchi DW , D’Alton ME . Maternal thyroid hypofunction and pregnancy outcome . Obstet Gynecol . 2008 ; 112 ( 1 ): 85 – 92 . 7. Männistö T , Vääräsmäki M , Pouta A , Hartikainen AL , Ruokonen A , Surcel HM , Bloigu A , Järvelin MR , Suvanto E . Thyroid dysfunction and autoantibodies during pregnancy as predictive factors of pregnancy complications and maternal morbidity in later life . J Clin Endocrinol Metab . 2010 ; 95 ( 3 ): 1084 – 1094 . 8. Chen LM , Du WJ , Dai J , Zhang Q , Si GX , Yang H , Ye EL , Chen QS , Yu LC , Zhang C , Lu XM . Effects of subclinical hypothyroidism on maternal and perinatal outcomes during pregnancy: a single-center cohort study of a Chinese population . PLoS One . 2014 ; 9 ( 10 ): e109364 . 9. Oguz A , Tuzun D , Sahin M , Usluogullari AC , Usluogullari B , Celik A , Gul K . Frequency of isolated maternal hypothyroxinemia in women with gestational diabetes mellitus in a moderately iodine-deficient area . Gynecol Endocrinol . 2015 ; 31 ( 10 ): 792 – 795 . 10. Yang S , Shi FT , Leung PC , Huang HF , Fan J . Low thyroid hormone in early pregnancy is associated with an increased risk of gestational diabetes mellitus . J Clin Endocrinol Metab . 2016 ; 101 ( 11 ): 4237 – 4243 . 11. Agarwal MM , Dhatt GS , Punnose J , Bishawi B , Zayed R . Thyroid function abnormalities and antithyroid antibody prevalence in pregnant women at high risk for gestational diabetes mellitus . Gynecol Endocrinol . 2006 ; 22 ( 5 ): 261 – 266 . 12. Casey BM , Dashe JS , Spong CY , McIntire DD , Leveno KJ , Cunningham GF . Perinatal significance of isolated maternal hypothyroxinemia identified in the first half of pregnancy . Obstet Gynecol . 2007 ; 109 ( 5 ): 1129 – 1135 . 13. Maia AL , Goemann IM , Meyer EL , Wajner SM . Type 1 iodothyronine deiodinase in human physiology and disease. Deiodinases: the balance of thyroid hormone . J Endocrinol . 2011 ; 209 ( 3 ): 283 – 297 . 14. Haddow JE , Craig WY , Neveux LM , Palomaki GE , Lambert-Messerlian G , Malone FD , D’Alton ME ; First and Second Trimester Risk of Aneuploidy (FaSTER) Research Consortium . Free thyroxine during early pregnancy and risk for gestational diabetes . PLoS One . 2016 ; 11 ( 2 ): e0149065 . 15. Nicoloff JT , Lum SM , Spencer CA , Morris R . Peripheral autoregulation of thyroxine to triiodothyronine conversion in man . Horm Metab Res Suppl . 1984 ; 14 : 74 – 79 . 16. Keck FS , Loos U . Peripheral autoregulation of thyromimetic activity in man . Horm Metab Res . 1988 ; 20 ( 2 ): 110 – 114 . 17. Bassols J , Prats-Puig A , Soriano-Rodríguez P , García-González MM , Reid J , Martínez-Pascual M , Mateos-Comerón F , de Zegher F , Ibáñez L , López-Bermejo A . Lower free thyroxin associates with a less favorable metabolic phenotype in healthy pregnant women . J Clin Endocrinol Metab . 2011 ; 96 ( 12 ): 3717 – 3723 . 18. Jing S , Xiaoying D , Ying X , Rui L , Mingyu G , Yuting C , Yanhua Y , Yufan W , Haiyan S , Yongde P . Different levels of thyroid hormones between impaired fasting glucose and impaired glucose tolerance: free T3 affects the prevalence of impaired fasting glucose and impaired glucose tolerance in opposite ways . Clin Endocrinol (Oxf) . 2014 ; 80 ( 6 ): 890 – 898 . 19. Knight BA , Shields BM , Hattersley AT , Vaidya B . Maternal hypothyroxinaemia in pregnancy is associated with obesity and adverse maternal metabolic parameters . Eur J Endocrinol . 2016 ; 174 ( 1 ): 51 – 57 . 20. Buck Louis GM , Grewal J , Albert PS , Sciscione A , Wing DA , Grobman WA , Newman RB , Wapner R , D’Alton ME , Skupski D , Nageotte MP , Ranzini AC , Owen J , Chien EK , Craigo S , Hediger ML , Kim S , Zhang C , Grantz KL . Racial/ethnic standards for fetal growth: the NICHD fetal growth studies . Am J Obstet Gynecol . 2015 ; 213 ( 4 ): 449.e1 – 449.e41 . 21. Committee on Practice Bulletins—Obstetrics . Practice Bulletin No. 137: Gestational diabetes mellitus . Obstet Gynecol . 2013 ; 122 ( 2 Pt 1 ): 406 – 416 . 22. Stagnaro-Green A , Abalovich M , Alexander E , Azizi F , Mestman J , Negro R , Nixon A , Pearce EN , Soldin OP , Sullivan S , Wiersinga W ; American Thyroid Association Taskforce on Thyroid Disease During Pregnancy and Postpartum . Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and postpartum . Thyroid . 2011 ; 21 ( 10 ): 1081 – 1125 . 23. Karakosta P , Alegakis D , Georgiou V , Roumeliotaki T , Fthenou E , Vassilaki M , Boumpas D , Castanas E , Kogevinas M , Chatzi L . Thyroid dysfunction and autoantibodies in early pregnancy are associated with increased risk of gestational diabetes and adverse birth outcomes . J Clin Endocrinol Metab . 2012 ; 97 ( 12 ): 4464 – 4472 . 24. Nair R , Mahadevan S , Muralidharan RS , Madhavan S . Does fasting or postprandial state affect thyroid function testing ? Indian J Endocrinol Metab . 2014 ; 18 ( 5 ): 705 – 707 . 25. Scobbo RR , VonDohlen TW , Hassan M , Islam S . Serum TSH variability in normal individuals: the influence of time of sample collection . W V Med J . 2004 ; 100 ( 4 ): 138 – 142 . 26. Kamat V , Hecht WL , Rubin RT . Influence of meal composition on the postprandial response of the pituitary-thyroid axis . Eur J Endocrinol . 1995 ; 133 ( 1 ): 75 – 79 . 27. Das DK , Bandyopadhyay D , Bandyopadhyay S , Neogi A . Thyroid hormone regulation of β-adrenergic receptors and catecholamine sensitive adenylate cyclase in foetal heart . Acta Endocrinol (Copenh) . 1984 ; 106 ( 4 ): 569 – 576 . 28. Kemp HF , Hundal HS , Taylor PM . Glucose transport correlates with GLUT2 abundance in rat liver during altered thyroid status . Mol Cell Endocrinol . 1997 ; 128 ( 1-2 ): 97 – 102 . 29. Bakker SJ , ter Maaten JC , Popp-Snijders C , Heine RJ , Gans RO . Triiodothyronine: a link between the insulin resistance syndrome and blood pressure ? J Hypertens . 1999 ; 17 ( 12 Pt 1 ): 1725 – 1729 . 30. Ortega E , Koska J , Pannacciulli N , Bunt JC , Krakoff J . Free triiodothyronine plasma concentrations are positively associated with insulin secretion in euthyroid individuals . Eur J Endocrinol . 2008 ; 158 ( 2 ): 217 – 221 . 31. Mentuccia D , Proietti-Pannunzi L , Tanner K , Bacci V , Pollin TI , Poehlman ET , Shuldiner AR , Celi FS . Association between a novel variant of the human type 2 deiodinase gene Thr92Ala and insulin resistance: evidence of interaction with the Trp64Arg variant of the β-3-adrenergic receptor . Diabetes . 2002 ; 51 ( 3 ): 880 – 883 . 32. Soldin OP , Tractenberg RE , Hollowell JG , Jonklaas J , Janicic N , Soldin SJ . Trimester-specific changes in maternal thyroid hormone, thyrotropin, and thyroglobulin concentrations during gestation: trends and associations across trimesters in iodine sufficiency . Thyroid . 2004 ; 14 ( 12 ): 1084 – 1090 . 33. Moncayo R , Zanon B , Heim K , Ortner K , Moncayo H . Thyroid function parameters in normal pregnancies in an iodine sufficient population . BBA Clin . 2015 ; 3 : 90 – 95 . Copyright © 2018 Endocrine Society

de Jesús Garduno-Garcia, José;Gastaldelli, Amalia;DeFronzo, Ralph A;Lertwattanarak, Raweewan;Holst, Jens J;Musi, Nicolas

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00260pmid: 29672742

Abstract Objective Insulin secretion (IS) declines with age, which increases the risk of impaired glucose tolerance (IGT) and type 2 diabetes mellitus (T2DM) in older adults. IS is regulated by the incretin hormones glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP). Here we tested the hypotheses that incretin release is lower in older adults and that this decline is associated with β-cell dysfunction. Research Design A total of 40 young (25 ± 3 years) and 53 older (74 ± 7 years) lean nondiabetic subjects underwent a 2-hour oral glucose tolerance test (OGTT). Based on the OGTT, subjects were divided into three groups: young subjects with normal glucose tolerance (Y-NGT; n = 40), older subjects with normal glucose tolerance (O-NGT; n = 32), and older subjects with IGT (O-IGT; n = 21). Main Outcome Measures Plasma insulin, C-peptide, GLP-1, and GIP concentrations were measured every 15 to 30 minutes. We quantitated insulin sensitivity (Matsuda index) and insulin secretory rate (ISR) by deconvolution of C-peptide with the calculation of β-cell glucose sensitivity. Results Matsuda index, early phase ISR (0 to 30 minutes), and parameters of β-cell function were lower in O-IGT than in Y-NGT subjects but not in O-NGT subjects. GLP-1 concentrations were elevated in both older groups [GLP-1 area under the curve (AUC)0–120 was 2.8 ± 0.1 in Y-NGT, 3.8 ± 0.5 in O-NGT, and 3.7 ± 0.4 nmol/L∙120 minutes in O-IGT subjects; P < 0.05], whereas GIP secretion was higher in O-NGT than in Y-NGT subjects (GIP AUC0–120 was 4.7 ± 0.3 in Y-NGT, 6.0 ± 0.4 in O-NGT, and 4.8 ± 0.3 nmol/L∙120 minutes in O-IGT subjects; P < 0.05). Conclusions Aging is associated with an exaggerated GLP-1 secretory response. However, it was not sufficient to increase insulin first-phase release in O-IGT and overcome insulin resistance. Aging is associated with major changes in glucose metabolism. Various studies indicate that increasing age is accompanied by impaired glucose tolerance (IGT). The 2-hour plasma glucose concentration during an oral glucose tolerance test (OGTT) rises on average 5.3 mg/dL per decade (1). The Baltimore Longitudinal Study of Aging showed a progressive decline in glucose tolerance from the third through the ninth decade of life (2). This decline in glucose tolerance with age also was evident in the National Health and Nutrition Examination Survey III, in agreement with recent estimates that approximately one-third of subjects aged ≥65 have diabetes (3). The cause for the high prevalence of glucose intolerance and type 2 diabetes mellitus (T2DM) in the older population is not clear. Yet many factors have been implicated, including changes in fat distribution (4), physical activity (5), muscle insulin sensitivity (6), and β-cell function (7). A decrease in β-cell function with advancing age has been previously documented (8). For example, β-cell responsiveness during a frequent sampled intravenous glucose tolerance test is lower in older nondiabetic subjects than in younger subjects (9, 10). In addition, β-cell insulin response upon arginine stimulation also is impaired in older subjects (9, 10). Despite these data indicating that aging leads to decreases in β-cell function, the cause of this age-dependent functional decline is not known. Some studies have found that β-cell mass declines with age (11), although other studies have not reported such decline (12). Incretin hormones glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) regulate β-cell function and mass (13). These peptides are secreted by the gut in response to nutrients, increasing insulin secretion rate (ISR) by β cells and reducing glucagon release (14). Although T2DM and adiposity are associated with altered incretin release and β-cell resistance to both GLP-1 and GIP (15, 16), it is unclear whether the decline in β-cell function seen in normal aging (i.e., in the absence of T2DM) also is associated with defects in incretin secretion. The purpose of this study was to evaluate incretin hormone secretion in lean older subjects with either normal glucose tolerance (NGT) or IGT, compared with lean young subjects with NGT. We also examined whether possible age-related differences in incretin release are associated with changes in β-cell function. Because incretins increase β-cell function and mass (15), we hypothesized that aging leads to a reduced or impaired incretin release in response to glucose load and that this impairment would be associated with reduced β-cell function. Methods Subjects We studied 40 young (18 to 30 years old) and 53 older (≥65 years old) nondiabetic, nonobese subjects. Each subject underwent a medical history, physical examination, screening laboratory tests, and a 75-g OGTT. All subjects were sedentary (not more than one session of exercise per week) and community-dwelling. Subjects were not obese (body mass index = 23 to 26 kg/m2) and did not have a family history (first-degree relative) of diabetes. Body weight was stable (±1 kg) for at least 3 months before enrollment. Subjects were not taking medication known to affect glucose metabolism. The study was approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio, and all subjects gave written voluntary consent. OGTT Plasma glucose, insulin, and C-peptide concentrations were measured at baseline and every 15 minutes for 2 hours after the ingestion of 75 g glucose; GLP-1 and GIP were measured every 30 minutes in samples collected in prechilled test tubes containing aprotinin and EDTA. Based on the OGTT, subjects from the older group were subdivided into NGT or IGT groups. All subjects in the young group had NGT. The incremental area under the curve (AUC) for plasma glucose and insulin during the OGTT was calculated with the trapezoidal rule (17). We calculated the homeostatic model assessment of insulin resistance (IR) index and the Matsuda index for insulin sensitivity, as previously described (18, 19). The primary stimulus for ISR is the increment in plasma glucose in the first minutes after the glucose load. Thus, we calculated the insulinogenic index as the incremental AUC for plasma insulin concentration (ΔI) divided by the incremental AUC for plasma glucose concentration (ΔG) from 0 to 30 minutes and the late insulin response as ΔI/ΔG from 30 to 120 minutes. The prehepatic ISR was calculated by plasma C-peptide deconvolution using MLAB (Civilized Software, Inc.; Silver Spring, MD) (20, 21). We also calculated the indexes of β-cell insulin secretion (IS), the rate sensitivity, that is, the IS in response to changes in glucose concentration, and the β-cell glucose sensitivity, defined as the slope of the dose response between ISR and glucose excursion, as previously described (22, 23). The disposition index during OGTT was calculated as ΔI/ΔG × Matsuda index from 0 to 30 and 0 to 120 minutes, respectively (21, 24). We also calculated the ratio of IS to IR, or IS/IR index, as ΔISR/ΔG × Matsuda index (21, 24). Laboratory analyses Plasma insulin and C-peptide concentrations were measured by radioimmunoassay (Diagnostic Products, Los Angeles, CA), glucose was measured with the oxidase method on a Beckman analyzer (Beckman Coulter, Inc., Brea, CA), and hemoglobin A1c was measured with a DCA 2000 analyzer (Bayer Corporation, Tarrytown, NY). GIP was measured by radioimmunoassay with a C-terminally directed antiserum code #867, raised against a synthetic peptide corresponding to the C-terminus of human GIP (University of Copenhagen, Copenhagen, Denmark), thus measuring “total” GIP (intact GIP + the primary metabolite GIP 3-42). Total GLP-1 (intact GLP-1 + the primary metabolite GLP-1 9-36 amide) was measured by radioimmunoassay according to standards of synthetic GLP-1 7-36 amide and antiserum code no. 89390, which is specific for the amidated C-terminus of GLP-1 (University of Copenhagen, Copenhagen, Denmark). Plasma concentrations of total cholesterol and triglyceride were measured enzymatically (Boehringer-Mannheim, Indianapolis, IN). Plasma high-density lipoprotein cholesterol was measured enzymatically on Hitachi 704 autoanalyzer (Boehringer-Mannheim) after precipitation of chylomicron and very low-density lipoprotein and low-density lipoprotein (LDL) cholesterol by phosphotungstic acid precipitation. LDL cholesterol was calculated from the Friedwald equation. Statistical methods All continuous data (mean ± standard error) and qualitative variables were expressed as percentages. The Kolmogorov-Smirnov test was performed to evaluate distribution of the variables. Comparison between groups (young vs old) was performed with t tests for quantitative variables with normal distribution and with Mann-Whitney U tests for those with nonnormal distribution. To compare more than two groups we used analysis of variance and Tukey’s test for normally distributed variables and the Kruskal-Wallis procedure for nonnormal data distribution. Correlations between continuous variables were carried out with Pearson correlations for variables with normal distribution and Spearman for those with nonnormal distribution. Results Subject characteristics We studied 93 subjects subdivided into three groups according to age and degree of glucose tolerance: young subjects with NGT (Y-NGT; n = 40, mean age 25 years), older subjects with NGT (O-NGT; n = 32, mean age 72), and older subjects with IGT (O-IGT; n = 21, mean age 77). Anthropometric and metabolic characteristics of the subjects are shown in Table 1. Sex distribution and body mass index were not statistically different between the three groups. Total and LDL cholesterol concentrations were higher in older IGT compared with NGT subjects (both Y-NGT and O-NGT), whereas high-densitylipoprotein cholesterol and triglycerides were similar in all groups. Fasting plasma insulin was similar in the three groups, whereas the fasting plasma glucose was slightly but significantly higher in both older groups (O-NGT, O-IGT) than in the Y-NGT group (Table 1). The glucose AUC increased progressively from Y-NGT to O-NGT to O-IGT (Fig. 1A). Table 1. Baseline Subject Characteristics Y-NGT O-NGT O-IGT Pa No. 40 32 21 Age, y 25.4 ± 3.4 71.9 ± 7.2b 76.6 ± 6.7b <0.0001 Sex, female/male 26/14 14/19 11/9 NS BMI, kg/m2 23.8 ± 2.5 25.2 ± 2.9 23.8 ± 2.8 0.16 Hemoglobin A1c, % 5.1 ± 0.3 5.5 ± 0.3 5.6 ± 0.3b <0.0001 Insulin, pmol/L 36.8 ± 4.9 35.9 ± 3.5 44.3 ± 7.8 0.46 Glucose, mmol/L 5.06 ± 0.55 5.36 ± 0.54b 5.52 ± 0.46b 0.001 Total cholesterol, mmol/L 3.89 ± 0.17 4.39 ± 0.23 4.40 ± 0.17b 0.04 HDL cholesterol, mmol/L 1.43 ± 0.08 1.51 ± 0.09 1.54 ± 0.11 0.39 LDL cholesterol, mmol/L 1.98 ± 0.18 2.43 ± 0.20 2.41 ± 0.13b 0.07 Triglycerides, mmol/L 1.07 ± 0.17 0.97 ± 0.18 1.05 ± 0.08c 0.91 Y-NGT O-NGT O-IGT Pa No. 40 32 21 Age, y 25.4 ± 3.4 71.9 ± 7.2b 76.6 ± 6.7b <0.0001 Sex, female/male 26/14 14/19 11/9 NS BMI, kg/m2 23.8 ± 2.5 25.2 ± 2.9 23.8 ± 2.8 0.16 Hemoglobin A1c, % 5.1 ± 0.3 5.5 ± 0.3 5.6 ± 0.3b <0.0001 Insulin, pmol/L 36.8 ± 4.9 35.9 ± 3.5 44.3 ± 7.8 0.46 Glucose, mmol/L 5.06 ± 0.55 5.36 ± 0.54b 5.52 ± 0.46b 0.001 Total cholesterol, mmol/L 3.89 ± 0.17 4.39 ± 0.23 4.40 ± 0.17b 0.04 HDL cholesterol, mmol/L 1.43 ± 0.08 1.51 ± 0.09 1.54 ± 0.11 0.39 LDL cholesterol, mmol/L 1.98 ± 0.18 2.43 ± 0.20 2.41 ± 0.13b 0.07 Triglycerides, mmol/L 1.07 ± 0.17 0.97 ± 0.18 1.05 ± 0.08c 0.91 Abbreviations: BMI, body mass index; HDL, high-density lipoprotein; NS, not significant. a Old vs young. b P < 0.05 vs Y-NGT. c P < 0.05 vs O-NGT. View Large Table 1. Baseline Subject Characteristics Y-NGT O-NGT O-IGT Pa No. 40 32 21 Age, y 25.4 ± 3.4 71.9 ± 7.2b 76.6 ± 6.7b <0.0001 Sex, female/male 26/14 14/19 11/9 NS BMI, kg/m2 23.8 ± 2.5 25.2 ± 2.9 23.8 ± 2.8 0.16 Hemoglobin A1c, % 5.1 ± 0.3 5.5 ± 0.3 5.6 ± 0.3b <0.0001 Insulin, pmol/L 36.8 ± 4.9 35.9 ± 3.5 44.3 ± 7.8 0.46 Glucose, mmol/L 5.06 ± 0.55 5.36 ± 0.54b 5.52 ± 0.46b 0.001 Total cholesterol, mmol/L 3.89 ± 0.17 4.39 ± 0.23 4.40 ± 0.17b 0.04 HDL cholesterol, mmol/L 1.43 ± 0.08 1.51 ± 0.09 1.54 ± 0.11 0.39 LDL cholesterol, mmol/L 1.98 ± 0.18 2.43 ± 0.20 2.41 ± 0.13b 0.07 Triglycerides, mmol/L 1.07 ± 0.17 0.97 ± 0.18 1.05 ± 0.08c 0.91 Y-NGT O-NGT O-IGT Pa No. 40 32 21 Age, y 25.4 ± 3.4 71.9 ± 7.2b 76.6 ± 6.7b <0.0001 Sex, female/male 26/14 14/19 11/9 NS BMI, kg/m2 23.8 ± 2.5 25.2 ± 2.9 23.8 ± 2.8 0.16 Hemoglobin A1c, % 5.1 ± 0.3 5.5 ± 0.3 5.6 ± 0.3b <0.0001 Insulin, pmol/L 36.8 ± 4.9 35.9 ± 3.5 44.3 ± 7.8 0.46 Glucose, mmol/L 5.06 ± 0.55 5.36 ± 0.54b 5.52 ± 0.46b 0.001 Total cholesterol, mmol/L 3.89 ± 0.17 4.39 ± 0.23 4.40 ± 0.17b 0.04 HDL cholesterol, mmol/L 1.43 ± 0.08 1.51 ± 0.09 1.54 ± 0.11 0.39 LDL cholesterol, mmol/L 1.98 ± 0.18 2.43 ± 0.20 2.41 ± 0.13b 0.07 Triglycerides, mmol/L 1.07 ± 0.17 0.97 ± 0.18 1.05 ± 0.08c 0.91 Abbreviations: BMI, body mass index; HDL, high-density lipoprotein; NS, not significant. a Old vs young. b P < 0.05 vs Y-NGT. c P < 0.05 vs O-NGT. View Large Figure 1. View largeDownload slide (A) Glucose and (B) insulin concentrations and (C) ISRs on OGTT in Y-NGT, O-NGT, and O-IGT nonobese subjects. *P < 0.05 O-IGT vs Y-NGT. §P < 0.05 O-IGT vs O-NGT. Figure 1. View largeDownload slide (A) Glucose and (B) insulin concentrations and (C) ISRs on OGTT in Y-NGT, O-NGT, and O-IGT nonobese subjects. *P < 0.05 O-IGT vs Y-NGT. §P < 0.05 O-IGT vs O-NGT. Indexes of insulin sensitivity The homeostatic model assessment of IR, which represents primarily hepatic insulin sensitivity, was not significantly different between the three groups (Table 2). Peripheral insulin sensitivity, calculated with the Matsuda index, was similar among NGT subjects (i.e., O-NGT and Y-NGT) but was significantly lower in the O-IGT group (P < 0.05 vs Y-NGT) (Table 2). ΔG0–120 was significantly higher in the O-IGT group than in both Y-NGT and O-NGT groups. The O-IGT group displayed a significant increase in glucose concentrations, particularly from 30 to 120 minutes (Table 2, Fig. 1A). Table 2. OGTT Indexes of Insulin Sensitivity and β-Cell Function Y-NGT O-NGT O-IGT Pa Glucose excursions during OGTT  ΔG0–30, mmol/L 39.3 ± 3.0 41.2 ± 3.1 39.6 ± 4.0 0.74  ΔG30–120, mmol/L 129.5 ± 17.2 156.8 ± 19.0 387.0 ± 21.4b,c 0.0001 Insulin sensitivity  HOMA-IR 1.21 ± 0.18 1.26 ± 0.13 1.56 ± 0.27 0.17  Matsuda index 11.0 ± 1.4 8.1 ± 0.7 6.5 ± 0.7b 0.009 Insulin concentrations and secretion during OGTT  ΔI0–30, nmol/L 4.6 ± 0.6 3.6 ± 0.5 2.5 ± 0.5b 0.03  ΔI30–120, nmol/L 24.9 ± 2.6 27.7 ± 3.6 31.0 ± 4.5b 0.48  ΔISR0–30, (nmol/min 22.6 ± 2.2 22.6 ± 2.4 13.7 ± 2.2b,c 0.02  ΔISR30–120, nmol/min 83.8 ± 5.2 101.6 ± 7.8 111.8 ± 8.4b 0.02  (ΔI/ΔG)0–30, pmol/mmol 127 ± 20 102 ± 17 68 ± 16b 0.02  (ΔI/ΔG)30–120, pmol/mmol 578 ± 479 415 ± 119 84 ± 12b,c 0.002  (ΔISR/ΔG)0–30, nmol/min/mmol/L 0.710 ± 0.099 0.683 ± 0.140 0.380 ± 0.073b,c 0.004  (ΔISR/ΔG)30–120, nmol/min/mmol/L 4.327 ± 4.429 1.480 ± 0.403 0.300 ± 0.025b,c <0.0001 β-Cell function  Kd, pmol/m2/mM 4097 ± 1011 3560 ± 1379 1919 ± 508 0.03  Glucose sensitivity, pmol/min/m2/mM 509 ± 55 479 ± 65 266 ± 28b,c 0.04  (ΔI/ΔG)0–30 × Matsuda 1134 ± 176 752 ± 141 358 ± 74b,c <0.0005  (ΔI/ΔG)0–120 × Matsuda 4026 ± 1864 1685 ± 351 462 ± 63b,c <0.001  (ΔISR/ΔG)0–30 × Matsuda 7.9 ± 1.7 5.1 ± 1.1 2.3 ± 0.4b,c <0.0002  (ΔISR/ΔG)0–120 × Matsuda 17.9 ± 5.9 7.4 ± 1.7 1.9 ± 0.3b,c <0.0001 Y-NGT O-NGT O-IGT Pa Glucose excursions during OGTT  ΔG0–30, mmol/L 39.3 ± 3.0 41.2 ± 3.1 39.6 ± 4.0 0.74  ΔG30–120, mmol/L 129.5 ± 17.2 156.8 ± 19.0 387.0 ± 21.4b,c 0.0001 Insulin sensitivity  HOMA-IR 1.21 ± 0.18 1.26 ± 0.13 1.56 ± 0.27 0.17  Matsuda index 11.0 ± 1.4 8.1 ± 0.7 6.5 ± 0.7b 0.009 Insulin concentrations and secretion during OGTT  ΔI0–30, nmol/L 4.6 ± 0.6 3.6 ± 0.5 2.5 ± 0.5b 0.03  ΔI30–120, nmol/L 24.9 ± 2.6 27.7 ± 3.6 31.0 ± 4.5b 0.48  ΔISR0–30, (nmol/min 22.6 ± 2.2 22.6 ± 2.4 13.7 ± 2.2b,c 0.02  ΔISR30–120, nmol/min 83.8 ± 5.2 101.6 ± 7.8 111.8 ± 8.4b 0.02  (ΔI/ΔG)0–30, pmol/mmol 127 ± 20 102 ± 17 68 ± 16b 0.02  (ΔI/ΔG)30–120, pmol/mmol 578 ± 479 415 ± 119 84 ± 12b,c 0.002  (ΔISR/ΔG)0–30, nmol/min/mmol/L 0.710 ± 0.099 0.683 ± 0.140 0.380 ± 0.073b,c 0.004  (ΔISR/ΔG)30–120, nmol/min/mmol/L 4.327 ± 4.429 1.480 ± 0.403 0.300 ± 0.025b,c <0.0001 β-Cell function  Kd, pmol/m2/mM 4097 ± 1011 3560 ± 1379 1919 ± 508 0.03  Glucose sensitivity, pmol/min/m2/mM 509 ± 55 479 ± 65 266 ± 28b,c 0.04  (ΔI/ΔG)0–30 × Matsuda 1134 ± 176 752 ± 141 358 ± 74b,c <0.0005  (ΔI/ΔG)0–120 × Matsuda 4026 ± 1864 1685 ± 351 462 ± 63b,c <0.001  (ΔISR/ΔG)0–30 × Matsuda 7.9 ± 1.7 5.1 ± 1.1 2.3 ± 0.4b,c <0.0002  (ΔISR/ΔG)0–120 × Matsuda 17.9 ± 5.9 7.4 ± 1.7 1.9 ± 0.3b,c <0.0001 Abbreviations: HOMA-IR, homeostatic model assessment of insulin resistance; Kd, rate sensitivity. a Old vs young. b P < 0.05 vs Y-NGT. c P < 0.05 vs O-NGT. View Large Table 2. OGTT Indexes of Insulin Sensitivity and β-Cell Function Y-NGT O-NGT O-IGT Pa Glucose excursions during OGTT  ΔG0–30, mmol/L 39.3 ± 3.0 41.2 ± 3.1 39.6 ± 4.0 0.74  ΔG30–120, mmol/L 129.5 ± 17.2 156.8 ± 19.0 387.0 ± 21.4b,c 0.0001 Insulin sensitivity  HOMA-IR 1.21 ± 0.18 1.26 ± 0.13 1.56 ± 0.27 0.17  Matsuda index 11.0 ± 1.4 8.1 ± 0.7 6.5 ± 0.7b 0.009 Insulin concentrations and secretion during OGTT  ΔI0–30, nmol/L 4.6 ± 0.6 3.6 ± 0.5 2.5 ± 0.5b 0.03  ΔI30–120, nmol/L 24.9 ± 2.6 27.7 ± 3.6 31.0 ± 4.5b 0.48  ΔISR0–30, (nmol/min 22.6 ± 2.2 22.6 ± 2.4 13.7 ± 2.2b,c 0.02  ΔISR30–120, nmol/min 83.8 ± 5.2 101.6 ± 7.8 111.8 ± 8.4b 0.02  (ΔI/ΔG)0–30, pmol/mmol 127 ± 20 102 ± 17 68 ± 16b 0.02  (ΔI/ΔG)30–120, pmol/mmol 578 ± 479 415 ± 119 84 ± 12b,c 0.002  (ΔISR/ΔG)0–30, nmol/min/mmol/L 0.710 ± 0.099 0.683 ± 0.140 0.380 ± 0.073b,c 0.004  (ΔISR/ΔG)30–120, nmol/min/mmol/L 4.327 ± 4.429 1.480 ± 0.403 0.300 ± 0.025b,c <0.0001 β-Cell function  Kd, pmol/m2/mM 4097 ± 1011 3560 ± 1379 1919 ± 508 0.03  Glucose sensitivity, pmol/min/m2/mM 509 ± 55 479 ± 65 266 ± 28b,c 0.04  (ΔI/ΔG)0–30 × Matsuda 1134 ± 176 752 ± 141 358 ± 74b,c <0.0005  (ΔI/ΔG)0–120 × Matsuda 4026 ± 1864 1685 ± 351 462 ± 63b,c <0.001  (ΔISR/ΔG)0–30 × Matsuda 7.9 ± 1.7 5.1 ± 1.1 2.3 ± 0.4b,c <0.0002  (ΔISR/ΔG)0–120 × Matsuda 17.9 ± 5.9 7.4 ± 1.7 1.9 ± 0.3b,c <0.0001 Y-NGT O-NGT O-IGT Pa Glucose excursions during OGTT  ΔG0–30, mmol/L 39.3 ± 3.0 41.2 ± 3.1 39.6 ± 4.0 0.74  ΔG30–120, mmol/L 129.5 ± 17.2 156.8 ± 19.0 387.0 ± 21.4b,c 0.0001 Insulin sensitivity  HOMA-IR 1.21 ± 0.18 1.26 ± 0.13 1.56 ± 0.27 0.17  Matsuda index 11.0 ± 1.4 8.1 ± 0.7 6.5 ± 0.7b 0.009 Insulin concentrations and secretion during OGTT  ΔI0–30, nmol/L 4.6 ± 0.6 3.6 ± 0.5 2.5 ± 0.5b 0.03  ΔI30–120, nmol/L 24.9 ± 2.6 27.7 ± 3.6 31.0 ± 4.5b 0.48  ΔISR0–30, (nmol/min 22.6 ± 2.2 22.6 ± 2.4 13.7 ± 2.2b,c 0.02  ΔISR30–120, nmol/min 83.8 ± 5.2 101.6 ± 7.8 111.8 ± 8.4b 0.02  (ΔI/ΔG)0–30, pmol/mmol 127 ± 20 102 ± 17 68 ± 16b 0.02  (ΔI/ΔG)30–120, pmol/mmol 578 ± 479 415 ± 119 84 ± 12b,c 0.002  (ΔISR/ΔG)0–30, nmol/min/mmol/L 0.710 ± 0.099 0.683 ± 0.140 0.380 ± 0.073b,c 0.004  (ΔISR/ΔG)30–120, nmol/min/mmol/L 4.327 ± 4.429 1.480 ± 0.403 0.300 ± 0.025b,c <0.0001 β-Cell function  Kd, pmol/m2/mM 4097 ± 1011 3560 ± 1379 1919 ± 508 0.03  Glucose sensitivity, pmol/min/m2/mM 509 ± 55 479 ± 65 266 ± 28b,c 0.04  (ΔI/ΔG)0–30 × Matsuda 1134 ± 176 752 ± 141 358 ± 74b,c <0.0005  (ΔI/ΔG)0–120 × Matsuda 4026 ± 1864 1685 ± 351 462 ± 63b,c <0.001  (ΔISR/ΔG)0–30 × Matsuda 7.9 ± 1.7 5.1 ± 1.1 2.3 ± 0.4b,c <0.0002  (ΔISR/ΔG)0–120 × Matsuda 17.9 ± 5.9 7.4 ± 1.7 1.9 ± 0.3b,c <0.0001 Abbreviations: HOMA-IR, homeostatic model assessment of insulin resistance; Kd, rate sensitivity. a Old vs young. b P < 0.05 vs Y-NGT. c P < 0.05 vs O-NGT. View Large Insulin response to a glucose load Plasma glucose concentrations were similar in all subjects in the first 30 minutes of the OGTT (Fig. 1A); glucose concentrations of O-NGT overlapped those of Y-NGT, whereas in O-IGT glucose concentrations were higher (Fig. 1A). The early incremental plasma insulin response (both ΔI0–30 and ΔISR0–30) to the OGTT was reduced only in O-IGT vs Y-NGT (Fig. 1B and 1C), whereas IS from 30 to 120 minutes (ΔI30–120 and ΔISR30–120) was increased in the O-IGT group (Table 2). Because older subjects had a late response in IS, both ΔI0–120 and ΔISR0–120 were similar in the three groups (Table 2). β-Cell function We have calculated rate sensitivity (i.e., the IS in response to changes in glucose concentration) and the β-cell glucose sensitivity (i.e., the slope of the dose response between ISR and glucose excursion) as previously described. The β-cell glucose sensitivity was reduced only in the O-IGT group. Also, the disposition index and the IS/IR index were reduced only in the O-IGT group, whether calculated in the first 30 minutes or throughout the OGTT (Table 2). Incretin secretion GLP-1 secretion was significantly higher in both older groups than in the Y-NGT group (GLP-1 AUC0–120 was 2.8 ± 0.1 in Y-NGT, 3.8 ± 0.5 in O-NGT, and 3.7 ± 0.4 nmol/L·120 minutes in O-IGT; P < 0.05). There was no difference in GLP-1 secretion between older groups (Fig. 2A). GIP secretion was significantly higher in the O-NGT group (AUC0–120 6.0 ± 0.4 nmol/L × 120 minutes) than in the Y-NGT and O-IGT groups (4.7 ± 0.3 and 4.8 ± 0.3 nmol/L × 120 minutes) (Fig. 2B). Even though the main effect of aging on β-cell function is most evident in the first 30 minutes of the OGTT, the GLP-1 AUC in the last 60 minutes of the curve remained significantly higher in O-NGT compared with Y-NGT subjects, suggesting a compensatory response of GLP-1 secretion in O-NGT. We analyzed the relationship between incretin secretion and insulin response by plotting the incremental incretin concentrations in the first 60 minutes vs the incremental ISR. We observed that O-IGT had higher GLP-1 secretion than Y-NGT but significantly lower ISR than both Y-NGT and O-NGT groups (Fig. 2C). O-NGT and Y-NGT had similar ISR response in the first hour, but GIP secretion was higher in O-NGT, perhaps indicating lower sensitivity to GIP in older subjects (Fig. 2D). On the other hand, O-IGT had lower GIP and ISR during the first 60 minutes of the OGTT compared with O-NGT (Fig. 2D). Figure 2. View largeDownload slide (A) GLP-1 and (B) GIP concentrations on OGTT in Y-NGT, O-NGT, and O-IGT nonobese subjects. *P < 0.05 O-IGT vs Y-NGT. §P < 0.05 O-NGT vs Y-NGT. P = ns O-IGT vs O-NGT. Comparison of incremental ISR between 0 and 60 minutes vs incremental (C) GLP-1 and (D) GIP concentrations. #P < 0.05 changes in ΔAUC ISR vs Y-NGT. ¶P < 0.05 changes in ΔAUC GLP-1 or GIP vs Y-NGT. ΔAUC, incremental area under the curve. Figure 2. View largeDownload slide (A) GLP-1 and (B) GIP concentrations on OGTT in Y-NGT, O-NGT, and O-IGT nonobese subjects. *P < 0.05 O-IGT vs Y-NGT. §P < 0.05 O-NGT vs Y-NGT. P = ns O-IGT vs O-NGT. Comparison of incremental ISR between 0 and 60 minutes vs incremental (C) GLP-1 and (D) GIP concentrations. #P < 0.05 changes in ΔAUC ISR vs Y-NGT. ¶P < 0.05 changes in ΔAUC GLP-1 or GIP vs Y-NGT. ΔAUC, incremental area under the curve. Discussion We investigated the relationship between age-related changes in β-cell function with incretin secretion in response to oral glucose. The loss of first-phase IS (first-phase ISR) is one of the earliest abnormalities observed in glucose-intolerant people (25–27). Chen et al. (9) demonstrated that older subjects may lose first-phase ISR and have a delayed insulin response in the first hour after an oral glucose load. In this study, the first-phase insulin response (measured as ΔISR/ΔG and ΔI/ΔG during the first 30 minutes of OGTT; Table 2) was lower in O-IGT but not in O-NGT compared with Y-NGT subjects. Similarly, O-IGT subjects had a lower 0- to 120-minute β-cell response. We investigated whether the impairment in insulin release seen in O-IGT subjects is caused by reduced incretin secretion. Against our prediction, we found that the incretin response was exaggerated in the older subjects (both O-NGT and O-IGT) compared with young. The higher GLP-1 response observed in both older groups might be interpreted as a physiological response to prevent loss of early-phase IS. Nonetheless, this exaggerated GLP-1 response was not sufficient to normalize first-phase IS in the O-IGT. Notably, β-cell sensitivity to glucose was altered only in older subjects with IGT but not in those with NGT. This suggests resistance of the β cells to the incretin effect such that the gut responds to the glucose load with an exaggerated GLP-1 release in an attempt to exert normal IS responses. A negative feedback relationship between insulin and GIP has been proposed to exist (28) but was never convincingly demonstrated, and the influence of IS on GLP-1 secretion is unclear. A deleterious effect of aging per se on β-cell function has not been consistently observed. For example, previous studies that used the hyperglycemic clamp technique have shown little or no decrease in IS, both first and second phase, with aging (29–31). Our results are therefore in line with previous findings showing that when early IS is preserved, older subjects have NGT. The exaggerated GLP-1 response shown in this study is in line with the study from Ranganath et al. (32). Yet others did not find any difference in GLP-1 concentration between younger healthy controls and older healthy controls and older subjects with T2DM (33). Unlike our study, these previous studies did not account for the level of IR (34) and did not exclude subjects with positive family history for T2DM, which is known to affect incretin responses (35). We found that the incretin response to the glucose load is higher in older subjects (NGT and IGT) than in young controls, but β-cell sensitivity to glucose is altered only in older subjects with IGT. This finding indicates a resistance of the β cell to the incretin effect, and therefore the gut may respond to the glucose load with an increased incretin release to obtain similar insulin responses, although the mechanism remains obscure. We propose that with aging the β cell becomes resistant to the incretin effect, thus needing an increased release of GLP-1 and GIP to stimulate adequate IS in response to the glucose load. Conclusions We conclude that the incretin response in older adults is not impaired but rather increased. The insulin response in the older NGT group is similar to that of the young NGT group, suggesting that resistance of the β cell to the incretin effect could contribute to the glucose intolerance seen with aging. Abbreviations: Abbreviations: AUC area under the curve GIP glucose-dependent insulinotropic peptide GLP-1 glucagon-like peptide 1 IGT impaired glucose tolerance IR insulin resistance IS insulin secretion ISR insulin secretory rate LDL low-density lipoprotein NGT normal glucose tolerance OGTT oral glucose tolerance test O-IGT older subjects with impaired glucose tolerance O-NGT older subjects with normal glucose tolerance T2DM type 2 diabetes mellitus Y-NGT young subjects with normal glucose tolerance Acknowledgments Financial Support: This work was supported by grants from the National Institutes of Health (R01-DK80157 and R01-DK089229) and the American Diabetes Association to N.M. This work also was supported by grants UL1TR000149 (CTSA), AG044271 (San Antonio Claude D. Pepper Older Americans Independence Center), and AG013319 (San Antonio Nathan Shock Center) and grants from the NovoNordisk Foundation (J.J.H.). A.G. has received research funds from the Italian Ministry of Research (MIUR) (Consiglio Nazionale delle Ricerche) for “Progetto Premiale” and “Ageing Project.” Author Contributions: J.d.J.G.-G. and R.L. performed the studies. J.d.J.G.-G., A.G., J.J.H., and N.M. analyzed the data and wrote the manuscript. J.J.H. and R.A.D. contributed to revising and reviewing the manuscript. N.M. is the guarantor of this work and takes full responsibility for the integrity of information and concepts presented in the manuscript. Disclosure Summary: J.d.J.G.-G. is a speaker for Novo-Nordisk, Sanofi Aventis, AstraZeneca, Boheringer Ingelheim, and Janssen. A.G. is a consultant for Eli-Lilly, Menarini, Gilead, Inventiva, and Sanofi. R.A.D. is a member of the Advisory Board of Takeda, Bristol Myers Squibb, Janssen, Boehringer Ingelheim, Novo Nordisk, and Amylin; he is a member of the Speaker Bureau of Novo Nordisk, Amylin, BMS, and Janssen; and he has grant support from Takeda, Amylin, and BMS. The salaries of N.M. and R.A.D. are paid in part by the South Texas Veterans Healthcare System. J.J.H. is a member of advisory boards for MSD and NovoNordisk. The remaining author has nothing to disclose. References 1. Scheen AJ . Diabetes mellitus in the elderly: insulin resistance and/or impaired insulin secretion ? Diabetes Metab . 2005 ; 31 : 5S27 – 5S34 . 2. Meigs JB , Muller DC , Nathan DM , Blake DR , Andres R ; Baltimore Longitudinal Study of Aging . The natural history of progression from normal glucose tolerance to type 2 diabetes in the Baltimore Longitudinal Study of Aging . Diabetes . 2003 ; 52 ( 6 ): 1475 – 1484 . 3. Kirkman MS , Briscoe VJ , Clark N , Florez H , Haas LB , Halter JB , Huang ES , Korytkowski MT , Munshi MN , Odegard PS , Pratley RE , Swift CS . Diabetes in older adults . Diabetes Care . 2012 ; 35 ( 12 ): 2650 – 2664 . 4. Lee CC , Glickman SG , Dengel DR , Brown MD , Supiano MA . Abdominal adiposity assessed by dual energy X-ray absorptiometry provides a sex-independent predictor of insulin sensitivity in older adults . J Gerontol A Biol Sci Med Sci . 2005 ; 60 ( 7 ): 872 – 877 . 5. Peterson MJ , Morey MC , Giuliani C , Pieper CF , Evenson KR , Mercer V , Visser M , Brach JS , Kritchevsky SB , Goodpaster BH , Rubin S , Satterfield S , Simonsick EM ; Health ABC Study . Walking in old age and development of metabolic syndrome: the health, aging, and body composition study . Metab Syndr Relat Disord . 2010 ; 8 ( 4 ): 317 – 322 . 6. Fink RI , Kolterman OG , Griffin J , Olefsky JM . Mechanisms of insulin resistance in aging . J Clin Invest . 1983 ; 71 ( 6 ): 1523 – 1535 . 7. Garcia GVFR , Freeman RV , Supiano MA , Smith MJ , Galecki AT , Halter JB . Glucose metabolism in older adults: a study including subjects more than 80 years of age . J Am Geriatr Soc . 1997 ; 45 ( 7 ): 813 – 817 . 8. Weir GC . Islet-cell biology in 2015: understanding secretion, ageing and death in β cells . Nat Rev Endocrinol . 2016 ; 12 ( 2 ): 72 – 74 . 9. Chen M , Bergman RN , Pacini G , Porte D Jr . Pathogenesis of age-related glucose intolerance in man: insulin resistance and decreased beta-cell function . J Clin Endocrinol Metab . 1985 ; 60 ( 1 ): 13 – 20 . 10. Pacini G , Beccaro F , Valerio A , Nosadini R , Crepaldi G . Reduced beta-cell secretion and insulin hepatic extraction in healthy elderly subjects . J Am Geriatr Soc . 1990 ; 38 ( 12 ): 1283 – 1289 . 11. Rahier J , Guiot Y , Goebbels RM , Sempoux C , Henquin JC . Pancreatic beta-cell mass in European subjects with type 2 diabetes . Diabetes Obes Metab . 2008 ; 10 ( suppl 4 ): 32 – 42 . 12. Saisho Y , Butler AE , Manesso E , Elashoff D , Rizza RA , Butler PC . β-cell mass and turnover in humans: effects of obesity and aging . Diabetes Care . 2013 ; 36 ( 1 ): 111 – 117 . 13. Ritzel R , Schulte M , Pørksen N , Nauck MS , Holst JJ , Juhl C , März W , Schmitz O , Schmiegel WH , Nauck MA . Glucagon-like peptide 1 increases secretory burst mass of pulsatile insulin secretion in patients with type 2 diabetes and impaired glucose tolerance . Diabetes . 2001 ; 50 ( 4 ): 776 – 784 . 14. Holst JJVT , Vilsbøll T , Deacon CF . The incretin system and its role in type 2 diabetes mellitus . Mol Cell Endocrinol . 2009 ; 297 ( 1–2 ): 127 – 136 . 15. Brubaker PL . Minireview: update on incretin biology: focus on glucagon-like peptide-1 . Endocrinology . 2010 ; 151 ( 5 ): 1984 – 1989 . 16. Muscelli E , Mari A , Casolaro A , Camastra S , Seghieri G , Gastaldelli A , Holst JJ , Ferrannini E . Separate impact of obesity and glucose tolerance on the incretin effect in normal subjects and type 2 diabetic patients . Diabetes . 2008 ; 57 ( 5 ): 1340 – 1348 . 17. Tai MM . A mathematical model for the determination of total area under glucose tolerance and other metabolic curves . Diabetes Care . 1994 ; 17 ( 2 ): 152 – 154 . 18. Matthews DR , Hosker JP , Rudenski AS , Naylor BA , Treacher DF , Turner RC . Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man . Diabetologia . 1985 ; 28 ( 7 ): 412 – 419 . 19. Matsuda M , DeFronzo RA . Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp . Diabetes Care . 1999 ; 22 ( 9 ): 1462 – 1470 . 20. Van Cauter E , Mestrez F , Sturis J , Polonsky KS . Estimation of insulin secretion rates from C-peptide levels. Comparison of individual and standard kinetic parameters for C-peptide clearance . Diabetes . 1992 ; 41 ( 3 ): 368 – 377 . 21. DeFronzo RA , Tripathy D , Abdul-Ghani M , Musi N , Gastaldelli A . The disposition index does not reflect β-cell function in IGT subjects treated with pioglitazone . J Clin Endocrinol Metab . 2014 ; 99 ( 10 ): 3774 – 3781 . 22. Gastaldelli A , Brodows RG , D’Alessio D . The effect of chronic twice daily exenatide treatment on β-cell function in new onset type 2 diabetes . Clin Endocrinol (Oxf) . 2014 ; 80 ( 4 ): 545 – 553 . 23. Salehi M , Gastaldelli A , D’Alessio DA . Altered islet function and insulin clearance cause hyperinsulinemia in gastric bypass patients with symptoms of postprandial hypoglycemia . J Clin Endocrinol Metab . 2014 ; 99 ( 6 ): 2008 – 2017 . 24. Gastaldelli A , Ferrannini E , Miyazaki Y , Matsuda M , DeFronzo RA ; San Antonio Metabolism Study . Beta-cell dysfunction and glucose intolerance: results from the San Antonio Metabolism (SAM) study . Diabetologia . 2004 ; 47 ( 1 ): 31 – 39 . 25. Osei K , Gaillard T , Schuster DP . Pathogenetic mechanisms of impaired glucose tolerance and type II diabetes in African-Americans. The significance of insulin secretion, insulin sensitivity, and glucose effectiveness . Diabetes Care . 1997 ; 20 ( 3 ): 396 – 404 . 26. Weyer C , Bogardus C , Pratley RE . Metabolic characteristics of individuals with impaired fasting glucose and/or impaired glucose tolerance . Diabetes . 1999 ; 48 ( 11 ): 2197 – 2203 . 27. Festa A , D’Agostino R Jr , Hanley AJ , Karter AJ , Saad MF , Haffner SM . Differences in insulin resistance in nondiabetic subjects with isolated impaired glucose tolerance or isolated impaired fasting glucose . Diabetes . 2004 ; 53 ( 6 ): 1549 – 1555 . 28. Creutzfeldt W , Talaulicar M , Ebert R , Willms B . Inhibition of gastric inhibitory polypeptide (GIP) release by insulin and glucose in juvenile diabetes . Diabetes . 1980 ; 29 ( 2 ): 140 – 145 . 29. Elahi D , Muller DC , McAloon-Dyke M , Tobin JD , Andres R . The effect of age on insulin response and glucose utilization during four hyperglycemic plateaus . Exp Gerontol . 1993 ; 28 ( 4-5 ): 393 – 409 . 30. Ahrén B , Pacini G . Age-related reduction in glucose elimination is accompanied by reduced glucose effectiveness and increased hepatic insulin extraction in man . J Clin Endocrinol Metab . 1998 ; 83 ( 9 ): 3350 – 3356 . 31. Bourey RE , Kohrt WM , Kirwan JP , Staten MA , King DS , Holloszy JO . Relationship between glucose tolerance and glucose-stimulated insulin response in 65-year-olds . J Gerontol . 1993 ; 48 ( 4 ): M122 – M127 . 32. Ranganath L , Sedgwick I , Morgan L , Wright J , Marks V . The ageing entero-insular axis . Diabetologia . 1998 ; 41 ( 11 ): 1309 – 1313 . 33. Korosi J , McIntosh CH , Pederson RA , Demuth HU , Habener JF , Gingerich R , Egan JM , Elahi D , Meneilly GS . Effect of aging and diabetes on the enteroinsular axis . J Gerontol A Biol Sci Med Sci . 2001 ; 56 ( 9 ): M575 – M579 . 34. Rask E , Olsson T , Söderberg S , Johnson O , Seckl J , Holst JJ , Ahrén B ; Northern Sweden Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) . Impaired incretin response after a mixed meal is associated with insulin resistance in nondiabetic men . Diabetes Care . 2001 ; 24 ( 9 ): 1640 – 1645 . 35. Matikainen N , Bogl LH , Hakkarainen A , Lundbom J , Lundbom N , Kaprio J , Rissanen A , Holst JJ , Pietiläinen KH . GLP-1 responses are heritable and blunted in acquired obesity with high liver fat and insulin resistance . Diabetes Care . 2014 ; 37 ( 1 ): 242 – 251 . Copyright © 2018 Endocrine Society

Dempster, David W;Brown, Jacques P;Fahrleitner-Pammer, Astrid;Kendler, David;Rizzo, Sebastien;Valter, Ivo;Wagman, Rachel B;Yin, Xiang;Yue, Susan V;Boivin, Georges

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02669pmid: 29672714

Abstract Context Denosumab is a potent antiresorptive agent that reduces fractures in postmenopausal women with osteoporosis. Objective Determine effects of up to 10 years of denosumab on bone histology, remodeling, and matrix mineralization characteristics. Design and Setting International, multicenter, randomized, double-blind trial [Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months (FREEDOM)] with a long-term open-label extension. Patients Postmenopausal women with osteoporosis (92 women in FREEDOM, 46 in extension) who provided iliac bone biopsies, including 11 who provided biopsies at multiple time points. Interventions FREEDOM subjects were randomized 1:1 to subcutaneous denosumab 60 mg or placebo every 6 months for 3 years. Long-term extension subjects continued receiving denosumab, open-label, for 7 additional years. Outcomes Bone histology, histomorphometry, matrix mineralization. Results Ten-year denosumab biopsies showed normal histology. Bone histomorphometry indicated normal bone structure and reduced bone remodeling after 10 years of denosumab, similar to levels after 2 and/or 3 and 5 years of denosumab. The degree of mineralization of bone was increased and mineralization heterogeneity was reduced in the denosumab years 2/3 group vs placebo. Changes in these mineralization variables progressed from years 2/3 to year 5 of denosumab, but not thereafter. Conclusions Denosumab for 2/3, 5, and 10 years was associated with normal histology, low bone remodeling rate, increased matrix mineralization, and lower mineralization heterogeneity compared with placebo. These variables were unchanged from year 5 to year 10. These data, in combination with the maintenance of low fracture rates for up to 10 years as previously reported with denosumab therapy, suggest that strong, prolonged remodeling inhibition does not impair bone strength. Denosumab is a potent antiresorptive agent that inhibits RANKL, an essential cytokine for osteoclast formation, activity, and survival (1). The pivotal 3-year placebo-controlled Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months (FREEDOM) fracture trial in postmenopausal women with osteoporosis showed that denosumab increased bone mineral density (BMD), decreased biochemical markers of bone remodeling, and reduced the risk of new vertebral, nonvertebral, and hip fractures (2). Bone histomorphometry data from iliac crest biopsies collected at years 2 and 3 of FREEDOM indicated marked reductions in bone remodeling parameters with denosumab, with a majority of biopsies exhibiting no fluorochrome labeling in cancellous bone (3). These findings are consistent with denosumab’s mechanism of action; rapid and strong osteoclast inhibition leads to reduced remodeling activation, and most remodeling sites that were active when denosumab therapy was initiated would refill with mineralized matrix before the scheduled administration of fluorochrome labeling agents that are used to identify sites of active bone formation (4). The open-label FREEDOM Extension study showed that years 5 to 10 of denosumab therapy were associated with further increases in BMD at the lumbar spine, total hip, and femoral neck; persistently low levels of biochemical and histomorphometric bone remodeling variables; and continued low rates of vertebral, hip, and nonvertebral fractures (5–9). This association between very low bone remodeling rates and reduced fracture risk with denosumab is consistent with nonclinical bone quality studies showing that denosumab increased vertebral, hip, and long bone strength, with the greatest bone strength consistently associated with the lowest levels of bone remodeling and fluorochrome labeling (4, 10). These results notwithstanding, concerns have been raised that long-term administration of potent antiresorptive agents, including denosumab, may lead to “oversuppression” of bone remodeling that results in impaired bone matrix quality and strength (11–13). One hypothetical mechanism by which oversuppression could impair bone biomechanical properties is through microdamage accumulation (14), although experimental evidence suggests that this mechanism is not an apparent concern with denosumab. OPG-Fc, a RANKL inhibitor with a similar mechanism of action as denosumab (15), significantly reduced microdamage levels in normal and fatigue-damaged bone (16). Although greater microdamage levels are sometimes associated with lower bone toughness (17), long-term denosumab administration did not reduce toughness in nonhuman primates, despite near-total inhibition of bone remodeling (10). Another proposed mechanism by which oversuppression could potentially impair bone biomechanics is by altering matrix mineralization (14). Antiresorptive agents, including bisphosphonates (BPs) and RANKL inhibitors (18–22), increase the degree of mineralization of bone (DMB) by affording bone remodeling units more time to mineralize (18). This effect of antiresorptives also typically leads to reductions in the heterogeneity of mineralization within bone, because more bone regions are able to achieve a higher degree of mineralization (19–21, 23). The increases in DMB that result from antiresorptive therapy contribute to gains in BMD and may also contribute to reductions in fracture risk by increasing bone matrix strength and stiffness (18, 19, 24–26). However, it has also been postulated that excessive increases in bone matrix mineralization (sometimes referred to as “hypermineralization”) and/or excessive reductions in mineralization heterogeneity could lead to bone brittleness and skeletal fragility (27–29). There is no operational definition of hypermineralization or insufficient heterogeneity, and there is little experimental evidence that antiresorptive agents can alter mineralization characteristics to the point of increasing skeletal fragility. Substantial to near-total inhibition of bone remodeling in animals treated with denosumab or other RANKL inhibitors was accompanied by increases in matrix mineralization and/or reductions in mineralization heterogeneity; these changes were associated with improvements in bone structural strength, without any impairments in bone material properties (21–23). However, several evidence gaps remain in this area of research, including a paucity of data on the long-term effects of denosumab or other antiresorptive agents on matrix mineralization variables, and minimal data on associations between treatment-related changes in matrix mineralization and long-term fracture outcomes. Long- and short-term data on the effects of denosumab on matrix mineralization may be of interest because of denosumab’s rapid and strong antiresorptive effects throughout the skeleton, including cancellous and cortical compartments (3, 4). The FREEDOM trial and its extension provided a unique opportunity to assess (1) the degree to which human bone matrix can increase its mineralization and decrease its mineralization heterogeneity as a result of denosumab treatment and (2) the potential implications of such changes on fracture rates and skeletal adverse events over an extended period of uninterrupted denosumab treatment. The current analyses from FREEDOM and its extension include bone histology, dynamic and static bone histomorphometry, and bone matrix mineralization characteristics with up to 10 years of denosumab therapy. These results are interpreted and discussed in the context of recently published data on fracture rates and other bone safety parameters in the FREEDOM extension study with up to 10 years of denosumab treatment (9). Subjects and Methods Study subjects Subjects included in this study were enrolled in the FREEDOM trial and its extension, details of which have been previously described (2, 6). Briefly, FREEDOM enrolled 7808 women aged 60 to 91 years with median BMD T scores of –2.9 at the lumbar spine and –1.9 at the total hip. All women who completed FREEDOM and did not miss more than one dose of study drug were eligible to enter the extension. Subjects were eligible to enroll in the bone biopsy substudy if they were enrolled at a clinical trial center that was participating in the bone biopsy substudy and had no sensitivity to tetracycline or its derivatives. Study design FREEDOM was a 3-year international, randomized, double-blind, placebo-controlled trial in postmenopausal women with osteoporosis (Supplemental Fig. 1) (2). Subjects received subcutaneous denosumab 60 mg or placebo every 6 months for 3 years, along with daily calcium (≥1000 mg) and vitamin D (≥400 IU) supplementation. All subjects enrolled in the FREEDOM extension were to receive open-label denosumab 60 mg every 6 months for 7 additional years. The current extension data are limited to subjects who received denosumab during FREEDOM (referred to as the long-term group), and do not include subjects who received placebo during FREEDOM (the cross-over group). Subsets of subjects underwent transiliac bone biopsies at year 2 or 3 in FREEDOM (encompassing 2 or 3 years of placebo or denosumab) and/or at year 2 or year 7 of the extension (encompassing 5 or 10 years of denosumab for subjects in the long-term group). Eleven subjects in the long-term denosumab group provided biopsies at more than one time point: one subject for denosumab year 2 and year 3; three subjects for denosumab year 3 and year 5; two subjects for denosumab year 2, year 3, and year 5; one subject for denosumab year 3, year 5, and year 10; one subject for denosumab year 3 and year 10; and three subjects for denosumab year 5 and year 10. Five subjects in the placebo group provided biopsies at year 2 and year 3. Only the year 3 data were used for subjects who had both year 2 and year 3 biopsy samples. These sequential biopsies, and a few biopsies that were not evaluable for all end points, explain occasional differences among numbers of subjects, samples, and observations. Bone biopsy procedures The methodology for bone biopsy procedure and analyses was similar to previously described approaches (3). Briefly, bone biopsies were obtained from the iliac crest near the end of the 6-monthly dosing interval, within 56 days of the year 2 and/or year 3 visit (for subjects in FREEDOM) or year 5 and/or year 10 visits (year 2 and year 7 of the extension, for the long-term group). A standard double-labeling procedure was used to identify sites undergoing primary mineralization of newly formed bone, as previously described (3). Briefly, tetracycline was administered on 3 successive days, followed 14 days later by the administration of demeclocycline on three successive days, followed by biopsy 5 to 14 days after the last demeclocycline dose. Urine samples for tetracycline measurements were collected within 24 hours of the last dose of the first tetracycline labeling period to confirm compliance. Bone biopsies were obtained from the anterior iliac crest using a Bordier/Meunier or Rochester-type trephine with internal diameter of 7 to 8 mm. Specimens were fixed and shipped in 70% ethanol and then dehydrated and embedded undecalcified in glycol methylmethacrylate (GMMA). Bone histology and histomorphometry analyses GMMA-embedded biopsy samples were sectioned at a 5-μm thickness at a central facility (Mayo Clinic, Rochester, MN) and mounted unstained for analyses of tetracycline labels under fluorescent microscopy. If no labels were present, a label search was performed at 5-μm intervals, as previously described (3). If fluorescent labels were present, adjacent sections were stained with toluidine blue or hematoxylin and eosin for qualitative analysis by a hematopathologist, and with Goldner trichrome stain for histomorphometric analyses of static parameters. Osteomeasure was used for histomorphometric analyses, using American Society for Bone and Mineral Research nomenclature (30). Where single fluorochrome labels were identified without evident double labels, a mineral apposition rate of 0.3 μm was imputed, per American Society for Bone and Mineral Research guidelines (30). Bone matrix mineralization analyses GMMA-embedded bone biopsy samples were cross-sectioned at 150-μm thickness and thinned to 100 ± 1 μm thickness by manual grinding between a frosted glass plate and a frosted glass slide using silicon carbide powder (Escil, Chassieu, France). Sections were then polished using a 1 μm alumina suspension (Escil) and cleaned with an ultrasonic device (Elma, Singen, Germany). Section thickness was measured with a precision micrometric thickness comparator (precision of 1 μm; Compac, Geneva, Switzerland). Bone sections were analyzed for matrix mineralization by digitized quantitative microradiography in a blinded manner, using code from a MATLAB program, as previously described (31). With this method, quantitative X-ray absorption by bone tissue is reflected in grayscale values that are converted at the pixel level into a DMB, expressed in grams of mineral per cm3 of bone tissue. DMB reflects the density of hydroxyapatite, the mineral component of bone matrix. This conversion was based on a calibration curve generated from an aluminum reference system with a known absorption coefficient. Regions with lower matrix mineralization have a darker grayscale appearance compared with clearer, more highly mineralized regions. Five regions of bone were analyzed: the cancellous region, the total cortical region (endocortical plus periosteal combined), the endocortical subregion, the periosteal subregion, and the total bone region (cancellous plus cortical). In most cases, both cortices were analyzed, and results for each cortex were averaged. DMB was measured for each bone region and subregion, as was the heterogeneity index (HI), which reflects the width of the DMB distribution curve at one-half of its maximum height (25, 31). Limited data on total bone DMB and HI for the denosumab years 2/3, 5, and 10 groups were reported previously (9). Iliac crest bone biopsies were also obtained from 42 nonosteoporotic untreated premenopausal women (mean age, 33.4 years; SD, 4.8 years; range, 25 to 41 years) from a previously described cohort (32, 33). These patients served as a premenopausal reference range for matrix mineralization analyses of the cancellous, cortical, and total bone compartments; the endocortical and periosteal subregions of this reference group were not analyzed. Statistical analyses For subjects who had both year 2 and year 3 biopsy samples in FREEDOM, only the year 3 samples were included in the calculation of group statistics. Comparisons for histology and histomorphometric data between FREEDOM placebo and FREEDOM denosumab data at years 2 and 3 combined have been published (3). For the histomorphometric parameters and bone matrix mineralization variables (DMB and HI), pairwise comparisons were performed among the four treatment groups (placebo years 2/3, denosumab years 2/3, denosumab year 5, and denosumab year 10). For DMB and HI, each of the four treatment groups was also compared with a premenopausal reference group. Two-sided Wilcoxon rank sum test was used for all comparisons between two groups without multiplicity adjustment. Results There were 21 and 22 biopsies available for histomorphometry and qualitative histology (respectively) for subjects from the long-term arm of FREEDOM extension with 10 years of denosumab exposure. Bone biopsy samples from 72 women in FREEDOM (30 placebo and 42 denosumab subjects at year 2 or 3), and 28 and 21 women in the extension who had received denosumab for a total of 5 and/or 10 years, respectively, were evaluated for matrix mineralization; four of these subjects provided more than one biopsy, and one biopsy was not evaluable for mineralization analyses. Baseline characteristics were similar for women in FREEDOM, the histology and histomorphometry substudy, and the bone matrix mineralization substudy (Supplemental Table 1). Histology data for the FREEDOM biopsies (placebo and denosumab years 2/3) and the denosumab year 5 biopsies showed no adverse histopathological findings (Table 1), as previously described (3, 5). Similarly, all 22 of the denosumab year 10 biopsies evaluated for histology showed normally mineralized lamellar bone, with no evidence of pathological findings including osteomalacia, woven bone, or marrow fibrosis (Table 1). The percentage of samples with any fluorochrome labeling of trabecular bone was observed to increase over time in the denosumab samples, from 34% at years 2/3 to 43% at year 5 to 77% at year 10 (Fig. 1). Cortical labeling was evident in most denosumab samples at all three time points, with no meaningful changes over time. Double fluorochrome labeling of trabecular or cortical bone was found in 7 (32%) denosumab year 10 samples (Supplemental Table 2). Additional details on fluorochrome labeling status are provided in Supplemental Table 2. Table 1. Bone Histology and Histopathology FREEDOM Extension Placebo Year 2 and/or 3 Denosumab Year 2 and/or 3 Denosumab Year 5 Denosumab Year 10 N = 45a N = 47a N = 28b N = 22b Evaluable biopsiesc 62 53 28 22 Normal lamellar bone, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Normal mineralization, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Present osteoid, n (%) 62 (100) 48 (91) 23 (82) 18 (82) No visible osteoid, n (%) 0 (0) 5 (9.4) 5 (17.9) 4 (18.2) Osteomalacia, n 0 0 0 0 Marrow fibrosis, n 0 0 0 0 Woven bone, n 0 0 0 0 FREEDOM Extension Placebo Year 2 and/or 3 Denosumab Year 2 and/or 3 Denosumab Year 5 Denosumab Year 10 N = 45a N = 47a N = 28b N = 22b Evaluable biopsiesc 62 53 28 22 Normal lamellar bone, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Normal mineralization, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Present osteoid, n (%) 62 (100) 48 (91) 23 (82) 18 (82) No visible osteoid, n (%) 0 (0) 5 (9.4) 5 (17.9) 4 (18.2) Osteomalacia, n 0 0 0 0 Marrow fibrosis, n 0 0 0 0 Woven bone, n 0 0 0 0 a Number of subjects who enrolled in the FREEDOM bone biopsy substudy, received ≥1 dose of investigational product during FREEDOM, and had an evaluable biopsy at year 2 or year 3. b Number of subjects who enrolled in the extension bone biopsy substudy, received ≥1 dose of investigational product during the extension, and had an evaluable biopsy at the time point(s) of interest. c Number of evaluable biopsies, which serves as the denominator for percentage values in parentheses; some subjects had ≥1 evaluable biopsy during the FREEDOM trial. View Large Table 1. Bone Histology and Histopathology FREEDOM Extension Placebo Year 2 and/or 3 Denosumab Year 2 and/or 3 Denosumab Year 5 Denosumab Year 10 N = 45a N = 47a N = 28b N = 22b Evaluable biopsiesc 62 53 28 22 Normal lamellar bone, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Normal mineralization, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Present osteoid, n (%) 62 (100) 48 (91) 23 (82) 18 (82) No visible osteoid, n (%) 0 (0) 5 (9.4) 5 (17.9) 4 (18.2) Osteomalacia, n 0 0 0 0 Marrow fibrosis, n 0 0 0 0 Woven bone, n 0 0 0 0 FREEDOM Extension Placebo Year 2 and/or 3 Denosumab Year 2 and/or 3 Denosumab Year 5 Denosumab Year 10 N = 45a N = 47a N = 28b N = 22b Evaluable biopsiesc 62 53 28 22 Normal lamellar bone, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Normal mineralization, n (%) 62 (100) 53 (100) 28 (100) 22 (100) Present osteoid, n (%) 62 (100) 48 (91) 23 (82) 18 (82) No visible osteoid, n (%) 0 (0) 5 (9.4) 5 (17.9) 4 (18.2) Osteomalacia, n 0 0 0 0 Marrow fibrosis, n 0 0 0 0 Woven bone, n 0 0 0 0 a Number of subjects who enrolled in the FREEDOM bone biopsy substudy, received ≥1 dose of investigational product during FREEDOM, and had an evaluable biopsy at year 2 or year 3. b Number of subjects who enrolled in the extension bone biopsy substudy, received ≥1 dose of investigational product during the extension, and had an evaluable biopsy at the time point(s) of interest. c Number of evaluable biopsies, which serves as the denominator for percentage values in parentheses; some subjects had ≥1 evaluable biopsy during the FREEDOM trial. View Large Figure 1. View largeDownload slide Percentage of bone biopsies with any fluorochrome on trabecular, cortical, and trabecular or cortical bone. n, number of biopsies with any label; N, number of evaluable biopsies. Figure 1. View largeDownload slide Percentage of bone biopsies with any fluorochrome on trabecular, cortical, and trabecular or cortical bone. n, number of biopsies with any label; N, number of evaluable biopsies. Histomorphometric analyses for FREEDOM and year 2 of the extension (5 years of denosumab) were presented previously (3, 5) and are included in Fig. 2 and Supplemental Table 3 for context. Year 10 histomorphometric data showed that the antiresorptive effect of denosumab was maintained over time, with no significant differences in cancellous eroded surface or osteoclast number compared with values for denosumab subjects at years 2/3 or year 5 (Fig. 2). Cancellous bone volume per tissue volume (BV/TV) was higher and eroded surface and osteoid width were lower in the year 10 samples compared with the years 2/3 placebo samples, with no significant differences for these variables between denosumab years 2/3, year 5, or year 10 (Fig. 2). Cancellous wall thickness was significantly lower in the denosumab year 10 group compared with the other three groups (Fig. 2). Cortical width remained similar in the denosumab samples over time, as did trabecular thickness (Fig. 2; Supplemental Table 3). Trabecular number was higher and trabecular separation was lower in the denosumab year 10 samples compared with placebo, denosumab years 2/3, and denosumab year 5 (Supplemental Table 3). The dynamic parameters mineralizing surface, mineral apposition rate, bone formation rate per bone volume, and activation frequency (9) were significantly lower in the year 10 denosumab group compared with placebo, but similar to values from the denosumab years 2/3 and year 5 groups (Fig. 2; Supplemental Table 3). Osteoid surface was lower in the year 5 and year 10 denosumab samples compared with the years 2/3 denosumab samples, whereas the percentages of denosumab biopsies that showed any osteoid remained similar over time (range, 82% to 91%) (Supplemental Table 4). Several samples that lacked visible osteoid did display tetracycline labels, indicating that the absence of visible osteoid does not necessarily mean lack of bone formation (Supplemental Table 4). Figure 2. View largeDownload slide Bone histomorphometry results for iliac crest bone biopsies from FREEDOM and the extension. All parameters except cortical width were obtained from cancellous bone. Data represent median and interquartile range, n = number of subjects with observed data; N = number of randomized subjects who enrolled in the bone biopsy substudy who received at least 1 dose of investigational product during FREEDOM (for the FREEDOM groups) and during extension (for the extension groups), and had at least one evaluable biopsy. *P < 0.05 vs placebo years 2/3; †P < 0.05 vs denosumab year 2/3; §P < 0.05 vs denosumab year 5, by two-sided Wilcoxon rank-sum test. BFR, bone formation rate; TV, tissue volume. Figure 2. View largeDownload slide Bone histomorphometry results for iliac crest bone biopsies from FREEDOM and the extension. All parameters except cortical width were obtained from cancellous bone. Data represent median and interquartile range, n = number of subjects with observed data; N = number of randomized subjects who enrolled in the bone biopsy substudy who received at least 1 dose of investigational product during FREEDOM (for the FREEDOM groups) and during extension (for the extension groups), and had at least one evaluable biopsy. *P < 0.05 vs placebo years 2/3; †P < 0.05 vs denosumab year 2/3; §P < 0.05 vs denosumab year 5, by two-sided Wilcoxon rank-sum test. BFR, bone formation rate; TV, tissue volume. Representative images depicting the degree and heterogeneity of matrix mineralization are shown in Fig. 3, which highlights single cortices from digitized microradiographs of samples that have cortical DMB and HI values similar to their group median. The heterogeneity of mineralization is visually reflected by the proportion and distribution of darker, less-mineralized osteons relative to lighter, more-mineralized osteons and interstitial bone. Quantified HI and DMB results show that between-group differences for the total bone region were generally reflective of changes observed in the cancellous, cortical, endocortical, and periosteal subregions (Fig. 4). DMB was significantly greater in the denosumab years 2/3 vs placebo group at years 2/3 for the total bone and for each subregion. The year 5 denosumab samples showed significantly greater DMB for total bone and for each subregion compared with the years 2/3 denosumab and placebo samples. For denosumab year 5, the median DMB value for total bone was 1.132 g/cm3 [interquartile range (IQR), 1.110 to 1.150], which was 7.3% higher compared with the median value of the placebo group (1.055 g/cm3; IQR, 1.034 to 1.070). DMB for total bone in the year 10 denosumab group (1.135 g/cm3; IQR, 1.122 to 1.152) was similar to that of the year 5 group, and significantly higher (by 7.6%) compared with the placebo group (P < 0.05). Compared with the premenopausal reference group, DMB values for total, cortical, and cancellous bone were significantly lower in the placebo group and significantly higher in the denosumab year 5 and year 10 groups (Fig. 4). Figure 3. View largeDownload slide Digitized microradiographs of single cortices of iliac crest bone biopsies representative of placebo (PBO) year 2 group and denosumab (DMAb) years 2, 5, and 10. Samples were selected based on cortical DMB and HI values (bottom) similar to their respective group’s median values (in parentheses). For the year 2 PBO and DMAb samples, the group median values (in parentheses) represent values for years 2 and 3 combined. Figure 3. View largeDownload slide Digitized microradiographs of single cortices of iliac crest bone biopsies representative of placebo (PBO) year 2 group and denosumab (DMAb) years 2, 5, and 10. Samples were selected based on cortical DMB and HI values (bottom) similar to their respective group’s median values (in parentheses). For the year 2 PBO and DMAb samples, the group median values (in parentheses) represent values for years 2 and 3 combined. Figure 4. View largeDownload slide DMB and HI for iliac crest bone biopsies. Note that the y-axis scales are truncated. For the box-and-whisker plots, the box’s lower bound represents the first quartile (Q1), its upper bound represents the third quartile (Q3), the line is the median, the diamond is the mean, and the circles are outliers. The gray bands with dashed lines represent the interquartile range (Q1–Q3) and median value, respectively, for the premenopausal reference group (n = 42), from which endocortical and periosteal subcompartment mineralization data were not obtained. *P < 0.05 vs placebo years 2/3, †P < 0.05 vs denosumab years 2/3, §P < 0.05 vs premenopausal reference group, by two-sided Wilcoxon rank-sum test for between-group comparisons. n = number of subjects with observed data. Figure 4. View largeDownload slide DMB and HI for iliac crest bone biopsies. Note that the y-axis scales are truncated. For the box-and-whisker plots, the box’s lower bound represents the first quartile (Q1), its upper bound represents the third quartile (Q3), the line is the median, the diamond is the mean, and the circles are outliers. The gray bands with dashed lines represent the interquartile range (Q1–Q3) and median value, respectively, for the premenopausal reference group (n = 42), from which endocortical and periosteal subcompartment mineralization data were not obtained. *P < 0.05 vs placebo years 2/3, †P < 0.05 vs denosumab years 2/3, §P < 0.05 vs premenopausal reference group, by two-sided Wilcoxon rank-sum test for between-group comparisons. n = number of subjects with observed data. HI for the total bone and for all subregions was significantly lower in the years 2/3 denosumab samples vs placebo (Fig. 4). The year 5 denosumab samples had significantly lower HI for total bone and for all subregions except endocortical compared with the years 2/3 denosumab samples. At year 5, median HI for total bone (0.116 g/cm3; IQR, 0.110 to 0.122) was 19.4% lower compared with years 2/3 placebo controls (0.144 g/cm3; IQR, 0.135 to 0.157). The group median value for HI in the year 10 denosumab group (0.114 g/cm3; IQR, 0.106 to 0.124) was similar to that for the denosumab year 5 group (0.116 g/cm3; IQR, 0.110 to 0.122), suggesting a steady state of heterogeneity was established by approximately year 5. Compared with the premenopausal reference group, HI values for total, cortical, and cancellous bone were significantly lower in the denosumab years 2/3, 5, and 10 groups (Fig. 4). For the FREEDOM and extension studies, changes in DMB and HI were similar for individual subjects from whom multiple sequential biopsies were obtained vs between-group differences observed for the larger cross-sectional sample set, which mostly comprised nonsequential biopsies (Fig. 5A). DMB and HI values were similar in denosumab subjects without fluorochrome labels compared with those with labels (Fig. 5B). Figure 5. View largeDownload slide Additional DMB and HI results for denosumab-treated subjects. Note that the x-axis scales are truncated. (A) Results for sequential biopsies from 11 denosumab-treated subjects that provided serial bone biopsy samples. (B) DMB and HI results in total bone for subjects with and without fluorochrome labels. The box-and-whisker plots and n values are explained in the Fig 4. legend. Figure 5. View largeDownload slide Additional DMB and HI results for denosumab-treated subjects. Note that the x-axis scales are truncated. (A) Results for sequential biopsies from 11 denosumab-treated subjects that provided serial bone biopsy samples. (B) DMB and HI results in total bone for subjects with and without fluorochrome labels. The box-and-whisker plots and n values are explained in the Fig 4. legend. Discussion This study showed that normal bone histology was maintained through 10 years of denosumab therapy in postmenopausal women with osteoporosis from the FREEDOM long-term extension bone biopsy substudy. These findings, including a lack of evidence for woven bone, marrow fibrosis, or impaired matrix mineralization, are similar to previous analyses conducted at year 5 of the FREEDOM long-term extension study (5). Bone histomorphometry findings for denosumab years 2/3, year 5, and year 10 are consistent with the mechanism of action of denosumab, which potently inhibits bone resorption and remodeling and increases bone mass and strength (4). The year 10 data show few differences compared with histomorphometry data for years 2/3 and year 5 of denosumab therapy (3, 5); these differences include higher trabecular number and lower trabecular spacing, although the year 10 findings were largely based on cross-sectional comparisons to earlier time points. There were only five sequential biopsies that bridged year 10 with earlier time points, which are too few to make meaningful conclusions about changes over time. The percentage of biopsies with cancellous fluorochrome labels was also observed to be higher at year 10 compared with earlier denosumab time points, but dynamic histomorphometry otherwise indicated very low levels of bone turnover at year 10, indicating that the strong antiremodeling effects of denosumab were maintained over this treatment duration. Systemic bone turnover markers also remained substantially reduced through 10 years of denosumab (9), although modest release of this inhibition tends to occur toward the end of the 6-month denosumab dosing interval (34). Recognizing that iliac crest bone biopsies represent a small, non–weight-bearing sample of the entire skeleton, these findings collectively suggest that denosumab markedly inhibits bone remodeling throughout much of the skeleton, an effect that is associated with persistently low fracture rates through at least a decade of therapy (9). Concerns have been raised that long-term administration of potent remodeling inhibitors might have deleterious effects on bone strength, potentially manifesting as rare atypical femur fractures (AFFs) (11, 13, 28, 35). These concerns, originally based on associations observed in BP studies, may also apply to denosumab, which has even greater antiremodeling effects (36, 37). Remodeling inhibitors may influence bone strength through changes in bone matrix mineralization, although it remains unclear whether such effects are biomechanically favorable or unfavorable (13, 14, 25, 38). The current study used microradiography to assess temporal changes in bone matrix mineralization characteristics during 10 years of denosumab therapy. Microradiography-based matrix mineralization data were previously shown to correlate with tissue-level bone strength (39). The digital microradiography method used in the current study was validated and used in previous clinical studies (31, 40, 41). As expected for a potent remodeling inhibitor, denosumab increased the overall degree of mineralization (i.e., DMB), and reduced the heterogeneity of mineralization (HI), compared with placebo. These DMB and HI changes were progressive with up to 5 years of denosumab therapy, with minimal, nonsignificant changes thereafter. The FREEDOM long-term extension population may have experienced the greatest overall bone remodeling inhibition of any group of treated postmenopausal women studied to date. Denosumab does not incorporate into bone matrix (15), which eliminates the potential for direct effects on mineralization characteristics or matrix material properties that may potentially arise from skeletal uptake of BPs (42). As such, the current changes in DMB and HI may be among the largest achieved through remodeling inhibition in a clinical trial setting, presenting a powerful opportunity to address true relationships between remodeling, matrix mineralization, and fracture risk. Previous quantitative backscattered electron imaging (qBEI) data showed that the degree of cancellous bone matrix mineralization in one study of BP-treated postmenopausal women with osteoporosis exceeded levels found in a skeletally healthy reference group (20), whereas in other studies, the heterogeneity of mineralization was similar in BP-treated subjects compared with this same reference group (20, 43). In the current study, DMB and HI values in the year 5 and 10 denosumab samples were both significantly different vs a premenopausal reference group, which further suggests a sizable treatment effect on the degree and heterogeneity of mineralization. It is therefore interesting to note that the overall study population from which the current biopsy subset was drawn (2343 women in the long-term arm of the extension study, 1343 of whom completed the study) showed persistently low rates of new vertebral, nonvertebral, and hip fractures through year 10 (9), with rates of nonvertebral fractures being lower during the extension compared with the rate for denosumab subjects during FREEDOM (44). Those findings indicate that the duration and degree of DMB and HI changes achieved here do not weaken bone or increase fracture risk at the population level. One subject from the long-term arm of the extension study (i.e., the group receiving denosumab since the beginning of FREEDOM) experienced an event consistent with the definition of AFF after 7 total years of denosumab (9), but this subject was not part of the biopsy substudy, and her bone matrix mineralization characteristics remain unknown. Other recent data indicated that bone matrix adjacent to AFF sites in BP-treated women had higher mineralization, as suggested by a greater mineral-to-organic-matrix ratio, compared with fracture samples from BP-treated and untreated women with typical femur fractures. However, the same samples showed that qBEI-derived variables that correspond to DMB (i.e., calcium mean) and HI (i.e., calcium width) were similar at AFF vs non-AFF sites from long-term BP users, and were also similar at AFF sites of BP users vs nonfractured femoral bone tissue from non-BP users (28). The current results, combined with long-term denosumab fracture data (9) and long-term denosumab bone quality data (10), suggest that the changes in matrix mineralization characteristics resulting from up to 10 years of denosumab treatment have favorable effects on bone structural strength. Similarly, the low levels or absence of fluorochrome labeling in bone biopsies do not appear to carry negative implications for bone strength in the current population and is likely a reflection of efficacy (4). Indeed, there was no difference in mineralization variables between subjects that had detectable fluorochrome labels and those that did not. There are limited data on the effects of antiresorptives on matrix mineralization beyond 3 years, with one BP study reporting that changes in the degree and heterogeneity of mineralization did not progress from year 3 to year 5 of treatment (19), and another showing no progressive changes in mineralization characteristics between years 2/3 and year 10 of treatment (43). Evidence of progressive changes in DMB and HI between denosumab years 2/3 and year 5 may be a unique finding among antiresorptive therapies and may have clinical implications. First, these findings suggest that the total remodeling period in these denosumab subjects, including secondary mineralization, is ∼5 years. A study in alendronate-treated postmenopausal women indicated that a new steady state of higher and less heterogeneous matrix mineralization was achieved within 2 to 3 years (18). This may indicate that secondary mineralization of preexisting remodeling units reached completion within 3 years after initiating alendronate, although ongoing residual remodeling may have influenced those results. Compared with alendronate, denosumab causes more rapid and more substantial inhibition of bone remodeling; bone resorption markers were greatly reduced within 3 days of initiating denosumab therapy (34), suggesting that changes in matrix mineralization begin within the first week of treatment. Based on that early trigger, and on observed changes in DMB and HI up to year 5, the time required to refill existing remodeling spaces and complete their primary and secondary mineralization appears to be ∼5 years. It is unclear whether this 5-year duration is unique to denosumab-treated postmenopausal women or whether it might apply to other populations as well. Second, these findings imply that increased matrix mineralization may contribute to progressive BMD gains with denosumab therapy for up to ∼5 years, but perhaps not thereafter. Denosumab causes continued BMD gains between year 5 and year 10 of treatment (9), and those gains may result from mineralization-independent phenomena, perhaps including modeling-based bone formation (MBBF). MBBF persisted in femoral neck of denosumab-treated ovariectomized cynomolgus monkeys despite continued strong inhibition of resorption and remodeling, and cancellous MBBF was significantly increased in iliac crest bone biopsies from denosumab-treated postmenopausal women (45, 46). A third implication of the mineralization findings relates to overall bone strength. Matrix mineralization and its tissue-level strength are strongly related (26), and increased matrix mineralization and/or reduced mineralization heterogeneity with denosumab and other RANKL inhibitors was associated with increased bone strength in animals (21–23). Those findings, and the current data, imply that changes in matrix mineralization may lead to progressive increases in bone strength for up to ∼5 years of denosumab therapy, which would represent the minimum duration of treatment for achieving the full biomechanical benefits conferred by denosumab’s antiremodeling effect. The reductions in HI in the 5- and 10-year samples were highly statistically significant vs placebo, and HI was also significantly lower for all denosumab treatment durations compared with a premenopausal reference range, yet the absolute reductions in HI compared with placebo seem modest in light of the degree and duration of remodeling inhibition associated with denosumab. As such, this study does not provide clear insights into the biomechanical implications of highly and homogenously mineralized matrix. It is reasonable to suspect that such a scenario, were it achievable, might represent a truly hypermineralized state characterized by inferior biomechanical properties that manifest through mechanisms previously proposed (13, 14, 28). As mentioned, biochemical markers of bone turnover tend to show some release of inhibition toward the end of the 6-monthly denosumab dosing interval, and we cannot exclude the possibility that “breakthrough” remodeling contributed to the lack of hypermineralization. However, these biopsies were collected near the end of the denosumab dosing interval, and dynamic histomorphometry nonetheless indicated very low remodeling rates in all denosumab groups, with activation frequencies at or near zero. These data suggest that the lack of hypermineralization in the year 10 biopsies is not for want of greater osteoclast inhibition, and may imply that factors besides remodeling activation could be functioning as self-regulatory mechanisms within bone that maintain key mineralization characteristics within biomechanically acceptable ranges. In support of this hypothesis, the remodeling period (47) and the rate of secondary matrix mineralization (39) both increase substantially with age, and antiresorptive therapies including denosumab can also increase the formation period (3). Each of these changes may serve to limit the potential for extreme increases in DMB and extreme decreases in HI when osteoclasts become inhibited. This study has several limitations, including a relatively small number of bone biopsies, which limits the ability to compare the results with specific patient outcomes, including fragility fractures and rare safety events such as AFF and osteonecrosis of the jaw. There were few paired (i.e., sequential) bone biopsies, although observed changes in matrix mineralization parameters for the paired biopsy subset generally align with intergroup differences observed for the entire sample set, which mostly comprised unpaired biopsies. The lack of placebo controls in FREEDOM extension limits firm conclusions regarding the effects of denosumab on histomorphometry and matrix mineralization characteristics at years 5 and 10. Finally, the microradiography method used to assess matrix mineralization provides end points that are analogous but not identical to those provided by other methods (e.g., qBEI, Raman spectroscopy), making it difficult to directly compare the current results against matrix mineralization data from some previous clinical trials of antiresorptive agents. In summary, iliac crest bone biopsies obtained from a subset of postmenopausal women from the long-term arm of FREEDOM extension study showed (1) maintenance of normal bone histology, (2) maintenance or improvements in bone microstructural parameters, (3) a persistently low state of bone resorption and remodeling, and (4) changes in bone matrix mineralization characteristics that are consistent with denosumab’s mechanism of action as a potent remodeling inhibitor. The degree and heterogeneity of bone matrix mineralization changed with up to 5 years of denosumab and remained similar between the year 5 and year 10 denosumab biopsies. These findings, when viewed in the light of 10-year data on fracture rates and bone safety assessments from the FREEDOM extension (9), indicate that denosumab maintains a favorable efficacy and bone safety profile for 10 years of uninterrupted therapy. Abbreviations: Abbreviations: AFF atypical femur fracture BMD bone mineral density BP bisphosphonate DMB degree of mineralization of bone FREEDOM Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months GMMA glycol methylmethacrylate HI heterogeneity index IQR interquartile range MBBF modeling-based bone formation qBEI quantitative backscattered electron imaging Acknowledgments Financial Support: Amgen Inc. sponsored this study. Paul Kostenuik (Phylon Pharma Services) and Lisa Humphries (Amgen Inc.) provided medical writing support. Clinical Trial Information: ClinicalTrials.gov no. NCT00089791 (registered 16 August 2004) and no. NCT00523341 (registered 31 August 2007). Author Contributions: D.W.D., X.Y., S.V.Y., and G.B. take responsibility for the data and accuracy of the analyses. D.W.D. wrote the initial manuscript draft with medical writing assistance support from the sponsor. All authors contributed to subsequent drafts of the manuscript, participated in the analysis and/or interpretation of the data, and in the critical review and revision of the report. All authors approved the final manuscript for submission. Disclosure Summary: D.W.D. has received grant/research support from Amgen Inc., Eli Lilly, and Radius Health; has consulted for Merck, Amgen Inc., Eli Lilly, Radius Health, and Tarsa; and is a member of the Speakers’ Bureau for Amgen Inc., Eli Lilly, and Radius Health. J.P.B. has received grant/research support from Amgen Inc. and Eli Lilly; has consulted for Amgen Inc., Eli Lilly, and Merck; and is a member of the Speakers’ Bureau for Amgen Inc. and Eli Lilly. A.F.-P. has received grant/research support from Amgen and Roche and is a member of the Speakers’ Bureau for Alexion, Amgen Inc., Eli Lilly, Merck, Shire, and UCB. D.K. has received grant/research support from Amgen Inc., Eli Lilly, and AstraZeneca; has consulted for Amgen Inc., Eli Lilly, and Pfizer; and is a member of the Speakers’ Bureau for Amgen Inc., Eli Lilly, and GSK. R.B.W., X.Y., and S.V.Y. are employed by Amgen Inc. and have stock/stock options in Amgen Inc. G.B. received grant/research support from Amgen Inc., Biom’up, Hoffman-La Roche, Eli Lilly, MSD, National Institutes of Health, Noraker, Procter & Gamble, and Servier, and has consulted for Amgen Inc, MSD, and Servier. The remaining authors have nothing to disclose. References 1. Lacey DL , Boyle WJ , Simonet WS , Kostenuik PJ , Dougall WC , Sullivan JK , San Martin J , Dansey R . Bench to bedside: elucidation of the OPG-RANK-RANKL pathway and the development of denosumab . Nat Rev Drug Discov . 2012 ; 11 ( 5 ): 401 – 419 . 2. Cummings SR , San Martin J , McClung MR , Siris ES , Eastell R , Reid IR , Delmas P , Zoog HB , Austin M , Wang A , Kutilek S , Adami S , Zanchetta J , Libanati C , Siddhanti S , Christiansen C ; FREEDOM Trial . Denosumab for prevention of fractures in postmenopausal women with osteoporosis . N Engl J Med . 2009 ; 361 ( 8 ): 756 – 765 . 3. Reid IR , Miller PD , Brown JP , Kendler DL , Fahrleitner-Pammer A , Valter I , Maasalu K , Bolognese MA , Woodson G , Bone H , Ding B , Wagman RB , San Martin J , Ominsky MS , Dempster DW ; Denosumab Phase 3 Bone Histology Study Group . Effects of denosumab on bone histomorphometry: the FREEDOM and STAND studies . J Bone Miner Res . 2010 ; 25 ( 10 ): 2256 – 2265 . 4. Kostenuik PJ , Smith SY , Jolette J , Schroeder J , Pyrah I , Ominsky MS . Decreased bone remodeling and porosity are associated with improved bone strength in ovariectomized cynomolgus monkeys treated with denosumab, a fully human RANKL antibody . Bone . 2011 ; 49 ( 2 ): 151 – 161 . 5. Brown JP , Reid IR , Wagman RB , Kendler D , Miller PD , Jensen JE , Bolognese MA , Daizadeh N , Valter I , Zerbini CA , Dempster DW . Effects of up to 5 years of denosumab treatment on bone histology and histomorphometry: the FREEDOM study extension . J Bone Miner Res . 2014 ; 29 ( 9 ): 2051 – 2056 . 6. Papapoulos S , Chapurlat R , Libanati C , Brandi ML , Brown JP , Czerwiński E , Krieg MA , Man Z , Mellström D , Radominski SC , Reginster JY , Resch H , Román Ivorra JA , Roux C , Vittinghoff E , Austin M , Daizadeh N , Bradley MN , Grauer A , Cummings SR , Bone HG . Five years of denosumab exposure in women with postmenopausal osteoporosis: results from the first two years of the FREEDOM extension . J Bone Miner Res . 2012 ; 27 ( 3 ): 694 – 701 . 7. Ferrari S , Adachi JD , Lippuner K , Zapalowski C , Miller PD , Reginster JY , Törring O , Kendler DL , Daizadeh NS , Wang A , O’Malley CD , Wagman RB , Libanati C , Lewiecki EM . Further reductions in nonvertebral fracture rate with long-term denosumab treatment in the FREEDOM open-label extension and influence of hip bone mineral density after 3 years . Osteoporos Int . 2015 ; 26 ( 12 ): 2763 – 2771 . 8. Papapoulos S , Lippuner K , Roux C , Lin CJ , Kendler DL , Lewiecki EM , Brandi ML , Czerwiński E , Franek E , Lakatos P , Mautalen C , Minisola S , Reginster JY , Jensen S , Daizadeh NS , Wang A , Gavin M , Libanati C , Wagman RB , Bone HG . The effect of 8 or 5 years of denosumab treatment in postmenopausal women with osteoporosis: results from the FREEDOM Extension study . Osteoporos Int . 2015 ; 26 ( 12 ): 2773 – 2783 . 9. Bone HG , Wagman RB , Brandi ML , Brown JP , Chapurlat R , Cummings SR , Czerwiński E , Fahrleitner-Pammer A , Kendler DL , Lippuner K , Reginster JY , Roux C , Malouf J , Bradley MN , Daizadeh NS , Wang A , Dakin P , Pannacciulli N , Dempster DW , Papapoulos S . 10 years of denosumab treatment in postmenopausal women with osteoporosis: results from the phase 3 randomised FREEDOM trial and open-label extension . Lancet Diabetes Endocrinol . 2017 ; 5 ( 7 ): 513 – 523 . 10. Ominsky MS , Stouch B , Schroeder J , Pyrah I , Stolina M , Smith SY , Kostenuik PJ . Denosumab, a fully human RANKL antibody, reduced bone turnover markers and increased trabecular and cortical bone mass, density, and strength in ovariectomized cynomolgus monkeys . Bone . 2011 ; 49 ( 2 ): 162 – 173 . 11. Odvina CV , Zerwekh JE , Rao DS , Maalouf N , Gottschalk FA , Pak CY . Severely suppressed bone turnover: a potential complication of alendronate therapy . J Clin Endocrinol Metab . 2005 ; 90 ( 3 ): 1294 – 1301 . 12. Miyazaki T , Tokimura F , Tanaka S . A review of denosumab for the treatment of osteoporosis . Patient Prefer Adherence . 2014 ; 8 : 463 – 471 . 13. Donnelly E , Meredith DS , Nguyen JT , Gladnick BP , Rebolledo BJ , Shaffer AD , Lorich DG , Lane JM , Boskey AL . Reduced cortical bone compositional heterogeneity with bisphosphonate treatment in postmenopausal women with intertrochanteric and subtrochanteric fractures . J Bone Miner Res . 2012 ; 27 ( 3 ): 672 – 678 . 14. Allen MR , Burr DB . Mineralization, microdamage, and matrix: how bisphosphonates influence material properties of bone . Bonekey Osteovision . 2007 ; 4 ( 2 ): 49 – 60 . 15. Kostenuik PJ , Nguyen HQ , McCabe J , Warmington KS , Kurahara C , Sun N , Chen C , Li L , Cattley RC , Van G , Scully S , Elliott R , Grisanti M , Morony S , Tan HL , Asuncion F , Li X , Ominsky MS , Stolina M , Dwyer D , Dougall WC , Hawkins N , Boyle WJ , Simonet WS , Sullivan JK . Denosumab, a fully human monoclonal antibody to RANKL, inhibits bone resorption and increases BMD in knock-in mice that express chimeric (murine/human) RANKL . J Bone Miner Res . 2009 ; 24 ( 2 ): 182 – 195 . 16. Bonnet N , Gerbaix M , Ominsky M , Ammann P , Kostenuik PJ , Ferrari SL . Influence of fatigue loading and bone turnover on bone strength and pattern of experimental fractures of the tibia in mice . Calcif Tissue Int . 2016 ; 99 ( 1 ): 99 – 109 . 17. Norman TL , Yeni YN , Brown CU , Wang Z . Influence of microdamage on fracture toughness of the human femur and tibia . Bone . 1998 ; 23 ( 3 ): 303 – 306 . 18. Boivin GY , Chavassieux PM , Santora AC , Yates J , Meunier PJ . Alendronate increases bone strength by increasing the mean degree of mineralization of bone tissue in osteoporotic women . Bone . 2000 ; 27 ( 5 ): 687 – 694 . 19. Zoehrer R , Roschger P , Paschalis EP , Hofstaetter JG , Durchschlag E , Fratzl P , Phipps R , Klaushofer K . Effects of 3- and 5-year treatment with risedronate on bone mineralization density distribution in triple biopsies of the iliac crest in postmenopausal women . J Bone Miner Res . 2006 ; 21 ( 7 ): 1106 – 1112 . 20. Misof BM , Roschger P , Gabriel D , Paschalis EP , Eriksen EF , Recker RR , Gasser JA , Klaushofer K . Annual intravenous zoledronic acid for three years increased cancellous bone matrix mineralization beyond normal values in the HORIZON biopsy cohort . J Bone Miner Res . 2013 ; 28 ( 3 ): 442 – 448 . 21. Valenta A , Roschger P , Fratzl-Zelman N , Kostenuik PJ , Dunstan CR , Fratzl P , Klaushofer K . Combined treatment with PTH (1-34) and OPG increases bone volume and uniformity of mineralization in aged ovariectomized rats . Bone . 2005 ; 37 ( 1 ): 87 – 95 . 22. Ross AB , Bateman TA , Kostenuik PJ , Ferguson VL , Lacey DL , Dunstan CR , Simske SJ . The effects of osteoprotegerin on the mechanical properties of rat bone . J Mater Sci Mater Med . 2001 ; 12 ( 7 ): 583 – 588 . 23. Misof BM , Roschger P , Ominsky MS , Messmer P , Kostenuik P , Klaushofer K . The effect of denosumab on bone matrix mineralization in mice [abstract] . J Bone Miner Res . 2011 ; 26 ( Suppl. 1 ):S94. Abstract 0058. 24. Meunier PJ , Boivin G . Bone mineral density reflects bone mass but also the degree of mineralization of bone: therapeutic implications . Bone . 1997 ; 21 ( 5 ): 373 – 377 . 25. Follet H , Boivin G , Rumelhart C , Meunier PJ . The degree of mineralization is a determinant of bone strength: a study on human calcanei . Bone . 2004 ; 34 ( 5 ): 783 – 789 . 26. Bala Y , Farlay D , Delmas PD , Meunier PJ , Boivin G . Time sequence of secondary mineralization and microhardness in cortical and cancellous bone from ewes . Bone . 2010 ; 46 ( 4 ): 1204 – 1212 . 27. Turner CH . Biomechanics of bone: determinants of skeletal fragility and bone quality . Osteoporos Int . 2002 ; 13 ( 2 ): 97 – 104 . 28. Lloyd AA , Gludovatz B , Riedel C , Luengo EA , Saiyed R , Marty E , Lorich DG , Lane JM , Ritchie RO , Busse B , Donnelly E . Atypical fracture with long-term bisphosphonate therapy is associated with altered cortical composition and reduced fracture resistance . Proc Natl Acad Sci USA . 2017 ; 114 ( 33 ): 8722 – 8727 . 29. Boskey AL , Spevak L , Weinstein RS . Spectroscopic markers of bone quality in alendronate-treated postmenopausal women . Osteoporos Int . 2009 ; 20 ( 5 ): 793 – 800 . 30. Dempster DW , Compston JE , Drezner MK , Glorieux FH , Kanis JA , Malluche H , Meunier PJ , Ott SM , Recker RR , Parfitt AM . Standardized nomenclature, symbols, and units for bone histomorphometry: a 2012 update of the report of the ASBMR Histomorphometry Nomenclature Committee . J Bone Miner Res . 2013 ; 28 ( 1 ): 2 – 17 . 31. Montagner F , Kaftandjian V , Farlay D , Brau D , Boivin G , Follet H . Validation of a novel microradiography device for characterization of bone mineralization . J XRay Sci Technol . 2015 ; 23 ( 2 ): 201 – 211 . 32. Parisien M , Cosman F , Morgan D , Schnitzer M , Liang X , Nieves J , Forese L , Luckey M , Meier D , Shen V , Lindsay R , Dempster DW . Histomorphometric assessment of bone mass, structure, and remodeling: a comparison between healthy black and white premenopausal women . J Bone Miner Res . 1997 ; 12 ( 6 ): 948 – 957 . 33. Roschger P , Gupta HS , Berzlanovich A , Ittner G , Dempster DW , Fratzl P , Cosman F , Parisien M , Lindsay R , Nieves JW , Klaushofer K . Constant mineralization density distribution in cancellous human bone . Bone . 2003 ; 32 ( 3 ): 316 – 323 . 34. McClung MR , Lewiecki EM , Cohen SB , Bolognese MA , Woodson GC , Moffett AH , Peacock M , Miller PD , Lederman SN , Chesnut CH , Lain D , Kivitz AJ , Holloway DL , Zhang C , Peterson MC , Bekker PJ ; AMG 162 Bone Loss Study Group . Denosumab in postmenopausal women with low bone mineral density . N Engl J Med . 2006 ; 354 ( 8 ): 821 – 831 . 35. Shane E , Burr D , Ebeling PR , Abrahamsen B , Adler RA , Brown TD , Cheung AM , Cosman F , Curtis JR , Dell R , Dempster D , Einhorn TA , Genant HK , Geusens P , Klaushofer K , Koval K , Lane JM , McKiernan F , McKinney R , Ng A , Nieves J , O’Keefe R , Papapoulos S , Sen HT , van der Meulen MC , Weinstein RS , Whyte M ; American Society for Bone and Mineral Research . Atypical subtrochanteric and diaphyseal femoral fractures: report of a task force of the American Society for Bone and Mineral Research . J Bone Miner Res . 2010 ; 25 ( 11 ): 2267 – 2294 . 36. Brown JP , Prince RL , Deal C , Recker RR , Kiel DP , de Gregorio LH , Hadji P , Hofbauer LC , Alvaro-Gracia JM , Wang H , Austin M , Wagman RB , Newmark R , Libanati C , San Martin J , Bone HG . Comparison of the effect of denosumab and alendronate on BMD and biochemical markers of bone turnover in postmenopausal women with low bone mass: a randomized, blinded, phase 3 trial . J Bone Miner Res . 2009 ; 24 ( 1 ): 153 – 161 . 37. Miller P , Pannacciulli N , Malouf-Sierra J , Singer A , Czerwinski E , Bone HG , Wang C , Wagman RB , Brown JP . A meta-analysis of four clinical trials of denosumab compared with bisphosphonates in postmenopausal women previously treated with oral bisphosphonates . J Bone Miner Res . 2017 ; 32 ( Suppl 1 ): S271 . 38. Currey JD . The effect of porosity and mineral content on the Young’s modulus of elasticity of compact bone . J Biomech . 1988 ; 21 ( 2 ): 131 – 139 . 39. Marotti G , Favia A , Zallone AZ . Quantitative analysis on the rate of secondary bone mineralization . Calcif Tissue Res . 1972 ; 10 ( 1 ): 67 – 81 . 40. Farlay D , Armas LA , Gineyts E , Akhter MP , Recker RR , Boivin G . Nonenzymatic glycation and degree of mineralization are higher in bone from fractured patients with type 1 diabetes mellitus . J Bone Miner Res . 2016 ; 31 ( 1 ): 190 – 195 . 41. Mailhot G , Dion N , Farlay D , Rizzo S , Bureau NJ , Jomphe V , Sankhe S , Boivin G , Lands LC , Ferraro P , Ste-Marie LG . Impaired rib bone mass and quality in end-stage cystic fibrosis patients . Bone . 2017 ; 98 : 9 – 17 . 42. Bala Y , Depalle B , Farlay D , Douillard T , Meille S , Follet H , Chapurlat R , Chevalier J , Boivin G . Bone micromechanical properties are compromised during long-term alendronate therapy independently of mineralization . J Bone Miner Res . 2012 ; 27 ( 4 ): 825 – 834 . 43. Roschger P , Lombardi A , Misof BM , Maier G , Fratzl-Zelman N , Fratzl P , Klaushofer K . Mineralization density distribution of postmenopausal osteoporotic bone is restored to normal after long-term alendronate treatment: qBEI and sSAXS data from the fracture intervention trial long-term extension (FLEX) . J Bone Miner Res . 2010 ; 25 ( 1 ): 48 – 55 . 44. Ferrari S , Butler P , Kendler DL , Miller PD , Roux C , Wang AT , Wagman RB , Lewiecki EM . Ten-year continued nonvertebral fracture reduction in postmenopausal women with denosumab treatment . J Bone Miner Res . 2017 ; 32 : S25 . 45. Ominsky MS , Libanati C , Niu QT , Boyce RW , Kostenuik PJ , Wagman RB , Baron R , Dempster DW . Sustained modeling-based bone formation during adulthood in cynomolgus monkeys may contribute to continuous BMD gains with denosumab . J Bone Miner Res . 2015 ; 30 ( 7 ): 1280 – 1289 . 46. Dempster D , Zhou H , Recker R , Brown J , Recknor C , Lewiecki EM , Miller P , Rao S , Kendler D , Lindsay R , Krege JH , Alam J , Taylor K , Ruff VA . Longitudinal changes in modeling- and remodeling-based bone formation with an anabolic vs. an antiresorptive agent in the AVA osteoporosis study . J Bone Miner Res . 2016 ; 31 : S51 . 47. Recker RR , Kimmel DB , Parfitt AM , Davies KM , Keshawarz N , Hinders S . Static and tetracycline-based bone histomorphometric data from 34 normal postmenopausal females . J Bone Miner Res . 1988 ; 3 ( 2 ): 133 – 144 . Copyright © 2018 Endocrine Society This article has been published under the terms of the Creative Commons Attribution License (CC BY; https://creativecommons.org/licenses/by/4.0/).

Soininen, Sonja;Eloranta, Aino-Maija;Viitasalo, Anna;Dion, Geneviève;Erkkilä, Arja;Sidoroff, Virpi;Lindi, Virpi;Mahonen, Anitta;Lakka, Timo A

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00335pmid: 29750416

Abstract Context The associations of serum 25-hydroxyvitamin D [25(OH)D] with plasma lipids remain controversial in children. Objective To examine the associations and interactions of 25(OH)D and related gene variants with lipids in children. Design Cross-sectional. Setting Kuopio, Finland. Participants Population sample of 419 prepubertal white children aged 6 to 8 years. Main Outcome Measures 25(OH)D, total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides. Results Serum 25(OH)D was negatively associated with total cholesterol (β = –0.141, P = 0.004), LDL cholesterol (β = –0.112, P = 0.023), HDL cholesterol (β = –0.150, P = 0.002), and triglycerides (β = –0.104, P = 0.035) adjusted for age and sex. Associations of 25(OH)D with total cholesterol, LDL cholesterol, and HDL cholesterol remained after adjustment for adiposity, physical activity, sedentary behavior, diet, daylight time, and parental education. Children in the highest quartile of 25(OH)D had the lowest total cholesterol (P = 0.022) and LDL cholesterol (P = 0.026) adjusted for age and sex. Cytochrome P450 family 2 subfamily R member 1 (CYP2R1) rs12794714, CYP2R1 rs10741657, and vitamin D binding protein (DBP) rs2282679 were associated with 25(OH)D adjusted for age and sex. CYP2R1 rs12794714 was associated with total cholesterol and LDL cholesterol and C10orf88 rs6599638 with HDL cholesterol adjusted for age, sex, and 25(OH)D. The gene variants did not explain or modify the associations of 25(OH)D with lipids. Conclusions 25(OH)D was independently and inversely associated with total cholesterol, LDL cholesterol, and HDL cholesterol. CYP2R1 rs12794714, CYP2R1 rs10741657, and DBP rs2282679 were associated with 25(OH)D. CYP2R1 rs12794714 was associated with total cholesterol and LDL cholesterol and chromosome 10 open reading frame 88 (C10orf88) rs6599638 with HDL cholesterol independent of 25(OH)D. None of the gene variants modified the associations of 25(OH)D with lipids. Further studies are needed to detect the mechanisms for the associations of 25(OH)D with lipids. Vitamin D regulates calcium, phosphorus, and bone metabolism, and its deficiency is associated with rickets in children and osteomalacia in adults (1). The knowledge of the other health effects of vitamin D is increasing, and low serum 25-hydroxyvitamin D [25(OH)D] has been associated with the components of metabolic syndrome and the increased risk of cardiovascular diseases in adults (2, 3). In children, decreased serum 25(OH)D has been associated with cardiometabolic risk factors in some studies, but the results have been inconsistent (4–8). Abnormalities in lipid and lipoprotein metabolism, especially increased plasma low-density lipoprotein (LDL) cholesterol, are risk factors for atherosclerosis already in childhood (9). In adults, serum 25(OH)D has been positively associated with plasma high-density lipoprotein (HDL) cholesterol and inversely associated with plasma triglycerides, but the associations with plasma total cholesterol and LDL cholesterol have been inconsistent (10). In children and adolescents, the associations of 25(OH)D with plasma levels of these lipids have been conflicting, and both positive and inverse associations have been found (4–6, 8, 11–13). The knowledge on the effects of vitamin D supplementation on lipid metabolism and cardiovascular health obtained from intervention studies in children and in adults is also insufficient and inconclusive (3, 4, 10). Increased serum 25(OH)D may be due to a healthy lifestyle, including regular exercise, spending plenty of time outdoors resulting in increased vitamin D production in the skin, and a healthy diet, all of which may also be associated with a more favorable plasma lipid profile. Therefore, causality between serum 25(OH)D and cardiovascular risk factors and diseases is not clear. Many studies in children lack information on several potential confounding factors for the associations of serum 25(OH)D with plasma lipids, such as pubertal status, adiposity, physical activity, socioeconomic status, and dietary factors. Also genetic factors may affect the associations of serum 25(OH)D with plasma lipids. Genome-wide association studies (GWASs) have identified several single nucleotide polymorphisms (SNPs) in genes linked with vitamin D metabolism to be associated with serum 25(OH)D (14, 15). However, there are few studies on the associations of SNPs related to serum 25(OH)D with plasma lipids (16–18). Vitamin D and cholesterol have a common precursor, 7-dehydrocholesterol (1), and vitamin D receptor complexes have been suggested to regulate cholesterol metabolism (19). We therefore hypothesized that genetic factors related to vitamin D metabolism may partly explain or modify the association between serum 25(OH)D and plasma lipids. As the process of atherosclerosis begins already in childhood (9), and cardiometabolic risk factors track from childhood to adulthood (20), it is important to understand the associations, mechanisms, and potential confounding factors between serum 25(OH)D and plasma lipids. We therefore studied the associations of serum 25(OH)D with plasma lipids, adjusting for a number of possible confounding factors in a population sample of prepubertal children 6 to 8 years of age. Moreover, we investigated whether SNPs previously related to serum 25(OH)D modify the associations of serum 25(OH)D with plasma lipids. Subjects and Methods Study design and participants The current study is based on the baseline data of the Physical Activity and Nutrition in Children (PANIC) study, which is a physical activity and dietary intervention study in a population sample of children 6 to 8 years of age from the city of Kuopio, Finland (ClinicalTrials.gov no. NCT01803776). Altogether 736 children from the primary schools of Kuopio were invited to participate in the baseline examinations from 2007 to 2009. Of the invited children, 512 (70%) participated in the baseline examinations. The participants did not differ in age, sex distribution, or body mass index standard deviation score (BMI-SDS) from all children who started the first grade in the city of Kuopio in 2007 to 2009 based on data from the standard school health examinations. We excluded children who had chronic diseases or medications that could affect serum 25(OH)D or plasma lipids, had entered puberty, or had race other than white to avoid confounding in statistical analyses. Complete data on the main variables were available for 419 children (195 girls, 224 boys) and valid data on dietary factors for 377 children (179 girls, 198 boys). The study was conducted according to the ethical guidelines laid down in the Declaration of Helsinki. The study protocol was approved by the Research Ethics Committee of the Hospital District of Northern Savo. Both children and their parents gave their written informed consent. The data that has been used is confidential, and therefore the data sources are not shared. Measurement of serum 25(OH)D and plasma lipids Venous blood samples for the measurement of 25(OH)D and lipids were taken after 12-hour overnight fasting. For 25(OH)D analyses, blood was immediately centrifuged and stored at a temperature of –75°C until biochemical analyses. Serum 25(OH)D concentration was analyzed by a chemiluminescence immunoassay called the LIAISON® 25 OH Vitamin D TOTAL Assay (DiaSorin, Stillwater, MN) using an automatic immunoanalyser (DiaSorin). Total variation, including intra-assay and interassay variation, for the assay was 8.2% to 11.0% in the concentration range of 21 to 123 nmol/L. The 25(OH)D analyses were performed in Eastern Finland Laboratory Centre Joint Authority Enterprise (ISLAB), which has been participating in the Vitamin D External Quality Assessment Scheme (DEQAS) since 2008 with DiaSorin LIAISON 25(OH)D assay meeting the performance targets, as described earlier (21). Lipids were measured from nonfrozen plasma samples. A colorimetric enzymatic assay was used to analyze plasma total cholesterol and triglyceride concentrations (Roche Diagnostics, Mannheim, Germany). The intra-assay and interassay coefficients of variation were 1.0% to 1.4% and 1.2% to 3.1% for total cholesterol and 0.9% to 4.2% and 1.5% to 1.8% for triglycerides, respectively. Homogeneous enzymatic colorimetric assays were used to analyze plasma HDL and LDL cholesterol concentrations (Roche Diagnostics). The intra-assay and interassay coefficients of variation were 1.1% to 1.3% and 1.3% to 4.3% for HDL cholesterol and 0.9% to 1.2% and 1.7% to 2.7% for LDL cholesterol, respectively. Genotyping and selection of SNPs DNA was isolated from the blood mononuclear cells using the QIAamp® DNA Blood Kit (Qiagen, Hilden, Germany). Genotyping was performed in the Institute for Molecular Medicine Finland (FIMM) using the Infinium® HumanCoreExome BeadChip (Illumina, San Diego, CA). The genotypes were determined using the GenomeStudio® software (Illumina). The final quality control was done using the PLINK® software, version 1.07. We selected SNPs that are located in genes involved in vitamin D metabolism and have been associated with 25(OH)D in two GWASs (14, 15). SNPs characterized by a minor allele frequency (MAF) <0.15 based on the 1000 Genomes data from the Single Nucleotide Polymorphism database (22) or a high linkage disequilibrium (R2 ≥ 0.8) based on the SNP Annotation and Proxy Search (23) were not considered. Overall seven SNPs from five regions were found in our data set: rs12785878 and rs3829251 in NAD synthetase 1 (NADSYN1), which is near a locus coding 7-dehydrocholesterol reductase (DHCR7) that converts 7-dehydrocholesterol to cholesterol but also is a substrate for vitamin D; rs6599638 in chromosome 10 open reading frame 88 (C10orf88) near gene coding acyl-Coenzyme A dehydrogenase involved in producing substrate for cholesterol synthesis; rs10741657 and rs12794714 in the locus of vitamin D-25-hydroxylase, cytochrome P450 family 2 subfamily R member 1 (CYP2R1), which converts vitamin D into 25(OH)D in liver; rs6013897 in the locus of vitamin D-24-hydroxylase, cytochrome P450 family 24 subfamily A member 1 (CYP24A1), which catabolizes both 25(OH)D and 1,25(OH)2D to prevent accumulation of toxic levels; and rs2282679 in the locus coding vitamin D binding protein (DBP). The genotype distributions of all these SNPs were within the Hardy-Weinberg equilibrium. Rs6013897 had a call rate <95% and was therefore excluded from the analyses. Other assessments Body weight was measured twice, with the children having fasted for 12 hours and emptied the bladder and standing in light underwear by the InBody® 720 bioelectrical impedance device (Biospace, Seoul, Korea) to accuracy of 0.1 kg. Body height was measured three times, with the children standing in the Frankfurt plane using a wall-mounted stadiometer to an accuracy of 0.1 cm. BMI-SDS was calculated using national reference values (24). Waist circumference was measured three times after expiration at middistance between the bottom of the rib cage and the top of the iliac crest with an unstretchable measuring tape to an accuracy of 0.1 cm. The means of the nearest two values of weight, height, and waist circumference were used. Body fat percentage was measured with the children being in the supine position, having emptied the bladder, and being in light clothing by dual-energy X-ray absorptiometry using the Lunar Prodigy Advance® dual-energy X-ray absorptiometry device (GE Medical Systems, Madison, WI) and the Encore® software, version 10.51.006 (GE Company, Madison, WI), using standardized protocols. Energy and nutrient intakes were assessed using food records as described in detail earlier (25). Valid food records of consecutive 3 (7.8%) or 4 (92.2%) days, including at least 1 weekend day, were accepted. Supplemental intakes of vitamin D or other nutrients are not included in the total intakes. Intakes of saturated, monounsaturated, and polyunsaturated fatty acids and carbohydrates were calculated as percentages of energy intake. Energy and nutrient intakes were assessed using Micro Nutrica® dietary analysis software, version 2.5 (Social Insurance Institution of Finland). Physical activity and sedentary behavior were assessed by the PANIC Physical Activity Questionnaire filled out by the parents with their children (26). The average daylight time from sunrise to sunset in Kuopio, Finland, at latitude 62·89°N, during 3 months before the blood sampling was obtained from the Almanac Office, University of Helsinki, Finland. Chronic diseases and allergies diagnosed by a physician, the use of medications, parental education, and annual household income were assessed using questionnaires administered by the parents. Parental education was defined as the highest completed or ongoing degree of the parents (vocational school or less; polytechnic or university). Annual household income was categorized as ≤30,000€ or >30,000€. A research physician carried out a medical examination and defined central puberty as breast development at Tanner stage ≥2 for girls and testicular volume ≥4 mL assessed using an orchidometer for boys. Statistical methods We performed statistical analyses using the IBM SPSS Statistics® software, version 23 (IBM Corp., Armonk, NY). The normality of distributions of the variables was verified visually and by the Kolmogorov-Smirnov test, and logarithmic transformation was performed when appropriate. The t test for independent samples and the Pearson χ2 test were used to examine differences in the basic characteristics between sexes. We selected factors that have earlier been associated with 25(OH)D or lipids (25, 27–33) as potential confounders for the associations of 25(OH)D with lipids. Of these factors, body fat percentage, parental education, physical activity, sedentary behavior, average daylight time before blood sampling, and the dietary intakes of fiber and carbohydrates correlated statistically significantly with at least one of the lipid variables and were included in the multivariate models. Of other dietary factors, the quality of dietary fat has been associated with lipids (29) and vitamin D intake with 25(OH)D (25), but the intakes of saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, and vitamin D were not correlated with lipids and were not included in the multivariate models. We did not include the use of vitamin D supplements in the analyses, because this information was missing for many children. Linear regression analysis was used to investigate the associations of 25(OH)D and possible confounding factors with plasma lipids. In Model 1, 25(OH)D and each confounding factor was entered separately with age and sex. In Model 2, age, sex, 25(OH)D, and all selected potential confounding factors were entered simultaneously using backward procedure. The means of total, LDL, and HDL cholesterol and triglycerides in the quartiles of serum 25(OH)D were compared using covariance analysis adjusted for age and sex. The pairwise comparisons of means in the quartiles were performed using the Sidak post hoc correction. These covariance analyses were repeated after additional adjustment for confounding factors that were statistically significantly associated with plasma lipids in backward linear regression analyses. The associations of the SNPs with 25(OH)D and lipids were studied using covariance analysis adjusted for age and sex. Moreover, the associations of SNPs with lipids were adjusted for 25(OH)D, and the associations of 25(OH)D with lipids were adjusted for the SNPs. The interaction between 25(OH)D and each SNP on lipids was studied using linear regression analysis. When dietary factors were used in the analyses, data on only 377 children with valid food records were included. Associations, differences, and interactions with P values <0.05 were considered statistically significant. Results Characteristics of children The boys were heavier and taller, had a higher waist circumference, a lower body fat percentage, and lower LDL cholesterol, were physically more active, and had a higher dietary intake of vitamin D than the girls (Table 1). Mean serum 25(OH)D was 68.1 nmol/L (Table 1). Of the children, 86 (20.5%) had serum 25(OH)D levels below 50 nmol/L, and only 4 (1.0%) had serum 25(OH)D below 30 nmol/L. Table 1. Characteristics of Children All (N = 419)a Girls (n = 195)a Boys (n = 224)a P Value Age, y 7.6 (0.4) 7.6 (0.4) 7.6 (0.4) 0.11 Parental education 0.15  Vocational school or less 81 (19.5%) 32 (16.5%) 49 (22.2%)  Polytechnic or university 334 (80.5%) 162 (83.5%) 172 (77.8%) Household income 0.62  ≤30,000 €/y 91 (22.4%) 45 (23.4%) 46 (21.4%)  >30,000 €/y 316 (77.6%) 147 (76.6%) 169 (78.6%) Body weight, kg 26.7 (4.7) 26.3 (4.8) 27.2 (4.6) 0.048 Body height, cm 128.7 (5.5) 127.6 (5.6) 129.6 (5.3) <0.001 BMI-SDS −0.19 (1.04) −0.19 (1.03) −0.19 (1.05) 0.95 Waist circumference, cm 56.5 (5.3) 55.8 (5.4) 57.1 (5.1) 0.008 Body fat percentage, % 19.5 (7.9) 22.1 (7.3) 17.2 (7.6) <0.001 25(OH)D, nmol/L 68.1 (22.5) 66.5 (18.9) 69.5 (25.2) 0.16 Total cholesterol, mmol/L 4.28 (0.61) 4.33 (0.61) 4.23 (0.61) 0.09 LDL cholesterol, mmol/L 2.36 (0.51) 2.42 (0.52) 2.31 (0.49) 0.032 HDL cholesterol, mmol/L 1.61 (0.31) 1.58 (0.31) 1.63 (0.31) 0.08 Triglycerides, mmol/L 0.60 (0.24) 0.62 (0.25) 0.58 (0.24) 0.08 Total physical activity, h/d 1.9 (0.7) 1.7 (0.6) 2.0 (0.72) <0.001 Total sedentary behavior, h/d 3.6 (1.6) 3.7 (1.6) 3.5 (1.6) 0.12 Average daylight time during 3 months before blood sampling, h/d 11.0 (3.8) 11.1 (3.9) 10.8 (3.7) 0.40 Vitamin D intake from food, µg/d 5.87 (2.16) 5.36 (1.66) 6.34 (2.45) <0.001 SFA intake, E% 12.1 (2.7) 12.0 (2.6) 12.2 (2.8) 0.49 MUFA intake, E% 10.0 (1.8) 9.9 (1.8) 10.1 (1.9) 0.29 PUFA intake, E% 4.9 (1.3) 4.9 (1.3) 5.0 (1.3) 0.66 Carbohydrate intake, E% 51.8 (5.0) 52.1 (4.6) 51.6 (5.3) 0.27 Fiber intake, g/1000 kcal 9.0 (2.5) 9.2 (2.4) 8.7 (2.5) 0.09 All (N = 419)a Girls (n = 195)a Boys (n = 224)a P Value Age, y 7.6 (0.4) 7.6 (0.4) 7.6 (0.4) 0.11 Parental education 0.15  Vocational school or less 81 (19.5%) 32 (16.5%) 49 (22.2%)  Polytechnic or university 334 (80.5%) 162 (83.5%) 172 (77.8%) Household income 0.62  ≤30,000 €/y 91 (22.4%) 45 (23.4%) 46 (21.4%)  >30,000 €/y 316 (77.6%) 147 (76.6%) 169 (78.6%) Body weight, kg 26.7 (4.7) 26.3 (4.8) 27.2 (4.6) 0.048 Body height, cm 128.7 (5.5) 127.6 (5.6) 129.6 (5.3) <0.001 BMI-SDS −0.19 (1.04) −0.19 (1.03) −0.19 (1.05) 0.95 Waist circumference, cm 56.5 (5.3) 55.8 (5.4) 57.1 (5.1) 0.008 Body fat percentage, % 19.5 (7.9) 22.1 (7.3) 17.2 (7.6) <0.001 25(OH)D, nmol/L 68.1 (22.5) 66.5 (18.9) 69.5 (25.2) 0.16 Total cholesterol, mmol/L 4.28 (0.61) 4.33 (0.61) 4.23 (0.61) 0.09 LDL cholesterol, mmol/L 2.36 (0.51) 2.42 (0.52) 2.31 (0.49) 0.032 HDL cholesterol, mmol/L 1.61 (0.31) 1.58 (0.31) 1.63 (0.31) 0.08 Triglycerides, mmol/L 0.60 (0.24) 0.62 (0.25) 0.58 (0.24) 0.08 Total physical activity, h/d 1.9 (0.7) 1.7 (0.6) 2.0 (0.72) <0.001 Total sedentary behavior, h/d 3.6 (1.6) 3.7 (1.6) 3.5 (1.6) 0.12 Average daylight time during 3 months before blood sampling, h/d 11.0 (3.8) 11.1 (3.9) 10.8 (3.7) 0.40 Vitamin D intake from food, µg/d 5.87 (2.16) 5.36 (1.66) 6.34 (2.45) <0.001 SFA intake, E% 12.1 (2.7) 12.0 (2.6) 12.2 (2.8) 0.49 MUFA intake, E% 10.0 (1.8) 9.9 (1.8) 10.1 (1.9) 0.29 PUFA intake, E% 4.9 (1.3) 4.9 (1.3) 5.0 (1.3) 0.66 Carbohydrate intake, E% 51.8 (5.0) 52.1 (4.6) 51.6 (5.3) 0.27 Fiber intake, g/1000 kcal 9.0 (2.5) 9.2 (2.4) 8.7 (2.5) 0.09 The values are means (standard deviations) or numbers (percentages) of children and P values for differences between girls and boys. Differences between girls and boys were tested with independent samples t test for continuous variables and Pearson χ2 test for categorical variables. Logarithmic transformation was performed for triglycerides before analysis. Abbreviations: E%, percentage of energy intake; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid. a Number of children (n) varies from 377 to 419 in different variables; n = 419, 195 girls and 224 boys: age, waist, weight, height, BMI-SDS, total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, and average daylight time; n = 408, 191 girls and 217 boys: body fat percentage; n = 415, 194 girls and 221 boys: parental education; n = 407, 192 girls and 215 boys: household income; n = 415, 194 girls and 221 boys: physical activity and sedentary behavior; and n = 377, 179 girls and 198 boys: intake of vitamin D, fiber, saturated fatty acid percentage of energy intake, monounsaturated fatty acid percentage of energy intake, polyunsaturated fatty acid percentage of energy intake, and carbohydrate percentage of energy intake. View Large Table 1. Characteristics of Children All (N = 419)a Girls (n = 195)a Boys (n = 224)a P Value Age, y 7.6 (0.4) 7.6 (0.4) 7.6 (0.4) 0.11 Parental education 0.15  Vocational school or less 81 (19.5%) 32 (16.5%) 49 (22.2%)  Polytechnic or university 334 (80.5%) 162 (83.5%) 172 (77.8%) Household income 0.62  ≤30,000 €/y 91 (22.4%) 45 (23.4%) 46 (21.4%)  >30,000 €/y 316 (77.6%) 147 (76.6%) 169 (78.6%) Body weight, kg 26.7 (4.7) 26.3 (4.8) 27.2 (4.6) 0.048 Body height, cm 128.7 (5.5) 127.6 (5.6) 129.6 (5.3) <0.001 BMI-SDS −0.19 (1.04) −0.19 (1.03) −0.19 (1.05) 0.95 Waist circumference, cm 56.5 (5.3) 55.8 (5.4) 57.1 (5.1) 0.008 Body fat percentage, % 19.5 (7.9) 22.1 (7.3) 17.2 (7.6) <0.001 25(OH)D, nmol/L 68.1 (22.5) 66.5 (18.9) 69.5 (25.2) 0.16 Total cholesterol, mmol/L 4.28 (0.61) 4.33 (0.61) 4.23 (0.61) 0.09 LDL cholesterol, mmol/L 2.36 (0.51) 2.42 (0.52) 2.31 (0.49) 0.032 HDL cholesterol, mmol/L 1.61 (0.31) 1.58 (0.31) 1.63 (0.31) 0.08 Triglycerides, mmol/L 0.60 (0.24) 0.62 (0.25) 0.58 (0.24) 0.08 Total physical activity, h/d 1.9 (0.7) 1.7 (0.6) 2.0 (0.72) <0.001 Total sedentary behavior, h/d 3.6 (1.6) 3.7 (1.6) 3.5 (1.6) 0.12 Average daylight time during 3 months before blood sampling, h/d 11.0 (3.8) 11.1 (3.9) 10.8 (3.7) 0.40 Vitamin D intake from food, µg/d 5.87 (2.16) 5.36 (1.66) 6.34 (2.45) <0.001 SFA intake, E% 12.1 (2.7) 12.0 (2.6) 12.2 (2.8) 0.49 MUFA intake, E% 10.0 (1.8) 9.9 (1.8) 10.1 (1.9) 0.29 PUFA intake, E% 4.9 (1.3) 4.9 (1.3) 5.0 (1.3) 0.66 Carbohydrate intake, E% 51.8 (5.0) 52.1 (4.6) 51.6 (5.3) 0.27 Fiber intake, g/1000 kcal 9.0 (2.5) 9.2 (2.4) 8.7 (2.5) 0.09 All (N = 419)a Girls (n = 195)a Boys (n = 224)a P Value Age, y 7.6 (0.4) 7.6 (0.4) 7.6 (0.4) 0.11 Parental education 0.15  Vocational school or less 81 (19.5%) 32 (16.5%) 49 (22.2%)  Polytechnic or university 334 (80.5%) 162 (83.5%) 172 (77.8%) Household income 0.62  ≤30,000 €/y 91 (22.4%) 45 (23.4%) 46 (21.4%)  >30,000 €/y 316 (77.6%) 147 (76.6%) 169 (78.6%) Body weight, kg 26.7 (4.7) 26.3 (4.8) 27.2 (4.6) 0.048 Body height, cm 128.7 (5.5) 127.6 (5.6) 129.6 (5.3) <0.001 BMI-SDS −0.19 (1.04) −0.19 (1.03) −0.19 (1.05) 0.95 Waist circumference, cm 56.5 (5.3) 55.8 (5.4) 57.1 (5.1) 0.008 Body fat percentage, % 19.5 (7.9) 22.1 (7.3) 17.2 (7.6) <0.001 25(OH)D, nmol/L 68.1 (22.5) 66.5 (18.9) 69.5 (25.2) 0.16 Total cholesterol, mmol/L 4.28 (0.61) 4.33 (0.61) 4.23 (0.61) 0.09 LDL cholesterol, mmol/L 2.36 (0.51) 2.42 (0.52) 2.31 (0.49) 0.032 HDL cholesterol, mmol/L 1.61 (0.31) 1.58 (0.31) 1.63 (0.31) 0.08 Triglycerides, mmol/L 0.60 (0.24) 0.62 (0.25) 0.58 (0.24) 0.08 Total physical activity, h/d 1.9 (0.7) 1.7 (0.6) 2.0 (0.72) <0.001 Total sedentary behavior, h/d 3.6 (1.6) 3.7 (1.6) 3.5 (1.6) 0.12 Average daylight time during 3 months before blood sampling, h/d 11.0 (3.8) 11.1 (3.9) 10.8 (3.7) 0.40 Vitamin D intake from food, µg/d 5.87 (2.16) 5.36 (1.66) 6.34 (2.45) <0.001 SFA intake, E% 12.1 (2.7) 12.0 (2.6) 12.2 (2.8) 0.49 MUFA intake, E% 10.0 (1.8) 9.9 (1.8) 10.1 (1.9) 0.29 PUFA intake, E% 4.9 (1.3) 4.9 (1.3) 5.0 (1.3) 0.66 Carbohydrate intake, E% 51.8 (5.0) 52.1 (4.6) 51.6 (5.3) 0.27 Fiber intake, g/1000 kcal 9.0 (2.5) 9.2 (2.4) 8.7 (2.5) 0.09 The values are means (standard deviations) or numbers (percentages) of children and P values for differences between girls and boys. Differences between girls and boys were tested with independent samples t test for continuous variables and Pearson χ2 test for categorical variables. Logarithmic transformation was performed for triglycerides before analysis. Abbreviations: E%, percentage of energy intake; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid. a Number of children (n) varies from 377 to 419 in different variables; n = 419, 195 girls and 224 boys: age, waist, weight, height, BMI-SDS, total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, and average daylight time; n = 408, 191 girls and 217 boys: body fat percentage; n = 415, 194 girls and 221 boys: parental education; n = 407, 192 girls and 215 boys: household income; n = 415, 194 girls and 221 boys: physical activity and sedentary behavior; and n = 377, 179 girls and 198 boys: intake of vitamin D, fiber, saturated fatty acid percentage of energy intake, monounsaturated fatty acid percentage of energy intake, polyunsaturated fatty acid percentage of energy intake, and carbohydrate percentage of energy intake. View Large Associations of serum 25(OH)D and other factors with plasma lipids Higher 25(OH)D was associated with lower total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides adjusted for age and sex (Table 2, Model 1). These negative associations of 25(OH)D with total, LDL, and HDL cholesterol, but not that with triglycerides, remained statistically significant after additional adjustment for other confounding factors (Table 2, Model 2). Table 2. Associations of Serum 25(OH)D and Other Factors With Plasma Lipids Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 25(OH)D −0.141 0.004 −0.130 0.012 −0.112 0.023 −0.109 0.032 −0.150 0.002 −0.143 0.004 −0.104 0.035 Body fat percentage 0.130 0.012 0.115 0.026 0.209 < 0.001 0.216 < 0.001 −0.195 < 0.001 −0.169 0.001 0.175 0.001 0.105 0.041 Parental education −0.024 0.63 −0.063 0.20 0.106 0.031 −0.155 0.002 −0.150 0.003 Total physical activity 0.029 0.57 −0.077 0.13 0.170 0.001 0.150 0.003 −0.158 0.002 −0.106 0.040 Total sedentary behavior 0.021 0.67 0.044 0.37 −0.043 0.39 0.106 0.041 Average daylight time 0.105 0.031 0.120 0.020 0.042 0.39 −0.020 0.69 0.139 0.004 0.113 0.026 Carbohydrate intake, E% −0.016 0.76 −0.024 0.64 −0.078 0.13 0.110 0.034 0.129 0.010 Fiber intake, g/1000 kcal −0.086 0.10 −0.034 0.52 −0.155 0.003 −0.154 0.002 −0.010 0.86 Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 25(OH)D −0.141 0.004 −0.130 0.012 −0.112 0.023 −0.109 0.032 −0.150 0.002 −0.143 0.004 −0.104 0.035 Body fat percentage 0.130 0.012 0.115 0.026 0.209 < 0.001 0.216 < 0.001 −0.195 < 0.001 −0.169 0.001 0.175 0.001 0.105 0.041 Parental education −0.024 0.63 −0.063 0.20 0.106 0.031 −0.155 0.002 −0.150 0.003 Total physical activity 0.029 0.57 −0.077 0.13 0.170 0.001 0.150 0.003 −0.158 0.002 −0.106 0.040 Total sedentary behavior 0.021 0.67 0.044 0.37 −0.043 0.39 0.106 0.041 Average daylight time 0.105 0.031 0.120 0.020 0.042 0.39 −0.020 0.69 0.139 0.004 0.113 0.026 Carbohydrate intake, E% −0.016 0.76 −0.024 0.64 −0.078 0.13 0.110 0.034 0.129 0.010 Fiber intake, g/1000 kcal −0.086 0.10 −0.034 0.52 −0.155 0.003 −0.154 0.002 −0.010 0.86 The values are standardized regression coefficients (β) and P values from linear regression models. Model 1: Each variable was entered separately in linear regression analysis with age and sex. Model 2: Age, sex, and all variables listed in the table were entered simultaneously in linear regression analysis using backward procedure. Abbreviation: E%, percentage of energy intake. Number of children (n) varies from 377 to 419 in different variables; n = 419, 195 girls and 224 boys: age, sex, cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, and average daylight time; n = 408, 191 girls and 217 boys: body fat percentage; n = 415, 194 girls and 221 boys: parental education; n = 415, 194 girls and 221 boys: physical activity and sedentary behavior; and n = 377, 179 girls and 198 boys: intake of fiber and carbohydrate percentage of energy intake. View Large Table 2. Associations of Serum 25(OH)D and Other Factors With Plasma Lipids Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 25(OH)D −0.141 0.004 −0.130 0.012 −0.112 0.023 −0.109 0.032 −0.150 0.002 −0.143 0.004 −0.104 0.035 Body fat percentage 0.130 0.012 0.115 0.026 0.209 < 0.001 0.216 < 0.001 −0.195 < 0.001 −0.169 0.001 0.175 0.001 0.105 0.041 Parental education −0.024 0.63 −0.063 0.20 0.106 0.031 −0.155 0.002 −0.150 0.003 Total physical activity 0.029 0.57 −0.077 0.13 0.170 0.001 0.150 0.003 −0.158 0.002 −0.106 0.040 Total sedentary behavior 0.021 0.67 0.044 0.37 −0.043 0.39 0.106 0.041 Average daylight time 0.105 0.031 0.120 0.020 0.042 0.39 −0.020 0.69 0.139 0.004 0.113 0.026 Carbohydrate intake, E% −0.016 0.76 −0.024 0.64 −0.078 0.13 0.110 0.034 0.129 0.010 Fiber intake, g/1000 kcal −0.086 0.10 −0.034 0.52 −0.155 0.003 −0.154 0.002 −0.010 0.86 Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 25(OH)D −0.141 0.004 −0.130 0.012 −0.112 0.023 −0.109 0.032 −0.150 0.002 −0.143 0.004 −0.104 0.035 Body fat percentage 0.130 0.012 0.115 0.026 0.209 < 0.001 0.216 < 0.001 −0.195 < 0.001 −0.169 0.001 0.175 0.001 0.105 0.041 Parental education −0.024 0.63 −0.063 0.20 0.106 0.031 −0.155 0.002 −0.150 0.003 Total physical activity 0.029 0.57 −0.077 0.13 0.170 0.001 0.150 0.003 −0.158 0.002 −0.106 0.040 Total sedentary behavior 0.021 0.67 0.044 0.37 −0.043 0.39 0.106 0.041 Average daylight time 0.105 0.031 0.120 0.020 0.042 0.39 −0.020 0.69 0.139 0.004 0.113 0.026 Carbohydrate intake, E% −0.016 0.76 −0.024 0.64 −0.078 0.13 0.110 0.034 0.129 0.010 Fiber intake, g/1000 kcal −0.086 0.10 −0.034 0.52 −0.155 0.003 −0.154 0.002 −0.010 0.86 The values are standardized regression coefficients (β) and P values from linear regression models. Model 1: Each variable was entered separately in linear regression analysis with age and sex. Model 2: Age, sex, and all variables listed in the table were entered simultaneously in linear regression analysis using backward procedure. Abbreviation: E%, percentage of energy intake. Number of children (n) varies from 377 to 419 in different variables; n = 419, 195 girls and 224 boys: age, sex, cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, and average daylight time; n = 408, 191 girls and 217 boys: body fat percentage; n = 415, 194 girls and 221 boys: parental education; n = 415, 194 girls and 221 boys: physical activity and sedentary behavior; and n = 377, 179 girls and 198 boys: intake of fiber and carbohydrate percentage of energy intake. View Large Children in the highest quartile of 25(OH)D (>79 nmol/L) had the lowest total cholesterol adjusted for age and sex and after additional adjustment for body fat percentage and average daylight time [Fig. 1(a)]. Children in the highest quartile of 25(OH)D also had the lowest LDL cholesterol adjusted for age and sex and after further adjustment for body fat percentage [Fig. 1(b)]. The differences in HDL cholesterol [Fig. 1(c)] or triglycerides [Fig. 1(d)] across the quartiles of 25(OH)D were not statistically significant. Figure 1. View largeDownload slide Mean (95% CI) plasma total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides in quartiles of serum 25(OH)D. (a) Mean (95% CI) plasma total cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 3.244, P = 0.022; difference between first and fourth quartile: P = 0.022. Gray: adjusted for age, sex, body fat percentage, and average daylight time 3 months before blood sampling. Difference across quartiles: F = 3.477, P = 0.016; difference between first and fourth quartile: P = 0.015. (b) Mean (95% CI) plasma LDL cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 3.122, P = 0.026; difference between first and fourth quartile: P = 0.020. Gray: adjusted for age, sex, and body fat percentage. Difference across quartiles: F = 2.881, P = 0.036; difference between first and fourth quartile: P = 0.029. (c) Mean (95% CI) plasma HDL cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 2.079, P = 0.102. Gray: adjusted for age, sex, body fat percentage, physical activity, and fiber intake. Difference across quartiles: F = 1.895, P = 0.130. (d) Mean (95% CI) plasma triglycerides in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 2.235, P = 0.084. Gray: adjusted for age, sex, body fat percentage, parental education, physical activity, average daylight time, and carbohydrate intake as percentage of energy intake. Difference across quartiles: F = 1.678, P = 0.171. Figure 1. View largeDownload slide Mean (95% CI) plasma total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides in quartiles of serum 25(OH)D. (a) Mean (95% CI) plasma total cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 3.244, P = 0.022; difference between first and fourth quartile: P = 0.022. Gray: adjusted for age, sex, body fat percentage, and average daylight time 3 months before blood sampling. Difference across quartiles: F = 3.477, P = 0.016; difference between first and fourth quartile: P = 0.015. (b) Mean (95% CI) plasma LDL cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 3.122, P = 0.026; difference between first and fourth quartile: P = 0.020. Gray: adjusted for age, sex, and body fat percentage. Difference across quartiles: F = 2.881, P = 0.036; difference between first and fourth quartile: P = 0.029. (c) Mean (95% CI) plasma HDL cholesterol in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 2.079, P = 0.102. Gray: adjusted for age, sex, body fat percentage, physical activity, and fiber intake. Difference across quartiles: F = 1.895, P = 0.130. (d) Mean (95% CI) plasma triglycerides in quartiles of serum 25(OH)D. Black: adjusted for age and sex. Difference across quartiles: F = 2.235, P = 0.084. Gray: adjusted for age, sex, body fat percentage, parental education, physical activity, average daylight time, and carbohydrate intake as percentage of energy intake. Difference across quartiles: F = 1.678, P = 0.171. Of other factors, higher body fat percentage was associated with higher total and LDL cholesterol, higher triglycerides and lower HDL cholesterol, higher levels of physical activity with higher HDL cholesterol and lower triglycerides, longer average daylight time with higher total cholesterol and triglycerides, a lower intake of dietary fiber with higher HDL cholesterol, and a higher intake of carbohydrates and lower parental education with higher triglycerides adjusted for confounding factors (Table 2, Model 2). Associations of gene variants with 25(OH)D and lipids The G allele of rs2282679 in DBP and the A allele of rs12794714 in CYP2R1 were negatively associated and the A allele of rs10741657 in CYP2R1 was positively associated with 25(OH)D adjusted for age and sex (Table 3). The G allele of rs6599638 in C10orf88 was positively associated with HDL cholesterol adjusted for age and sex (Table 3) and after additional adjustment for 25(OH)D (P for linear trend = 0.021). The A allele of rs12794714 in CYP2R1 was negatively associated with total and LDL cholesterol adjusted for age and sex (Table 3). The associations of rs12794714 in CYP2R1 with total cholesterol (P for linear trend <0.001) and LDL cholesterol (P for linear trend = 0.007) remained after further adjustment for 25(OH)D. The associations of 25(OH)D with total, LDL, and HDL cholesterol and triglycerides remained after further adjustments for the SNPs. There was no interaction between any SNP and 25(OH)D on lipids. Table 3. Associations of Gene Variants With Serum 25(OH)D and Plasma Lipids SNP, Nearest Gene(s), Chromosome Genotypes n 25(OH)D Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Rs12785878 T/T 163 (40.2%) 69.4 (65.9–72.9) 4.25 (4.15–4.34) 2.35 (2.28–2.43) 1.60 (1.55–1.65) 0.59 (0.55–0.63) NADSYN1/DHCR7 T/G 178 (44.0%) 66.5 (63.2–69.8) 4.36 (4.27–4.45) 2.42 (2.35–2.45) 1.62 (1.57–1.66) 0.61 (0.57–0.64) chr 11 G/G 63 (15.8%) MAF: 0.362 67.4 (61.9–73.9) 4.14 (3.99–4.29) 2.23 (2.11–2.36) 1.61 (1.53–1.68) 0.55 (0.49–0.62) p1 = 0.54 p1 = 0.040 p1 = 0.037 p1 = 0.91 p1 = 0.19 p2 = 0.47 p2 = 0.69 p2 = 0.37 p2 = 0.80 p2 = 0.41 Rs3829251 G/G 215 (53.3%) 68.4 (65.3–71.4) 4.25 (4.18–4.34) 2.37 (2.30–2.44) 1.59 (1.55–1.63) 0.59 (0.56–0.63) NADSYN1/DHCR7 G/A 151 (37.5%) 66.9 (63.4–70.5) 4.33 (4.23–4.43) 2.40 (2.32–2.48) 1.61 (1.56–1.66) 0.61 (0.57–0.64) chr 11 A/A 37 (9.2%) MAF: 0.279 68.7 (61.4–76.0) 4.13 (3.93–4.33) 2.13 (1.98–2.31) 1.68 (1.58.1.78) 0.50 (0.42–0.58) p1 = 0.78 p1 = 0.18 p1 = 0.020 p1 = 0.27 p1 = 0.022 p2 = 0.96 p2 = 0.69 p2 = 0.11 p2 = 0.13 p2 = 0.10 rs6599638 A/A 129 (31.9%) 68.7 (64.8–72.6) 4.23 (4.13–4.34) 2.38 (2.29–2.46) 1.56 (1.51–1.61) 0.59 (0.55–0.64) C10orf88 A/G 208 (51.4%) 67.4 (64.4–70.5) 4.30 (4.21–4.38) 2.38 (2.31–2.45) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 10 G/G 68 (16.8%) MAF: 0.377 68.4 (63.1–73.8) 4.29 (4.15–4.44) 2.29 (2.17–2.41) 1.66 (1.59–1.73) 0.60 (0.53–0.65) p1 = 0.85 p1 = 0.61 p1 = 0.46 p1 = 0.07 p1 = 0.93 p2 = 0.73 p2 = 0.39 p2 = 0.36 p2 = 0.020 p2 = 0.75 rs2282679 T/T 277 (66.1%) 70.0 (67.3–72.6) 4.30 (4.23–4.37) 2.38 (2.32–2.44) 1.61 (1.57–1.64) 0.60 (0.57–0.62) DBP T/G 112 (27.7%) 63.7 (59.5–67.9) 4.20 (4.08–4.31) 2.29 (2.20–2.39) 1.62 (1.56–1.68) 0.56 (0.52–0.61) chr 4 G/G 16 (3.8%) MAF: 0.178 65.6 (54.7–76.5) 4.45 (4.14–4.75) 2.61 (2.36–2.86) 1.51 (1.35–1.66) 0.78 (0.66–0.90) p1 = 0.022 p1 = 0.19 p1 = 0.05 p1 = 0.37 p1 = 0.006 p2 = 0.004 p2 = 0.52 p2 = 0.99 p2 = 0.43 p2 = 0.33 rs10741657 G/G 123 (30.4%) 66.0 (62.1–70.0) 4.26 (4.15–4.37) 2.37 (2.28–2.46) 1.60 (1.54–1.65) 0.59 (0.55–0.63) CYP2R1 G/A 214 (52.8%) 66.8 (63.8–69.8) 4.27 (4.19–4.36) 2.35 (2.28–2.42) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 11 A/A 68 (16.8%) MAF: 0.432 76.2 (70.7–81.6) 4.33 (4.18–4.48) 2.39 (2.27–2.52) 1.60 (1.52–1.67) 0.60 (0.54–0.66) p1 = 0.004 p1 = 0.75 p1 = 0.82 p1 = 0.81 p1 = 0.97 p2 = 0.006 p2 = 0.60 p2 = 0.94 p2 = 0.90 p2 = 0.80 rs12794714 G/G 152 (36.3%) 72.0 (68.5–75.6) 4.38 (4.28–4.48) 2.44 (2.36–2.52) 1.60 (1.56–1.65) 0.61 (0.58–0.65) CYP2R1 G/A 204 (50.4%) 66.3 (63.2–69.3) 4.25 (4.17–4.33) 2.33 (2.26–2.40) 1.63 (1.59–1.67) 0.58 (0.55–0.62) chr 11 A/A 49 (12.1%) MAF: 0.373 63.0 (56.7–69.3) 4.09 (3.92–4.26) 2.28 (2.13–2.42) 1.52 (1.44–1.61) 0.58 (0.51–0.65) p1 = 0.023 p1 = 0.010 p1 = 0.06 p1 = 0.09 p1 = 0.35 p2 = 0.005 p2 = 0.003 p2 = 0.019 p2 = 0.38 p2 = 0.23 SNP, Nearest Gene(s), Chromosome Genotypes n 25(OH)D Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Rs12785878 T/T 163 (40.2%) 69.4 (65.9–72.9) 4.25 (4.15–4.34) 2.35 (2.28–2.43) 1.60 (1.55–1.65) 0.59 (0.55–0.63) NADSYN1/DHCR7 T/G 178 (44.0%) 66.5 (63.2–69.8) 4.36 (4.27–4.45) 2.42 (2.35–2.45) 1.62 (1.57–1.66) 0.61 (0.57–0.64) chr 11 G/G 63 (15.8%) MAF: 0.362 67.4 (61.9–73.9) 4.14 (3.99–4.29) 2.23 (2.11–2.36) 1.61 (1.53–1.68) 0.55 (0.49–0.62) p1 = 0.54 p1 = 0.040 p1 = 0.037 p1 = 0.91 p1 = 0.19 p2 = 0.47 p2 = 0.69 p2 = 0.37 p2 = 0.80 p2 = 0.41 Rs3829251 G/G 215 (53.3%) 68.4 (65.3–71.4) 4.25 (4.18–4.34) 2.37 (2.30–2.44) 1.59 (1.55–1.63) 0.59 (0.56–0.63) NADSYN1/DHCR7 G/A 151 (37.5%) 66.9 (63.4–70.5) 4.33 (4.23–4.43) 2.40 (2.32–2.48) 1.61 (1.56–1.66) 0.61 (0.57–0.64) chr 11 A/A 37 (9.2%) MAF: 0.279 68.7 (61.4–76.0) 4.13 (3.93–4.33) 2.13 (1.98–2.31) 1.68 (1.58.1.78) 0.50 (0.42–0.58) p1 = 0.78 p1 = 0.18 p1 = 0.020 p1 = 0.27 p1 = 0.022 p2 = 0.96 p2 = 0.69 p2 = 0.11 p2 = 0.13 p2 = 0.10 rs6599638 A/A 129 (31.9%) 68.7 (64.8–72.6) 4.23 (4.13–4.34) 2.38 (2.29–2.46) 1.56 (1.51–1.61) 0.59 (0.55–0.64) C10orf88 A/G 208 (51.4%) 67.4 (64.4–70.5) 4.30 (4.21–4.38) 2.38 (2.31–2.45) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 10 G/G 68 (16.8%) MAF: 0.377 68.4 (63.1–73.8) 4.29 (4.15–4.44) 2.29 (2.17–2.41) 1.66 (1.59–1.73) 0.60 (0.53–0.65) p1 = 0.85 p1 = 0.61 p1 = 0.46 p1 = 0.07 p1 = 0.93 p2 = 0.73 p2 = 0.39 p2 = 0.36 p2 = 0.020 p2 = 0.75 rs2282679 T/T 277 (66.1%) 70.0 (67.3–72.6) 4.30 (4.23–4.37) 2.38 (2.32–2.44) 1.61 (1.57–1.64) 0.60 (0.57–0.62) DBP T/G 112 (27.7%) 63.7 (59.5–67.9) 4.20 (4.08–4.31) 2.29 (2.20–2.39) 1.62 (1.56–1.68) 0.56 (0.52–0.61) chr 4 G/G 16 (3.8%) MAF: 0.178 65.6 (54.7–76.5) 4.45 (4.14–4.75) 2.61 (2.36–2.86) 1.51 (1.35–1.66) 0.78 (0.66–0.90) p1 = 0.022 p1 = 0.19 p1 = 0.05 p1 = 0.37 p1 = 0.006 p2 = 0.004 p2 = 0.52 p2 = 0.99 p2 = 0.43 p2 = 0.33 rs10741657 G/G 123 (30.4%) 66.0 (62.1–70.0) 4.26 (4.15–4.37) 2.37 (2.28–2.46) 1.60 (1.54–1.65) 0.59 (0.55–0.63) CYP2R1 G/A 214 (52.8%) 66.8 (63.8–69.8) 4.27 (4.19–4.36) 2.35 (2.28–2.42) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 11 A/A 68 (16.8%) MAF: 0.432 76.2 (70.7–81.6) 4.33 (4.18–4.48) 2.39 (2.27–2.52) 1.60 (1.52–1.67) 0.60 (0.54–0.66) p1 = 0.004 p1 = 0.75 p1 = 0.82 p1 = 0.81 p1 = 0.97 p2 = 0.006 p2 = 0.60 p2 = 0.94 p2 = 0.90 p2 = 0.80 rs12794714 G/G 152 (36.3%) 72.0 (68.5–75.6) 4.38 (4.28–4.48) 2.44 (2.36–2.52) 1.60 (1.56–1.65) 0.61 (0.58–0.65) CYP2R1 G/A 204 (50.4%) 66.3 (63.2–69.3) 4.25 (4.17–4.33) 2.33 (2.26–2.40) 1.63 (1.59–1.67) 0.58 (0.55–0.62) chr 11 A/A 49 (12.1%) MAF: 0.373 63.0 (56.7–69.3) 4.09 (3.92–4.26) 2.28 (2.13–2.42) 1.52 (1.44–1.61) 0.58 (0.51–0.65) p1 = 0.023 p1 = 0.010 p1 = 0.06 p1 = 0.09 p1 = 0.35 p2 = 0.005 p2 = 0.003 p2 = 0.019 p2 = 0.38 p2 = 0.23 SNPs, nearest genes, genotypes, and MAFs. The values are numbers of subjects (percentages) and means (95% CIs) from analysis of variances adjusted for age and sex. p1 signifies P value for the difference across groups, p2 signifies P value for linear trend. All SNPs were in Hardy-Weinberg equilibrium. Abbreviation: chr, chromosome. View Large Table 3. Associations of Gene Variants With Serum 25(OH)D and Plasma Lipids SNP, Nearest Gene(s), Chromosome Genotypes n 25(OH)D Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Rs12785878 T/T 163 (40.2%) 69.4 (65.9–72.9) 4.25 (4.15–4.34) 2.35 (2.28–2.43) 1.60 (1.55–1.65) 0.59 (0.55–0.63) NADSYN1/DHCR7 T/G 178 (44.0%) 66.5 (63.2–69.8) 4.36 (4.27–4.45) 2.42 (2.35–2.45) 1.62 (1.57–1.66) 0.61 (0.57–0.64) chr 11 G/G 63 (15.8%) MAF: 0.362 67.4 (61.9–73.9) 4.14 (3.99–4.29) 2.23 (2.11–2.36) 1.61 (1.53–1.68) 0.55 (0.49–0.62) p1 = 0.54 p1 = 0.040 p1 = 0.037 p1 = 0.91 p1 = 0.19 p2 = 0.47 p2 = 0.69 p2 = 0.37 p2 = 0.80 p2 = 0.41 Rs3829251 G/G 215 (53.3%) 68.4 (65.3–71.4) 4.25 (4.18–4.34) 2.37 (2.30–2.44) 1.59 (1.55–1.63) 0.59 (0.56–0.63) NADSYN1/DHCR7 G/A 151 (37.5%) 66.9 (63.4–70.5) 4.33 (4.23–4.43) 2.40 (2.32–2.48) 1.61 (1.56–1.66) 0.61 (0.57–0.64) chr 11 A/A 37 (9.2%) MAF: 0.279 68.7 (61.4–76.0) 4.13 (3.93–4.33) 2.13 (1.98–2.31) 1.68 (1.58.1.78) 0.50 (0.42–0.58) p1 = 0.78 p1 = 0.18 p1 = 0.020 p1 = 0.27 p1 = 0.022 p2 = 0.96 p2 = 0.69 p2 = 0.11 p2 = 0.13 p2 = 0.10 rs6599638 A/A 129 (31.9%) 68.7 (64.8–72.6) 4.23 (4.13–4.34) 2.38 (2.29–2.46) 1.56 (1.51–1.61) 0.59 (0.55–0.64) C10orf88 A/G 208 (51.4%) 67.4 (64.4–70.5) 4.30 (4.21–4.38) 2.38 (2.31–2.45) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 10 G/G 68 (16.8%) MAF: 0.377 68.4 (63.1–73.8) 4.29 (4.15–4.44) 2.29 (2.17–2.41) 1.66 (1.59–1.73) 0.60 (0.53–0.65) p1 = 0.85 p1 = 0.61 p1 = 0.46 p1 = 0.07 p1 = 0.93 p2 = 0.73 p2 = 0.39 p2 = 0.36 p2 = 0.020 p2 = 0.75 rs2282679 T/T 277 (66.1%) 70.0 (67.3–72.6) 4.30 (4.23–4.37) 2.38 (2.32–2.44) 1.61 (1.57–1.64) 0.60 (0.57–0.62) DBP T/G 112 (27.7%) 63.7 (59.5–67.9) 4.20 (4.08–4.31) 2.29 (2.20–2.39) 1.62 (1.56–1.68) 0.56 (0.52–0.61) chr 4 G/G 16 (3.8%) MAF: 0.178 65.6 (54.7–76.5) 4.45 (4.14–4.75) 2.61 (2.36–2.86) 1.51 (1.35–1.66) 0.78 (0.66–0.90) p1 = 0.022 p1 = 0.19 p1 = 0.05 p1 = 0.37 p1 = 0.006 p2 = 0.004 p2 = 0.52 p2 = 0.99 p2 = 0.43 p2 = 0.33 rs10741657 G/G 123 (30.4%) 66.0 (62.1–70.0) 4.26 (4.15–4.37) 2.37 (2.28–2.46) 1.60 (1.54–1.65) 0.59 (0.55–0.63) CYP2R1 G/A 214 (52.8%) 66.8 (63.8–69.8) 4.27 (4.19–4.36) 2.35 (2.28–2.42) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 11 A/A 68 (16.8%) MAF: 0.432 76.2 (70.7–81.6) 4.33 (4.18–4.48) 2.39 (2.27–2.52) 1.60 (1.52–1.67) 0.60 (0.54–0.66) p1 = 0.004 p1 = 0.75 p1 = 0.82 p1 = 0.81 p1 = 0.97 p2 = 0.006 p2 = 0.60 p2 = 0.94 p2 = 0.90 p2 = 0.80 rs12794714 G/G 152 (36.3%) 72.0 (68.5–75.6) 4.38 (4.28–4.48) 2.44 (2.36–2.52) 1.60 (1.56–1.65) 0.61 (0.58–0.65) CYP2R1 G/A 204 (50.4%) 66.3 (63.2–69.3) 4.25 (4.17–4.33) 2.33 (2.26–2.40) 1.63 (1.59–1.67) 0.58 (0.55–0.62) chr 11 A/A 49 (12.1%) MAF: 0.373 63.0 (56.7–69.3) 4.09 (3.92–4.26) 2.28 (2.13–2.42) 1.52 (1.44–1.61) 0.58 (0.51–0.65) p1 = 0.023 p1 = 0.010 p1 = 0.06 p1 = 0.09 p1 = 0.35 p2 = 0.005 p2 = 0.003 p2 = 0.019 p2 = 0.38 p2 = 0.23 SNP, Nearest Gene(s), Chromosome Genotypes n 25(OH)D Total Cholesterol LDL Cholesterol HDL Cholesterol Triglycerides Rs12785878 T/T 163 (40.2%) 69.4 (65.9–72.9) 4.25 (4.15–4.34) 2.35 (2.28–2.43) 1.60 (1.55–1.65) 0.59 (0.55–0.63) NADSYN1/DHCR7 T/G 178 (44.0%) 66.5 (63.2–69.8) 4.36 (4.27–4.45) 2.42 (2.35–2.45) 1.62 (1.57–1.66) 0.61 (0.57–0.64) chr 11 G/G 63 (15.8%) MAF: 0.362 67.4 (61.9–73.9) 4.14 (3.99–4.29) 2.23 (2.11–2.36) 1.61 (1.53–1.68) 0.55 (0.49–0.62) p1 = 0.54 p1 = 0.040 p1 = 0.037 p1 = 0.91 p1 = 0.19 p2 = 0.47 p2 = 0.69 p2 = 0.37 p2 = 0.80 p2 = 0.41 Rs3829251 G/G 215 (53.3%) 68.4 (65.3–71.4) 4.25 (4.18–4.34) 2.37 (2.30–2.44) 1.59 (1.55–1.63) 0.59 (0.56–0.63) NADSYN1/DHCR7 G/A 151 (37.5%) 66.9 (63.4–70.5) 4.33 (4.23–4.43) 2.40 (2.32–2.48) 1.61 (1.56–1.66) 0.61 (0.57–0.64) chr 11 A/A 37 (9.2%) MAF: 0.279 68.7 (61.4–76.0) 4.13 (3.93–4.33) 2.13 (1.98–2.31) 1.68 (1.58.1.78) 0.50 (0.42–0.58) p1 = 0.78 p1 = 0.18 p1 = 0.020 p1 = 0.27 p1 = 0.022 p2 = 0.96 p2 = 0.69 p2 = 0.11 p2 = 0.13 p2 = 0.10 rs6599638 A/A 129 (31.9%) 68.7 (64.8–72.6) 4.23 (4.13–4.34) 2.38 (2.29–2.46) 1.56 (1.51–1.61) 0.59 (0.55–0.64) C10orf88 A/G 208 (51.4%) 67.4 (64.4–70.5) 4.30 (4.21–4.38) 2.38 (2.31–2.45) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 10 G/G 68 (16.8%) MAF: 0.377 68.4 (63.1–73.8) 4.29 (4.15–4.44) 2.29 (2.17–2.41) 1.66 (1.59–1.73) 0.60 (0.53–0.65) p1 = 0.85 p1 = 0.61 p1 = 0.46 p1 = 0.07 p1 = 0.93 p2 = 0.73 p2 = 0.39 p2 = 0.36 p2 = 0.020 p2 = 0.75 rs2282679 T/T 277 (66.1%) 70.0 (67.3–72.6) 4.30 (4.23–4.37) 2.38 (2.32–2.44) 1.61 (1.57–1.64) 0.60 (0.57–0.62) DBP T/G 112 (27.7%) 63.7 (59.5–67.9) 4.20 (4.08–4.31) 2.29 (2.20–2.39) 1.62 (1.56–1.68) 0.56 (0.52–0.61) chr 4 G/G 16 (3.8%) MAF: 0.178 65.6 (54.7–76.5) 4.45 (4.14–4.75) 2.61 (2.36–2.86) 1.51 (1.35–1.66) 0.78 (0.66–0.90) p1 = 0.022 p1 = 0.19 p1 = 0.05 p1 = 0.37 p1 = 0.006 p2 = 0.004 p2 = 0.52 p2 = 0.99 p2 = 0.43 p2 = 0.33 rs10741657 G/G 123 (30.4%) 66.0 (62.1–70.0) 4.26 (4.15–4.37) 2.37 (2.28–2.46) 1.60 (1.54–1.65) 0.59 (0.55–0.63) CYP2R1 G/A 214 (52.8%) 66.8 (63.8–69.8) 4.27 (4.19–4.36) 2.35 (2.28–2.42) 1.62 (1.58–1.66) 0.59 (0.56–0.63) chr 11 A/A 68 (16.8%) MAF: 0.432 76.2 (70.7–81.6) 4.33 (4.18–4.48) 2.39 (2.27–2.52) 1.60 (1.52–1.67) 0.60 (0.54–0.66) p1 = 0.004 p1 = 0.75 p1 = 0.82 p1 = 0.81 p1 = 0.97 p2 = 0.006 p2 = 0.60 p2 = 0.94 p2 = 0.90 p2 = 0.80 rs12794714 G/G 152 (36.3%) 72.0 (68.5–75.6) 4.38 (4.28–4.48) 2.44 (2.36–2.52) 1.60 (1.56–1.65) 0.61 (0.58–0.65) CYP2R1 G/A 204 (50.4%) 66.3 (63.2–69.3) 4.25 (4.17–4.33) 2.33 (2.26–2.40) 1.63 (1.59–1.67) 0.58 (0.55–0.62) chr 11 A/A 49 (12.1%) MAF: 0.373 63.0 (56.7–69.3) 4.09 (3.92–4.26) 2.28 (2.13–2.42) 1.52 (1.44–1.61) 0.58 (0.51–0.65) p1 = 0.023 p1 = 0.010 p1 = 0.06 p1 = 0.09 p1 = 0.35 p2 = 0.005 p2 = 0.003 p2 = 0.019 p2 = 0.38 p2 = 0.23 SNPs, nearest genes, genotypes, and MAFs. The values are numbers of subjects (percentages) and means (95% CIs) from analysis of variances adjusted for age and sex. p1 signifies P value for the difference across groups, p2 signifies P value for linear trend. All SNPs were in Hardy-Weinberg equilibrium. Abbreviation: chr, chromosome. View Large Discussion In our population study among prepubertal children, higher serum 25(OH)D was associated with lower plasma total, LDL, and HDL cholesterol and triglycerides. The associations of 25(OH)D with total, LDL, and HDL cholesterol but not with triglycerides remained after controlling for all confounding factors. The A allele of rs12794714 in CYP2R1 was negatively associated and the A allele of rs10741657 in CYP2R1 was positively associated with 25(OH)D, and the G allele of rs2282679 in DBP was negatively associated with 25(OH)D. However, these SNPs did not explain or modify the associations of 25(OH)D with lipids. Moreover, the allele A of rs12794714 in CYP2R1 was negatively associated with total and LDL cholesterol, and the G allele of rs6599638 in C10orf88 was positively associated with HDL cholesterol even when adjusted for 25(OH)D. Serum 25(OH)D levels below 30 to 50 nmol/L have been determined as vitamin D deficiency (34, 35), and some authors have suggested that the lower limit for the sufficient level could be as high as 75 nmol/L (35). A review and meta-analysis in adults found an inverse association between 25(OH)D and the risk of cardiovascular diseases at 25(OH)D of 20 to 60 nmol/L but not above this level (3). We found the lowest total and LDL cholesterol levels above 79 nmol/L, representing the highest quartile of 25(OH)D, that is consistent with the higher suggested limit for the sufficient level of 25(OH)D (35). Many studies in children have not found an association between 25(OH)D and total cholesterol (7, 11, 36). An inverse association between 25(OH)D and total cholesterol has been reported in some studies among children (8, 37), whereas one study observed a positive relationship in girls (13). In most of the studies among children and adolescents, there has been no association between 25(OH)D and LDL cholesterol (4–7, 13, 36, 37). However, 25(OH)D has been inversely associated with LDL cholesterol in some pediatric studies (8, 11) and was positively related to LDL cholesterol in one study among obese female adolescents (38). A review and meta-analysis including mainly children and adolescents found weak inverse associations of 25(OH)D with total and LDL cholesterol (39). Our study confirms the inverse associations of 25(OH)D with total and LDL cholesterol among children. Importantly, the associations remained even though several confounding factors, including body fat percentage, dietary factors, physical activity, sedentary behavior, daylight time, and socioeconomic status, were taken into account. One reason for the discrepancy between the results of some previous studies may be that confounding factors have not been taken into account in all studies. In addition, some studies have not measured serum lipids using fasting samples. The association between 25(OH)D and HDL cholesterol has been positive in many studies among children and adolescents (5, 6, 12, 40), but several studies have not observed such an association (8, 11, 13, 36). A review and meta-analysis that included mainly children and adolescents found a weak positive association between 25(OH)D and HDL cholesterol (39). The inverse association between 25(OH)D and HDL cholesterol that was found in the current study has previously been reported only in infants (37). It is possible that the association is different in older children with more advanced puberty. In most of the pediatric studies, 25(OH)D has been inversely associated with triglycerides (6–8, 37, 40), but one study found a positive association in girls (13), and several studies have observed no association (5, 11, 12, 36). In line with many previous studies, we found that 25(OH)D was inversely associated with triglycerides, but the relationship was partly explained by confounding factors. Cholesterol and vitamin D are synthesized from a common precursor, 7-dehydrocholesterol. DHCR7 converts 7-dehydrocholesterol to cholesterol. However, in the presence of ultraviolet B radiation in the skin, 7-dehydrocholesterol can be converted to previtamin D3 and further to vitamin D3 (1). Based on this metabolic pathway, one could have expected lower cholesterol levels in summer when daylight time is longer and that this could be one of the reasons for the inverse association between 25(OH)D and cholesterol. However, we found weak positive associations of daylight time with total cholesterol and triglycerides, and daylight time did not explain the association of 25(OH)D with total, LDL, or HDL cholesterol. One of the reasons for this may be that daylight time is a less important determinant of 25(OH)D than dietary intake of vitamin D in our study population from the northern latitude (25). Moreover, SNPs related to DHCR7 involved in cholesterol and vitamin D synthesis in the skin were not associated with 25(OH)D or lipid levels. Altogether, these findings suggest that vitamin D metabolism in the skin may not explain the association between 25(OH)D and lipids in the current study. The inverse associations of 25(OH)D with total, LDL, and HDL cholesterol could also be related to liver metabolism. Vitamin D receptor has been shown to downregulate the small-heterodimer partner and increase cholesterol 7α-hydroxylase in the liver, leading to a higher metabolism of cholesterol to bile acids and thus lower cholesterol levels (19). Moreover, genetic factors could partly explain the associations of 25(OH)D with lipids. An SNP in APOA5 that is involved in cholesterol metabolism has modified the association between 25(OH)D and HDL cholesterol (41), but this gene variant was not included in the current analyses. Finally, there may also be some indirect mechanisms for the associations of 25(OH)D with lipids, such as the effects of parathyroid hormone and calcium metabolism. Further studies on the mechanisms between 25(OH)D and lipids are needed. We investigated six SNPs in genes in the vitamin D pathway that have been associated with serum 25(OH)D in GWASs (14, 15). Consistent with the GWAS results, rs10741657 and rs12794714 in CYP2R1 and rs2282679 in DBP were associated with 25(OH)D. However, these gene variants did not explain the association between 25(OH)D and lipids. Moreover, we observed that rs12794714 in CYP2R1 was associated with total cholesterol and LDL cholesterol and that rs6599638 near gene C10orf88 was associated with HDL cholesterol even after controlling for 25(OH)D. The associations of 25(OH)D with lipids did not depend on the SNPs, and the associations of rs12794714 and rs6599638 with lipids were independent of 25(OH)D. CYP2R1 is the main enzyme converting vitamin D into 25(OH)D in the liver (1) and is a member of CYP450 family of enzymes, some of which are involved in cholesterol synthesis (42). One of the explanations for the associations of rs12794714 in CYP2R1 with total and LDL cholesterol could be that CYP2R1 is also such an enzyme. C10orf88 is near a gene coding acyl-coenzyme A dehydrogenase involved in producing substrates for cholesterol synthesis. This could be a mechanism for the association between the SNP in C10orf88 and HDL cholesterol in the current study. Gene variants in DBP and NADSYN/DHCR7 were associated with increased risk of dyslipidemia in adults of African descent (16). However, we found no associations of these gene variants with lipids in children. The strength of our study is a population sample of children with a low prevalence of diseases and medications possibly affecting the association between 25(OH)D and lipids. Moreover, we excluded children who had such diseases or medications or had entered puberty to avoid associated confounding. We took several possible confounding factors, including body fat percentage, physical activity, sedentary behavior dietary factors, daylight time, and socioeconomic status, into account in the analyses. However, we cannot exclude residual confounding due to some unmeasured factors. The number of children who were homozygous for the rare allele of rs2282679 in DBP was small, which limited statistical power in the analyses. Finally, our results are based on cross-sectional analyses, and it is therefore not possible to draw a conclusion on the causality of the associations. Conclusion Serum 25(OH)D was associated with lower total, LDL, and HDL cholesterol independent of body fat percentage, dietary factors, physical activity, sedentary behavior, daylight time, and socioeconomic status. Children having serum 25(OH)D over 79 nmol/L had the lowest total and LDL cholesterol. Consistent with earlier findings (15), rs12794714 and rs10741657 in CYP2R1 and rs2282679 in DBP were associated with 25(OH)D. A new observation of the study is that rs12794714 in CYP2R1 was also associated with total and LDL cholesterol and rs6599638 in C10orf88 with HDL cholesterol even after controlling for 25(OH)D. However, none of the gene variants explained or modified the associations of 25(OH)D with lipids. Further studies are needed to confirm our findings and to detect mechanisms for the associations between 25(OH)D and lipids. Abbreviations: Abbreviations: 25(OH)D 25-hydroxyvitamin D BMI-SDS body mass index standard deviation score C10orf88 chromosome 10 open reading frame 88 CYP2R1 cytochrome P450 family 2 subfamily R member 1 DBP vitamin D binding protein DHCR7 7-dehydrocholesterol reductase GWAS genome-wide association study HDL high-density lipoprotein LDL low-density lipoprotein MAF minor allele frequency NADSYN1 NAD synthetase 1 PANIC Physical Activity and Nutrition in Children SNP single nucleotide polymorphism Acknowledgments The authors are grateful to all the children and their parents for participating in the PANIC study. The authors are also indebted to the members of the PANIC research team for their skillful contribution in performing the study. We also thank Sami Heikkinen for help with genetic data and Juuso Väistö for help with editing the figures. Financial Support: This work was supported by grants from Ministry of Social Affairs and Health of Finland, Ministry of Education and Culture of Finland, Finnish Innovation Fund Sitra, Social Insurance Institution of Finland, Finnish Cultural Foundation, Juho Vainio Foundation, Foundation for Pediatric Research, Doctoral Programs in Public Health, Paavo Nurmi Foundation, Paulo Foundation, Diabetes Research Foundation, Yrjö Jahnsson Foundation, Finnish Foundation for Cardiovascular Research, Orion Research Foundation sr, Research Committee of the Kuopio University Hospital Catchment Area (State Research Funding), Kuopio University Hospital [previous state research funding (EVO), funding no. 5031343], and the city of Kuopio. Clinical Trial Information: ClinicalTrials.gov no. NCT01803776 (registered 4 March 2013). Author Contributions: S.S. participated in the collection of data, conducted the statistical analyses, and wrote the draft of the manuscript. A.-M.E., V.L., and A.V. participated in data collection and contributed to the critical revision of the manuscript. G.D., A.E., and V.S. contributed to the critical revision of the manuscript. A.M. contributed to the interpretation of the data and critical revision of the manuscript and provided funding for the study. T.A.L. was responsible for planning the study, funding, statistical analyses, and the interpretation of the data and also contributed to the critical revision of the manuscript. All the authors read and approved the final version of the manuscript. Disclosure Summary: The authors have nothing to disclose. References 1. Hossein-nezhad A , Holick MF . Vitamin D for health: a global perspective . Mayo Clin Proc . 2013 ; 88 ( 7 ): 720 – 755 . 2. Brenner DR , Arora P , Garcia-Bailo B , Wolever TM , Morrison H , El-Sohemy A , Karmali M , Badawi A . Plasma vitamin D levels and risk of metabolic syndrome in Canadians . Clin Invest Med . 2011 ; 34 ( 6 ): E377 . 3. Wang L , Song Y , Manson JE , Pilz S , März W , Michaëlsson K , Lundqvist A , Jassal SK , Barrett-Connor E , Zhang C , Eaton CB , May HT , Anderson JL , Sesso HD . Circulating 25-hydroxy-vitamin D and risk of cardiovascular disease: a meta-analysis of prospective studies . Circ Cardiovasc Qual Outcomes . 2012 ; 5 ( 6 ): 819 – 829 . 4. Dolinsky DH , Armstrong S , Mangarelli C , Kemper AR . The association between vitamin D and cardiometabolic risk factors in children: a systematic review . Clin Pediatr (Phila) . 2013 ; 52 ( 3 ): 210 – 223 . 5. Williams DM , Fraser A , Sayers A , Fraser WD , Hyppönen E , Smith GD , Sattar N , Lawlor DA . Associations of childhood 25-hydroxyvitamin D2 and D3 and cardiovascular risk factors in adolescence: prospective findings from the Avon Longitudinal Study of Parents and Children . Eur J Prev Cardiol . 2014 ; 21 ( 3 ): 281 – 290 . 6. Williams DM , Fraser A , Sayers A , Fraser WD , Hingorani A , Deanfield J , Davey Smith G , Sattar N , Lawlor DA . Associations of 25-hydroxyvitamin D2 and D3 with cardiovascular risk factors in childhood: cross-sectional findings from the Avon Longitudinal Study of Parents and Children . J Clin Endocrinol Metab . 2012 ; 97 ( 5 ): 1563 – 1571 . 7. Kwon JH , Lee SE , Lee HA , Kim YJ , Lee HY , Gwak HS , Park EA , Cho SJ , Oh SY , Ha EH , Park H , Kim HS . Relationship of serum 25-Hydroxyvitamin D (25[OH]D) levels and components of metabolic syndrome in prepubertal children . Nutrition . 2015 ; 31 ( 11-12 ): 1324 – 1327 . 8. Petersen RA , Dalskov S-M , Sørensen LB , Hjorth MF , Andersen R , Tetens I , Krarup H , Ritz C , Astrup A , Michaelsen KF , Mølgaard C , Damsgaard CT . Vitamin D status is associated with cardiometabolic markers in 8-11-year-old children, independently of body fat and physical activity . Br J Nutr . 2015 ; 114 ( 10 ): 1647 – 1655 . 9. Berenson GS , Srinivasan SR , Bao W , Newman WP III , Tracy RE , Wattigney WA . Association between multiple cardiovascular risk factors and atherosclerosis in children and young adults. The Bogalusa Heart Study . N Engl J Med . 1998 ; 338 ( 23 ): 1650 – 1656 . 10. Jorde R , Grimnes G . Vitamin D and metabolic health with special reference to the effect of vitamin D on serum lipids . Prog Lipid Res . 2011 ; 50 ( 4 ): 303 – 312 . 11. Birken CS , Lebovic G , Anderson LN , McCrindle BW , Mamdani M , Kandasamy S , Khovratovich M , Parkin PC , Maguire JL . TARGet Kids! collaboration. Association between Vitamin D and circulating lipids in early childhood . PLoS One . 2015 ; 10 ( 7 ): e0131938 . 12. Ganji V , Zhang X , Shaikh N , Tangpricha V . Serum 25-hydroxyvitamin D concentrations are associated with prevalence of metabolic syndrome and various cardiometabolic risk factors in US children and adolescents based on assay-adjusted serum 25-hydroxyvitamin D data from NHANES 2001-2006 . Am J Clin Nutr . 2011 ; 94 ( 1 ): 225 – 233 . 13. Delvin EE , Lambert M , Levy E , O’Loughlin J , Mark S , Gray-Donald K , Paradis G , Vitamin D . Vitamin D status is modestly associated with glycemia and indicators of lipid metabolism in French-Canadian children and adolescents . J Nutr . 2010 ; 140 ( 5 ): 987 – 991 . 14. Ahn J , Yu K , Stolzenberg-Solomon R , Simon KC , McCullough ML , Gallicchio L , Jacobs EJ , Ascherio A , Helzlsouer K , Jacobs KB , Li Q , Weinstein SJ , Purdue M , Virtamo J , Horst R , Wheeler W , Chanock S , Hunter DJ , Hayes RB , Kraft P , Albanes D . Genome-wide association study of circulating vitamin D levels . Hum Mol Genet . 2010 ; 19 ( 13 ): 2739 – 2745 . 15. Wang TJ , Zhang F , Richards JB , Kestenbaum B , van Meurs JB , Berry D , Kiel DP , Streeten EA , Ohlsson C , Koller DL , Peltonen L , Cooper JD , O’Reilly PF , Houston DK , Glazer NL , Vandenput L , Peacock M , Shi J , Rivadeneira F , McCarthy MI , Anneli P , de Boer IH , Mangino M , Kato B , Smyth DJ , Booth SL , Jacques PF , Burke GL , Goodarzi M , Cheung C-L , Wolf M , Rice K , Goltzman D , Hidiroglou N , Ladouceur M , Wareham NJ , Hocking LJ , Hart D , Arden NK , Cooper C , Malik S , Fraser WD , Hartikainen A-L , Zhai G , Macdonald HM , Forouhi NG , Loos RJF , Reid DM , Hakim A , Dennison E , Liu Y , Power C , Stevens HE , Jaana L , Vasan RS , Soranzo N , Bojunga J , Psaty BM , Lorentzon M , Foroud T , Harris TB , Hofman A , Jansson J-O , Cauley JA , Uitterlinden AG , Gibson Q , Järvelin M-R , Karasik D , Siscovick DS , Econs MJ , Kritchevsky SB , Florez JC , Todd JA , Dupuis J , Hyppönen E , Spector TD . Common genetic determinants of vitamin D insufficiency: a genome-wide association study . Lancet . 2010 ; 376 ( 9736 ): 180 – 188 . 16. Foucan L , Vélayoudom-Céphise F-L , Larifla L , Armand C , Deloumeaux J , Fagour C , Plumasseau J , Portlis M-L , Liu L , Bonnet F , Ducros J . Polymorphisms in GC and NADSYN1 genes are associated with vitamin D status and metabolic profile in non-diabetic adults . BMC Endocr Disord . 2013 ; 13 ( 1 ): 36 . 17. Vimaleswaran KS , Power C , Hyppönen E . Interaction between vitamin D receptor gene polymorphisms and 25-hydroxyvitamin D concentrations on metabolic and cardiovascular disease outcomes . Diabetes Metab . 2014 ; 40 ( 5 ): 386 – 389 . 18. Jorde R , Grimnes G . Exploring the association between serum 25-hydroxyvitamin D and serum lipids-more than confounding ? Eur J Clin Nutr . 2018 ; 72 ( 4 ): 526 – 533 . 19. Chow ECY , Magomedova L , Quach HP , Patel R , Durk MR , Fan J , Maeng HJ , Irondi K , Anakk S , Moore DD , Cummins CL , Pang KS . Vitamin D receptor activation down-regulates the small heterodimer partner and increases CYP7A1 to lower cholesterol . Gastroenterology . 2014 ; 146 ( 4 ): 1048 – 1059 . 20. Nicklas TA , von Duvillard SP , Berenson GS . Tracking of serum lipids and lipoproteins from childhood to dyslipidemia in adults: the Bogalusa Heart Study . Int J Sports Med . 2002 ; 23 ( S1 , Suppl 1 ): S39 – S43 . 21. Soininen S , Eloranta A-M , Lindi V , Lakka TA . Response: food fortification as a means to increase vitamin D intake . Br J Nutr . 2016 ; 116 ( 6 ): 1134 – 1135 . 22. National Center for Biotechnology Information. Database of single nucleotide polymorphisms (dbSNP). Available at: www.ncbi.nlm.nih.gov/SNP/. Accessed 6 April 2018. 23. Johnson AD , Handsaker RE , Pulit SL , Nizzari MM , O’Donnell CJ , de Bakker PIW . SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap . Bioinformatics . 2008 ; 24 ( 24 ): 2938 – 2939 . 24. Saari A , Sankilampi U , Hannila M-L , Kiviniemi V , Kesseli K , Dunkel L . New Finnish growth references for children and adolescents aged 0 to 20 years: length/height-for-age, weight-for-length/height, and body mass index-for-age . Ann Med . 2011 ; 43 ( 3 ): 235 – 248 . 25. Soininen S , Eloranta A-M , Lindi V , Venäläinen T , Zaproudina N , Mahonen A , Lakka TA . Determinants of serum 25-hydroxyvitamin D concentration in Finnish children: the Physical Activity and Nutrition in Children (PANIC) study . Br J Nutr . 2016 ; 115 ( 6 ): 1080 – 1091 . 26. Haapala EA , Poikkeus A-M , Tompuri T , Kukkonen-Harjula K , Leppänen PHT , Lindi V , Lakka TA . Associations of motor and cardiovascular performance with academic skills in children . Med Sci Sports Exerc . 2014 ; 46 ( 5 ): 1016 – 1024 . 27. Lamb MM , Ogden CL , Carroll MD , Lacher DA , Flegal KM . Association of body fat percentage with lipid concentrations in children and adolescents: United States, 1999-2004 . Am J Clin Nutr . 2011 ; 94 ( 3 ): 877 – 883 . 28. Sun C , Pezic A , Tikellis G , Ponsonby A-L , Wake M , Carlin JB , Cleland V , Dwyer T . Effects of school-based interventions for direct delivery of physical activity on fitness and cardiometabolic markers in children and adolescents: a systematic review of randomized controlled trials . Obes Rev . 2013 ; 14 ( 10 ): 818 – 838 . 29. Te Morenga L , Montez JM . Health effects of saturated and trans-fatty acid intake in children and adolescents: systematic review and meta-analysis . PLoS One . 2017 ; 12 ( 11 ): e0186672 . 30. Hauner H , Bechthold A , Boeing H , Brönstrup A , Buyken A , Leschik-Bonnet E , Linseisen J , Schulze M , Strohm D , Wolfram G ; German Nutrition Society . Evidence-based guideline of the German Nutrition Society: carbohydrate intake and prevention of nutrition-related diseases . Ann Nutr Metab . 2012 ; 60 ( s1 , Suppl 1 ): 1 – 58 . 31. Väistö J , Eloranta A-M , Viitasalo A , Tompuri T , Lintu N , Karjalainen P , Lampinen E-K , Ågren J , Laaksonen DE , Lakka H-M , Lindi V , Lakka TA . Physical activity and sedentary behaviour in relation to cardiometabolic risk in children: cross-sectional findings from the Physical Activity and Nutrition in Children (PANIC) Study . Int J Behav Nutr Phys Act . 2014 ; 11 ( 1 ): 55 . 32. Howe LD , Lawlor DA , Propper C . Trajectories of socioeconomic inequalities in health, behaviours and academic achievement across childhood and adolescence . J Epidemiol Community Health . 2013 ; 67 ( 4 ): 358 – 364 . 33. Eloranta AM , Lindi V , Schwab U , Kiiskinen S , Venäläinen T , Lakka HM , Laaksonen DE , Lakka TA . Dietary factors associated with metabolic risk score in Finnish children aged 6-8 years: the PANIC study . Eur J Nutr . 2014 ; 53 ( 6 ): 1431 – 1439 . 34. IOM (Institute of Medicine) . Dietary Reference Intakes for Calcium and Vitamin D . Washington, D.C. : National Academies Press ; 2011 . 35. Holick MF , Binkley NC , Bischoff-Ferrari HA , Gordon CM , Hanley DA , Heaney RP , Murad MH , Weaver CM ; Endocrine Society . Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline . J Clin Endocrinol Metab . 2011 ; 96 ( 7 ): 1911 – 1930 . 36. Sacheck J , Goodman E , Chui K , Chomitz V , Must A , Economos C . Vitamin D deficiency, adiposity, and cardiometabolic risk in urban schoolchildren . J Pediatr . 2011 ; 159 ( 6 ): 945 – 950 . 37. Arnberg K , Østergård M , Madsen AL , Krarup H , Michaelsen KF , Mølgaard C . Associations between vitamin D status in infants and blood lipids, body mass index and waist circumference . Acta Paediatr . 2011 ; 100 ( 9 ): 1244 – 1248 . 38. Ashraf AP , Alvarez JA , Gower BA , Saenz KH , McCormick KL . Associations of serum 25-hydroxyvitamin D and components of the metabolic syndrome in obese adolescent females . Obesity (Silver Spring) . 2011 ; 19 ( 11 ): 2214 – 2221 . 39. Kelishadi R , Farajzadegan Z , Bahreynian M . Association between vitamin D status and lipid profile in children and adolescents: a systematic review and meta-analysis . Int J Food Sci Nutr . 2014 ; 65 ( 4 ): 404 – 410 . 40. Pacifico L , Anania C , Osborn JF , Ferraro F , Bonci E , Olivero E , Chiesa C . Low 25(OH)D3 levels are associated with total adiposity, metabolic syndrome, and hypertension in Caucasian children and adolescents . Eur J Endocrinol . 2011 ; 165 ( 4 ): 603 – 611 . 41. Vimaleswaran KS , Cavadino A , Hyppönen E . APOA5 genotype influences the association between 25-hydroxyvitamin D and high density lipoprotein cholesterol . Atherosclerosis . 2013 ; 228 ( 1 ): 188 – 192 . 42. Nebert DW , Wikvall K , Miller WL . Human cytochromes P450 in health and disease . Philos Trans R Soc Lond B Biol Sci . 2013 ; 368 ( 1612 ): 20120431 . Copyright © 2018 Endocrine Society

Gaskins, Audrey J;Sundaram, Rajeshwari;Louis, Germaine M Buck;Chavarro, Jorge E

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00385pmid: 29800287

Abstract Context Marine long-chain omega-3 fatty acids have been positively related to markers of fecundity in both men and women. However, seafood, their primary food source, can also be a source of toxicants, which could counteract the reproductive benefits. Objective To examine the relationship of male and female seafood intake with time to pregnancy (TTP). Design Our prospective cohort study included 501 couples planning pregnancy, who participated in the Longitudinal Investigation of Fertility and the Environment study (2005 to 2009) and were followed up for ≤1 year or until pregnancy was detected. Seafood intake was collected daily during follow-up in journals. Setting Couples residing in Michigan and Texas were recruited using population-based sampling frameworks. Main Outcome Measures The primary outcome was the TTP, determined using an in-home pregnancy test. A secondary outcome was sexual intercourse frequency (SIF) as recorded in the daily journals. Results Couples with male and female partners who consumed eight or more seafood servings per cycle had 47% (95% CI, 7% to 103%) and 60% (95% CI, 15% to 122%) greater fecundity (shorter TTP) than couples with male and female partners who consumed one or fewer seafood servings per cycle. Couples with both partners consuming eight or more seafood servings per cycle had 61% (95% CI, 17% to 122%) greater fecundity than couples consuming less. Male and female partners with the highest seafood intake (eight or more servings per cycle) also had 22% greater SIF. Conclusions Greater male and female seafood intake was associated with a higher SIF and fecundity among a large prospective cohort of couples attempting pregnancy. Infertility, the failure to achieve pregnancy after 12 months of unprotected sexual intercourse, affects 15% to 25% of couples (1, 2). Although infertility treatments exist, their costs (3), limited geographic accessibility (4), and modest success (5) justify identifying modifiable factors that increase a couple’s chance of conceiving without medical assistance. Seafood is a recommended component of many healthy eating patterns (6, 7). In the context of fertility, however, seafood has largely been studied as a potential harm, representing a primary source of exposure to reproductive toxicants such as organochlorines, dioxins, and mercury (8–11). In contrast, some studies have found reproductive benefits with higher marine long-chain omega-3 fatty acid intake, such as increased progesterone levels, a shorter time to pregnancy (TTP), and better semen quality (12–14). For the average US adult, the current recommendation is to eat at least two seafood servings per week (6); however, in January 2017, the US Food and Drug Administration and Environmental Protection Agency recommended that women who are pregnant or might become pregnant should eat no more than three servings per week (15). This guideline was established to limit fetal methyl-mercury exposure, which has been linked to adverse neurocognitive consequences. However, to the best of our knowledge, these guidelines did not consider the potential reproductive benefits of seafood intake. To address this gap, we used data from a prospective cohort of couples attempting to become pregnant, with information on daily seafood intake and sexual intercourse collected in journals, to investigate whether male and female seafood intake was associated with the TTP and whether this association could be due to differences in sexual activity. Materials and Methods The Longitudinal Investigation of Fertility and the Environment (LIFE) study is a prospective cohort of 501 couples attempting to conceive in two geographic areas (Texas and Michigan) from 2005 to 2009. Using population-based sampling frameworks, households were contacted to identify eligible couples in a committed relationship. Female partners were required to be aged 18 to 44 years, to have menstrual cycles of 21 to 42 days, and to have had no hormonal birth control injections during the previous year. Male partners were required to be aged ≥18 years. Both partners were required to have the ability to communicate in English or Spanish and to have undergone no sterilization procedures or have physician-diagnosed infertility. The couples were also excluded if they had not been using contraception for >2 months. A complete description of the study’s methods has been previously reported (16). In brief, of the 1188 eligible couples, 501 (42%) were enrolled in the present study and followed up for ≤12 months, with monthly pregnancy tests. The institutional review boards at each institution approved the protocol. All participants provided written informed consent. Research assistants traveled to the couples’ homes and completed baseline in-person interviews separately with each partner. Both partners were asked how often during the previous 12 months they had eaten canned tuna fish; fish caught in unknown locations; crab, shrimp, or other shellfish caught in an unknown location; fish caught in local waters; and crab, shrimp, or shellfish caught in local waters. The five response options ranged from “never or almost never” to “two or more times per week.” The selected frequency category for each seafood item was then converted to a monthly intake, and all items were summed to find the total baseline seafood intake. In the daily journals, the male and female participants were asked to report the number of 4-oz servings of fish or shellfish consumed. These daily responses were then summed across the cycle to determine their cycle-specific seafood intake. For the analysis, the baseline and daily male and female seafood intake were classified into categories that approximated quartiles. During the enrollment interview, each partner reported their age, level of education, ethnicity, race, household income, and use of cigarettes. Participants were asked whether they had followed a regular vigorous exercise program in the previous 12 months and, if so, how many days per week. The four-item Cohen perceived stress scale was also administered (17). The men and women reported whether they had consumed ≥12 alcoholic drinks in the previous year, and, if so, how often they had consumed alcoholic beverages. All participants had their weight and height measured using standardized procedures, and the body mass index was calculated as the weight in kilograms divided by the height in square meters. The primary outcome was fecundity, as measured by TTP. We used daily journal information supplemented with fertility monitors to define the menstrual cycles, defined as the interval (in days) from the onset of bleeding that increased in intensity and lasted ≥2 days to the onset of the next similar bleeding episode. Because couples were allowed to enroll in the LIFE study midcycle, we defined this as cycle 0 to differentiate it from cycle 1, which denoted the first fully observed menstrual cycle. Pregnancy was defined as a positive study-provided home pregnancy test, which was sensitive for 25 mIU/mL human chorionic gonadotropin. A secondary outcome was the frequency of vaginal–penial intercourse, as recorded by the men and women in their daily journals. For each cycle of follow-up, the sexual intercourse frequency (SIF) reports were summed across all days to find the total SIF per cycle. The correlation between the SIF per cycle as reported by the male and female partners was 0.98, and the average difference between the two reports was −0.02 times per month. Because of the slightly lower amount of missing data in the female diaries, the female report of SIF was used as the main outcome variable. We classified each partner as having high (nine times or more per month; 75th percentile) or low-to-average (less than nine times per month) seafood intake. The male and female demographic data and lifestyle characteristics were then compared using ANOVA for continuous variables or χ2 tests for categorical variables. The correlation within and between male and female seafood intake at baseline and during follow-up was calculated using Spearman correlation coefficients. Cox proportional odds models for discrete survival data accounting for left truncation (to account for the time without contraception before enrollment) and right censoring (to account for the loss to follow-up or the end of the study) were used to estimate the fecundability ORs (FORs), and their 95% CIs, as a measure of fecundity. FORs represent the relative odds of achieving pregnancy conditional on not becoming pregnant in the previous cycle, such that an FOR <1 indicates diminished fecundity as measured by a longer TTP. Seafood intake was initially considered as quartiles of intake, and in a supplemental analysis, it was modeled continuously using linear and quadratic terms. To analyze the association between seafood intake and SIF per cycle during the follow-up period, we used generalized linear mixed models with the Poisson distribution. Effect estimates and 95% CIs are presented as the percentage of difference in SIF for a particular group compared with the reference group. We also explored the association between day-level seafood intake and SIF using a generalized linear mixed model with logit link. The results are presented as the ORs and 95% CIs of sexual intercourse in a given day. We imputed the SIF values for cycles with >50% of the days missing information on SIF and any cycle with <14 days of follow-up (n = 159 cycles) using Markov chain Monte Carlo methods (PROC MI in SAS; SAS Institute, Cary, NC) with five multiple imputations based on menstrual cycle length, cycle number of follow-up, female age, the difference between the couple’s ages, female race and education level, and male exercise. Effect estimates from models using multiply imputed values for SIF were estimated using Rubin’s formula for combining estimates across imputations (PROC MIANALYZE in SAS). Confounding was evaluated using previous knowledge and descriptive statistics from our cohort through the use of directed acyclic graphs. Variables retained in the final multivariable models were female age (in years), the difference between couple’s ages (in years), female race (non-Hispanic white vs other), male exercise (yes vs no), and male and female alcohol intake (one or more time per week vs less than one time per week). Additional models were run further, adjusting for male and female partner seafood intake owing to the high amount of concordance within a couple. The fecundity models were also further adjusted for SIF to evaluate the extent to which this variable explained any observed associations. A P value for trend was calculated across the categories of seafood intake using the median intake level in each category as a continuous variable. In the main analysis, missing data on seafood intake in the daily journals were considered as no intake, which is common for dietary analyses. Sensitivity analyses were performed in which missing seafood intake was imputed for cycles that were missing 100% and >50% of days of seafood intake data using Markov chain Monte Carlo methods with five multiple imputations and Rubin formula to combine estimates across imputations. To address concerns of residual confounding, we also calculated propensity scores and ran the final model adjusting for this variable and stratified by quintiles of this variable. To quantify the effect of unmeasured confounding, we calculated the e-value, which estimates the minimum strength of an association that an unmeasured confounder would need to have with both the exposure and outcome to fully explain a specific exposure–outcome association (18). SAS, version 9.4 (SAS Institute) was used for all statistical analyses. Results Male partners who reported the highest usual seafood intake were less likely to have a non-Hispanic white partner and more likely to exercise regularly and consume alcohol one or more time per week compared with men with lower intake (Table 1). Female partners with the greatest usual seafood intake were, on average, older, had older partners, were less likely to be non-Hispanic white, and were more likely to consume alcohol one or more time per week compared with females with lower intake. Seafood intake was not associated with body mass index, education level of either partner, or household income. Male and female seafood intake within a couple correlated moderately at baseline (r = 0.46) and correlated highly during the follow-up period (r = 0.70; Supplemental Table 1). Within men and women, the baseline seafood intake correlated moderately with the intake during follow-up (r = 0.47 and r = 0.53, respectively). Table 1. Demographic and Lifestyle Characteristics Stratified by Seafood Intake at Baseline in the LIFE Study (n = 501 Couples) Variable Male Baseline Seafood Intake Female Baseline Seafood Intake Less Than Nine Times per Month (n = 406) Nine Times or More per Month (n = 95) P Valuea Less Than Nine Times per Month (n = 419) Nine Times or More per Month (n = 82) P Valuea Female demographic data  Age, y 30.1 ± 4.1 29.7 ± 4.3 0.41 29.7 ± 4.0 31.2 ± 4.4 0.003  Non-Hispanic white 339 (83.5) 68 (71.6) 0.007 350 (83.5) 57 (69.5) 0.003  College education 309 (76.1) 71 (74.7) 0.78 319 (76.1) 61 (74.4) 0.74 Male demographic data  Age, y 31.8 ± 4.8 31.7 ± 5.2 0.87 31.4 ± 4.8 33.5 ± 5.2 0.004  Non-Hispanic white 340 (83.7) 72 (75.8) 0.07 347 (82.8) 65 (79.3) 0.44  College education 256 (63.1) 55 (57.9) 0.35 264 (63.0) 47 (57.3) 0.33  Couple income 0.14 0.29   <$29,999 15 (3.8) 6 (6.3) 19 (4.6) 2 (2.4)   $30,000–$49,999 51 (12.8) 5 (5.3) 51 (12.4) 5 (6.1)   $50,000–$69,999 67 (16.8) 19 (20.0) 71 (17.3) 15 (18.3)   ≥$70,000 265 (66.6) 65 (68.4) 270 (65.7) 60 (73.2) Female lifestyle factors  BMI, kg/m2 27.6 ± 7.2 27.0 ± 6.4 0.46 27.3 ± 7.1 28.5 ± 6.9 0.15  Current smoker 45 (11.1) 11 (11.6) 0.89 46 (11.0) 10 (12.2) 0.75  Exercises regularly 164 (40.4) 36 (37.9) 0.65 168 (40.1) 32 (39.0) 0.86  Seafood intake, times per month 4.7 ± 4.4 7.7 ± 4.4 <0.001 3.7 ± 2.7 13.1 ± 4.0 <0.001  Alcohol intake one or more times per week 116 (28.6) 38 (40.4) 0.13 114 (27.3) 40 (48.8) <0.001  Stress in previous month 3.6 ± 2.6 3.6 ± 2.4 0.90 3.5 ± 2.5 3.9 ± 2.6 0.28 Male lifestyle factors  BMI, kg/m2 29.4 ± 4.9 29.8 ± 5.2 0.52 29.5 ± 5.1 29.3 ± 4.2 0.71  Current smoker 56 (13.8) 18 (19.0) 0.20 60 (14.3) 14 (17.1) 0.52  Exercises regularly 162 (39.9) 49 (51.6) 0.04 170 (40.6) 41 (50.0) 0.11  Seafood intake, times per month 3.9 ± 2.7 13.1 ± 4.4 <0.001 5.1 ± 4.3 8.7 ± 5.6 <0.001  Alcohol intake one or more times per week 212 (52.2) 65 (68.4) 0.03 221 (52.7) 56 (68.3) 0.07  Stress in previous month 3.0 ± 2.4 3.3 ± 2.4 0.32 3.1 ± 2.3 2.8 ± 2.5 0.39 Variable Male Baseline Seafood Intake Female Baseline Seafood Intake Less Than Nine Times per Month (n = 406) Nine Times or More per Month (n = 95) P Valuea Less Than Nine Times per Month (n = 419) Nine Times or More per Month (n = 82) P Valuea Female demographic data  Age, y 30.1 ± 4.1 29.7 ± 4.3 0.41 29.7 ± 4.0 31.2 ± 4.4 0.003  Non-Hispanic white 339 (83.5) 68 (71.6) 0.007 350 (83.5) 57 (69.5) 0.003  College education 309 (76.1) 71 (74.7) 0.78 319 (76.1) 61 (74.4) 0.74 Male demographic data  Age, y 31.8 ± 4.8 31.7 ± 5.2 0.87 31.4 ± 4.8 33.5 ± 5.2 0.004  Non-Hispanic white 340 (83.7) 72 (75.8) 0.07 347 (82.8) 65 (79.3) 0.44  College education 256 (63.1) 55 (57.9) 0.35 264 (63.0) 47 (57.3) 0.33  Couple income 0.14 0.29   <$29,999 15 (3.8) 6 (6.3) 19 (4.6) 2 (2.4)   $30,000–$49,999 51 (12.8) 5 (5.3) 51 (12.4) 5 (6.1)   $50,000–$69,999 67 (16.8) 19 (20.0) 71 (17.3) 15 (18.3)   ≥$70,000 265 (66.6) 65 (68.4) 270 (65.7) 60 (73.2) Female lifestyle factors  BMI, kg/m2 27.6 ± 7.2 27.0 ± 6.4 0.46 27.3 ± 7.1 28.5 ± 6.9 0.15  Current smoker 45 (11.1) 11 (11.6) 0.89 46 (11.0) 10 (12.2) 0.75  Exercises regularly 164 (40.4) 36 (37.9) 0.65 168 (40.1) 32 (39.0) 0.86  Seafood intake, times per month 4.7 ± 4.4 7.7 ± 4.4 <0.001 3.7 ± 2.7 13.1 ± 4.0 <0.001  Alcohol intake one or more times per week 116 (28.6) 38 (40.4) 0.13 114 (27.3) 40 (48.8) <0.001  Stress in previous month 3.6 ± 2.6 3.6 ± 2.4 0.90 3.5 ± 2.5 3.9 ± 2.6 0.28 Male lifestyle factors  BMI, kg/m2 29.4 ± 4.9 29.8 ± 5.2 0.52 29.5 ± 5.1 29.3 ± 4.2 0.71  Current smoker 56 (13.8) 18 (19.0) 0.20 60 (14.3) 14 (17.1) 0.52  Exercises regularly 162 (39.9) 49 (51.6) 0.04 170 (40.6) 41 (50.0) 0.11  Seafood intake, times per month 3.9 ± 2.7 13.1 ± 4.4 <0.001 5.1 ± 4.3 8.7 ± 5.6 <0.001  Alcohol intake one or more times per week 212 (52.2) 65 (68.4) 0.03 221 (52.7) 56 (68.3) 0.07  Stress in previous month 3.0 ± 2.4 3.3 ± 2.4 0.32 3.1 ± 2.3 2.8 ± 2.5 0.39 Data presented as mean ± SD or n (%). a P values presented from χ2 tests for categorical variables and Kruskal-Wallis nonparametric tests for continuous variables. View Large Table 1. Demographic and Lifestyle Characteristics Stratified by Seafood Intake at Baseline in the LIFE Study (n = 501 Couples) Variable Male Baseline Seafood Intake Female Baseline Seafood Intake Less Than Nine Times per Month (n = 406) Nine Times or More per Month (n = 95) P Valuea Less Than Nine Times per Month (n = 419) Nine Times or More per Month (n = 82) P Valuea Female demographic data  Age, y 30.1 ± 4.1 29.7 ± 4.3 0.41 29.7 ± 4.0 31.2 ± 4.4 0.003  Non-Hispanic white 339 (83.5) 68 (71.6) 0.007 350 (83.5) 57 (69.5) 0.003  College education 309 (76.1) 71 (74.7) 0.78 319 (76.1) 61 (74.4) 0.74 Male demographic data  Age, y 31.8 ± 4.8 31.7 ± 5.2 0.87 31.4 ± 4.8 33.5 ± 5.2 0.004  Non-Hispanic white 340 (83.7) 72 (75.8) 0.07 347 (82.8) 65 (79.3) 0.44  College education 256 (63.1) 55 (57.9) 0.35 264 (63.0) 47 (57.3) 0.33  Couple income 0.14 0.29   <$29,999 15 (3.8) 6 (6.3) 19 (4.6) 2 (2.4)   $30,000–$49,999 51 (12.8) 5 (5.3) 51 (12.4) 5 (6.1)   $50,000–$69,999 67 (16.8) 19 (20.0) 71 (17.3) 15 (18.3)   ≥$70,000 265 (66.6) 65 (68.4) 270 (65.7) 60 (73.2) Female lifestyle factors  BMI, kg/m2 27.6 ± 7.2 27.0 ± 6.4 0.46 27.3 ± 7.1 28.5 ± 6.9 0.15  Current smoker 45 (11.1) 11 (11.6) 0.89 46 (11.0) 10 (12.2) 0.75  Exercises regularly 164 (40.4) 36 (37.9) 0.65 168 (40.1) 32 (39.0) 0.86  Seafood intake, times per month 4.7 ± 4.4 7.7 ± 4.4 <0.001 3.7 ± 2.7 13.1 ± 4.0 <0.001  Alcohol intake one or more times per week 116 (28.6) 38 (40.4) 0.13 114 (27.3) 40 (48.8) <0.001  Stress in previous month 3.6 ± 2.6 3.6 ± 2.4 0.90 3.5 ± 2.5 3.9 ± 2.6 0.28 Male lifestyle factors  BMI, kg/m2 29.4 ± 4.9 29.8 ± 5.2 0.52 29.5 ± 5.1 29.3 ± 4.2 0.71  Current smoker 56 (13.8) 18 (19.0) 0.20 60 (14.3) 14 (17.1) 0.52  Exercises regularly 162 (39.9) 49 (51.6) 0.04 170 (40.6) 41 (50.0) 0.11  Seafood intake, times per month 3.9 ± 2.7 13.1 ± 4.4 <0.001 5.1 ± 4.3 8.7 ± 5.6 <0.001  Alcohol intake one or more times per week 212 (52.2) 65 (68.4) 0.03 221 (52.7) 56 (68.3) 0.07  Stress in previous month 3.0 ± 2.4 3.3 ± 2.4 0.32 3.1 ± 2.3 2.8 ± 2.5 0.39 Variable Male Baseline Seafood Intake Female Baseline Seafood Intake Less Than Nine Times per Month (n = 406) Nine Times or More per Month (n = 95) P Valuea Less Than Nine Times per Month (n = 419) Nine Times or More per Month (n = 82) P Valuea Female demographic data  Age, y 30.1 ± 4.1 29.7 ± 4.3 0.41 29.7 ± 4.0 31.2 ± 4.4 0.003  Non-Hispanic white 339 (83.5) 68 (71.6) 0.007 350 (83.5) 57 (69.5) 0.003  College education 309 (76.1) 71 (74.7) 0.78 319 (76.1) 61 (74.4) 0.74 Male demographic data  Age, y 31.8 ± 4.8 31.7 ± 5.2 0.87 31.4 ± 4.8 33.5 ± 5.2 0.004  Non-Hispanic white 340 (83.7) 72 (75.8) 0.07 347 (82.8) 65 (79.3) 0.44  College education 256 (63.1) 55 (57.9) 0.35 264 (63.0) 47 (57.3) 0.33  Couple income 0.14 0.29   <$29,999 15 (3.8) 6 (6.3) 19 (4.6) 2 (2.4)   $30,000–$49,999 51 (12.8) 5 (5.3) 51 (12.4) 5 (6.1)   $50,000–$69,999 67 (16.8) 19 (20.0) 71 (17.3) 15 (18.3)   ≥$70,000 265 (66.6) 65 (68.4) 270 (65.7) 60 (73.2) Female lifestyle factors  BMI, kg/m2 27.6 ± 7.2 27.0 ± 6.4 0.46 27.3 ± 7.1 28.5 ± 6.9 0.15  Current smoker 45 (11.1) 11 (11.6) 0.89 46 (11.0) 10 (12.2) 0.75  Exercises regularly 164 (40.4) 36 (37.9) 0.65 168 (40.1) 32 (39.0) 0.86  Seafood intake, times per month 4.7 ± 4.4 7.7 ± 4.4 <0.001 3.7 ± 2.7 13.1 ± 4.0 <0.001  Alcohol intake one or more times per week 116 (28.6) 38 (40.4) 0.13 114 (27.3) 40 (48.8) <0.001  Stress in previous month 3.6 ± 2.6 3.6 ± 2.4 0.90 3.5 ± 2.5 3.9 ± 2.6 0.28 Male lifestyle factors  BMI, kg/m2 29.4 ± 4.9 29.8 ± 5.2 0.52 29.5 ± 5.1 29.3 ± 4.2 0.71  Current smoker 56 (13.8) 18 (19.0) 0.20 60 (14.3) 14 (17.1) 0.52  Exercises regularly 162 (39.9) 49 (51.6) 0.04 170 (40.6) 41 (50.0) 0.11  Seafood intake, times per month 3.9 ± 2.7 13.1 ± 4.4 <0.001 5.1 ± 4.3 8.7 ± 5.6 <0.001  Alcohol intake one or more times per week 212 (52.2) 65 (68.4) 0.03 221 (52.7) 56 (68.3) 0.07  Stress in previous month 3.0 ± 2.4 3.3 ± 2.4 0.32 3.1 ± 2.3 2.8 ± 2.5 0.39 Data presented as mean ± SD or n (%). a P values presented from χ2 tests for categorical variables and Kruskal-Wallis nonparametric tests for continuous variables. View Large Higher male (but not female) baseline seafood intake was associated with higher SIF during follow-up after multivariable adjustment (Table 2). Men who usually consumed seafood nine or more times per month had a 22.9% (95% CI, 6.8% to 41.5%) greater SIF compared with men who usually consumed seafood two times or less per month (P for trend = 0.007). A positive association was found between baseline female seafood intake and SIF that became attenuated after adjustment for male partner intake. During follow-up, both male and female seafood intake was independently associated with SIF, with slightly stronger associations observed for male intake. Furthermore, when both partners consumed eight or more servings per cycle, SIF was increased by 21.9% (95% CI, 15.2% to 29.0%) compared with couples consuming less. In the day-level analyses, the odds of sexual intercourse was 39% (95% CI, 29% to 50%) greater if both partners consumed seafood the same day, 3% (95% CI, −5% to 11%) greater if only the woman consumed seafood, and 2% (95% CI, −6% to 10%) greater if only the man consumed seafood compared with couples with neither partner consuming seafood. The associations were identical when the male report of SIF was used (instead of the female report). Table 2. Associations Between Male and Female Seafood Intake at Baseline and During Follow-Up and Frequency of Sexual Intercourse (n = 501 Couples; 2372 Follow-Up Cycles) Variable Subjects or Cycles, n (%) % Difference in SIF (95% CI) Model 1a Model 2b Male baseline seafood intake  Two times or less per month 153 (31) Reference Reference  Three to four times per month 118 (24) 3.0 (−9.9 to 17.7) 2.5 (−10.5 to 17.4)  Five to eight times per month 135 (27) 7.5 (−5.3 to 22.0) 7.1 (−6.4 to 22.4)  Nine times or more per month 95 (19) 22.9 (6.8 to 41.5) 21.7 (4.6 to 41.7)  P for trend 0.007 0.02 Female baseline seafood intake  Two times or less per month 177 (35) Reference Reference  Three to four times per month 95 (19) 1.1 (−11.9 to 16.0) −2.8 (−15.5 to 11.8)  Five to eight times per month 147 (29) 1.4 (−10.0 to 14.1) −3.8 (−15.1 to 9.1)  Nine times or more per month 82 (16) 17.3 (1.2 to 36.0) 9.7 (−6.3 to 28.4)  P for trend 0.12 0.60 Male daily journal seafood intake  One serving or less per cycle 814 (34) Reference Reference  One to three servings per cycle 422 (18) 4.8 (−0.8 to 10.9) 4.8 (−1.1 to 11.0)  Four to seven servings per cycle 575 (24) 16.7 (10.8 to 23.0) 15.2 (8.8 to 22.1)  Eight servings or more per cycle 561 (24) 32.6 (25.4 to 40.2) 26.9 (18.7 to 35.6)  P for trend <0.001 <0.001 Female daily journal seafood intake  One serving or less per cycle 835 (35) Reference Reference  One to three servings per cycle 446 (19) 2.5 (−3.2 to 8.5) −1.5 (−7.2 to 4.5)  Four to seven servings per cycle 605 (26) 9.4 (3.7 to 15.4) 0.5 (-5.2 to 6.6)  Eight servings or more per cycle 486 (20) 26.4 (19.2 to 34.0) 10.2 (2.9 to 18.2)  P for trend <0.001 0.009 Couple daily journal seafood intake  At least one partner consumed fewer than 8 servings per cycle 2057 (87) Reference  Both partners consumed eight servings or more per cycle 315 (13) 21.9 (15.2 to 29.0) Variable Subjects or Cycles, n (%) % Difference in SIF (95% CI) Model 1a Model 2b Male baseline seafood intake  Two times or less per month 153 (31) Reference Reference  Three to four times per month 118 (24) 3.0 (−9.9 to 17.7) 2.5 (−10.5 to 17.4)  Five to eight times per month 135 (27) 7.5 (−5.3 to 22.0) 7.1 (−6.4 to 22.4)  Nine times or more per month 95 (19) 22.9 (6.8 to 41.5) 21.7 (4.6 to 41.7)  P for trend 0.007 0.02 Female baseline seafood intake  Two times or less per month 177 (35) Reference Reference  Three to four times per month 95 (19) 1.1 (−11.9 to 16.0) −2.8 (−15.5 to 11.8)  Five to eight times per month 147 (29) 1.4 (−10.0 to 14.1) −3.8 (−15.1 to 9.1)  Nine times or more per month 82 (16) 17.3 (1.2 to 36.0) 9.7 (−6.3 to 28.4)  P for trend 0.12 0.60 Male daily journal seafood intake  One serving or less per cycle 814 (34) Reference Reference  One to three servings per cycle 422 (18) 4.8 (−0.8 to 10.9) 4.8 (−1.1 to 11.0)  Four to seven servings per cycle 575 (24) 16.7 (10.8 to 23.0) 15.2 (8.8 to 22.1)  Eight servings or more per cycle 561 (24) 32.6 (25.4 to 40.2) 26.9 (18.7 to 35.6)  P for trend <0.001 <0.001 Female daily journal seafood intake  One serving or less per cycle 835 (35) Reference Reference  One to three servings per cycle 446 (19) 2.5 (−3.2 to 8.5) −1.5 (−7.2 to 4.5)  Four to seven servings per cycle 605 (26) 9.4 (3.7 to 15.4) 0.5 (-5.2 to 6.6)  Eight servings or more per cycle 486 (20) 26.4 (19.2 to 34.0) 10.2 (2.9 to 18.2)  P for trend <0.001 0.009 Couple daily journal seafood intake  At least one partner consumed fewer than 8 servings per cycle 2057 (87) Reference  Both partners consumed eight servings or more per cycle 315 (13) 21.9 (15.2 to 29.0) Generalized linear mixed models with Poisson distribution and log link were used to estimate the percentage of difference (95% CIs); cycles with >50% of days with missing information on SIF and cycles with <14 days of follow-up (n = 159 cycles) had their SIF values imputed (using five multiple imputations). a Model 1 adjusted for cycle length, female age, difference in male and female age, female race (non-Hispanic white vs other), male exercise (yes vs no), and male and female alcohol intake (one or more times per week vs less than one time per week). b Model 2 adjusted for variables in model 1 plus male or female partner seafood intake (using the same assessment method). View Large Table 2. Associations Between Male and Female Seafood Intake at Baseline and During Follow-Up and Frequency of Sexual Intercourse (n = 501 Couples; 2372 Follow-Up Cycles) Variable Subjects or Cycles, n (%) % Difference in SIF (95% CI) Model 1a Model 2b Male baseline seafood intake  Two times or less per month 153 (31) Reference Reference  Three to four times per month 118 (24) 3.0 (−9.9 to 17.7) 2.5 (−10.5 to 17.4)  Five to eight times per month 135 (27) 7.5 (−5.3 to 22.0) 7.1 (−6.4 to 22.4)  Nine times or more per month 95 (19) 22.9 (6.8 to 41.5) 21.7 (4.6 to 41.7)  P for trend 0.007 0.02 Female baseline seafood intake  Two times or less per month 177 (35) Reference Reference  Three to four times per month 95 (19) 1.1 (−11.9 to 16.0) −2.8 (−15.5 to 11.8)  Five to eight times per month 147 (29) 1.4 (−10.0 to 14.1) −3.8 (−15.1 to 9.1)  Nine times or more per month 82 (16) 17.3 (1.2 to 36.0) 9.7 (−6.3 to 28.4)  P for trend 0.12 0.60 Male daily journal seafood intake  One serving or less per cycle 814 (34) Reference Reference  One to three servings per cycle 422 (18) 4.8 (−0.8 to 10.9) 4.8 (−1.1 to 11.0)  Four to seven servings per cycle 575 (24) 16.7 (10.8 to 23.0) 15.2 (8.8 to 22.1)  Eight servings or more per cycle 561 (24) 32.6 (25.4 to 40.2) 26.9 (18.7 to 35.6)  P for trend <0.001 <0.001 Female daily journal seafood intake  One serving or less per cycle 835 (35) Reference Reference  One to three servings per cycle 446 (19) 2.5 (−3.2 to 8.5) −1.5 (−7.2 to 4.5)  Four to seven servings per cycle 605 (26) 9.4 (3.7 to 15.4) 0.5 (-5.2 to 6.6)  Eight servings or more per cycle 486 (20) 26.4 (19.2 to 34.0) 10.2 (2.9 to 18.2)  P for trend <0.001 0.009 Couple daily journal seafood intake  At least one partner consumed fewer than 8 servings per cycle 2057 (87) Reference  Both partners consumed eight servings or more per cycle 315 (13) 21.9 (15.2 to 29.0) Variable Subjects or Cycles, n (%) % Difference in SIF (95% CI) Model 1a Model 2b Male baseline seafood intake  Two times or less per month 153 (31) Reference Reference  Three to four times per month 118 (24) 3.0 (−9.9 to 17.7) 2.5 (−10.5 to 17.4)  Five to eight times per month 135 (27) 7.5 (−5.3 to 22.0) 7.1 (−6.4 to 22.4)  Nine times or more per month 95 (19) 22.9 (6.8 to 41.5) 21.7 (4.6 to 41.7)  P for trend 0.007 0.02 Female baseline seafood intake  Two times or less per month 177 (35) Reference Reference  Three to four times per month 95 (19) 1.1 (−11.9 to 16.0) −2.8 (−15.5 to 11.8)  Five to eight times per month 147 (29) 1.4 (−10.0 to 14.1) −3.8 (−15.1 to 9.1)  Nine times or more per month 82 (16) 17.3 (1.2 to 36.0) 9.7 (−6.3 to 28.4)  P for trend 0.12 0.60 Male daily journal seafood intake  One serving or less per cycle 814 (34) Reference Reference  One to three servings per cycle 422 (18) 4.8 (−0.8 to 10.9) 4.8 (−1.1 to 11.0)  Four to seven servings per cycle 575 (24) 16.7 (10.8 to 23.0) 15.2 (8.8 to 22.1)  Eight servings or more per cycle 561 (24) 32.6 (25.4 to 40.2) 26.9 (18.7 to 35.6)  P for trend <0.001 <0.001 Female daily journal seafood intake  One serving or less per cycle 835 (35) Reference Reference  One to three servings per cycle 446 (19) 2.5 (−3.2 to 8.5) −1.5 (−7.2 to 4.5)  Four to seven servings per cycle 605 (26) 9.4 (3.7 to 15.4) 0.5 (-5.2 to 6.6)  Eight servings or more per cycle 486 (20) 26.4 (19.2 to 34.0) 10.2 (2.9 to 18.2)  P for trend <0.001 0.009 Couple daily journal seafood intake  At least one partner consumed fewer than 8 servings per cycle 2057 (87) Reference  Both partners consumed eight servings or more per cycle 315 (13) 21.9 (15.2 to 29.0) Generalized linear mixed models with Poisson distribution and log link were used to estimate the percentage of difference (95% CIs); cycles with >50% of days with missing information on SIF and cycles with <14 days of follow-up (n = 159 cycles) had their SIF values imputed (using five multiple imputations). a Model 1 adjusted for cycle length, female age, difference in male and female age, female race (non-Hispanic white vs other), male exercise (yes vs no), and male and female alcohol intake (one or more times per week vs less than one time per week). b Model 2 adjusted for variables in model 1 plus male or female partner seafood intake (using the same assessment method). View Large Baseline seafood intake was not associated with fecundity after multivariable adjustment (Supplemental Table 2). Also, no differences were found in the associations according to whether the seafood was caught in local vs unknown waters or whether the seafood was shellfish vs fish (data not shown). However, the prospectively collected male and female seafood intake from the daily journals was related to increased fecundity (shorter TTP; Table 3). Specifically, men and women who consumed eight or more seafood servings per cycle had 47% (95% CI, 7% to 103%) and 60% (95% CI, 15% to 122%) greater fecundity compared with the men and women who consumed one or fewer seafood servings per cycle after multivariable adjustment. These associations were attenuated with further adjustment for partner seafood intake (model 2), most likely owing to the high correlation between intake during the follow-up period. When modeling intake as a continuous variable, both male and female seafood intake was associated with greater fecundity, plateauing at ~14 to 16 seafood servings per cycle (>90th percentile of intake; Supplemental Fig. 1). Table 3. Associations Between Male and Female Seafood Intake During Follow-Up and TTP (n = 501 Couples) Variable Pregnancies/Cycles FOR (95% CI) Unadjusted Model 1a Model 2b Model 3c Male daily journal seafood intake  One serving or less per cycle 99/814 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 58/422 1.15 (0.81–1.64) 1.17 (0.82–1.68) 1.18 (0.81–1.71) 1.06 (0.74–1.52)  Four to seven servings per cycle 90/575 1.25 (0.91–1.71) 1.28 (0.93–1.76) 1.23 (0.87–1.74) 1.19 (0.86–1.85)  Eight servings or more per cycle 98/561 1.37 (1.01–1.87) 1.47 (1.07–2.03) 1.24 (0.84–1.83) 1.33 (0.96–1.85)  P for trend 0.04 0.02 0.28 0.07 Female daily journal seafood intake  One serving or less per cycle 108/835 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 61/446 0.99 (0.70–1.39) 1.04 (0.73–1.47) 0.98 (0.68–1.40) 1.00 (0.70–1.42)  Four to seven servings per cycle 83/605 1.00 (0.73–1.36) 1.07 (0.78–1.47) 0.97 (0.68–1.38) 1.03 (0.75–1.42)  Eight servings or more per cycle 93/486 1.38 (1.02–1.89) 1.60 (1.15–2.22) 1.42 (0.96–2.11) 1.44 (1.04–2.01)  P for trend 0.06 0.01 0.13 0.04 Variable Pregnancies/Cycles FOR (95% CI) Unadjusted Model 1a Model 2b Model 3c Male daily journal seafood intake  One serving or less per cycle 99/814 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 58/422 1.15 (0.81–1.64) 1.17 (0.82–1.68) 1.18 (0.81–1.71) 1.06 (0.74–1.52)  Four to seven servings per cycle 90/575 1.25 (0.91–1.71) 1.28 (0.93–1.76) 1.23 (0.87–1.74) 1.19 (0.86–1.85)  Eight servings or more per cycle 98/561 1.37 (1.01–1.87) 1.47 (1.07–2.03) 1.24 (0.84–1.83) 1.33 (0.96–1.85)  P for trend 0.04 0.02 0.28 0.07 Female daily journal seafood intake  One serving or less per cycle 108/835 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 61/446 0.99 (0.70–1.39) 1.04 (0.73–1.47) 0.98 (0.68–1.40) 1.00 (0.70–1.42)  Four to seven servings per cycle 83/605 1.00 (0.73–1.36) 1.07 (0.78–1.47) 0.97 (0.68–1.38) 1.03 (0.75–1.42)  Eight servings or more per cycle 93/486 1.38 (1.02–1.89) 1.60 (1.15–2.22) 1.42 (0.96–2.11) 1.44 (1.04–2.01)  P for trend 0.06 0.01 0.13 0.04 Cox models for discrete survival time accounting for left truncation were used to calculate the FORs and 95% CIs. a Model 1 adjusted for female age, difference in male and female age, female race (non-Hispanic white vs other), male exercise (yes vs no), male and female alcohol intake (one or more times per week vs less than one time per week), and cycle length. b Model 2 adjusted for variables in model 1 plus male or female partner seafood intake during follow-up. c Model 3 adjusted for variables in model 1 plus SIF (modeled with a linear and squared term). View Large Table 3. Associations Between Male and Female Seafood Intake During Follow-Up and TTP (n = 501 Couples) Variable Pregnancies/Cycles FOR (95% CI) Unadjusted Model 1a Model 2b Model 3c Male daily journal seafood intake  One serving or less per cycle 99/814 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 58/422 1.15 (0.81–1.64) 1.17 (0.82–1.68) 1.18 (0.81–1.71) 1.06 (0.74–1.52)  Four to seven servings per cycle 90/575 1.25 (0.91–1.71) 1.28 (0.93–1.76) 1.23 (0.87–1.74) 1.19 (0.86–1.85)  Eight servings or more per cycle 98/561 1.37 (1.01–1.87) 1.47 (1.07–2.03) 1.24 (0.84–1.83) 1.33 (0.96–1.85)  P for trend 0.04 0.02 0.28 0.07 Female daily journal seafood intake  One serving or less per cycle 108/835 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 61/446 0.99 (0.70–1.39) 1.04 (0.73–1.47) 0.98 (0.68–1.40) 1.00 (0.70–1.42)  Four to seven servings per cycle 83/605 1.00 (0.73–1.36) 1.07 (0.78–1.47) 0.97 (0.68–1.38) 1.03 (0.75–1.42)  Eight servings or more per cycle 93/486 1.38 (1.02–1.89) 1.60 (1.15–2.22) 1.42 (0.96–2.11) 1.44 (1.04–2.01)  P for trend 0.06 0.01 0.13 0.04 Variable Pregnancies/Cycles FOR (95% CI) Unadjusted Model 1a Model 2b Model 3c Male daily journal seafood intake  One serving or less per cycle 99/814 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 58/422 1.15 (0.81–1.64) 1.17 (0.82–1.68) 1.18 (0.81–1.71) 1.06 (0.74–1.52)  Four to seven servings per cycle 90/575 1.25 (0.91–1.71) 1.28 (0.93–1.76) 1.23 (0.87–1.74) 1.19 (0.86–1.85)  Eight servings or more per cycle 98/561 1.37 (1.01–1.87) 1.47 (1.07–2.03) 1.24 (0.84–1.83) 1.33 (0.96–1.85)  P for trend 0.04 0.02 0.28 0.07 Female daily journal seafood intake  One serving or less per cycle 108/835 1.0 (Reference) 1.0 (Reference) 1.0 (Reference) 1.0 (Reference)  One to three servings per cycle 61/446 0.99 (0.70–1.39) 1.04 (0.73–1.47) 0.98 (0.68–1.40) 1.00 (0.70–1.42)  Four to seven servings per cycle 83/605 1.00 (0.73–1.36) 1.07 (0.78–1.47) 0.97 (0.68–1.38) 1.03 (0.75–1.42)  Eight servings or more per cycle 93/486 1.38 (1.02–1.89) 1.60 (1.15–2.22) 1.42 (0.96–2.11) 1.44 (1.04–2.01)  P for trend 0.06 0.01 0.13 0.04 Cox models for discrete survival time accounting for left truncation were used to calculate the FORs and 95% CIs. a Model 1 adjusted for female age, difference in male and female age, female race (non-Hispanic white vs other), male exercise (yes vs no), male and female alcohol intake (one or more times per week vs less than one time per week), and cycle length. b Model 2 adjusted for variables in model 1 plus male or female partner seafood intake during follow-up. c Model 3 adjusted for variables in model 1 plus SIF (modeled with a linear and squared term). View Large The positive association between seafood intake and fecundity was stronger for couples when both partners consumed high amounts of seafood during follow-up (Fig. 1). On average, the estimated percentages of couples who were pregnant by 6 and 12 months among the couples who consumed eight or more seafood servings per cycle were 81% and 92% compared with 64% and 79% among the couples consuming less. This translated into an adjusted FOR of 1.61 (95% CI, 1.17 to 2.22; Supplemental Table 3) and a 13% lower absolute difference in the incidence of infertility. The positive associations between male, female, and couple seafood intake and fecundity were slightly attenuated after adjustment for SIF; however, the FORs for the highest female and couple seafood consumers remained statistically significant. Figure 1. View largeDownload slide Interaction between male and female partner seafood intake during follow-up on TTP (n = 501 couples). Cox models for discrete survival time accounting for left truncation were used to calculate the adjusted cumulative probabilities of pregnancy at the average value for continuous covariates and the most common value for categorical covariates (female age, 30 years; difference in male and female age, 1.8 years; female non-Hispanic white race; no male exercise; male alcohol intake one or more times per week; female alcohol intake less than one time per week; and cycle length, 30 days). Figure 1. View largeDownload slide Interaction between male and female partner seafood intake during follow-up on TTP (n = 501 couples). Cox models for discrete survival time accounting for left truncation were used to calculate the adjusted cumulative probabilities of pregnancy at the average value for continuous covariates and the most common value for categorical covariates (female age, 30 years; difference in male and female age, 1.8 years; female non-Hispanic white race; no male exercise; male alcohol intake one or more times per week; female alcohol intake less than one time per week; and cycle length, 30 days). In sensitivity analyses aimed at testing our assumption that the days with seafood intake data missing in the journals represented days with zero seafood intake, the associations between male, female, and couple daily journal seafood intake and fecundity were attenuated but still positively related to fecundity (Supplemental Table 4). The association between higher seafood intake for both partners and TTP was also robust in the sensitivity analyses with further adjustment and stratification by propensity score (Supplemental Table 5). Finally, the unmeasured confounding analyses showed that the observed FOR of 1.61 could only be explained by an unmeasured confounder that was associated with both the exposure and the outcome by a risk ratio of ≥2.13-fold, above and beyond the measured confounders. Discussion In the present prospective cohort study with preconception enrollment and daily follow-up of couples, the seafood intake in both partners was associated with a greater frequency of sexual intercourse and fecundity. Specifically, the daily odds of sexual intercourse were 39% greater when both partners consumed seafood on the same day. Also, for couples in which both partners consumed eight or more seafood servings per cycle had 61% greater fecundity and a 13% lower absolute difference in the incidence of infertility compared with couples consuming less seafood. The reported data on seafood intake and TTP, although sparse, are conflicting (19, 20). Several reasons exist for this heterogeneity, including differences in study designs, primary sources, types and range of seafood consumed, and outcomes assessments. Both previous studies relied on a retrospective report of seafood intake, which could have introduced a substantial measurement error. Our study has illustrated this point well, because we found no associations between baseline seafood intake, assessed using a typical retrospective questionnaire, and fecundity, despite the moderate positive correlations with the prospectively collected intake. Insufficient power could also be an issue. A retrospective cohort study from Sweden found no differences in the TTP comparing women differentially exposed to fatty fish contaminated with persistent organochlorines. However, within each group, the consumption of locally caught fatty fish had a marginally important, positive relation with fecundity (success OR, 1.27; 95% CI, 0.96 to 1.69; and success OR, 1.36; 95% CI, 0.96 to 1.94) (20). Only one study found a detrimental effect of female seafood intake on fecundity. That retrospective TTP study of recently pregnant female anglers found that women who consumed one or more fish meal per month from Lake Ontario (a highly contaminated source of fish) had reduced fecundity (fecundability ratio, 0.73; 95% CI, 0.54 to 0.98). No associations were found between male partner intake and TTP (19). A previous report from the LIFE study found inverse relations between select male and female serum organochlorines and dioxins and fecundity (21) and no associations with blood mercury concentrations and TTP (22). These findings suggest that the fertility benefits of seafood consumption might outweigh the potential harms of environmental pollutants carried by these foods. In agreement with our findings, a recent prospective study found that among US women trying to get pregnant who did not use fish oil supplements, intake of omega-3 fatty acids, a primary nutrient found in seafood, was associated with higher fecundity (fecundability ratio, 1.40; 95% CI, 1.13 to 1.73 for quartile 4 vs quartile 1) (12). Similarly, in a prospective cohort study of women undergoing infertility treatment with assisted reproductive technologies, higher serum levels and intake of long-chain omega-3 fatty acids were associated with a higher probability of achieving pregnancy and a live birth (23). We observed a positive association between seafood intake and SIF, supporting popular beliefs of the aphrodisiac properties of seafood. This association did not completely explain the relationship with fecundity, suggesting that the effects of seafood could result from mechanisms other than increased sexual activity. Several studies have found positive associations between omega fatty acid intake (24, 25), seafood intake (26, 27), and dietary patterns prioritizing seafood intake (28–31) and semen quality parameters, lending support to the idea that higher seafood intake could increase the quantity and quality of sperm. Among women, dietary intake of docosapentaenoic acid was associated with a lower risk of anovulation and dietary intake of total marine omega-3 polyunsaturated fats was associated with increased luteal-phase progesterone concentrations (13), suggesting beneficial effects of seafood on ovulation and menstrual cycle function. Finally, two separate infertility cohort studies have shown that embryo quality measures were improved among women with higher fish (32) and docosahexaenoic acid (33) intake, supporting a favorable role of seafood intake on early embryo development. Although these previous studies support our findings that greater seafood intake might promote fecundity through various biological mechanisms, it is important to consider alternate explanations. First, individuals with higher seafood intake could have healthier diets overall, which we were unable to account for in the present study. However, the estimated e-value of 2.13 decreases the likelihood of this explanation. Although it is conceivable that unmeasured dietary factors are associated with seafood intake by a risk ratio of >2.13-fold, we are unaware of any studies linking any specific dietary or lifestyle factors to fecundity by a risk ratio (or FOR) >2.13-fold. For context, the FOR comparing women aged <27 years to those aged ≥35 years in this cohort was 1.96. Thus, although residual confounding is possible, it is unlikely to explain the entire association. Second, couples who consume greater amounts of seafood together may share more meals and thus more time together (including nights), which might explain the association between sexual activity and subsequently fecundity. However, after we adjusted for SIF in our models, the association between seafood intake and fecundity remained, suggesting that this behavioral pathway cannot completely explain the association. Our study had other limitations. First, it consisted solely of couples planning pregnancy without medical assistance. All couples were also given fertility monitors to help time intercourse relative to ovulation and instructed to use them throughout the follow-up period. Because women using fertility monitors are more likely to get pregnant within two cycles than those who do not use a monitor (34), our results might not generalize to all women of reproductive age. However, the use of fertility monitors in our study was also a strength, because it removed any confounding by use of this or similar devices. We were able to assess differences in type (shellfish vs fish) and source (local vs unknown waters) of seafood using the baseline assessment tool and did not find any differences. However, we did not collect these details during follow-up, which limited our ability to distinguish between specific types of seafood and their potential reproductive effects. Finally, although it would have been ideal to include an assessment of daily diet, this was not feasible given the high participant burden of daily 24-hour recalls; therefore, residual confounding by other dietary factors, including dietary supplements, is possible, although unlikely to explain the entire association as discussed. Our study had multiple strengths, including the reference standard assessment of TTP through the prospective use of fertility monitors and daily journals combined with in-home pregnancy testing. In addition, we had daily, prospective assessment of seafood intake. We were also able to reduce the likelihood of residual confounding by adjusting for many demographic and lifestyle factors. Because our sampling frameworks, which in Texas used the Parks and Wildlife Department's angler database for recruitment and in Michigan used a commercially available marketing database with recruitment filters to identify individuals with fishing interests, we were also able to study a unique population in which seafood intake was not tightly correlated with socioeconomic status. Moreover, despite the overrecruitment of anglers, the average seafood intake of our cohort was very similar to that of men and women from a representative US sample (35). Our population was also recruited because of presumed exposure to persistent environmental chemicals that have been linked to fecundity impairments (36). Thus, it is possible that our results regarding seafood consumption would be even stronger in a population unexposed to sources of seafood contamination. Finally, by including the male partners, we were able to evaluate the separate and joint effects of male and female seafood consumption, which is rare in fecundity studies. In conclusion, couples in which both partners consumed eight or more seafood servings per cycle, or approximately two or more seafood servings per week, had a significantly greater SIF and higher fecundity. These findings highlight the importance of a couples’ diet for fecundity and the need for appropriate preconception guidance. Future research is needed that specifically evaluates the potential harms associated with predatory fish intake, because such fish tends to contain greater levels of persistent environmental chemicals and mercury. Abbreviations: Abbreviations: FOR fecundability OR LIFE Longitudinal Investigation of Fertility and the Environment SIF sexual intercourse frequency TTP time to pregnancy Acknowledgments We thank the participants and staff of the Longitudinal Investigation of Fertility and the Environment study for their valuable contributions to this research. Financial Support: The Longitudinal Investigation of Fertility and the Environment study was funded by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grants N01-HD-3-3355, N01-HD-3-3356, and NOH-HD-3-3358 to G.M. Buck Louis). Dr. Gaskins is supported by a career development award from the National Institute of Environmental Health Sciences, National Institutes of Health (grant K99ES026648). Disclosure Summary: The authors have nothing to disclose. References 1. Thoma ME , McLain AC , Louis JF , King RB , Trumble AC , Sundaram R , Buck Louis GM . Prevalence of infertility in the United States as estimated by the current duration approach and a traditional constructed approach . Fertil Steril . 2013 ; 99 ( 5 ): 1324 – 1331.e1 . 2. Slama R , Hansen OK , Ducot B , Bohet A , Sorensen D , Giorgis Allemand L , Eijkemans MJ , Rosetta L , Thalabard JC , Keiding N , Bouyer J . Estimation of the frequency of involuntary infertility on a nation-wide basis . Hum Reprod . 2012 ; 27 ( 5 ): 1489 – 1498 . 3. Katz P , Showstack J , Smith JF , Nachtigall RD , Millstein SG , Wing H , Eisenberg ML , Pasch LA , Croughan MS , Adler N . Costs of infertility treatment: results from an 18-month prospective cohort study . Fertil Steril . 2011 ; 95 ( 3 ): 915 – 921 . 4. Harris JA , Menke MN , Haefner JK , Moniz MH , Perumalswami CR . Geographic access to assisted reproductive technology health care in the United States: a population-based cross-sectional study . Fertil Steril . 2017 ; 107 ( 4 ): 1023 – 1027 . 5. Centers for Disease Control and Prevention (CDC), American Society for Reproductive Medicine (ASRM), Society of Assisted Reproductive Technology (SART) . 2010 Assisted Reproductive Technology Report: National Summary of Fertility Clinic Reports . Atlanta, GA : U.S. Department of Health and Human Services ; 2012 . 6. United States Department of Agriculture (USDA). Dietary Guidelines for Americans 2015-2020. Available at: health.gov/dietaryguidelines/2015/resources/2015-2020_Dietary_Guidelines.pdf. Accessed 8 August 2017. 7. Willett WC , Sacks F , Trichopoulou A , Drescher G , Ferro-Luzzi A , Helsing E , Trichopoulos D . Mediterranean diet pyramid: a cultural model for healthy eating . Am J Clin Nutr . 1995 ; 61 ( 6 , Suppl ): 1402S – 1406S . 8. Vandermeersch G , Lourenço HM , Alvarez-Muñoz D , Cunha S , Diogène J , Cano-Sancho G , Sloth JJ , Kwadijk C , Barcelo D , Allegaert W , Bekaert K , Fernandes JO , Marques A , Robbens J . Environmental contaminants of emerging concern in seafood—European database on contaminant levels . Environ Res . 2015 ; 143 ( Pt B ): 29 – 45 . 9. Rattan S , Zhou C , Chiang C , Mahalingam S , Brehm E , Flaws JA . Exposure to endocrine disruptors during adulthood: consequences for female fertility . J Endocrinol . 2017 ; 233 ( 3 ): R109 – R129 . 10. Bloom MS , Parsons PJ , Kim D , Steuerwald AJ , Vaccari S , Cheng G , Fujimoto VY . Toxic trace metals and embryo quality indicators during in vitro fertilization (IVF) . Reprod Toxicol . 2011 ; 31 ( 2 ): 164 – 170 . 11. Rignell-Hydbom A , Axmon A , Lundh T , Jönsson BA , Tiido T , Spano M . Dietary exposure to methyl mercury and PCB and the associations with semen parameters among Swedish fishermen . Environ Health . 2007 ; 6 ( 1 ): 14 . 12. Wise LA , Wesselink AK , Tucker KL , Saklani S , Mikkelsen EM , Cueto H , Riis AH , Trolle E , McKinnon CJ , Hahn KA , Rothman KJ , Sørensen HT , Hatch EE . Dietary fat intake and fecundability in two preconception cohort studies . Am J Epidemiol . 2018 ; 187 ( 1 ): 60 – 74 . 13. Mumford SL , Chavarro JE , Zhang C , Perkins NJ , Sjaarda LA , Pollack AZ , Schliep KC , Michels KA , Zarek SM , Plowden TC , Radin RG , Messer LC , Frankel RA , Wactawski-Wende J . Dietary fat intake and reproductive hormone concentrations and ovulation in regularly menstruating women . Am J Clin Nutr . 2016 ; 103 ( 3 ): 868 – 877 . 14. Esmaeili V , Shahverdi AH , Moghadasian MH , Alizadeh AR . Dietary fatty acids affect semen quality: a review . Andrology . 2015 ; 3 ( 3 ): 450 – 461 . 15. The Food and Drug Administration (FDA) and the Environmental Protection Agency (EPA). FDA and EPA Issue Final Fish Consumption Advice. 2017. Available at: www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm537362.htm. Accessed 8 August 2017. 16. Buck Louis GM , Schisterman EF , Sweeney AM , Wilcosky TC , Gore-Langton RE , Lynch CD , Boyd Barr D , Schrader SM , Kim S , Chen Z , Sundaram R . Designing prospective cohort studies for assessing reproductive and developmental toxicity during sensitive windows of human reproduction and development—the LIFE study . Paediatr Perinat Epidemiol . 2011 ; 25 ( 5 ): 413 – 424 . 17. Cohen S, Williamson G. Perceived stress in a probability sample of the United States. In: Spacapan S, Scamp S, eds. The Social Psychology of Health: Claremont Symposium on Applied Social Psychology. Newbury Park, CA: Sage; 1998. 18. VanderWeele TJ , Ding P . Sensitivity analysis in observational research: introducing the e-value . Ann Intern Med . 2017 ; 167 ( 4 ): 268 – 274 . 19. Buck GM , Vena JE , Schisterman EF , Dmochowski J , Mendola P , Sever LE , Fitzgerald E , Kostyniak P , Greizerstein H , Olson J . Parental consumption of contaminated sport fish from Lake Ontario and predicted fecundability . Epidemiology . 2000 ; 11 ( 4 ): 388 – 393 . 20. Axmon A , Rylander L , Strömberg U , Hagmar L . Female fertility in relation to the consumption of fish contaminated with persistent organochlorine compounds . Scand J Work Environ Health . 2002 ; 28 ( 2 ): 124 – 132 . 21. Buck Louis GM , Sundaram R , Schisterman EF , Sweeney AM , Lynch CD , Gore-Langton RE , Maisog J , Kim S , Chen Z , Barr DB . Persistent environmental pollutants and couple fecundity: the LIFE study . Environ Health Perspect . 2013 ; 121 ( 2 ): 231 – 236 . 22. Buck Louis GM , Sundaram R , Schisterman EF , Sweeney AM , Lynch CD , Gore-Langton RE , Chen Z , Kim S , Caldwell KL , Barr DB . Heavy metals and couple fecundity, the LIFE study . Chemosphere . 2012 ; 87 ( 11 ): 1201 – 1207 . 23. Chiu YH , Karmon AE , Gaskins AJ , Arvizu M , Williams PL , Souter I , Rueda BR , Hauser R , Chavarro JE ; EARTH Study Team . Serum omega-3 fatty acids and treatment outcomes among women undergoing assisted reproduction . Hum Reprod . 2018 ; 33 ( 1 ): 156 – 165 . 24. Attaman JA , Toth TL , Furtado J , Campos H , Hauser R , Chavarro JE . Dietary fat and semen quality among men attending a fertility clinic . Hum Reprod . 2012 ; 27 ( 5 ): 1466 – 1474 . 25. Safarinejad MR . Effect of omega-3 polyunsaturated fatty acid supplementation on semen profile and enzymatic anti-oxidant capacity of seminal plasma in infertile men with idiopathic oligoasthenoteratospermia: a double-blind, placebo-controlled, randomised study . Andrologia . 2011 ; 43 ( 1 ): 38 – 47 . 26. Afeiche MC , Gaskins AJ , Williams PL , Toth TL , Wright DL , Tanrikut C , Hauser R , Chavarro JE . Processed meat intake is unfavorably and fish intake favorably associated with semen quality indicators among men attending a fertility clinic . J Nutr . 2014 ; 144 ( 7 ): 1091 – 1098 . 27. Eslamian G , Amirjannati N , Rashidkhani B , Sadeghi MR , Hekmatdoost A . Intake of food groups and idiopathic asthenozoospermia: a case-control study . Hum Reprod . 2012 ; 27 ( 11 ): 3328 – 3336 . 28. Vujkovic M , de Vries JH , Dohle GR , Bonsel GJ , Lindemans J , Macklon NS , van der Spek PJ , Steegers EA , Steegers-Theunissen RP . Associations between dietary patterns and semen quality in men undergoing IVF/ICSI treatment . Hum Reprod . 2009 ; 24 ( 6 ): 1304 – 1312 . 29. Gaskins AJ , Colaci DS , Mendiola J , Swan SH , Chavarro JE . Dietary patterns and semen quality in young men . Hum Reprod . 2012 ; 27 ( 10 ): 2899 – 2907 . 30. Cutillas-Tolín A , Mínguez-Alarcón L , Mendiola J , López-Espín JJ , Jørgensen N , Navarrete-Muñoz EM , Torres-Cantero AM , Chavarro JE . Mediterranean and western dietary patterns are related to markers of testicular function among healthy men . Hum Reprod . 2015 ; 30 ( 12 ): 2945 – 2955 . 31. Karayiannis D , Kontogianni MD , Mendorou C , Douka L , Mastrominas M , Yiannakouris N . Association between adherence to the Mediterranean diet and semen quality parameters in male partners of couples attempting fertility . Hum Reprod . 2017 ; 32 ( 1 ): 215 – 222 . 32. Braga DP , Halpern G , Setti AS , Figueira RC , Iaconelli A Jr , Borges E Jr . The impact of food intake and social habits on embryo quality and the likelihood of blastocyst formation . Reprod Biomed Online . 2015 ; 31 ( 1 ): 30 – 38 . 33. Hammiche F , Vujkovic M , Wijburg W , de Vries JH , Macklon NS , Laven JS , Steegers-Theunissen RP . Increased preconception omega-3 polyunsaturated fatty acid intake improves embryo morphology . Fertil Steril . 2011 ; 95 ( 5 ): 1820 – 1823 . 34. Robinson JE , Wakelin M , Ellis JE . Increased pregnancy rate with use of the Clearblue Easy Fertility Monitor . Fertil Steril . 2007 ; 87 ( 2 ): 329 – 334 . 35. Papanikolaou Y , Brooks J , Reider C , Fulgoni VL III . U.S. adults are not meeting recommended levels for fish and omega-3 fatty acid intake: results of an analysis using observational data from NHANES 2003-2008 . Nutr J . 2014 ; 13 ( 1 ): 31 . 36. Buck Louis GM . Persistent environmental pollutants and couple fecundity: an overview . Reproduction . 2014 ; 147 ( 4 ): R97 – R104 . Copyright © 2018 Endocrine Society

Djelmis, Josip;Ivanišević, Marina;Desoye, Gernot;van Poppel, Mireille;Berberović, Edina;Soldo, Dragan;Oreskovic, Slavko

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00272pmid: 29722816

Abstract Context Type 1 diabetes mellitus (T1DM) is associated with a disturbance of carbohydrate and lipid metabolism. Objective To determine whether T1DM alters maternal and neonatal fatty acid (FA) levels. Design Observational study. Setting Academic hospital. Patients Sixty pregnant women (30 women with T1DM with good glycemic control and 30 healthy women) were included in the study. Maternal blood, umbilical vein, and artery blood samples were collected immediately upon delivery. Following lipid extraction, the FA profiles of the total FA pool of maternal serum and umbilical vein and artery serum were determined by gas chromatography. Results Total FA concentration in maternal serum did not differ between the study groups; it was significantly higher in umbilical vein serum of the T1DM group compared with that in the control group [median (interquartile range)]: T1DM 2126.2 (1446.4 to 3181.3) and control 1073.8 (657.5 to 2226.0; P < 0.001), and in umbilical artery vein serum: T1DM 1805.7 (1393.1 to 2125.0) and control 990.0 (643.3 to 1668.0; P < 0.001). Composition of FAs in umbilical vein serum showed significantly higher concentrations of saturated, monounsaturated, and polyunsaturated FAs (SFAs, MUFAs, and PUFAs, respectively) in the T1DM group than compared with those in the control group (P = 0.001). Furthermore, cord blood levels of leptin (P < 0.001), C-peptide (P < 0.001), and insulin resistance (P = 0.015) were higher in the T1DM group compared with controls. Conclusion The neonates born to mothers with T1DM had higher concentrations of total FAs, SFAs and MUFAs, as well as PUFAs, compared with control newborns. In pregnant women with type 1 diabetes mellitus (T1DM), endogenous insulin production is absent or minimal, thus requiring exogenous insulin for glycemic control and prevention of ketoacidosis (1). In addition to disturbances of carbohydrate metabolism, T1DM is associated with changes in the content and composition of maternal lipids (2). Whether T1DM also influences content and composition of fetal lipids in the umbilical circulation is currently unknown. Maternal sources of fetal lipids include lipoprotein-associated triacylglycerols, phospholipids, and cholesterol esters, as well as free fatty acids (FAs). The current understanding of maternal lipid supply for the fetus involves the transfer of free FAs and free cholesterol, after which, both moieties are esterified in the fetal liver and packaged into lipoproteins (2). Despite the presence of FA transporters in the syncytiotrophoblast (3), free FAs are likely to cross this membrane by diffusion (4, 5). Nonessential FAs can also be formed de novo from carbohydrates and acetate in the fetus under the control of fetal insulin (6, 7). In normal circumstances, free FAs have a minor role in energy provision; however, they generally can serve as energy deposits in the neonatal period (6). Particularly important for the fetus are the essential FAs, i.e., long-chain polyunsaturated FA (LC-PUFAs) (8). One of those LC-PUFAs, docosahexaenoic acid (DHA), has a major role in the development of nervous system cells, and its cord blood concentrations are associated with psychomotor development at 6 months of age (9, 10). In pregnant women, there is an increased maternal-fetal transfer of essential over nonessential FAs (11, 12). The aim of the study was to investigate the impact of T1DM on materno-fetal levels of FAs. Specifically, we tested the hypothesis of whether T1DM alters the FA profiles of maternal blood, the umbilical vein, and artery blood. Materials and Methods Study population This prospective study included a total of 60 pregnant women with singleton pregnancies who had either T1DM (T1DM group, n = 30) or were healthy (control group, n = 30). Pregnant women with T1DM were followed up at the Department of Obstetrics and Gynecology, Zagreb University Hospital Center, from 2012 to 2016. Study women had a T1DM duration of 5 to 10 years, were without diabetes complications, had a normal course of pregnancy, and delivered healthy, eutrophic newborns within the 25th to 75th birthweight percentile, according to weeks of pregnancy, sex, and parity (13). Pregnant women with T1DM were on intensified insulin analogs therapy and had well-controlled glycemia. Healthy women underwent gestational diabetes (GDM) screening, according to International Association of Diabetes and Pregnancy Study Groups criteria, with 75 g oral glucose tolerance test between gestational weeks 24 and 34 (14). Women with preterm delivery, multiple pregnancies, diabetic complications (retinopathy, nephropathy, neuropathy, and chronic hypertension), fetal chromosomal anomalies, or malformations were excluded from the study. Additional exclusion criteria were smoking by self-report, n-3 and n-6 FA supplementation, gestational hypertension/pre-eclampsia, and GDM. Gestational age was determined by the first day of last menstruation and verified by ultrasound examination between 6 and 10 weeks of gestation. All pregnancies were terminated by elective Caesarean section (CS) between gestational weeks 39 and 40. The mothers fasted for at least 10 hours before the CS. Maternal blood glucose during delivery was maintained in the range of 4 to 5 mM by IV infusion of 500 mL 5% glucose with insulin. Pregnancies in mothers with T1DM were terminated by CS as a standard procedure, whereas in the control group, indications for CS were breach presentation, narrow pelvis, post-CS state, or CS on mother’s demand. Maternal venous blood samples (5 mL) were collected during delivery. Umbilical cord blood was obtained during the CS, immediately after delivery but before the placenta was removed through puncture of the umbilical vessels. Blood was centrifuged immediately after collection, and the sera were stored at −80°C until analysis. C-Peptide and leptin concentrations and HbA1c percentage were determined in maternal and umbilical vein serum. Free FA concentrations were quantified in maternal and umbilical vein and arterial serum. Ethical statement The study was approved by the Ethics Committee of the Clinical Department of Obstetrics and Gynecology, Zagreb University Hospital Center, School of Medicine, University of Zagreb (No. 021-1/117A-2012). All women included in the study provided written, informed consent for themselves and their newborns. Data collection The following parameters were recorded: maternal height (centimeters) and weight (kilograms) measured before pregnancy; gestational weight gain as the difference between weight before pregnancy (self-reported) and at time of delivery; and prepregnancy body mass index (BMI; kilograms per square meter; BMI), calculated from prepregnancy values. Maternal HbA1c was determined in each trimester in the T1DM group and between the 36th and 38th weeks of pregnancy in the control group. Neonatal birth weight (grams), birth length (centimeters), and 1 and 5 minute Apgar scores were measured. Blood sample analyses FA quantification Total lipids were extracted by a mixture of chloroform/methanol solvent, a method of increasing polarity, according to Folch et al. (15). Heptadecanoic acid (C17:0) was used as the internal standard. FAs from lipid extracts were converted to methyl esters by trans-esterification with methanolic HCl (International Organization for Standardization standard). The FA profile was determined by gas chromatography-mass spectrometry on a Varian 3400 (Varian, Palo Alto, CA), equipped with a Saturn II ion trap mass spectrometer operating in the electron impact mode, as previously described (12). The concentration of each FA (micrograms per milliliter) was quantified by comparing the area of the internal standard peak (C17:0) with the peak area surface of a single FA. The results are expressed as percent (micrograms per 100 µg FAs) of arachidonic acid (AA) and DHA. Glucose, HbA1c, C-peptide, and leptin quantification Glucose levels were quantified by the hexokinase method on a Cobas C301 analyzer, and HbA1c levels in whole blood were measured by the Turbidimetric Inhibition Immunoassay on a Cobas C501 analyzer with reagents from the manufacturer (Roche, Basel, Switzerland). C-peptide concentration was determined by ELISA (C-peptide ELISA; Mercodia, Uppsala, Sweden). Serum leptin concentration was determined by ELISA (Quantikine Human Leptin Immunoassay; R&D Systems, Minneapolis, MN). Neonatal insulin resistance was calculated according to homeostasis model assessment 2 (HOMA2) (16), using software available online (http://software.informer.com/landing/). Statistical analysis Statistical analyses were performed with SPSS statistical package (release by SPSS version 24 software; IBM, Chicago, IL). Continuous variables are expressed as the means ± SD or medians (25th to 75th percentile) for a skewed distribution, and qualitative variables are presented as frequencies and percentages. Between-group differences in normally distributed continuous variables were assessed with a t test. The Mann-Whitney U test was used for variables with a skewed distribution, and a χ2-test was used for proportions. For repeated measurements of continuous data, the Wilcoxon signed-rank test was used. Pearson correlation coefficient (r) was used to assess linear dependence between normally distributed variables. Linear regression was performed between the concentration of total FAs in maternal and umbilical vein serum and between docosahexaenoic and arachidonic FAs. Data that were not normally distributed were log transformed before analyses. Regression with Spearman correlation coefficient (rs) was performed for non-normality data. Statistical tests were two sided. Statistical analyses were considered significant when P < 0.05. Results Maternal and neonatal characteristics Women with T1DM and controls did not differ (P > 0.05) in maternal age, BMI, or gestational weight gain (Table 1). Neonates did not differ in gestational age at delivery, neonatal birth weight, or Apgar scores at 1 and 5 minutes. Neonates born to mothers with T1DM were shorter (P = 0.001) and thus, had a higher ponderal index (P < 0.001) compared with control group neonates. The HbA1c values in the T1DM group dropped from 6.6% to 5.9% from the first to the third trimester (P = 0.022) but were still higher (P < 0.001) than in the control group (4.8%; Table 1). Table 1. Maternal and Neonatal Characteristics T1DM (n = 30) Control (n = 30) P Maternal characteristics  Age, y 30.0 ± 4.4 31.8 ± 5.3 0.184  Prepregnancy BMI, kg/m2 22.8 ± 3.9 22.9 ± 3.1 0.989  Gestational weight gain, kg 13.9 ± 3.7 16.0 ± 4.5 0.063  HbA1C first trimester, %a 6.6 ± 0.9 NA  HbA1C second trimester, % 5.6 ± 1.2 NA  HbA1C third trimester, %a 5.9 ± 1.0 4.8 ± 0.6 0.006  Gestational age at delivery, wk 39.0 ± 0.6 39.1 ± 0.4 0.451 Maternal serum measurements  C-Peptide, pM 235.5 (94–554.8) 570.0 (405–1089.6) 0.007  Glucose, mM 4.0 (2.9–6.4) 3.5 (3.0–4.1) 0.105  Leptin, ng/mL 25.0 (15.6–38.7) 19.7 (13.0–32.2) 0.276 Neonatal characteristics  Birth weight, g 3245.0 ± 366.4 3347.0 ± 338.8 0.264  Birth length, cm 48.5 ± 1.7 50.2 ± 2.0 0.001  Ponderal index, g, × 100/(cm)3 2.8 (2.7–3.0) 2.6 (2.5–2.7) <0.001  Apgar score at 1 min 9.6 ± 0.03 9.5 ± 0.1 0.156  Apgar score at 5 min 9.9 ± 0.2 9.9 ± 0.2 >0.99  Sex ratio: male/female, n (%)b 14/16 (46.7/53.3) 20/10 (66.7/33.3) 0.098 Umbilical vein blood measurements  C-Peptide, pM 867.3 (546–1149.8) 323.4 (214–447.3) <0.001  Insulin resistance HOMA2 1.5 (1.1–2.6) 0.7 (0.6–1.0) 0.015  Glucose, mM 3.2 (2.2–4.4) 2.9 (2.6–3.19) 0.108  Leptin, ng/mL 15.8 (8.4–24.9) 6.8 (3.9–11.6) <0.001 T1DM (n = 30) Control (n = 30) P Maternal characteristics  Age, y 30.0 ± 4.4 31.8 ± 5.3 0.184  Prepregnancy BMI, kg/m2 22.8 ± 3.9 22.9 ± 3.1 0.989  Gestational weight gain, kg 13.9 ± 3.7 16.0 ± 4.5 0.063  HbA1C first trimester, %a 6.6 ± 0.9 NA  HbA1C second trimester, % 5.6 ± 1.2 NA  HbA1C third trimester, %a 5.9 ± 1.0 4.8 ± 0.6 0.006  Gestational age at delivery, wk 39.0 ± 0.6 39.1 ± 0.4 0.451 Maternal serum measurements  C-Peptide, pM 235.5 (94–554.8) 570.0 (405–1089.6) 0.007  Glucose, mM 4.0 (2.9–6.4) 3.5 (3.0–4.1) 0.105  Leptin, ng/mL 25.0 (15.6–38.7) 19.7 (13.0–32.2) 0.276 Neonatal characteristics  Birth weight, g 3245.0 ± 366.4 3347.0 ± 338.8 0.264  Birth length, cm 48.5 ± 1.7 50.2 ± 2.0 0.001  Ponderal index, g, × 100/(cm)3 2.8 (2.7–3.0) 2.6 (2.5–2.7) <0.001  Apgar score at 1 min 9.6 ± 0.03 9.5 ± 0.1 0.156  Apgar score at 5 min 9.9 ± 0.2 9.9 ± 0.2 >0.99  Sex ratio: male/female, n (%)b 14/16 (46.7/53.3) 20/10 (66.7/33.3) 0.098 Umbilical vein blood measurements  C-Peptide, pM 867.3 (546–1149.8) 323.4 (214–447.3) <0.001  Insulin resistance HOMA2 1.5 (1.1–2.6) 0.7 (0.6–1.0) 0.015  Glucose, mM 3.2 (2.2–4.4) 2.9 (2.6–3.19) 0.108  Leptin, ng/mL 15.8 (8.4–24.9) 6.8 (3.9–11.6) <0.001 Data are means (SD) unless otherwise indicated; data are median (first and third quartiles); t test was used to test statistical difference between two groups or Mann-Whitney U test. P values in bold are significant. Abbreviation: NA, not available. a Wilcoxon signed-rank test; HbA1c between first and third trimester in T1DM; P = 0.022. b The χ2 test was used on group comparison. View Large Table 1. Maternal and Neonatal Characteristics T1DM (n = 30) Control (n = 30) P Maternal characteristics  Age, y 30.0 ± 4.4 31.8 ± 5.3 0.184  Prepregnancy BMI, kg/m2 22.8 ± 3.9 22.9 ± 3.1 0.989  Gestational weight gain, kg 13.9 ± 3.7 16.0 ± 4.5 0.063  HbA1C first trimester, %a 6.6 ± 0.9 NA  HbA1C second trimester, % 5.6 ± 1.2 NA  HbA1C third trimester, %a 5.9 ± 1.0 4.8 ± 0.6 0.006  Gestational age at delivery, wk 39.0 ± 0.6 39.1 ± 0.4 0.451 Maternal serum measurements  C-Peptide, pM 235.5 (94–554.8) 570.0 (405–1089.6) 0.007  Glucose, mM 4.0 (2.9–6.4) 3.5 (3.0–4.1) 0.105  Leptin, ng/mL 25.0 (15.6–38.7) 19.7 (13.0–32.2) 0.276 Neonatal characteristics  Birth weight, g 3245.0 ± 366.4 3347.0 ± 338.8 0.264  Birth length, cm 48.5 ± 1.7 50.2 ± 2.0 0.001  Ponderal index, g, × 100/(cm)3 2.8 (2.7–3.0) 2.6 (2.5–2.7) <0.001  Apgar score at 1 min 9.6 ± 0.03 9.5 ± 0.1 0.156  Apgar score at 5 min 9.9 ± 0.2 9.9 ± 0.2 >0.99  Sex ratio: male/female, n (%)b 14/16 (46.7/53.3) 20/10 (66.7/33.3) 0.098 Umbilical vein blood measurements  C-Peptide, pM 867.3 (546–1149.8) 323.4 (214–447.3) <0.001  Insulin resistance HOMA2 1.5 (1.1–2.6) 0.7 (0.6–1.0) 0.015  Glucose, mM 3.2 (2.2–4.4) 2.9 (2.6–3.19) 0.108  Leptin, ng/mL 15.8 (8.4–24.9) 6.8 (3.9–11.6) <0.001 T1DM (n = 30) Control (n = 30) P Maternal characteristics  Age, y 30.0 ± 4.4 31.8 ± 5.3 0.184  Prepregnancy BMI, kg/m2 22.8 ± 3.9 22.9 ± 3.1 0.989  Gestational weight gain, kg 13.9 ± 3.7 16.0 ± 4.5 0.063  HbA1C first trimester, %a 6.6 ± 0.9 NA  HbA1C second trimester, % 5.6 ± 1.2 NA  HbA1C third trimester, %a 5.9 ± 1.0 4.8 ± 0.6 0.006  Gestational age at delivery, wk 39.0 ± 0.6 39.1 ± 0.4 0.451 Maternal serum measurements  C-Peptide, pM 235.5 (94–554.8) 570.0 (405–1089.6) 0.007  Glucose, mM 4.0 (2.9–6.4) 3.5 (3.0–4.1) 0.105  Leptin, ng/mL 25.0 (15.6–38.7) 19.7 (13.0–32.2) 0.276 Neonatal characteristics  Birth weight, g 3245.0 ± 366.4 3347.0 ± 338.8 0.264  Birth length, cm 48.5 ± 1.7 50.2 ± 2.0 0.001  Ponderal index, g, × 100/(cm)3 2.8 (2.7–3.0) 2.6 (2.5–2.7) <0.001  Apgar score at 1 min 9.6 ± 0.03 9.5 ± 0.1 0.156  Apgar score at 5 min 9.9 ± 0.2 9.9 ± 0.2 >0.99  Sex ratio: male/female, n (%)b 14/16 (46.7/53.3) 20/10 (66.7/33.3) 0.098 Umbilical vein blood measurements  C-Peptide, pM 867.3 (546–1149.8) 323.4 (214–447.3) <0.001  Insulin resistance HOMA2 1.5 (1.1–2.6) 0.7 (0.6–1.0) 0.015  Glucose, mM 3.2 (2.2–4.4) 2.9 (2.6–3.19) 0.108  Leptin, ng/mL 15.8 (8.4–24.9) 6.8 (3.9–11.6) <0.001 Data are means (SD) unless otherwise indicated; data are median (first and third quartiles); t test was used to test statistical difference between two groups or Mann-Whitney U test. P values in bold are significant. Abbreviation: NA, not available. a Wilcoxon signed-rank test; HbA1c between first and third trimester in T1DM; P = 0.022. b The χ2 test was used on group comparison. View Large Glucose, C-peptide, leptin, glucose levels, and insulin resistance in maternal and neonatal blood Maternal C-peptide concentration was higher (P = 0.007) in the control group than in the T1DM group (Table 1). Women with T1DM had median C-peptide levels of 235.5 (interquartile range (94 to 554.8) pM, indicating that their β-cell function and endogenous insulin production were, in part, preserved. Maternal plasma glucose and leptin concentrations were similar in both groups. In the neonates, the umbilical cord vein concentrations of C-peptide (P < 0.001) and leptin (P < 0.001), as well as HOMA2 insulin resistance (P = 0.015), were higher in the T1DM group, whereas plasma glucose concentrations were similar in both groups (Table 1). FAs in maternal serum The total FA concentration in serum of women with T1DM did not differ significantly (P = 0.074) when compared with the control group (Table 2). Although the concentration of saturated FAs (SFAs) in mothers with T1DM was not significantly different from controls (P = 0.071), the percentage of SFAs was significantly lower in T1DM than in the control group (34.8% vs 37.6%, P = 0.006). The concentration of both monounsaturated FAs (MUFAs) and PUFAs was significantly higher in serum of women with T1DM compared with controls (P = 0.016 and P = 0.027, respectively). However, the percentage of MUFAs and PUFAs was not significantly different between the groups (Table 2). Table 2. Concentration of Total, SFA, MUFA, and PUFA and Percent of FAs (micrograms per 100 µg FAs) in Maternal and Umbilical Vein and Artery Blood Maternal Serum Umbilical Vein Umbilical Artery FAs T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Total FAs, µg/mL 5921.3 (3255.0–8124.3)a 3161.6 (113.9–7993.6)b 0.074 2126.2 (1446.4–3181.3)a,c 1073.8 (657.5–2226.0)b,d <0.001 1805.7 (1393.1–2125.0)c 990.0 (643.3–1668.0)d 0.001 SFAs, µg/mL 1892.4 (901.0–2530.6)e 916.8 (237.2–2372.4)f 0.071 529.1 (404.5–770.0)e,g 223.5 (78.5–551.4)f,h 0.001 201.8 (170.8–254.8)g 229.9 (108.6–310.1)h 0.011 Percent SFAs, µg/100 µg FA 34.8 (32.7–37.5)i 37.6 (35.8–39.9)j 0.006 34.9 (33.1–38.5)i,k 44.3 (33.49–49.0)j,l 0.002 34.7 (29.5–37.3)k 44.0 (33.9–49.0)l <0.001 MUFAs, µg/mL 1250.6 (554.1–1646.4)m 500.2 (97.5–1154.9)n 0.016 530.7 (408.8–778.9)m,o 221.2 (114–556.5)n,p 0.001 326.6 (208.1–430.9)o 221.3 (129.0–340.5)p 0.033 Percent MUFAs, µg/100 µg FA 21.3 (19.4–22.5)q 20.3 (16.6–24.4)r 0.237 31.9 (27.9–36.4)q,s 19.9 (14.1–29.5)r,t 0.002 33.8 (27.4–38.)s 30.7 (19.5–34.7)t 0.135 PUFAs (n-3 + n-6 FAs), µg/mL 2209.0 (1248.7–3452.6)u 1061.3 (270.6–2884.6)v 0.027 489.1 (368.5–858.9)u,w 140.3 (36.2–491.8)v,x 0.001 282.9 (222.6–408.3)w 186,1 (126.0–250.6)x 0.004 Percent of PUFA (n-3 + n-6 FAs), µg/100 µg FA 43.3 (40.6–45.3)y 41.7 (35.5–44.1)z 0.089 32.8 (29.8–35.7)y,aa 31.4 (28.3–35.7)z,bb 0.525 31.3 (30.0–36.9)aa 28.8 (22.1–34.9)bb 0.008 Maternal Serum Umbilical Vein Umbilical Artery FAs T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Total FAs, µg/mL 5921.3 (3255.0–8124.3)a 3161.6 (113.9–7993.6)b 0.074 2126.2 (1446.4–3181.3)a,c 1073.8 (657.5–2226.0)b,d <0.001 1805.7 (1393.1–2125.0)c 990.0 (643.3–1668.0)d 0.001 SFAs, µg/mL 1892.4 (901.0–2530.6)e 916.8 (237.2–2372.4)f 0.071 529.1 (404.5–770.0)e,g 223.5 (78.5–551.4)f,h 0.001 201.8 (170.8–254.8)g 229.9 (108.6–310.1)h 0.011 Percent SFAs, µg/100 µg FA 34.8 (32.7–37.5)i 37.6 (35.8–39.9)j 0.006 34.9 (33.1–38.5)i,k 44.3 (33.49–49.0)j,l 0.002 34.7 (29.5–37.3)k 44.0 (33.9–49.0)l <0.001 MUFAs, µg/mL 1250.6 (554.1–1646.4)m 500.2 (97.5–1154.9)n 0.016 530.7 (408.8–778.9)m,o 221.2 (114–556.5)n,p 0.001 326.6 (208.1–430.9)o 221.3 (129.0–340.5)p 0.033 Percent MUFAs, µg/100 µg FA 21.3 (19.4–22.5)q 20.3 (16.6–24.4)r 0.237 31.9 (27.9–36.4)q,s 19.9 (14.1–29.5)r,t 0.002 33.8 (27.4–38.)s 30.7 (19.5–34.7)t 0.135 PUFAs (n-3 + n-6 FAs), µg/mL 2209.0 (1248.7–3452.6)u 1061.3 (270.6–2884.6)v 0.027 489.1 (368.5–858.9)u,w 140.3 (36.2–491.8)v,x 0.001 282.9 (222.6–408.3)w 186,1 (126.0–250.6)x 0.004 Percent of PUFA (n-3 + n-6 FAs), µg/100 µg FA 43.3 (40.6–45.3)y 41.7 (35.5–44.1)z 0.089 32.8 (29.8–35.7)y,aa 31.4 (28.3–35.7)z,bb 0.525 31.3 (30.0–36.9)aa 28.8 (22.1–34.9)bb 0.008 Data presented are median and first and third quartiles. The Mann-Whitney U test was used to test significant differences between the two groups. Wilcoxon signed-rank test (see footnotes). P values in bold are significant. Abbreviations: MUFAs, monounsaturated FAs; SFAs, saturated FAs. a Maternal and umbilical vein serum concentration of total FAs in T1DM, P < 0.001. b Maternal and umbilical vein serum concentration of total FAs in control, P < 0.001. c Umbilical vein and umbilical artery serum concentration of total FAs in T1DM, P < 0.016. d Umbilical vein and umbilical artery serum concentration of total FAs in control, P = 0.892. e Maternal and umbilical vein serum concentration of SFAs in T1DM, P < 0.001. f Maternal and umbilical vein serum concentration of SFAs in control, P < 0.001. g Umbilical vein and umbilical artery serum concentration of SFAs in T1DM, P < 0.001. h Umbilical vein and umbilical artery serum concentration of SFAs in control, P = 0.765. i Maternal and umbilical vein serum percentage of SFAs in T1DM, P = 0.230. j Maternal and umbilical vein serum percentage of SFAs in control, P = 0.004. k Umbilical vein and umbilical artery percentage of SFAs in T1DM, P = 0.315. l Umbilical vein and umbilical artery percentage of SFAs in control, P = 0.271. m Maternal and umbilical vein serum concentration of MUFAs in T1DM, P < 0.001). n Maternal and umbilical vein serum concentration of MUFAs in control, P < 0.001. o Umbilical vein and umbilical artery serum concentration of MUFAs in T1DM, P < 0.001. p Umbilical vein and umbilical artery serum concentration of MUFAs in control, P = 0.299. q Maternal and umbilical vein serum percentage of MUFAs in T1DM, P < 0.001. r Maternal and umbilical vein serum percentage of MUFAs in control, P < 0.001. s Umbilical vein and umbilical artery percentage of MUFAs in T1DM, P = 0.222. t Umbilical vein and umbilical artery serum percentage of MUFAs in control, P = 0.003. u Maternal and umbilical vein serum concentration of PUFAs in T1DM, P < 0.001. v Maternal and umbilical vein serum concentration of PUFAs in control, P < 0.001. w Umbilical vein and umbilical artery serum concentration of PUFAs in T1DM, P < 0.001. x Umbilical vein and umbilical artery serum concentration of PUFAs in control, P < 0.016. y Maternal and umbilical vein serum percentage of PUFAs in T1DM, P < 0.001. z Maternal and umbilical vein serum percentage of PUFAs in control, P < 0.001. aa Umbilical vein and umbilical artery serum percentage of PUFAs in T1DM, P = 0.719. bb Umbilical vein and umbilical artery serum percentage of PUFAs in control, P = 0.028. View Large Table 2. Concentration of Total, SFA, MUFA, and PUFA and Percent of FAs (micrograms per 100 µg FAs) in Maternal and Umbilical Vein and Artery Blood Maternal Serum Umbilical Vein Umbilical Artery FAs T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Total FAs, µg/mL 5921.3 (3255.0–8124.3)a 3161.6 (113.9–7993.6)b 0.074 2126.2 (1446.4–3181.3)a,c 1073.8 (657.5–2226.0)b,d <0.001 1805.7 (1393.1–2125.0)c 990.0 (643.3–1668.0)d 0.001 SFAs, µg/mL 1892.4 (901.0–2530.6)e 916.8 (237.2–2372.4)f 0.071 529.1 (404.5–770.0)e,g 223.5 (78.5–551.4)f,h 0.001 201.8 (170.8–254.8)g 229.9 (108.6–310.1)h 0.011 Percent SFAs, µg/100 µg FA 34.8 (32.7–37.5)i 37.6 (35.8–39.9)j 0.006 34.9 (33.1–38.5)i,k 44.3 (33.49–49.0)j,l 0.002 34.7 (29.5–37.3)k 44.0 (33.9–49.0)l <0.001 MUFAs, µg/mL 1250.6 (554.1–1646.4)m 500.2 (97.5–1154.9)n 0.016 530.7 (408.8–778.9)m,o 221.2 (114–556.5)n,p 0.001 326.6 (208.1–430.9)o 221.3 (129.0–340.5)p 0.033 Percent MUFAs, µg/100 µg FA 21.3 (19.4–22.5)q 20.3 (16.6–24.4)r 0.237 31.9 (27.9–36.4)q,s 19.9 (14.1–29.5)r,t 0.002 33.8 (27.4–38.)s 30.7 (19.5–34.7)t 0.135 PUFAs (n-3 + n-6 FAs), µg/mL 2209.0 (1248.7–3452.6)u 1061.3 (270.6–2884.6)v 0.027 489.1 (368.5–858.9)u,w 140.3 (36.2–491.8)v,x 0.001 282.9 (222.6–408.3)w 186,1 (126.0–250.6)x 0.004 Percent of PUFA (n-3 + n-6 FAs), µg/100 µg FA 43.3 (40.6–45.3)y 41.7 (35.5–44.1)z 0.089 32.8 (29.8–35.7)y,aa 31.4 (28.3–35.7)z,bb 0.525 31.3 (30.0–36.9)aa 28.8 (22.1–34.9)bb 0.008 Maternal Serum Umbilical Vein Umbilical Artery FAs T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Total FAs, µg/mL 5921.3 (3255.0–8124.3)a 3161.6 (113.9–7993.6)b 0.074 2126.2 (1446.4–3181.3)a,c 1073.8 (657.5–2226.0)b,d <0.001 1805.7 (1393.1–2125.0)c 990.0 (643.3–1668.0)d 0.001 SFAs, µg/mL 1892.4 (901.0–2530.6)e 916.8 (237.2–2372.4)f 0.071 529.1 (404.5–770.0)e,g 223.5 (78.5–551.4)f,h 0.001 201.8 (170.8–254.8)g 229.9 (108.6–310.1)h 0.011 Percent SFAs, µg/100 µg FA 34.8 (32.7–37.5)i 37.6 (35.8–39.9)j 0.006 34.9 (33.1–38.5)i,k 44.3 (33.49–49.0)j,l 0.002 34.7 (29.5–37.3)k 44.0 (33.9–49.0)l <0.001 MUFAs, µg/mL 1250.6 (554.1–1646.4)m 500.2 (97.5–1154.9)n 0.016 530.7 (408.8–778.9)m,o 221.2 (114–556.5)n,p 0.001 326.6 (208.1–430.9)o 221.3 (129.0–340.5)p 0.033 Percent MUFAs, µg/100 µg FA 21.3 (19.4–22.5)q 20.3 (16.6–24.4)r 0.237 31.9 (27.9–36.4)q,s 19.9 (14.1–29.5)r,t 0.002 33.8 (27.4–38.)s 30.7 (19.5–34.7)t 0.135 PUFAs (n-3 + n-6 FAs), µg/mL 2209.0 (1248.7–3452.6)u 1061.3 (270.6–2884.6)v 0.027 489.1 (368.5–858.9)u,w 140.3 (36.2–491.8)v,x 0.001 282.9 (222.6–408.3)w 186,1 (126.0–250.6)x 0.004 Percent of PUFA (n-3 + n-6 FAs), µg/100 µg FA 43.3 (40.6–45.3)y 41.7 (35.5–44.1)z 0.089 32.8 (29.8–35.7)y,aa 31.4 (28.3–35.7)z,bb 0.525 31.3 (30.0–36.9)aa 28.8 (22.1–34.9)bb 0.008 Data presented are median and first and third quartiles. The Mann-Whitney U test was used to test significant differences between the two groups. Wilcoxon signed-rank test (see footnotes). P values in bold are significant. Abbreviations: MUFAs, monounsaturated FAs; SFAs, saturated FAs. a Maternal and umbilical vein serum concentration of total FAs in T1DM, P < 0.001. b Maternal and umbilical vein serum concentration of total FAs in control, P < 0.001. c Umbilical vein and umbilical artery serum concentration of total FAs in T1DM, P < 0.016. d Umbilical vein and umbilical artery serum concentration of total FAs in control, P = 0.892. e Maternal and umbilical vein serum concentration of SFAs in T1DM, P < 0.001. f Maternal and umbilical vein serum concentration of SFAs in control, P < 0.001. g Umbilical vein and umbilical artery serum concentration of SFAs in T1DM, P < 0.001. h Umbilical vein and umbilical artery serum concentration of SFAs in control, P = 0.765. i Maternal and umbilical vein serum percentage of SFAs in T1DM, P = 0.230. j Maternal and umbilical vein serum percentage of SFAs in control, P = 0.004. k Umbilical vein and umbilical artery percentage of SFAs in T1DM, P = 0.315. l Umbilical vein and umbilical artery percentage of SFAs in control, P = 0.271. m Maternal and umbilical vein serum concentration of MUFAs in T1DM, P < 0.001). n Maternal and umbilical vein serum concentration of MUFAs in control, P < 0.001. o Umbilical vein and umbilical artery serum concentration of MUFAs in T1DM, P < 0.001. p Umbilical vein and umbilical artery serum concentration of MUFAs in control, P = 0.299. q Maternal and umbilical vein serum percentage of MUFAs in T1DM, P < 0.001. r Maternal and umbilical vein serum percentage of MUFAs in control, P < 0.001. s Umbilical vein and umbilical artery percentage of MUFAs in T1DM, P = 0.222. t Umbilical vein and umbilical artery serum percentage of MUFAs in control, P = 0.003. u Maternal and umbilical vein serum concentration of PUFAs in T1DM, P < 0.001. v Maternal and umbilical vein serum concentration of PUFAs in control, P < 0.001. w Umbilical vein and umbilical artery serum concentration of PUFAs in T1DM, P < 0.001. x Umbilical vein and umbilical artery serum concentration of PUFAs in control, P < 0.016. y Maternal and umbilical vein serum percentage of PUFAs in T1DM, P < 0.001. z Maternal and umbilical vein serum percentage of PUFAs in control, P < 0.001. aa Umbilical vein and umbilical artery serum percentage of PUFAs in T1DM, P = 0.719. bb Umbilical vein and umbilical artery serum percentage of PUFAs in control, P = 0.028. View Large FAs in umbilical vein serum In umbilical vein serum, the concentration of total FAs, SFAs, MUFAs, and PUFAs was higher in T1DM compared with controls (Table 2). The percentage of SFAs in umbilical vein was lower in the T1DM group (34.9% vs 44.3%, P = 0.002). In contrast, the percentage of MUFAs was higher in T1DM (31.9% vs 19.9%, P = 0.002), whereas the percentage of PUFAs did not differ (32.8% vs 31.4%, P = 0.53; Table 2). FAs in umbilical artery serum In umbilical artery serum, the concentration of total FAs, MUFAs, and PUFAs was higher in T1DM when compared with the controls (Table 2). Compared with controls, the concentration of SFAs and the percentage of SFAs were lower in T1DM (201.8 vs 229.9, P = 0.011; 34.7% vs 44.0%, P < 0.001). The percentage of MUFAs did not differ between the two groups, whereas the percentage of PUFAs was higher in T1DM than in the control group (31.3% vs 28.8%, P = 0.008). Comparisons of FAs between maternal and umbilical vein serum The concentrations of total FAs, SFAs, MUFAs, and PUFAs were higher in maternal serum of both T1DM and control women when compared the corresponding umbilical vein serum (all P < 0.001; Table 2). In controls, maternal serum was less rich in SFAs compared with umbilical vein serum (37.6% vs 44.3%, P = 0.004), but this difference was not found in T1DM (34.8% vs 34.9%, P = 0.230). A lower percentage of MUFAs was found in maternal serum of T1DM (21.3% vs 31.9%) when compared with umbilical vein serum (P < 0.001), whereas in the control group, the percentage MUFAs was higher in maternal (20.3%) than in umbilical vein serum (19.9%, P < 0.001). A higher percentage of PUFAs was found in maternal serum than in umbilical vein serum in both groups (T1DM: 43.3% vs 32.8%; controls: 41.7% vs 31.4%). Comparisons of FAs between umbilical vein and artery serum In T1DM, the concentrations of total FAs, SFAs, MUFAs, and PUFAs were all significantly higher in umbilical vein compared with the corresponding umbilical artery serum, but the percentages were not different (Table 2). In controls, total FA concentration was higher in umbilical vein serum compared with artery serum. No differences in the concentration of SFAs and MUFAs were found between umbilical vein and artery serum of controls. The PUFA concentration was lower in umbilical vein than in artery serum. The percentage of SFAs was not different between venous and arterial serum in controls, whereas the percentage of MUFAs and PUFAs was lower and higher, respectively, in umbilical vein than in artery serum. FA profile in maternal, umbilical vein, and umbilical artery serum In maternal vein serum, the concentrations of all particular FAs tended to be higher in T1DM, but only the concentrations of oleic (C18:1n-9), DHA (C22:6n-3), and linoleic acid (C18:3n-6) were significantly different from controls (Table 3). Table 3. FA Profile in Maternal Vein Serum, Umbilical Vein, and Umbilical Artery Serum Maternal Serum Umbilical Vein Umbilical Artery FAs, µg/mL T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Myristic acid (C14:0) 41.8 (16.2–70.0) 8.8 (0.7–93.6) 0.143 8.8 (0.6–14.1) 2.2 (0.1–12.4) 0.069 12.5 (8.4–17.2) 4.2 (0.4–16.6) 0.058 Pentadecanoic acid (C15:0) 7.4 (1.4–13.7) 1.9 (0.0–14.5) 0.255 4.4 (0.4–8.1) 0.5 (0.0–4.9) 0.033 3.9 (0.4–7.5) 1.0 (0.0–6.8) 0.387 Palmitic acid (C16:0) 1481.5 (704.1–2490.1) 631.1 (190.2–1806.4) 0.051 321.3 (281.4–438.1) 119.1 (46.1–333.1) 0.001 313.0 (278.9–428.1) 316.7 (190.7–360.) 0.156 Stearic acid (C18:0) 260.3 (96.8–348.5) 62.7 (27.5–404.9) 0.084 152.8 (75.0–234.4) 36.7 (9.7–165.9) 0.001 155.1 (127.6–185.8) 106.2 (61.1–169.8) 0.098 Palmitoleic acid (C16:1n-7) 48.9 (23.3–66.0) 35.3 (7.0–114.9) 0.261 36.0 (25.4–53.1) 8.1 (2.2–40.7) 0.001 40.7 (34.2–55.9) 28.6 (19.9–44.8) 0.003 Oleic acid (C18:1n-9) 909.1 (396.9–1217.9) 260.5 (90.2–778.8) 0.012 240.0 (89.5–298.2) 65.9 (18.7–226.1) <0.001 237.2 (160.0–296.7) 122.7 (57.9–200.0) <0.001 Vaccenic acid (C18:1n-7) 6.7 (1.6–25.9) 0.0 (0.0–20.5 0.072 41.2 (0.0–54.6) 0.0 (0.0–39.2) 0.017 40.7 (0.0–51.2) 12.8 (0.0–75.0) 0.946 α-Linoleic acid (C18:3n-3) 10.2 (4.7–14.9) 4.4 (1.3–14.6) 0.133 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND Eicosapentanoic acid (C20:5n-3) 4.0 (0.2–8.0) 0.3 (0.0–8.8) 0.247 0.2 (0.0–1.1) 0.0 (0.0–0.0) NA 0.0 (0.0–0.5) 0.0 (0.0–0.0) NA DHA (C22:6n-3) 110.2 (55.8–143.4) 70.7 (21.1–117.6) 0.019 45.9 (15.7–78.1) 19.9 (3.3–44.5) 0.003 43.5 (34.8–56.2) 36.2 (24.9–58.4) 0.391 Linoleic acid (C18:3n-6) 12.9 (0.7–18.7) 0.8 (0.0–6.0) 0.006 6.6 (0.6–20.5) 0.3 (0.0–7.6) <0.001 7.2 (0.5–27.5) 1.1 (0.1–7.8) 0.009 AA (C20:4n-6) 259.4 (148.1–408.2) 163.5 (48.5–317.3) 0.051 202.9 (142.4–305.8) 59.2 (16.4–176.1) <0.001 193.1 (149.5–225.4) 124.4 (89.5–169.5) 0.016 AA:(DHA + EPA) 2.8 (2.5–3.1) 2.4 (1.5–3.5) 0.383 4.2 (3.7–4.6) 4.0 (3.5–4.8) 0.524 4.1 (3.4–4.9) 3.7 (2.9–4.1) 0.107 Percent AA, µg/100 µg FA 5.8 (5.2–6.7) 5.9 (4.2–7.1) 0.929 13.4 (12.0–15.3) 12.8 (10.2–14.7) 0.383 12.6 (9.8–14.3) 13.5 (8.3–15.5) 0.802 Percent DHA, µg/100 µg FA 2.0 (1.8–2.5) 1.8 (1.4–2.5) 0.336 3.1 (2.8–3.5) 3.1 (2.1–3.9) 0.913 2.7 (2.3–3.2) 3.4 (3.0–3.8) 0.022 Maternal Serum Umbilical Vein Umbilical Artery FAs, µg/mL T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Myristic acid (C14:0) 41.8 (16.2–70.0) 8.8 (0.7–93.6) 0.143 8.8 (0.6–14.1) 2.2 (0.1–12.4) 0.069 12.5 (8.4–17.2) 4.2 (0.4–16.6) 0.058 Pentadecanoic acid (C15:0) 7.4 (1.4–13.7) 1.9 (0.0–14.5) 0.255 4.4 (0.4–8.1) 0.5 (0.0–4.9) 0.033 3.9 (0.4–7.5) 1.0 (0.0–6.8) 0.387 Palmitic acid (C16:0) 1481.5 (704.1–2490.1) 631.1 (190.2–1806.4) 0.051 321.3 (281.4–438.1) 119.1 (46.1–333.1) 0.001 313.0 (278.9–428.1) 316.7 (190.7–360.) 0.156 Stearic acid (C18:0) 260.3 (96.8–348.5) 62.7 (27.5–404.9) 0.084 152.8 (75.0–234.4) 36.7 (9.7–165.9) 0.001 155.1 (127.6–185.8) 106.2 (61.1–169.8) 0.098 Palmitoleic acid (C16:1n-7) 48.9 (23.3–66.0) 35.3 (7.0–114.9) 0.261 36.0 (25.4–53.1) 8.1 (2.2–40.7) 0.001 40.7 (34.2–55.9) 28.6 (19.9–44.8) 0.003 Oleic acid (C18:1n-9) 909.1 (396.9–1217.9) 260.5 (90.2–778.8) 0.012 240.0 (89.5–298.2) 65.9 (18.7–226.1) <0.001 237.2 (160.0–296.7) 122.7 (57.9–200.0) <0.001 Vaccenic acid (C18:1n-7) 6.7 (1.6–25.9) 0.0 (0.0–20.5 0.072 41.2 (0.0–54.6) 0.0 (0.0–39.2) 0.017 40.7 (0.0–51.2) 12.8 (0.0–75.0) 0.946 α-Linoleic acid (C18:3n-3) 10.2 (4.7–14.9) 4.4 (1.3–14.6) 0.133 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND Eicosapentanoic acid (C20:5n-3) 4.0 (0.2–8.0) 0.3 (0.0–8.8) 0.247 0.2 (0.0–1.1) 0.0 (0.0–0.0) NA 0.0 (0.0–0.5) 0.0 (0.0–0.0) NA DHA (C22:6n-3) 110.2 (55.8–143.4) 70.7 (21.1–117.6) 0.019 45.9 (15.7–78.1) 19.9 (3.3–44.5) 0.003 43.5 (34.8–56.2) 36.2 (24.9–58.4) 0.391 Linoleic acid (C18:3n-6) 12.9 (0.7–18.7) 0.8 (0.0–6.0) 0.006 6.6 (0.6–20.5) 0.3 (0.0–7.6) <0.001 7.2 (0.5–27.5) 1.1 (0.1–7.8) 0.009 AA (C20:4n-6) 259.4 (148.1–408.2) 163.5 (48.5–317.3) 0.051 202.9 (142.4–305.8) 59.2 (16.4–176.1) <0.001 193.1 (149.5–225.4) 124.4 (89.5–169.5) 0.016 AA:(DHA + EPA) 2.8 (2.5–3.1) 2.4 (1.5–3.5) 0.383 4.2 (3.7–4.6) 4.0 (3.5–4.8) 0.524 4.1 (3.4–4.9) 3.7 (2.9–4.1) 0.107 Percent AA, µg/100 µg FA 5.8 (5.2–6.7) 5.9 (4.2–7.1) 0.929 13.4 (12.0–15.3) 12.8 (10.2–14.7) 0.383 12.6 (9.8–14.3) 13.5 (8.3–15.5) 0.802 Percent DHA, µg/100 µg FA 2.0 (1.8–2.5) 1.8 (1.4–2.5) 0.336 3.1 (2.8–3.5) 3.1 (2.1–3.9) 0.913 2.7 (2.3–3.2) 3.4 (3.0–3.8) 0.022 Data presented are median and first and third quartiles. Mann-Whitney U test was used to test statistical difference between two groups. P values in bold are significant. Abbreviations: EPA, eicosapentaenoic acid; NA, not available; ND, not detected. View Large Table 3. FA Profile in Maternal Vein Serum, Umbilical Vein, and Umbilical Artery Serum Maternal Serum Umbilical Vein Umbilical Artery FAs, µg/mL T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Myristic acid (C14:0) 41.8 (16.2–70.0) 8.8 (0.7–93.6) 0.143 8.8 (0.6–14.1) 2.2 (0.1–12.4) 0.069 12.5 (8.4–17.2) 4.2 (0.4–16.6) 0.058 Pentadecanoic acid (C15:0) 7.4 (1.4–13.7) 1.9 (0.0–14.5) 0.255 4.4 (0.4–8.1) 0.5 (0.0–4.9) 0.033 3.9 (0.4–7.5) 1.0 (0.0–6.8) 0.387 Palmitic acid (C16:0) 1481.5 (704.1–2490.1) 631.1 (190.2–1806.4) 0.051 321.3 (281.4–438.1) 119.1 (46.1–333.1) 0.001 313.0 (278.9–428.1) 316.7 (190.7–360.) 0.156 Stearic acid (C18:0) 260.3 (96.8–348.5) 62.7 (27.5–404.9) 0.084 152.8 (75.0–234.4) 36.7 (9.7–165.9) 0.001 155.1 (127.6–185.8) 106.2 (61.1–169.8) 0.098 Palmitoleic acid (C16:1n-7) 48.9 (23.3–66.0) 35.3 (7.0–114.9) 0.261 36.0 (25.4–53.1) 8.1 (2.2–40.7) 0.001 40.7 (34.2–55.9) 28.6 (19.9–44.8) 0.003 Oleic acid (C18:1n-9) 909.1 (396.9–1217.9) 260.5 (90.2–778.8) 0.012 240.0 (89.5–298.2) 65.9 (18.7–226.1) <0.001 237.2 (160.0–296.7) 122.7 (57.9–200.0) <0.001 Vaccenic acid (C18:1n-7) 6.7 (1.6–25.9) 0.0 (0.0–20.5 0.072 41.2 (0.0–54.6) 0.0 (0.0–39.2) 0.017 40.7 (0.0–51.2) 12.8 (0.0–75.0) 0.946 α-Linoleic acid (C18:3n-3) 10.2 (4.7–14.9) 4.4 (1.3–14.6) 0.133 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND Eicosapentanoic acid (C20:5n-3) 4.0 (0.2–8.0) 0.3 (0.0–8.8) 0.247 0.2 (0.0–1.1) 0.0 (0.0–0.0) NA 0.0 (0.0–0.5) 0.0 (0.0–0.0) NA DHA (C22:6n-3) 110.2 (55.8–143.4) 70.7 (21.1–117.6) 0.019 45.9 (15.7–78.1) 19.9 (3.3–44.5) 0.003 43.5 (34.8–56.2) 36.2 (24.9–58.4) 0.391 Linoleic acid (C18:3n-6) 12.9 (0.7–18.7) 0.8 (0.0–6.0) 0.006 6.6 (0.6–20.5) 0.3 (0.0–7.6) <0.001 7.2 (0.5–27.5) 1.1 (0.1–7.8) 0.009 AA (C20:4n-6) 259.4 (148.1–408.2) 163.5 (48.5–317.3) 0.051 202.9 (142.4–305.8) 59.2 (16.4–176.1) <0.001 193.1 (149.5–225.4) 124.4 (89.5–169.5) 0.016 AA:(DHA + EPA) 2.8 (2.5–3.1) 2.4 (1.5–3.5) 0.383 4.2 (3.7–4.6) 4.0 (3.5–4.8) 0.524 4.1 (3.4–4.9) 3.7 (2.9–4.1) 0.107 Percent AA, µg/100 µg FA 5.8 (5.2–6.7) 5.9 (4.2–7.1) 0.929 13.4 (12.0–15.3) 12.8 (10.2–14.7) 0.383 12.6 (9.8–14.3) 13.5 (8.3–15.5) 0.802 Percent DHA, µg/100 µg FA 2.0 (1.8–2.5) 1.8 (1.4–2.5) 0.336 3.1 (2.8–3.5) 3.1 (2.1–3.9) 0.913 2.7 (2.3–3.2) 3.4 (3.0–3.8) 0.022 Maternal Serum Umbilical Vein Umbilical Artery FAs, µg/mL T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P T1DM (n = 30) Control (n = 30) P Myristic acid (C14:0) 41.8 (16.2–70.0) 8.8 (0.7–93.6) 0.143 8.8 (0.6–14.1) 2.2 (0.1–12.4) 0.069 12.5 (8.4–17.2) 4.2 (0.4–16.6) 0.058 Pentadecanoic acid (C15:0) 7.4 (1.4–13.7) 1.9 (0.0–14.5) 0.255 4.4 (0.4–8.1) 0.5 (0.0–4.9) 0.033 3.9 (0.4–7.5) 1.0 (0.0–6.8) 0.387 Palmitic acid (C16:0) 1481.5 (704.1–2490.1) 631.1 (190.2–1806.4) 0.051 321.3 (281.4–438.1) 119.1 (46.1–333.1) 0.001 313.0 (278.9–428.1) 316.7 (190.7–360.) 0.156 Stearic acid (C18:0) 260.3 (96.8–348.5) 62.7 (27.5–404.9) 0.084 152.8 (75.0–234.4) 36.7 (9.7–165.9) 0.001 155.1 (127.6–185.8) 106.2 (61.1–169.8) 0.098 Palmitoleic acid (C16:1n-7) 48.9 (23.3–66.0) 35.3 (7.0–114.9) 0.261 36.0 (25.4–53.1) 8.1 (2.2–40.7) 0.001 40.7 (34.2–55.9) 28.6 (19.9–44.8) 0.003 Oleic acid (C18:1n-9) 909.1 (396.9–1217.9) 260.5 (90.2–778.8) 0.012 240.0 (89.5–298.2) 65.9 (18.7–226.1) <0.001 237.2 (160.0–296.7) 122.7 (57.9–200.0) <0.001 Vaccenic acid (C18:1n-7) 6.7 (1.6–25.9) 0.0 (0.0–20.5 0.072 41.2 (0.0–54.6) 0.0 (0.0–39.2) 0.017 40.7 (0.0–51.2) 12.8 (0.0–75.0) 0.946 α-Linoleic acid (C18:3n-3) 10.2 (4.7–14.9) 4.4 (1.3–14.6) 0.133 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND 0.0 (0.0–0.0) 0.0 (0.0–0.0) ND Eicosapentanoic acid (C20:5n-3) 4.0 (0.2–8.0) 0.3 (0.0–8.8) 0.247 0.2 (0.0–1.1) 0.0 (0.0–0.0) NA 0.0 (0.0–0.5) 0.0 (0.0–0.0) NA DHA (C22:6n-3) 110.2 (55.8–143.4) 70.7 (21.1–117.6) 0.019 45.9 (15.7–78.1) 19.9 (3.3–44.5) 0.003 43.5 (34.8–56.2) 36.2 (24.9–58.4) 0.391 Linoleic acid (C18:3n-6) 12.9 (0.7–18.7) 0.8 (0.0–6.0) 0.006 6.6 (0.6–20.5) 0.3 (0.0–7.6) <0.001 7.2 (0.5–27.5) 1.1 (0.1–7.8) 0.009 AA (C20:4n-6) 259.4 (148.1–408.2) 163.5 (48.5–317.3) 0.051 202.9 (142.4–305.8) 59.2 (16.4–176.1) <0.001 193.1 (149.5–225.4) 124.4 (89.5–169.5) 0.016 AA:(DHA + EPA) 2.8 (2.5–3.1) 2.4 (1.5–3.5) 0.383 4.2 (3.7–4.6) 4.0 (3.5–4.8) 0.524 4.1 (3.4–4.9) 3.7 (2.9–4.1) 0.107 Percent AA, µg/100 µg FA 5.8 (5.2–6.7) 5.9 (4.2–7.1) 0.929 13.4 (12.0–15.3) 12.8 (10.2–14.7) 0.383 12.6 (9.8–14.3) 13.5 (8.3–15.5) 0.802 Percent DHA, µg/100 µg FA 2.0 (1.8–2.5) 1.8 (1.4–2.5) 0.336 3.1 (2.8–3.5) 3.1 (2.1–3.9) 0.913 2.7 (2.3–3.2) 3.4 (3.0–3.8) 0.022 Data presented are median and first and third quartiles. Mann-Whitney U test was used to test statistical difference between two groups. P values in bold are significant. Abbreviations: EPA, eicosapentaenoic acid; NA, not available; ND, not detected. View Large In the umbilical vein serum, the concentrations of all particular FAs, except myristic acid (C14:0), were significantly higher in the T1DM group compared with the control group. In the umbilical artery serum, differences between T1DM and controls were smaller, and only concentrations for palmitoleic (C16:1n-7), oleic (C18:1n-9), linoleic acid (C18:3n-6), and AA (C20:4n-6) were higher in the T1DM group compared with the control group. The FA concentrations not higher in T1DM were palmitic acid (C16:0) and stearic acid (C18:0) in umbilical artery serum, and in addition, the percentage DHA was significantly lower in T1DM compared with controls (2.7% vs 3.4%, P = 0.02). Correlation between maternal serum and umbilical vein serum of total FAs and among neonatal ponderal index, glucose, and insulin resistance (HOMA2) Maternal and umbilical vein total FAs were correlated with Pearson correlation coefficient (r) (r = 0.685; P < 0.001; Fig. 1). Umbilical artery and umbilical vein total FAs were strongly correlated (rs = 0.883; P < 0.001). Figure 1. View largeDownload slide Linear regression between total FAs in maternal and umbilical vein serum (r = 0.685; P < 0.001). Figure 1. View largeDownload slide Linear regression between total FAs in maternal and umbilical vein serum (r = 0.685; P < 0.001). Maternal glucose concentrations were strongly correlated with glucose in umbilical vein (rs = 0.763; P < 0.001; (Table 4). Table 4. Significant Spearman Correlations Between Umbilical Vein and Umbilical Artery Serum Total FAs and Among Neonatal Ponderal Index, Glucose, C-Peptide, Leptin, HbA1c, and Insulin Resistance (HOMA2) Correlations rs P Glucose concentration between maternal and umbilical vein 0.763 <0.001 Ponderal index and insulin resistance (HOMA2) 0.477 0.012 C-Peptide and total FAs in umbilical vein 0.328 0.011 Leptin and ponderal index 0.263 0.043 Total FAs between umbilical artery and umbilical vein 0.883 <0.001 AA and DHA in umbilical vein 0.513 <0.001 C-Peptide and total FAs in umbilical artery 0.379 0.003 Insulin resistance (HOMA2) and total FAs in umbilical artery 0.456 0.017 Leptin and C-peptide in umbilical vein serum 0.526 0.005 Correlations rs P Glucose concentration between maternal and umbilical vein 0.763 <0.001 Ponderal index and insulin resistance (HOMA2) 0.477 0.012 C-Peptide and total FAs in umbilical vein 0.328 0.011 Leptin and ponderal index 0.263 0.043 Total FAs between umbilical artery and umbilical vein 0.883 <0.001 AA and DHA in umbilical vein 0.513 <0.001 C-Peptide and total FAs in umbilical artery 0.379 0.003 Insulin resistance (HOMA2) and total FAs in umbilical artery 0.456 0.017 Leptin and C-peptide in umbilical vein serum 0.526 0.005 All P values are significant. View Large Table 4. Significant Spearman Correlations Between Umbilical Vein and Umbilical Artery Serum Total FAs and Among Neonatal Ponderal Index, Glucose, C-Peptide, Leptin, HbA1c, and Insulin Resistance (HOMA2) Correlations rs P Glucose concentration between maternal and umbilical vein 0.763 <0.001 Ponderal index and insulin resistance (HOMA2) 0.477 0.012 C-Peptide and total FAs in umbilical vein 0.328 0.011 Leptin and ponderal index 0.263 0.043 Total FAs between umbilical artery and umbilical vein 0.883 <0.001 AA and DHA in umbilical vein 0.513 <0.001 C-Peptide and total FAs in umbilical artery 0.379 0.003 Insulin resistance (HOMA2) and total FAs in umbilical artery 0.456 0.017 Leptin and C-peptide in umbilical vein serum 0.526 0.005 Correlations rs P Glucose concentration between maternal and umbilical vein 0.763 <0.001 Ponderal index and insulin resistance (HOMA2) 0.477 0.012 C-Peptide and total FAs in umbilical vein 0.328 0.011 Leptin and ponderal index 0.263 0.043 Total FAs between umbilical artery and umbilical vein 0.883 <0.001 AA and DHA in umbilical vein 0.513 <0.001 C-Peptide and total FAs in umbilical artery 0.379 0.003 Insulin resistance (HOMA2) and total FAs in umbilical artery 0.456 0.017 Leptin and C-peptide in umbilical vein serum 0.526 0.005 All P values are significant. View Large Longer-term, higher maternal glucose levels, reflected in higher HbA1c (percentage) levels, were correlated with neonatal ponderal index (rs = 0.507; P = 0.003). Ponderal index was positively correlated with both leptin (rs = 0.263; P = 0.043) and insulin resistance (HOMA2; rs = 0.526; P = 0.005) in umbilical cord blood. Both C-peptide (rs = 0.379; P = 0.003) and insulin resistance (HOMA2; rs = 0.456; P = 0.02) in umbilical cord blood were correlated to total FAs in umbilical artery serum. Discussion The current study compared FA concentrations and the FA profile in maternal and umbilical venous and arterial serum between well-controlled T1DM and control mothers. Adiposity, which is a confounder of lipid and FAs levels, was comparable in the mothers, as there was no between-group difference in BMI and gestational weight gain. Similar to other studies (17), also, leptin levels did not differ between the two study groups. The most important finding was that well-controlled T1DM affects the FA levels, mostly in the neonates, in which the levels of total FAs, as well as most of the SFAs, MUFAs and PUFAs, were elevated compared with healthy controls. Importantly, also, the cord blood concentrations of the essential FAs, and in particular of DHA, were higher in the T1DM neonates than in the controls. Total FA concentrations were not significantly different between pregnant T1DM with good metabolic control and healthy, nondiabetic women. Large variation in the data may have precluded significance, as the concentrations of some FAs and FA classes were significantly elevated in maternal serum in the women with T1DM compared with control women. Consistent with the present results, also in GDM pregnancies, maternal lipids did not differ from control pregnancies (18). T1DM effects on neonatal anthropometrics and hormones The cord blood levels of C-peptide, leptin, and insulin resistance were also significantly higher in the T1DM group compared with the control group. T1DM, similar to GDM and T2DM, is a known factor that predisposes women to have hypertrophic newborns through an enhanced fetal glucose-insulin axis. This eventually leads to more fetal adipose tissue, as reflected by higher leptin levels (19, 20). Although the placenta also produces leptin, almost all leptin is released into the maternal circulation (21), thus making its contribution to the circulating leptin pool in the fetus small. Insulin and leptin in the cord blood are correlated, especially in neonates in the highest birthweight category (22). Leptin increases with the amount of adipose tissue and regulates fetal weight. Hence, children born to diabetic mothers have higher cord blood leptin levels than those born to nondiabetic mothers. Our leptin results are consistent with previous studies (23). Although we have not directly measured neonatal fat mass or relative body fat, the positive correlation (rs = 0.263; P = 0.043) between leptin and ponderal index may suggest an increase in neonatal fat in the T1DM pregnancies. The higher insulin resistance HOMA2 (P = 0.015) in neonates born to mothers with T1DM compared with control mothers is an important finding. Insulin resistance (HOMA2) correlated with the neonatal ponderal index. The higher insulin resistance in T1DM neonates is likely the result of the higher maternal glucose levels, reflected by higher HbA1c values in the third trimester than in the controls, but effects of early hyperglycemia cannot be ruled out (24). Interestingly, despite neonatal hyperinsulinemia in the T1DM neonates, their birthweight was not different from control neonates. However, the shorter stature in T1DM neonates, which is another important and unexpected finding, may be associated with reduced lean mass. Thus, T1DM appears to affect the lean and fat compartments disproportionally, although this must be verified by more direct measurements of body composition. A shorter length of neonates born to mothers of diabetic pregnancies was already found in GDM (25). In that study, the interpretation was offered that fetal FAs, through the transcription factor peroxisome proliferator-activated receptor-γ, may drive mesenchymal stem cell differentiation to a more adipogenic than osteoblastic phenotype. Whereas this is speculative, the elevated cord blood levels of FAs, found here in T1DM, may support this notion, which certainly warrants further studies. T1DM effect on neonatal FAs In well-controlled T1DM, maternal FA levels were higher than in controls, although this did not reach statistical significance for each species. This was accompanied by higher concentrations of total FAs, SFAs, MUFAs, and PUFAs in the neonate. Steady-state levels are determined by the following: transplacental transfer, de novo synthesis out of glucose, uptake into tissue, and lipolysis. The study design does not allow for distinguishing among these possibilities. Our study confirmed earlier findings (4, 26) of correlations in total FA concentration between maternal and umbilical vein serum and extends this to mothers with T1DM. Essential FAs are important for normal fetal growth and development (27). Transplacental transport of DHA, eicosapentaenoic acid (EPA), and AA occurs in several steps, from cell transmembrane transport through intracellular transport modulated by protein carriers to further passage through the cell membrane toward the fetus (8, 28). In our previous study, we demonstrated that there was no difference in LC-PUFA n-3 percentage between T1DM and control groups of pregnant women and their fetuses (12), which is in line with the results presented here. Our results differ from an earlier study that found lower percentage of AA and DHA in neonates born to mothers with T1DM (29). Here, the levels of AA and DHA were higher in women with T1DM and significantly higher in their fetuses compared with their control counterparts (Table 3). The major difference between the two studies is that we measured FA percentage in serum, whereas the lower AA and DHA percentage was found in neonatal phospholipids. This would suggest that T1DM modifies the incorporation of AA and DHA into phospholipids. Whether DHA concentrations in cord blood serum or in the phospholipid fraction are more relevant for neonatal development remains to be established, but it is pertinent that the brain DHA transporter major facilitator superfamily domain-containing protein 2 transfers DHA in phospholipids across the blood brain barrier (30). The placenta is important for selective AA and DHA canalization from the mother to the fetus; the evidence is a high coefficient of correlation between AA and DHA in the maternal and the umbilical vein blood (Fig. 2). The median maternal percent of AA (5.8%) and DHA (2.0%) in serum of mothers with T1DM was similar to the control group (AA, 5.9%; DHA, 1.8%). AA serves as a precursor of proinflammatory eicosanoids, whereas DHA is a precursor of anti-inflammatory resolvins and protectins (31); this could suggest a proinflammatory environment in T1DM neonates that is perhaps associated with hyperleptinemia. In a recent randomized controlled trial, we found that DHA supplementation stimulated the production of endogenous insulin in women with T1DM and was accompanied by lower C-peptide levels in the newborn (32). In our study, neonates born by mothers with T1DM had an elevated AA/DHA + EPA ratio (median 4.1:1). This imbalance may be an indication of their increased risk to develop obesity later in life (33). It may well be that a beneficial effect of DHA supplementation may lie in improving the neonatal glucose-insulin axis and perhaps indirectly also, in DHA uptake into the brain across the blood brain barrier. This may also indirectly decrease the risk for childhood obesity. Figure 2. View largeDownload slide Linear regression between DHA and AA in umbilical vein serum (r = 0.924; P < 0.001). Figure 2. View largeDownload slide Linear regression between DHA and AA in umbilical vein serum (r = 0.924; P < 0.001). Strengths of the study This study investigates FA profiles and FA concentration in T1DM in arterial and venous cord blood. We have combined the FA measurements with those of key hormones known to influence lipid and FA levels, i.e., insulin and leptin. Cord blood was obtained with the placenta still in situ and the neonate not separated, reducing the influences of the postpartum period and allowing the results to represent the in vivo situation as closely as possible. As all women delivered by elective CS and were denied access to food, the levels of FAs and other parameters measured in umbilical serum were not influenced by mode of delivery or nutritional status of the women. The measurement of the arterial and venous cord blood is another strength. There has been only one study in GDM that also determined the FA profile in both arms of the umbilical circulation (34). The combination of these measurements with concentrations of hormonal regulators in the umbilical cord allows conclusions about potential mechanisms that determine the neonatal FA profile, although these must be confirmed in larger studies and other populations. We must acknowledge some limitations: The small sample size precluded significance of higher maternal FA levels in mothers with T1DM. The study included only well-controlled women with T1DM. The results may be different, especially for neonatal outcomes, in poorly controlled T1DM or in women without residual β-cell function, because adaptive and protective placental mechanisms may become exhausted under more extreme metabolic conditions (35). Future studies should measure DHA levels in serum, as well as in the phospholipids of newborn cord blood, and test which of these two fractions is more important for brain development and cognitive function in the children born to T1DM pregnancies. Conclusion In conclusion, the levels of leptin, C-peptide, insulin resistance, total FAs, SFAs, MUFAs, PUFAs, DHA, and AA were significantly higher in T1DM umbilical vein serum compared with those in the control group. Abbreviations: Abbreviations: AA arachidonic acid BMI body mass index CS Caesarean section DHA docosahexaenoic acid EPA eicosapentaenoic acid FA fatty acid GDM gestational diabetes HOMA homeostasis model assessment LC-PUFA long-chain polyunsaturated fatty acid MUFA monounsaturated fatty acid PUFA polyunsaturated fatty acid r Pearson correlation coefficient rs Spearman correlation coefficient SFA saturated fatty acid T1/2DM type 1/2 diabetes mellitus Acknowledgments Financial Support: The study was part of the scientific project approved by the Ministry of Science, Education and Technology of the Republic of Croatia, entitled Metabolic and Endocrine Changes in Pregnant Patients with Diabetes (No. 108-1080401-0386). Author Contributions: J.D. designed the study, developed the statistical analysis plan, wrote the manuscript, is the guarantor of this work, and takes responsibility for the integrity of the data and the accuracy of the data analysis. M.I., S.O., and D.S. collected data. E.B. extracted lipids and determined the fatty acid profile by gas chromatography-mass spectrometry. G.D. and M.v.P. reviewed and edited the manuscript and contributed to the discussion. Disclosure Summary: The authors have nothing to disclose. References 1. Atkinson MA , Eisenbarth GS . Type 1 diabetes: new perspectives on disease pathogenesis and treatment . Lancet . 2001 ; 358 ( 9277 ): 221 – 229 . 2. Herrera E . Metabolic changes in diabetic pregnancy. In: Djelmis J , Desoye G , Ivanisevic M , eds. Diabetology of Pregnancy . Basel : Karger ; 2005 : 34 – 45 . 3. Campbell FM , Taffesse S , Gordon MJ . Dutta-Roy AK . Plasma membrane fatty-acid-binding protein in human placenta: identification and characterization . Biochem Biophys Res Commun . 1995 ; 209 ( 3 ): 1011 – 1017 . 4. Hendrickse W , Stammers JP , Hull D . The transfer of free fatty acids across the human placenta . Br J Obstet Gynaecol . 1985 ; 92 ( 9 ): 945 – 952 . 5. Desoye G , Gauster M , Wadsack C . Placental transport in pregnancy pathologies . Am J Clin Nutr . 2011 ; 94 ( Suppl 6 ): 1896S – 1902S . 6. Herrera E , Desoye G . Maternal and fetal lipid metabolism under normal and gestational diabetic conditions . Horm Mol Biol Clin Investig . 2016 ; 26 ( 2 ): 109 – 127 . 7. Catalano PM , Thomas A , Huston-Presley L , Amini SB . Increased fetal adiposity: a very sensitive marker of abnormal in utero development . Am J Obstet Gynecol . 2003 ; 189 ( 6 ): 1698 – 1704 . 8. Hanebutt FL , Demmelmair H , Schiessl B , Larqué E , Koletzko B . Long-chain polyunsaturated fatty acid (LC-PUFA) transfer across the placenta . Clin Nutr . 2008 ; 27 ( 5 ): 685 – 693 . 9. Neuringer M , Reisbick S , Janowsky J . The role of n-3 fatty acids in visual and cognitive development: current evidence and methods of assessment . J Pediatr . 1994 ; 125 ( 5 ): S39 – S47 . 10. Zornoza-Moreno M , Fuentes-Hernández S , Carrión V , Alcántara-López MV , Madrid JA , López-Soler C , Sánchez-Solís M , Larqué E . Is low docosahexaenoic acid associated with disturbed rhythms and neurodevelopment in offsprings of diabetic mothers ? Eur J Clin Nutr . 2014 ; 68 ( 8 ): 931 – 937 . 11. Larqué E , Demmelmair H , Berger B , Hasbargen U , Koletzko B . In vivo investigation of the placental transfer of (13)C-labeled fatty acids in humans . J Lipid Res . 2003 ; 44 ( 1 ): 49 – 55 . 12. Berberovic E , Ivanisevic M , Juras J , Horvaticek M , Delas I , Djelmis J . Arachidonic and docosahexaenoic acid in the blood of a mother and umbilical vein in diabetic pregnant women . J Matern Fetal Neonatal Med . 2013 ; 26 ( 13 ): 1287 – 1291 . 13. Kolcić I , Polašek O , Pfeifer D , Smolej-Narancić N , Ilijić M , Bljajić D , Biloglav Z , Ivanisević M , Delmis J . Birth weight of healthy newborns in Zagreb area, Croatia . Coll Antropol . 2005 ; 29 ( 1 ): 257 – 262 . 14. International Association of Diabetes and Pregnancy Study Groups Consensus Panel , Metzger BE , Gabbe SG , Persson B , Buchanan TA , Catalano PA , Damm P , Dyer AR , Leiva A , Hod M , Kitzmiler JL , Lowe LP , McIntyre HD , Oats JJ , Omori Y , Schmidt MI . International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy . Diabetes Care . 2010 ; 33 ( 3 ): 676 – 682 . 15. Folch J , Lees M , Sloane Stanley GH . A simple method for the isolation and purification of total lipides from animal tissues . J Biol Chem . 1957 ; 226 ( 1 ): 497 – 509 . 16. Levy JC , Matthews DR , Hermans MP . Correct homeostasis model assessment (HOMA) evaluation uses the computer program . Diabetes Care . 1998 ; 21 ( 12 ): 2191 – 2192 . 17. Stock SM , Bremme KA . Elevation of plasma leptin levels during pregnancy in normal and diabetic women . Metabolism . 1998 ; 47 ( 7 ): 840 – 843 . 18. Schaefer-Graf UM , Meitzner K , Ortega-Senovilla H , Graf K , Vetter K , Abou-Dakn M , Herrera E . Differences in the implications of maternal lipids on fetal metabolism and growth between gestational diabetes mellitus and control pregnancies . Diabet Med . 2011 ; 28 ( 9 ): 1053 – 1059 . 19. Nielsen LR , Rehfeld JF , Pedersen-Bjergaard U , Damm P , Mathiesen ER . Pregnancy-induced rise in serum C-peptide concentrations in women with type 1 diabetes . Diabetes Care . 2009 ; 32 ( 6 ): 1052 – 1057 . 20. Maffei M , Volpe L , Di Cianni G , Bertacca A , Ferdeghini M , Murru S , Teti G , Casadidio I , Cecchetti P , Navalesi R , Benzi L . Plasma leptin levels in newborns from normal and diabetic mothers . Horm Metab Res . 1998 ; 30 ( 9 ): 575 – 580 . 21. Linnemann K , Malek A , Sager R , Blum WF , Schneider H , Fusch C . Leptin production and release in the dually in vitro perfused human placenta . J Clin Endocrinol Metab . 2000 ; 85 ( 11 ): 4298 – 4301 . 22. Wolf HJ , Ebenbichler CF , Huter O , Bodner J , Lechleitner M , Föger B , Patsch JR , Desoye G . Fetal leptin and insulin levels only correlate inlarge-for-gestational age infants . Eur J Endocrinol . 2000 ; 142 ( 6 ): 623 – 629 . 23. Persson B , Westgren M , Celsi G , Nord E , Ortqvist E . Leptin concentrations in cord blood in normal newborn infants and offspring of diabetic mothers . Horm Metab Res . 1999 ; 31 ( 8 ): 467 – 471 . 24. Desoye G , Nolan CJ . The fetal glucose steal: an underappreciated phenomenon in diabetic pregnancy . Diabetologia . 2016 ; 59 ( 6 ): 1089 – 1094 . 25. Lampl M , Jeanty P . Exposure to maternal diabetes is associated with altered fetal growth patterns: A hypothesis regarding metabolic allocation to growth under hyperglycemic-hypoxemic conditions . Am J Hum Biol . 2004 ; 16 ( 3 ): 237 – 263 . 26. Wolfe MD , Chuang LT , Rayburn WF , Wen PC , VanderJagt DJ , Glew RH . Low fatty acid concentrations in neonatal cord serum correlate with maternal serum . J Matern Fetal Neonatal Med . 2012 ; 25 ( 8 ): 1292 – 1296 . 27. Innis SM . Essential fatty acids in growth and development . Prog Lipid Res . 1991 ; 30 ( 1 ): 39 – 103 . 28. Campbell FM , Dutta-Roy AK . Plasma membrane fatty acid-binding protein (FABPpm) is exclusively located in the maternal facing membranes of the human placenta . FEBS Lett . 1995 ; 375 ( 3 ): 227 – 230 . 29. Ghebremeskel K , Thomas B , Lowy C , Min Y , Crawford MA . Type 1 diabetes compromises plasma arachidonic and docosahexaenoic acids in newborn babies . Lipids . 2004 ; 39 ( 4 ): 335 – 342 . 30. Nguyen LN , Ma D , Shui G , Wong P , Cazenave-Gassiot A , Zhang X , Wenk MR , Goh EL , Silver DL . Mfsd2a is a transporter for the essential omega-3 fatty acid docosahexaenoic acid . Nature . 2014 ; 509 ( 7501 ): 503 – 506 . 31. Serhan CN , Chiang N , Dalli J , Levy BD . Lipid mediators in the resolution of inflammation . Cold Spring Harb Perspect Biol . 2014 ; 7 ( 2 ): a016311 . 32. Horvaticek M , Djelmis J , Ivanisevic M , Oreskovic S , Herman M . Effect of eicosapentaenoic acid and docosahexaenoic acid supplementation on C-peptide preservation in pregnant women with type-1 diabetes: randomized placebo controlled clinical trial . Eur J Clin Nutr . 2017 ; 71 ( 8 ): 968 – 972 . 33. Simopoulos AP . An increase in the omega-6/omega-3 fatty acid ratio increases the risk for obesity . Nutrients . 2016 ; 8 ( 3 ): 128 – 145 . 34. Ortega-Senovilla H , Alvino G , Taricco E , Cetin I , Herrera E . Gestational diabetes mellitus upsets the proportion of fatty acids in umbilical arterial but not venous plasma . Diabetes Care . 2009 ; 32 ( 1 ): 120 – 122 . 35. Gauster M , Hiden U , van Poppel M , Frank S , Wadsack C , Hauguel-de Mouzon S , Desoye G . Dysregulation of placental endothelial lipase in obese women with gestational diabetes mellitus . Diabetes . 2011 ; 60 ( 10 ): 2457 – 2464 . Copyright © 2018 Endocrine Society

Saito, Hiroshi;Noda, Hiroshi;Gatault, Philippe;Bockenhauer, Detlef;Loke, Kah Yin;Hiort, Olaf;Silve, Caroline;Sharwood, Erin;Martin, Regina Matsunaga;Dillon, Michael J;Gillis, David;Harris, Mark;Rao, Sudhaker D;Pauli, Richard M;Gardella, Thomas J;Jüppner, Harald

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00332pmid: 29788189

Abstract Context Five different activating PTH/PTH-related peptide (PTHrP) receptor (PTHR1) mutations have been reported as causes of Jansen metaphyseal chondrodysplasia (JMC), a rare disorder characterized by severe growth plate abnormalities and PTH-independent hypercalcemia. Objectives Assess the natural history of clinical and laboratory findings in 24 patients with JMC and characterize the disease-causing mutant receptors in vitro. Patients and Methods The H223R mutation occurred in 18 patients. T410P, I458R and I458K each occurred in single cases; T410R was present in a father and his two sons. Laboratory records were analyzed individually and in aggregate. Results Postnatal calcium levels were normal in most patients, but elevated between 0.15 and 10 years (11.8 ± 1.37 mg/dL) and tended to normalize in adults (10.0 ± 1.03 mg/dL). Mean phosphate levels were at the lower end of the age-specific normal ranges. Urinary calcium/creatinine (mg/mg) were consistently elevated (children, 0.80 ± 0.40; adults, 0.28 ± 0.19). Adult heights were well below the 3rd percentile for all patients, except for those with the T410R mutation. Most patients with JMC had undergone orthopedic surgical procedures, most had nephrocalcinosis, and two had advanced chronic kidney disease. The five PTHR1 mutants showed varying degrees of constitutive and PTH-stimulated cAMP signaling activity when expressed in HEK293 reporter cells. The inverse agonist [L11,dW12,W23,Y36]PTHrP(7–36) reduced basal cAMP signaling for each PTHR1 mutant. Conclusions Except for T410R, the other PTHR1 mutations were associated with indistinguishable mineral ion abnormalities and cause similarly severe growth impairment. Hypercalciuria persisted into adulthood. An inverse agonist ligand effectively reduced in vitro PTH-independent cAMP formation at all five PTHR1 mutants, suggesting a potential path toward therapy. The PTH/PTH-related peptide (PTHrP) receptor (PTHR1) mediates the actions of two peptides, PTH and PTHrP, which stimulate at least two signaling pathways, cAMP/PKA and Ca2+/IP3/PKC. The PTHR1, a class B G protein–coupled receptor, is abundantly expressed in kidney and bone, as well as in the metaphyseal growth plates (1). In growth plate chondrocytes, activation of the PTHR1 by PTHrP slows the differentiation of chondrocytes, thus contributing importantly to normal bone growth and elongation (2). In bone, activation of the PTHR1 by PTH directly affects osteoblast and osteocyte activity, and indirectly affects, through the RANK/RANKL system, osteoclast maturation and activity. In distal renal tubules, the PTHR1 mediates the PTH-dependent reabsorption of calcium, whereas in the proximal tubules it enhances excretion of phosphate and the expression of 1α-hydroxylase (3). Jansen metaphyseal chondrodysplasia (JMC) is a rare autosomal-dominant disease caused by heterozygous, activating PTHR1 mutations (4–6). Thus far, five different PTHR1 mutations affecting one of three different amino acid residues have been identified in patients with JMC; these mutations, H223R, T410P/R, and I458K/R, are each located at the intracellular end of transmembrane helices, namely 2, 6, and 7, respectively (7). The constitutive activity of the PTHR1 mutants slows chondrocyte maturation, leading to marked growth plate abnormalities that resemble severe rachitic changes (8, 9). In addition to short stature and bowing of long bones, patients with JMC often exhibit micrognathia, hypertelorism, high-arched palate, delayed tooth eruption or impaction, and premature closure of cranial sutures. However, this information is based on anecdotal reports, as a comprehensive natural history profile of JMC has yet to be established (7, 10–16). Prominent laboratory abnormalities reported for patients with JMC include severe PTH- and PTHrP-independent hypercalcemia and hypophosphatemia that are associated with high rates of bone turnover, cortical thinning, and excessive hypomineralized osteoid (14). Severe metaphyseal changes associated with life-long hypercalcemia were thought to be the hallmarks of JMC (7, 11, 13). However, recent reports revealed that some patients, diagnosed radiographically and genetically with JMC, did not show overt hypercalcemia or hypophosphatemia (13, 17). It is thus currently uncertain as to the extent that radiographic, height, and biochemical abnormalities in JMC can vary due, for example, to patient age and/or type of PTHR1 mutation. Additionally, even in the absence of obvious hypercalcemia, urinary calcium excretion may be elevated. Patients affected by JMC can thus be at risk for developing nephrocalcinosis and possibly impaired renal function. The purpose of the current study was, therefore, to assess the natural history and long-term outcome of multiple patients with documented, disease-causing PTHR1 mutations. We report blood and urinary calcium levels in newborns, children, and adults affected by JMC; adult heights, need for surgical intervention, and other biochemical abnormalities and renal function are also assessed. Additionally, we characterize the different JMC-causing PTHR1 variants in cell-based functional assays and investigate in vitro their response to a PTH agonist and a PTHrP-based inverse agonist ligand. Subjects and Methods Patients and data collection Clinical and laboratory information of previously reported patients was obtained from earlier publications (5, 6, 10–19). No additional patients with a confirmed molecular defect were identified by searching PubMed (Public/Publisher MEDLINE, National Center for Biotechnology Information, Bethesda, MD) electronic database on 27 September 2017 using the query “Jansen type metaphyseal chondrodysplasia” (MeSH Terms) OR “Jansen metaphyseal chondrodysplasia” (All Fields). Whenever possible, follow-up data were obtained from the primary care physician or specialist involved in the care of the patient. Additionally, we collected clinical and laboratory information for five patients not previously reported, for whom a disease-causing genetic PTHR1 mutation was identified. Laboratory data are listed according to four age groups; birth until the age of 1.5 months, 0.15 to 10 years, 17 to 38 years, and >49 years. Furthermore, we were able to obtain the final adult height for a subset of 13 patients, as well as information on renal function and calcifications, major skeletal abnormalities, use of bisphosphonates, and surgical interventions. The z scores for height in children and adults were calculated based on the data from the World Health Organization Child Growth Standard, National Health and Nutrition Survey, and from Centers for Disease Control and Prevention/National Center for Health Statistics. Case reports As examples of the natural course of laboratory abnormalities in Jansen disease, findings are presented for three previously unreported patients, H223R-15, H223R-16, and H223R-17. Laboratory findings as well as major radiographic and physical abnormalities are also provided for two other unreported patients, H223R-9 and H223R-18 (Supplemental Table 1). Patients H223R-4, H223R-13, H223R-14, T410R-2, and T410R-3 each inherited the PTHR1 mutation from an affected parent; all other patients with JMC have healthy parents, suggesting that their PTH1R mutation occurred de novo. Patient H223R-15 This 4-year-old boy, the first child of healthy parents, presented at birth with breathing difficulties due to micrognathia and bilateral choanal stenosis. He was noted to have hypertelorism, an elongated and high arched palate, downsloping palpebral fissures, large open fontanelles with widely spaced sagittal sutures, and palpable rachitic rosary. Investigations in the neonatal period showed serum calcium levels at the upper end of normal (9.64 to 11.4 mg/dL), with mildly decreased serum phosphate (1.62 mmol/L, normal range at this age, 1.8 to 3.0 mmol/L) and low serum PTH (12 pg/mL; normal range at this age, 20 to 95 pg/mL). During the subsequent months his serum calcium increased (see Fig. 1, green filled circles), with associated hypercalciuria, elevated serum alkaline phosphatase activity, elevated serum 1,25(OH)2 vitamin D levels (101 pg/mL; range, 63 to 136 pg/mL; normal range, 19 to 76 pg/mL), and progressive suppression of PTH concentration to <1 pg/mL. His skeletal survey showed markedly abnormal bones with typical JMC features; the H223R mutation was identified at 7 months of age. Serial renal ultrasound examinations, performed during infancy to investigate persistent hypertension, revealed nephrocalcinosis by 8 months of age. His hypertension resolved without treatment. Figure 1. View largeDownload slide Serum calcium concentrations of multiple patients with different PTHR1 mutations from the newborn period until the sixth decade of life; eight patients with measurements within the first 1.5 mo of life are indicated at the left of the axis break. Patients are depicted by open or closed symbols of different colors. The patients are identified in Supplemental Table 1. Patients with the H223R mutation are represented by open or closed circles; black-filled circles represent patients for whom only one measurement was available; colored open or closed circles represent patients for whom multiple measurements were available. Consecutive measurements for patient H223R-17 are depicted with red circles/line. Data for three patients with the T410R-PTHR1 mutation at different ages (father, black triangle; his two sons, blue and red triangles, respectively), and measurements for the patients with the T410P-PTHR1 mutation (diamonds), I458K-PTHR1 mutation (trapezoids), and I458R-PTHR1 mutation (pentagons) are shown. Dashed lines represent the upper/lower end of the adult normal range for total calcium levels (8.6 to 10.2 mg/dL). The reference range for infants is 8.4 to 10.6 mg/dL. Figure 1. View largeDownload slide Serum calcium concentrations of multiple patients with different PTHR1 mutations from the newborn period until the sixth decade of life; eight patients with measurements within the first 1.5 mo of life are indicated at the left of the axis break. Patients are depicted by open or closed symbols of different colors. The patients are identified in Supplemental Table 1. Patients with the H223R mutation are represented by open or closed circles; black-filled circles represent patients for whom only one measurement was available; colored open or closed circles represent patients for whom multiple measurements were available. Consecutive measurements for patient H223R-17 are depicted with red circles/line. Data for three patients with the T410R-PTHR1 mutation at different ages (father, black triangle; his two sons, blue and red triangles, respectively), and measurements for the patients with the T410P-PTHR1 mutation (diamonds), I458K-PTHR1 mutation (trapezoids), and I458R-PTHR1 mutation (pentagons) are shown. Dashed lines represent the upper/lower end of the adult normal range for total calcium levels (8.6 to 10.2 mg/dL). The reference range for infants is 8.4 to 10.6 mg/dL. Patient H223R-17 This 25-year-old female was recognized as having abnormal long bone radiographic features on the first day of life; hypercalcemia was noted on day 5. A diagnosis of JMC was made on the basis of clinical, radiographic, and biochemical findings at the age of 4 months. Medical interventions included a low-calcium and low-salt diet, as well as oral phosphate supplementation for much of her childhood. Her early growth was slow with lengths/heights below the 3rd percentile and further slowing was noted at 3 years of age. She had severe and recurrent alignment abnormalities of her legs (primarily varus deformity and anterior bowing of both the tibiae and femora); multiple osteotomies of both tibiae and both femora were performed between ages of 2.5 and 14 years (at 2.5, 5, 7, 10, and 14 years). Progressive kyphoscoliosis required posterior spinal fusion from T2 to L3 at age 11 years. Her maximal adult height is 116.9 cm. Most recent laboratory studies showed a total serum calcium level of 10.5 mg/dL (upper end of normal) with suppressed PTH (<4 pg/mL). Serum phosphate was at the lower end of the normal range (0.81 mmol/L) and the 1,25(OH)2 vitamin D level was 70.1 pg/mL, which is at the upper end of the normal range, although the 25 vitamin D level was only 13 ng/mL (i.e., well below the recommended level of 32 ng/mL). The serum creatinine was 0.39 mg/dL, which yields, based on the Schwartz equation (20, 21), a calculated glomerular filtration rate of 108.9 mL/min/1.73 m2. Time course of her serum levels from infancy until adulthood are shown in Fig. 1 (red open circles) and in Supplemental Fig. 1 along with urinary calcium/creatinine ratios; note that the serum calcium level was extremely elevated throughout childhood, but decreased to the upper end of the normal range during adulthood; nevertheless, hypercalciuria and an elevated urinary calcium/creatinine ratios persisted. Medullary nephrocalcinosis was documented in early childhood. Patient H223R-16 The 56-year-old male had reached a maximal adult height of 133 cm. At that age, his laboratory studies revealed a normal serum calcium level (9.4 mg/dL) with an elevated PTH (312 pg/mL) and a slightly elevated serum phosphate level (1.55 mmol/L), that is, laboratory findings not typically observed in Jansen disease. However, his serum creatinine was abnormal at 4.04 mg/dL and the estimated glomerular filtration rate was only 22 mL/min/1.73 m2, as calculated by the Schwartz formula. A progressive decline in renal function had been noted since his late 30s (Fig. 2A). The most recent serum alkaline phosphatase activity was above the upper end of normal (155 IU/L; reference range, 30 to 120 IU/L), the 1,25(OH)2 vitamin D level was at the lower end of normal (19.2 pg/mL), and the 25 vitamin D level was well below the recommended range (6.4 ng/mL). His most recent urinary calcium/creatinine ratio was 0.03, and his renal function was significantly impaired. Nephrocalcinosis had been known since early childhood, and current imaging by CT revealed marked bilateral renal calcifications with staghorn calculi (Fig. 2B). Figure 2. View largeDownload slide (A) Estimated glomerular filtration rates (eGFRs) as calculated by the Schwartz formula are presented for eight adult patients with three different PTHR1 mutations. For the patient with the T410P mutation (diamonds), three measurements are shown that were obtained during adulthood prior to hemodialysis that was initiated at age 37 y. For patient H223R-16 (filled circles) numerous measurements were performed after the age of 38 y showing the progressive decline in renal function. (B) Latest abdominal CT of patient H223R-16 at age 55 y showing extensive renal calcifications (arrows). Figure 2. View largeDownload slide (A) Estimated glomerular filtration rates (eGFRs) as calculated by the Schwartz formula are presented for eight adult patients with three different PTHR1 mutations. For the patient with the T410P mutation (diamonds), three measurements are shown that were obtained during adulthood prior to hemodialysis that was initiated at age 37 y. For patient H223R-16 (filled circles) numerous measurements were performed after the age of 38 y showing the progressive decline in renal function. (B) Latest abdominal CT of patient H223R-16 at age 55 y showing extensive renal calcifications (arrows). Cell culture and in vitro studies For characterization of wild-type and mutant PTH/PTHrP receptors, GS22A cells, an HEK293-derived cell line that stably expresses the luciferase-based pGlosensor-22F (Glosensor) cAMP reporter plasmid (22, 23), were cultured at 37°C in a humidified atmosphere containing 5% CO2 in DMEM (Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum. Cells were seeded in 96-well plates at a density of 2 × 104 cells per well. The following day, transfections were performed with varying amounts of each plasmid DNA (pcDNA3.1 empty vector, wild-type human PTHR1, or one of the five JMC mutants, i.e., H223R, I458K, I458R, T410P, or T410R) using FuGENE® HD transfection reagent (Promega, Madison, WI) according to the manufacturer’s instructions. For assessment of receptor expression an antibody was used that specifically recognizes the human PTHR1 (rabbit polyclonal anti-hPTHR1 antibody, PRB-640P (BioLegend; formerly Covance Antibody Products; Dedham, MA) and goat anti-rabbit IgG(H+L) antibody (HRP conjugate, product no. 31460, lot no. RJ242536, Invitrogen, Carlsbad, CA) was performed with ELISA. Basal level of cAMP accumulation and ligand effects on PTHR1-mediated cAMP signaling were assessed 48 hours after transfection via the Glosensor cAMP reporter (Promega). Confluent cells in 96-well plates were loaded with luciferin (0.5 mM) for 25 minutes at room temperature. Subsequently, varying concentrations of agonist peptides or vehicle were added and incubations were continued for an additional period of up to 90 minutes. Luminescence arising in response to intracellular cAMP binding to the Glosensor reporter enzyme was measured at 2-minute intervals during both the pretreatment and ligand-addition phases using a PerkinElmer Envision plate reader (PerkinElmer, Inc., Waltham, MA). The area under the curve (AUC) of the luminescence response during a 25-minute preligand phase (basal) and during a subsequent 90-minute ligand treatment phase was calculated to determine cAMP generation in cells expressing mutant or wild-type PTHR1 and to establish agonist dose-response curves. For the ligand treatment experiments, vehicle or PTH(1–34) at varying concentrations (from 1 × 10−7 to 1 × 10−11 M) were added to GS22A cells transfected with 100 ng of each plasmid DNA. Aggregate data of the AUC of the luminescence response are expressed as mean ± SEM of five experiments, each performed in duplicate. For the inverse agonist experiments, vehicle or [L11,dW12,W23,Y36]PTHrP(7–36) (1 × 10−6 M) was added to GS22A cells transfected with 100 ng of each plasmid DNA. The decrease in the ratio from the start point (time 0) of each luminescence response was calculated. Aggregate data are expressed as mean ± SEM of two experiments, each performed in quadruplicate. Data were processed using Excel for Mac (Microsoft, Redmond, WA) and Prism 7.0 (GraphPad Software, La Jolla, CA). Curves were fit to the data using a four-parameter, nonlinear regression function. Results The H223R mutation was identified in 18 patients with JMC [(5, 6, 10, 11, 13, 15, 16, 18, 19) and unpublished cases], whereas the T410P, I458R, and I458K mutations were each reported in a single case (6, 12, 14, 19); the T410R mutation was found in a father and his two sons (17). With the exception of H223R-4, H223R-13, H223R-14, T410R-2, and T410R-3, who inherited the mutant allele from an affected parent, each other patient with JMC was born to healthy parents; thus, most patients with JMC acquired the mutation de novo. Three patients with JMC have children (n = 5), all five of whom inherited the parental PTHR1 mutation; one affected female parent (H223R-12) has two affected sons (13), the other affected female parent (H223R-3) has an affected daughter (6), and the one affected male parent (T410R-1) has two affected sons (17). Most patients were diagnosed with JMC during childhood. However, the affected male patient T410R-1 was not diagnosed until the age of 33 years when his two affected sons, both with the same PTHR1 mutation, presented with typical radiographic findings; these patients exhibit less severe clinical and biochemical abnormalities than do most other patients with JMC (17). Similarly, one female patient (H223R-3) was not diagnosed until the age of 37 years, when her daughter was found to have the JMC mutation following evaluation for achondroplasia (6). Another female patient (H223R-12), a 38-year-old mother with two affected sons, had been noted to have severe short stature since early childhood and abnormal radiographic findings, but was not overtly hypercalcemic (13); thus, the JMC diagnosis was not considered until her two sons were confirmed to have the disease. Laboratory measurements were obtained for eight patients during the first 1.5 months of life because of respiratory difficulties and/or skeletal abnormalities (see Fig. 1 and Supplemental Table 1). When excluding patient H223R-17, who had a total calcium level of 13.7 mg/dL at the age of 5 days, most patients with JMC evaluated during the neonatal period (n = 7) had calcium levels that were within the normal range [9.6 ± 0.64 mg/dL (mean ± SD)]. During infancy and childhood (0.15 to 10 years), patients with JMC with the H223R mutation (n = 17) had significantly elevated total serum calcium levels [12.0 ± 1.34 mg/dL (mean ± SD); range, 9.3 to 14.8 mg/dL); similar degrees of hypercalcemia were observed also for cases with other PTHR1 mutations. The three patients with the T410R mutation had lower calcium levels at each measurement (Fig. 3A). Figure 3. View largeDownload slide Serum and urinary calcium measurements for multiple children [0.15 to 10 y; n = 22 for serum calcium, n = 15 for urinary calcium/creatinine (Ca/Cr) ratio] and multiple adults (17 to 38 y; n = 11 for serum calcium, n = 8 for urinary Ca/Cr ratio) with Jansen disease due to different PTHR1 mutations. Each data point represents the mean when a patient had multiple measurements during the two observation periods. (A) Total calcium levels; dashed lines represent the upper/lower end of the adult normal range (8.6 to 10.2 mg/dL). (B) Urinary Ca/Cr ratio; all individual data points are shown. Means ± SD are for patients with the H223R mutation. Dashed line represents the upper end of normal for adult patients (<0.2). Children and adults showed no significant difference in the urinary Ca/Cr ratio. Figure 3. View largeDownload slide Serum and urinary calcium measurements for multiple children [0.15 to 10 y; n = 22 for serum calcium, n = 15 for urinary calcium/creatinine (Ca/Cr) ratio] and multiple adults (17 to 38 y; n = 11 for serum calcium, n = 8 for urinary Ca/Cr ratio) with Jansen disease due to different PTHR1 mutations. Each data point represents the mean when a patient had multiple measurements during the two observation periods. (A) Total calcium levels; dashed lines represent the upper/lower end of the adult normal range (8.6 to 10.2 mg/dL). (B) Urinary Ca/Cr ratio; all individual data points are shown. Means ± SD are for patients with the H223R mutation. Dashed line represents the upper end of normal for adult patients (<0.2). Children and adults showed no significant difference in the urinary Ca/Cr ratio. The average total serum calcium level for adult patients with JMC (17 to 38 years; n = 7) with the H223R mutation was 10.3 ± 0.67 mg/dL, which is significantly lower than that for children affected by this disorder (infancy/childhood vs adult, P < 0.005). Thus, hypercalcemia in JMC is clearly more pronounced during infancy/childhood, with average calcium levels reaching the upper end of the normal range by adulthood (see Fig. 3A). The average urinary calcium/creatinine ratio (mg/mg) was 0.90 ± 0.45 (range, 0.32 to 1.40) for infants/children with the H223R mutation; the ratios for children with other JMC mutations were 0.80 (T410P), 0.45 ± 0.09 (T410R), 0.71 (I458K), and 0.61 (I458R) (Fig. 3B). There was a strong correlation between serum calcium and the urinary calcium/creatinine ratio (Supplemental Fig. 2). Adults with the H223R mutation showed a lower, but still elevated, urinary calcium excretion with an average calcium/creatinine ratio of 0.51 ± 0.09 (infancy/childhood vs adult, P = 0.25). These data show that urinary calcium excretion remained above the normal range even after total serum calcium levels had improved. The serum phosphate concentrations were at the lower end of the age-specific normal range in both childhood and adulthood (Fig. 4A). Serum PTH concentrations for each of the different PTHR1 mutations were below or at the lower end of the reference range, except for case H223R-12 and the adult patients with the T410R mutation. PTH levels were not significantly different for children and adults (infancy/childhood vs adult, P = 0.44) (Fig. 4B). The serum alkaline phosphatase concentrations were above the age-specific normal range, except for one adult with the H223R mutation (H223R-17) and one of the two brothers with the T410R mutation. Few patients had measurements of serum 1,25(OH)2 vitamin D concentrations; these were within or slightly above the reference range (see Supplemental Table 1). Figure 4. View largeDownload slide Serum phosphate levels and PTH levels at different ages for multiple patients affected by Jansen disease due to different PTHR1 mutations. The means are shown when patients had multiple measurements during the two observation periods. (A) Phosphate levels for infants (<1 y), children between 1 and 12 y of age, and patients >15 y of age). The lower limits of the age-dependent reference ranges for phosphate are: 0 to 6 mo, 1.8 mmol/L (5.6 mg/dL); 6 to 12 mo, 1.6 mmol/L (4.9 mg/dL); 1 to 10 y, 1.2 mmol/L (3.8 mg/dL); and >15 y, 0.8 mmol/L (2.5 mg/dL). Individual data points are shown. Means ± SD are for patients with the H223R mutation. (B) PTH levels for children (0.15 to 10 y) and adults (17 to 38 y). Lower end of the adult reference range is 10 pg/mL (dashed line). Individual data points and means ± SD for patients with the H223R mutation are shown. Serum PTH levels were not significantly different for affected children and adults. Figure 4. View largeDownload slide Serum phosphate levels and PTH levels at different ages for multiple patients affected by Jansen disease due to different PTHR1 mutations. The means are shown when patients had multiple measurements during the two observation periods. (A) Phosphate levels for infants (<1 y), children between 1 and 12 y of age, and patients >15 y of age). The lower limits of the age-dependent reference ranges for phosphate are: 0 to 6 mo, 1.8 mmol/L (5.6 mg/dL); 6 to 12 mo, 1.6 mmol/L (4.9 mg/dL); 1 to 10 y, 1.2 mmol/L (3.8 mg/dL); and >15 y, 0.8 mmol/L (2.5 mg/dL). Individual data points are shown. Means ± SD are for patients with the H223R mutation. (B) PTH levels for children (0.15 to 10 y) and adults (17 to 38 y). Lower end of the adult reference range is 10 pg/mL (dashed line). Individual data points and means ± SD for patients with the H223R mutation are shown. Serum PTH levels were not significantly different for affected children and adults. Twelve of 14 patients for whom follow-up ultrasound data were available demonstrated nephrocalcinosis; only two patients, H223R-1 and T410R-1, showed no evidence of renal calcifications when evaluated at the age of 3 and 33 years, respectively (17, 18). Two patients, H223R-16 and T410P, both older than 50 years, exhibited severe chronic kidney disease (see Fig. 2A) secondary to long-standing nephrocalcinosis or renal calculi, as well as urinary tract obstructions and recurrent pyelonephritis (14). Eight patients are known to have developed kyphoscoliosis, and three patients revealed craniosynostosis. Eight patients had been treated with a bisphosphonate and 13 patients had undergone surgical interventions for correction of long-bone deformities, progressive scoliosis, cranial vault reconstruction, or nephrolithotomy (see Supplemental Table 1). The mean final adult height for patients with the H223R mutation was 127.0 ± 6.0 cm for males (n = 4) and 120.4 ± 10.3 cm for females (n = 5) (Fig. 5A). The mean adult height of the three male patients with T410R mutation was 157.7 ± 6.4 cm, which is significantly taller than that of adult males with the H223R mutation (P < 0.002); the final height of the single patient with the T410P mutation was 96 cm. The SD scores for height of the pediatric patients with JMC were at least 2 z scores below the normal mean (Fig. 5B). Figure 5. View largeDownload slide Height data for different patients affected by Jansen disease due to different PTHR1 mutations. (A) Individual final heights for 13 adult patients with JMC. Means ± SD are shown for the final heights of patients with the H223R mutation; the red broken lines indicates the 3rd percentile for normal adult heights. (B) Individual height z scores for eight children. Figure 5. View largeDownload slide Height data for different patients affected by Jansen disease due to different PTHR1 mutations. (A) Individual final heights for 13 adult patients with JMC. Means ± SD are shown for the final heights of patients with the H223R mutation; the red broken lines indicates the 3rd percentile for normal adult heights. (B) Individual height z scores for eight children. Long-term clinical outcomes of patients affected by Jansen disease Only two previous reports provided long-term follow up of patients with JMC, who are both females with either the T410P (14) or the H223R mutation (11). For patient H223R-11 additional data became available showing that CTX levels decreased during the 11 years of bisphosphonate treatment from a maximum of 0.79 ng/mL to ∼0.2 ng/mL. After discontinuation of alendronate at the age of 31 years, her serum calcium level increased to 11.3 to 11.9 mg/dL and serum CTX increased to 0.30 to 0.37 ng/mL. The urinary calcium/creatinine ratio, which had been between 0.22 and 0.33 during bisphosphonate treatment, increased after discontinuation of this medication to 0.44 to 0.53, despite increasing the dose of hydrochlorothiazide to 50 mg/d. At the age of 30 years, a renal CT showed stable bilateral microcalculi (up to 6 mm in size), but no nephrocalcinosis; serum creatinine levels remained between 0.4 and 0.5 mg/dL. Additional retrospectively collected clinical and laboratory findings for several other patients with JMC are provided in Fig. 1 and Supplemental Table 1. Characterization of the PTHR1 mutants in HEK293-derived reporter cells GS22A cells (HEK293 cells stably transfected with the Glosensor cAMP reporter) were transiently transfected with increasing amounts of plasmid DNA (10, 20, 40, 80, and 160 ng per well) encoding either a mutant or the wild-type PTHR1. The PTHR1 mutants showed dose-dependent increases in basal cAMP levels that reached a plateau at 160 ng DNA per well. All mutant receptors showed agonist-independent cAMP generation; the T410R mutant revealed the lowest constitutive activity, whereas I458K-PTHR1 and I458R-PTHR1 generated a much higher basal cAMP level; there was no readily detectable increase in basal cAMP generation in cells expressing the wild-type PTHR1 (Fig. 6A). Similar to previously reported findings (6), cell surface expression of all mutant receptors (100 ng per well), as determined by anti-PTHR1 antibody binding, was significantly reduced in comparison with the wild-type PTHR1 (data not shown). Each PTHR1 mutant mediated a cAMP response to increasing concentrations of PTH(1–34) that was reduced as compared with that mediated by the WT-PTHR1, except for the I458K mutant, which exhibited an increased sensitivity to the agonist ligand (Fig. 6B). Treatment of cells expressing the different PTHR1 mutants with the ligand analog [L11,dW12,W23,Y36]PTHrP(7–36) (10−6 M) resulted a rapid and persistent reduction in basal cAMP signaling, consistent with the notion that this N-terminally truncated antagonist peptide can function as an inverse agonist and thus cause a decrease in the proportion of mutant receptors that are in the active-state conformation (Fig. 6C). Figure 6. View largeDownload slide Functional evaluation of the wild-type and different PTHR1 mutants in HEK-293/Glosensor (GS22A) cells. For some data points, the error bars are small and thus within the height of the symbol. (A) Basal cAMP production in GS22A cells that were transiently transfected with increasing amounts of plasmid DNA (10, 20, 40, 80, and 160 ng per well) encoding either a mutant or the wild-type PTHR1. (B) PTH-stimulated cAMP accumulation in cells transfected with plasmid DNA (100 ng per well) encoding either wild-type or mutant PTHR1s. Data are shown as the AUC of cAMP accumulation (mean ± SEM. (C) Functional evaluation of the inverse agonist [L11,dW12,W23,Y36]PTHrP(7–36) in GS22A cells expressing the wild-type PTHR1 or different JMC mutants. The cAMP-dependent luminescence responses in cells transfected with plasmid DNA (100 ng per well) encoding either wild-type or mutant receptor are shown. Data are shown as the AUC of cAMP-dependent luminescence measured over time after addition (t = 0) of either buffer (open symbols) or inverse agonist (filled symbols); all data were corrected for time 0 (mean ± SEM). Figure 6. View largeDownload slide Functional evaluation of the wild-type and different PTHR1 mutants in HEK-293/Glosensor (GS22A) cells. For some data points, the error bars are small and thus within the height of the symbol. (A) Basal cAMP production in GS22A cells that were transiently transfected with increasing amounts of plasmid DNA (10, 20, 40, 80, and 160 ng per well) encoding either a mutant or the wild-type PTHR1. (B) PTH-stimulated cAMP accumulation in cells transfected with plasmid DNA (100 ng per well) encoding either wild-type or mutant PTHR1s. Data are shown as the AUC of cAMP accumulation (mean ± SEM. (C) Functional evaluation of the inverse agonist [L11,dW12,W23,Y36]PTHrP(7–36) in GS22A cells expressing the wild-type PTHR1 or different JMC mutants. The cAMP-dependent luminescence responses in cells transfected with plasmid DNA (100 ng per well) encoding either wild-type or mutant receptor are shown. Data are shown as the AUC of cAMP-dependent luminescence measured over time after addition (t = 0) of either buffer (open symbols) or inverse agonist (filled symbols); all data were corrected for time 0 (mean ± SEM). Discussion We report on clinical and laboratory observations for 24 patients affected by JMC with information collected from shortly after birth up to the age of 56 years; serial measurements are presented for several cases. Our goal was to help assess the natural history profile for JMC, an ultra-rare, high-impact disease. We found that all but one patient had blood calcium levels that were within the reference range during the first 1.5 months of life, indicating that the development of hypercalcemia depends largely on postnatal mechanisms, which could include enhanced 1,25(OH)2 vitamin D–dependent intestinal calcium absorption and enhanced resorption of mineralized bone. Hypercalcemia was variable, but typically became pronounced during infancy/childhood and improved significantly by adulthood; ionized calcium was normal in the few adult cases in whom it was measured. Importantly, however, hypercalciuria with suppressed PTH secretion persisted into adulthood and likely contributed to the progressive decline in renal function that was encountered in the two older patients. In contrast, serum phosphate levels remained at the lower end of the age-specific normal range. We also noted considerable variability in the clinical findings among different patients with JMC, even in those carrying the same PTHR1 mutation. For example, female patient H223R-12 had never shown overt abnormalities of mineral ion homeostasis, whereas her two affected children were hypercalcemic by age 2 years (13). The reason for such variations in blood calcium levels is unknown, but could involve differences in dietary intake of calcium and/or vitamin D, or some unknown genetic modifiers affecting calcium homeostasis. Twelve of fourteen patients, for whom results of ultrasonographic studies were available, showed nephrocalcinosis. The T410R mutation, present in three members of one family (17), appears to cause a relatively milder form of JMC, as it was not associated with major elevations in blood calcium levels, one of the three patients had normal renal ultrasound images, and the adult heights were at or close to the 3rd percentile, despite radiographic growth plate changes typical of the disease. Consistent with the less severe clinical and biochemical abnormalities associated with the T410R mutation, in vitro studies showed only a low level of constitutive cAMP formation for this mutant allele (17). The findings in this family with the T410R mutation make it evident that certain PTHR1 activating mutations can cause changes in the growth plates without causing major abnormalities in mineral ion homeostasis. The I458K mutation, which had been identified only in a single pediatric case (12), showed elevated basal activity and full responsiveness to PTH(1–34) Mineral ion abnormalities and impairment of growth revealed no obvious difference when compared with patients with other PTHR1 mutations at the same age, but it will be necessary to determine whether differences can be observed later in life. It remains uncertain as to why hypercalcemia ameliorates with age and why hypercalciuria persists in most adult patients with JMC without overt hypercalcemia. Several mechanisms most likely contribute to the blood calcium elevation observed at certain times in affected individuals, namely increased bone resorption, enhanced intestinal calcium absorption, and possibly enhanced calcium reabsorption in the distal renal tubules. With the exception of a few adult patients, serum levels of alkaline phosphatase, a marker of osteoblast activity, remained above the reference range (see Supplemental Table 1). It is therefore conceivable that increased bone turnover with increased bone resorption persists during adulthood. Only few published reports discuss the possibility of impaired renal calcium handling in JMC. In fact, only Parfitt et al. (14) investigated the relationship between fractional calcium excretion and serum calcium levels in the patient with JMC with the T410P mutation, and the authors had shown normalization of tubular calcium reabsorption with age. However, when the studies were performed, the patient already had significantly impaired renal function, which may have contributed to the decline in calcium excretion. Nonetheless, it appears possible that decreased serum 1,25(OH)2 vitamin D concentrations during adulthood, combined with reduced expression of the PTHR1 mutant in distal renal tubules and thus reduced constitutive calcium reabsorption, leads to amelioration of hypercalcemia, albeit with enhanced bone resorption and urinary calcium excretion persisting. PTH levels in older patients remained suppressed at or below the lower limit of the reference range despite improved serum calcium levels. Circulating PTH levels are regulated mainly by the concentration of blood ionized calcium, which activates the calcium-sensing receptors expressed on the surface of parathyroid cells to thereby reduce hormone secretion (24). Although blood ionized calcium levels were available only for three adult patients [H223R-4, 1.28 (normal range: 1.08 to 1.34) (6); H223R-11, 1.43 (normal range: 1.15 to 1.33) (11); H223R-12, 1.25 (normal range: 1.14 to 1.29) (13)], the measurements were above or at the upper end of the normal range. Hence, ionized calcium may be elevated intermittently, thus activating the calcium-sensing receptor on the parathyroid cells sufficiently to reduce PTH secretion. Importantly, low or low-normal PTH levels prevent activation of PTHR1 expressed from the normal allele, thus limiting most likely distal tubular calcium reabsorption and contributing to the hypercalciuria and nephrocalcinosis. Consequently, a decreased blood PTH level combined with an increased urine calcium excretion and typical skeletal findings may be a more reliable indicator of JMC than the blood calcium level alone, which has been normal in some patients of the current study. Most patients with JMC, whose ultrasonographic studies were available, revealed nephrocalcinosis early in life and two older patients developed severe chronic kidney disease. These complications of the disease are probably caused or accelerated by a tendency toward hypercalcemia combined with markedly increased urinary calcium and phosphate excretion. In the patient with the T410P mutation, nephrocalcinosis contributed to the chronic urinary tract obstructions, making her prone to infections (14). It is therefore important to routinely monitor renal function in adult patients with JMC, as it appears to decline considerably with age, especially with recurrent pyelonephritis or obstructive uropathy. To slow or prevent deterioration of kidney function, treatment with a bisphosphonate and the subsequent addition of a thiazide diuretic has been reported to normalize blood calcium levels and to markedly reduce urine calcium excretion in patients with JMC (11, 19). Onuchi et al. (11) documented in one patient, H223R-11, that the combination of alendronate (10 mg/d), initiated at 20 years of age, and hydrochlorothiazide, initiated at 26 years of age (initially 12.5 mg/d, subsequently increased to 25 mg/d), normalized urinary calcium excretion. Discontinuation of alendronate at the age of 31 years led to an increase in serum and urine calcium, despite treatment with a higher dose of hydrochlorothiazide (50 mg/d), but her renal function has thus far remained stable. Although long-term outcome data for five additional patients with the H223R mutation, who had been treated with a bisphosphonate, are not yet available, it appears plausible that limiting urinary calcium excretion will help preserve renal function. Although JMC is very rare, the impact of the disease on patient quality of life and the associated long-term health care burden emphasize the need for an effective form of therapy. No specific treatment of JMC is currently available, however. Amino-terminally truncated PTH and PTHrP analogs with the Gly12→dTrp substitution, originally developed as PTH antagonists (25), function in vitro as inverse agonists on the constitutively active PTHR1 mutants of JMC (26, 27) (see Fig. 6B) and also in a transgenic mouse model of JMC (28). Whether such an inverse agonist ligand could be developed so as to suppress the elevated signaling activity of the mutant PTHR1 in bone cells, growth plate chondrocytes, and kidney cells of patients with JMC remains to be investigated. In conclusion, findings in 24 patients with JMC reveal that the final adult height of most patients is markedly reduced; only individuals with the T410R mutation, a PTHR1 mutation with only limited constitutive activity when tested in vitro, showed better growth. Hypercalcemia in JMC varies with age and depends at least to some extent on the intrinsic signaling properties of the specific PTHR1 mutant. Hypercalcemia improves with age, but most patients continue to exhibit long-standing hypercalciuria and thus nephrocalcinosis, which likely contributes to progressively impaired renal function. Findings in vitro suggest that PTHR1 inverse agonist ligands are worth exploring as a potential means of therapy for JMC. Abbreviations: Abbreviations: AUC area under the curve Glosensor pGlosensor-22F JMC Jansen metaphyseal chondrodysplasia PTHR1 PTH/PTH-related peptide receptor PTHrP PTH-related peptide Acknowledgments We thank the patients and their families. Financial Support: This work was supported by National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases Grants DK11794 (to H.J. and T.J.G.) DK46718 (to H.J.), and R01DK113039 (to H.J. and T.J.G.). Current Affiliation: C. Silve’s current affiliation is INSERM Unité 1169, Hôpital Bicêtre, Le Kremlin Bicêtre, 94270 Paris, France. Disclosure Summary: H.N. is an appointee of MGH and employee of Chugai Pharmaceutical Co., Ltd. The remaining authors have nothing to disclose. References 1. Lee K , Deeds JD , Segre GV . Expression of parathyroid hormone-related peptide and its receptor messenger ribonucleic acids during fetal development of rats . Endocrinology . 1995 ; 136 ( 2 ): 453 – 463 . 2. Maes C , Kronenberg HM . Bone development and remodeling. In: DeGroot LJ and Jameson JL , eds. Endocrinology . 7th ed. Philadelphia, PA : W.B. Saunders ; 2016 : 1038 – 1062 . 3. Gardella TJ , Jüppner H , Brown EM , Kronenberg HM , Potts JT Jr . Parathyroid hormone and parathyroid hormone receptor type 1 in the regulation of calcium and phosphate homeostasis and bone metabolism. In: DeGroot LJ and Jameson JL , eds. Endocrinology . 7th ed. Philadelphia, PA : W.B. Saunders ; 2016 : 969 – 990 . 4. Jansen M . Über atypische Chondrodystrophie (Achondroplasie) und über eine noch nicht beschriebene angeborene Wachstumsstörung des Knochensystems: Metaphysäre Dysostosis . Zeitschr Orthop Chir . 1934 ; 61 : 253 – 286 . 5. Schipani E , Kruse K , Jüppner H . A constitutively active mutant PTH-PTHrP receptor in Jansen-type metaphyseal chondrodysplasia . Science . 1995 ; 268 ( 5207 ): 98 – 100 . 6. Schipani E , Langman CB , Parfitt AM , Jensen GS , Kikuchi S , Kooh SW , Cole WG , Jüppner H . Constitutively activated receptors for parathyroid hormone and parathyroid hormone-related peptide in Jansen’s metaphyseal chondrodysplasia . N Engl J Med . 1996 ; 335 ( 10 ): 708 – 714 . 7. Silve C , Jüppner H . Genetic disorders caused by mutations in the PTH/PTHrP receptor and down-stream effector molecules. In: Bilezikian J , ed. The Parathyroids: Basic and Clinical Concepts . San Diego, CA : Academic Press ; 2015 : 587 – 605 . 8. Frame B , Poznanski AK . Conditions that may be confused with rickets. In: DeLuca HF and Anast CS , eds. Pediatric Diseases Related to Calcium . New York, NY : Elsevier ; 1980 : 269 – 289 . 9. Rao DS , Frame B , Reynolds WA , Parfitt AM . Hypercalcemia in metaphyseal chondrodysplasia of Jansen (MCD): an enigma. In: Norman AW , Schaefer K , von Herrath D , Grigoleit HG , Coburn JW , DeLuca HF , Mawer EB , and Suda T , eds. Vitamin D, Basic Research and Its Clinical Application . Berlin, Germany : Walter de Gruyter ; 1979 : 1173 – 1176 . 10. Brown WW , Jüppner H , Langman CB , Price H , Farrow EG , White KE , McCormick KL . Hypophosphatemia with elevations in serum fibroblast growth factor 23 in a child with Jansen’s metaphyseal chondrodysplasia . J Clin Endocrinol Metab . 2009 ; 94 ( 1 ): 17 – 20 . 11. Onuchic L , Ferraz-de-Souza B , Mendonca BB , Correa PH , Martin RM . Potential effects of alendronate on fibroblast growth factor 23 levels and effective control of hypercalciuria in an adult with Jansen’s metaphyseal chondrodysplasia . J Clin Endocrinol Metab . 2012 ; 97 ( 4 ): 1098 – 1103 . 12. Savoldi G , Izzi C , Signorelli M , Bondioni MP , Romani C , Lanzi G , Moratto D , Verdoni L , Pinotti M , Prefumo F , Superti-Furga A , Pilotta A . Prenatal presentation and postnatal evolution of a patient with Jansen metaphyseal dysplasia with a novel missense mutation in PTH1R . Am J Med Genet A . 2013 ; 161A ( 10 ): 2614 – 2619 . 13. Nampoothiri S , Fernández-Rebollo E , Yesodharan D , Gardella TJ , Rush ET , Langman CB , Jüppner H . Jansen metaphyseal chondrodysplasia due to heterozygous H223R-PTH1R mutations with or without overt hypercalcemia . J Clin Endocrinol Metab . 2016 ; 101 ( 11 ): 4283 – 4289 . 14. Parfitt AM , Schipani E , Rao DS , Kupin W , Han Z-H , Jüppner H . Hypercalcemia due to constitutive activity of the parathyroid hormone (PTH)/PTH-related peptide receptor: comparison with primary hyperparathyroidism . J Clin Endocrinol Metab . 1996 ; 81 ( 10 ): 3584 – 3588 . 15. Minagawa M , Arakawa K , Takeuchi S , Minamitani K , Yasuda T , Niimi H . Jansen-type metaphyseal chondrodysplasia: analysis of PTH/PTH-related protein receptor messenger RNA by the reverse transcriptase-polymerase chain method . Endocr J . 1997 ; 44 ( 4 ): 493 – 499 . 16. Silverthorn KG , Houston CS , Duncan BP . Murk Jansen’s metaphyseal chondrodysplasia with long-term followup . Pediatr Radiol . 1987 ; 17 ( 2 ): 119 – 123 . 17. Bastepe M , Raas-Rothschild A , Silver J , Weissman I , Jüppner H , Gillis D . A form of Jansen’s metaphyseal chondrodysplasia with limited metabolic and skeletal abnormalities is caused by a novel activating parathyroid hormone (PTH))/PTH-related peptide receptor mutation . J Clin Endocrinol Metab . 2004 ; 89 ( 7 ): 3595 – 3600 . 18. Kruse K , Schütz C . Calcium metabolism in the Jansen type of metaphyseal dysplasia . Eur J Pediatr . 1993 ; 152 ( 11 ): 912 – 915 . 19. Schipani E , Langman CB , Hunzelman J , LeMerrer M , Loke KY , Dillon MJ , Silve C , Jüppner H . A novel parathyroid hormone (PTH)/PTH-related peptide receptor mutation in Jansen’s metaphyseal chondrodysplasia . J Clin Endocrinol Metab . 1999 ; 84 ( 9 ): 3052 – 3057 . 20. Schwartz GJ , Haycock GB , Edelmann CM Jr , Spitzer A . A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine . Pediatrics . 1976 ; 58 ( 2 ): 259 – 263 . 21. Fadrowski JJ , Neu AM , Schwartz GJ , Furth SL . Pediatric GFR estimating equations applied to adolescents in the general population . Clin J Am Soc Nephrol . 2011 ; 6 ( 6 ): 1427 – 1435 . 22. Carter PH , Dean T , Bhayana B , Khatri A , Rajur R , Gardella TJ . Actions of the small molecule ligands SW106 and AH-3960 on the type-1 parathyroid hormone receptor . Mol Endocrinol . 2015 ; 29 ( 2 ): 307 – 321 . 23. Cheloha RW , Watanabe T , Dean T , Gellman SH , Gardella TJ . Backbone modification of a parathyroid hormone receptor-1 antagonist/inverse agonist . ACS Chem Biol . 2016 ; 11 ( 10 ): 2752 – 2762 . 24. Brown EM . Control of parathyroid hormone secretion by its key physiological regulators. In: Bilezikian J , ed. The Parathyroids: Basic and Clinical Concepts . San Diego, CA : Academic Press ; 2015 : 101 – 118 . 25. Rosen HN , Lim M , Garber J , Moreau S , Bhargava HN , Pallotta J , Spark R , Greenspan S , Rosenblatt M , Chorev M . The effect of PTH antagonist BIM-44002 on serum calcium and PTH levels in hypercalcemic hyperparathyroid patients . Calcif Tissue Int . 1997 ; 61 ( 6 ): 455 – 459 . 26. Gardella TJ , Luck MD , Jensen GS , Schipani E , Potts JT Jr , Jüppner H . Inverse agonism of amino-terminally truncated parathyroid hormone (PTH) and PTH-related peptide (PTHrP) analogs revealed with constitutively active mutant PTH/PTHrP receptors . Endocrinology . 1996 ; 137 ( 9 ): 3936 – 3941 . 27. Carter PH , Petroni BD , Gensure RC , Schipani E , Potts JT Jr , Gardella TJ . Selective and nonselective inverse agonists for constitutively active type-1 parathyroid hormone receptors: evidence for altered receptor conformations . Endocrinology . 2001 ; 142 ( 4 ): 1534 – 1545 . 28. Guo J , Noda H , Reyes M , Armanini M , Martins J , Bouxsein M , Demay M , Jüppner H , Gardella T . Inverse agonist infusion mitigates bone remodeling abnormalities in the Col1-PTHR-H223R mouse model of Jansen’s metaphyseal chondrodysplasia . J Bone Miner Res . 2017 ; 32 : S34 . Copyright © 2018 Endocrine Society

Qi, Jia;Wang, Wangsheng;Zhu, Qinling;He, Yaqiong;Lu, Yao;Wang, Yuan;Li, Xiaoxue;Chen, Zi-jiang;Sun, Yun

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02459pmid: 29618067

Abstract Context Endometrial insulin resistance (IR) may account for the endometrial dysfunction in polycystic ovary syndrome (PCOS). The underlying mechanism remains to be elucidated. Objective To investigate whether the abundance of 11β-hydroxysteroid dehydrogenases (11β-HSDs) 1 and 2 and cortisol as well as the insulin signaling pathway are altered in PCOS endometrium and to clarify the relationship between endometrial IR and local cortisol. Design We measured cortisol and cortisone concentrations, 11β-HSD1 and 11β-HSD2, and core insulin signaling molecules in endometrial biopsies collected from non-PCOS and PCOS with or without IR patients on the seventh day after human chorionic gonadotropin injection. We also studied the effects of cortisol on glucose uptake and the insulin signaling pathway in primary cultured endometrial epithelial cells (EECs). Results The cortisol concentration was elevated, whereas 11β-HSD2 expression was diminished in endometrial biopsies obtained from PCOS with IR patients compared with those from non-PCOS and PCOS without IR patients. The implantation rate was relatively impaired and the endometrial insulin signaling pathway was defective in PCOS with IR patients. In addition, cortisol attenuated insulin-stimulated glucose uptake in EECs, which was mediated by inhibition of Akt phosphorylation and glucose transporter type 4 translocation via induction of phosphatase and tensin homolog deleted on chromosome ten (PTEN). Conclusions Decreased oxidation of cortisol and defects of insulin signaling in endometrium were observed in PCOS with IR patients. The excessive cortisol level, derived from the reduction of 11β-HSD2, might contribute to the development of endometrial IR by inhibiting the insulin signaling pathway via induction of PTEN expression in EECs. Polycystic ovary syndrome (PCOS), a common and complex endocrine disorder, affects 5% to 20% of reproductive-age women with short- and long-term effects (1). PCOS is characterized by polycystic ovaries, ovulatory dysfunction, and hyperandrogenism (2). In addition, a high percentage of patients with PCOS have symptoms of insulin resistance (IR) (2–4). The abnormal endocrine and metabolic characteristics of PCOS might be detrimental to endometrial function, manifesting as endometrial hyperplasia or cancer and reduction of receptivity (5–7). Patients with PCOS have been reported to have adverse reproductive outcomes, including higher abortion rates compared with the unaffected population (7). The reduction in fertility might not only be attributed to ovulatory dysfunction, but also to endometrial defects. When compared with fertile endometrium, several alterations of the insulin signaling pathway in PCOS endometrium have been reported, such as increased phosphatase and tensin homolog deleted on chromosome ten (PTEN) as well as decreased insulin receptor substrate 1 (IRS-1) (8, 9). Some studies have linked endometrial IR to decreased endometrial receptivity (10) and tumorigenesis (11, 12). However, the etiology of local endometrial IR remains to be elucidated. Endogenous glucocorticoids play a crucial role in many areas, including in the pathophysiology of IR. The conversion of inactive glucocorticoids and active glucocorticoids is catalyzed by 11β-hydroxysteroid dehydrogenases (11β-HSDs). There are two types of 11β-HSDs: 11β-HSD1 and 11β-HSD2 (13, 14). 11β-HSD1 has both reductase and oxidase functions, bidirectionally converting biologically inactive cortisone and active cortisol; 11β-HSD2 only has an oxidase function, converting active cortisol to inactive cortisone (15). Several studies have demonstrated that cortisol concentration and 11β-HSDs were changed in serum, adipose tissue, and granulosa cells in PCOS patients (16–18). In addition, our previous study showed that cortisol generated locally by 11β-HSD1 contributed to IR in granulosa cells in PCOS (19). However, it is unclear whether the local generation of cortisol in endometrium exerts a role in the process of endometrial IR in PCOS. Considering the ovarian 11β-HSD alterations in women with PCOS, our primary aim in this study was to clarify whether there is an imbalanced state of glucocorticoid and its metabolic enzymes in endometria from patients with PCOS with or without IR. We also sought to establish a possible correlation between cortisol and local IR. A secondary aim of this study was to clarify in vitro whether cortisol could abolish insulin-stimulated glucose uptake and insulin signaling pathway in primary cultured endometrial epithelial cells (EECs). Materials and Methods Patients and tissue collection The endometrial biopsies were collected with endometrial suction curettes (Runting) from PCOS and non-PCOS patients undergoing gonadotropin-releasing hormone antagonist stimulation cycle without fresh embryo transfer. The biopsies were collected on the seventh day after human chorionic gonadotropin (hCG) injection, known as the window of implantation (WOI) phase. The diagnosis of PCOS was established according to the revised Rotterdam consensus (20). The subgroups of PCOS with IR and PCOS without IR were subdivided according to the homeostasis model assessment of IR index (HOMA-IR [fasting serum insulin (μIU/mL) × fasting serum glucose (mmol/L)/22.5]), with 3.15 selected as a cutoff point (21). Non-PCOS patients were women with regular menstrual cycles, normal body mass index (18.5 to 23.9 kg/m2), and only tubal infertile conditions without IR. The endometrial biopsies from non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR (n = 18) were snap frozen in liquid nitrogen for extraction of glucocorticoids, messenger RNA (mRNA), and protein to detect cortisol, cortisone, 11β-HSD mRNA and protein, phosphorylated IRS-1 and IRS-1, phosphorylated Akt and Akt, and PTEN mRNA and protein. The EECs were isolated from other patients including non-PCOS (n = 29) and PCOS (n = 21). All procedures were performed at the Center for Reproductive Medicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University. Written informed consent was obtained, and approval of the ethics protocol was granted from the Ethics Committee of Renji Hospital (2017081109). Demographic features and clinical outcomes Baseline serum hormonal profiles, including follicle-stimulating hormone, luteinizing hormone, testosterone, estradiol, and anti-Müllerian hormone, were determined using chemiluminescence assay kits (Beckman Access Health Co.). Serum fasting insulin and fasting glucose were measured using a chemiluminescence assay kit (Beckman Access Health Co.) and a standard glucose oxidase method (Roche), respectively. Quantitative insulin sensitivity check index was calculated as 1/[logI0(μIU/mL) + logG0(mg/dL)]. The implantation rate was defined as the number of gestational sacs per number of embryos transferred in every frozen embryo transfer cycle. Extraction and measurement of cortisol and cortisone in endometrium The endometrium was ground in liquid nitrogen and extracted with ethyl acetate. After evaporation, the extract was resuspended in phosphate-buffered saline. The suspension was equally divided and reconstituted in the assay buffer provided by the manufacturers and then measured using a cortisol assay kit (R&D Systems) and a cortisone chemiluminescent immunoassay kit (Innovative Research) following manufacturer instructions. Immunohistochemical and immunofluorescent staining Protein expression of 11β-HSD2 was assessed in paraffin-embedded endometrial tissue sections. Immunostaining was performed on 5-μm-thick tissue sections as previously described (22). Briefly, the endogenous peroxidase activity was quenched with 3% H2O2, and then 11β-HSD2 or preimmune serum at 1:500 dilution was used as a primary or negative control for overnight incubation at 4°C followed by secondary antibody for 30 minutes at 37°C, respectively. The colorimetric reactions were developed using a standard diaminobenzidine kit (ZSGB-BIO). Immunofluorescent staining was performed on cultured cells fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100. After blocking and incubation with primary antibodies, cells were incubated with Alexa Fluor 488- and 594-labeled secondary antibodies (Proteintech). Nuclei were stained with 4´,6´-diamino-2-phenylindole (1 μg/mL). Images were obtained with a microscope and camera connected to a computer with an image analysis system (Zeiss). The primary antibodies are listed in Supplemental Table 1. Cell culture and treatment EECs were enzymatically isolated from human endometrium curettage samples according to a selective attachment method (23) with minor modifications. Briefly, endometrial samples were digested with collagenase type I and deoxyribonuclease and then sequentially size fractionated with 180- and 40-μm sieves. Epithelial glands were retained on the 40-μm sieve and were collected by backwashing the 40-μm filter paper and resuspended in Dulbecco’s modified Eagle medium/Hams F12 containing 10% fetal bovine serum (Gibco) and 1% antibiotic-antimycotic solution (Gibco). Cell purity was tested routinely by immunofluorescence staining for cytokeratin and vimentin. To compare the glucose uptake and glucose transporter type 4 (GLUT4) translocation among non-PCOS, PCOS without IR, and PCOS with IR patients, the cells were cultured in phenol red-free and serum-free medium 1 day after plating and stimulated with insulin (100 nM; Sigma) for 20 and 30 minutes. To study the role of cortisol on the insulin-stimulated glucose uptake, Akt phosphorylation, and GLUT4 translocation, EECs were treated in phenol red-free and serum-free culture medium after plating for 3 days. After treatment with cortisol (1 μM; Sigma) for 24 hours, the cells were stimulated with insulin for 20, 15, and 30 minutes, respectively. The measurement of glucose uptake is described below. To detect the effect of cortisol on PTEN expression, cells were treated with cortisol (0.1 and 1 μM) for 24 hours before analysis with Western blotting or quantitative real-time polymerase chain reaction (PCR) as described below. To determine the involvement of PTEN, the PTEN inhibitor bpV (phen) (1 μM; Sigma) was added to EECs for 30 minutes before insulin stimulation. Glucose uptake Glucose uptake in EECs was measured after insulin stimulation (100 nM, 20 minutes) using Glucose Uptake-Glo Assay (Promega) according to manufacturer instructions. Quantitative real-time PCR Total RNA from cells and endometrial biopsies was extracted using a total RNA Kit (Omega Bio-Tek) according to manufacturer instructions and was reverse transcribed to complementary DNA using the PrimeScript reverse transcription kit (TaKaRa) with appropriate controls. Quantitative real-time PCR was performed and analyzed with ABI Prism System (Applied Biosystems) using SYBR Premix (TaKaRa) in triplicate. Relative mRNA expression was calculated by the comparative cycle threshold method with ACTB as the housekeeping gene. The primer sequences are listed in Supplemental Table 2. Protein extraction and Western blotting Total protein was extracted from cells and endometrial biopsies using ice-cold radio-immunoprecipitation assay lysis buffer (CWBIO) containing protease inhibitor cocktail (Roche) and phosphatase inhibitor (Active Motif). Membrane and cytoplasmic proteins were extracted using the Membrane and Cytoplasmic Protein Extraction Kit (Sangon Biotech). Protein was quantified with a Bradford assay, and 20 μg protein of each sample was electrophoresed in 10% sodium dodecyl sulfate–polyacrylamide gel and transferred to a nitrocellulose blot. After blocking and incubation with primary antibodies, the membranes were incubated with the respective secondary antibody conjugated with horseradish peroxidase (Proteintech) for 1 hour. Bands with peroxidase activity were detected by an enhanced chemiluminescent detection kit (Merck Millipore) and visualized with a G-Box chemiluminescence image capture system (Syngene). Primary antibodies are listed in Supplemental Table 1. Statistical analysis All data are reported as the mean ± standard deviation (SD). Analyses were performed using the Statistical Package for Social Science (version 16.0; SPSS) and Graphpad Prism statistical software (version 5.0, Graphpad). The data were initially subjected to Kolmogorov-Smirnov tests to assess deviation from Gaussian distribution. For normally distributed data, we applied unpaired t test and one-way analysis of variance followed by Bonferroni tests. For data not normally distributed, we applied Kruskal-Wallis test followed by Dunn’s multiple comparison test. Correlation between variables was performed using Pearson correlation analysis. P < 0.05 was considered to be statistically significant. Results Clinical characteristics and implantation outcomes The demographic characteristics and implantation outcomes of recruited participants are displayed in Table 1. Fifty-four patients were classified into three groups: non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR (n = 18). Patient age, basal follicle-stimulating hormone, estradiol, and hormones on hCG injection day were comparable among the three groups. Basal levels of anti-Müllerian hormone, luteinizing hormone, and testosterone were significantly higher in PCOS with or without IR than in non-PCOS patients. The fasting insulin, HOMA-IR, and quantitative insulin sensitivity check index were significantly higher in PCOS with IR patients than in non-PCOS and PCOS without IR patients. This is a characteristic of IR syndrome. Although there was not a significant difference, there was a tendency for the implantation rate in PCOS with IR women to be lower than that of non-PCOS (P = 0.06). Table 1. Demographic Features and Clinical Outcomes of Recruited Patients Non-PCOS (n = 18) PCOS Without IR (n = 18) PCOS With IR (n = 18) Age, y 27.67 ± 2.74 28.28 ± 3.39 28.56 ± 2.20 BMI, kg/m2 20.84 ± 2.19 23.54 ± 4.48 25.79 ± 3.84a Basal FSH, mIU/mL 7.13 ± 1.71 6.04 ± 1.23 6.13 ± 1.17 Basal LH, mIU/mL 4.87 ± 1.67 6.88 ± 5.77a 9.66 ± 7.36a Basal E2, pg/mL 42.40 ± 17.94 38.68 ± 20.68 42.24 ± 17.45 Basal T, nmol/L 1.01 ± 0.58 1.66 ± 0.67a 1.84 ± 0.83a Hormones on hCG day  LH, mIU/mL 2.04 ± 1.69 1.48 ± 0.85 2.13 ± 1.52  E2, pg/mL 2988.44 ± 1613.16 3992.64 ± 1825.50 3157.94 ± 1436.53  P4, ng/mL 0.92 ± 0.27 1.29 ± 0.67 1.21 ± 1.14 Fasting glucose, mmol/L 4.46 ± 0.37 4.88 ± 0.38 5.42 ± 1.44a Fasting insulin, μIU/mL 6.49 ± 1.99 8.91 ± 2.00 17.31 ± 5.91a,b HOMA-IR 1.30 ± 0.45 1.91 ± 0.39 4.18 ± 1.90a,b QUICKI 0.37 ± 0.03 0.35 ± 0.01a 0.31 ± 0.01a,b AMH, ng/mL 5.29 ± 1.44 12.99 ± 4.15a 10.94 ± 5.04a Frozen embryo transfer cycle 27 24 24 Implantation rate 50.0% ± 46.0% 37.5% ± 39.7% 27.1% ± 39.0% Non-PCOS (n = 18) PCOS Without IR (n = 18) PCOS With IR (n = 18) Age, y 27.67 ± 2.74 28.28 ± 3.39 28.56 ± 2.20 BMI, kg/m2 20.84 ± 2.19 23.54 ± 4.48 25.79 ± 3.84a Basal FSH, mIU/mL 7.13 ± 1.71 6.04 ± 1.23 6.13 ± 1.17 Basal LH, mIU/mL 4.87 ± 1.67 6.88 ± 5.77a 9.66 ± 7.36a Basal E2, pg/mL 42.40 ± 17.94 38.68 ± 20.68 42.24 ± 17.45 Basal T, nmol/L 1.01 ± 0.58 1.66 ± 0.67a 1.84 ± 0.83a Hormones on hCG day  LH, mIU/mL 2.04 ± 1.69 1.48 ± 0.85 2.13 ± 1.52  E2, pg/mL 2988.44 ± 1613.16 3992.64 ± 1825.50 3157.94 ± 1436.53  P4, ng/mL 0.92 ± 0.27 1.29 ± 0.67 1.21 ± 1.14 Fasting glucose, mmol/L 4.46 ± 0.37 4.88 ± 0.38 5.42 ± 1.44a Fasting insulin, μIU/mL 6.49 ± 1.99 8.91 ± 2.00 17.31 ± 5.91a,b HOMA-IR 1.30 ± 0.45 1.91 ± 0.39 4.18 ± 1.90a,b QUICKI 0.37 ± 0.03 0.35 ± 0.01a 0.31 ± 0.01a,b AMH, ng/mL 5.29 ± 1.44 12.99 ± 4.15a 10.94 ± 5.04a Frozen embryo transfer cycle 27 24 24 Implantation rate 50.0% ± 46.0% 37.5% ± 39.7% 27.1% ± 39.0% All data are mean ± standard deviation values. Abbreviations: AMH, anti-Müllerian hormone; BMI, body mass index; E2, estradiol; FSH, follicle-stimulating hormone; LH, luteinizing hormone; P4, progesterone; QUICKI, quantitative insulin sensitivity check index; T, testosterone. a P < 0.05 vs non-PCOS. b P < 0.05 vs PCOS without IR. View Large Table 1. Demographic Features and Clinical Outcomes of Recruited Patients Non-PCOS (n = 18) PCOS Without IR (n = 18) PCOS With IR (n = 18) Age, y 27.67 ± 2.74 28.28 ± 3.39 28.56 ± 2.20 BMI, kg/m2 20.84 ± 2.19 23.54 ± 4.48 25.79 ± 3.84a Basal FSH, mIU/mL 7.13 ± 1.71 6.04 ± 1.23 6.13 ± 1.17 Basal LH, mIU/mL 4.87 ± 1.67 6.88 ± 5.77a 9.66 ± 7.36a Basal E2, pg/mL 42.40 ± 17.94 38.68 ± 20.68 42.24 ± 17.45 Basal T, nmol/L 1.01 ± 0.58 1.66 ± 0.67a 1.84 ± 0.83a Hormones on hCG day  LH, mIU/mL 2.04 ± 1.69 1.48 ± 0.85 2.13 ± 1.52  E2, pg/mL 2988.44 ± 1613.16 3992.64 ± 1825.50 3157.94 ± 1436.53  P4, ng/mL 0.92 ± 0.27 1.29 ± 0.67 1.21 ± 1.14 Fasting glucose, mmol/L 4.46 ± 0.37 4.88 ± 0.38 5.42 ± 1.44a Fasting insulin, μIU/mL 6.49 ± 1.99 8.91 ± 2.00 17.31 ± 5.91a,b HOMA-IR 1.30 ± 0.45 1.91 ± 0.39 4.18 ± 1.90a,b QUICKI 0.37 ± 0.03 0.35 ± 0.01a 0.31 ± 0.01a,b AMH, ng/mL 5.29 ± 1.44 12.99 ± 4.15a 10.94 ± 5.04a Frozen embryo transfer cycle 27 24 24 Implantation rate 50.0% ± 46.0% 37.5% ± 39.7% 27.1% ± 39.0% Non-PCOS (n = 18) PCOS Without IR (n = 18) PCOS With IR (n = 18) Age, y 27.67 ± 2.74 28.28 ± 3.39 28.56 ± 2.20 BMI, kg/m2 20.84 ± 2.19 23.54 ± 4.48 25.79 ± 3.84a Basal FSH, mIU/mL 7.13 ± 1.71 6.04 ± 1.23 6.13 ± 1.17 Basal LH, mIU/mL 4.87 ± 1.67 6.88 ± 5.77a 9.66 ± 7.36a Basal E2, pg/mL 42.40 ± 17.94 38.68 ± 20.68 42.24 ± 17.45 Basal T, nmol/L 1.01 ± 0.58 1.66 ± 0.67a 1.84 ± 0.83a Hormones on hCG day  LH, mIU/mL 2.04 ± 1.69 1.48 ± 0.85 2.13 ± 1.52  E2, pg/mL 2988.44 ± 1613.16 3992.64 ± 1825.50 3157.94 ± 1436.53  P4, ng/mL 0.92 ± 0.27 1.29 ± 0.67 1.21 ± 1.14 Fasting glucose, mmol/L 4.46 ± 0.37 4.88 ± 0.38 5.42 ± 1.44a Fasting insulin, μIU/mL 6.49 ± 1.99 8.91 ± 2.00 17.31 ± 5.91a,b HOMA-IR 1.30 ± 0.45 1.91 ± 0.39 4.18 ± 1.90a,b QUICKI 0.37 ± 0.03 0.35 ± 0.01a 0.31 ± 0.01a,b AMH, ng/mL 5.29 ± 1.44 12.99 ± 4.15a 10.94 ± 5.04a Frozen embryo transfer cycle 27 24 24 Implantation rate 50.0% ± 46.0% 37.5% ± 39.7% 27.1% ± 39.0% All data are mean ± standard deviation values. Abbreviations: AMH, anti-Müllerian hormone; BMI, body mass index; E2, estradiol; FSH, follicle-stimulating hormone; LH, luteinizing hormone; P4, progesterone; QUICKI, quantitative insulin sensitivity check index; T, testosterone. a P < 0.05 vs non-PCOS. b P < 0.05 vs PCOS without IR. View Large Cortisol and cortisone concentrations in human endometrial tissues No statistically significant differences were found in the summed concentrations of cortisol plus cortisone among the three groups [Fig. 1(a)]. Concentrations of cortisol in PCOS with IR patients were significantly higher compared with non-PCOS and PCOS without IR patients [Fig. 1(b)]. Concentrations of cortisone in PCOS with IR patients were significantly lower than those in non-PCOS patients and PCOS without IR patients [Fig. 1(c)]. Ratios of cortisol to cortisone were significantly elevated in PCOS with IR patients compared with non-PCOS and PCOS without IR patients [Fig. 1(d)]. The quantitative detection showed an imbalanced metabolic state between cortisol and cortisone in the endometria of PCOS with IR patients. Figure 1. View largeDownload slide The abundance of cortisol, cortisone, and 11β-HSDs in endometrium of non-PCOS, PCOS without IR, and PCOS with IR. (a–d) Endometrial concentrations of (a) cortisol plus cortisone, (b) cortisol, (c) cortisone, and (d) ratio of cortisol to cortisone in endometrial biopsies obtained from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; **P < 0.01; ***P < 0.001. (e) The representative blot of 11β-HSD1 and 11β-HSD2 in endometrial biopsies from non-PCOS patients (n = 7), PCOS without IR patients (n = 7), and PCOS with IR patients (n = 7). (f and g) Quantification of 11β-HSD 1 and 11β-HSD2 in endometrial biopsies from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; ***P < 0.001 vs non-PCOS; #P < 0.05 vs PCOS without IR. (h and i) Correlation of the abundance of 11β-HSD2 mRNA with (h) local cortisone and (i) cortisol in endometrial biopsies. Circles represent data points for non-PCOS (n = 18); squares represent data points for PCOS without IR (n = 18); triangles represent data points for PCOS with IR (n = 18). Data are mean ± SD values. Figure 1. View largeDownload slide The abundance of cortisol, cortisone, and 11β-HSDs in endometrium of non-PCOS, PCOS without IR, and PCOS with IR. (a–d) Endometrial concentrations of (a) cortisol plus cortisone, (b) cortisol, (c) cortisone, and (d) ratio of cortisol to cortisone in endometrial biopsies obtained from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; **P < 0.01; ***P < 0.001. (e) The representative blot of 11β-HSD1 and 11β-HSD2 in endometrial biopsies from non-PCOS patients (n = 7), PCOS without IR patients (n = 7), and PCOS with IR patients (n = 7). (f and g) Quantification of 11β-HSD 1 and 11β-HSD2 in endometrial biopsies from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; ***P < 0.001 vs non-PCOS; #P < 0.05 vs PCOS without IR. (h and i) Correlation of the abundance of 11β-HSD2 mRNA with (h) local cortisone and (i) cortisol in endometrial biopsies. Circles represent data points for non-PCOS (n = 18); squares represent data points for PCOS without IR (n = 18); triangles represent data points for PCOS with IR (n = 18). Data are mean ± SD values. 11β-HSD1 and 11β-HSD2 mRNA and protein abundance in endometrial tissues No significant difference in 11β-HSD1 mRNA and protein was observed among the three groups [Fig. 1(e) and 1(f)], but both mRNA and protein abundance of 11β-HSD2 in the endometria of PCOS with IR patients were significantly decreased in comparison with those in non-PCOS and PCOS without IR patients [Fig. 1(e) and 1(g)]. Consistently, Pearson analysis showed that 11β-HSD2 mRNA level in endometrium was positively correlated with endometrial cortisone level but negatively correlated with endometrial cortisol level [Fig. 1(h) and 1(i)]. These data suggested that the decrease in 11β-HSD2 expression may account for the increased cortisol level and decreased cortisone level in the endometria obtained from PCOS with IR patients. The abundance of IRS-1, phosphorylated IRS-1, Akt, phosphorylated Akt, and PTEN in endometrium Quantitative Western blotting revealed that IRS-1, p-IRS-1 Ser307, and Akt protein level were comparable among three groups [Fig. 2(a)–2(c) and 2(f)]. The phosphorylation of IRS-1 at Ser318 was significantly higher in the PCOS with IR group compared with non-PCOS [Fig. 2(a) and 2(d)], and the phosphorylation of IRS-1 at Ser612 was significantly higher in PCOS with IR group compared with non-PCOS and PCOS without IR [Fig. 2(a) and 2(e)]. Akt phosphorylation at Ser473 was significantly lower in PCOS with IR patients compared with non-PCOS and PCOS without IR [Fig. 2(a) and 2(g)]. Furthermore, PTEN mRNA and protein levels were significantly increased in PCOS with IR patients [Fig. 2(h) and 2(i)]. These data suggested that the insulin signaling pathway was abolished in the endometria of PCOS with IR patients. Figure 2. View largeDownload slide The abundance of IRS-1, p-IRS-1 Ser307, p-IRS-1 Ser318, p-IRS-1 Ser612, Akt, p-Akt Ser473, and PTEN in endometrium. (a) The representative blot of the protein abundance of IRS-1, p-IRS-1 Ser307, p-IRS-1 Ser318, p-IRS-1 Ser612, Akt, p-Akt Ser473, and PTEN in endometria from non-PCOS (n = 7), PCOS without IR (n = 7), and PCOS with IR patients (n = 7). (b–h) Quantification of the Western blotting assays of (b) IRS-1, (c) p-IRS-1 Ser307, (d) p-IRS-1 Ser318, (e) p-IRS-1 Ser612, (f) Akt, (g) p-Akt Ser473, and (h) PTEN in endometria from non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR patients (n = 18). (i) The mRNA level of PTEN in endometria from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; **P < 0.01, ***P < 0.001. Data are mean ± SD values. Figure 2. View largeDownload slide The abundance of IRS-1, p-IRS-1 Ser307, p-IRS-1 Ser318, p-IRS-1 Ser612, Akt, p-Akt Ser473, and PTEN in endometrium. (a) The representative blot of the protein abundance of IRS-1, p-IRS-1 Ser307, p-IRS-1 Ser318, p-IRS-1 Ser612, Akt, p-Akt Ser473, and PTEN in endometria from non-PCOS (n = 7), PCOS without IR (n = 7), and PCOS with IR patients (n = 7). (b–h) Quantification of the Western blotting assays of (b) IRS-1, (c) p-IRS-1 Ser307, (d) p-IRS-1 Ser318, (e) p-IRS-1 Ser612, (f) Akt, (g) p-Akt Ser473, and (h) PTEN in endometria from non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR patients (n = 18). (i) The mRNA level of PTEN in endometria from non-PCOS patients (n = 18), PCOS without IR patients (n = 18), and PCOS with IR patients (n = 18). *P < 0.05; **P < 0.01, ***P < 0.001. Data are mean ± SD values. Impaired glucose uptake and GLUT4 translocation in EECs derived from PCOS-IR patients Immunohistochemistry revealed strong staining of 11β-HSD2 on surface and glandular epithelial cells and low-intensity staining in stromal cells [Fig. 3(a)]. This suggested that the reduction of cortisol occurred mainly in EECs. The isolated EECs were identified by immunofluorescence staining with cytokeratin 7 before further experiments [Fig. 3(b)]. To clarify the glucose uptake capacity in EECs in PCOS patients, we obtained EECs from non-PCOS, PCOS without IR, and PCOS with IR patients and measured the GLUT4 translocation and glucose uptake under insulin stimulation. Insulin could stimulate both GLUT4 translocation from cytoplasm to membrane [Fig. 3(c) and 3(d)] and glucose uptake [Fig. 3(e)] among three groups, but the GLUT4 translocation and glucose uptake were diminished in EECs from PCOS with IR patients compared with EECs from non-PCOS patients [Fig. 3(c)–3(e)]. These data suggested that glucose uptake and GLUT4 translocation capacity were impaired in PCOS with IR patients. Figure 3. View largeDownload slide The GLUT4 translocation and glucose uptake in EECs derived from non-PCOS, PCOS without IR, and PCOS with IR patients. (a) Immunohistochemical staining of 11β-HSD2 in human uterine endometrium. (b) Immunofluorescence staining of cytokeratin 7 (CK7) (red) and Vimentin (green). The nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI) (blue). (c) The representative blot of GLUT4 translocation from cytoplasm to membrane in EECs derived from non-PCOS, PCOS without IR, and PCOS with IR endometria. Na+K+ATPase is used as a housekeeping protein located in membrane. (d) Quantification of the Western blotting assays of GLUT4 translocation from cytoplasm to membrane (membrane GLUT4/cytoplasm GLUT4) in non-PCOS (n = 4), PCOS without IR (n = 4), and PCOS with IR (n = 4). (e) Fold change of glucose uptake in non-PCOS (n = 4), PCOS without IR (n = 4), and PCOS with IR (n = 4). *P < 0.05; **P < 0.01. Data are mean ± SD values. Figure 3. View largeDownload slide The GLUT4 translocation and glucose uptake in EECs derived from non-PCOS, PCOS without IR, and PCOS with IR patients. (a) Immunohistochemical staining of 11β-HSD2 in human uterine endometrium. (b) Immunofluorescence staining of cytokeratin 7 (CK7) (red) and Vimentin (green). The nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI) (blue). (c) The representative blot of GLUT4 translocation from cytoplasm to membrane in EECs derived from non-PCOS, PCOS without IR, and PCOS with IR endometria. Na+K+ATPase is used as a housekeeping protein located in membrane. (d) Quantification of the Western blotting assays of GLUT4 translocation from cytoplasm to membrane (membrane GLUT4/cytoplasm GLUT4) in non-PCOS (n = 4), PCOS without IR (n = 4), and PCOS with IR (n = 4). (e) Fold change of glucose uptake in non-PCOS (n = 4), PCOS without IR (n = 4), and PCOS with IR (n = 4). *P < 0.05; **P < 0.01. Data are mean ± SD values. Cortisol attenuates insulin-stimulated glucose uptake by inhibition of Akt phosphorylation and GLUT4 translocation via induction of PTEN expression in EECs Prior treatment with cortisol (1 μM, 24 hours) could attenuate the insulin-induced glucose uptake [Fig. 4(a)], Akt phosphorylation [Fig. 4(b)], and GLUT4 translocation from cytoplasm to membrane [Fig. 4(c)] in EECs derived from non-PCOS patients. The cortisol-mediated impairment of glucose uptake, Akt phosphorylation, and GLUT4 translocation were also observed in EECs derived from PCOS patients [Fig. 4(d)–4(f)]. Figure 4. View largeDownload slide The effects of cortisol on glucose uptake and insulin signaling in EECs derived from non-PCOS and PCOS patients. (a) Insulin-stimulated glucose uptake, and the effects of cortisol on insulin-stimulated glucose uptake in EECs from non-PCOS patients (n = 4). (b) Effects of cortisol on Akt phosphorylation (Ser473) in EECs from non-PCOS patients (n = 5). (c) Effects of cortisol on insulin-stimulated GLUT4 translocation in EECs from non-PCOS patients (n = 4). (d–f) Effects of cortisol on insulin-stimulated glucose uptake (n= 4), Akt phosphorylation (n = 5), and GLUT4 translocation (n = 4) in EECs from PCOS patients. *P < 0.05; **P < 0.01; ***P < 0.001 vs control without insulin and cortisol; #P < 0.05; ##P < 0.01; ###P < 0.001 vs insulin alone. Data are mean ± SD with representative blots. Figure 4. View largeDownload slide The effects of cortisol on glucose uptake and insulin signaling in EECs derived from non-PCOS and PCOS patients. (a) Insulin-stimulated glucose uptake, and the effects of cortisol on insulin-stimulated glucose uptake in EECs from non-PCOS patients (n = 4). (b) Effects of cortisol on Akt phosphorylation (Ser473) in EECs from non-PCOS patients (n = 5). (c) Effects of cortisol on insulin-stimulated GLUT4 translocation in EECs from non-PCOS patients (n = 4). (d–f) Effects of cortisol on insulin-stimulated glucose uptake (n= 4), Akt phosphorylation (n = 5), and GLUT4 translocation (n = 4) in EECs from PCOS patients. *P < 0.05; **P < 0.01; ***P < 0.001 vs control without insulin and cortisol; #P < 0.05; ##P < 0.01; ###P < 0.001 vs insulin alone. Data are mean ± SD with representative blots. In addition, PTEN mRNA and protein levels were induced by cortisol in cultured EECs obtained from non-PCOS patients [Fig. 5(a)]. Treatment of EECs with the PTEN inhibitor bPV (phen) rescued the cortisol-induced suppression of Akt phosphorylation [Fig. 5(b)], as well as GLUT4 translocation [Fig. 5(c)]. Furthermore, Pearson analysis showed that cortisol levels positively, but 11β-HSD2 abundance negatively, correlated with PTEN mRNA [Fig. 5(d) and 5(e)]. These data suggested that cortisol attenuated insulin-stimulated glucose uptake via the suppression of Akt phosphorylation and GLUT4 translocation in EECs from non-PCOS and PCOS patients. PTEN was involved in cortisol-induced attenuation of Akt phosphorylation and GLUT4 translocation. Figure 5. View largeDownload slide The involvement of PTEN in the effects of cortisol in EECs. (a) Effects of cortisol on PTEN mRNA and protein abundance in EECs obtained from non-PCOS patients (n = 4). *P < 0.05 and ***P < 0.001 vs control (cortisol = 0). (b) The amount of insulin-stimulated phosphorylated Akt in response to cortisol in the presence or absence of bPV(phen), the PTEN inhibitor (n = 4). **P < 0.01 vs control without cortisol and bPV; ###P < 0.001 vs cortisol alone. (c) The amount of insulin-stimulated GLUT4 in membrane and cytoplasm in response to cortisol in the presence or absence of bPV(phen) (n = 4). *P < 0.05 vs membrane control without insulin and bPV; #P < 0.05 vs membrane cortisol alone. (d and e) Correlation of PTEN mRNA with (d) cortisol and (e) 11β-HSD2 in non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR patients (n = 18). Circles represent data points for non-PCOS; squares represent data points for PCOS without IR; triangles represent data points for PCOS with IR. Data are mean ± SD with representative blots. Figure 5. View largeDownload slide The involvement of PTEN in the effects of cortisol in EECs. (a) Effects of cortisol on PTEN mRNA and protein abundance in EECs obtained from non-PCOS patients (n = 4). *P < 0.05 and ***P < 0.001 vs control (cortisol = 0). (b) The amount of insulin-stimulated phosphorylated Akt in response to cortisol in the presence or absence of bPV(phen), the PTEN inhibitor (n = 4). **P < 0.01 vs control without cortisol and bPV; ###P < 0.001 vs cortisol alone. (c) The amount of insulin-stimulated GLUT4 in membrane and cytoplasm in response to cortisol in the presence or absence of bPV(phen) (n = 4). *P < 0.05 vs membrane control without insulin and bPV; #P < 0.05 vs membrane cortisol alone. (d and e) Correlation of PTEN mRNA with (d) cortisol and (e) 11β-HSD2 in non-PCOS (n = 18), PCOS without IR (n = 18), and PCOS with IR patients (n = 18). Circles represent data points for non-PCOS; squares represent data points for PCOS without IR; triangles represent data points for PCOS with IR. Data are mean ± SD with representative blots. Discussion To our knowledge, this study is the first to evaluate endometrial cortisol, cortisone, and 11β-HSDs levels in PCOS patients. Endometria from PCOS with IR patients had increased cortisol, decreased cortisone, and diminished 11β-HSD2 in the WOI phase compared with those from non-PCOS and PCOS without IR patients. PCOS with IR patients also exhibited endometrial insulin signaling pathway defects and relatively lower implantation rates. Further in vitro studies clarified that excessive cortisol might attenuate insulin sensitivity by decreasing the phosphorylation of Akt (Ser473) and the translocation of GLUT4 via induction of PTEN expression. Thus, this study uncovered the imbalanced state of cortisol and cortisone in endometrium of PCOS with IR patients and established a correlation between the elevated local cortisol and endometrial IR. With a prevalence of 44% to 70%, IR is common among PCOS patients. IR is traditionally defined as the insensitivity or unresponsiveness to insulin and thus requires an increased insulin level (4). In addition to systemic dysfunction of hyperinsulinemia, IR also exhibits as peripheral IR affecting insulin target tissue including endometrium. Li et al. (24) found that maternal hyperinsulinemia impaired mouse endometrial receptivity in early pregnancy. Chang et al. (10) found that PCOS patients with IR had compromised implantation rates, suggesting the effects of IR on endometrial function and receptivity. Consistently, we also observed a lower implantation rate in PCOS with IR patients, although it was not statistically significant, possibly due to the small number of patients in our study. All the evidence suggested that IR might play an important role on human endometrial receptivity in PCOS. Successful implantation requires the endometrium to undergo changes and be receptive to embryos in a short period known as the WOI phase (25). This demands a large amount of energy, mainly from glucose uptake, for the endometrium to properly differentiate to a receptive state (26, 27). Therefore, local endometrial IR might diminish the glucose uptake in endometrial cells and interfere with endometrial receptivity. The metabolic actions of insulin are mainly mediated through activation of IRS-1, phosphatidylinositol-3,4,5-trisphosphate, and Akt, resulting in the translocation of GLUT4 from intracellular vesicles to the plasma membrane. Defects in this signaling pathway could induce IR (4). Previous studies have reported the diminished endometrial IRS-1 in PCOS with hyperinsulinemia and increased endometrial PTEN expression in PCOS (8, 9, 28). In this study, we also revealed significantly elevated PTEN and relatively decreased IRS-1 in PCOS with IR patients. Furthermore, we demonstrated that IRS-1 phosphorylation at Ser318 and Ser612 was increased and Akt phosphorylation at Ser473 was diminished in the endometria from PCOS with IR patients. These evidences further clarified the defective insulin signaling pathway in the endometria from PCOS with IR patients. Because serine phosphorylation of IRS-1 acts as a negative feedback signal for insulin effect, we provided supporting evidence for the former study of Fornes et al. (9) that the IRS-1 activating tyrosine phosphorylation (Y612) was decreased in endometria of PCOS. In addition, the diminished GLUT4 translocation and glucose uptake in the EECs of PCOS with IR patients further clarified the local IR in endometrium. Excessive glucocorticoid exposure has correlated with whole-body IR for decades, and the local effect of cortisol is exaggerated in insulin-target tissue (29, 30). It has been documented that excessive cortisol generated by 11β-HSD1 contributes to the development of IR in adipose, skeletal muscle, and granulosa cells (19, 31, 32). However, quantitative measurements of cortisol and its metabolic enzymes in the endometria of PCOS women have not been performed previously. In endometrium, 11β-HSD1 expression was low except during the menstrual and decidua phases, whereas 11β-HSD2 was the codominant metabolic glucocorticoid enzyme in the secretory phase (13, 14). Here we detailed the elevated cortisol and ratios of cortisol to cortisone in the endometria obtained from PCOS with IR patients. This might reflect either an increase in local reduction of cortisone or a decrease in local oxidation of cortisol. Upon further exploration, we found a significant decrease of 11β-HSD2 in the endometria from PCOS with IR patients, whereas 11β-HSD1 was comparable among the three groups. Furthermore, 11β-HSD2 was negatively correlated with cortisol and positively correlated with cortisone levels. Based on these data, we attributed the elevated ratios of cortisol to cortisone in the endometria of PCOS with IR patients mainly to the decreased local cortisol oxidation caused by 11β-HSD2. This was consistent with the codominant role for 11β-HSD2 in secretory phase. An optimal amount of cortisol is indispensable to embryo implantation (33). However, excessive cortisol caused by decreased oxidation of 11β-HSD2 may disrupt glucocorticoid homeostasis and impair insulin sensitivity in endometrium. As embryo implantation is initiated by embryo attachment to endometrial epithelium (34) and 11β-HSD2 is mainly expressed in endometrial epithelium, we focused on the effects of cortisol on insulin sensitivity in EECs. In the insulin-stimulated glucose uptake pathway, GLUT4 translocation is critical and mediated by Akt phosphorylation. Defects in GLUT4 translocation and Akt phosphorylation indicated IR (35, 36). In this study, we demonstrated that glucose uptake, Akt phosphorylation, and exocytosis of GLUT4 to the plasma membrane were inhibited by cortisol in EECs from both non-PCOS and PCOS patients. This suggested that cortisol was involved in endometrial IR. It is understood that PTEN negatively regulates Akt activity by dephosphorylating phosphatidylinositol-3,4,5-trisphosphate (37, 38) and the induction of PTEN expression could attenuate insulin sensitivity (37, 39). Our in vitro study revealed that PTEN was induced by cortisol in EECs. Additionally, the inhibitor of PTEN restored the cortisol-caused depression of Akt phosphorylation and GLUT4 translocation, suggesting a crucial role of PTEN in cortisol-stimulated endometrial IR. Because PTEN expression in endometrium was also closely correlated with cortisol level, PTEN might be recognized as an important contributor to cortisol-induced IR in endometrium. In this study, we intended to compare the endometria from non-PCOS, PCOS without IR, and PCOS with IR, but we did not exclude the possibility that glucocorticoid alteration might also exist in non-PCOS IR patients. Furthermore, the causes of diminished 11β-HSD2 and the specific underlying molecular mechanisms involving PTEN require further investigation. Our study provides vital preliminary evidence for direct future research on cortisol-induced endometrial IR. In conclusion, we demonstrated that PCOS with IR patients had increased cortisol, diminished 11β-HSD2, and impaired insulin sensitivity in endometrium in the WOI phase compared with non-PCOS and PCOS without IR patients. This indicated that a decreased local inactivation of cortisol by 11β-HSD2 might be a possible cause of endometrial IR in PCOS. Our in vitro study suggested a detrimental role of cortisol in insulin sensitivity in endometrium. We propose that maintaining cortisol at an optimal level in PCOS patients might be beneficial for embryo implantation. Abbreviations: Abbreviations: 11β-HSD 11β-hydroxysteroid dehydrogenase EEC endometrial epithelial cell GLUT4 glucose transporter type 4 hCG human chorionic gonadotropin HOMA-IR homeostasis model assessment of insulin resistance index IR insulin resistance IRS-1 insulin receptor substrate 1 mRNA messenger RNA PCOS polycystic ovary syndrome PCR polymerase chain reaction PTEN phosphatase and tensin homolog deleted on chromosome ten WOI window of implantation Acknowledgments We thank Huiliang Xie, Xiaoming Zhao, Yan Hong, and Minzhi Gao for help with patient recruitment and endometrial biopsies collection and Jianjun Liu, Xiaoping Zhao, Li Zhao, and Panli Li for skillful technical support. Financial Support: This work was supported by National Natural Science Foundation of China Grant 81771648 (to Y.S.), National Key R&D Program of China Grant 2017YFC1001403 (to Y.S.), National Natural Science Foundation of China Grant 81571499 (to Y.S.), Chinese National Key Basic Research Projects Grant 2014CB943300 (to Y.S.), Program of Shanghai Academic Research Leader in Shanghai Municipal Commission of Health and Family Planning Grant 2017BR015 (to Y.S.), Clinical Skills Improvement Project of Major Disorders Hospital Development Center of Shanghai Grant 16CR1022A (to Y.S.), Shanghai Technological Innovation Plan Grant 18140902400 (to Y.S.), and Shanghai Municipal Education Commission-Gaofeng Clinical Medicine Grant 20161413 (to Y.S.). Disclosure Summary: The authors have nothing to disclose. References 1. Azziz R , Carmina E , Chen Z , Dunaif A , Laven JS , Legro RS , Lizneva D , Natterson-Horowtiz B , Teede HJ , Yildiz BO . Polycystic ovary syndrome . Nat Rev Dis Primers . 2016 ; 2 : 16057 . 2. Rotterdam E ; Rotterdam ESHRE/ASRM-Sponsored PCOS consensus workshop group . Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS) . Hum Reprod . 2004 ; 19 ( 1 ): 41 – 47 . 3. Dunaif A , Segal KR , Futterweit W , Dobrjansky A . Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome . Diabetes . 1989 ; 38 ( 9 ): 1165 – 1174 . 4. Diamanti-Kandarakis E , Dunaif A . Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications . Endocr Rev . 2012 ; 33 ( 6 ): 981 – 1030 . 5. Bellver J , Martínez-Conejero JA , Labarta E , Alamá P , Melo MA , Remohí J , Pellicer A , Horcajadas JA . Endometrial gene expression in the window of implantation is altered in obese women especially in association with polycystic ovary syndrome . Fertil Steril . 2011 ; 95 ( 7 ): 2335 – 2341, 2341.e1–2341.e8 . 6. Boomsma CM , Eijkemans MJ , Hughes EG , Visser GH , Fauser BC , Macklon NS . A meta-analysis of pregnancy outcomes in women with polycystic ovary syndrome . Hum Reprod Update . 2006 ; 12 ( 6 ): 673 – 683 . 7. Giudice LC . Endometrium in PCOS: implantation and predisposition to endocrine CA . Best Pract Res Clin Endocrinol Metab . 2006 ; 20 ( 2 ): 235 – 244 . 8. Shafiee MN , Seedhouse C , Mongan N , Chapman C , Deen S , Abu J , Atiomo W . Up-regulation of genes involved in the insulin signalling pathway (IGF1, PTEN and IGFBP1) in the endometrium may link polycystic ovarian syndrome and endometrial cancer . Mol Cell Endocrinol . 2016 ; 424 : 94 – 101 . 9. Fornes R , Ormazabal P , Rosas C , Gabler F , Vantman D , Romero C , Vega M . Changes in the expression of insulin signaling pathway molecules in endometria from polycystic ovary syndrome women with or without hyperinsulinemia . Mol Med . 2010 ; 16 ( 3-4 ): 129 – 136 . 10. Chang EM , Han JE , Seok HH , Lee DR , Yoon TK , Lee WS . Insulin resistance does not affect early embryo development but lowers implantation rate in in vitro maturation-in vitro fertilization-embryo transfer cycle . Clin Endocrinol (Oxf) . 2013 ; 79 ( 1 ): 93 – 99 . 11. Haoula Z , Salman M , Atiomo W . Evaluating the association between endometrial cancer and polycystic ovary syndrome . Hum Reprod . 2012 ; 27 ( 5 ): 1327 – 1331 . 12. Fearnley EJ , Marquart L , Spurdle AB , Weinstein P , Webb PM ; Australian Ovarian Cancer Study Group and Australian National Endometrial Cancer Study Group . Polycystic ovary syndrome increases the risk of endometrial cancer in women aged less than 50 years: an Australian case-control study . Cancer Causes Control . 2010 ; 21 ( 12 ): 2303 – 2308 . 13. McDonald SE , Henderson TA , Gomez-Sanchez CE , Critchley HO , Mason JI . 11Beta-hydroxysteroid dehydrogenases in human endometrium . Mol Cell Endocrinol . 2006 ; 248 ( 1-2 ): 72 – 78 . 14. Smith RE , Salamonsen LA , Komesaroff PA , Li KX , Myles KM , Lawrence M , Krozowski Z . 11Beta-hydroxysteroid dehydrogenase type II in the human endometrium: localization and activity during the menstrual cycle . J Clin Endocrinol Metab . 1997 ; 82 ( 12 ): 4252 – 4257 . 15. Chapman K , Holmes M , Seckl J . 11β-Hydroxysteroid dehydrogenases: intracellular gate-keepers of tissue glucocorticoid action . Physiol Rev . 2013 ; 93 ( 3 ): 1139 – 1206 . 16. Shabir I , Ganie MA , Praveen EP , Khurana ML , John J , Gupta N , Kumar G , Ammini AC . Morning plasma cortisol is low among obese women with polycystic ovary syndrome . Gynecol Endocrinol . 2013 ; 29 ( 12 ): 1045 – 1047 . 17. Michael AE , Glenn C , Wood PJ , Webb RJ , Pellatt L , Mason HD . Ovarian 11β-hydroxysteroid dehydrogenase (11βHSD) activity is suppressed in women with anovulatory polycystic ovary syndrome (PCOS): apparent role for ovarian androgens . J Clin Endocrinol Metab . 2013 ; 98 ( 8 ): 3375 – 3383 . 18. Li S , Tao T , Wang L , Mao X , Zheng J , Zhao A , Liu W . The expression of 11β-HSDs, GR, and H6PDH in subcutaneous adipose tissue from polycystic ovary syndrome subjects . Horm Metab Res . 2013 ; 45 ( 11 ): 802 – 807 . 19. Zhu Q , Zuo R , He Y , Wang Y , Chen ZJ , Sun Y , Sun K . Local regeneration of cortisol by 11β-HSD1 contributes to insulin resistance of the granulosa cells in PCOS . J Clin Endocrinol Metab . 2016 ; 101 ( 5 ): 2168 – 2177 . 20. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group . Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome . Fertil Steril . 2004 ; 81 ( 1 ): 19 – 25 . 21. Alebić MS , Bulum T , Stojanović N , Duvnjak L . Definition of insulin resistance using the homeostasis model assessment (HOMA-IR) in IVF patients diagnosed with polycystic ovary syndrome (PCOS) according to the Rotterdam criteria . Endocrine . 2014 ; 47 ( 2 ): 625 – 630 . 22. Li J , Wang W , Liu C , Wang W , Li W , Shu Q , Chen ZJ , Sun K . Critical role of histone acetylation by p300 in human placental 11β-HSD2 expression . J Clin Endocrinol Metab . 2013 ; 98 ( 7 ): E1189 – E1197 . 23. Kirk D , King RJ , Heyes J , Peachey L , Hirsch PJ , Taylor RW . Normal human endometrium in cell culture. I. Separation and characterization of epithelial and stromal components in vitro . In Vitro . 1978 ; 14 ( 8 ): 651 – 662 . 24. Li R , Wu J , He J , Wang Y , Liu X , Chen X , Tong C , Ding Y , Su Y , Chen W , Zhang C , Gao R . Mice endometrium receptivity in early pregnancy is impaired by maternal hyperinsulinemia . Mol Med Rep . 2017 ; 15 ( 5 ): 2503 – 2510 . 25. Schulte MM , Tsai JH , Moley KH . Obesity and PCOS: the effect of metabolic derangements on endometrial receptivity at the time of implantation . Reprod Sci . 2015 ; 22 ( 1 ): 6 – 14 . 26. von Wolff M , Ursel S , Hahn U , Steldinger R , Strowitzki T . Glucose transporter proteins (GLUT) in human endometrium: expression, regulation, and function throughout the menstrual cycle and in early pregnancy . J Clin Endocrinol Metab . 2003 ; 88 ( 8 ): 3885 – 3892 . 27. Medina RA , Meneses AM , Vera JC , Gúzman C , Nualart F , Rodriguez F , de los Angeles Garcia M , Kato S , Espinoza N , Monsó C , Carvajal A , Pinto M , Owen GI . Differential regulation of glucose transporter expression by estrogen and progesterone in Ishikawa endometrial cancer cells . J Endocrinol . 2004 ; 182 ( 3 ): 467 – 478 . 28. Oróstica L , Rosas C , Plaza-Parrochia F , Astorga I , Gabler F , García V , Romero C , Vega M . Altered steroid metabolism and insulin signaling in PCOS endometria: impact in tissue function . Curr Pharm Des . 2016 ; 22 ( 36 ): 5614 – 5624 . 29. Geer EB , Islam J , Buettner C . Mechanisms of glucocorticoid-induced insulin resistance: focus on adipose tissue function and lipid metabolism . Endocrinol Metab Clin North Am . 2014 ; 43 ( 1 ): 75 – 102 . 30. Morgan SA , Sherlock M , Gathercole LL , Lavery GG , Lenaghan C , Bujalska IJ , Laber D , Yu A , Convey G , Mayers R , Hegyi K , Sethi JK , Stewart PM , Smith DM , Tomlinson JW . 11Beta-hydroxysteroid dehydrogenase type 1 regulates glucocorticoid-induced insulin resistance in skeletal muscle . Diabetes . 2009 ; 58 ( 11 ): 2506 – 2515 . 31. Dube S , Slama MQ , Basu A , Rizza RA , Basu R . Glucocorticoid excess increases hepatic 11β-HSD-1 activity in humans: implications in steroid-induced diabetes . J Clin Endocrinol Metab . 2015 ; 100 ( 11 ): 4155 – 4162 . 32. Brands M , van Raalte DH , João Ferraz M , Sauerwein HP , Verhoeven AJ , Aerts JM , Diamant M , Serlie MJ . No difference in glycosphingolipid metabolism and mitochondrial function in glucocorticoid-induced insulin resistance in healthy men . J Clin Endocrinol Metab . 2013 ; 98 ( 3 ): 1219 – 1225 . 33. Whirledge SD , Oakley RH , Myers PH , Lydon JP , DeMayo F , Cidlowski JA . Uterine glucocorticoid receptors are critical for fertility in mice through control of embryo implantation and decidualization . Proc Natl Acad Sci USA . 2015 ; 112 ( 49 ): 15166 – 15171 . 34. Aplin JD , Ruane PT . Embryo-epithelium interactions during implantation at a glance . J Cell Sci . 2017 ; 130 ( 1 ): 15 – 22 . 35. Hribal ML , Federici M , Porzio O , Lauro D , Borboni P , Accili D , Lauro R , Sesti G . The Gly-->Arg972 amino acid polymorphism in insulin receptor substrate-1 affects glucose metabolism in skeletal muscle cells . J Clin Endocrinol Metab . 2000 ; 85 ( 5 ): 2004 – 2013 . 36. Leto D , Saltiel AR . Regulation of glucose transport by insulin: traffic control of GLUT4 . Nat Rev Mol Cell Biol . 2012 ; 13 ( 6 ): 383 – 396 . 37. Song MS , Salmena L , Pandolfi PP . The functions and regulation of the PTEN tumour suppressor . Nat Rev Mol Cell Biol . 2012 ; 13 ( 5 ): 283 – 296 . 38. He X , Saji M , Radhakrishnan D , Romigh T , Ngeow J , Yu Q , Wang Y , Ringel MD , Eng C . PTEN lipid phosphatase activity and proper subcellular localization are necessary and sufficient for down-regulating AKT phosphorylation in the nucleus in Cowden syndrome . J Clin Endocrinol Metab . 2012 ; 97 ( 11 ): E2179 – E2187 . 39. Gupta A , Dey CS . PTEN, a widely known negative regulator of insulin/PI3K signaling, positively regulates neuronal insulin resistance . Mol Biol Cell . 2012 ; 23 ( 19 ): 3882 – 3898 . Copyright © 2018 Endocrine Society

Sugiyama, Toshihiro

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00886pmid: 29771348

Recent results from the Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months (FREEDOM) trial and its extension regarding bone histology, histomorphometry, and matrix mineralization (1), as well as areal bone mineral density (BMD), fracture incidence, and adverse events (2), showed efficacy and safety of long-term denosumab treatment of osteoporosis in postmenopausal women. Interestingly, areal BMD at the lumbar spine, total hip, and femoral neck but not the one-third radius continuously increased for up to 10 years (2). Based on data from iliac bone biopsy specimens that the mean degree of mineralization and the heterogeneity index were similar at 5 and 10 years (1), the authors suggested that modeling-based bone formation shown in adult monkeys treated with denosumab would be one possible mechanism, as previously discussed (2–4). Here I would like to provide further mechanistic insights into the authors’ suggestion. The above-mentioned continuous but site-specific effects of denosumab treatment on areal BMD (2) indicate that long-term efficacy of denosumab treatment can be expected in both trabecular and cortical bone compartments at the weight-bearing sites. This is likely to support modeling-based bone formation uncoupled with bone resorption because the primary determinant of bone modeling is elastic deformation (strain) of the skeleton engendered by physical activity (5). The noncontinuous effect on areal BMD at the non-weight-bearing radius (2) agrees with the similar bone mineralization characteristics at 5 and 10 years (1), which could further support that long-term efficacy of denosumab treatment is independent of remodeling-based coupled bone resorption and formation, although iliac crest analyzed is also a non-weight-bearing bone as the authors pointed out. Here it is important to note that modeling-based bone formation in the adult skeleton would generally require an increase in the level of mechanical strain itself and/or the resultant skeletal response, whereas an increase in the degree of bone mineralization caused by denosumab treatment of up to 5 years (1) acts to decrease the level of mechanical strain, and denosumab treatment is unlikely to directly enhance skeletal response to mechanical strain. We previously suggested that the increased level of mechanical strain might be expected through an increase in physical activity (6); the incidence of falls that did not cause fractures in the denosumab group (4.5%) was significantly lower than that in the placebo group (5.7%) in the FREEDOM trial. However, even if an increase in physical activity after denosumab treatment is possible, the increased level of mechanical strain alone is difficult to fully explain the long-term continuous increases in areal BMD because skeletal adaptation to change in mechanical strain is relatively rapid. The homeostatic system to maintain the level of mechanical strain in the skeleton indicates the negative feedback control against the decreased level of bone strain resulting from an increase in bone strength associated with osteoporosis therapy (7). Consequently, long-term continuous modeling-based bone formation could be theoretically realized if denosumab treatment can indirectly and gradually increase the skeletal response to mechanical strain during habitual physical activity. Considering the osteocyte lacunocanalicular system (8), long-term denosumab treatment might enhance mechanical strain–related stimuli by narrowing the lacunocanalicular space and increasing fluid flow. Abbreviation: Abbreviation: BMD bone mineral density Acknowledgments Disclosure Summary: The author has nothing to disclose. References 1. Dempster DW , Brown JP , Fahrleitner-Pammer A , Kendler D , Rizzo S , Valter I , Wagman RB , Yin X , Yue SV , Boivin G . Effects of long-term denosumab on bone histomorphometry and mineralization in women with postmenopausal osteoporosis . J Clin Endocrinol Metab . 2018 ; 103 ( 7 ): 2498 – 2509 . 2. Bone HG , Wagman RB , Brandi ML , Brown JP , Chapurlat R , Cummings SR , Czerwiński E , Fahrleitner-Pammer A , Kendler DL , Lippuner K , Reginster JY , Roux C , Malouf J , Bradley MN , Daizadeh NS , Wang A , Dakin P , Pannacciulli N , Dempster DW , Papapoulos S . 10 Years of denosumab treatment in postmenopausal women with osteoporosis: results from the phase 3 randomised FREEDOM trial and open-label extension . Lancet Diabetes Endocrinol . 2017 ; 5 ( 7 ): 513 – 523 . 3. Portal-Núñez S , Mediero A , Esbrit P , Sánchez-Pernaute O , Largo R , Herrero-Beaumont G . Unexpected bone formation produced by RANKL blockade . Trends Endocrinol Metab . 2017 ; 28 ( 10 ): 695 – 704 . 4. Dempster DW , Zhou H , Recker RR , Brown JP , Recknor CP , Lewiecki EM , Miller PD , Rao SD , Kendler DL , Lindsay R , Krege JH , Alam J , Taylor KA , Melby TE , Ruff VA . Remodeling- and modeling-based bone formation with teriparatide versus denosumab: a longitudinal analysis from baseline to 3 months in the AVA study . J Bone Miner Res . 2018 ; 33 ( 2 ): 298 – 306 . 5. Sugiyama T , Oda H . Osteoporosis therapy: bone modeling during growth and aging . Front Endocrinol (Lausanne) . 2017 ; 8 : 46 . 6. Sugiyama T , Kim YT , Oda H . A possible mechanism of denosumab treatment for fracture prevention . J Clin Endocrinol Metab . 2016 ; 101 ( 2 ): L15 – L16 . 7. Sugiyama T . Treatment of low bone density or osteoporosis to prevent fractures in men and women . Ann Intern Med . 2017 ; 167 ( 12 ): 899 – 900 . 8. Dallas SL , Prideaux M , Bonewald LF . The osteocyte: an endocrine cell ... and more . Endocr Rev . 2013 ; 34 ( 5 ): 658 – 690 . Copyright © 2018 Endocrine Society

Winer, Karen K

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02593pmid: 29897557

Hypoparathyroidism is a rare endocrine disorder treated with vitamin D analogs and calcium supplements. Although conventional therapy effectively raises serum calcium levels, it does not fully restore normal mineral homeostasis because it bypasses the PTH effects on the kidney and bone and relies entirely on calcium transport across the gastrointestinal tract to normalize blood calcium. Without the renal calcium-retaining effects of PTH, conventional therapy often leads to abnormally elevated urine calcium excretion and long-term complications, including decreased renal function and nephrocalcinosis (1, 2). Hypoparathyroidism is the only classic hormonal insufficiency state where hormonal replacement is not standard therapy. Recently, the full-length molecule, recombinant human PTH (rhPTH) 1–84, has emerged as a potential replacement therapy and is currently approved by the US Food and Drug Administration as an adjunct to conventional therapy for adult patients who are refractory to conventional therapy. Many adults with hypoparathyroidism treated with conventional therapy have health-related quality-of-life (HRQoL) impairments and lower quality-of-life (QoL) scores, using validated tools, compared with reference norms (3–9). Patients treated with conventional therapy have a wide range of complaints, including anxiety, depression, fatigue, muscle weakness, exercise intolerance, and various cognitive deficits (4, 5, 10–12). Controlled studies have examined the impact of PTH 1–84 therapy on QoL but have not demonstrated significant improvements compared with controls receiving exclusively conventional therapy (3, 7). Most studies of PTH effects on HRQoL use the Short Form Health Survey 36 (SF-36), a subjective, self-administered 36-item questionnaire that is divided into eight domains. Two summary measures, physical and mental health component scores, are calculated from the individual domain scores. The SF-36 measures QoL in chronic illness. This questionnaire does not adequately measure cognitive deficits, fatigue, physical endurance, or muscle strength. Therefore, no single instrument can exclusively characterize and precisely measure the burden of disease experienced by patients with hypoparathyroidism. HRQoL deficits, including increased fatigue, anxiety, and decreased muscle strength, may be a feature of PTH deficiency, but the ability to relieve these symptoms with PTH may depend on whether the physiologic replacement regimen adequately restores normal mineral levels with minimal fluctuation throughout the day. Although PTH 1–34 and PTH 1–84 have identical biological effects, studies of these two peptides take two divergent approaches to replacement therapy. From its very early investigative stages (13), synthetic human PTH (hPTH) 1–34 has been given in multiple subcutaneous (SC) injections titrated with small incremental dose changes to normalize both blood and urine calcium levels. This approach allows for lower individual doses and a more steady-state physiologic profile of serum and urine minerals, which is most evident in the later pump studies (14, 15). PTH 1–84, on the other hand, has been given as fixed daily or every-other-day doses as an adjunct to flexible doses of conventional therapy (16). Large, fixed PTH doses may be associated with transient hypercalcemia and hypercalciuria (17–20), which may produce symptoms associated with elevated calcium such as nausea, bone pain, poor concentration, and polyuria. Such recurrent symptoms, however transient, will contribute to decreased well-being. Two studies, which appeared in earlier 2018 issues of JCEM, by Vokes et al. (7) and Palermo et al. (8), explored HRQoL in adults with hypoparathyroidism in response to PTH 1–84 and rhPTH 1–34 therapy, respectively. Both studies used the fixed-dose approach to replacement therapy with little or no titration of the PTH dose. Vokes et al. (7) investigated the impact of PTH 1–84 on HRQoL as measured by the SF-36 during the 6-month double-blind, randomized, controlled multicenter study (16), including predominantly adult patients with postsurgical (70%) hypoparathyroidism. Most patients received PTH 1–84 plus conventional therapy (n = 83) compared with a smaller group randomized to placebo injections (n = 39) plus conventional therapy. A subset of patients had magnesium deficiency and received magnesium supplements. All patients on the PTH treatment arm initially received 50 μg PTH 1–84 daily. The protocol provided the option to titrate the PTH dose up, first to 75 μg and subsequently to 100 μg at preset intervals. SF-36 scores were abnormally low at study baseline. At 24 weeks, there were no significant differences in SF-36 scores between the two treatment arms, PTH 1–84 vs placebo. At 24 weeks, when comparing each treatment arm to baseline, patients receiving PTH 1–84 with conventional therapy had a significant improvement in SF-36 scores in several domains in two (North America and Western Europe) of the three study site geographic areas. Similar improvements were not observed in the patients on conventional therapy and placebo from those same sites. Palermo et al. (8) describe results from a 2-year open-label uncontrolled study of 42 adult subjects (38 women) with postsurgical hypoparathyroidism. Patients with hypomagnesemia were excluded. Patients received fixed doses of rhPTH 1–34, in amounts previously approved for the treatment of osteoporosis (20 μg), by SC injection twice daily. The twice-daily 20-μg rhPTH 1–34 dose remained unchanged for the 2-year duration of the study. HRQoL measures at 6 and 12 months of PTH treatment were improved in all domains compared with baseline. The study results, however, do not inform us if the improved SF-36 scores are a result of the reduction of calcitriol or the addition of rhPTH 1–34 or the synergy of the two regimens. Furthermore, baseline values reflect the results of treatment management by referring physicians, in most cases, not the investigators. Because conventional therapy was not optimized by study investigators prior to starting PTH, the baseline biochemical profile is not an adequate substitute for a conventional therapy control group. Sikjaer et al. (3) examined the effects of PTH 1–84 on muscle function and QoL. Their 6-month randomized controlled study included 62 adult patients with hypoparathyroidism who were randomized to either a daily dose of 100 μg PTH 1–84 by SC injection or to placebo injection; both treatment groups received conventional therapy. At baseline, patients had impaired HRQoL and no evidence of myopathy. At 6 months, PTH 1–84 compared with placebo did not show a beneficial effect on HRQoL and produced a significant decrease in muscle strength in the upper extremities. Intermittent periods of hypercalcemia and PTH excess (17, 18) may explain the apparent myopathy at 6 months because primary hyperparathyroidism is associated with chronic muscle fatigue (21). In a large Norwegian cross-sectional study of HRQoL in hypoparathyroidism, patients with postsurgical hypoparathyroidism had lower SF-36 scores compared with those who had nonsurgical hypoparathyroidism (22). A recent study of nonsurgical hypoparathyroidism (11) reported greater neuropsychiatric dysfunction compared with controls and lower SF-36 scores compared with postsurgical hypoparathyroidsim in three subdomains (physical function, social function, and mental health). Both studies (11, 22) reported no association between the biochemical markers of mineral metabolism and SF-36 scores. These findings suggest that the disordered sense of well-being may result from direct effects of PTH deficiency rather than the effects of abnormal mineral homeostasis. To further understand QoL deficits in patients who had thyroid surgery, Sikjaer et al. (4), in a cross-sectional study, compared three age-matched groups of 22 adult patients with (1) postsurgical hypothyroidism, (2) a combination of postsurgical hypoparathyroidism and hypothyroidism, and (3) healthy controls. Patients with isolated postsurgical hypothyroidism had abnormally low SF-36 scores in several domains, including vitality, pain, and mental health. The combination of hypothyroidism and hypoparathyroidism (treated with conventional therapy) led to more profound deficits in HRQoL and also decreased muscle strength compared with controls. Complaints of muscle weakness and fatigue were common in our patients with hypoparathyroidism when they were referred to us for therapy. We studied fatigue and physical endurance in a subset (seven patients) of a larger controlled study of twice-daily hPTH 1–34 doses compared with conventional therapy in 27 adults with hypoparathyroidism (23). At study baseline, patients had scores on a multifaceted fatigue assessment in the mild-fatigue range. A 9-minute walk test performed at baseline and after 6 months of twice-daily hPTH 1–34 injections revealed no significant differences in endurance comparing hPTH 1-34 vs conventional therapy. At 6 months, two of the four patients in the hPTH 1–34 group showed a 50% improvement in their scores on a self-reported fatigue scale. In a study comparing hPTH 1–34 delivery by an insulin pump compared with twice-daily injections in adults with postsurgical hypoparathyroidism, we showed at baseline, on conventional therapy, decreased muscle strength in eight adults with postsurgical hypoparathyroidism (14). The average maximal isometric strength, measured with Biodex 3 dynamometer (Biodex Medical Systems, Shirley, NY), was only 50% of normal age-matched reference values (14). Paired comparisons of muscle strength, however, revealed no improvement of muscle function in 3 or 6 months after twice-daily SC hPTH 1–34 injections or pump delivery of hPTH 1–34. Recovery of muscle strength likely requires PTH replacement therapy occurring over a longer period of time than that reported in this study. Furthermore, the ability of PTH replacement therapy to overcome deficits may be age or disease duration dependent. It is possible that decreased muscle strength may be related to fatigue and depression, which affect many adult patients with hypoparathyroidism (4, 5, 12). In addition, exercise often triggers symptoms of hypocalcemia, which can lead to the avoidance of even low levels of exercise, thus contributing to muscle weakness and atrophy, not easily reversed during short-term replacement therapy. Although numerous studies examine QoL, muscle weakness, and fatigue, controlled studies do not provide evidence that PTH therapy, compared with controls, reverses these deficiencies. Future studies should investigate effects of PTH 1–84 given alone instead of in combination with conventional therapy, to further demonstrate its efficacy as a replacement therapy. We need long-term safety data (24) and longitudinal studies in which steady-state normal levels of minerals in the blood and urine are achieved and the implementation of more precise, disease-specific techniques to measure QoL. Most important, future studies should be designed in which PTH is given in a physiologic manner with attention to the fine details of calcium homeostasis such as the amount of fluctuation of serum and urine calcium. A device for monitoring calcium to allow for real-time dose adjustments in response to fluctuations in blood calcium would facilitate such titration and give patients a greater sense of control over their health and, perhaps, a greater sense of well-being. Abbreviations: Abbreviations: hPTH synthetic human PTH HRQoL health-related quality of life QoL quality of life rhPTH recombinant human PTH SC subcutaneous SF-36 Short Form Health Survey 36 Acknowledgments Disclosure Summary: The author has nothing to disclose. References 1. Winer KK , Yanovski JA , Cutler GB Jr . Synthetic human parathyroid hormone 1-34 vs calcitriol and calcium in the treatment of hypoparathyroidism . JAMA . 1996 ; 276 ( 8 ): 631 – 636 . 2. Mannstadt M , Bilezikian JP , Thakker RV , Hannan FM , Clarke BL , Rejnmark L , Mitchell DM , Vokes TJ , Winer KK , Shoback DM . Hypoparathyroidism . Nat Rev Dis Primers . 2017 ; 3 : 17055 . 3. Sikjaer T , Rolighed L , Hess A , Fuglsang-Frederiksen A , Mosekilde L , Rejnmark L . Effects of PTH(1-84) therapy on muscle function and quality of life in hypoparathyroidism: results from a randomized controlled trial . Osteoporos Int . 2014 ; 25 ( 6 ): 1717 – 1726 . 4. Sikjaer T , Moser E , Rolighed L , Underbjerg L , Bislev LS , Mosekilde L , Rejnmark L . Concurrent hypoparathyroidism is associated with impaired physical function and quality of life in hypothyroidism . J Bone Miner Res . 2016 ; 31 ( 7 ): 1440 – 1448 . 5. Arlt W , Fremerey C , Callies F , Reincke M , Schneider P , Timmermann W , Allolio B. Well-being, mood and calcium homeostasis in patients with hypoparathyroidism receiving standard treatment with calcium and vitamin D . Eur J Endocrinol . 2002 ; 146 : 215 – 222 . 6. Cusano NE , Rubin MR , McMahon DJ , Irani D , Tulley A , Sliney J Jr , Bilezikian JP . The effect of PTH(1-84) on quality of life in hypoparathyroidism . J Clin Endocrinol Metab . 2013 ; 98 ( 6 ): 2356 – 2361 . 7. Vokes TJ , Mannstadt M , Levine MA , Clarke BL , Lakatos P , Chen K , Piccolo R , Krasner A , Shoback DM , Bilezikian JP . Recombinant human parathyroid hormone effect on health-related quality of life in adults with chronic hypoparathyroidism . J Clin Endocrinol Metab . 2017 ; 103 ( 2 ): 722 – 731 . 8. Palermo A , Santonati A, Tabacco G , Bosco D , Spada A , Pedone C , Raggiunti B , Doris T , Maggi D , Grimaldi F , Manfrini S , Vescini F . PTH(1-34) for surgical hypoparathyroidism: a 2 year prospective, open-label investigation of efficacy and quality of life . J Clin Endocrinol Metab . 2018 ; 103 ( 1 ): 271 – 280 . 9. Büttner M , Musholt TJ , Singer S . Quality of life in patients with hypoparathyroidism receiving standard treatment: a systematic review . Endocrine . 2017 ; 58 ( 1 ): 14 – 20 . 10. Underbjerg L , Sikjaer T , Mosekilde L , Rejnmark L . The epidemiology of nonsurgical hypoparathyroidism in Denmark: a nationwide case finding study . J Bone Miner Res . 2015 ; 30 ( 9 ): 1738 – 1744 . 11. Underbjerg L , Sikjaer T , Rejnmark L . Health-related quality of life in patients with nonsurgical hypoparathyroidism and pseudohypoparathyroidism [published online ahead of print March 9, 2018] . Clin Endocrinol (Oxf) . 12. Underbjerg L , Sikjaer T , Mosekilde L , Rejnmark L . Postsurgical hypoparathyroidism—risk of fractures, psychiatric diseases, cancer, cataract, and infections . J Bone Miner Res . 2014 ; 29 ( 11 ): 2504 – 2510 . 13. Winer KK , Yanovski JA , Sarani B , Cutler GB Jr . A randomized, cross-over trial of once-daily versus twice-daily parathyroid hormone 1-34 in treatment of hypoparathyroidism . J Clin Endocrinol Metab . 1998 ; 83 ( 10 ): 3480 – 3486 . 14. Winer KK , Zhang B , Shrader JA , Peterson D , Smith M , Albert PS , Cutler GB Jr . Synthetic human parathyroid hormone 1-34 replacement therapy: a randomized crossover trial comparing pump versus injections in the treatment of chronic hypoparathyroidism . J Clin Endocrinol Metab . 2012 ; 97 ( 2 ): 391 – 399 . 15. Winer KK , Fulton KA , Albert PS , Cutler GB Jr. Effects of pump versus twice-daily injection delivery of synthetic parathyroid hormone 1-34 in children with severe congenital hypoparathyroidism . J Pediatr . 2014 ; 165 : 556 – 563 . 16. Mannstadt M , Clarke BL , Vokes T , Brandi ML , Ranganath L , Fraser WD , Lakatos P , Bajnok L , Garceau R , Mosekilde L , Lagast H , Shoback D , Bilezikian JP . Efficacy and safety of recombinant human parathyroid hormone (1-84) in hypoparathyroidism (REPLACE): a double-blind, placebo-controlled, randomised, phase 3 study . Lancet Diabetes Endocrinol . 2013 ; 1 ( 4 ): 275 – 283 . 17. Sikjaer T , Rejnmark L , Rolighed L , Heickendorff L , Mosekilde L ; Hypoparathyroid Study Group . The effect of adding PTH(1-84) to conventional treatment of hypoparathyroidism: a randomized, placebo-controlled study . J Bone Miner Res . 2011 ; 26 ( 10 ): 2358 – 2370 . 18. Sikjaer T , Amstrup AK , Rolighed L , Kjaer SG , Mosekilde L , Rejnmark L . PTH(1-84) replacement therapy in hypoparathyroidism: a randomized controlled trial on pharmacokinetic and dynamic effects after 6 months of treatment . J Bone Miner Res . 2013 ; 28 ( 10 ): 2232 – 2243 . 19. Rubin MR , Cusano NE , Fan WW , Delgado Y , Zhang C , Costa AG , Cremers S , Dworakowski E , Bilezikian JP . Therapy of hypoparathyroidism with PTH(1-84): a prospective six year investigation of efficacy and safety . J Clin Endocrinol Metab . 2016 ; 101 ( 7 ): 2742 – 2750 . 20. Bilezikian JP , Clarke BL , Mannstadt M , Rothman J , Vokes T , Lee HM , Krasner A . Safety and efficacy of recombinant human parathyroid hormone in adults with hypoparathyroidism randomly assigned to receive fixed 25-μg or 50-μg daily doses . Clin Ther . 2017 ; 39 ( 10 ): 2096 – 2102 . 21. Reppe S , Stilgren L , Abrahamsen B , Olstad OK , Cero F , Brixen K , Nissen-Meyer LS , Gautvik KM . Abnormal muscle and hematopoietic gene expression may be important for clinical morbidity in primary hyperparathyroidism . Am J Physiol Endocrinol Metab . 2007 ; 292 ( 5 ): E1465 – E1473 . 22. Astor MC , Løvås K , Debowska A , Eriksen EF , Evang JA , Fossum C , Fougner KJ , Holte SE , Lima K , Moe RB , Myhre AG , Kemp EH , Nedrebø BG , Svartberg J , Husebye ES . Epidemiology and health-related quality of life in hypoparathyroidism in Norway . J Clin Endocrinol Metab . 2016 ; 101 ( 8 ): 3045 – 3053 . 23. Winer KK , Ko CW , Reynolds JC , Dowdy K , Keil M , Peterson D , Gerber LH , McGarvey C , Cutler GB Jr . Long-term treatment of hypoparathyroidism: a randomized controlled study comparing parathyroid hormone-(1-34) versus calcitriol and calcium . J Clin Endocrinol Metab . 2003 ; 88 ( 9 ): 4214 – 4220 . 24. Maracucci G , Pepa GD , Brandi ML . Drug safety evaluation of parathyroid hormone for hypocalcemia in patients with hypoparathyroidism . Expert Opin Drug Saf . 2017 ; 16 ( 5 ): 617 – 625 .

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-01268pmid: 29905821

In the above-named article by Hembree WC, Cohen-Kettenis PT, Gooren L, Hannema SE, Meyer WJ, Murad MH, Rosenthal SM, Safer JD, Tangpricha V, and T’Sjoen GG (J Clin Endocrinol Metab. 2017;102(11):3869–3903; doi: 10.1210/jc.2017-01658), the following errors occurred. In the Summary of Recommendations, 1.0 Evaluation of youth and adults, Recommendation 1.1 on page 3870 and in Recommendations for Those Involved in the Gender-Affirming Hormone Treatment of Individuals With GD/Gender Incongruence on page 3877, the Recommendation was originally: 1.1. We advise that only trained mental health professionals (MHPs) who meet the following criteria should diagnose gender dysphoria (GD)/gender incongruence in adults: (1) competence in using the Diagnostic and Statistical Manual of Mental Disorders (DSM) and/or the International Statistical Classification of Diseases and Related Health Problems (ICD) for diagnostic purposes, (2) the ability to diagnose GD/gender incongruence and make a distinction between GD/gender incongruence and conditions that have similar features (e.g., body dysmorphic disorder), (3) training in diagnosing psychiatric conditions, (4) the ability to undertake or refer for appropriate treatment, (5) the ability to psychosocially assess the person’s understanding, mental health, and social conditions that can impact gender-affirming hormone therapy, and (6) a practice of regularly attending relevant professional meetings. (Ungraded Good Practice Statement) The Recommendation should read: 1.1. We advise that only trained mental health professionals (MHPs) AND/OR TRAINED PHYSICIANS who meet the following criteria should diagnose GD/gender incongruence in adults: (1) competence in using the DSM and/or the ICD for diagnostic purposes, (2) the ability to diagnose GD/gender incongruence and make a distinction between GD/gender incongruence and conditions that have similar features (e.g., body dysmorphic disorder), (3) training in diagnosing RELATED psychiatric conditions, (4) the ability to undertake or refer for appropriate treatment, (5) the ability to psychosocially assess the person’s understanding, mental health, and social conditions that can impact gender-affirming hormone therapy, and (6) a practice of regularly attending relevant professional meetings. (Ungraded Good Practice Statement) In Section 4.0 Adverse Outcome Prevention and Long-Term Care, Subsections Evidence and Transgender females, on pp. 3891–3892 a citation was omitted from the text. The paragraph and citation were originally: “There have been no studies to determine whether clinicians should use the sex assigned at birth or affirmed gender for assessing osteoporosis (e.g., when using the FRAX tool). Although some researchers use the sex assigned at birth (with the assumption that bone mass has usually peaked for transgender people who initiate hormones in early adulthood), this should be assessed on a case-by-case basis until there are more data available. This assumption will be further complicated by the increasing prevalence of transgender people who undergo hormonal transition at a pubertal age or soon after puberty. Sex for comparison within risk assessment tools may be based on the age at which hormones were initiated and the length of exposure to hormones. In some cases, it may be reasonable to assess risk using both the male and female calculators and using an intermediate value. Because all subjects underwent normal pubertal development, with known effects on bone size, reference values for birth sex were used for all participants (154).” The paragraph and citations should read: “There have been no studies to determine whether clinicians should use the sex assigned at birth or affirmed gender for assessing osteoporosis (e.g., when using the FRAX tool). Although some researchers use the sex assigned at birth (with the assumption that bone mass has usually peaked for transgender people who initiate hormones in early adulthood), this should be assessed on a case-by-case basis until there are more data available. This assumption will be further complicated by the increasing prevalence of transgender people who undergo hormonal transition at a pubertal age or soon after puberty. Sex for comparison within risk assessment tools may be based on the age at which hormones were initiated and the length of exposure to hormones. In some cases, it may be reasonable to assess risk using both the male and female calculators and using an intermediate value. Because all subjects underwent normal pubertal development, with known effects on bone size, reference values for birth sex were used for all participants (154, 267).” 267. Radix A and Deutsch MB. Bone health and osteoporosis. In: Deutsch MB, ed. Guidelines for the Primary and Gender-Affirming Care of Transgender and Gender Nonbinary People; second edition. Center of Excellence for Transgender Health, Department of Family and Community Medicine, University of California San Francisco; June 2016. Available at www.transhealth.ucsf.edu/guidelines. Copyright © 2018 Endocrine Society

Bertocchio, Jean-Philippe;Tafflet, Muriel;Koumakis, Eugénie;Maruani, Gérard;Vargas-Poussou, Rosa;Silve, Caroline;Nissen, Peter H;Baron, Stéphanie;Prot-Bertoye, Caroline;Courbebaisse, Marie;Souberbielle, Jean-Claude;Rejnmark, Lars;Cormier, Catherine;Houillier, Pascal

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02773pmid: 29727008

Abstract Context Parathyroid-related hypercalcemia is due to primary hyperparathyroidism (PHPT) or to familial hypocalciuric hypercalcemia (FHH). PHPT can lead to complications that necessitate parathyroidectomy. FHH is a rare genetic disease resembling PHPT; surgery is ineffective. A reliable method for distinguishing FHH from PHPT is needed. Objective To develop an easy-to-use tool to predict if a patient has PHPT. Design Retrospective analysis of two prospective cohorts. Development of an unsupervised risk equation (Pro-FHH). Setting University hospitals in Paris, France, and Aarhus, Denmark. Participants Patients (Paris: 65 with FHH, 85 with PHPT; Aarhus: 38 with FHH, 55 with PHPT) were adults with hypercalcemia and PTH concentration within normal range. Main Outcome Measures Performance of Pro-FHH to predict PHPT. Results Pro-FHH takes into account plasma calcium, PTH, and serum osteocalcin concentrations, and calcium-to-creatinine clearance ratio calculated from 24-hour urine collection (24h-CCCR). In the Paris cohort, area under the receiver operating characteristic curve (AUROC) of Pro-FHH was 0.961, higher than that of 24h-CCCR. With a cutoff value of 0.928, Pro-FHH had 100% specificity and 100% positive predictive value for the diagnosis of PHPT; it correctly categorized 51 of 85 patients with PHPT; the remaining 34 were recommended to undergo genetic testing. No patients with FHH were wrongly categorized. In an independent cohort from Aarhus, AUROC of Pro-FHH was 0.951, higher than that of 24h-CCCR. Conclusion Pro-FHH effectively predicted whether a patient has PHPT. A prospective trial is necessary to assess its usefulness in a larger population and in patients with elevated PTH concentration. Primary hyperparathyroidism (PHPT) is an endocrine disease with an estimated prevalence of one per 1000 in men and two per 1000 in women (1). PHPT may actually be of a higher prevalence (~1%), because of undiagnosed cases (2). The typical biological presentation of PHPT has changed over the last several decades. Currently, most patients with PHPT have mild hypercalcemia with slightly increased or even normal PTH concentration. Although PHPT is frequently asymptomatic at the time of diagnosis, it is often (in ≥50% of patients) associated with renal and/or bone damage (3). The only cure for PHPT is parathyroidectomy (4, 5). Recurrent laryngeal nerve injury, transient or persistent hypoparathyroidism, and/or hungry bone syndrome (6) can complicate parathyroidectomy. Familial hypocalciuric hypercalcemia (FHH) was first reported by Foley et al. (7) in 1972. Its prevalence in the general population is unknown. FHH makes the diagnosis of parathyroid-related hypercalcemia more complex. FHH is an autosomal dominant disease with a high penetrance due to a defect of extracellular calcium sensing in the parathyroid glands and the kidney (8). FHH has been causally traced to loss-of-function mutations in three genes: CASR, encoding the calcium-sensing receptor (CaSR) in FHH1 (9, 10); GNA11, encoding the Gα11 protein in FHH2 (11); and AP2S1, encoding the adaptor-related protein complex 2, sigma-2 subunit in FHH3 (12). Parathyroid gland surgery does not cure FHH and must be avoided (13). A recent report showed that patients with PHPT and those with FHH frequently have quite similar biological presentations (14): blood calcium, magnesium, and PTH concentrations, and urinary calcium excretion in patients with PHPT or FHH considerably overlap. Therefore, distinguishing between PHPT and FHH may be extremely challenging, especially when serum PTH concentration is within the normal range, as it is in up to 48% of patients with PHPT (1, 15) and ~80% of patients with FHH (14). The latest guidelines on the diagnosis of PHPT state that calcium-to-creatinine clearance ratio (CCCR) calculated from 24-hour urine collection (24h-CCCR) can help distinguish between FHH and PHPT (16): On average, urinary calcium excretion is lower in patients with FHH than in those with PHPT. As with other clinical measures in these patients, however, 24h-CCCR values overlap in FHH and PHPT. It remains unclear whether phenotypic characteristics discriminate patients with FHH or PHPT on an individual basis. Our goal was to develop an easy-to-use tool, which we called Pro-FHH, to accurately predict whether a patient with parathyroid-related hypercalcemia has PHPT or FHH. Pro-FHH stands for “to protect FHH patients”; that is, to avoid unnecessary surgery in patients with FHH by diagnosing PHPT safely in patients with high Pro-FHH values and by performing genetic testing in all others. We studied patients with genetically proven FHH and with surgically proven PHPT and assessed the performance of Pro-FHH in two independent groups of patients with FHH or PHPT. Patients and Methods Development study subjects Data were prospectively collected from January 1992 to May 2016 and analyzed from March 2015 to December 2016. Included patients met all the following inclusion criteria: (1) high, fasting, serum ionized calcium concentration (≥1.32 mM); (2) normal, fasting, concomitant PTH concentration; (3) referral to one of the inclusion centers (southwestern area of Paris: European Georges Pompidou and Cochin Hospitals). Only adults were included (≥18 years old). This study was conducted in accordance with the declaration of Helsinki and approved by the French National regulatory board (CNIL 1887067v0). All tested patients gave their informed written consent for gene analysis (Supplemental Materials). Complications (i.e., nephrolithiasis, osteoporosis, and fracture) were recorded at the time of referral. The Paris cohort included 65 patients with FHH and 85 with PHPT. Biological assessments Treatments with loop diuretics or thiazides and calcium supplements were withdrawn prior to the study. Morning blood and urine samples were collected after an overnight fast. Ionized calcium concentration was determined in serum. In plasma, total calcium (PCa), PTH, phosphate (PPi), magnesium (PMg), 25(OH)vitamin D, osteocalcin (Ocn), and creatinine (PCr) concentrations were measured. In the second morning urine, calcium and creatinine were quantified. In urine calcium (UCa), phosphate, sodium, urea, and creatinine (UCr) concentrations in a 24-hour collection sample were quantified. Estimated glomerular filtration rate (eGFR) was estimated with the Modification of Diet in Renal Disease formula: 186×(PCr × 0.0113)−1.154 × age−0.203, adapted to sex as described previously (17). Because various analytical methods for the measurement of PTH, PMg, and Ocn concentrations were used over the time (Supplemental Materials), those variables are expressed as ratio of measured value to the upper limit of normal. 24h-CCCR was calculated as previously published (18): 24-hour UCa x PCrPCa x 24-hour UCr. Statistical analyses Baseline characteristics are described by median (interquartile range) or as a percentage for quantitative and qualitative data, respectively. Values were compared by Mann-Whitney or χ2 test using Prism, version 7.0b, for MacOSX (GraphPad Software) where appropriate. For additional analyses, SAS software, version 9.4 (SAS Institute) was used. Variables were log-transformed if not normally distributed. All variables were log-transformed except age, PTH ratio, PPi, and PCr. As an exploratory study, a principal component analysis was performed to highlight similarities and/or redundancies between variables to limit the number of subsequent multivariate analyses (Supplemental Fig. 1A and 1B). A heatmap was used to visualize pairwise correlations (Supplemental Fig. 1C). Models and risk equation We created two different logistic regression models to predict the risk of having PHPT. Model 1 was based on the standard recommendations of care (16): the 24-hour CCCR. Model 2 (M2) was an unsupervised model: From all data, without any supervised choice, a stepwise regression selected all nonredundant variables that reached a sufficient importance (P < 0.20) in univariate analysis as well as age, sex, history of osteoporosis, nephrolithiasis, and fracture. Effects were entered step by step into the model when P values were <0.10 and were removed when P values were >0.05. We ensured that referral to one hospital or the other did not change the estimations of the final model. Pro-FHH, the unsupervised risk equation, was then developed using the selected variables of the multivariate model (i.e., M2) as follows: P = 11+e−Xβ, where X is the vector of the selected variables and β is the vector of regression coefficients of the logistic regression. Internal validation Internal validation was performed by the leave-one-out cross-validation technique (Supplemental Fig. 2). We dropped data of one subject and re-estimated the parameter as many times as there were subjects. We then graphically controlled for the similarity between the cross-validated individual predicted to the individual predicted probability. We evaluated the discrimination ability by calculating the area under the receiver operating characteristic curve (AUROC) for each model. We obtained 95% CIs after 1000 bootstrapped replications. We tested differences between each AUROC of the two models using a paired Student t test. The concordance between predicted and observed number of patients with PHPT by decile of estimated risk was evaluated by the Hosmer-Lemeshow goodness-of-fit test; P > 0.20 indicated an adequate calibration. External validation For external validation, 93 patients (55 with PHPT and 38 with FHH) from an independent, well-described (19) prospective cohort in Aarhus, Denmark, were included. They met the same inclusion criteria as patients included in the development study. Ocn was not measured in this cohort; therefore, serum alkaline phosphatase concentration as a marker of bone remodeling (MBR) was used instead, expressed as a ratio to the upper limit of the normal range. The individual probability was calculated for each of those patients, and AUROC was calculated to evaluate the capacity of the equation to discriminate correctly the patients. P < 0.05 was considered significant. Results Patient characteristics The study flowchart is shown in Fig. 1. A total of 220 patients met the inclusion criteria, of whom 12 were lost to follow-up (LTF). A total of 116 patients were screened first for genetic abnormality; of these, 56 had FHH and 36 of the 116 patients were LTF or declined surgery. Of the 24 patients who underwent parathyroid surgery, 19 were cured, 4 were not cured, and 1 was LTF. Ninety-two patients underwent parathyroidectomy as a first-line treatment; of those, 66 were cured and 26 were not. Of the patients who were not cured by surgery, 13 underwent genetic testing, which was positive in 9, negative in 4, and 13 had an uncertain diagnosis (i.e., no genetic testing). Overall, 65 patients (30%) had genetically proven FHH (59 with FHH1, 1 with FHH2, and 5 with FHH3; gene mutations are reported in Supplemental Table 1); 85 patients (39%) had surgically proven PHPT, according the usual criteria (20); and 70 patients (32%) had uncertain diagnosis. Figure 1. View largeDownload slide Flowchart of the study. A total of 220 patients met the inclusion criteria during the period of inclusion and were screened for biological and clinical history. Patients underwent genetic testing or surgery; if results were negative, they were proposed for the other arm of the study, if not LTF. Patients with FHH (n = 65) came from groups A and F, as shown in the figure; patients with PHPT (n = 85) came from groups B and E. Some patients were left with uncertain diagnosis (n = 70) because they did not recover after surgery and were negative for genetic testing (n = 8 from groups C and G), because data were missing (LTF), or because they declined surgery or genetic testing (n = 62 from groups D, H, and I). iCa, fasting ionized serum calcium concentration; LTF, lost to follow-up; N, normal. Figure 1. View largeDownload slide Flowchart of the study. A total of 220 patients met the inclusion criteria during the period of inclusion and were screened for biological and clinical history. Patients underwent genetic testing or surgery; if results were negative, they were proposed for the other arm of the study, if not LTF. Patients with FHH (n = 65) came from groups A and F, as shown in the figure; patients with PHPT (n = 85) came from groups B and E. Some patients were left with uncertain diagnosis (n = 70) because they did not recover after surgery and were negative for genetic testing (n = 8 from groups C and G), because data were missing (LTF), or because they declined surgery or genetic testing (n = 62 from groups D, H, and I). iCa, fasting ionized serum calcium concentration; LTF, lost to follow-up; N, normal. Diagnoses were uncertain because subjects were LTF, did not undergo genetic testing, declined surgery, or were not cured by surgery. We compared the clinical and biological characteristics of the 70 patients with uncertain diagnosis with those of included patients (i.e., those with FHH or PHPT). The former were slightly older, less often had a diagnosis of osteoporosis, and had lower PCa and PMg levels and eGFR than included patients (Supplemental Table 2). In the group of patients with uncertain diagnosis, patients were older, reported more history of nephrolithiasis, and had a lower PCa but a higher 24h-CCCR than included patients with FHH. None of the other available data differed. The clinical and biological data of patients with FHH and those with PHPT are shown in Table 1. Patients with FHH were younger than those with PHPT. Most patients (78%) were women, and this sex disequilibrium was more marked in the PHPT group than in the FHH group. At diagnosis, 32 patients (21%) reported a history of nephrolithiasis; 41 (27%), osteoporosis; and 20 (13%), a history of fracture. These findings were predominantly observed in the PHPT group. PCa concentration was higher in patients with FHH than in those with PHPT. The median PTH concentration was in the upper part of the normal range for both groups and higher in the PHPT group, as was UCa excretion [assessed on 24-hour urine collection or corrected to UCr concentration (i.e., 24-hour UCa/UCr and fasting UCa/UCr)], and body weight. PPi and renal threshold of phosphate (21) were in the lower part of the normal range, with no difference between groups. Daily urinary sodium and urea excretion levels were similar between patients with FHH and those with PHPT. 25(OH)vitamin D concentration was similar in both groups. The median PMg ratio was within the normal range in both groups and slightly, but significantly, higher in patients with FHH. Median MBR (Ocn) ratio was significantly higher in patients with PHPT than in those with FHH. eGFR was within the normal range and similar in both groups. Table 1. Characteristics of Patients Included in the Development Study FHH (n = 65) PHPT (n = 85) P Age, y 49 (39–62) 59 (52–68) <0.001 Women, no. (%) 44 (68) 73 (86) 0.01 Postmenopausal women, no. (%) 26 (59) 58 (79) 0.02 BMI, kg/m2 24.6 (21.3–27.5) 23.4 (21.5–26.5) 0.31 History of nephrolithiasis, no. (%) 8 (12) 24 (28) 0.02 History of osteoporosis, no. (%) 8 (12) 33 (35) 0.001 History of fracture, no. (%) 4 (6) 16 (19) 0.02 Basal biology  Fasting plasma calcium, mM 2.62 (2.54–2.71) 2.53 (2.46–2.58) <0.001  Fasting ionized calcium, mM 1.41 (1.37–1.47) 1.37 (1.34–1.41) <0.001  Fasting serum PTH ratioa 0.70 (0.55–0.87) 0.86 (0.75–0.91) <0.001  Fasting UCa/UCr, mmol/mmol 0.15 (0.10–0.25) 0.45 (0.32–0.65) <0.001  Fasting plasma phosphate, mM 0.86 (0.75–0.96) 0.86 (0.76–0.96) 0.72  TmPi/GFR, mmol/L GF 0.83 (0.70–0.98) 0.80 (0.68–0.89) 0.24  Urinary creatinine excretion, mmol/d 10.60 (8.60–13.00) 9.50 (8.40–11.15) 0.08  Urinary phosphate excretion, mmol/d 20.9 (17.1–25.6) 22.0 (18.5–27.9) 0.21  Urinary calcium excretion, mmol/d 2.24 (1.20–3.67) 4.86 (3.69–7.23) <0.001  Urinary sodium excretion, mmol/d 124 (92–158) 132 (96–167) 0.47  Urinary urea excretion, mmol/d 319 (223–407) 301 (240–366) 0.66  Plasma 25(OH)vitamin D, nM 59 (42–88) 67 (45–84) 0.78  Fasting plasma magnesium ratioa 0.90 (0.84–0.97) 0.86 (0.80–0.88) <0.001  Plasma osteocalcin ratioa 0.81 (0.62–1.07) 1.24 (0.97–1.64) <0.001  Plasma creatinine, μM 74 (61–85) 68 (58–76) 0.003  eGFR, mL/min/1.73 m2b 85 (70–109) 88 (77–106) 0.42  Fasting CCCR, % 0.43 (0.26–0.73) 1.17 (0.89–1.53) <0.001  24h-CCCR, % 0.56 (0.34–0.87) 1.34 (0.99–1.82) <0.001  UCa/body weight, mmol/kg 0.03 (0.02–0.06) 0.08 (0.06–0.12) <0.001  24-hour UCa/UCr, mmol/mmol 0.21 (0.12–0.33) 0.51 (0.38–0.73) <0.001 FHH (n = 65) PHPT (n = 85) P Age, y 49 (39–62) 59 (52–68) <0.001 Women, no. (%) 44 (68) 73 (86) 0.01 Postmenopausal women, no. (%) 26 (59) 58 (79) 0.02 BMI, kg/m2 24.6 (21.3–27.5) 23.4 (21.5–26.5) 0.31 History of nephrolithiasis, no. (%) 8 (12) 24 (28) 0.02 History of osteoporosis, no. (%) 8 (12) 33 (35) 0.001 History of fracture, no. (%) 4 (6) 16 (19) 0.02 Basal biology  Fasting plasma calcium, mM 2.62 (2.54–2.71) 2.53 (2.46–2.58) <0.001  Fasting ionized calcium, mM 1.41 (1.37–1.47) 1.37 (1.34–1.41) <0.001  Fasting serum PTH ratioa 0.70 (0.55–0.87) 0.86 (0.75–0.91) <0.001  Fasting UCa/UCr, mmol/mmol 0.15 (0.10–0.25) 0.45 (0.32–0.65) <0.001  Fasting plasma phosphate, mM 0.86 (0.75–0.96) 0.86 (0.76–0.96) 0.72  TmPi/GFR, mmol/L GF 0.83 (0.70–0.98) 0.80 (0.68–0.89) 0.24  Urinary creatinine excretion, mmol/d 10.60 (8.60–13.00) 9.50 (8.40–11.15) 0.08  Urinary phosphate excretion, mmol/d 20.9 (17.1–25.6) 22.0 (18.5–27.9) 0.21  Urinary calcium excretion, mmol/d 2.24 (1.20–3.67) 4.86 (3.69–7.23) <0.001  Urinary sodium excretion, mmol/d 124 (92–158) 132 (96–167) 0.47  Urinary urea excretion, mmol/d 319 (223–407) 301 (240–366) 0.66  Plasma 25(OH)vitamin D, nM 59 (42–88) 67 (45–84) 0.78  Fasting plasma magnesium ratioa 0.90 (0.84–0.97) 0.86 (0.80–0.88) <0.001  Plasma osteocalcin ratioa 0.81 (0.62–1.07) 1.24 (0.97–1.64) <0.001  Plasma creatinine, μM 74 (61–85) 68 (58–76) 0.003  eGFR, mL/min/1.73 m2b 85 (70–109) 88 (77–106) 0.42  Fasting CCCR, % 0.43 (0.26–0.73) 1.17 (0.89–1.53) <0.001  24h-CCCR, % 0.56 (0.34–0.87) 1.34 (0.99–1.82) <0.001  UCa/body weight, mmol/kg 0.03 (0.02–0.06) 0.08 (0.06–0.12) <0.001  24-hour UCa/UCr, mmol/mmol 0.21 (0.12–0.33) 0.51 (0.38–0.73) <0.001 Data are given as median (interquartile range), unless otherwise indicated. Abbreviations: BMI, body mass index; GF, glomerular filtrate; GFR, glomerular filtration rate; TmPi, renal threshold for phosphate. a More than one type of assay was used, and data are expressed as the ratio of measured value to upper limit of normal. b By Modification of Diet in Renal Disease formula. View Large Table 1. Characteristics of Patients Included in the Development Study FHH (n = 65) PHPT (n = 85) P Age, y 49 (39–62) 59 (52–68) <0.001 Women, no. (%) 44 (68) 73 (86) 0.01 Postmenopausal women, no. (%) 26 (59) 58 (79) 0.02 BMI, kg/m2 24.6 (21.3–27.5) 23.4 (21.5–26.5) 0.31 History of nephrolithiasis, no. (%) 8 (12) 24 (28) 0.02 History of osteoporosis, no. (%) 8 (12) 33 (35) 0.001 History of fracture, no. (%) 4 (6) 16 (19) 0.02 Basal biology  Fasting plasma calcium, mM 2.62 (2.54–2.71) 2.53 (2.46–2.58) <0.001  Fasting ionized calcium, mM 1.41 (1.37–1.47) 1.37 (1.34–1.41) <0.001  Fasting serum PTH ratioa 0.70 (0.55–0.87) 0.86 (0.75–0.91) <0.001  Fasting UCa/UCr, mmol/mmol 0.15 (0.10–0.25) 0.45 (0.32–0.65) <0.001  Fasting plasma phosphate, mM 0.86 (0.75–0.96) 0.86 (0.76–0.96) 0.72  TmPi/GFR, mmol/L GF 0.83 (0.70–0.98) 0.80 (0.68–0.89) 0.24  Urinary creatinine excretion, mmol/d 10.60 (8.60–13.00) 9.50 (8.40–11.15) 0.08  Urinary phosphate excretion, mmol/d 20.9 (17.1–25.6) 22.0 (18.5–27.9) 0.21  Urinary calcium excretion, mmol/d 2.24 (1.20–3.67) 4.86 (3.69–7.23) <0.001  Urinary sodium excretion, mmol/d 124 (92–158) 132 (96–167) 0.47  Urinary urea excretion, mmol/d 319 (223–407) 301 (240–366) 0.66  Plasma 25(OH)vitamin D, nM 59 (42–88) 67 (45–84) 0.78  Fasting plasma magnesium ratioa 0.90 (0.84–0.97) 0.86 (0.80–0.88) <0.001  Plasma osteocalcin ratioa 0.81 (0.62–1.07) 1.24 (0.97–1.64) <0.001  Plasma creatinine, μM 74 (61–85) 68 (58–76) 0.003  eGFR, mL/min/1.73 m2b 85 (70–109) 88 (77–106) 0.42  Fasting CCCR, % 0.43 (0.26–0.73) 1.17 (0.89–1.53) <0.001  24h-CCCR, % 0.56 (0.34–0.87) 1.34 (0.99–1.82) <0.001  UCa/body weight, mmol/kg 0.03 (0.02–0.06) 0.08 (0.06–0.12) <0.001  24-hour UCa/UCr, mmol/mmol 0.21 (0.12–0.33) 0.51 (0.38–0.73) <0.001 FHH (n = 65) PHPT (n = 85) P Age, y 49 (39–62) 59 (52–68) <0.001 Women, no. (%) 44 (68) 73 (86) 0.01 Postmenopausal women, no. (%) 26 (59) 58 (79) 0.02 BMI, kg/m2 24.6 (21.3–27.5) 23.4 (21.5–26.5) 0.31 History of nephrolithiasis, no. (%) 8 (12) 24 (28) 0.02 History of osteoporosis, no. (%) 8 (12) 33 (35) 0.001 History of fracture, no. (%) 4 (6) 16 (19) 0.02 Basal biology  Fasting plasma calcium, mM 2.62 (2.54–2.71) 2.53 (2.46–2.58) <0.001  Fasting ionized calcium, mM 1.41 (1.37–1.47) 1.37 (1.34–1.41) <0.001  Fasting serum PTH ratioa 0.70 (0.55–0.87) 0.86 (0.75–0.91) <0.001  Fasting UCa/UCr, mmol/mmol 0.15 (0.10–0.25) 0.45 (0.32–0.65) <0.001  Fasting plasma phosphate, mM 0.86 (0.75–0.96) 0.86 (0.76–0.96) 0.72  TmPi/GFR, mmol/L GF 0.83 (0.70–0.98) 0.80 (0.68–0.89) 0.24  Urinary creatinine excretion, mmol/d 10.60 (8.60–13.00) 9.50 (8.40–11.15) 0.08  Urinary phosphate excretion, mmol/d 20.9 (17.1–25.6) 22.0 (18.5–27.9) 0.21  Urinary calcium excretion, mmol/d 2.24 (1.20–3.67) 4.86 (3.69–7.23) <0.001  Urinary sodium excretion, mmol/d 124 (92–158) 132 (96–167) 0.47  Urinary urea excretion, mmol/d 319 (223–407) 301 (240–366) 0.66  Plasma 25(OH)vitamin D, nM 59 (42–88) 67 (45–84) 0.78  Fasting plasma magnesium ratioa 0.90 (0.84–0.97) 0.86 (0.80–0.88) <0.001  Plasma osteocalcin ratioa 0.81 (0.62–1.07) 1.24 (0.97–1.64) <0.001  Plasma creatinine, μM 74 (61–85) 68 (58–76) 0.003  eGFR, mL/min/1.73 m2b 85 (70–109) 88 (77–106) 0.42  Fasting CCCR, % 0.43 (0.26–0.73) 1.17 (0.89–1.53) <0.001  24h-CCCR, % 0.56 (0.34–0.87) 1.34 (0.99–1.82) <0.001  UCa/body weight, mmol/kg 0.03 (0.02–0.06) 0.08 (0.06–0.12) <0.001  24-hour UCa/UCr, mmol/mmol 0.21 (0.12–0.33) 0.51 (0.38–0.73) <0.001 Data are given as median (interquartile range), unless otherwise indicated. Abbreviations: BMI, body mass index; GF, glomerular filtrate; GFR, glomerular filtration rate; TmPi, renal threshold for phosphate. a More than one type of assay was used, and data are expressed as the ratio of measured value to upper limit of normal. b By Modification of Diet in Renal Disease formula. View Large Development and assessment of performance of Pro-FHH The principal component analysis is described in Supplemental Material and Supplemental Fig. 1. The nonredundant quantitative variables entered in the stepwise analysis were age; 24-hour UCr, fasting PCa, fasting CCCR, and 24h-CCCR levels; MBR (Ocn) ratio, PMg ratio, and PTH ratio. The qualitative variables were sex, history of fracture, nephrolithiasis, and osteoporosis. Age, PTH ratio, PCa level, MBR (Ocn) ratio, and 24h-CCCR were independently associated with the risk of having PHPT and were kept in the development of the unsupervised risk equation M2 (Table 2). Pro-FHH (i.e., probability of having PHPT) was constructed as follows: Table 2. Characteristics and Performances of the Two Models Model Variables No. OR (95%CI) AUROC AUROC Bootstrap HL-χ2 1: Recommendations Ln(24h-CCCR) per 0.1 148a 1.32 (1.20–1.45) 0.862 0.862 (0.860–0.844) 0.26 2: Unsupervised (i.e., Pro-FHH) PCa per 0.01 147a 0.89 (0.84–0.96) 0.961 0.961 (0.960–0.962) 0.90 PTH ratio per 0.1 2.17 (1.42–3.34) Ln(24h-CCCR) per 0.1 1.36 (1.21–1.53) Ln(MBR ratio) per 0.1 1.33 (1.14–1.56) Model Variables No. OR (95%CI) AUROC AUROC Bootstrap HL-χ2 1: Recommendations Ln(24h-CCCR) per 0.1 148a 1.32 (1.20–1.45) 0.862 0.862 (0.860–0.844) 0.26 2: Unsupervised (i.e., Pro-FHH) PCa per 0.01 147a 0.89 (0.84–0.96) 0.961 0.961 (0.960–0.962) 0.90 PTH ratio per 0.1 2.17 (1.42–3.34) Ln(24h-CCCR) per 0.1 1.36 (1.21–1.53) Ln(MBR ratio) per 0.1 1.33 (1.14–1.56) Abbreviation: Ln, natural log. a Due to missing data, the number of patients is lower than the number included in the study. Table 2. Characteristics and Performances of the Two Models Model Variables No. OR (95%CI) AUROC AUROC Bootstrap HL-χ2 1: Recommendations Ln(24h-CCCR) per 0.1 148a 1.32 (1.20–1.45) 0.862 0.862 (0.860–0.844) 0.26 2: Unsupervised (i.e., Pro-FHH) PCa per 0.01 147a 0.89 (0.84–0.96) 0.961 0.961 (0.960–0.962) 0.90 PTH ratio per 0.1 2.17 (1.42–3.34) Ln(24h-CCCR) per 0.1 1.36 (1.21–1.53) Ln(MBR ratio) per 0.1 1.33 (1.14–1.56) Model Variables No. OR (95%CI) AUROC AUROC Bootstrap HL-χ2 1: Recommendations Ln(24h-CCCR) per 0.1 148a 1.32 (1.20–1.45) 0.862 0.862 (0.860–0.844) 0.26 2: Unsupervised (i.e., Pro-FHH) PCa per 0.01 147a 0.89 (0.84–0.96) 0.961 0.961 (0.960–0.962) 0.90 PTH ratio per 0.1 2.17 (1.42–3.34) Ln(24h-CCCR) per 0.1 1.36 (1.21–1.53) Ln(MBR ratio) per 0.1 1.33 (1.14–1.56) Abbreviation: Ln, natural log. a Due to missing data, the number of patients is lower than the number included in the study. p=11+e−23.19+11.17×PCa−7.77×PTH ratio−3.09 ×  Ln(24h−CCCR)−2.89×Ln(MBR ratio) We compared Pro-FHH performance to predict PHPT with the currently recommended criterion for diagnosis, which relies only on 24h-CCCR (Fig. 2). The AUROC of Pro-FHH was significantly higher [0.961 (0.960 to 0.962)] than that of 24h-CCCR [0.862 (0.844 to 0.862), P < 0.001]. Using cutoff values of 0.928 and 2% for Pro-FHH and 24h-CCCR, respectively, no patient with FHH was incorrectly categorized as having PHPT by Pro-FHH and two patients with FHH (10%) were incorrectly categorized as having PHPT by 24h-CCCR (Table 3). The specificity and the positive predictive value of Pro-FHH for the diagnosis of PHPT were both 100%. Using cutoff values of 0.062 and 1% for Pro-FHH and 24h-CCCR, respectively, no patient with PHPT was incorrectly categorized as having FHH by Pro-FHH, but 22 patients with PHPT were incorrectly categorized as having FHH by 24h-CCCR (Table 3). The specificity and the positive predictive value of Pro-FHH for the diagnosis of FHH were both 100%. Figure 2. View largeDownload slide AUROC curves of 24h-CCCR and Pro-FHH. In the development study, model 1 was based on standard recommendations of care (16): the 24h-CCCR. M2 was the unsupervised Pro-FHH. The AUROC of M2 (Pro-FHH, red curve) is significantly (P < 0.0001) higher than that of model 1 (24h-CCCR, blue curve). Figure 2. View largeDownload slide AUROC curves of 24h-CCCR and Pro-FHH. In the development study, model 1 was based on standard recommendations of care (16): the 24h-CCCR. M2 was the unsupervised Pro-FHH. The AUROC of M2 (Pro-FHH, red curve) is significantly (P < 0.0001) higher than that of model 1 (24h-CCCR, blue curve). Table 3. Categorization of Patients According to 24h-CCCR and Pro-FHH No. of Patients With FHH No. of Patients With PHPT Total No. Positive Predictive Value a , % Negative Predictive Value, % Sensitivity, % Specificity a , % Performance in diagnosis of PHPT  24h-CCCR in the development cohort   ≤2% 61 67 128   >2% 2 18 20 90.0 47.7 21.2 96.8   Total 63b 85 148  p-Pro-FHHc in the development cohort   ≤0.928 62 34 96   >0.928 0 51 51 100.0 64.6 60.0 100.0   Total 62b 85 147  24h-CCCR in validation cohort   ≤2% 38 35 73   >2% 0 20 20 100.0 52.0 36.4 100.0   Total 38 55 93  p-Pro-FHH in validation cohort   ≤0.928 38 46 84   >0.928 0 9 9 100.0 45.2 16.4 100.0   Total 38 55 93 Performance in diagnosis of FHH  24h-CCCR in the development cohort   <1% 50 22 72 69.4 82.9 79.4 74.1   ≥1% 13 63 76   Total 63b 85 148  p-Pro-FHH in the development cohort   <0.062 38 0 38 100.0 78.0 61.3 100.0   ≥0.062 24 85 109   Total 62b 85 147  24h-CCCR in validation cohort   <1% 30 8 38 78.9 85.4 78.9 85.4   ≥1% 8 47 55   Total 38 55 93  p-Pro-FHH in validation cohort   <0.062 33 5 38 86.8 90.9 86.8 90.9   ≥0.062 5 50 55   Total 38 55 93 No. of Patients With FHH No. of Patients With PHPT Total No. Positive Predictive Value a , % Negative Predictive Value, % Sensitivity, % Specificity a , % Performance in diagnosis of PHPT  24h-CCCR in the development cohort   ≤2% 61 67 128   >2% 2 18 20 90.0 47.7 21.2 96.8   Total 63b 85 148  p-Pro-FHHc in the development cohort   ≤0.928 62 34 96   >0.928 0 51 51 100.0 64.6 60.0 100.0   Total 62b 85 147  24h-CCCR in validation cohort   ≤2% 38 35 73   >2% 0 20 20 100.0 52.0 36.4 100.0   Total 38 55 93  p-Pro-FHH in validation cohort   ≤0.928 38 46 84   >0.928 0 9 9 100.0 45.2 16.4 100.0   Total 38 55 93 Performance in diagnosis of FHH  24h-CCCR in the development cohort   <1% 50 22 72 69.4 82.9 79.4 74.1   ≥1% 13 63 76   Total 63b 85 148  p-Pro-FHH in the development cohort   <0.062 38 0 38 100.0 78.0 61.3 100.0   ≥0.062 24 85 109   Total 62b 85 147  24h-CCCR in validation cohort   <1% 30 8 38 78.9 85.4 78.9 85.4   ≥1% 8 47 55   Total 38 55 93  p-Pro-FHH in validation cohort   <0.062 33 5 38 86.8 90.9 86.8 90.9   ≥0.062 5 50 55   Total 38 55 93 a For Pro-FHH, positive predictive value and specificity are 100% at the cutoff values of 0.062 and 0.928, respectively. The standard recommendations of care for diagnosis, 24h-CCCR, misclassified 24 patients (17%). In the validation cohort, Pro-FHH equation misclassified five patients (5.4%), whereas the standard misclassified eight patients (8.6%). b Due to missing data, the number of patients is lower than the number of included in the study. c The probability of PHPT yielded by the unsupervised pro-FHH equation. View Large Table 3. Categorization of Patients According to 24h-CCCR and Pro-FHH No. of Patients With FHH No. of Patients With PHPT Total No. Positive Predictive Value a , % Negative Predictive Value, % Sensitivity, % Specificity a , % Performance in diagnosis of PHPT  24h-CCCR in the development cohort   ≤2% 61 67 128   >2% 2 18 20 90.0 47.7 21.2 96.8   Total 63b 85 148  p-Pro-FHHc in the development cohort   ≤0.928 62 34 96   >0.928 0 51 51 100.0 64.6 60.0 100.0   Total 62b 85 147  24h-CCCR in validation cohort   ≤2% 38 35 73   >2% 0 20 20 100.0 52.0 36.4 100.0   Total 38 55 93  p-Pro-FHH in validation cohort   ≤0.928 38 46 84   >0.928 0 9 9 100.0 45.2 16.4 100.0   Total 38 55 93 Performance in diagnosis of FHH  24h-CCCR in the development cohort   <1% 50 22 72 69.4 82.9 79.4 74.1   ≥1% 13 63 76   Total 63b 85 148  p-Pro-FHH in the development cohort   <0.062 38 0 38 100.0 78.0 61.3 100.0   ≥0.062 24 85 109   Total 62b 85 147  24h-CCCR in validation cohort   <1% 30 8 38 78.9 85.4 78.9 85.4   ≥1% 8 47 55   Total 38 55 93  p-Pro-FHH in validation cohort   <0.062 33 5 38 86.8 90.9 86.8 90.9   ≥0.062 5 50 55   Total 38 55 93 No. of Patients With FHH No. of Patients With PHPT Total No. Positive Predictive Value a , % Negative Predictive Value, % Sensitivity, % Specificity a , % Performance in diagnosis of PHPT  24h-CCCR in the development cohort   ≤2% 61 67 128   >2% 2 18 20 90.0 47.7 21.2 96.8   Total 63b 85 148  p-Pro-FHHc in the development cohort   ≤0.928 62 34 96   >0.928 0 51 51 100.0 64.6 60.0 100.0   Total 62b 85 147  24h-CCCR in validation cohort   ≤2% 38 35 73   >2% 0 20 20 100.0 52.0 36.4 100.0   Total 38 55 93  p-Pro-FHH in validation cohort   ≤0.928 38 46 84   >0.928 0 9 9 100.0 45.2 16.4 100.0   Total 38 55 93 Performance in diagnosis of FHH  24h-CCCR in the development cohort   <1% 50 22 72 69.4 82.9 79.4 74.1   ≥1% 13 63 76   Total 63b 85 148  p-Pro-FHH in the development cohort   <0.062 38 0 38 100.0 78.0 61.3 100.0   ≥0.062 24 85 109   Total 62b 85 147  24h-CCCR in validation cohort   <1% 30 8 38 78.9 85.4 78.9 85.4   ≥1% 8 47 55   Total 38 55 93  p-Pro-FHH in validation cohort   <0.062 33 5 38 86.8 90.9 86.8 90.9   ≥0.062 5 50 55   Total 38 55 93 a For Pro-FHH, positive predictive value and specificity are 100% at the cutoff values of 0.062 and 0.928, respectively. The standard recommendations of care for diagnosis, 24h-CCCR, misclassified 24 patients (17%). In the validation cohort, Pro-FHH equation misclassified five patients (5.4%), whereas the standard misclassified eight patients (8.6%). b Due to missing data, the number of patients is lower than the number of included in the study. c The probability of PHPT yielded by the unsupervised pro-FHH equation. View Large External validation of Pro-FHH Patients from the independent cohort were 61 (51 to 72) years old, and median 24h-CCCR was 1.20 (0.70 to 1.90). AUROC of Pro-FHH was significantly higher [0.951 (0.950 to 0.952)] than that of 24h-CCCR [0.878 (0.877 to 0.881), P < 0.001]. Using cutoff values of 0.928 and 2% for Pro-FHH and 24h-CCCR, respectively, no patient with FHH was incorrectly categorized as having PHPT by Pro-FHH or by 24h-CCCR. The specificity and the positive predictive value of Pro-FHH for the diagnosis of PHPT were both 100%. Using cutoff values of 0.062 and 1% for Pro-FHH and 24h-CCCR, respectively, five patients with PHPT (13%) were incorrectly categorized as having FHH by Pro-FHH, and eight patients with PHPT (21%) were incorrectly categorized as having FHH by 24h-CCCR (Table 3). The specificity and the positive predictive value of Pro-FHH for the diagnosis of FHH were both 100%. Overall, 24h-CCCR categorized 40 patients (17% of total study population) in the PHPT group (of whom two had FHH). Pro-FHH categorized significantly (P = 0.01) more patients (n = 60; 25% of total study population) in the PHPT group; none was misdiagnosed (Fig. 3). Figure 3. View largeDownload slide Performances of 24h-CCCR and Pro-FHH. In merged cohorts (from Paris and Aarhus), when 24h-CCCR was applied, PHPT was suspected in 17% of patients, two were misdiagnosed (FHH), and 83% had to be tested for FHH genes. When Pro-FHH was applied, PHPT was suspected in 25% of patients, none were misdiagnosed, and 75% had to be tested for FHH genes. iCa, fasting ionized serum calcium concentration; N, normal. Figure 3. View largeDownload slide Performances of 24h-CCCR and Pro-FHH. In merged cohorts (from Paris and Aarhus), when 24h-CCCR was applied, PHPT was suspected in 17% of patients, two were misdiagnosed (FHH), and 83% had to be tested for FHH genes. When Pro-FHH was applied, PHPT was suspected in 25% of patients, none were misdiagnosed, and 75% had to be tested for FHH genes. iCa, fasting ionized serum calcium concentration; N, normal. Discussion Parathyroid surgery is the only means to cure PHPT. Parathyroidectomy can fail to cure PHPT (4), however, and can be complicated (6). Parathyroidectomy must not be performed in patients with FHH. Because it is challenging to distinguish FHH from PHPT when plasma PTH concentration lies within the normal range, we focused on this population. An option is to perform genetic testing in all patients with hypercalcemia who have a normal PTH level, because “normocalcemic” FHH does exist but seems to be extremely rare (22). This, however, would increase significantly the number of tests performed, increasing costs, overloading medical genetic departments, and delaying the time to surgery for patients with PHPT. In addition, it is likely that one or more genes causing FHH are currently unknown and, therefore, a negative genetic test might not be able to rule out FHH (14, 23). Moreover, genetic analyses could find variants of uncertain pathogenicity; here, we applied the recommendations by the international guidelines (24), even if there is no definitive substitute to functional analysis. We chose to develop a statistical test to reliably predict whether a patient with hypercalcemia has PHPT, which obviates the need for genetic testing. In development of the risk assessment, we analyzed patients with FHH and with PHPT resembling FHH because plasma PTH level was within the normal range. Seventy patients LTF or with uncertain diagnoses could not be included in the study, and because the patients who were not included had 24h-CCCR, plasma PTH, and MBR values similar to those of included patients, it is unlikely that the main conclusions of this study would have been different if these patients had been included. Data from this group were closer to those of patients with PHPT than to those of patients with FHH, suggesting that the prevalence of PHPT in nonincluded patients was higher than that of FHH. FHH was previously reported to be due to a mutation of the CASR gene in two thirds of patients (25). In our cohort, the percentage was higher. About 90% of patients in our cohorts had FHH1, as recently published (14). The results should not have been different if the distribution of different FHH subtypes had been different, because Pro-FHH values do not differ between FHH type 1 and 3. PHPT is complicated by nephrolithiasis, osteoporosis, and fractures in 36%, 63%, and 14% of patients, respectively (3). In our cohort, the prevalence of complications seems similar or slightly lower, maybe due to a less severe disease, because the PTH level was within the normal range. On average, clinical (i.e., age, sex, and medical history) and biological (i.e., PCa, UCa, PMg, PTH, and Ocn) data significantly differ between the FHH and PHPT groups. However, the individual performance of those variables to distinguish FHH from PHPT is clinically insufficient because values greatly overlap between groups. A familial history of hypercalcemia could be helpful for the diagnosis of FHH, because this disease is inherited as an autosomal dominant trait; however, a familial history of hypercalcemia was known in only 24 of patients with FHH (37%) in our cohort. In addition, mutations could also occur de novo, meaning that there would be no familial history of hypercalcemia. Moreover, familial PHPT inherited as an autosomal dominant trait also exists [e.g., type 1 multiple endocrine neoplasia (26) or hyperparathyroidism-jaw tumor (27) syndromes and isolated familial PHPT]. Further complicating diagnosis, FHH and PHPT (28) or FHH and multiple endocrine neoplasia (29) may coexist in the same patients; it was not the case in our cohort. The personal history of normal PCa is relevant: New onset of hypercalcemia indicates PHPT, whereas patients with FHH have a lifelong hypercalcemia. In most patients with PHPT in this study, however, there was no previous analysis of PCa. PMg concentration and 24-hour magnesium-to-creatinine clearance ratio differ between patients with FHH and those with PHPT (30); like 24h-CCCR, the 24-hour magnesium-to-creatinine clearance ratio is lower in patients with FHH than in those with PHPT. Unfortunately, due to the retrospective design, 24-hour urinary magnesium excretion was not measured in most patients in this study. Even if PMg was significantly higher in the FHH group in our cohort, multivariate analysis did not identify PMg as an independent variable useful to distinguish patients with FHH and those with PHPT. Moreover, in previous studies, the relations between plasma and urinary magnesium in patients with FHH or PHPT differed less than the relations between PMg and UCa (14, 30). The latest guidelines recommend that 24h-CCCR be used to distinguish FHH from PHPT: A low 24h-CCCR (<1%) favors FHH, whereas high 24h-CCCR (>2%) favors PHPT (16). These cutoff values were determined on the basis of the earliest and the latest (18, 30, 31) systematic comparison studies between patients with FHH and patients with PHPT. Because 24h-CCCR can be higher than 2% in patients with FHH and lower than 1% in those with PHPT (18), this measure cannot reliably distinguish patients with FHH from those with PHPT. In the present study, two patients with FHH (2%) had 24h-CCCR values >2%, and 22 patients with PHPT (26%) had 24h-CCCR values <1%. Deficiency in 25(OH)vitamin D, which can decrease 24h-CCCR (32), was rare in our cohort: Only eight patients (5%) had a concentration <25 nM, and deficiency frequency was similar between FHH and PHPT groups and unlikely to explain the only fair performance of 24h-CCCR. Moreover, because the glomerular filtration rate is included in the calculation of 24h-CCCR, renal insufficiency could affect 24h-CCCR. In our cohorts, most patients had normal eGFR, and no difference was seen between groups. Therefore, the caveats of 24h-CCCR are likely more intrinsic than extrinsic. The difference in 24h-CCCR between patients with FHH and the patients with PHPT mainly reflects the different renal handling of calcium, because renal tubular expression of CaSR is activated in the latter and inactivated in the former. However, bone remodeling also differs; as previously reported by others (33), patients with PHPT have higher bone turnover than patients with FHH. Therefore, we included a MBR (specifically, Ocn in the French cohort, and alkaline phosphatase in the Danish cohort, depending on the available data) into the risk equation. Doing so increases the diagnostic performance of Pro-FHH over that of 24h-CCCR. Bone remodeling is increased in postmenopausal women; therefore, menopausal status could have affected the performance of Pro-FHH. Actually, the circulating level of Ocn and alkaline phosphatase was slightly higher in postmenopausal than in premenopausal women, but this did not affect the performance of Pro-FHH (data not shown). The development of Pro-FHH suffers some weaknesses. First, its development was based on retrospective design, because our patients were not systematically tested. Due to this design (and the scarcity of FHH), some assays changed over the period of inclusion (e.g., those for PMg, Ocn, and PTH); therefore, we expressed the results as a ratio to the upper limit of the normal range. Our findings should now be tested in an independent cohort and a same assay used throughout. Second, only patients having a PTH level within the normal range were included. Hyperparathyroidism occurs less frequently in patients with FHH than in patients with PHPT (i.e., only 20% of patients with FHH (14) and >50% of those with PHPT have hyperparathyroidism), it is a less commonly questioning presentation. Whether Pro-FHH is valid in patients with high PTH values remains unknown and must be now assessed; the prevalence of PHPT should be higher again in this population; therefore, the performances of Pro-FHH (to diagnose PHPT with a cutoff value >0.928) should not be affected. Strengths are that data were collected from expert centers and analyzed without any supervision, and that the performance of Pro-FHH was validated in an independent cohort. Moreover, Pro-FHH has been made as simple as possible and could be easily adopted by all practitioners because it requires only a measure of calcium, creatinine (in blood and urine), PTH, and a BMR. In an era in which PHPT is underdiagnosed (2), our machine-learning (34) approach resulted in a tool that should increase the diagnosis of PHPT without risk of confusion between FHH and PHPT. In our cohorts, the use of Pro-FHH instead of 24h-CCCR would have spared unnecessary surgery in two of 100 patients with FHH and unnecessary genetic testing in 21 of 140 patients with PHPT. Pro-FHH had higher AUROC (~0.95) compared with 24h-CCCR (~0.80) and 100% specificity and 100% positive predictive value for the diagnosis of PHPT. Therefore, use of Pro-FHH will result in a better safety profile and will spare time and money compared with the current standard. Abbreviations: Abbreviations: 24h-CCCR calcium-to-creatinine clearance ratio calculated from 24-hour urine collection AUROC area under the receiver operating characteristic curve CaSR calcium-sensing receptor CCCR calcium-to-creatinine clearance ratio eGFR estimated glomerular filtration rate FHH familial hypocalciuric hypercalcemia LTF lost to follow-up M2 model 2 Ocn osteocalcin MBR marker of bone remodeling PCa total calcium in plasma PCr total creatinine in plasma PHPT primary hyperparathyroidism PMg total magnesium in plasma PPi total phosphate in plasma UCa calcium in 24-hour urine collection sample UCr creatinine in 24-hour urine collection sample Acknowledgments We thank all patients and families and the staff from the genetic laboratories and the clinic departments of European Georges Pompidou and Cochin Hospitals, France, and Aarhus Hospital, Denmark. We thank all the physicians who referred patients for the diagnosis of hypercalcemia. We also thank Prof. Alexandre Loupy for his meaningful remarks and recommendations. Author Contributions: J.-P.B., M.T., and P.H. designed the study. J.-P.B. conducted the study. J.-P.B., E.K., and L.R. collected the data. J.-P.B., M.T., P.H.N., and P.H. analyzed the data. J.-P.B., M.T., G.M., P.H.N., and P.H. interpreted the data. J.-P.B., M.T., and P.H. drafted the manuscript. J.-P.B., M.T., G.M., R.V.-P., C.S., S.B., C.P.-B., M.C., L.R., and P.H. revised the manuscript. All authors approved the final version of manuscript. J.-P.B, M.T., and P.H. take responsibility for the integrity of the data analysis. Disclosure Summary: The authors have nothing to disclose. References 1. Yeh MW , Ituarte PH , Zhou HC , Nishimoto S , Liu IL , Harari A , Haigh PI , Adams AL . Incidence and prevalence of primary hyperparathyroidism in a racially mixed population . J Clin Endocrinol Metab . 2013 ; 98 ( 3 ): 1122 – 1129 . 2. Press DM , Siperstein AE , Berber E , Shin JJ , Metzger R , Monteiro R , Mino J , Swagel W , Mitchell JC . The prevalence of undiagnosed and unrecognized primary hyperparathyroidism: a population-based analysis from the electronic medical record . Surgery . 2013 ; 154 ( 6 ): 1232 – 1237, discussion 1237–1238 . 3. Cipriani C , Biamonte F , Costa AG , Zhang C , Biondi P , Diacinti D , Pepe J , Piemonte S , Scillitani A , Minisola S , Bilezikian JP . Prevalence of kidney stones and vertebral fractures in primary hyperparathyroidism using imaging technology . J Clin Endocrinol Metab . 2015 ; 100 ( 4 ): 1309 – 1315 . 4. Campbell MJ . The definitive management of primary hyperparathyroidism: who needs an operation ? JAMA . 2017 ; 317 ( 11 ): 1167 – 1168 . 5. Wilhelm SM , Wang TS , Ruan DT , Lee JA , Asa SL , Duh QY , Doherty GM , Herrera MF , Pasieka JL , Perrier ND , Silverberg SJ , Solórzano CC , Sturgeon C , Tublin ME , Udelsman R , Carty SE . The American Association of Endocrine Surgeons Guidelines for definitive management of primary hyperparathyroidism . JAMA Surg . 2016 ; 151 ( 10 ): 959 – 968 . 6. Udelsman R , Åkerström G , Biagini C , Duh QY , Miccoli P , Niederle B , Tonelli F . The surgical management of asymptomatic primary hyperparathyroidism: proceedings of the Fourth International Workshop . J Clin Endocrinol Metab . 2014 ; 99 ( 10 ): 3595 – 3606 . 7. Foley TP Jr , Harrison HC , Arnaud CD , Harrison HE . Familial benign hypercalcemia . J Pediatr . 1972 ; 81 ( 6 ): 1060 – 1067 . 8. Firek AF , Kao PC , Heath H III . Plasma intact parathyroid hormone (PTH) and PTH-related peptide in familial benign hypercalcemia: greater responsiveness to endogenous PTH than in primary hyperparathyroidism . J Clin Endocrinol Metab . 1991 ; 72 ( 3 ): 541 – 546 . 9. Pollak MR , Brown EM , Chou YH , Hebert SC , Marx SJ , Steinmann B , Levi T , Seidman CE , Seidman JG . Mutations in the human Ca(2+)-sensing receptor gene cause familial hypocalciuric hypercalcemia and neonatal severe hyperparathyroidism . Cell . 1993 ; 75 ( 7 ): 1297 – 1303 . 10. Chou YH , Brown EM , Levi T , Crowe G , Atkinson AB , Arnqvist HJ , Toss G , Fuleihan GE , Seidman JG , Seidman CE . The gene responsible for familial hypocalciuric hypercalcemia maps to chromosome 3q in four unrelated families . Nat Genet . 1992 ; 1 ( 4 ): 295 – 300 . 11. Nesbit MA , Hannan FM , Howles SA , Babinsky VN , Head RA , Cranston T , Rust N , Hobbs MR , Heath H III , Thakker RV . Mutations affecting G-protein subunit α11 in hypercalcemia and hypocalcemia . N Engl J Med . 2013 ; 368 ( 26 ): 2476 – 2486 . 12. Nesbit MA , Hannan FM , Howles SA , Reed AA , Cranston T , Thakker CE , Gregory L , Rimmer AJ , Rust N , Graham U , Morrison PJ , Hunter SJ , Whyte MP , McVean G , Buck D , Thakker RV . Mutations in AP2S1 cause familial hypocalciuric hypercalcemia type 3 . Nat Genet . 2013 ; 45 ( 1 ): 93 – 97 . 13. Hannan FM , Babinsky VN , Thakker RV . Disorders of the calcium-sensing receptor and partner proteins: insights into the molecular basis of calcium homeostasis . J Mol Endocrinol . 2016 ; 57 ( 3 ): R127 – R142 . 14. Vargas-Poussou R , Mansour-Hendili L , Baron S , Bertocchio JP , Travers C , Simian C , Treard C , Baudouin V , Beltran S , Broux F , Camard O , Cloarec S , Cormier C , Debussche X , Dubosclard E , Eid C , Haymann JP , Kiando SR , Kuhn JM , Lefort G , Linglart A , Lucas-Pouliquen B , Macher MA , Maruani G , Ouzounian S , Polak M , Requeda E , Robier D , Silve C , Souberbielle JC , Tack I , Vezzosi D , Jeunemaitre X , Houillier P . Familial hypocalciuric hypercalcemia types 1 and 3 and primary hyperparathyroidism: similarities and differences . J Clin Endocrinol Metab . 2016 ; 101 ( 5 ): 2185 – 2195 . 15. Wallace LB , Parikh RT , Ross LV , Mazzaglia PJ , Foley C , Shin JJ , Mitchell JC , Berber E , Siperstein AE , Milas M . The phenotype of primary hyperparathyroidism with normal parathyroid hormone levels: how low can parathyroid hormone go ? Surgery . 2011 ; 150 ( 6 ): 1102 – 1112 . 16. Eastell R , Brandi ML , Costa AG , D’Amour P , Shoback DM , Thakker RV . Diagnosis of asymptomatic primary hyperparathyroidism: proceedings of the Fourth International Workshop . J Clin Endocrinol Metab . 2014 ; 99 ( 10 ): 3570 – 3579 . 17. Levey AS , Stevens LA , Schmid CH , Zhang YL , Castro AF III , Feldman HI , Kusek JW , Eggers P , Van Lente F , Greene T , Coresh J ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) . A new equation to estimate glomerular filtration rate . Ann Intern Med . 2009 ; 150 ( 9 ): 604 – 612 . 18. Christensen SE , Nissen PH , Vestergaard P , Heickendorff L , Brixen K , Mosekilde L . Discriminative power of three indices of renal calcium excretion for the distinction between familial hypocalciuric hypercalcaemia and primary hyperparathyroidism: a follow-up study on methods . Clin Endocrinol (Oxf) . 2008 ; 69 ( 5 ): 713 – 720 . 19. Jakobsen NF , Rolighed L , Nissen PH , Mosekilde L , Rejnmark L . Muscle function and quality of life are not impaired in familial hypocalciuric hypercalcemia: a cross-sectional study on physiological effects of inactivating variants in the calcium-sensing receptor gene (CASR) . Eur J Endocrinol . 2013 ; 169 : 349 – 357 . 20. Maruani G , Hertig A , Paillard M , Houillier P . Normocalcemic primary hyperparathyroidism: evidence for a generalized target-tissue resistance to parathyroid hormone . J Clin Endocrinol Metab . 2003 ; 88 ( 10 ): 4641 – 4648 . 21. Walton RJ , Bijvoet OL . Nomogram for derivation of renal threshold phosphate concentration . Lancet . 1975 ; 2 ( 7929 ): 309 – 310 . 22. Lietman SA , Tenenbaum-Rakover Y , Jap TS , Yi-Chi W , De-Ming Y , Ding C , Kussiny N , Levine MA . A novel loss-of-function mutation, Gln459Arg, of the calcium-sensing receptor gene associated with apparent autosomal recessive inheritance of familial hypocalciuric hypercalcemia . J Clin Endocrinol Metab . 2009 ; 94 ( 11 ): 4372 – 4379 . 23. Hovden S , Rejnmark L , Ladefoged SA , Nissen PH . AP2S1 and GNA11 mutations - not a common cause of familial hypocalciuric hypercalcemia . Eur J Endocrinol . 2017 ; 176 : 177 – 185 . 24. Richards S , Aziz N , Bale S , Bick D , Das S , Gastier-Foster J , Grody WW , Hegde M , Lyon E , Spector E , Voelkerding K , Rehm HL ; ACMG Laboratory Quality Assurance Committee . Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology . Genet Med . 2015 ; 17 ( 5 ): 405 – 424 . 25. Shinall MC Jr ., Dahir KM , Broome JT . Differentiating familial hypocalciuric hypercalcemia from primary hyperparathyroidism . Endocr Pract . 2013 ; 19 : 697 – 702 . 26. Thakker RV . Multiple endocrine neoplasia type 1 (MEN1) and type 4 (MEN4) . Mol Cell Endocrinol . 2014 ; 386 ( 1-2 ): 2 – 15 . 27. Li Y , Simonds WF . Endocrine neoplasms in familial syndromes of hyperparathyroidism . Endocr Relat Cancer . 2016 ; 23 ( 6 ): R229 – R247 . 28. Brachet C , Boros E , Tenoutasse S , Lissens W , Andry G , Martin P , Bergmann P , Heinrichs C . Association of parathyroid adenoma and familial hypocalciuric hypercalcaemia in a teenager . Eur J Endocrinol . 2009 ; 161 : 207 – 210 . 29. Hovden S , Jespersen ML , Nissen PH , Poulsen PL , Rolighed L , Ladefoged SA , Rejnmark L . Multiple endocrine neoplasia phenocopy revealed as a co-occurring neuroendocrine tumor and familial hypocalciuric hypercalcemia type 3 . Clin Case Rep . 2016 ; 4 ( 10 ): 922 – 927 . 30. Marx SJ , Spiegel AM , Brown EM , Koehler JO , Gardner DG , Brennan MF , Aurbach GD . Divalent cation metabolism. Familial hypocalciuric hypercalcemia versus typical primary hyperparathyroidism . Am J Med . 1978 ; 65 ( 2 ): 235 – 242 . 31. Marx SJ . Letter to the editor: Distinguishing typical primary hyperparathyroidism from familial hypocalciuric hypercalcemia by using an index of urinary calcium . J Clin Endocrinol Metab . 2015 ; 100 : L29 – 30 . Google Scholar PubMed PubMed 32. Jayasena CN , Mahmud M , Palazzo F , Donaldson M , Meeran K , Dhillo WS . Utility of the urine calcium-to-creatinine ratio to diagnose primary hyperparathyroidism in asymptomatic hypercalcaemic patients with vitamin D deficiency . Ann Clin Biochem . 2011 ; 48 ( Pt 2 ): 126 – 129 . 33. Christensen SE , Nissen PH , Vestergaard P , Heickendorff L , Rejnmark L , Brixen K , Mosekilde L . Skeletal consequences of familial hypocalciuric hypercalcaemia vs. primary hyperparathyroidism . Clin Endocrinol (Oxf) . 2009 ; 71 ( 6 ): 798 – 807 . 34. Somnay YR , Craven M , McCoy KL , Carty SE , Wang TS , Greenberg CC , Schneider DF . Improving diagnostic recognition of primary hyperparathyroidism with machine learning . Surgery . 2017 ; 161 ( 4 ): 1113 – 1121 . Copyright © 2018 Endocrine Society

Nicolson, Norman G;Murtha, Timothy D;Dong, Weilai;Paulsson, Johan O;Choi, Jungmin;Barbieri, Andrea L;Brown, Taylor C;Kunstman, John W;Larsson, Catharina;Prasad, Manju L;Korah, Reju;Lifton, Richard P;Juhlin, C Christofer;Carling, Tobias

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00277pmid: 29726952

Abstract Context Follicular thyroid carcinoma (FTC) is classified into minimally invasive (miFTC), encapsulated angioinvasive (eaFTC), and widely invasive (wiFTC) subtypes, according to the 2017 World Health Organization guidelines. The genetic signatures of these subtypes may be crucial for diagnosis, prognosis, and treatment but have not been described. Objective Identify and describe the genetic underpinnings of subtypes of FTC. Methods Thirty-nine tumors, comprising 12 miFTCs, 17 eaFTCs, and 10 wiFTCs, were whole-exome sequenced and analyzed. Somatic mutations, constitutional sequence variants, somatic copy number alterations, and mutational signatures were described. Clinicopathologic parameters and mutational profiles were assessed for associations with patient outcomes. Results Total mutation burden was consistent across FTC subtypes, with a median of 10 (range 1 to 44) nonsynonymous somatic mutations per tumor. Overall, 20.5% of specimens had a mutation in the RAS subfamily (HRAS, KRAS, or NRAS), with no notable difference between subtypes. Mutations in TSHR, DICER1, EIF1AX, KDM5C, NF1, PTEN, and TP53 were also noted to be recurrent across the cohort. Clonality analysis demonstrated more subclones in wiFTC. Survival analysis demonstrated worse disease-specific survival in the eaFTC and wiFTC cohorts, with no recurrences or deaths for patients with miFTC. Mutation burden was associated with worse prognosis, independent of histopathological classification. Conclusions Though the number and variety of somatic variants are similar in the different histopathological subtypes of FTC in our study, mutational burden was an independent predictor of mortality and recurrence. Follicular thyroid carcinoma (FTC) is a well-differentiated endocrine malignancy that accounts for 10% of all thyroid cancers (1–3). FTC is ∼2.5 times more common in women, similar to papillary thyroid carcinoma (PTC) (2, 4). The 5-year survival of FTC is ∼88%, but drops to 78% at 10 years (5). FTCs have traditionally been classified as minimally invasive (miFTC) or widely invasive FTC (wiFTC) based on the presence of vascular and/or capsular infiltration (6). In 2017, the World Health Organization (WHO) unveiled new guidelines that include an intermediate histopathologic type: encapsulated angioinvasive FTC (eaFTC) (7). Although determination of invasive status is critical for prognostication, even the new classification relies on subjective pathological evaluation of the degree of invasion. The earlier system, despite its limitations, has been demonstrated to predict considerable differences in tumor recurrence, metastatic potential, and mortality, though the 2017 criteria have not yet been similarly validated (8–11). Although the genomic landscape and driver events in PTC have been well described (12), the molecular etiology of FTC is less well known, particularly wiFTC and eaFTC. Recurrent somatic mutations in the RAS family (NRAS, KRAS, and HRAS) have been reported in FTC, usually in the 61st codon (13–15). Most common is NRAS, mutated in 15% to 40% of FTCs (16, 17). RAS family mutations hold clinical significance, as they have been shown to increase metastatic potential and disease-specific mortality (18). The PAX8-PPARγ fusion gene is identified in about one-third of FTCs, with estimates ranging from 12% to 56% (19, 20). Although PAX8-PPARγ likely contributes to follicular tumorigenesis, it does not appear to impact prognosis (20). The increased availability and accuracy of next-generation sequencing technology has allowed recent advances in understanding the mutational landscape of FTC and the major differences between FTC and follicular thyroid adenoma (FTA) (13–15). However, nearly all prior sequencing studies in FTC have been performed on miFTC, with limited applicability to the rarer but much deadlier wiFTC. Moreover, there have been few data characterizing the molecular underpinnings of invasion or the genetic distinctions between the different categories of FTC, particularly since the recent introduction of eaFTC in the WHO 2017 guidelines. The lack of genetic markers associated with each category limits the use of the histopathological classification of follicular tumors to surgical specimens rather than fine-needle aspiration (FNA) biopsies. Our study uses next-generation sequencing techniques and bioinformatics tools to investigate the genomic landscape of FTC, with particular attention to the recently described WHO 2017 histopathological categories of invasiveness. We investigate whether the divergent behavior of these tumors in clinical practice is associated with distinct molecular profiles. Materials and Methods Patient cohort and sample acquisition The 39 patients recruited for this study received surgical treatment at Yale New Haven Hospital (n = 24; FTC600 series) or Karolinska University Hospital (n = 15; FTC1 series) between 2002 and 2013 (Supplemental Table 1). All samples were independently reviewed by a minimum of two experienced endocrine pathologists for histopathological confirmation, and poorly differentiated thyroid cancer was ruled out by the Turin criteria, in accordance with the 2017 recommendations from the WHO. Matched normal samples for each tumor were obtained from adjacent histologically normal thyroid or blood leukocyte DNA. The diagnosis and degree of invasion were confirmed according to the 2017 guidelines established by the WHO (7). Although the 2017 WHO guidelines distinguish Hürthle cell (oxyphilic) carcinomas from FTCs, both Hürthle cell and conventional FTCs were included in this study, as both exhibit invasive behavior. Informed consent was obtained from all patients involved in this study. The acquisition and use of protected health information and tissue specimens were performed as specified by the Health Insurance Portability and Accountability Act (Yale) or Swedish Act on Biobanks (Karolinska). The study was approved by the Yale University and Karolinska Institutet Institutional Review Boards. Whole-exome sequencing Genomic DNA was extracted from formalin-fixed paraffin-embedded (FFPE) or fresh-frozen tissue. Three 1-mm–thick tissue cores were obtained per block for FFPE samples, with paraffin enzymatically removed and genomic DNA prepared using a proprietary in-house method at the Yale Center for Genome Analysis. Before and after coring, FFPE blocks were sectioned, stained, and analyzed via light microscopy to ensure tissue identity. For fresh-frozen samples, after tissue disruption via sonication, genomic DNA was isolated using the DNeasy Blood & Tissue Kit (Qiagen) in accordance with the manufacturer’s instructions. The quality and quantity of isolated DNA were then measured via spectrophotometry (NanoDrop Technologies). Tumor and normal genomic DNA pairs were then subjected to exome capture and sequencing as previously described (21). Adaptors of known sequence were ligated to genomic DNA fragments that were amplified by ligation-mediated PCR. Specimens were then subjected to capture using the NimbleGen 2.1M human exome array. Exome-specific DNA was eluted and then underwent 74-base paired-end sequencing on the HiSeq 2000 instrument per the specifications of the manufacturer (Illumina). Removal of PCR duplicates was performed with Picard Tools (http://broadinstitute.github.io/picard). Though 42 tumor-normal pairs originally generated libraries for whole-exome sequencing (WES), 39 of these passed quality metrics and were included in subsequent analyses. The Burrow-Wheeler Aligner-MEM program was used to map reads to the human reference genome GRCh37/hg19 (22). Somatic and germline single nucleotide variants and short insertions or deletions were called by MuTect2 using the Bayesian classifier followed by in-house filtering scripts to increase variant calling specificity (23). Likely false-positives were excluded using the D-ToxoG filter, and variant allele frequencies <0.10 were manually censored for variant calling (24, 25). Tumor purity was calculated with a nested approach, taking into account B-allele frequency in loss-of-heterozygosity regions, followed by tumor driver allele frequency and average overall variant allele frequency (26). Known variants in annotated databases [1000 Genomes (27, 28); and NHLBI ESP6500 (Exome Variant Server, NHLBI GO Exome Sequencing Project, Seattle, WA; http://evs.gs.washington.edu/EVS/) and 2577 noncancer exomes sequenced at Yale] were excluded, and novel exonic variants were evaluated for impact on transcriptional and/or translational processing. Copy-number variation detection EXCAVATOR was used to examine somatic copy number variation (CNV) in tumors (29). Briefly, EXCAVATOR normalizes the nonuniform WES read depth taking guanine/cytosine content into account and thereby calculates the ratios of normalized read depth between tumor and normal for the exome capture intervals and performs segmentation of copy number (CN) data using a novel heterogeneous hidden Markov model algorithm. Somatic CN gain and loss were defined relative to the matched normal samples. The significance of individual CNV peaks was evaluated using GISTIC (30). Fusion gene detection in FTC Fresh-frozen tumor samples (n = 13) were assayed for the presence of three chromosomal translocations common to thyroid carcinoma: PAX8-PPARγ, RET-PTC1, or RET-PTC3. The remaining samples from the WES cohort were excluded due to low RNA yield or quality or were in FFPE and not amenable to reliable RNA gene fusion detection. After isolating RNA using an RNeasy Plus Mini Kit (Qiagen), spectrophotometry was used to determine the concentration and purity of RNA (NanoDrop Technologies). We then assayed the samples for the defined chromosomal translocations with the Thyroid Cancer Fusion Gene Detection Kit (THRNA-RT32; EntroGen, Inc.). In this protocol, cDNA synthesis and gene amplification are performed simultaneously. The products are exposed to dual-labeled hydrolysis probes designed to detect the amplification products. Fusion gene amplification was detected via a FAM-BHQ–labeled probe in a CFX96 Real-Time System thermocycler (Bio-Rad). Additionally, exome sequencing data were investigated for evidence of somatic gene fusions using the DELLY2 and Manta packages and filtered for recurrent or previously described fusion events (31, 32). Mutational signatures, clonality, and pathway analyses Mutational signatures, as described by Alexandrov et al. (33), were also analyzed. Signatures of the individual samples and across each type of tumor were constructed through the compilation and quantification of somatic variant patterns using the deconstructSigs algorithm (34). SciClone was used to detect clonal populations of tumor cells in each sample (35). In brief, this algorithm clusters mutations by variant allele frequency to predict subclonal populations within the tumor, focusing on CN neutral regions. For clonality analysis, mutations at a more generous variant allele frequency cutoff of 0.05 were included. Additionally, using the network and pathway analysis platforms from STRING (https://string-db.org), GeNets (https://apps.broadinstitute.org/genets#), and Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp/kegg/mapper.html), all called somatic alterations were evaluated to determine if genetic events were enriched in specific biologic pathways. Networks were mapped with any disconnected nodes excluded for visual clarity. The STRING database maps connections between proteins based on experimental data as well as text mining of major biomedical databases (36). GeNets analysis uses a meta-network of several other protein interaction tools and databases to map connections between genes of interest, whereas KEGG lists all pathways enriched in each set of genes, ranked by number of genes altered per pathway (37). Statistical analysis Clinicopathologic parameters including patient age, sex, tumor size (largest diameter), and American Joint Committee on Cancer (AJCC) stage (7th and 8th editions) (38) and genetic parameters including total count of somatic variants, clonality, and number of cancer-associated driver events were assessed for association with invasive status. All variables were assumed to be nonparametrically distributed due to small sample size. Continuous variables were analyzed using the Kruskal-Wallis test and/or two-tailed Mann-Whitney test, as appropriate, with Dunn correction for multiple testing. Fisher exact test with Freeman-Halton extension was used for categorical variables. Disease-specific survival (i.e., time to death due to FTC) was illustrated with Kaplan-Meier curves, and Cox proportional hazards regression was used for univariate and multivariate survival analysis. All statistical analyses were performed using Prism 7 (GraphPad Software), except for three-category Fisher exact tests, which were performed using VassarStats (vassarstats.net) and Cox proportional hazards analyses, which were performed in R (39). Results Patient and tumor characteristics Twenty-six of the 39 patients included in the exome analysis were female (67%), and 13 were male (33%). The mean age of the population was 55.2 years old (median 54; range 14 to 86 years). The overall median tumor diameter was 3.6 cm, and the majority of patients had stage I or II disease (92%; 36 out of 39) based on AJCC 8th edition. Following surgery, patients were clinically monitored for disease progression or recurrence for a median duration of 5.8 years. The overall recurrence rate and disease-specific mortality rate were both 15% (6 out of 39), as all patients in our cohort who had disease recurrence ultimately died of FTC; these figures are concordant with previously reported trends (5). All patients had surgery, and most had radioactive iodine treatment, whereas only one had cytotoxic or targeted chemotherapy (Supplemental Table 1). On average, miFTCs were smaller and found in younger patients, though these differences were not statistically significant. Relative to miFTC and eaFTC, wiFTC had a higher AJCC stage at time of surgery (P = 0.013) and were more likely to recur or be the cause of death [P = 0.0017; Table 1; Fig. 1(a)]. Table 1. Clinicopathologic Features of Patients With miFTC, eaFTC, and wiFTC Variable All (N = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Age at diagnosis, y, mean ± SD 55.2 ± 16.1 48.1 ± 15.6 58.0 ± 15.9 58.9 ± 15.7 0.24 Female sex, n/total n (%) 26/39 (67) 8/12 (67) 12/17 (71) 6/10 (60) 0.91 Tumor diameter, cm, mean ± SD 3.6 ± 1.5 3.0 ± 0.9 3.9 ± 2.0 3.8 ± 0.9 0.29 AJCC stage greater than II, n/total n (%) 3/39 (8) 0/12 (0) 0/17 (0) 3/10 (30) 0.013 (Pmi-ea > 0.99; Pmi-wi = 0.078; Pea-wi = 0.041) Disease recurrence, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Death due to FTC, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Disease-specific survival, 5-year, n/total n (%) 22/24 (92) 7/7 (100) 9/9 (100) 7/9 (78) 0.31 Disease-specific survival, 10-year, n/total n (%) 8/14 (57) 4/4 (100) 3/4 (75) 1/6 (17) 0.052 Variable All (N = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Age at diagnosis, y, mean ± SD 55.2 ± 16.1 48.1 ± 15.6 58.0 ± 15.9 58.9 ± 15.7 0.24 Female sex, n/total n (%) 26/39 (67) 8/12 (67) 12/17 (71) 6/10 (60) 0.91 Tumor diameter, cm, mean ± SD 3.6 ± 1.5 3.0 ± 0.9 3.9 ± 2.0 3.8 ± 0.9 0.29 AJCC stage greater than II, n/total n (%) 3/39 (8) 0/12 (0) 0/17 (0) 3/10 (30) 0.013 (Pmi-ea > 0.99; Pmi-wi = 0.078; Pea-wi = 0.041) Disease recurrence, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Death due to FTC, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Disease-specific survival, 5-year, n/total n (%) 22/24 (92) 7/7 (100) 9/9 (100) 7/9 (78) 0.31 Disease-specific survival, 10-year, n/total n (%) 8/14 (57) 4/4 (100) 3/4 (75) 1/6 (17) 0.052 P values <0.05 shown in bold. a Kruskal-Wallis followed by corrected Dunn test for continuous variables and Fisher exact test (with Freeman-Halton extension) for categorical variables. P values in parentheses are pairwise comparisons between miFTC, eaFTC, and wiFTC when significance was found in the overall three-way analysis. View Large Table 1. Clinicopathologic Features of Patients With miFTC, eaFTC, and wiFTC Variable All (N = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Age at diagnosis, y, mean ± SD 55.2 ± 16.1 48.1 ± 15.6 58.0 ± 15.9 58.9 ± 15.7 0.24 Female sex, n/total n (%) 26/39 (67) 8/12 (67) 12/17 (71) 6/10 (60) 0.91 Tumor diameter, cm, mean ± SD 3.6 ± 1.5 3.0 ± 0.9 3.9 ± 2.0 3.8 ± 0.9 0.29 AJCC stage greater than II, n/total n (%) 3/39 (8) 0/12 (0) 0/17 (0) 3/10 (30) 0.013 (Pmi-ea > 0.99; Pmi-wi = 0.078; Pea-wi = 0.041) Disease recurrence, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Death due to FTC, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Disease-specific survival, 5-year, n/total n (%) 22/24 (92) 7/7 (100) 9/9 (100) 7/9 (78) 0.31 Disease-specific survival, 10-year, n/total n (%) 8/14 (57) 4/4 (100) 3/4 (75) 1/6 (17) 0.052 Variable All (N = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Age at diagnosis, y, mean ± SD 55.2 ± 16.1 48.1 ± 15.6 58.0 ± 15.9 58.9 ± 15.7 0.24 Female sex, n/total n (%) 26/39 (67) 8/12 (67) 12/17 (71) 6/10 (60) 0.91 Tumor diameter, cm, mean ± SD 3.6 ± 1.5 3.0 ± 0.9 3.9 ± 2.0 3.8 ± 0.9 0.29 AJCC stage greater than II, n/total n (%) 3/39 (8) 0/12 (0) 0/17 (0) 3/10 (30) 0.013 (Pmi-ea > 0.99; Pmi-wi = 0.078; Pea-wi = 0.041) Disease recurrence, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Death due to FTC, n/total n (%) 6/39 (15) 0/12 (0) 1/17 (6) 5/9 (56) 0.0017 (Pmi-ea > 0.99; Pmi-wi = 0.0062; Pea-wi = 0.0097) Disease-specific survival, 5-year, n/total n (%) 22/24 (92) 7/7 (100) 9/9 (100) 7/9 (78) 0.31 Disease-specific survival, 10-year, n/total n (%) 8/14 (57) 4/4 (100) 3/4 (75) 1/6 (17) 0.052 P values <0.05 shown in bold. a Kruskal-Wallis followed by corrected Dunn test for continuous variables and Fisher exact test (with Freeman-Halton extension) for categorical variables. P values in parentheses are pairwise comparisons between miFTC, eaFTC, and wiFTC when significance was found in the overall three-way analysis. View Large Figure 1. View largeDownload slide Clinicopathological and mutational characteristics of FTC. (a) Clinical and pathological parameters of FTCs, grouped by histologic invasion subtype. Disease-specific survival is shown; cases lost to follow-up or who died for another reason were censored. (b) Somatic mutation profiles of recurrent and driver mutations as generated by WES. The top row contains the total number of nonsynonymous variants per tumor, including single nucleotide variants and short insertion/deletion events. Each subsequent row contains the mutations in a specific gene. Bars to the right illustrate prevalence in our cohort. Driver genes were selected based on Vogelstein et al. (41) and supplemented with purported FTC drivers from the three most recent sequencing studies in FTC: BRAF, BRIP1, CNOT1, DICER1, EIF1AX, EZH1, HRAS, IDH1, IGF2BP3, KDM5C, KMT2C, KRAS–MAP4K3, NF1, NRAS, PTEN, SOS1, SPOP, STAG2, TCF12, TP53, and TSHR (13–15). Recurrent mutations were noted in at least two samples, whereas private mutations were identified in only one sample in our cohort. (c) Recurrent arm-level CN events. Cases unable to have CNV analysis performed due to low DNA quality are marked with an “X.” F, female; M, male. Figure 1. View largeDownload slide Clinicopathological and mutational characteristics of FTC. (a) Clinical and pathological parameters of FTCs, grouped by histologic invasion subtype. Disease-specific survival is shown; cases lost to follow-up or who died for another reason were censored. (b) Somatic mutation profiles of recurrent and driver mutations as generated by WES. The top row contains the total number of nonsynonymous variants per tumor, including single nucleotide variants and short insertion/deletion events. Each subsequent row contains the mutations in a specific gene. Bars to the right illustrate prevalence in our cohort. Driver genes were selected based on Vogelstein et al. (41) and supplemented with purported FTC drivers from the three most recent sequencing studies in FTC: BRAF, BRIP1, CNOT1, DICER1, EIF1AX, EZH1, HRAS, IDH1, IGF2BP3, KDM5C, KMT2C, KRAS–MAP4K3, NF1, NRAS, PTEN, SOS1, SPOP, STAG2, TCF12, TP53, and TSHR (13–15). Recurrent mutations were noted in at least two samples, whereas private mutations were identified in only one sample in our cohort. (c) Recurrent arm-level CN events. Cases unable to have CNV analysis performed due to low DNA quality are marked with an “X.” F, female; M, male. WES Thirty-nine FTC samples and their respective matched normal samples passed quality metrics after WES. Mean depth of coverage for tumor and normal samples was 258.6 and 115.4, respectively. The proportion of bases with a minimum of 20 times coverage was 96.9% and 93.1% for tumor and normal samples, respectively, with associated error rates of 0.3% and 0.3%. Mean tumor purity was calculated at 70% and not associated with the number of called somatic mutations (Supplemental Fig. 1). Overall, there was no considerable difference in total or cancer-specific somatic mutational burden across histologic subtypes, though there was a trend toward more mutations in eaFTC and wiFTC (Table 2). Mutations in RAS family genes were noted in 20.5% of samples (8 out of 39), all in codon 61. NRAS mutations were found in 5 out of 39 samples, whereas KRAS and HRAS variants were found in 2 and 1 specimens, respectively, with no difference across histologic categories [Fig. 1(b)]. TSHR mutations were identified in four tumors. Among known cancer- or FTC-specific driver genes, DICER1, EIF1AX, KDM5C, NF1, PRDM1, PTEN, and TP53 were recurrently mutated in two samples each. Although these were all too rare to be distributed in a statistically significant manner, TP53 mutations were only noted in wiFTC (P = 0.06). There was a trend toward more driver gene mutations in the wiFTC cohort. An additional set of recurrently mutated genes in our cohort have not been previously described as drivers, and though two of these (CAMTA1 and SFPQ) are in the Catalogue of Somatic Mutations in Cancer (COSMIC) cancer gene census, the specific mutations identified in our study are not listed in COSMIC (http://cancer.sanger.ac.uk/cosmic). Table 2. Genetic Features of miFTC, eaFTC, and wiFTC Variable All (n = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Nonsynonymous somatic variants, mean ± SD 12.5 ± 8.9 8.2 ± 5.3 15.4 ± 10.7 12.8 ± 7.6 0.08 Transition/transversion ratio per sample, mean ± SD 1.3 ± 1.2 1.4 ± 2.0 1.1 ± 0.7 1.5 ± 1.0 0.42 RAS family mutation, n/total n (%) 8/39 (21) 1/12 (8) 5/17 (29) 2/10 (20) 0.37 Cancer driver gene nonsynonymous mutations, mean ± SD 0.8 ± 1.0 0.5 ± 0.5 0.8 ± 0.8 1.2 ± 1.5 0.62 Number of clones, mean ± SD 1.9 ± 0.7 1.8 ± 0.6 1.8 ± 0.6 2.4 ± 0.7 <0.05 (Pmi-ea > 0.99; Pmi-wi = 0.15; Pea-wi = 0.05) Variable All (n = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Nonsynonymous somatic variants, mean ± SD 12.5 ± 8.9 8.2 ± 5.3 15.4 ± 10.7 12.8 ± 7.6 0.08 Transition/transversion ratio per sample, mean ± SD 1.3 ± 1.2 1.4 ± 2.0 1.1 ± 0.7 1.5 ± 1.0 0.42 RAS family mutation, n/total n (%) 8/39 (21) 1/12 (8) 5/17 (29) 2/10 (20) 0.37 Cancer driver gene nonsynonymous mutations, mean ± SD 0.8 ± 1.0 0.5 ± 0.5 0.8 ± 0.8 1.2 ± 1.5 0.62 Number of clones, mean ± SD 1.9 ± 0.7 1.8 ± 0.6 1.8 ± 0.6 2.4 ± 0.7 <0.05 (Pmi-ea > 0.99; Pmi-wi = 0.15; Pea-wi = 0.05) P values <0.05 shown in bold. a Kruskal-Wallis test followed by corrected Dunn test for continuous variables or Fisher exact test (with Freeman-Halton extension) for categorical variables. P values in parentheses are pairwise comparisons between miFTC, eaFTC, and wiFTC when significance was found in the overall three-way analysis. View Large Table 2. Genetic Features of miFTC, eaFTC, and wiFTC Variable All (n = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Nonsynonymous somatic variants, mean ± SD 12.5 ± 8.9 8.2 ± 5.3 15.4 ± 10.7 12.8 ± 7.6 0.08 Transition/transversion ratio per sample, mean ± SD 1.3 ± 1.2 1.4 ± 2.0 1.1 ± 0.7 1.5 ± 1.0 0.42 RAS family mutation, n/total n (%) 8/39 (21) 1/12 (8) 5/17 (29) 2/10 (20) 0.37 Cancer driver gene nonsynonymous mutations, mean ± SD 0.8 ± 1.0 0.5 ± 0.5 0.8 ± 0.8 1.2 ± 1.5 0.62 Number of clones, mean ± SD 1.9 ± 0.7 1.8 ± 0.6 1.8 ± 0.6 2.4 ± 0.7 <0.05 (Pmi-ea > 0.99; Pmi-wi = 0.15; Pea-wi = 0.05) Variable All (n = 39) miFTC (n = 12) eaFTC (n = 17) wiFTC (n = 10) P Value a Nonsynonymous somatic variants, mean ± SD 12.5 ± 8.9 8.2 ± 5.3 15.4 ± 10.7 12.8 ± 7.6 0.08 Transition/transversion ratio per sample, mean ± SD 1.3 ± 1.2 1.4 ± 2.0 1.1 ± 0.7 1.5 ± 1.0 0.42 RAS family mutation, n/total n (%) 8/39 (21) 1/12 (8) 5/17 (29) 2/10 (20) 0.37 Cancer driver gene nonsynonymous mutations, mean ± SD 0.8 ± 1.0 0.5 ± 0.5 0.8 ± 0.8 1.2 ± 1.5 0.62 Number of clones, mean ± SD 1.9 ± 0.7 1.8 ± 0.6 1.8 ± 0.6 2.4 ± 0.7 <0.05 (Pmi-ea > 0.99; Pmi-wi = 0.15; Pea-wi = 0.05) P values <0.05 shown in bold. a Kruskal-Wallis test followed by corrected Dunn test for continuous variables or Fisher exact test (with Freeman-Halton extension) for categorical variables. P values in parentheses are pairwise comparisons between miFTC, eaFTC, and wiFTC when significance was found in the overall three-way analysis. View Large The range of somatic mutational burden across the cohort was 1 to 44 nonsynonymous variants per tumor (mean 12.5; median 10), with no marked difference in mutational signature between subtypes (Fig. 2; Supplemental Table 2). Although 55 constitutional variants were noted in potential cancer-associated genes, none of these was in BROCA-curated thyroid cancer susceptibility genes (tests.labmed.washington.edu/BROCA) (Supplemental Table 3). There was a higher number of subclones per tumor in the wiFTC category (P = 0.05), though the number of somatic variants in our cohort is low, limiting the reliability of the clonality analysis (Supplemental Fig. 2). Figure 2. View largeDownload slide Somatic mutational burden and signature of FTC. (a) Burden of somatic variants per tumor, displayed according to histological subtype. (b) Base-change spectra of individual tumors. (c) Mutational signatures [as originally described in Alexandrov et al. (33) and curated now by COSMIC] in each subtype of FTC. Signatures were generated for each tumor individually using deconstructSigs and summed to generate the distribution shown for each subtype. “Unknown” represents base changes that could not be mapped to a known mutational signature. Figure 2. View largeDownload slide Somatic mutational burden and signature of FTC. (a) Burden of somatic variants per tumor, displayed according to histological subtype. (b) Base-change spectra of individual tumors. (c) Mutational signatures [as originally described in Alexandrov et al. (33) and curated now by COSMIC] in each subtype of FTC. Signatures were generated for each tumor individually using deconstructSigs and summed to generate the distribution shown for each subtype. “Unknown” represents base changes that could not be mapped to a known mutational signature. CN analysis Overall, FTCs were defined in our study by a general pattern of CN gain, with differences at some loci between the different tumor subtypes. Arm-level CN events were identified across the genome [Fig. 1(c); Supplemental Fig. 3A and 3B]. Although the most common arm-level CN events were gains of 5q, 7p, and 12q, recurrent CN losses at 22q were also identified. The regions with the most substantial CN gains were 7q22.3-36.3 and 12q24.33, whereas 1p36.11 demonstrated CN loss (Supplemental Fig. 3C–3E). The 7q22.3-36.3 region contains 306 genes, including the oncogene PRKAR2B (40). Interestingly, the eaFTC cohort carried recurrent CN gains in region 5q31.3-q35.3, which were not seen in the other tumor types; this region carries the genes CSF1R and NPM1, both of which are implicated as cancer drivers in other studies (41). Survival analysis Invasive status and AJCC stage were significantly associated with disease-specific survival, whereas, surprisingly, tumor size had no statistically noteworthy effect on survival in our cohort [Fig. 3(a)–3(d)]. Higher-than-mean total nonsynonymous mutation burden was associated with worse survival, as was burden of cancer driver gene mutations (41) [Fig. 3(e) and 3(f)], but not mutational burden of COSMIC census genes or clones per tumor [Fig. 3(g) and 3(h)]. Univariate Cox proportional hazards analysis demonstrated that patient age, histologic subtype, AJCC stage, total nonsynonymous mutational burden (as a continuous variable), and number of driver events were all associated with disease-specific survival. Although the number of events in our cohort was too low to support multivariate analysis of all of these factors at once, total mutational burden was found to predict mortality in a limited multivariate analysis that controlled for stage and histologic subtype (Table 3). Figure 3. View largeDownload slide Kaplan-Meier survival curves. Disease-specific survival is plotted against time to event, out to 10 years of follow-up. Tick marks along survival curves indicate cases censored at that time point. Cases were stratified on the basis of: (a) invasion (n = 12 miFTC, 17 eaFTC, and 10 wiFTC); (b) tumor size (n = 30 tumors ≤4 cm and 9 tumors >4 cm); AJCC stage: (c) 7th edition (n = 25 stage I or II, 10 stage III, and 4 stage IV) or (d) 8th edition (n = 36 stage I or II and 3 stage IV); (e) total nonsynonymous mutational burden (n = 24 with <13 variants and 15 with ≥13 variants); (f) cancer driver gene mutations as proposed by Vogelstein et al. (41) (n = 19 with 0 drivers, 17 with 1 driver, and 3 with 2 or more drivers); (g) COSMIC census gene mutations (n = 9 with 0 genes mutated, 11 with 1 gene mutated, 10 with 2 genes mutated, and 9 with 3 or more genes mutated); and (h) subclones per tumor as determined by SciClone (35) (n = 8 with 1 clone, 25 with 2 clones, and 5 with 3 or more clones). Figure 3. View largeDownload slide Kaplan-Meier survival curves. Disease-specific survival is plotted against time to event, out to 10 years of follow-up. Tick marks along survival curves indicate cases censored at that time point. Cases were stratified on the basis of: (a) invasion (n = 12 miFTC, 17 eaFTC, and 10 wiFTC); (b) tumor size (n = 30 tumors ≤4 cm and 9 tumors >4 cm); AJCC stage: (c) 7th edition (n = 25 stage I or II, 10 stage III, and 4 stage IV) or (d) 8th edition (n = 36 stage I or II and 3 stage IV); (e) total nonsynonymous mutational burden (n = 24 with <13 variants and 15 with ≥13 variants); (f) cancer driver gene mutations as proposed by Vogelstein et al. (41) (n = 19 with 0 drivers, 17 with 1 driver, and 3 with 2 or more drivers); (g) COSMIC census gene mutations (n = 9 with 0 genes mutated, 11 with 1 gene mutated, 10 with 2 genes mutated, and 9 with 3 or more genes mutated); and (h) subclones per tumor as determined by SciClone (35) (n = 8 with 1 clone, 25 with 2 clones, and 5 with 3 or more clones). Table 3. Predictors of Disease-Specific Mortality Variable Hazard Ratio 95% CI P Value a Univariate analysis  Age 1.1 1.0–1.2 0.043  Sex, male 1.3 0.24–7.2 0.76  Size, cm 1.7 0.74–3.9 0.21  Subtypeb 8.2 1.1–58.6 0.036  AJCC 8th edition stageb 2.3 1.3–4.0 0.0081  Total nonsynonymous variants 1.1 1.0–1.2 0.013  Clones 1.8 0.67–4.7 0.28  Cancer driver gene mutationsc 2.8 1.4–5.4 0.0018  FTC driver gene mutationsd 3.3 1.2–9.3 0.018  RAS gene mutation 0.53 0.06–4.5 0.54  CN loss of arm 22q 0.97 0.09–11 0.98 Multivariate analysis  Subtypeb 174 0.66–45,546 0.07  AJCC 8th edition stageb 1.8 0.64–5.0 0.27  Total nonsynonymous variants 1.4 1.06–2.0 0.02 Variable Hazard Ratio 95% CI P Value a Univariate analysis  Age 1.1 1.0–1.2 0.043  Sex, male 1.3 0.24–7.2 0.76  Size, cm 1.7 0.74–3.9 0.21  Subtypeb 8.2 1.1–58.6 0.036  AJCC 8th edition stageb 2.3 1.3–4.0 0.0081  Total nonsynonymous variants 1.1 1.0–1.2 0.013  Clones 1.8 0.67–4.7 0.28  Cancer driver gene mutationsc 2.8 1.4–5.4 0.0018  FTC driver gene mutationsd 3.3 1.2–9.3 0.018  RAS gene mutation 0.53 0.06–4.5 0.54  CN loss of arm 22q 0.97 0.09–11 0.98 Multivariate analysis  Subtypeb 174 0.66–45,546 0.07  AJCC 8th edition stageb 1.8 0.64–5.0 0.27  Total nonsynonymous variants 1.4 1.06–2.0 0.02 P values <0.05 shown in bold. a Likelihood ratio test for univariate analysis or Wald test for multivariate analysis. b Per each increment of subtype (miFTC, eaFTC, or wiFTC) or stage (I–IV). c Cancer driver gene mutations as curated by Vogelstein et al. (41). d Drawn from recent exome sequencing studies in FTC: BRAF, BRIP1, CNOT1, DICER1, EIF1AX, EZH1, HRAS, IDH1, IGF2BP3, KDM5C, KMT2C, KRAS, MAP4K3, NF1, NRAS, PTEN, SOS1, SPOP, STAG2, TCF12, TP53, and TSHR (13–15). View Large Table 3. Predictors of Disease-Specific Mortality Variable Hazard Ratio 95% CI P Value a Univariate analysis  Age 1.1 1.0–1.2 0.043  Sex, male 1.3 0.24–7.2 0.76  Size, cm 1.7 0.74–3.9 0.21  Subtypeb 8.2 1.1–58.6 0.036  AJCC 8th edition stageb 2.3 1.3–4.0 0.0081  Total nonsynonymous variants 1.1 1.0–1.2 0.013  Clones 1.8 0.67–4.7 0.28  Cancer driver gene mutationsc 2.8 1.4–5.4 0.0018  FTC driver gene mutationsd 3.3 1.2–9.3 0.018  RAS gene mutation 0.53 0.06–4.5 0.54  CN loss of arm 22q 0.97 0.09–11 0.98 Multivariate analysis  Subtypeb 174 0.66–45,546 0.07  AJCC 8th edition stageb 1.8 0.64–5.0 0.27  Total nonsynonymous variants 1.4 1.06–2.0 0.02 Variable Hazard Ratio 95% CI P Value a Univariate analysis  Age 1.1 1.0–1.2 0.043  Sex, male 1.3 0.24–7.2 0.76  Size, cm 1.7 0.74–3.9 0.21  Subtypeb 8.2 1.1–58.6 0.036  AJCC 8th edition stageb 2.3 1.3–4.0 0.0081  Total nonsynonymous variants 1.1 1.0–1.2 0.013  Clones 1.8 0.67–4.7 0.28  Cancer driver gene mutationsc 2.8 1.4–5.4 0.0018  FTC driver gene mutationsd 3.3 1.2–9.3 0.018  RAS gene mutation 0.53 0.06–4.5 0.54  CN loss of arm 22q 0.97 0.09–11 0.98 Multivariate analysis  Subtypeb 174 0.66–45,546 0.07  AJCC 8th edition stageb 1.8 0.64–5.0 0.27  Total nonsynonymous variants 1.4 1.06–2.0 0.02 P values <0.05 shown in bold. a Likelihood ratio test for univariate analysis or Wald test for multivariate analysis. b Per each increment of subtype (miFTC, eaFTC, or wiFTC) or stage (I–IV). c Cancer driver gene mutations as curated by Vogelstein et al. (41). d Drawn from recent exome sequencing studies in FTC: BRAF, BRIP1, CNOT1, DICER1, EIF1AX, EZH1, HRAS, IDH1, IGF2BP3, KDM5C, KMT2C, KRAS, MAP4K3, NF1, NRAS, PTEN, SOS1, SPOP, STAG2, TCF12, TP53, and TSHR (13–15). View Large Fusion gene analysis Of the available samples tested directly for fusion genes, only one tumor was found to have the PAX8-PPARγ fusion gene (7.7%; 1 out of 13). This widely invasive, stage II tumor was successfully resected, and the patient had 10 years of recurrence-free survival prior to the conclusion of this study. This patient had no other known cancer drivers identified in our study. The observed rate of PAX8-PPARγ fusion in this study is considerably lower than previously reported (20), most likely due to the smaller cohort sizes in each FTC category. As expected for FTC, no RET-PTC1 or RET-PTC3 fusion events were identified in our cohort. The algorithmic approach to fusion gene identification from exome sequencing data did not reveal any previously described FTC fusions or any recurrent fusion events across our cohort. Pathway analyses KEGG pathway analysis demonstrated enrichment of mutations in the MAPK signaling pathway, as expected, and several generic malignancy-associated pathways, but these were not specific to particular categories of invasiveness (Supplemental Table 4). The wiFTC cohort did carry more mutations in several pathways known to be associated with more aggressive malignancies, including cAMP signaling, mitophagy, and longevity regulation. Similarly, protein-protein interaction analysis demonstrated networks of mutations in Ras signaling, cell cycle regulation, and RNA processing, which were not specific to any of the FTC subtypes (Supplemental Fig. 4). Interestingly, each subtype of tumor often carried mutations in different genes within similar networks, hinting at a final common pathway of FTC tumorigenesis. Discussion Follicular thyroid tumors present a unique dilemma for clinicians and patients. Minimally invasive FTCs carry an excellent prognosis with low risk of recurrence or disease-specific mortality, whereas widely invasive tumors are much more aggressive even with maximally intensive treatment (8–10). The new category of eaFTC provides additional granularity in histologic diagnosis, but its effect on prognosis remains unclear. Additionally, the histologic classification requires a surgical specimen in which the capsule can be examined thoroughly, which limits the utility of FNA as a diagnostic tool. Unfortunately, even with the advent of newer molecular markers from thyroid FNA specimens, many patients must still undergo partial or total thyroidectomy to arrive at a definitive diagnosis when faced with a follicular thyroid neoplasm. Although wiFTCs have a distinctly aggressive clinical phenotype, the molecular evolution leading from a normal thyroid follicular cell to wiFTC is incompletely understood. wiFTC is relatively rare, and most next-generation sequencing studies in FTC have analyzed few widely invasive tumors. Furthermore, it is not clear based on currently available evidence whether follicular tumors progress from FTA to wiFTC in a stepwise fashion through miFTC or eaFTC intermediates or whether these subtypes represent distinct entities arising from disparate genetic backgrounds. To determine whether genetic distinctions determine the different histologic categories of FTC as defined in the 2017 WHO guidelines, our group preformed WES analysis of 39 FTC specimens across a spectrum of miFTC, eaFTC, and wiFTC cases. Our cohort is characterized by relatively modest mutational burden [mean 12.5 nonsynonymous variants per tumor or 0.38 per megabase (Mb)], concordant with other recent studies of well-differentiated thyroid cancer (mean 0.41/Mb for PTC or 0.31/Mb in FTC); although the wiFTCs behave aggressively in many cases, they do not seem to carry the large mutational burden described in anaplastic thyroid cancers (mean 2.7/Mb in anaplastic thyroid cancer or 0.72/Mb excluding hypermutators) (12, 13, 42). Mutations within RAS genes were noted in 20.5% of our cohort overall. This frequency of RAS mutations is slightly lower than generally reported in FTC, but within the range reported across prior studies. Mutations in the TSH receptor gene TSHR were common in our cohort, found in 10.3% of samples. TSHR mutations occurred in both the miFTC and wiFTC groups. Mutations in TSHR and the RAS family were mutually exclusive and did not appear to occur at significantly different rates based on invasive status in our cohort. Cancer driver genes EIF1AX and DICER1 carried mutations in two samples each; recurrent mutations in these genes were demonstrated in the recent genomic analyses of FTC (13, 15), and recurrent EIF1AX mutations were noted in a recent WES analysis of anaplastic thyroid carcinoma, the most aggressive thyroid carcinoma (42). In this cohort, one wiFTC sample was found to have a BRAFK601E mutation, without any other FTC driver events. Although somatic BRAFV600E mutations are found in ∼60% of PTCs, they are not historically associated with follicular thyroid tumors (12). This FTC had no nuclear features of PTC, confirming the diagnosis of FTC rather than follicular variant of PTC. Indeed, recent reports have identified BRAF mutations in a small proportion of FTAs (14) and specifically one BRAFK601E mutation in FTC (13). Although only a single driver gene fusion event was identified in this study, WES-indiscernible fusion events such as fusions involving introns playing a driving role in some FTCs cannot be ruled out. Future comprehensive analyses in the field will hopefully identify additional gene fusions in FTC in addition to the well-described PAX8-PPARγ fusion gene and other less common fusion events such as THADA (19, 43). Our cohort was characterized by numerous somatic arm-level copy changes, including recurrent CN loss of 22q. Losses of 22q have also been identified in recent next-generation sequencing studies of FTCs (13, 15). The functional consequences of 22q deletion have not been fully elucidated, though one study has shown that these tumors behave similarly to RAS-family mutated tumors on a transcriptional level (15). The recurrent CN gains identified in 5q in eaFTC and 7q in miFTC and wiFTC may be tumor-driving events in some cases, though the molecular mechanisms have not yet been definitively identified. Survival analysis demonstrated considerable differences in outcome based on AJCC stage and invasion subtype. Interestingly, the AJCC 7th edition staging system seemed to perform better in our cohort than the 8th edition system, likely due to our cohort’s enrichment in more aggressive widely invasive cases, which were downstaged under the new system. The genetics of the tumors also informed the survival analysis, with total mutation burden, cancer driver burden, and FTC driver burden all associated with worse prognosis. Most patients who had recurrence or mortality due to FTC had a more advanced genetic profile as well as more invasive histopathological phenotype, raising the question of which factor is more important. Our multivariate survival analysis suggests that total mutational burden may be a strong prognostic indicator independent of histopathology. Total mutational burden is already in clinical use as a biomarker to guide immunomodulatory treatment in melanoma and other cancers (44). Together with histological classification, the genetic profile of invasive follicular thyroid tumors can be used to make informed treatment decisions for patients facing this diverse group of cancers. In this study, the genetic profile seems to predict survival as a complement to the traditional histological approach, and next-generation sequencing is fast becoming affordable and practical for clinical use, particularly in high-volume centers. Whether similar results could be obtained using less-invasive biopsy specimens remains to be seen, but the approach shows promise of obviating the need for diagnostic surgery for some patients with follicular lesions. Conclusion Overall, our data demonstrate that the 2017 WHO subtypes of invasive FTC arise from comparable genetic backgrounds, in spite of the dramatic differences in clinical outcome. The differences in tumor behavior may be due to time elapsed since starting on a shared pathway toward malignancy, as tumors have time to accumulate more mutations and become more aggressive over time, without a signature genetic event defining invasiveness. Alternatively, the difference may lie in the epigenetics or noncoding genetic profile of the tumor or a background thyroid milieu that encourages the more invasive and aggressive behavior even with similar somatic mutational landscape. Some tumors have an aggressive phenotype in spite of relatively bland mutational profiles, lending further credence to the idea of nongenetic driver events in some cases. In the absence of specific genetic markers to guide prognosis in FTC, clinicians will need to rely on an integrated approach to this disease, incorporating clinical, pathological, and genetic factors to guide treatment decisions. Abbreviations: Abbreviations: AJCC American Joint Committee on Cancer CN copy number CNV copy number variation COSMIC Catalogue of Somatic Mutations in Cancer eaFTC encapsulated angioinvasive follicular thyroid carcinoma FFPE formalin-fixed paraffin-embedded FNA fine-needle aspiration FTA follicular thyroid adenoma FTC follicular thyroid carcinoma KEGG Kyoto Encyclopedia of Genes and Genomes Mb megabase miFTC minimally invasive follicular thyroid carcinoma PTC papillary thyroid carcinoma WES whole-exome sequencing WHO World Health Organization wiFTC widely invasive follicular thyroid carcinoma Acknowledgments The authors thank the Ohse Research Foundation at the Yale School of Medicine, the Stockholm County Council, the Swedish Cancer Society, and the Swedish Society for Medical Research for support of this research. Financial Support: This study was supported by the Yale School of Medicine (to T.C.), Stockholms Läns Landsting (to C.C.J.), Svenska Sällskapet för Medicinsk Forskning (to C.C.J.), Cancerfonden (to C.C.J.), and the Damon Runyon Cancer Research Foundation (to T.C.). Disclosure Summary: The authors have nothing to disclose. References 1. Aschebrook-Kilfoy B , Ward MH , Sabra MM , Devesa SS . Thyroid cancer incidence patterns in the United States by histologic type, 1992-2006 . Thyroid . 2011 ; 21 ( 2 ): 125 – 134 . 2. Enewold L , Zhu K , Ron E , Marrogi AJ , Stojadinovic A , Peoples GE , Devesa SS . Rising thyroid cancer incidence in the United States by demographic and tumor characteristics, 1980-2005 . Cancer Epidemiol Biomarkers Prev . 2009 ; 18 ( 3 ): 784 – 791 . 3. Lim H , Devesa SS , Sosa JA , Check D , Kitahara CM . Trends in thyroid cancer incidence and mortality in the United States, 1974-2013 . JAMA . 2017 ; 317 ( 13 ): 1338 – 1348 . 4. Aschebrook-Kilfoy B , Grogan RH , Ward MH , Kaplan E , Devesa SS . Follicular thyroid cancer incidence patterns in the United States, 1980-2009 . Thyroid . 2013 ; 23 ( 8 ): 1015 – 1021 . 5. James BC , Aschebrook-Kilfoy B , Cipriani N , Kaplan EL , Angelos P , Grogan RH . The incidence and survival of rare cancers of the thyroid, parathyroid, adrenal, and pancreas . Ann Surg Oncol . 2016 ; 23 ( 2 ): 424 – 433 . 6. DeLellis RA , Lloyd RV , Heitz PU , Eng C , eds. Pathology & Genetics of Tumours of Endocrine Organs. World Health Organization Classification of Tumours . Vol 8. 3rd ed . Lyon : IARC Press ; 2004 . 7. Lloyd RV , Osamura RY , Klöppel G , Rosai J , eds. WHO Classification of Tumours of Endocrine Organs . Vol 10. Geneva : WHO Press ; 2017 . 8. D’Avanzo A , Treseler P , Ituarte PH , Wong M , Streja L , Greenspan FS , Siperstein AE , Duh QY , Clark OH . Follicular thyroid carcinoma: histology and prognosis . Cancer . 2004 ; 100 ( 6 ): 1123 – 1129 . 9. Sugino K , Ito K , Nagahama M , Kitagawa W , Shibuya H , Ohkuwa K , Yano Y , Uruno T , Akaishi J , Kameyama K , Ito K . Prognosis and prognostic factors for distant metastases and tumor mortality in follicular thyroid carcinoma . Thyroid . 2011 ; 21 ( 7 ): 751 – 757 . 10. Asari R , Koperek O , Scheuba C , Riss P , Kaserer K , Hoffmann M , Niederle B . Follicular thyroid carcinoma in an iodine-replete endemic goiter region: a prospectively collected, retrospectively analyzed clinical trial . Ann Surg . 2009 ; 249 ( 6 ): 1023 – 1031 . 11. Cipriani NA , Nagar S , Kaplan SP , White MG , Antic T , Sadow PM , Aschebrook-Kilfoy B , Angelos P , Kaplan EL , Grogan RH . Follicular thyroid carcinoma: how have histologic diagnoses changed in the last half-century and what are the prognostic implications ? Thyroid . 2015 ; 25 ( 11 ): 1209 – 1216 . 12. Cancer Genome Atlas Research Network . Integrated genomic characterization of papillary thyroid carcinoma . Cell . 2014 ; 159 ( 3 ): 676 – 690 . 13. Jung SH , Kim MS , Jung CK , Park HC , Kim SY , Liu J , Bae JS , Lee SH , Kim TM , Lee SH , Chung YJ . Mutational burdens and evolutionary ages of thyroid follicular adenoma are comparable to those of follicular carcinoma . Oncotarget . 2016 ; 7 ( 43 ): 69638 – 69648 . 14. Swierniak M , Pfeifer A , Stokowy T , Rusinek D , Chekan M , Lange D , Krajewska J , Oczko-Wojciechowska M , Czarniecka A , Jarzab M , Jarzab B , Wojtas B . Somatic mutation profiling of follicular thyroid cancer by next generation sequencing . Mol Cell Endocrinol . 2016 ; 433 : 130 – 137 . 15. Yoo SK , Lee S , Kim SJ , Jee HG , Kim BA , Cho H , Song YS , Cho SW , Won JK , Shin JY , Park J , Kim JI , Lee KE , Park YJ , Seo JS . Comprehensive analysis of the transcriptional and mutational landscape of follicular and papillary thyroid cancers . PLoS Genet . 2016 ; 12 ( 8 ): e1006239 . 16. Fukahori M , Yoshida A , Hayashi H , Yoshihara M , Matsukuma S , Sakuma Y , Koizume S , Okamoto N , Kondo T , Masuda M , Miyagi Y . The associations between RAS mutations and clinical characteristics in follicular thyroid tumors: new insights from a single center and a large patient cohort . Thyroid . 2012 ; 22 ( 7 ): 683 – 689 . 17. Vuong HG , Kondo T , Oishi N , Nakazawa T , Mochizuki K , Inoue T , Tahara I , Kasai K , Hirokawa M , Tran TM , Katoh R . Genetic alterations of differentiated thyroid carcinoma in iodine-rich and iodine-deficient countries . Cancer Med . 2016 ; 5 ( 8 ): 1883 – 1889 . 18. Garcia-Rostan G , Zhao H , Camp RL , Pollan M , Herrero A , Pardo J , Wu R , Carcangiu ML , Costa J , Tallini G . ras mutations are associated with aggressive tumor phenotypes and poor prognosis in thyroid cancer . J Clin Oncol . 2003 ; 21 ( 17 ): 3226 – 3235 . 19. Nikiforova MN , Biddinger PW , Caudill CM , Kroll TG , Nikiforov YE . PAX8-PPARgamma rearrangement in thyroid tumors: RT-PCR and immunohistochemical analyses . Am J Surg Pathol . 2002 ; 26 ( 8 ): 1016 – 1023 . 20. Boos LA , Dettmer M , Schmitt A , Rudolph T , Steinert H , Moch H , Sobrinho-Simões M , Komminoth P , Perren A . Diagnostic and prognostic implications of the PAX8-PPARγ translocation in thyroid carcinomas-a TMA-based study of 226 cases . Histopathology . 2013 ; 63 ( 2 ): 234 – 241 . 21. Choi M , Scholl UI , Ji W , Liu T , Tikhonova IR , Zumbo P , Nayir A , Bakkaloğlu A , Ozen S , Sanjad S , Nelson-Williams C , Farhi A , Mane S , Lifton RP . Genetic diagnosis by whole exome capture and massively parallel DNA sequencing . Proc Natl Acad Sci USA . 2009 ; 106 ( 45 ): 19096 – 19101 . 22. Li H , Durbin R . Fast and accurate short read alignment with Burrows-Wheeler transform . Bioinformatics . 2009 ; 25 ( 14 ): 1754 – 1760 . 23. Cibulskis K , Lawrence MS , Carter SL , Sivachenko A , Jaffe D , Sougnez C , Gabriel S , Meyerson M , Lander ES , Getz G . Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples . Nat Biotechnol . 2013 ; 31 ( 3 ): 213 – 219 . 24. Costello M , Pugh TJ , Fennell TJ , Stewart C , Lichtenstein L , Meldrim JC , Fostel JL , Friedrich DC , Perrin D , Dionne D , Kim S , Gabriel SB , Lander ES , Fisher S , Getz G . Discovery and characterization of artifactual mutations in deep coverage targeted capture sequencing data due to oxidative DNA damage during sample preparation . Nucleic Acids Res . 2013 ; 41 ( 6 ): e67 . 25. Chen G , Mosier S , Gocke CD , Lin MT , Eshleman JR . Cytosine deamination is a major cause of baseline noise in next-generation sequencing . Mol Diagn Ther . 2014 ; 18 ( 5 ): 587 – 593 . 26. Zhao S , Choi M , Overton JD , Bellone S , Roque DM , Cocco E , Guzzo F , English DP , Varughese J , Gasparrini S , Bortolomai I , Buza N , Hui P , Abu-Khalaf M , Ravaggi A , Bignotti E , Bandiera E , Romani C , Todeschini P , Tassi R , Zanotti L , Carrara L , Pecorelli S , Silasi DA , Ratner E , Azodi M , Schwartz PE , Rutherford TJ , Stiegler AL , Mane S , Boggon TJ , Schlessinger J , Lifton RP , Santin AD . Landscape of somatic single-nucleotide and copy-number mutations in uterine serous carcinoma . Proc Natl Acad Sci USA . 2013 ; 110 ( 8 ): 2916 – 2921 . 27. Sudmant PH , Rausch T , Gardner EJ , Handsaker RE , Abyzov A , Huddleston J , Zhang Y , Ye K , Jun G , Fritz MH , Konkel MK , Malhotra A , Stütz AM , Shi X , Casale FP , Chen J , Hormozdiari F , Dayama G , Chen K , Malig M , Chaisson MJP , Walter K , Meiers S , Kashin S , Garrison E , Auton A , Lam HYK , Mu XJ , Alkan C , Antaki D , Bae T , Cerveira E , Chines P , Chong Z , Clarke L , Dal E , Ding L , Emery S , Fan X , Gujral M , Kahveci F , Kidd JM , Kong Y , Lameijer EW , McCarthy S , Flicek P , Gibbs RA , Marth G , Mason CE , Menelaou A , Muzny DM , Nelson BJ , Noor A , Parrish NF , Pendleton M , Quitadamo A , Raeder B , Schadt EE , Romanovitch M , Schlattl A , Sebra R , Shabalin AA , Untergasser A , Walker JA , Wang M , Yu F , Zhang C , Zhang J , Zheng-Bradley X , Zhou W , Zichner T , Sebat J , Batzer MA , McCarroll SA , Mills RE , Gerstein MB , Bashir A , Stegle O , Devine SE , Lee C , Eichler EE , Korbel JO ; 1000 Genomes Project Consortium . An integrated map of structural variation in 2,504 human genomes . Nature . 2015 ; 526 ( 7571 ): 75 – 81 . 28. Auton A , Brooks LD , Durbin RM , Garrison EP , Kang HM , Korbel JO , Marchini JL , McCarthy S , McVean GA , Abecasis GR ; 1000 Genomes Project Consortium . A global reference for human genetic variation . Nature . 2015 ; 526 ( 7571 ): 68 – 74 . 29. Magi A , Tattini L , Cifola I , D’Aurizio R , Benelli M , Mangano E , Battaglia C , Bonora E , Kurg A , Seri M , Magini P , Giusti B , Romeo G , Pippucci T , De Bellis G , Abbate R , Gensini GF . EXCAVATOR: detecting copy number variants from whole-exome sequencing data . Genome Biol . 2013 ; 14 ( 10 ): R120 . 30. Mermel CH , Schumacher SE , Hill B , Meyerson ML , Beroukhim R , Getz G . GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers . Genome Biol . 2011 ; 12 ( 4 ): R41 . 31. Chen X , Schulz-Trieglaff O , Shaw R , Barnes B , Schlesinger F , Källberg M , Cox AJ , Kruglyak S , Saunders CT . Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications . Bioinformatics . 2016 ; 32 ( 8 ): 1220 – 1222 . 32. Rausch T , Zichner T , Schlattl A , Stütz AM , Benes V , Korbel JO . DELLY: structural variant discovery by integrated paired-end and split-read analysis . Bioinformatics . 2012 ; 28 ( 18 ): i333 – i339 . 33. Alexandrov LB , Nik-Zainal S , Wedge DC , Aparicio SA , Behjati S , Biankin AV , Bignell GR , Bolli N , Borg A , Børresen-Dale AL , Boyault S , Burkhardt B , Butler AP , Caldas C , Davies HR , Desmedt C , Eils R , Eyfjörd JE , Foekens JA , Greaves M , Hosoda F , Hutter B , Ilicic T , Imbeaud S , Imielinski M , Jäger N , Jones DT , Jones D , Knappskog S , Kool M , Lakhani SR , López-Otín C , Martin S , Munshi NC , Nakamura H , Northcott PA , Pajic M , Papaemmanuil E , Paradiso A , Pearson JV , Puente XS , Raine K , Ramakrishna M , Richardson AL , Richter J , Rosenstiel P , Schlesner M , Schumacher TN , Span PN , Teague JW , Totoki Y , Tutt AN , Valdés-Mas R , van Buuren MM , van ’t Veer L , Vincent-Salomon A , Waddell N , Yates LR , Zucman-Rossi J , Futreal PA , McDermott U , Lichter P , Meyerson M , Grimmond SM , Siebert R , Campo E , Shibata T , Pfister SM , Campbell PJ , Stratton MR ; Australian Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC MMML-Seq Consortium; ICGC PedBrain . Signatures of mutational processes in human cancer [published correction appears in Nature. 2013;502(7470):258]. Nature . 2013 ; 500 ( 7463 ): 415 – 421 . 34. Rosenthal R , McGranahan N , Herrero J , Taylor BS , Swanton C . DeconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution . Genome Biol . 2016 ; 17 ( 1 ): 31 . 35. Miller CA , White BS , Dees ND , Griffith M , Welch JS , Griffith OL , Vij R , Tomasson MH , Graubert TA , Walter MJ , Ellis MJ , Schierding W , DiPersio JF , Ley TJ , Mardis ER , Wilson RK , Ding L . SciClone: inferring clonal architecture and tracking the spatial and temporal patterns of tumor evolution . PLOS Comput Biol . 2014 ; 10 ( 8 ): e1003665 . 36. Szklarczyk D , Morris JH , Cook H , Kuhn M , Wyder S , Simonovic M , Santos A , Doncheva NT , Roth A , Bork P , Jensen LJ , von Mering C . The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible . Nucleic Acids Res . 2017 ; 45 ( D1 ): D362 – D368 . 37. Kanehisa M , Goto S , Sato Y , Furumichi M , Tanabe M . KEGG for integration and interpretation of large-scale molecular data sets . Nucleic Acids Res . 2012 ; 40 ( Database issue ): D109 – D114 . 38. Tuttle RM , Haugen B , Perrier ND . Updated American Joint Committee on Cancer/tumor-node-metastasis staging system for differentiated and anaplastic thyroid cancer (eighth edition): what changed and why ? Thyroid 2017 ; 27 ( 6 ): 751 – 756 39. R: A language and environment for statistical computing [computer program]. Vienna, Austria: R Foundation for Statistical Computing: 2017. 40. Sha J , Xue W , Dong B , Pan J , Wu X , Li D , Liu D , Huang Y . PRKAR2B plays an oncogenic role in the castration-resistant prostate cancer . Oncotarget . 2017 ; 8 ( 4 ): 6114 – 6129 . 41. Vogelstein B , Papadopoulos N , Velculescu VE , Zhou S , Diaz LA Jr , Kinzler KW . Cancer genome landscapes . Science . 2013 ; 339 ( 6127 ): 1546 – 1558 . 42. Kunstman JW , Juhlin CC , Goh G , Brown TC , Stenman A , Healy JM , Rubinstein JC , Choi M , Kiss N , Nelson-Williams C , Mane S , Rimm DL , Prasad ML , Höög A , Zedenius J , Larsson C , Korah R , Lifton RP , Carling T . Characterization of the mutational landscape of anaplastic thyroid cancer via whole-exome sequencing . Hum Mol Genet . 2015 ; 24 ( 8 ): 2318 – 2329 . 43. Panebianco F , Kelly LM , Liu P , Zhong S , Dacic S , Wang X , Singhi AD , Dhir R , Chiosea SI , Kuan SF , Bhargava R , Dabbs D , Trivedi S , Gandhi M , Diaz R , Wald AI , Carty SE , Ferris RL , Lee AV , Nikiforova MN , Nikiforov YE . THADA fusion is a mechanism of IGF2BP3 activation and IGF1R signaling in thyroid cancer . Proc Natl Acad Sci USA . 2017 ; 114 ( 9 ): 2307 – 2312 . 44. Khagi Y , Goodman AM , Daniels GA , Patel SP , Sacco AG , Randall JM , Bazhenova LA , Kurzrock R . Hypermutated circulating tumor DNA: correlation with response to checkpoint inhibitor-based immunotherapy . Clin Cancer Res . 2017 ; 23 ( 19 ): 5729 – 5736 . Copyright © 2018 Endocrine Society

Pinnaduwage, Lakmini;Ye, Chang;Hanley, Anthony J;Connelly, Philip W;Sermer, Mathew;Zinman, Bernard;Retnakaran, Ravi

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2018-00306pmid: 29897453

Abstract Context Serum concentrations of liver enzymes and the hepatokine fetuin-A have been linked to the risk of type 2 diabetes, but their longitudinal impact on insulin resistance and β-cell dysfunction is unclear. Objective To evaluate the impact of changes over 2 years in fetuin-A and the liver enzymes alanine aminotransferase (ALT), aspartate aminotransferase (AST), and γ-glutamyltransferase (GGT) on changes in insulin sensitivity, β-cell function, and glycemia in women with varying degrees of previous gestational dysglycemia, reflecting a range of future diabetic risk. Design/Setting/Participants In total, 336 women underwent glucose challenge test (GCT) and oral glucose tolerance test (OGTT) in pregnancy, followed by repeat OGTT and measurement of ALT/AST/GGT/fetuin-A at both 1 year and 3 years postpartum. The antepartum GCT/OGTT identified four gestational glucose tolerance groups: gestational diabetes (n = 104), gestational impaired glucose tolerance (n = 59), abnormal GCT with normal OGTT (n = 98), and normal GCT/OGTT (n = 75). Results At 1 and 3 years postpartum, ALT, AST, GGT, and fetuin-A did not differ across the four groups, but the intervening change in ALT/AST ratio was greater in the gestational dysglycemia groups (P = 0.05). Higher baseline ALT/AST (t = −1.99, P = 0.05) and fetuin-A (t = −3.17, P = 0.002) predicted lower insulin sensitivity (Matsuda) at 3 years, as did their respective changes from 1 to 3 years (ALT/AST: t = −5.47, P < 0.0001; fetuin-A: t = −3.56, P = 0.0004). Change in ALT/AST predicted lower β-cell function (t = −2.33, P = 0.02) and higher fasting glucose at 3 years (t = 2.55, P = 0.01). Moreover, baseline fetuin-A predicted prediabetes/diabetes at 3 years (OR, 1.38; 95% CI, 1.01 to 1.88). Conclusion Circulating hepatic markers, particularly ALT/AST ratio and fetuin-A, track with changes in insulin sensitivity and β-cell function, supporting a pathophysiologic basis in their prediction of diabetic risk. Over the past 15 years, several large prospective epidemiologic studies have shown that modest elevations of serum liver enzymes, such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), and γ-glutamyl transferase (GGT), are associated with an increased risk of subsequently developing type 2 diabetes (T2DM) (1–3). More recently, similar associations with incident T2DM have been reported with fetuin-A, a novel serum glycoprotein that is secreted by the liver into the circulation at high concentration (a “hepatokine”) (4–7). Collectively, it is believed that perturbations of these markers may be reflecting hepatic fat, the deposition of which is often enhanced in those who develop T2DM (8, 9). However, despite an abundance of epidemiologic studies in which baseline measurement of these hepatic markers has been associated with the subsequent development of T2DM, evidence supporting causality in this regard is limited. Notably, there is a relative paucity of human data linking changes in hepatic markers with changes in insulin sensitivity and β-cell function over time, as might be expected for causal mediators. Reasons for this evidence gap include (1) the fact that most studies have measured these markers at baseline only, (2) few have assessed insulin sensitivity and β-cell function, and (3) even fewer have evaluated changes over time in these factors. Thus, recognizing these limitations, we sought to evaluate the longitudinal relationship over time between hepatic markers and concurrent changes in insulin sensitivity and β-cell function in a population at risk for T2DM. Glucose intolerance in pregnancy provides the unique opportunity to identify just such a patient population. Indeed, each degree of antepartum glucose intolerance [ranging from gestational diabetes mellitus (GDM) to milder gestational impaired glucose tolerance (GIGT) to lesser dysglycemia to normoglycemia] identifies a proportionate level of future risk of T2DM (one that is highest in GDM, followed by GIGT, etc) (10–12). Moreover, each degree of gestational glucose intolerance predicts distinct trajectories of insulin sensitivity, β-cell function, and glycemia in the first 3 years postpartum (12). Accordingly, the early postpartum years following gestational dysglycemia can provide a model of the early natural history of T2DM with which to evaluate the impact of hepatic markers on glucose homeostasis. In this setting, we hypothesized that changes in hepatic makers would be associated with changes in insulin sensitivity and/or β-cell function. Thus, our objective in this study was to characterize the impact of changes between 1 and 3 years postpartum in liver enzymes (ALT, AST, GGT) and fetuin-A on changes in insulin sensitivity, β-cell function, and glycemia in women with varying degrees of previous gestational dysglycemia and hence a range of future diabetic risk. Methods The study population consisted of women participating in a prospective observational cohort study in which we are evaluating the relationship between glucose tolerance in pregnancy and metabolic function in the years after delivery. The study protocol has been described in detail (12). In brief, women are first recruited at the time of antepartum screening for GDM in the late second/early third trimester and undergo metabolic characterization at recruitment and again at 3 months and 1 year postpartum. At the latter visit, they are recruited into an ongoing long-term observational cohort study in which participating women undergo serial metabolic characterization biannually thereafter. This analysis reports on the associations of hepatic markers with metabolic function in 336 women who have completed their 3-year postpartum visit. The study protocol has been approved by the Mount Sinai Hospital Research Ethics Board, and all women have provided written informed consent for their participation. Recruitment and determination of glucose tolerance status in pregnancy At our institution, all pregnant women are screened for GDM by a 50-g glucose challenge test (GCT), followed by referral for a diagnostic oral glucose tolerance test (OGTT) if the GCT is abnormal (plasma glucose ≥7.8 mmol/L at 1 hour after ingestion of a 50-g glucose load). For this study, women are recruited either before or after the GCT, and all participants undergo a 3-hour 100-g OGTT for ascertainment of gestational glucose tolerance status, regardless of the GCT result (i.e., even if normal). As previously described (12), the recruitment of women after an abnormal GCT serves to enrich the study population for those with varying degrees of glucose intolerance. The GCT and OGTT enable stratification of participants into the following gestational glucose tolerance groups: (1) GDM, defined by National Diabetes Data Group (NDDG) criteria, which require at least two of the following on the OGTT: fasting blood glucose ≥5.8 mmol/L, 1-hour glucose ≥10.6 mmol/L, 2-hour glucose ≥9.2 mmol/L, or 3-hour glucose ≥8.1 mmol/L (2) GIGT, defined by meeting only one of the above NDDG criteria (3) Abnormal GCT with normal glucose tolerance (NGT), defined by an abnormal GCT followed by NGT on the OGTT (i.e., meeting none of the NDDG criteria) (4) Normal GCT NGT, defined by a normal GCT followed by NGT on the OGTT These four groups reflect the full spectrum of future diabetic risk and have been previously shown to predict distinct trajectories of insulin sensitivity, β-cell function, and glycemia in the first 3 years postpartum (12). Assessments at 1 and 3 years postpartum Participants return to the clinical investigation unit at 1 and 3 years for metabolic characterization, including 2-hour 75-g OGTT on both occasions (12). On each OGTT, current glucose tolerance status was defined according to Canadian Diabetes Association (CDA) guidelines (13). CDA guidelines define prediabetes as impaired glucose tolerance (IGT), impaired fasting glucose (IFG), or combined IFG/IGT (13). All OGTTs were performed in the morning after an overnight fast, with venous blood samples drawn for measurement of glucose and specific insulin at fasting and at 30, 60, and 120 minutes following ingestion of the glucose load. Specific insulin was measured with the Roche-Elecsys-1010 immunoassay analyzer and electrochemiluminescence immunoassay kit (Roche Diagnostics, Laval, QB, Canada). On each OGTT, the primary measure of insulin sensitivity was the Matsuda index (14), with the homeostasis model assessment of insulin resistance (HOMA-IR) providing a secondary measure (15). The primary measure of β-cell function was the insulinogenic index (IGI)/HOMA-IR (16), with the Insulin Secretion-Sensitivity Index–2 (ISSI-2) providing a secondary measure (17, 18). Fetuin-A was measured by an enzyme-linked immunosorbent assay (ALPCO, Salem, NH), with a lower limit of detection of 5.0 μg/L and an upper limit of detection of 216 μg/L. ALT, AST, and GGT were measured by the routine clinical biochemistry laboratory at Mount Sinai Hospital. Statistical analyses All analyses were conducted using SAS 9.2 (SAS Institute, Cary, NC). All tests were two-sided and performed at a significance level of P < 0.05. Characteristics of the gestational glucose tolerance groups were compared at 1 year and 3 years postpartum by either one-way analysis of variance or Kruskal-Wallis test for continuous variables, or either χ2 or Fisher exact test for categorical variables (Table 1). In particular, changes in hepatic markers from 1 to 3 years were compared between these groups (i.e., groups that reflect different degrees of diabetic risk). Table 1. Demographic, Clinical, and Metabolic Characteristics of Study Population at 1 Year and 3 Years Postpartum, Stratified Into the Following Four Groups Based on Gestational Glucose Tolerance Status: Normal GCT NGT, Abnormal GCT NGT, GIGT, and GDM Characteristic Normal GCT NGT (n = 75) Abnormal GCT NGT (n = 98) GIGT (n = 59) GDM (n = 104) P Value At 1 y postpartum  Age, y 36 ± 4 36 ± 4 36 ± 4 36 ± 4 0.82  Ethnicity, n (%) 0.50   White 57 (76.0) 73 (74.5) 40 (67.8) 69 (66.4)   Asian 5 (6.7) 9 (9.2) 7 (11.9) 17 (16.4)   Other 13 (17.3) 16 (16.3) 12 (20.3) 18 (17.3)  Family history of T2DM, n (%) 40 (53.3) 58 (59.2) 38 (64.4) 68 (65.4) 0.38  Months breastfeeding, mo 11 (6–12) 10 (6–12) 9 (4–12) 10 (3–12) 0.43  Current smoking, n (%) 2 (2.7) 3 (3.1) 5 (8.8) 2 (2.0) 0.21  BMI, kg/m2 24.4 (21.5–28.4) 23.7 (21.8–27.8) 25.4 (23.1–30.1) 25.5 (22.5–29.5) 0.12  Waist circumference, cm 87 ± 12 85 ± 12 89 ± 13 90 ± 14 0.04  Insulin sensitivity/resistance   Matsuda index 11.0 (6.4–15.5) 10.2 (6.1–13.8) 7.5 (4.9–12.9) 7.7 (4.4–11.0) 0.003   HOMA-IR 1.1 (0.7–1.7) 1.1 (0.6–1.7) 1.4 (0.8–2.0) 1.2 (0.7–2.0) 0.14  β-Cell function   ISSI-2 877 ± 292 841 ± 324 655 ± 227 673 ± 267 <0.0001   Insulinogenic index/HOMA-IR 13.7 (8.7–22.8) 10.6 (6.6–16.4) 7.0 (4.0–9.1) 7.3 (4.5–12.1) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.3 4.6 ± 0.4 4.9 ± 0.6 4.9 ± 0.5 <0.0001   2-h glucose, mmol/L 5.4 ± 1.2 6.1 ± 1.6 6.4 ± 1.8 6.9 ± 1.9 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 69 (97.2) 82 (87.2) 42 (77.8) 69 (70.4)   Prediabetes/diabetes 2 (2.8) 12 (12.8) 12 (22.2) 29 (29.6)  ALT, IU/L 11 (9–15) 11 (9–15) 12 (9–17) 12 (9–14) 0.81  AST, IU/L 18 (17–22) 19 (17–21) 19 (16–24) 19 (16–23) 0.97  ALT/AST ratio 0.71 ± 0.38 0.63 ± 0.21 0.65 ± 0.29 0.64 ± 0.28 0.41  GGT, IU/L 10 (8.0–15) 11 (8–14) 12 (8–17) 12 (8–17) 0.32  Fetuin-A, g/L 0.5 (0.4–0.5) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.10 At 3 y postpartum  BMI, kg/m2 25.5 ± 4.5 25.4 ± 5.0 26.6 ± 4.8 26.9 ± 6.1 0.14  Waist circumference, cm 88 ± 12 86 ± 12 89 ± 12 90 ± 13 0.14  Insulin sensitivity/resistance   Matsuda index 10.0 (7.4–13.0) 8.2 (5.5–12.3) 7.3 (4.3–10.8) 6.2 (4.1–9.6) <0.0001   HOMA-IR 1.1 (0.7–1.8) 1.2 (0.7–1.9) 1.4 (1.0–2.1) 1.5 (1.0–2.4) 0.01  β-Cell function   ISSI-2 925 ± 331 898 ± 397 759 ± 333 668 ± 365 <0.0001   Insulinogenic index/HOMA-IR 13.1 (9.0–22.6) 10.7 (8.1–18.4) 8.7 (4.6–13.4) 6.8 (4.4–11.3) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.4 4.6 ± 0.5 4.8 ± 0.5 4.9 ± 0.6 <0.0001   2-h glucose, mmol/L 5.6 ± 1.2 6.1 ± 1.7 6.6 ± 2.1 7.4 ± 2.2 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 70 (93.3) 84 (85.7) 47 (80.0) 66 (63.5)   Prediabetes/diabetes 5 (6.7) 14 (14.3) 12 (20.0) 38 (36.5)  ALT, IU/L 15 (12–19) 15 (12–19) 15 (10–18) 16 (13–20) 0.69  AST, IU/L 18 (16–23) 18 (16–22) 17 (15–20) 18 (15–22) 0.34  ALT/AST ratio 0.86 ± 0.26 0.87 ± 0.28 0.90 ± 0.33 0.88 ± 0.22 0.80  GGT, IU/L 12 (9–17) 12 (10–17) 13 (10–18) 13 (11–22) 0.21  Fetuin-A, g/L 0.6 (0.5–0.7) 0.6 (0.5–0.8) 0.6 (0.5–0.8) 0.6 (0.5–0.7) 0.24 Changes from 1 to 3 y  Change in ALT 1.3 ± 11.7 4.4 ± 6.2 2.4 ± 10.7 3.8 ± 9.3 0.23  Change in AST −1.0 ± 8.6 −0.9 ± 5.5 −3.9 ± 8.1 −1.5 ± 7.5 0.11  Change in ALT/AST ratio 0.14 ± 0.36 0.26 ± 0.25 0.28 ± 0.32 0.25 ± 0.28 0.05  Change in GGT 0.8 ± 9.8 2.3 ± 7.1 4.6 ± 17.6 1.5 ± 12.6 0.38  Change in fetuin-A 0.1 ± 0.2 0.1 ± 0.3 0.1 ± 0.4 0.1 ± 0.3 0.98 Characteristic Normal GCT NGT (n = 75) Abnormal GCT NGT (n = 98) GIGT (n = 59) GDM (n = 104) P Value At 1 y postpartum  Age, y 36 ± 4 36 ± 4 36 ± 4 36 ± 4 0.82  Ethnicity, n (%) 0.50   White 57 (76.0) 73 (74.5) 40 (67.8) 69 (66.4)   Asian 5 (6.7) 9 (9.2) 7 (11.9) 17 (16.4)   Other 13 (17.3) 16 (16.3) 12 (20.3) 18 (17.3)  Family history of T2DM, n (%) 40 (53.3) 58 (59.2) 38 (64.4) 68 (65.4) 0.38  Months breastfeeding, mo 11 (6–12) 10 (6–12) 9 (4–12) 10 (3–12) 0.43  Current smoking, n (%) 2 (2.7) 3 (3.1) 5 (8.8) 2 (2.0) 0.21  BMI, kg/m2 24.4 (21.5–28.4) 23.7 (21.8–27.8) 25.4 (23.1–30.1) 25.5 (22.5–29.5) 0.12  Waist circumference, cm 87 ± 12 85 ± 12 89 ± 13 90 ± 14 0.04  Insulin sensitivity/resistance   Matsuda index 11.0 (6.4–15.5) 10.2 (6.1–13.8) 7.5 (4.9–12.9) 7.7 (4.4–11.0) 0.003   HOMA-IR 1.1 (0.7–1.7) 1.1 (0.6–1.7) 1.4 (0.8–2.0) 1.2 (0.7–2.0) 0.14  β-Cell function   ISSI-2 877 ± 292 841 ± 324 655 ± 227 673 ± 267 <0.0001   Insulinogenic index/HOMA-IR 13.7 (8.7–22.8) 10.6 (6.6–16.4) 7.0 (4.0–9.1) 7.3 (4.5–12.1) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.3 4.6 ± 0.4 4.9 ± 0.6 4.9 ± 0.5 <0.0001   2-h glucose, mmol/L 5.4 ± 1.2 6.1 ± 1.6 6.4 ± 1.8 6.9 ± 1.9 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 69 (97.2) 82 (87.2) 42 (77.8) 69 (70.4)   Prediabetes/diabetes 2 (2.8) 12 (12.8) 12 (22.2) 29 (29.6)  ALT, IU/L 11 (9–15) 11 (9–15) 12 (9–17) 12 (9–14) 0.81  AST, IU/L 18 (17–22) 19 (17–21) 19 (16–24) 19 (16–23) 0.97  ALT/AST ratio 0.71 ± 0.38 0.63 ± 0.21 0.65 ± 0.29 0.64 ± 0.28 0.41  GGT, IU/L 10 (8.0–15) 11 (8–14) 12 (8–17) 12 (8–17) 0.32  Fetuin-A, g/L 0.5 (0.4–0.5) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.10 At 3 y postpartum  BMI, kg/m2 25.5 ± 4.5 25.4 ± 5.0 26.6 ± 4.8 26.9 ± 6.1 0.14  Waist circumference, cm 88 ± 12 86 ± 12 89 ± 12 90 ± 13 0.14  Insulin sensitivity/resistance   Matsuda index 10.0 (7.4–13.0) 8.2 (5.5–12.3) 7.3 (4.3–10.8) 6.2 (4.1–9.6) <0.0001   HOMA-IR 1.1 (0.7–1.8) 1.2 (0.7–1.9) 1.4 (1.0–2.1) 1.5 (1.0–2.4) 0.01  β-Cell function   ISSI-2 925 ± 331 898 ± 397 759 ± 333 668 ± 365 <0.0001   Insulinogenic index/HOMA-IR 13.1 (9.0–22.6) 10.7 (8.1–18.4) 8.7 (4.6–13.4) 6.8 (4.4–11.3) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.4 4.6 ± 0.5 4.8 ± 0.5 4.9 ± 0.6 <0.0001   2-h glucose, mmol/L 5.6 ± 1.2 6.1 ± 1.7 6.6 ± 2.1 7.4 ± 2.2 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 70 (93.3) 84 (85.7) 47 (80.0) 66 (63.5)   Prediabetes/diabetes 5 (6.7) 14 (14.3) 12 (20.0) 38 (36.5)  ALT, IU/L 15 (12–19) 15 (12–19) 15 (10–18) 16 (13–20) 0.69  AST, IU/L 18 (16–23) 18 (16–22) 17 (15–20) 18 (15–22) 0.34  ALT/AST ratio 0.86 ± 0.26 0.87 ± 0.28 0.90 ± 0.33 0.88 ± 0.22 0.80  GGT, IU/L 12 (9–17) 12 (10–17) 13 (10–18) 13 (11–22) 0.21  Fetuin-A, g/L 0.6 (0.5–0.7) 0.6 (0.5–0.8) 0.6 (0.5–0.8) 0.6 (0.5–0.7) 0.24 Changes from 1 to 3 y  Change in ALT 1.3 ± 11.7 4.4 ± 6.2 2.4 ± 10.7 3.8 ± 9.3 0.23  Change in AST −1.0 ± 8.6 −0.9 ± 5.5 −3.9 ± 8.1 −1.5 ± 7.5 0.11  Change in ALT/AST ratio 0.14 ± 0.36 0.26 ± 0.25 0.28 ± 0.32 0.25 ± 0.28 0.05  Change in GGT 0.8 ± 9.8 2.3 ± 7.1 4.6 ± 17.6 1.5 ± 12.6 0.38  Change in fetuin-A 0.1 ± 0.2 0.1 ± 0.3 0.1 ± 0.4 0.1 ± 0.3 0.98 P values are for overall comparison across groups by one-way analysis of variance or Kruskal-Wallis test for continuous variables, or either χ2 or Fisher exact test for categorical variables. Continuous variables are presented as mean ± SD (if normally distributed) or median with interquartile range (if skewed). View Large Table 1. Demographic, Clinical, and Metabolic Characteristics of Study Population at 1 Year and 3 Years Postpartum, Stratified Into the Following Four Groups Based on Gestational Glucose Tolerance Status: Normal GCT NGT, Abnormal GCT NGT, GIGT, and GDM Characteristic Normal GCT NGT (n = 75) Abnormal GCT NGT (n = 98) GIGT (n = 59) GDM (n = 104) P Value At 1 y postpartum  Age, y 36 ± 4 36 ± 4 36 ± 4 36 ± 4 0.82  Ethnicity, n (%) 0.50   White 57 (76.0) 73 (74.5) 40 (67.8) 69 (66.4)   Asian 5 (6.7) 9 (9.2) 7 (11.9) 17 (16.4)   Other 13 (17.3) 16 (16.3) 12 (20.3) 18 (17.3)  Family history of T2DM, n (%) 40 (53.3) 58 (59.2) 38 (64.4) 68 (65.4) 0.38  Months breastfeeding, mo 11 (6–12) 10 (6–12) 9 (4–12) 10 (3–12) 0.43  Current smoking, n (%) 2 (2.7) 3 (3.1) 5 (8.8) 2 (2.0) 0.21  BMI, kg/m2 24.4 (21.5–28.4) 23.7 (21.8–27.8) 25.4 (23.1–30.1) 25.5 (22.5–29.5) 0.12  Waist circumference, cm 87 ± 12 85 ± 12 89 ± 13 90 ± 14 0.04  Insulin sensitivity/resistance   Matsuda index 11.0 (6.4–15.5) 10.2 (6.1–13.8) 7.5 (4.9–12.9) 7.7 (4.4–11.0) 0.003   HOMA-IR 1.1 (0.7–1.7) 1.1 (0.6–1.7) 1.4 (0.8–2.0) 1.2 (0.7–2.0) 0.14  β-Cell function   ISSI-2 877 ± 292 841 ± 324 655 ± 227 673 ± 267 <0.0001   Insulinogenic index/HOMA-IR 13.7 (8.7–22.8) 10.6 (6.6–16.4) 7.0 (4.0–9.1) 7.3 (4.5–12.1) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.3 4.6 ± 0.4 4.9 ± 0.6 4.9 ± 0.5 <0.0001   2-h glucose, mmol/L 5.4 ± 1.2 6.1 ± 1.6 6.4 ± 1.8 6.9 ± 1.9 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 69 (97.2) 82 (87.2) 42 (77.8) 69 (70.4)   Prediabetes/diabetes 2 (2.8) 12 (12.8) 12 (22.2) 29 (29.6)  ALT, IU/L 11 (9–15) 11 (9–15) 12 (9–17) 12 (9–14) 0.81  AST, IU/L 18 (17–22) 19 (17–21) 19 (16–24) 19 (16–23) 0.97  ALT/AST ratio 0.71 ± 0.38 0.63 ± 0.21 0.65 ± 0.29 0.64 ± 0.28 0.41  GGT, IU/L 10 (8.0–15) 11 (8–14) 12 (8–17) 12 (8–17) 0.32  Fetuin-A, g/L 0.5 (0.4–0.5) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.10 At 3 y postpartum  BMI, kg/m2 25.5 ± 4.5 25.4 ± 5.0 26.6 ± 4.8 26.9 ± 6.1 0.14  Waist circumference, cm 88 ± 12 86 ± 12 89 ± 12 90 ± 13 0.14  Insulin sensitivity/resistance   Matsuda index 10.0 (7.4–13.0) 8.2 (5.5–12.3) 7.3 (4.3–10.8) 6.2 (4.1–9.6) <0.0001   HOMA-IR 1.1 (0.7–1.8) 1.2 (0.7–1.9) 1.4 (1.0–2.1) 1.5 (1.0–2.4) 0.01  β-Cell function   ISSI-2 925 ± 331 898 ± 397 759 ± 333 668 ± 365 <0.0001   Insulinogenic index/HOMA-IR 13.1 (9.0–22.6) 10.7 (8.1–18.4) 8.7 (4.6–13.4) 6.8 (4.4–11.3) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.4 4.6 ± 0.5 4.8 ± 0.5 4.9 ± 0.6 <0.0001   2-h glucose, mmol/L 5.6 ± 1.2 6.1 ± 1.7 6.6 ± 2.1 7.4 ± 2.2 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 70 (93.3) 84 (85.7) 47 (80.0) 66 (63.5)   Prediabetes/diabetes 5 (6.7) 14 (14.3) 12 (20.0) 38 (36.5)  ALT, IU/L 15 (12–19) 15 (12–19) 15 (10–18) 16 (13–20) 0.69  AST, IU/L 18 (16–23) 18 (16–22) 17 (15–20) 18 (15–22) 0.34  ALT/AST ratio 0.86 ± 0.26 0.87 ± 0.28 0.90 ± 0.33 0.88 ± 0.22 0.80  GGT, IU/L 12 (9–17) 12 (10–17) 13 (10–18) 13 (11–22) 0.21  Fetuin-A, g/L 0.6 (0.5–0.7) 0.6 (0.5–0.8) 0.6 (0.5–0.8) 0.6 (0.5–0.7) 0.24 Changes from 1 to 3 y  Change in ALT 1.3 ± 11.7 4.4 ± 6.2 2.4 ± 10.7 3.8 ± 9.3 0.23  Change in AST −1.0 ± 8.6 −0.9 ± 5.5 −3.9 ± 8.1 −1.5 ± 7.5 0.11  Change in ALT/AST ratio 0.14 ± 0.36 0.26 ± 0.25 0.28 ± 0.32 0.25 ± 0.28 0.05  Change in GGT 0.8 ± 9.8 2.3 ± 7.1 4.6 ± 17.6 1.5 ± 12.6 0.38  Change in fetuin-A 0.1 ± 0.2 0.1 ± 0.3 0.1 ± 0.4 0.1 ± 0.3 0.98 Characteristic Normal GCT NGT (n = 75) Abnormal GCT NGT (n = 98) GIGT (n = 59) GDM (n = 104) P Value At 1 y postpartum  Age, y 36 ± 4 36 ± 4 36 ± 4 36 ± 4 0.82  Ethnicity, n (%) 0.50   White 57 (76.0) 73 (74.5) 40 (67.8) 69 (66.4)   Asian 5 (6.7) 9 (9.2) 7 (11.9) 17 (16.4)   Other 13 (17.3) 16 (16.3) 12 (20.3) 18 (17.3)  Family history of T2DM, n (%) 40 (53.3) 58 (59.2) 38 (64.4) 68 (65.4) 0.38  Months breastfeeding, mo 11 (6–12) 10 (6–12) 9 (4–12) 10 (3–12) 0.43  Current smoking, n (%) 2 (2.7) 3 (3.1) 5 (8.8) 2 (2.0) 0.21  BMI, kg/m2 24.4 (21.5–28.4) 23.7 (21.8–27.8) 25.4 (23.1–30.1) 25.5 (22.5–29.5) 0.12  Waist circumference, cm 87 ± 12 85 ± 12 89 ± 13 90 ± 14 0.04  Insulin sensitivity/resistance   Matsuda index 11.0 (6.4–15.5) 10.2 (6.1–13.8) 7.5 (4.9–12.9) 7.7 (4.4–11.0) 0.003   HOMA-IR 1.1 (0.7–1.7) 1.1 (0.6–1.7) 1.4 (0.8–2.0) 1.2 (0.7–2.0) 0.14  β-Cell function   ISSI-2 877 ± 292 841 ± 324 655 ± 227 673 ± 267 <0.0001   Insulinogenic index/HOMA-IR 13.7 (8.7–22.8) 10.6 (6.6–16.4) 7.0 (4.0–9.1) 7.3 (4.5–12.1) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.3 4.6 ± 0.4 4.9 ± 0.6 4.9 ± 0.5 <0.0001   2-h glucose, mmol/L 5.4 ± 1.2 6.1 ± 1.6 6.4 ± 1.8 6.9 ± 1.9 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 69 (97.2) 82 (87.2) 42 (77.8) 69 (70.4)   Prediabetes/diabetes 2 (2.8) 12 (12.8) 12 (22.2) 29 (29.6)  ALT, IU/L 11 (9–15) 11 (9–15) 12 (9–17) 12 (9–14) 0.81  AST, IU/L 18 (17–22) 19 (17–21) 19 (16–24) 19 (16–23) 0.97  ALT/AST ratio 0.71 ± 0.38 0.63 ± 0.21 0.65 ± 0.29 0.64 ± 0.28 0.41  GGT, IU/L 10 (8.0–15) 11 (8–14) 12 (8–17) 12 (8–17) 0.32  Fetuin-A, g/L 0.5 (0.4–0.5) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.5 (0.4–0.6) 0.10 At 3 y postpartum  BMI, kg/m2 25.5 ± 4.5 25.4 ± 5.0 26.6 ± 4.8 26.9 ± 6.1 0.14  Waist circumference, cm 88 ± 12 86 ± 12 89 ± 12 90 ± 13 0.14  Insulin sensitivity/resistance   Matsuda index 10.0 (7.4–13.0) 8.2 (5.5–12.3) 7.3 (4.3–10.8) 6.2 (4.1–9.6) <0.0001   HOMA-IR 1.1 (0.7–1.8) 1.2 (0.7–1.9) 1.4 (1.0–2.1) 1.5 (1.0–2.4) 0.01  β-Cell function   ISSI-2 925 ± 331 898 ± 397 759 ± 333 668 ± 365 <0.0001   Insulinogenic index/HOMA-IR 13.1 (9.0–22.6) 10.7 (8.1–18.4) 8.7 (4.6–13.4) 6.8 (4.4–11.3) <0.0001  OGTT   Fasting glucose, mmol/L 4.6 ± 0.4 4.6 ± 0.5 4.8 ± 0.5 4.9 ± 0.6 <0.0001   2-h glucose, mmol/L 5.6 ± 1.2 6.1 ± 1.7 6.6 ± 2.1 7.4 ± 2.2 <0.0001  Current glucose tolerance, n (%) <0.0001   Normal 70 (93.3) 84 (85.7) 47 (80.0) 66 (63.5)   Prediabetes/diabetes 5 (6.7) 14 (14.3) 12 (20.0) 38 (36.5)  ALT, IU/L 15 (12–19) 15 (12–19) 15 (10–18) 16 (13–20) 0.69  AST, IU/L 18 (16–23) 18 (16–22) 17 (15–20) 18 (15–22) 0.34  ALT/AST ratio 0.86 ± 0.26 0.87 ± 0.28 0.90 ± 0.33 0.88 ± 0.22 0.80  GGT, IU/L 12 (9–17) 12 (10–17) 13 (10–18) 13 (11–22) 0.21  Fetuin-A, g/L 0.6 (0.5–0.7) 0.6 (0.5–0.8) 0.6 (0.5–0.8) 0.6 (0.5–0.7) 0.24 Changes from 1 to 3 y  Change in ALT 1.3 ± 11.7 4.4 ± 6.2 2.4 ± 10.7 3.8 ± 9.3 0.23  Change in AST −1.0 ± 8.6 −0.9 ± 5.5 −3.9 ± 8.1 −1.5 ± 7.5 0.11  Change in ALT/AST ratio 0.14 ± 0.36 0.26 ± 0.25 0.28 ± 0.32 0.25 ± 0.28 0.05  Change in GGT 0.8 ± 9.8 2.3 ± 7.1 4.6 ± 17.6 1.5 ± 12.6 0.38  Change in fetuin-A 0.1 ± 0.2 0.1 ± 0.3 0.1 ± 0.4 0.1 ± 0.3 0.98 P values are for overall comparison across groups by one-way analysis of variance or Kruskal-Wallis test for continuous variables, or either χ2 or Fisher exact test for categorical variables. Continuous variables are presented as mean ± SD (if normally distributed) or median with interquartile range (if skewed). View Large Next, in Table 2, Spearman partial correlation analyses were conducted to assess the relationships between baseline-adjusted changes in each of the hepatic markers from 1 to 3 years with concurrent baseline-adjusted changes in metabolic factors (anthropometrics, insulin sensitivity/resistance, β-cell function, glycemia). We then proceeded to multiple linear regression analyses to determine whether changes in hepatic markers from 1 to 3 years were independently associated with the Matsuda index at 3 years (Table 3) and IGI/HOMA-IR at 3 years (Table 4) as per our main hypothesis, after complete adjustment for covariates. For each of these outcomes, we constructed the following five models to test the hepatic markers in turn: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each regression model was constructed with the following covariates: (1) clinical risk factors for diabetes [age, ethnicity, family history of diabetes, body mass index (BMI) at 1 year, change in BMI from 1 to 3 years, and duration of breastfeeding], (2) the baseline measure of the outcome variable at 1 year, and (3) the indicated hepatic marker at 1 year and its change from 1 to 3 years. Multiple linear regression analyses were similarly conducted for the secondary measures of insulin sensitivity (HOMA-IR) and β-cell function (ISSI-2) in Supplemental Tables 1 and 2, respectively. We then performed multiple linear regression analyses to determine whether changes in hepatic markers from 1 to 3 years were independently associated with fasting glucose at 3 years (Table 5), with model construction performed in the same way. Table 2. Partial Spearman Correlations of Baseline-Adjusted Changes in Hepatic Markers (ALT, AST, ALT/AST Ratio, GGT, Fetuin-A) With Baseline-Adjusted Changes in Metabolic Factors From 1 to 3 Years Postpartum Baseline-Adjusted Changes Between 1 and 3 y Baseline-Adjusted Change in ALT Between 1 and 3 y Baseline-Adjusted Change in AST Between 1 and 3 y Baseline-Adjusted Change in ALT/AST Ratio Between 1 and 3 y Baseline-Adjusted Change in GGT Between 1 and 3 y Baseline-Adjusted Change in Fetuin-A Between 1 and 3 y r P r P r P r P r P BMI 0.10 0.10 0.02 0.78 0.11 0.06 0.22 0.0003 0.06 0.33 Waist 0.15 0.01 0.05 0.42 0.17 0.006 0.13 0.03 0.04 0.53 Matsuda index −0.06 0.31 0.16 0.007 −0.20 0.001 −0.29 <0.0001 −0.24 <0.0001 HOMA-IR 0.07 0.22 −0.12 0.04 0.21 0.0005 0.20 0.0009 0.20 0.0005 Insulinogenic index/HOMA-IR −0.03 0.69 0.04 0.55 −0.08 0.23 −0.06 0.31 −0.11 0.07 ISSI-2 −0.10 0.08 −0.03 0.58 −0.10 0.10 −0.07 0.24 0.01 0.85 Fasting glucose 0.08 0.19 −0.01 0.82 0.09 0.15 0.12 0.04 0.06 0.30 Baseline-Adjusted Changes Between 1 and 3 y Baseline-Adjusted Change in ALT Between 1 and 3 y Baseline-Adjusted Change in AST Between 1 and 3 y Baseline-Adjusted Change in ALT/AST Ratio Between 1 and 3 y Baseline-Adjusted Change in GGT Between 1 and 3 y Baseline-Adjusted Change in Fetuin-A Between 1 and 3 y r P r P r P r P r P BMI 0.10 0.10 0.02 0.78 0.11 0.06 0.22 0.0003 0.06 0.33 Waist 0.15 0.01 0.05 0.42 0.17 0.006 0.13 0.03 0.04 0.53 Matsuda index −0.06 0.31 0.16 0.007 −0.20 0.001 −0.29 <0.0001 −0.24 <0.0001 HOMA-IR 0.07 0.22 −0.12 0.04 0.21 0.0005 0.20 0.0009 0.20 0.0005 Insulinogenic index/HOMA-IR −0.03 0.69 0.04 0.55 −0.08 0.23 −0.06 0.31 −0.11 0.07 ISSI-2 −0.10 0.08 −0.03 0.58 −0.10 0.10 −0.07 0.24 0.01 0.85 Fasting glucose 0.08 0.19 −0.01 0.82 0.09 0.15 0.12 0.04 0.06 0.30 Bold indicates P < 0.05. View Large Table 2. Partial Spearman Correlations of Baseline-Adjusted Changes in Hepatic Markers (ALT, AST, ALT/AST Ratio, GGT, Fetuin-A) With Baseline-Adjusted Changes in Metabolic Factors From 1 to 3 Years Postpartum Baseline-Adjusted Changes Between 1 and 3 y Baseline-Adjusted Change in ALT Between 1 and 3 y Baseline-Adjusted Change in AST Between 1 and 3 y Baseline-Adjusted Change in ALT/AST Ratio Between 1 and 3 y Baseline-Adjusted Change in GGT Between 1 and 3 y Baseline-Adjusted Change in Fetuin-A Between 1 and 3 y r P r P r P r P r P BMI 0.10 0.10 0.02 0.78 0.11 0.06 0.22 0.0003 0.06 0.33 Waist 0.15 0.01 0.05 0.42 0.17 0.006 0.13 0.03 0.04 0.53 Matsuda index −0.06 0.31 0.16 0.007 −0.20 0.001 −0.29 <0.0001 −0.24 <0.0001 HOMA-IR 0.07 0.22 −0.12 0.04 0.21 0.0005 0.20 0.0009 0.20 0.0005 Insulinogenic index/HOMA-IR −0.03 0.69 0.04 0.55 −0.08 0.23 −0.06 0.31 −0.11 0.07 ISSI-2 −0.10 0.08 −0.03 0.58 −0.10 0.10 −0.07 0.24 0.01 0.85 Fasting glucose 0.08 0.19 −0.01 0.82 0.09 0.15 0.12 0.04 0.06 0.30 Baseline-Adjusted Changes Between 1 and 3 y Baseline-Adjusted Change in ALT Between 1 and 3 y Baseline-Adjusted Change in AST Between 1 and 3 y Baseline-Adjusted Change in ALT/AST Ratio Between 1 and 3 y Baseline-Adjusted Change in GGT Between 1 and 3 y Baseline-Adjusted Change in Fetuin-A Between 1 and 3 y r P r P r P r P r P BMI 0.10 0.10 0.02 0.78 0.11 0.06 0.22 0.0003 0.06 0.33 Waist 0.15 0.01 0.05 0.42 0.17 0.006 0.13 0.03 0.04 0.53 Matsuda index −0.06 0.31 0.16 0.007 −0.20 0.001 −0.29 <0.0001 −0.24 <0.0001 HOMA-IR 0.07 0.22 −0.12 0.04 0.21 0.0005 0.20 0.0009 0.20 0.0005 Insulinogenic index/HOMA-IR −0.03 0.69 0.04 0.55 −0.08 0.23 −0.06 0.31 −0.11 0.07 ISSI-2 −0.10 0.08 −0.03 0.58 −0.10 0.10 −0.07 0.24 0.01 0.85 Fasting glucose 0.08 0.19 −0.01 0.82 0.09 0.15 0.12 0.04 0.06 0.30 Bold indicates P < 0.05. View Large Table 3. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of the Matsuda Index at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.78 0.44  Change in ALT from 1 to 3 y −0.01 −2.61 0.01 II  AST at 1 y 0.002 0.28 0.78  Change in AST from 1 to 3 y 0.02 3.23 0.001 III  ALT/AST ratio at 1 y −0.24 −1.99 0.05  Change in ALT/AST ratio from 1 to 3 y −0.59 −5.47 <0.0001 IV  GGT at 1 y −0.004 −1.42 0.16  Change in GGT from 1 to 3 y −0.008 −3.20 0.002 V  Fetuin-A at 1 y −0.47 −3.17 0.002  Change in fetuin-A from 1 to 3 y −0.40 −3.56 0.0004 Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.78 0.44  Change in ALT from 1 to 3 y −0.01 −2.61 0.01 II  AST at 1 y 0.002 0.28 0.78  Change in AST from 1 to 3 y 0.02 3.23 0.001 III  ALT/AST ratio at 1 y −0.24 −1.99 0.05  Change in ALT/AST ratio from 1 to 3 y −0.59 −5.47 <0.0001 IV  GGT at 1 y −0.004 −1.42 0.16  Change in GGT from 1 to 3 y −0.008 −3.20 0.002 V  Fetuin-A at 1 y −0.47 −3.17 0.002  Change in fetuin-A from 1 to 3 y −0.40 −3.56 0.0004 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and the Matsuda index at 1 y. View Large Table 3. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of the Matsuda Index at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.78 0.44  Change in ALT from 1 to 3 y −0.01 −2.61 0.01 II  AST at 1 y 0.002 0.28 0.78  Change in AST from 1 to 3 y 0.02 3.23 0.001 III  ALT/AST ratio at 1 y −0.24 −1.99 0.05  Change in ALT/AST ratio from 1 to 3 y −0.59 −5.47 <0.0001 IV  GGT at 1 y −0.004 −1.42 0.16  Change in GGT from 1 to 3 y −0.008 −3.20 0.002 V  Fetuin-A at 1 y −0.47 −3.17 0.002  Change in fetuin-A from 1 to 3 y −0.40 −3.56 0.0004 Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.78 0.44  Change in ALT from 1 to 3 y −0.01 −2.61 0.01 II  AST at 1 y 0.002 0.28 0.78  Change in AST from 1 to 3 y 0.02 3.23 0.001 III  ALT/AST ratio at 1 y −0.24 −1.99 0.05  Change in ALT/AST ratio from 1 to 3 y −0.59 −5.47 <0.0001 IV  GGT at 1 y −0.004 −1.42 0.16  Change in GGT from 1 to 3 y −0.008 −3.20 0.002 V  Fetuin-A at 1 y −0.47 −3.17 0.002  Change in fetuin-A from 1 to 3 y −0.40 −3.56 0.0004 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and the Matsuda index at 1 y. View Large Table 4. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of IGI/HOMA-IR at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.31 0.75  Change in ALT from 1 to 3 y −0.01 −1.78 0.08 II  AST at 1 y 0.02 1.90 0.06  Change in AST from 1 to 3 y 0.01 1.52 0.13 III  ALT/AST ratio at 1 y −0.59 −2.45 0.01  Change in ALT/AST ratio from 1 to 3 y −0.73 −3.35 0.001 IV  GGT at 1 y 0.0006 0.12 0.90  Change in GGT from 1 to 3 y −0.002 −0.39 0.70 V  Fetuin-A at 1 y −0.26 −0.85 0.40  Change in fetuin-A from 1 to 3 y −0.25 −1.08 0.28 Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.31 0.75  Change in ALT from 1 to 3 y −0.01 −1.78 0.08 II  AST at 1 y 0.02 1.90 0.06  Change in AST from 1 to 3 y 0.01 1.52 0.13 III  ALT/AST ratio at 1 y −0.59 −2.45 0.01  Change in ALT/AST ratio from 1 to 3 y −0.73 −3.35 0.001 IV  GGT at 1 y 0.0006 0.12 0.90  Change in GGT from 1 to 3 y −0.002 −0.39 0.70 V  Fetuin-A at 1 y −0.26 −0.85 0.40  Change in fetuin-A from 1 to 3 y −0.25 −1.08 0.28 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and IGI/HOMA-IR at 1 y. View Large Table 4. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of IGI/HOMA-IR at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.31 0.75  Change in ALT from 1 to 3 y −0.01 −1.78 0.08 II  AST at 1 y 0.02 1.90 0.06  Change in AST from 1 to 3 y 0.01 1.52 0.13 III  ALT/AST ratio at 1 y −0.59 −2.45 0.01  Change in ALT/AST ratio from 1 to 3 y −0.73 −3.35 0.001 IV  GGT at 1 y 0.0006 0.12 0.90  Change in GGT from 1 to 3 y −0.002 −0.39 0.70 V  Fetuin-A at 1 y −0.26 −0.85 0.40  Change in fetuin-A from 1 to 3 y −0.25 −1.08 0.28 Model/Hepatic Markers β t P I  ALT at 1 y −0.003 −0.31 0.75  Change in ALT from 1 to 3 y −0.01 −1.78 0.08 II  AST at 1 y 0.02 1.90 0.06  Change in AST from 1 to 3 y 0.01 1.52 0.13 III  ALT/AST ratio at 1 y −0.59 −2.45 0.01  Change in ALT/AST ratio from 1 to 3 y −0.73 −3.35 0.001 IV  GGT at 1 y 0.0006 0.12 0.90  Change in GGT from 1 to 3 y −0.002 −0.39 0.70 V  Fetuin-A at 1 y −0.26 −0.85 0.40  Change in fetuin-A from 1 to 3 y −0.25 −1.08 0.28 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and IGI/HOMA-IR at 1 y. View Large Table 5. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of Fasting Glucose at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y 0.004 1.01 0.31  Change in ALT from 1 to 3 y 0.007 1.98 0.05 II  AST at 1 y −0.003 −0.55 0.59  Change in AST from 1 to 3 y −0.002 −0.45 0.65 III  ALT/AST ratio at 1 y 0.20 1.62 0.11  Change in ALT/AST ratio from 1 to 3 y 0.29 2.55 0.01 IV  GGT at 1 y −0.003 −1.17 0.24  Change in GGT from 1 to 3 y 0.0008 0.32 0.75 V  Fetuin-A at 1 y −0.08 −0.53 0.59  Change in fetuin-A from 1 to 3 y 0.01 0.10 0.92 Model/Hepatic Markers β t P I  ALT at 1 y 0.004 1.01 0.31  Change in ALT from 1 to 3 y 0.007 1.98 0.05 II  AST at 1 y −0.003 −0.55 0.59  Change in AST from 1 to 3 y −0.002 −0.45 0.65 III  ALT/AST ratio at 1 y 0.20 1.62 0.11  Change in ALT/AST ratio from 1 to 3 y 0.29 2.55 0.01 IV  GGT at 1 y −0.003 −1.17 0.24  Change in GGT from 1 to 3 y 0.0008 0.32 0.75 V  Fetuin-A at 1 y −0.08 −0.53 0.59  Change in fetuin-A from 1 to 3 y 0.01 0.10 0.92 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and fasting glucose at 1 y. View Large Table 5. Hepatic Markers at 1 Year and Their Changes From 1 to 3 Years as Predictors of Fasting Glucose at 3 Years Postpartum Model/Hepatic Markers β t P I  ALT at 1 y 0.004 1.01 0.31  Change in ALT from 1 to 3 y 0.007 1.98 0.05 II  AST at 1 y −0.003 −0.55 0.59  Change in AST from 1 to 3 y −0.002 −0.45 0.65 III  ALT/AST ratio at 1 y 0.20 1.62 0.11  Change in ALT/AST ratio from 1 to 3 y 0.29 2.55 0.01 IV  GGT at 1 y −0.003 −1.17 0.24  Change in GGT from 1 to 3 y 0.0008 0.32 0.75 V  Fetuin-A at 1 y −0.08 −0.53 0.59  Change in fetuin-A from 1 to 3 y 0.01 0.10 0.92 Model/Hepatic Markers β t P I  ALT at 1 y 0.004 1.01 0.31  Change in ALT from 1 to 3 y 0.007 1.98 0.05 II  AST at 1 y −0.003 −0.55 0.59  Change in AST from 1 to 3 y −0.002 −0.45 0.65 III  ALT/AST ratio at 1 y 0.20 1.62 0.11  Change in ALT/AST ratio from 1 to 3 y 0.29 2.55 0.01 IV  GGT at 1 y −0.003 −1.17 0.24  Change in GGT from 1 to 3 y 0.0008 0.32 0.75 V  Fetuin-A at 1 y −0.08 −0.53 0.59  Change in fetuin-A from 1 to 3 y 0.01 0.10 0.92 Five multiple linear regression models were constructed to evaluate the following hepatic markers at 1 y and their changes from 1 to 3 y: ALT (model I), AST (model II), ALT/AST ratio (model III), GGT (model IV), and fetuin-A (model V). Each model was adjusted for age, ethnicity, family history of T2DM, BMI at 1 y, change in BMI from 1 to 3 y, breastfeeding, and fasting glucose at 1 y. View Large Finally, we conducted forward selection logistic regression analyses to determine independent predictors of prediabetes/diabetes at 3 years [defined as IFG, IGT, combined IFG/IGT, or diabetes, as per CDA guidelines (13)]. The covariates for selection were as follows: (1) diabetes risk factors (age, ethnicity, family history of diabetes, BMI at 1 year, change in BMI from 1 to 3 years, duration of breastfeeding), (2) glucose tolerance status at 1 year, and (3) all of the hepatic markers at 1 year (ALT, AST, GGT, fetuin-A) and the change in each hepatic marker from 1 to 3 years (Fig. 1). Age, ethnicity, family history of diabetes, BMI at 1 year, and duration of breastfeeding were forced into the model as major clinical risk factors for diabetes. The forward selection model was repeated with the inclusion of ALT/AST ratio and its change, in place of ALT and AST and their changes. Figure 1. View largeDownload slide Forward selection logistic regression model of (outcome) prediabetes/diabetes at 3 y postpartum, with the following covariates available for selection: age, ethnicity, family history of diabetes, BMI at 1 y, change in BMI from 1 to 3 y, duration of breastfeeding, glucose tolerance status at 1 y, all of the hepatic markers at 1 y (ALT, AST, GGT, and fetuin-A), and change in each hepatic marker from 1 to 3 y. Major clinical risk factors for diabetes (age, ethnicity, family history of diabetes, BMI at 1 y, and duration of breastfeeding) were forced into the model. Figure 1. View largeDownload slide Forward selection logistic regression model of (outcome) prediabetes/diabetes at 3 y postpartum, with the following covariates available for selection: age, ethnicity, family history of diabetes, BMI at 1 y, change in BMI from 1 to 3 y, duration of breastfeeding, glucose tolerance status at 1 y, all of the hepatic markers at 1 y (ALT, AST, GGT, and fetuin-A), and change in each hepatic marker from 1 to 3 y. Major clinical risk factors for diabetes (age, ethnicity, family history of diabetes, BMI at 1 y, and duration of breastfeeding) were forced into the model. Results Table 1 shows the characteristics of the study population at 1 year and 3 years postpartum, stratified into the following four groups based on their preceding gestational glucose tolerance status: (1) normal GCT NGT (n = 75), (2) abnormal GCT NGT (n = 98), (3) GIGT (n = 59), and (4) GDM (n = 104). At 1 year postpartum, the groups did not differ in age, ethnicity, family history of diabetes, duration of breastfeeding, smoking status, or BMI. As expected, at 1 year postpartum, there was a progressive decrease in insulin sensitivity (Matsuda index: P = 0.003) and β-cell function (IGI/HOMA-IR: P < 0.0001) from normal GCT NGT to abnormal GCT NGT to GIGT to GDM. Accordingly, there was a concomitant progressive rise in fasting glucose and 2-hour glucose across the four groups (both P < 0.0001), coupled with a stepwise increase in the prevalence of dysglycemia at 1 year from 2.8% to 12.8% to 22.2% to 29.6% (P < 0.0001), most of which (95%) was prediabetes. Of note, the hepatic markers ALT, AST, ALT/AST ratio, GGT, and fetuin-A did not differ across groups. At 3 years postpartum, the same patterns were apparent across the four gestational glucose tolerance groups for insulin sensitivity, β-cell function, fasting glucose, 2-hour glucose, and current glucose tolerance status. As before, the hepatic markers did not differ among the groups at 3 years. However, the change in ALT/AST ratio from 1 to 3 years differed across the groups (P = 0.05), with normal GCT NGT showing a smaller change than the three gestational dysglycemia groups. Changes in the other hepatic markers did not differ across the groups. After adjustment for diabetes risk factors (age, ethnicity, family history of diabetes, BMI, duration of breastfeeding) and current glucose tolerance, there were no statistically significant differences between the four gestational glucose tolerance groups in mean adjusted levels of any of the hepatic markers at 3 years (data not shown). Thus, between 1 and 3 years postpartum, the interval change in ALT/AST ratio was the only apparent hepatic marker difference between these four groups that reflected different degrees of future diabetic risk. Changes in hepatic markers and metabolic outcomes over time We next performed Spearman partial correlation analyses to evaluate the relationships between baseline-adjusted changes in hepatic markers and baseline-adjusted changes in metabolic factors from 1 to 3 years postpartum (Table 2). These analyses revealed that baseline-adjusted changes in ALT/AST ratio, GGT, and fetuin-A were inversely associated with baseline-adjusted changes in the Matsuda index, suggesting that increasing concentrations of these hepatic markers over time were accompanied by worsening insulin sensitivity. We next performed multiple linear regression analyses to determine whether any of the hepatic markers or their changes over time were independent predictors of the Matsuda index at 3 years, after adjustment for diabetes risk factors (age, ethnicity, family history of T2DM, breastfeeding, BMI at 1 year, change in BMI from 1 to 3 years) and the Matsuda index at 1 year (Table 3). Importantly, for each hepatic marker, its respective change from 1 to 3 years was an independent predictor of insulin sensitivity (Matsuda) at 3 years (Table 3). These relationships were strongest for ALT/AST ratio (t = −5.47, P < 0.0001) and fetuin-A (t = −3.56, P < 0.0001), both of which also showed independent associations of their baseline measures at 1 year with lower insulin sensitivity at 3 years. Similarly, the baseline measures at 1 year and changes from 1 to 3 years in ALT/AST ratio and fetuin-A both emerged as independent predictors of HOMA-IR at 3 years (Supplemental Table 1). Using the same modeling approach, we performed similar multiple linear regression analyses of our primary and secondary measures of β-cell function in Table 4 and Supplemental Table 2, respectively. These analyses showed that an increase in ALT/AST ratio from 1 to 3 years predicted poorer β-cell function as measured by either IGI/HOMA-IR (t = −3.35, P = 0.001) (Table 4) or ISSI-2 (t = −2.33, P = 0.02) (Supplemental Table 2). Moreover, similar multiple linear regression analyses of fasting glucose revealed that the increase in ALT/AST ratio from 1 to 3 years predicted higher fasting glycemia at 3 years (t = 2.55, P = 0.01) (Table 5). To evaluate the robustness of the findings from these multiple linear regression models in Tables 3 to 5 and Supplemental Tables 1 to 2, we performed a series of sensitivity analyses. The findings were largely unchanged in sensitivity analyses in which waist circumference at 1 year and the change in waist circumference from 1 to 3 years were included in the models, in place of baseline BMI and its change, respectively (data not shown). To address the possibility of false discovery, we also applied the correction method of Storey and confirmed that the findings remained unchanged (data not shown). Thus, from all of these analyses, ALT/AST ratio and fetuin-A emerged as hepatic markers that predicted pathophysiologic determinants of diabetes. Hepatic markers and prediabetes/diabetes Finally, we performed logistic regression analyses to determine whether any of the hepatic markers or their changes over time were independent predictors of prediabetes/diabetes at 3 years (Fig. 1). In a forward selection analysis that included baseline ALT, AST, GGT, and fetuin-A and their respective changes from 1 to 3 years as potential covariates, fetuin-A at 1 year emerged as a significant predictor of prediabetes/diabetes at 3 years (OR, 1.38; 95% CI, 1.01 to 1.88), accompanied by glucose intolerance at 1 year (OR, 11.0; 95% CI, 5.03 to 24.14), BMI at 1 year (OR, 1.44; 95% CI, 1.00 to 2.06), and the change in BMI from 1 to 3 years (OR, 1.55; 95% CI, 1.09 to 2.22). Furthermore, on sensitivity analyses with inclusion of ALT/AST ratio in place of ALT and AST, these findings were unchanged, with fetuin-A at 1 year again predicting prediabetes/diabetes (OR, 1.38; 95% CI, 1.01 to 1.88) (data not shown). Discussion In this study, we demonstrate three main findings. First, at both 1 and 3 years postpartum, the hepatic markers ALT, AST, GGT, and fetuin-A did not differ across four recent gestational tolerance groups reflecting different degrees of future diabetic risk. However, the intervening change in the ALT/AST ratio did differ between the groups. Second, changes over time in hepatic markers, particularly the ALT/AST ratio and fetuin-A, tracked with changes in insulin sensitivity and β-cell function. Third, and most important, at this early point in the natural history of T2DM, the ALT/AST ratio and fetuin-A emerged as independent predictors of fasting glycemia and prediabetes/diabetes, respectively. Taken together, these findings support a pathophysiologic basis in the relationship between hepatic markers and subsequent risk of T2DM. Although previous reports have linked fatty liver with insulin resistance in women with a history of GDM (19–21), few studies have evaluated circulating hepatic markers and their implications for diabetic risk in this population. In a cross-sectional study at mean 8 to 9 months postpartum, Rottenkolber et al. (22) reported that circulating fetuin-A was higher in 96 women with recent GDM compared with 51 controls. Forbes et al. (20) found that ALT and GGT did not differ between 110 women with previous GDM and 113 without such a history, at a mean 6 to 7 years postpartum. However, limitations of the studies to date have included modest samples sizes, cross-sectional evaluation at a single point in time, and potential heterogeneity in comparators (i.e., categorized as non-GDM, with identification after delivery). In this context, strengths of the current study are the prospective ascertainment of gestational glucose tolerance, yielding a well-characterized cohort of 336 women across the full spectrum of gestational glucose tolerance (from normal to GDM), coupled with assessment of both hepatic markers and metabolic outcomes on two occasions 2 years apart. With this approach, we demonstrate that, at both 1 and 3 years postpartum, ALT, AST, GGT, and fetuin-A did not differ across four recent gestational glucose tolerance groups that reflect distinct degrees of future diabetic risk. This study design also made it possible to evaluate the changes in hepatic markers over 2 years in these risk groups. In this regard, we observed that the three groups with higher future diabetic risk (GDM, GIGT, abnormal GCT NGT) experienced a greater increase in the ALT/AST ratio between 1 and 3 years than did the group with the lowest such risk (normal GCT NGT). Although an increased AST/ALT ratio (De Ritis ratio) has long been discussed clinically as a marker of liver disease (23), there has been limited previous evaluation in relation to the risk of T2DM. In a recent cross-sectional analysis from the Korean National Health and Nutrition Examination Survey, Ko et al. (24) reported that a lower AST/ALT ratio within the physiological range (i.e., hence higher ALT/AST ratio) was associated with a greater likelihood of IFG and T2DM. Similarly, in a cross-sectional study of nonobese Japanese adults, the ALT/AST ratio was associated with insulin resistance, as measured by HOMA-IR (25). Interestingly, when the Insulin Resistance and Atherosclerosis Study reported that this ratio predicted metabolic syndrome, the only component disorder thereof with which it was significantly associated in a series of separate models was IFG (26). Against this background, our findings suggest that, in this study population that is very early in the natural history of T2DM, the ALT/AST ratio may amplify a signal of future diabetic risk that is not yet detectable with individual liver enzymes alone. In this context, our evaluation of hepatic makers and metabolic function at two points in time is particularly informative by enabling not only the assessment of cross-sectional and longitudinal associations between these exposures and outcomes but also the relationship between their respective changes over time. We show that, although changes in all of the liver markers related to changes in insulin sensitivity (Table 3), the strongest predictors thereof were ALT/AST ratio and fetuin-A. Moreover, an increase in the ALT/AST ratio was the only hepatic marker that independently predicted lower β-cell function at 3 years, and it did so for both β-cell measures (Table 4, Supplemental Table 2). Finally, consistent with its associations with insulin resistance and β-cell dysfunction, rising ALT/AST ratio emerged as an independent predictor of higher fasting glucose at this early point in the natural history of diabetes (Table 5). Few previous studies have evaluated serial measurements of liver enzymes in relation to diabetic risk, although a secondary case-control analysis of the West of Scotland Coronary Prevention Study found elevations in ALT at 18 months, 12 months, and 6 months prior to the diagnosis of diabetes in 86 middle-aged male participants (27). The current study extends this concept by showing that a rising ALT/AST ratio tracks with insulin resistance, β-cell dysfunction, and fasting glucose early in the natural history of diabetes in young women. Although it has generally been believed that hepatic fat underlies the association between liver enzymes and diabetic risk (8, 9), a growing body of evidence suggests that this mechanistic basis may not fully apply to fetuin-A. First, in vitro studies have shown that fetuin-A reversibly binds the insulin receptor tyrosine kinase in peripheral tissues and can thereby directly induce insulin resistance (28, 29). Second, the Multiethnic Study of Atherosclerosis recently reported that baseline fetuin-A independently predicted incident diabetes in women, after adjustment for liver fat content (measured by computed tomography) (7). Third, in the Nurses’ Health Study, the circulating fetuin-A concentration predicted the subsequent risk of T2DM even after adjustment for liver enzymes (ALT and GGT) (6). Our findings are consistent with these observations and extend this literature in two ways. First, we demonstrate that both baseline fetuin-A and its change from 1 to 3 years postpartum were independently associated with lower whole-body insulin sensitivity (Table 3). Second, fetuin-A at 1 year emerged as an independent predictor of prediabetes/diabetes at 3 years in a forward selection analysis that included all of the liver enzymes, both at baseline and their change over time. Of note, it did so in the absence of an independent association with fasting glycemia, consistent with previous observations suggesting that elevated fetuin-A is a feature of IGT but not IFG (30, 31). Overall, our findings link both ALT/AST ratio and fetuin-A to early dysglycemia in young women but support the concept that they have distinct mechanistic bases in this regard. A limitation of this study is that some factors that could affect liver enzymes have not been evaluated. These include maternal infections, diet, and lifestyle practices. However, the current analyses were adjusted for major diabetes risk factors (age, ethnicity, family history, BMI) and breastfeeding history [which has been shown to have sustained effects on insulin sensitivity and long-term diabetic risk (32)]. Another limitation is that insulin sensitivity and β-cell function were assessed with OGTT-based surrogate indices rather than clamp studies. However, their time-consuming and invasive nature would have made clamp studies difficult to complete on two occasions over 2 years in 336 new mothers. Moreover, the Matsuda index, HOMA-IR, ISSI-2, and IGI/HOMA-IR are validated measures that have been widely used in previous studies (10–12, 14–18), and the serial OGTTs on which they were determined made it possible to also assess glucose tolerance status. In conclusion, at both 1 and 3 years postpartum, ALT, AST, GGT, and fetuin-A did not differ across four recent gestational tolerance groups reflecting different degrees of future diabetic risk. However, the interval change in the ALT/AST ratio differed between these groups. Moreover, change in the ALT/AST ratio and fetuin-A tracked with changes in insulin sensitivity and β-cell function, thereby linking these hepatic markers to the pathologic determinants of T2DM. Most important, the ALT/AST ratio and fetuin-A emerged as independent predictors of fasting glycemia and prediabetes/diabetes, respectively. Taken together, these findings thus support a pathophysiologic basis in the relationship between changes in circulating hepatic markers and diabetic risk very early in the natural history of T2DM in women. Abbreviations: Abbreviations: ALT alanine aminotransferase AST aspartate aminotransferase BMI body mass index CDA Canadian Diabetes Association GCT glucose challenge test GDM gestational diabetes mellitus GGT γ-glutamyltransferase GIGT gestational impaired glucose tolerance HOMA-IR homeostasis model assessment of insulin resistance IFG impaired fasting glucose IGI insulinogenic index IGT impaired glucose tolerance ISSI-2 Insulin Secretion-Sensitivity Index–2 NDDG National Diabetes Data Group NGT normal glucose tolerance OGTT oral glucose tolerance test T2DM type 2 diabetes mellitus Acknowledgments Financial Support: This study was supported by operating grants from the Canadian Institutes of Health Research (CIHR)(MOP-84206) and Canadian Diabetes Association (CDA)(CDA-OG-3-15-4924-RR) to R.R. A.J.H. holds a Tier-II Canada Research Chair in Diabetes Epidemiology. B.Z. holds the Sam and Judy Pencer Family Chair in Diabetes Research at Mount Sinai Hospital and University of Toronto. R.R. is supported by a Heart and Stroke Foundation of Ontario Mid-Career Investigator Award and holds the Boehringer Ingelheim Chair in Beta-cell Preservation, Function and Regeneration at Mount Sinai Hospital. Author Contributions: R.R., A.J.H., P.W.C., M.S., and B.Z. designed and implemented the study. R.R. and C.Y. contributed to the analysis plan and interpretation of the data. C.Y. performed the statistical analyses. L.P. wrote the first draft. All authors critically revised the manuscript for important intellectual content. All authors approved the final manuscript. R.R. is guarantor, had full access to all of the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis. Disclosure Summary: The authors have nothing to disclose. References 1. Hanley AJ , Williams K , Festa A , Wagenknecht LE , D’Agostino RB Jr , Kempf J , Zinman B , Haffner SM ; Insulin Resistance Atherosclerosis Study . Elevations in markers of liver injury and risk of type 2 diabetes: the insulin resistance atherosclerosis study . Diabetes . 2004 ; 53 ( 10 ): 2623 – 2632 . 2. Fraser A , Harris R , Sattar N , Ebrahim S , Davey Smith G , Lawlor DA . Alanine aminotransferase, γ-glutamyltransferase, and incident diabetes: the British Women’s Heart and Health Study and meta-analysis . Diabetes Care . 2009 ; 32 : 741 – 750 . 3. Schneider AL , Lazo M , Ndumele CE , Pankow JS , Coresh J , Clark JM , Selvin E . Liver enzymes, race, gender and diabetes risk: the Atherosclerosis Risk in Communities (ARIC) Study . Diabet Med . 2013 ; 30 ( 8 ): 926 – 933 . 4. Ix JH , Wassel CL , Kanaya AM , Vittinghoff E , Johnson KC , Koster A , Cauley JA , Harris TB , Cummings SR , Shlipak MG ; Health ABC Study . Fetuin-A and incident diabetes mellitus in older persons . JAMA . 2008 ; 300 ( 2 ): 182 – 188 . 5. Stefan N , Fritsche A , Weikert C , Boeing H , Joost HG , Häring HU , Schulze MB . Plasma fetuin-A levels and the risk of type 2 diabetes . Diabetes . 2008 ; 57 ( 10 ): 2762 – 2767 . 6. Sun Q , Cornelis MC , Manson JE , Hu FB . Plasma levels of fetuin-A and hepatic enzymes and risk of type 2 diabetes in women in the U.S . Diabetes . 2012 ; 62 ( 1 ): 49 – 55 . 7. Aroner SA , Mukamal KJ , St-Jules DE , Budoff MJ , Katz R , Criqui MH , Allison MA , de Boer IH , Siscovick DS , Ix JH , Jensen MK . Fetuin-A and risk of diabetes independent of liver fat content: the Multi-Ethnic Study of Atherosclerosis . Am J Epidemiol . 2016 ; 185 ( 1 ): 54 – 64 . 8. Kotronen A , Juurinen L , Hakkarainen A , Westerbacka J , Cornér A , Bergholm R , Yki-Järvinen H . Liver fat is increased in type 2 diabetic patients and underestimated by serum alanine aminotransferase compared with equally obese nondiabetic subjects . Diabetes Care . 2008 ; 31 : 165 – 169 . 9. Ballestri S , Zona S , Targher G , Romagnoli D , Baldelli E , Nascimbeni F , Roverato A , Guaraldi G , Lonardo A . Nonalcoholic fatty liver disease is associated with an almost twofold increased risk of incident type 2 diabetes and metabolic syndrome: evidence from a systematic review and meta-analysis . J Gastroenterol Hepatol . 2016 ; 31 ( 5 ): 936 – 944 . 10. Retnakaran R . Glucose tolerance status in pregnancy: a window to the future risk of diabetes and cardiovascular disease in young women . Curr Diabetes Rev . 2009 ; 5 ( 4 ): 239 – 244 . 11. Retnakaran R , Qi Y , Sermer M , Connelly PW , Hanley AJ , Zinman B . Glucose intolerance in pregnancy and future risk of pre-diabetes or diabetes . Diabetes Care . 2008 ; 31 ( 10 ): 2026 – 2031 . 12. Kramer CK , Swaminathan B , Hanley AJ , Connelly PW , Sermer M , Zinman B , Retnakaran R . Each degree of glucose intolerance in pregnancy predicts distinct trajectories of β-cell function, insulin sensitivity, and glycemia in the first 3 years postpartum . Diabetes Care . 2014 ; 37 ( 12 ): 3262 – 3269 . 13. Canadian Diabetes Association Clinical Practice Guidelines Expert Committee . Definition, classification and diagnosis of diabetes, prediabetes and metabolic syndrome . Can J Diabetes . 2013 ; 37 ( Suppl 1 ): S8 – S11 . 14. Matsuda M , DeFronzo RA . Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp . Diabetes Care . 1999 ; 22 ( 9 ): 1462 – 1470 . 15. Matthews DR , Hosker JP , Rudenski AS , Naylor BA , Treacher DF , Turner RC . Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man . Diabetologia . 1985 ; 28 ( 7 ): 412 – 419 . 16. Kahn SE . The relative contributions of insulin resistance and beta-cell dysfunction to the pathophysiology of type 2 diabetes . Diabetologia . 2003 ; 46 ( 1 ): 3 – 19 . 17. Retnakaran R , Shen S , Hanley AJ , Vuksan V , Hamilton JK , Zinman B . Hyperbolic relationship between insulin secretion and sensitivity on oral glucose tolerance test . Obesity (Silver Spring) . 2008 ; 16 ( 8 ): 1901 – 1907 . 18. Retnakaran R , Qi Y , Goran MI , Hamilton JK . Evaluation of proposed oral disposition index measures in relation to the actual disposition index . Diabet Med . 2009 ; 26 ( 12 ): 1198 – 1203 . 19. Tiikkainen M , Tamminen M , Häkkinen AM , Bergholm R , Vehkavaara S , Halavaara J , Teramo K , Rissanen A , Yki-Järvinen H . Liver-fat accumulation and insulin resistance in obese women with previous gestational diabetes . Obes Res . 2002 ; 10 ( 9 ): 859 – 867 . 20. Forbes S , Taylor-Robinson SD , Patel N , Allan P , Walker BR , Johnston DG . Increased prevalence of non-alcoholic fatty liver disease in European women with a history of gestational diabetes . Diabetologia . 2011 ; 54 : 641 – 647 . 21. Foghsgaard S , Andreasen C , Vedtofte L , Andersen ES , Bahne E , Strandberg C , Buhl T , Holst JJ , Svare JA , Clausen TD , Mathiesen ER , Damm P , Gluud LL , Knop FK , Vilsbøll T . Nonalcoholic fatty liver disease is prevalent in women with prior gestational diabetes mellitus and independently associated with insulin resistance and waist circumference . Diabetes Care . 2016 ; 40 ( 1 ): 109 – 116 . 22. Rottenkolber M , Ferrari U , Holland L , Aertsen S , Kammer NN , Hetterich H , Fugmann M , Banning F , Weise M , Sacco V , Kohn D , Freibothe I , Hutter S , Hasbargen U , Lehmann R , Grallert H , Parhofer KG , Seissler J , Lechner A . The diabetes risk phenotype of young women with recent gestational diabetes . J Clin Endocrinol Metab . 2015 ; 100 ( 6 ): E910 – E918 . 23. Botros M , Sikaris KA . The de Ritis ratio: the test of time . Clin Biochem Rev . 2013 ; 34 ( 3 ): 117 – 130 . 24. Ko SH , Baeg MK , Han KD , Ko SH , Ahn YB . Increased liver markers are associated with higher risk of type 2 diabetes . World J Gastroenterol . 2015 ; 21 ( 24 ): 7478 – 7487 . 25. Kawamoto R , Kohara K , Kusunoki T , Tabara Y , Abe M , Miki T . Alanine aminotransferase/aspartate aminotransferase ratio is the best surrogate marker for insulin resistance in non-obese Japanese adults . Cardiovasc Diabetol . 2012 ; 11 ( 1 ): 117 . 26. Hanley AJ , Williams K , Festa A , Wagenknecht LE , D’Agostino RB Jr , Haffner SM . Liver markers and development of the metabolic syndrome: the insulin resistance atherosclerosis study . Diabetes . 2005 ; 54 ( 11 ): 3140 – 3147 . 27. Sattar N , McConnachie A , Ford I , Gaw A , Cleland SJ , Forouhi NG , McFarlane P , Shepherd J , Cobbe S , Packard C . Serial metabolic measurements and conversion to type 2 diabetes in the west of Scotland coronary prevention study: specific elevations in alanine aminotransferase and triglycerides suggest hepatic fat accumulation as a potential contributing factor . Diabetes . 2007 ; 56 ( 4 ): 984 – 991 . 28. Rauth G , Pöschke O , Fink E , Eulitz M , Tippmer S , Kellerer M , Häring HU , Nawratil P , Haasemann M , Jahnen-Dechent W , Muller-Esterl W . The nucleotide and partial amino acid sequences of rat fetuin: identity with the natural tyrosine kinase inhibitor of the rat insulin receptor . Eur J Biochem . 1992 ; 204 ( 2 ): 523 – 529 . 29. Mathews ST , Srinivas PR , Leon MA , Grunberger G . Bovine fetuin is an inhibitor of insulin receptor tyrosine kinase . Life Sci . 1997 ; 61 ( 16 ): 1583 – 1592 . 30. Ou HY , Yang YC , Wu HT , Wu JS , Lu FH , Chang CJ . Increased fetuin-A concentrations in impaired glucose tolerance with or without nonalcoholic fatty liver disease, but not impaired fasting glucose . J Clin Endocrinol Metab . 2012 ; 97 ( 12 ): 4717 – 4723 . 31. Laughlin GA , Barrett-Connor E , Cummins KM , Daniels LB , Wassel CL , Ix JH . Sex-specific association of fetuin-A with type 2 diabetes in older community-dwelling adults: the Rancho Bernardo study . Diabetes Care . 2013 ; 36 ( 7 ): 1994 – 2000 . 32. Bajaj H , Ye C , Hanley AJ , Connelly PW , Sermer M , Zinman B , Retnakaran R . Prior lactation reduces future diabetic risk through sustained postweaning effects on insulin sensitivity . Am J Physiol Endocrinol Metab . 2017 ; 312 ( 3 ): E215 – E223 . Copyright © 2018 Endocrine Society

Geoffron, Sophie;Habib, Walid Abi;Chantot-Bastaraud, Sandra;Dubern, Béatrice;Steunou, Virginie;Azzi, Salah;Afenjar, Alexandra;Busa, Tiffanny;Canton, Ana Pinheiro;Chalouhi, Christel;Dufourg, Marie-Noëlle;Esteva, Blandine;Fradin, Mélanie;Geneviève, David;Heide, Solveig;Isidor, Bertrand;Linglart, Agnès;Picard, Fanny Morice;Naud-Saudreau, Catherine;Petit, Isabelle Oliver;Philip, Nicole;Pienkowski, Catherine;Rio, Marlène;Rossignol, Sylvie;Tauber, Maithé;Thevenon, Julien;Vu-Hong, Thuy-Ai;Harbison, Madeleine D;Salem, Jennifer;Brioude, Frédéric;Netchine, Irène;Giabicani, Eloïse

2018 Journal of Clinical Endocrinology and Metabolism

doi: 10.1210/jc.2017-02152pmid: 29659920

Abstract Context Silver-Russell syndrome (SRS) (mainly secondary to 11p15 molecular disruption) and Temple syndrome (TS) (secondary to 14q32.2 molecular disruption) are imprinting disorders with phenotypic (prenatal and postnatal growth retardation, early feeding difficulties) and molecular overlap. Objective To describe the clinical overlap between SRS and TS and extensively study the molecular aspects of TS. Patients We retrospectively collected data on 28 patients with disruption of the 14q32.2 imprinted region, identified in our center, and performed extensive molecular analysis. Results Seventeen (60.7%) patients showed loss of methylation of the MEG3/DLK1 intergenic differentially methylated region by epimutation. Eight (28.6%) patients had maternal uniparental disomy of chromosome 14 and three (10.7%) had a paternal deletion in 14q32.2. Most patients (72.7%) had a Netchine-Harbison SRS clinical scoring system ≥4/6, and consistent with a clinical diagnosis of SRS. The mean age at puberty onset was 7.2 years in girls and 9.6 years in boys; 37.5% had premature pubarche. The body mass index of all patients increased before pubarche and/or the onset of puberty. Multilocus analysis identified multiple methylation defects in 58.8% of patients. We identified four potentially damaging genetic variants in genes encoding proteins involved in the establishment or maintenance of DNA methylation. Conclusions Most patients with 14q32.2 disruption fulfill the criteria for a clinical diagnosis of SRS. These clinical data suggest similar management of patients with TS and SRS, with special attention to their young age at the onset of puberty and early increase of body mass index. Impaired fetal growth is associated with an increased risk of perinatal morbidity and mortality and metabolic problems later in life, according to the Developmental Origins of Health and Disease theory (1). Imprinted regions are known to play an important role in fetal growth (2). Paternally expressed genes are mostly involved in growth promotion, whereas maternally expressed genes repress it. Most imprinted regions are methylated on the maternal allele. In humans, only two regions are methylated in the male germ line (3, 4), the 11p15 H19/IGF2 intergenic differentially methylated region (IG-DMR) and the 14q32.2 MEG3/DLK1:IG-DMR, involved in Silver-Russell syndrome (SRS)/Beckwith-Wiedemann syndrome, and Temple syndrome (TS)/Kagami-Ogata syndrome, respectively. SRS is characterized by fetal and postnatal growth retardation and feeding difficulties (5–9). Epimutation, resulting in the loss of methylation (LOM) of H19/IGF2:IG-DMR on the paternal allele, is identified in 50% of SRS cases (10–12). In this region, the imprinting center H19/IGF2:IG-DMR is methylated on the paternal allele, resulting in IGF2 expression. When unmethylated, as on the maternal allele, it allows H19 expression, a long noncoding RNA (Fig. 1). The key role of IGF2 in prenatal growth is well-established. Maternal uniparental disomy for chromosome 7 [upd(7)mat] is seen in ∼5% to 10% of patients with SRS (13). However, for 35% to 40% of patients with SRS, the molecular etiology remains unknown. Figure 1. View largeDownload slide Scheme of normal 11p15 and 14q32.2 chromosomal regions. DIO3, iodothyronine deiodinase 3; DLK1, delta-like homolog 1; H19, long noncoding RNA; IGF2, insulin-like growth factor 2; MEG3, MEG8, maternally expressed genes 3 and 8; RTL1, retrotransposon-like 1;snoRNAs, small nucleolar RNAs; miRNAs, microRNAs. Figure 1. View largeDownload slide Scheme of normal 11p15 and 14q32.2 chromosomal regions. DIO3, iodothyronine deiodinase 3; DLK1, delta-like homolog 1; H19, long noncoding RNA; IGF2, insulin-like growth factor 2; MEG3, MEG8, maternally expressed genes 3 and 8; RTL1, retrotransposon-like 1;snoRNAs, small nucleolar RNAs; miRNAs, microRNAs. TS, first clinically and molecularly described in 1991, associates fetal and postnatal growth retardation, hypotonia, obesity, and early puberty (14). TS is caused by disruption of the 14q32.2 imprinted region, where MEG3/DLK1:IG-DMR is methylated on the paternal allele. MEG3/DLK1:IG-DMR methylation results in DLK1, RTL1, and DIO3 expression, whereas long noncoding RNAs (MEG3 and MEG8), microRNAs, and small nucleolar RNAs are expressed when it is unmethylated (as on the maternal allele) (Fig. 1). In a meta-analysis of 51 patients with TS, the molecular anomalies identified consisted mostly of maternal uniparental disomy of chromosome 14 [upd(14)mat] (78.4%), epimutation of MEG3/DLK1:IG-DMR on the paternal allele (11.8%), and paternal deletion of the MEG3/DLK1 domain (9.8%) (15). A cohort of 32 patients with 14q32.2 anomalies has recently been reported and, again, most had upd(14)mat (71.9%), whereas only 18.8% had epimutations (16). Clinical overlap between SRS and TS has been previously highlighted in reports of patients presenting with a clinical diagnosis of SRS with no 11p15 disruption or upd(7)mat, but for whom chromosome 14q32.2 anomalies were identified (17–19). Thus, these syndromes overlap in terms of phenotype and may be caused by anomalies of imprinted regions sharing similar molecular organization, both methylated in the male germ line. We identified 28 patients with chromosome 14q32.2 disruption in our molecular diagnostic laboratory. The purpose of this study was to clinically and molecularly characterize these patients to determine the clinical overlap with patients with SRS. Furthermore, we sought to identify the mechanism involved in the onset of 14q32.2 epimutation. Patients and Methods Study population The study population consisted of 28 patients with chromosome 14q32.2 disruption. The molecular diagnosis of 25 patients was performed in our laboratory and three upd(14)mat were identified without methylation analysis in other diagnostic laboratories. All patients were either followed in our clinic or were referred by other clinical centers for molecular analysis. A clinical file, including extensive clinical data, growth charts, a detailed phenotypic description, and pictures was completed for all patients. Each patient had been examined by a geneticist and/or a pediatric endocrinologist. Written informed consent for participation was received either from the patients themselves or their parents, in accordance with French national ethics rules for patients recruited in France (Assistance Publique – Hôpitaux de Paris authorization no. 681) and with the institutional review board I00000204 of the Mount Sinai School of Medicine, New York, for patients recruited in the United States. Clinical assessment and definitions The Netchine-Harbison clinical scoring system (NH-CSS), recently adopted as the clinical definition of SRS by the first international consensus on this syndrome (9, 11), was applied to each of the 28 patients. This scoring system defines a suspicion of SRS if at least four of the six following criteria are met: (1) being born small for gestational age (SGA) [birth weight and/or birth length ≤−2 standard deviation score (SDS) for gestational age], (2) postnatal growth failure (height at 24 ± 1 months ≤−2 SDS or height ≤−2 SDS from midparental target height), (3) relative macrocephaly at birth (head circumference at birth ≥1.5 SDS above birth weight and/or length SDS), (4) protruding forehead (forehead projecting beyond the facial plane on a side view as a toddler), (5) body asymmetry [leg length discrepancy ≥0.5 cm or arm asymmetry or leg length discrepancy <0.5 cm with at least two other asymmetrical body parts (one nonface)], and (6) low body mass index (BMI) (BMI ≤−2 SDS at 24 months) and/or feeding difficulties defined by the use of a feeding tube and/or cyproheptadine for appetite stimulation. See the Supplemental Materials and Methods for auxologic methods. Premature pubarche was defined by the appearance of pubic or axillary hair occurring before eight years in girls and nine years in boys (20). Precocious puberty was defined by breast development (thelarche) before age 8 years in girls and testicular enlargement before age 9 years in boys (21). Exaggerated adrenarche was defined by high levels of serum dehydroepiandrosterone sulfate for age (after other diseases were excluded) (22). Molecular analysis Methylation studies at both 11p15 H19/IGF2:IG-DMR and MEG3/DLK1:IG-DMR loci are described in the Supplemental Data. All patients had hypomethylation at MEG3/DLK1:IG-DMR. We distinguished three different mechanisms: upd(14)mat, deletion, and LOM by epimutation. Single nucleotide polymorphism microarray analysis We analyzed the DNA samples using Illumina CytoSNP-12 arrays (Illumina, San Diego, CA) to distinguish between MEG3/DLK1:IG-DMR epimutation, upd(14)mat, and large copy number variations. See the Supplemental Materials and Methods for details. IG-DMR and exome variant sequencing Whole-exome sequencing Library preparation, exome capture, sequencing, and data analysis were performed by IntegraGen SA (Evry, France). The sequencing methods and bioinformatics analysis are detailed in the Supplemental Materials and Methods. Statistical analysis The characteristics of the population are described as percentages for qualitative variables or as SDS and mean (range) for continuous variables. For subgroup comparisons, we used the Wilcoxon Mann-Whitney test and the Fischer test. Results Patients Our cohort was composed of 28 patients (17 girls). Three patients (10.7%) were conceived with the aid of medically assisted procreation: two in vitro fecundations and one intrauterine insemination. The median maternal age was 29.1 (19.8–41.5) years and paternal age was 30.7 (25.8-44.8) years. Molecular diagnosis Classical molecular anomalies found in SRS [i.e., 11p15 epimutation and upd(7)mat] were ruled out for all but three patients with upd(14)mat not identified in our laboratory. All patients presented with chromosome 14q32.2 hypomethylation at the MEG3/DLK1:IG-DMR, which was secondary to upd(14)mat in eight (28.6%) patients or to a paternal deletion of DLK1/MEG3 region in three (10.7%), whereas 17 (60.7%) had MEG3/DLK1:IG-DMR LOM caused by epimutation on the paternal allele. This was ascertained after ruling out a upd(14)mat or deletion of the DLK1/MEG3 region by single nucleotide polymorphism microarray. Clinical features The median age at the end of the study was 7.5 (1.3 to 21.6) years. Birth parameters, postnatal growth, dysmorphic anomalies, psychomotor development, and associated malformations are summarized in Table 1. Dysmorphic features such as protruding forehead, prominent heel, tented appearance of the mouth, and acromicria are presented in Supplemental Fig. 1. Table 1. Main Clinical Features of the Cohort All Patients Epimutations (1) Upd(14)mat (2) (1) vs (2) Paternal Deletion n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) P n (%) Mean (Min–Max) <−2 SDS (%) Birth  Term (wk of amenorrhea) 28 37.5 (28.0–40.0) 17 37.2 (28.0–40.0) 8 37.6 (36.0–39.0) 1 3 38.7 (37.0–40.0)  Birth length (SDS) 28 −2.3 (−5.2 to 0.8) 16 (57.1) 17 −2.3 (−5.2 to −0.9) 9 (52.9) 8 −2.4 (−3.7 to 0.8) 6 (75.0) 0.37 3 −1.6 (−2.2 to −1.2) 1 (33.3)  Birth weight (SDS) 28 −2.4 (−4.0 to −0.7) 24 (85.7) 17 −2.4 (−4.0 to −0.9) 14 (82.4) 8 −2.8 (−3.3 to −2.3) 8 (100) 0.22 3 −2.0 (−3.1 to −0.7) 2 (66.7)  Intrauterine growth retardation 25/28 (89.3) 15/17 (88.2) 8/8 (100) 1 2/3 (66.7)  Birth head circumference (SDS) 25 −1.3 (−3.7 to 1.0) 15 −1.3 (−3.7 to 1.0) 7 −1.5 (−2.0 to −0.7) 0.46 3 −0.7 (−0.6 to −0.7)  Relative macrocephaly 14/25 (56.0) 7/15 (46.7) 5/7 (71.4) 0.73 2/3 (66.7) Growth  Height at 2 y (SDS) 25 −2.0 (−4.4 to −0.2) 14/25 (56.0) 15 −2.0 (−4.4 to −0.2) 9 (60.0) 8 −1.8 (−4.1 to −0.6) 3 (37.5) 0.56 2 −2.5 (−2.6 to −2.4) 2/2 (100)  BMI at 2 y (SDS) 24 −1.5 (−3.2 to 0.0) 8 (33.3) 15 −1.4 (−3.2 to 0.0) 5 (33.3) 7 −1.8 (−2.9 to −0.6) 3 (42.8) 0.27 2 −1.7 (−1.9 to −1.4) 0  Early feeding difficulties 17/28 (60.7) 9/17 (52.9) 6/8 (75.0) 0.71 2/3 (66.7) NH-CSS  NH-CSS ≥4 16/22 (72.7) 9/14 (64.3) 5/5 (100.0) 0.71 2/3 (66.7) Clinical signs  Protruding forehead 25/28 (89.3) 17/17 (100) 6/8 (75.0) 0.76 2/3 (66.7)  Body asymmetry 7/26 (26.9) 6/17 (35.3) 1/6 (16.7) 1 0/3 (0)  Acromicria 19/26 (73.1) 12/17 (70.6) 6/7 (85.7) 1 1/2 (50.0)  Downturned mouth 18/24 (75.0) 12/15 (80.0) 5/7 (71.4) 1 1/2 (50.0)  Low-set posteriorly rotated ears 14/25 (56.0) 11/16 (68.8) 2/7 (28.6) 0.44 1/2 (50.0)  Clinodactyly 16/26 (61.5) 10/16 (62.5) 5/8 (62.5) 1 1/2 (50.0) Development  Neonatal hypotonia 14/20 (70.0) 5/10 (50.0) 6/7 (85.7) 0.70 3/3 (100)  Motor delay 16/25 (64.0) 8/15 (53.3) 5/7 (71.4) 0.72 3/3 (100)  Speech delay 12/23 (52.2) 7/15 (46.7) 2/5 (40.0) 1 3/3 (100)  Normal schooling 13/19 (68.4) 10/12 (83.3) 3/5 (60.0) 1 0/2 (0.0)  Behavioral disorders 3/23 (13.0) 2/15 (13.3) 1/5 (20.0) 0.5 0/3 (0.0) Associated malformations  Urogenitala 5/23 (21.7) 3/12 (25.0) 1/8 (12.5) 1 1/3 (33.3)  Stomatologyb 7/21 (33.3) 6/13 (46.2) 1/5 (20.0%) 0.64 0/3 (0.0)  Heartc 2/17 (11.8) 0/9 (0.0) 1/6 (16.7%) 0.44 1/2 (50.0) All Patients Epimutations (1) Upd(14)mat (2) (1) vs (2) Paternal Deletion n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) P n (%) Mean (Min–Max) <−2 SDS (%) Birth  Term (wk of amenorrhea) 28 37.5 (28.0–40.0) 17 37.2 (28.0–40.0) 8 37.6 (36.0–39.0) 1 3 38.7 (37.0–40.0)  Birth length (SDS) 28 −2.3 (−5.2 to 0.8) 16 (57.1) 17 −2.3 (−5.2 to −0.9) 9 (52.9) 8 −2.4 (−3.7 to 0.8) 6 (75.0) 0.37 3 −1.6 (−2.2 to −1.2) 1 (33.3)  Birth weight (SDS) 28 −2.4 (−4.0 to −0.7) 24 (85.7) 17 −2.4 (−4.0 to −0.9) 14 (82.4) 8 −2.8 (−3.3 to −2.3) 8 (100) 0.22 3 −2.0 (−3.1 to −0.7) 2 (66.7)  Intrauterine growth retardation 25/28 (89.3) 15/17 (88.2) 8/8 (100) 1 2/3 (66.7)  Birth head circumference (SDS) 25 −1.3 (−3.7 to 1.0) 15 −1.3 (−3.7 to 1.0) 7 −1.5 (−2.0 to −0.7) 0.46 3 −0.7 (−0.6 to −0.7)  Relative macrocephaly 14/25 (56.0) 7/15 (46.7) 5/7 (71.4) 0.73 2/3 (66.7) Growth  Height at 2 y (SDS) 25 −2.0 (−4.4 to −0.2) 14/25 (56.0) 15 −2.0 (−4.4 to −0.2) 9 (60.0) 8 −1.8 (−4.1 to −0.6) 3 (37.5) 0.56 2 −2.5 (−2.6 to −2.4) 2/2 (100)  BMI at 2 y (SDS) 24 −1.5 (−3.2 to 0.0) 8 (33.3) 15 −1.4 (−3.2 to 0.0) 5 (33.3) 7 −1.8 (−2.9 to −0.6) 3 (42.8) 0.27 2 −1.7 (−1.9 to −1.4) 0  Early feeding difficulties 17/28 (60.7) 9/17 (52.9) 6/8 (75.0) 0.71 2/3 (66.7) NH-CSS  NH-CSS ≥4 16/22 (72.7) 9/14 (64.3) 5/5 (100.0) 0.71 2/3 (66.7) Clinical signs  Protruding forehead 25/28 (89.3) 17/17 (100) 6/8 (75.0) 0.76 2/3 (66.7)  Body asymmetry 7/26 (26.9) 6/17 (35.3) 1/6 (16.7) 1 0/3 (0)  Acromicria 19/26 (73.1) 12/17 (70.6) 6/7 (85.7) 1 1/2 (50.0)  Downturned mouth 18/24 (75.0) 12/15 (80.0) 5/7 (71.4) 1 1/2 (50.0)  Low-set posteriorly rotated ears 14/25 (56.0) 11/16 (68.8) 2/7 (28.6) 0.44 1/2 (50.0)  Clinodactyly 16/26 (61.5) 10/16 (62.5) 5/8 (62.5) 1 1/2 (50.0) Development  Neonatal hypotonia 14/20 (70.0) 5/10 (50.0) 6/7 (85.7) 0.70 3/3 (100)  Motor delay 16/25 (64.0) 8/15 (53.3) 5/7 (71.4) 0.72 3/3 (100)  Speech delay 12/23 (52.2) 7/15 (46.7) 2/5 (40.0) 1 3/3 (100)  Normal schooling 13/19 (68.4) 10/12 (83.3) 3/5 (60.0) 1 0/2 (0.0)  Behavioral disorders 3/23 (13.0) 2/15 (13.3) 1/5 (20.0) 0.5 0/3 (0.0) Associated malformations  Urogenitala 5/23 (21.7) 3/12 (25.0) 1/8 (12.5) 1 1/3 (33.3)  Stomatologyb 7/21 (33.3) 6/13 (46.2) 1/5 (20.0%) 0.64 0/3 (0.0)  Heartc 2/17 (11.8) 0/9 (0.0) 1/6 (16.7%) 0.44 1/2 (50.0) Abbreviations: max, maximum; min, minimum. a Bicornurate uterus (n = 1), renal agenesis (n = 1), bilateral cryptorchidism (n = 2), nephrocalcinosis (n = 1). b Multiple agenesis (n = 3), crowded teeth (n = 3), delayed tooth eruption (n = 2). c Aneurysm of the interatrial septum (n = 2), interatrial communication (n = 1). View Large Table 1. Main Clinical Features of the Cohort All Patients Epimutations (1) Upd(14)mat (2) (1) vs (2) Paternal Deletion n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) P n (%) Mean (Min–Max) <−2 SDS (%) Birth  Term (wk of amenorrhea) 28 37.5 (28.0–40.0) 17 37.2 (28.0–40.0) 8 37.6 (36.0–39.0) 1 3 38.7 (37.0–40.0)  Birth length (SDS) 28 −2.3 (−5.2 to 0.8) 16 (57.1) 17 −2.3 (−5.2 to −0.9) 9 (52.9) 8 −2.4 (−3.7 to 0.8) 6 (75.0) 0.37 3 −1.6 (−2.2 to −1.2) 1 (33.3)  Birth weight (SDS) 28 −2.4 (−4.0 to −0.7) 24 (85.7) 17 −2.4 (−4.0 to −0.9) 14 (82.4) 8 −2.8 (−3.3 to −2.3) 8 (100) 0.22 3 −2.0 (−3.1 to −0.7) 2 (66.7)  Intrauterine growth retardation 25/28 (89.3) 15/17 (88.2) 8/8 (100) 1 2/3 (66.7)  Birth head circumference (SDS) 25 −1.3 (−3.7 to 1.0) 15 −1.3 (−3.7 to 1.0) 7 −1.5 (−2.0 to −0.7) 0.46 3 −0.7 (−0.6 to −0.7)  Relative macrocephaly 14/25 (56.0) 7/15 (46.7) 5/7 (71.4) 0.73 2/3 (66.7) Growth  Height at 2 y (SDS) 25 −2.0 (−4.4 to −0.2) 14/25 (56.0) 15 −2.0 (−4.4 to −0.2) 9 (60.0) 8 −1.8 (−4.1 to −0.6) 3 (37.5) 0.56 2 −2.5 (−2.6 to −2.4) 2/2 (100)  BMI at 2 y (SDS) 24 −1.5 (−3.2 to 0.0) 8 (33.3) 15 −1.4 (−3.2 to 0.0) 5 (33.3) 7 −1.8 (−2.9 to −0.6) 3 (42.8) 0.27 2 −1.7 (−1.9 to −1.4) 0  Early feeding difficulties 17/28 (60.7) 9/17 (52.9) 6/8 (75.0) 0.71 2/3 (66.7) NH-CSS  NH-CSS ≥4 16/22 (72.7) 9/14 (64.3) 5/5 (100.0) 0.71 2/3 (66.7) Clinical signs  Protruding forehead 25/28 (89.3) 17/17 (100) 6/8 (75.0) 0.76 2/3 (66.7)  Body asymmetry 7/26 (26.9) 6/17 (35.3) 1/6 (16.7) 1 0/3 (0)  Acromicria 19/26 (73.1) 12/17 (70.6) 6/7 (85.7) 1 1/2 (50.0)  Downturned mouth 18/24 (75.0) 12/15 (80.0) 5/7 (71.4) 1 1/2 (50.0)  Low-set posteriorly rotated ears 14/25 (56.0) 11/16 (68.8) 2/7 (28.6) 0.44 1/2 (50.0)  Clinodactyly 16/26 (61.5) 10/16 (62.5) 5/8 (62.5) 1 1/2 (50.0) Development  Neonatal hypotonia 14/20 (70.0) 5/10 (50.0) 6/7 (85.7) 0.70 3/3 (100)  Motor delay 16/25 (64.0) 8/15 (53.3) 5/7 (71.4) 0.72 3/3 (100)  Speech delay 12/23 (52.2) 7/15 (46.7) 2/5 (40.0) 1 3/3 (100)  Normal schooling 13/19 (68.4) 10/12 (83.3) 3/5 (60.0) 1 0/2 (0.0)  Behavioral disorders 3/23 (13.0) 2/15 (13.3) 1/5 (20.0) 0.5 0/3 (0.0) Associated malformations  Urogenitala 5/23 (21.7) 3/12 (25.0) 1/8 (12.5) 1 1/3 (33.3)  Stomatologyb 7/21 (33.3) 6/13 (46.2) 1/5 (20.0%) 0.64 0/3 (0.0)  Heartc 2/17 (11.8) 0/9 (0.0) 1/6 (16.7%) 0.44 1/2 (50.0) All Patients Epimutations (1) Upd(14)mat (2) (1) vs (2) Paternal Deletion n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) n (%) Mean (Min–Max) <−2 SDS (%) P n (%) Mean (Min–Max) <−2 SDS (%) Birth  Term (wk of amenorrhea) 28 37.5 (28.0–40.0) 17 37.2 (28.0–40.0) 8 37.6 (36.0–39.0) 1 3 38.7 (37.0–40.0)  Birth length (SDS) 28 −2.3 (−5.2 to 0.8) 16 (57.1) 17 −2.3 (−5.2 to −0.9) 9 (52.9) 8 −2.4 (−3.7 to 0.8) 6 (75.0) 0.37 3 −1.6 (−2.2 to −1.2) 1 (33.3)  Birth weight (SDS) 28 −2.4 (−4.0 to −0.7) 24 (85.7) 17 −2.4 (−4.0 to −0.9) 14 (82.4) 8 −2.8 (−3.3 to −2.3) 8 (100) 0.22 3 −2.0 (−3.1 to −0.7) 2 (66.7)  Intrauterine growth retardation 25/28 (89.3) 15/17 (88.2) 8/8 (100) 1 2/3 (66.7)  Birth head circumference (SDS) 25 −1.3 (−3.7 to 1.0) 15 −1.3 (−3.7 to 1.0) 7 −1.5 (−2.0 to −0.7) 0.46 3 −0.7 (−0.6 to −0.7)  Relative macrocephaly 14/25 (56.0) 7/15 (46.7) 5/7 (71.4) 0.73 2/3 (66.7) Growth  Height at 2 y (SDS) 25 −2.0 (−4.4 to −0.2) 14/25 (56.0) 15 −2.0 (−4.4 to −0.2) 9 (60.0) 8 −1.8 (−4.1 to −0.6) 3 (37.5) 0.56 2 −2.5 (−2.6 to −2.4) 2/2 (100)  BMI at 2 y (SDS) 24 −1.5 (−3.2 to 0.0) 8 (33.3) 15 −1.4 (−3.2 to 0.0) 5 (33.3) 7 −1.8 (−2.9 to −0.6) 3 (42.8) 0.27 2 −1.7 (−1.9 to −1.4) 0  Early feeding difficulties 17/28 (60.7) 9/17 (52.9) 6/8 (75.0) 0.71 2/3 (66.7) NH-CSS  NH-CSS ≥4 16/22 (72.7) 9/14 (64.3) 5/5 (100.0) 0.71 2/3 (66.7) Clinical signs  Protruding forehead 25/28 (89.3) 17/17 (100) 6/8 (75.0) 0.76 2/3 (66.7)  Body asymmetry 7/26 (26.9) 6/17 (35.3) 1/6 (16.7) 1 0/3 (0)  Acromicria 19/26 (73.1) 12/17 (70.6) 6/7 (85.7) 1 1/2 (50.0)  Downturned mouth 18/24 (75.0) 12/15 (80.0) 5/7 (71.4) 1 1/2 (50.0)  Low-set posteriorly rotated ears 14/25 (56.0) 11/16 (68.8) 2/7 (28.6) 0.44 1/2 (50.0)  Clinodactyly 16/26 (61.5) 10/16 (62.5) 5/8 (62.5) 1 1/2 (50.0) Development  Neonatal hypotonia 14/20 (70.0) 5/10 (50.0) 6/7 (85.7) 0.70 3/3 (100)  Motor delay 16/25 (64.0) 8/15 (53.3) 5/7 (71.4) 0.72 3/3 (100)  Speech delay 12/23 (52.2) 7/15 (46.7) 2/5 (40.0) 1 3/3 (100)  Normal schooling 13/19 (68.4) 10/12 (83.3) 3/5 (60.0) 1 0/2 (0.0)  Behavioral disorders 3/23 (13.0) 2/15 (13.3) 1/5 (20.0) 0.5 0/3 (0.0) Associated malformations  Urogenitala 5/23 (21.7) 3/12 (25.0) 1/8 (12.5) 1 1/3 (33.3)  Stomatologyb 7/21 (33.3) 6/13 (46.2) 1/5 (20.0%) 0.64 0/3 (0.0)  Heartc 2/17 (11.8) 0/9 (0.0) 1/6 (16.7%) 0.44 1/2 (50.0) Abbreviations: max, maximum; min, minimum. a Bicornurate uterus (n = 1), renal agenesis (n = 1), bilateral cryptorchidism (n = 2), nephrocalcinosis (n = 1). b Multiple agenesis (n = 3), crowded teeth (n = 3), delayed tooth eruption (n = 2). c Aneurysm of the interatrial septum (n = 2), interatrial communication (n = 1). View Large Eight of the 23 patients, for whom data were available, were treated with recombinant growth hormone (GH) therapy from a mean age of 4.7 (1.1 to 11.3) years according to the SGA indication and posology (23). NH-CSS Among patients for whom all items of the NH-CSS were available, 72.7% (16/22) had a score ≥4/6, and consistent with a clinical diagnosis of SRS. One item was missing in six patients, of whom two had an NH-CSS score of 4/5 (compatible with a diagnosis of SRS), one had a score of 3/5 and three had scores of 2/5, which does not fulfill the criteria for a clinical diagnosis of SRS (Supplemental Fig. 2). Among the six patients who did not fulfill the NH-CSS criteria, five (83.37%) had an epimutation and one a deletion (case 26). Puberty and pubarche We collected data on puberty and pubarche for all patients but one (one girl for whom data were not available). At the end of the study, 11 patients had gone into puberty, eight girls and three boys; the oldest among the other nonpubertal patients was an 8.7-year-old girl. Of these 11 patients, six (54.5%) had precocious puberty, including five (62.5%) girls and one (33.3%) boy. Four (66.7%) had epimutations and two (33.3%) had upd(14)mat. Puberty occurred early for the other three girls, before age 9 years, and was rapidly progressive, with menarche <1 year after breast development for two of them. Puberty also started early for the other two boys, at 10.0 and 10.2 years (Table 2). Six (54.5%) patients were treated with gonadotrophin-releasing hormone analogs (aGnRHs) to suppress puberty at a mean age of 7.9 (5.0 to 10.3) years. Six (54.5%) patients had exaggerated adrenarche; four among them were treated with cyproterone acetate. Table 2. Characteristics of Pubarche and Puberty for 11 Patients >9 Years of Age Total Girls (n = 8) Boys (n = 3) Pubarche  Premature pubarche 3/8 (37.5%) 2/6 (33.3%) 1/2 (50.0%)  Exaggerated adrenarche 6/10 (60.0%) 4/7 (57.1%) 2/3 (66.7%)  Age at pubarche onset, y 8.6 (6.3–12.0) 9.0 (8.0–10.0)  Bone age advancement during puberty 9/9 (100.0%) 8/8 (100.0%) 1/1 (100.0%) Central puberty  Age at thelarche or testicle enlargement, y — 7.2 (4.0–8.5) 9.6 (8.6–10.2)  Age at menarche, ya — 10.2 (8.9–12.0) —  Delay between thelarche and menarche, ya — 1.9 (0.5–3.5) — BMI  At pubarche onset (SDS) 2.1 (0.2–6.1) 2.1 (0.2-6.1) 2.3 (1.0–3.6)  At central puberty onset (SDS) 1.9 (−0.9 to 6.1) 1.8 (−0.9 to 6.1) 2.4 (1.1–3.6) Total Girls (n = 8) Boys (n = 3) Pubarche  Premature pubarche 3/8 (37.5%) 2/6 (33.3%) 1/2 (50.0%)  Exaggerated adrenarche 6/10 (60.0%) 4/7 (57.1%) 2/3 (66.7%)  Age at pubarche onset, y 8.6 (6.3–12.0) 9.0 (8.0–10.0)  Bone age advancement during puberty 9/9 (100.0%) 8/8 (100.0%) 1/1 (100.0%) Central puberty  Age at thelarche or testicle enlargement, y — 7.2 (4.0–8.5) 9.6 (8.6–10.2)  Age at menarche, ya — 10.2 (8.9–12.0) —  Delay between thelarche and menarche, ya — 1.9 (0.5–3.5) — BMI  At pubarche onset (SDS) 2.1 (0.2–6.1) 2.1 (0.2-6.1) 2.3 (1.0–3.6)  At central puberty onset (SDS) 1.9 (−0.9 to 6.1) 1.8 (−0.9 to 6.1) 2.4 (1.1–3.6) a Without GnRH analogs. View Large Table 2. Characteristics of Pubarche and Puberty for 11 Patients >9 Years of Age Total Girls (n = 8) Boys (n = 3) Pubarche  Premature pubarche 3/8 (37.5%) 2/6 (33.3%) 1/2 (50.0%)  Exaggerated adrenarche 6/10 (60.0%) 4/7 (57.1%) 2/3 (66.7%)  Age at pubarche onset, y 8.6 (6.3–12.0) 9.0 (8.0–10.0)  Bone age advancement during puberty 9/9 (100.0%) 8/8 (100.0%) 1/1 (100.0%) Central puberty  Age at thelarche or testicle enlargement, y — 7.2 (4.0–8.5) 9.6 (8.6–10.2)  Age at menarche, ya — 10.2 (8.9–12.0) —  Delay between thelarche and menarche, ya — 1.9 (0.5–3.5) — BMI  At pubarche onset (SDS) 2.1 (0.2–6.1) 2.1 (0.2-6.1) 2.3 (1.0–3.6)  At central puberty onset (SDS) 1.9 (−0.9 to 6.1) 1.8 (−0.9 to 6.1) 2.4 (1.1–3.6) Total Girls (n = 8) Boys (n = 3) Pubarche  Premature pubarche 3/8 (37.5%) 2/6 (33.3%) 1/2 (50.0%)  Exaggerated adrenarche 6/10 (60.0%) 4/7 (57.1%) 2/3 (66.7%)  Age at pubarche onset, y 8.6 (6.3–12.0) 9.0 (8.0–10.0)  Bone age advancement during puberty 9/9 (100.0%) 8/8 (100.0%) 1/1 (100.0%) Central puberty  Age at thelarche or testicle enlargement, y — 7.2 (4.0–8.5) 9.6 (8.6–10.2)  Age at menarche, ya — 10.2 (8.9–12.0) —  Delay between thelarche and menarche, ya — 1.9 (0.5–3.5) — BMI  At pubarche onset (SDS) 2.1 (0.2–6.1) 2.1 (0.2-6.1) 2.3 (1.0–3.6)  At central puberty onset (SDS) 1.9 (−0.9 to 6.1) 1.8 (−0.9 to 6.1) 2.4 (1.1–3.6) a Without GnRH analogs. View Large Epiphyseal fusion occurred early in four patients, at 12.7 (11.2 to 13.8) years in girls (n = 3) and 13.8 years in one boy, without aGnRH treatment. The mean final height of the girls (n = 4) was 143.5 (141.0 to 145.0) cm, corresponding to −3.6 (−4.0 to −3.3) SDS, according to Sempé (24), with a mean pubertal growth spurt of 12.8 (10.0 to 17.0) cm without aGnRH and 25.5 cm for the girl who was treated. One boy had a final height of 150.0 cm (−3.9 SDS), with a pubertal growth spurt of 19.3 cm; the second had a final height of 169 cm (−0.8 SDS), far from his target height (+2.7 SDS). None of these patients received recombinant GH treatment. All clinical data concerning puberty and pubarche of these 11 patients are summarized in Supplemental Table 1. Metabolic outcomes The age of adiposity rebound was precocious for 93.8% (15/16 for whom data were available) of the patients, with a mean at 2.1 (1.0 to 6.5) years. Twelve patients (75.0%) had adiposity rebound by the age of 2 years. Among patients with precocious adiposity rebound, only one needed enteral feeding. For this patient, nutrition intake is on the decrease but she experienced complete anorexia. For all other patients, BMI had spontaneously grown precociously. The BMI of all patients for whom puberty had started increased markedly (>1 SDS) before the onset of pubarche and/or central puberty. Extensive molecular analysis Methylation analysis of 18 imprinted loci We studied the methylation levels at 18 imprinted loci, using TaqMan allele-specific methylated multiplex real-time quantitative polymerase chain