Androgens, Irregular Menses, and Risk of Diabetes and Coronary Artery Calcification in the Diabetes Prevention Program

Androgens, Irregular Menses, and Risk of Diabetes and Coronary Artery Calcification in the... Abstract Context It is unclear whether relative elevations in androgens or irregular menses (IM) are associated with greater cardiometabolic risk among women who are already overweight and glucose intolerant. Research Design and Methods We conducted a secondary analysis of the Diabetes Prevention Program (DPP) and the Diabetes Prevention Program Outcomes Study (DPPOS). Participants included women with sex hormone measurements who did not use exogenous estrogen (n = 1422). We examined whether free androgen index (FAI) or IM was associated with diabetes risk during the DPP/DPPOS or with coronary artery calcification (CAC) at DPPOS year 10. Models were adjusted for menopausal status, age, race or ethnicity, randomization arm, body mass index (BMI), and hemoglobin A1c. Results Women had an average age of 48.2 ± 9.9 years. Elevations in FAI and IM were associated with greater BMI, waist circumference, and blood pressure and lower adiponectin. FAI was not associated with diabetes risk during the DPP/DPPOS [hazard ratio (HR) 0.97; 95% confidence interval (CI), 0.93 to 1.02] or increased odds of CAC [odds ratio (OR) 1.06; 95% CI, 0.92 to 1.23]. IM was also not associated with diabetes risk during the DPP/DPPOS (HR 1.07; 95% CI, 0.87 to 1.31) or increased odds of CAC (OR 0.89; 95% CI, 0.53 to 1.49). Women who had both relative elevations in FAI and IM had similar diabetes risk and odds of CAC as women without these conditions. Differences by treatment arm and menopausal status were not observed. Conclusions Among midlife women who were already glucose intolerant and overweight, androgen concentrations and IM did not additionally contribute to increased risk for diabetes or CAC. Polycystic ovary syndrome (PCOS) is a common endocrinopathy that may affect up to one in five women of reproductive age (1). Existing definitions are based on the presence of two or more of the following criteria: relative elevations in androgens, irregular menses (IM), and polycystic ovary morphology (1). Women with PCOS have greater insulin resistance compared with age- and weight-matched controls (2–4) and subsequently higher risk of type 2 diabetes (5–9). However, the extent to which central PCOS components (i.e., androgens and IM) increase cardiometabolic risk among women without a diagnosis of PCOS is unclear. Specifically, it is unknown whether these characteristics increase risk beyond the presence of impaired glucose tolerance and overweight alone. Although one meta-analysis noted that women with elevated total testosterone concentrations had higher risk of diabetes in cross-sectional analyses (10), more recent longitudinal studies of women with a range of body mass indices (BMIs) and degrees of glucose tolerance have not observed associations between hyperandrogenism and diabetes (11, 12). Reports also conflict as to whether hyperandrogenism is associated with coronary artery calcification (CAC) (13, 14) or carotid intima media thickness (15–18), both of which are indicators of subclinical atherosclerosis. Similarly, some studies report an association of IM with incident diabetes (19, 20) and incident coronary disease (21) in populations with a range of BMIs and degrees of glucose tolerance, but others have noted that these associations are no longer significant after adjustment for BMI (13, 22, 23). The Diabetes Prevention Program (DPP) randomly assigned overweight, nondiabetic glucose-intolerant participants to a program of intensive lifestyle modification (ILS), metformin, or placebo (24). The follow-up observational cohort, the Diabetes Prevention Program Outcomes Study (DPPOS), has continued to ascertain incident diabetes semiannually and has also measured CAC. Thus, we were able to assess whether androgens or history of IM at study entry were associated with increased risk factors for cardiometabolic disease beyond their known associations with impaired glucose tolerance and increased weight. For the present report, we examined traditional and nontraditional cardiovascular risk factors, progression to diabetes, and presence of CAC. We hypothesized that women with relative androgen elevations or IM would have adverse risk factor profiles compared with other overweight and glucose-intolerant women without these conditions. We also hypothesized that the free androgen index (FAI), a measure of androgen levels, and IM would be associated with increased risk of incident diabetes and a higher prevalence of detectable CAC. Research Design and Methods Population and setting The design, methods, and baseline characteristics of the DPP have been previously described (24). Between 1996 and 1999, participants were recruited across 27 clinical centers located throughout the United States. Inclusion criteria were age ≥25 years, BMI ≥24 kg/m2 (≥22 kg/m2 for Asian Americans), a fasting plasma glucose level of 5.3 to 7.0 mmol/L (95 to 125 mg/dL), and a 2-hour plasma glucose level of 7.8 to 11.1 mmol/L (140 to 199 mg/dL) after an oral 75-g glucose load. Eligible participants were randomly assigned to one of three interventions: 850 mg metformin twice daily, placebo twice daily, or ILS. The goals of ILS were to achieve and maintain a weight reduction of ≥7% through consumption of a low-calorie, low-fat diet, plus moderate physical activity for ≥150 minutes per week. Each participating institution was overseen by its respective ethics review board, and written informed consent was obtained from all participants. IM, pregnancy, and menopause Standardized interviewer-administered questionnaires were used to obtain other demographic and clinical data. At the time of DPP randomization, a screening questionnaire inquired about women’s menses. Women were asked, “During most of your life, were your periods regular? That is, did they occur once a month?” (Do not include any time when you were pregnant or taking birth control pills.) Women who answered “no” or “sometimes regular, sometimes irregular” were classified as having IM. Questionnaires assessed previous use of estrogen, consisting of birth control pills or estrogen therapy, as well as menopausal status. Women were classified as postmenopausal if they reported absence of menses for a year, bilateral oophorectomy, or hysterectomy and age ≥55 years. Finally, current estrogen use was ascertained by study staff examination of pill bottles. Sex hormone measurements Ancillary investigations were conducted to ascertain associations between sex hormones and incident diabetes. Eligibility criteria included consent to participate in ancillary studies and plasma availability. Of the 2191 women originally randomly assigned, 2002 women underwent sex hormone measures on prerandomization DPP samples. For the purposes of this report, we excluded women who used estrogen therapy or oral contraceptive pills at baseline, resulting in 1422 participants. Women who were excluded were more likely to have been randomly assigned to the placebo arm and were older, more likely to be postmenopausal, white, thinner, and had lower testosterone, higher sex hormone binding globulin (SHBG), and lower FAI compared with women who were included (Supplemental Table 1). Of the 1422 women included in this report, 143 women did not have testosterone or SHBG measures; these women were included in analyses of IM but not of FAI. Blood samples for sex hormone measurements were collected fasting before 10 am. Sex hormones were measured by Endoceutics (Quebec City, Canada) using liquid chromatography–mass spectrometry. SHBG was measured with an enzyme-linked immunosorbent assay (Bioline, Taunton, MA) with interassay coefficients of variation (CVs) of 7.8% and 5.0% at 18.2 and 63.1 nmol/L, respectively. The lower limits of detection were 20 ng/dL for dehydroepiandrosterone sulfate (DHEAS) and 2, 10, 0.2, and 10 pg/mL for dihydrotestosterone, testosterone, estradiol, and estrone sulfate, respectively. The lower limits of quantification were 100 ng/mL for DHEAS and 10, 50, 1, and 50 pg/mL for dihydrotestosterone, testosterone, estradiol, and estrone sulfate, respectively. Interassay variation (CV) was 10.2%, 10.7%, 7.0%, and 12.5% for DHEAS, testosterone, estradiol, and estrone, respectively, at the lower limits of quantification level. FAI was calculated as 100 × (total testosterone/SHBG). Traditional and nontraditional cardiovascular risk factors Blood pressure and anthropometrics were measured with standardized techniques. Aside from sex hormone measures, other assays were performed at a central biochemistry laboratory (Northwest Lipid Research Laboratories, University of Washington, Seattle). For the purposes of this analysis, we used measures from DPP baseline. Plasma glucose was measured on a chemistry autoanalyzer by the glucokinase method. Insulin measurements were performed by a polyethylene glycol–accelerated double antibody radioimmunoassay method. Insulin sensitivity was assessed by the surrogate measurement 1/fasting insulin (1/FI). Lipoprotein fractions were separated with preparative ultracentrifugation by β-quantification for triglycerides of ≥4.52 mmol/L. For the high-density lipoprotein (HDL) measurement, apolipoprotein B–containing lipoproteins were removed from plasma by precipitation with dextran sulfate, and cholesterol in HDL particles was measured with Roche reagent on a Roche Modular P autoanalyzer (Roche Diagnostics, Indianapolis, IN). Plasma triglycerides were determined enzymatically with methods standardized to the Centers for Disease Control and Prevention Reference Methods. Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald equation. A nonequilibrium density gradient ultracentrifugation method was used to characterize LDL floatation rate. Total circulating adiponectin was measured with a latex particle-enhanced turbidimetric assay (Otsuka Pharmaceutical, Tokyo, Japan). The within-run and between-run CVs for this assay are 6.21% and 9.25%, respectively. Plasma high-sensitivity C-reactive protein (CRP) and fibrinogen were measured immunochemically with Dade-Behring reagent on the Behring Nephelometer autoanalyzer (Dade-Behring, Glasgow, DE), and the two CVs were 2.10% and 3.10% for CRP and 2.70% and 2.60% for fibrinogen, respectively. Tissue plasminogen activator (tPA) levels were measured in citrated plasma with an enzyme-linked immunosorbent assay (Asserachrom tPA; Diagnostica, Stago, France), which measures total tPA antigen, with CVs of 6.45% and 6.70% respectively. DPPOS incident diabetes At the conclusion of the DPP, participants had been followed for an average of 3.2 years. The placebo and metformin groups were unmasked as to their treatment assignment, and all participants were offered the ILS in a group format during a 1-year bridge period (25). The surviving consent members (n = 3149) of the three original treatment arms were invited to participate in an observational follow-up, the DPPOS. Eighty-eight percent (n = 2776) joined. Maintenance group lifestyle sessions were offered quarterly to all DPPOS participants, and metformin continued to be provided to participants originally randomly assigned to metformin who remained eligible. As in the DPP, the development of diabetes was determined with a 75-g oral glucose tolerance test performed annually and fasting plasma glucose tests every 6 months. For diagnosis of diabetes, fasting plasma glucose ≥7.0 mmol/L or 2-hour levels ≥11.1 mmol/L had to be confirmed by a repeat test within 6 weeks; to keep diagnostic criteria similar between the trial and the follow-up, hemoglobin A1c was not used to diagnose diabetes. The DPPOS is ongoing, and participants in this analysis had been followed for ~10 years. DPPOS CAC CAC was measured during year 10 of DPPOS in 2029 participants (74% of the DPPOS cohort) at all sites (26). Participants who weighed >350 lb at the time of scanning were ineligible to participate because of the inability to acquire the relevant images on conventional equipment; 201 refused consent, 57 had a stent, 32 had atrial fibrillation, and 1 was pregnant. At the DPP baseline examination, participants who later underwent CAC studies had slightly higher minority race or ethnicity representation, were slightly younger, and had modestly lower BMI and lower systolic blood pressure, and slightly fewer were smokers compared with those who did not undergo the CAC assessment. Baseline metabolic measures were not different between tested and untested groups, except for slightly higher HDL levels in tested participants. The proportion with CAC measurements did not differ between treatment groups. Sixty-five percent of the 813 women with CAC measures (n = 532) underwent sex hormone measures and were not using exogenous estrogen and thus were included for the subanalysis of CAC. Methods for CAC assessment have been previously reported (26). Briefly, chest computed tomography was performed by certified technologists at each site. Subjects were scanned twice, and measurement of CAC was calibrated against a phantom of known physical calcium concentration. A radiologist or cardiologist read all computed tomography scans at the central reading center (Los Angeles Biomedical Research Institute at Harbor–University of California, Los Angles in Torrance, California) in a manner blinded to patient characteristics and treatment assignment. Discrepancies were reviewed and agreement obtained through consensus. For each scan, a total phantom-adjusted averaged Agatston score was calculated, defined as the sum of calcium measures from the left main, left anterior descending, circumflex, and right coronary arteries. For the purposes of the current report where the burden of CAC was low, we examined the presence and absence of CAC. Statistical analyses Baseline characteristics were described as n (percentages) for categorical variables and mean [standard deviation (SD)] or median (interquartile range) for quantitative variables with normal and skewed distributions, respectively. Differences between women in the upper 10th percentile and lower 90th percentile of FAI (Table 1) and with and without histories of IM (Table 2) were tested via the χ2 test for independence for categorical variables and the t test or the nonparametric Wilcoxon test, as appropriate, for continuous variables. There is no agreement on what FAI cutoff point is consistent with hyperandrogenism, because of variability in testosterone levels and poor standardization of assays. Consequently, the upper quartile or upper 5th or 10th percentile of FAI is commonly and arbitrarily used to define androgen elevation within a particular study population. We examined whether another FAI cutoff point existed by using k-means clustering (27) and also whether there were cutoff points in FAI at which the hazard of diabetes changed markedly by using classification and regression trees (28); neither method revealed another cutoff point, so the 10th percentile was used. Analyses were stratified by menopausal status (Tables 1 and 2 ) or adjusted for menopausal status (Tables 3 and 4 ). To examine the relationship between the variables of FAI and IM with incident diabetes during the DPP and the DPPOS, we constructed Cox proportional hazards models that adjusted for baseline age, race or ethnicity, randomization arm, BMI, and A1c (Table 3). Analyses of incident diabetes throughout the DPPOS included incident diabetes in the DPP. Additional evaluations for interactions with randomization arm when randomization arm was a covariate were performed by adding interaction terms between randomization arm with IM and randomization arm with FAI; interaction terms were not significant, and thus models combined data across randomization arms. Similarly, evaluations for interactions with menopause in models examining diabetes risk were performed by adding interaction terms between FAI and menopause and IM and menopause; neither of these was significant, and thus models combined data across menopausal status. To examine the relationship between FAI and IM with the odds of any CAC at year 10 of the DPPOS, we constructed logistic regression models that adjusted for baseline age, race or ethnicity, randomization arm, BMI, and A1c (Table 4). Table 1. Baseline Participant Characteristics by Menopausal Status and FAI   Premenopausal Women   Postmenopausal Women     FAI, Lower 90th Percentile (n = 761)  FAI, Upper 10th Percentile (n = 85)  P  FAI, Lower 90th Percentile (n = 389)  FAI, Upper 10th Percentile (n = 44)  P  Treatment group, n (%)      0.087      0.16   Placebo  282 (37.1%)  30 (35.3%)    151 (38.8%)  14 (31.8%)     Metformin  248 (32.6%)  20 (23.5%)    106 (27.2%)  18 (40.9%)     ILS  231 (30.4%)  35 (41.2%)    132 (33.9%)  12 (27.3%)    Age group, n (%)      <0.001      0.020   25 to 44 y  405 (53.2%)  73 (85.9%)    33 (8.5%)  9 (20.5%)     45 to 59 y  356 (46.8%)  12 (14.1%)    231 (59.4%)  19 (43.2%)     ≥60 y        125 (32.1%)  16 (36.4%)    Race or ethnicity, n (%)      0.079      0.29   Non-Hispanic white  364 (47.8%)  50 (58.8%)    199 (51.2%)  22 (50%)     African American  178 (23.4%)  22 (25.9%)    125 (32.1%)  14 (31.8%)     Hispanic  157 (20.6%)  7 (8.2%)    52 (13.4%)  4 (9.1%)     Asian  25 (3.3%)  2 (2.4%)    7 (1.8%)  3 (6.8%)     American Indian  37 (4.9%)  4 (4.7%)    6 (1.5%)  1 (2.3%)    Number of pregnancies  3 [2, 4]  2 [2, 3]  0.082  3 [2, 4]  3 [2, 4]  0.92   No pregnancies  113 (14.8%)  22 (25.9%)  0.013  46 (11.8%)  7 (15.9%)  0.59  IM, n (%)  126 (16.6%)  29 (34.1%)  <0.001  64 (16.5%)  11 (25%)  0.23  Past estrogen use (birth control pills or estrogen therapy)  582 (76.5%)  65 (76.5%)  1.000  290 (74.6%)  26 (59.1%)  0.044  BMI, kg/m2  35.9 ± 7.1  37.8 ± 6.7  0.021  35.1 ± 6.7  36.8 ± 7.7  0.11  Waist circumference, cm  104.9 ± 15.6  108.7 ± 15.2  0.032  104.2 ± 14.2  108.8 ± 14.8  0.045  Fasting glucose, mg/dL  106.5 ± 8.3  107.1 ± 8.2  0.53  106.9 ± 7.9  105.7 ± 6.4  0.33  2-h glucose, mg/dL  163.8 ± 17  167.8 ± 16.4  0.038  163.5 ± 16.7  163.4 ± 17.4  0.98  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.6  0.94  6 ± 0.5  5.9 ± 0.4  0.23  DHEAS, μmol/L  2.19 [1.44, 3.22]  2.95 [1.9, 4.21]  <0.001  1.71 [1.07, 2.56]  1.62 [1.16, 2.82]  0.88  Testosterone, nmol/L  0.58 [0.45, 0.79]  1.05 [0.89, 1.39]  <0.001  0.52 [0.41, 0.66]  1.3 [0.82, 1.94]  <0.001  FAI, mean ± SD  1.5 ± 0.7  4.8 ± 1.8  <0.001  1.6 ± 0.8  9.2 ± 14.1  <0.001  Estradiol, pmol/L  210.6 [91.6, 384]  305 [139.4, 443.5]  0.006  37.3 [24.6, 74.7]  46.3 [31.8, 66.8]  0.18  Estrone sulfate, pmol/L  3281.7 [1710.2, 6223.1]  4490.1 [3193, 7224.6]  <0.001  1225.3 [619.7, 2226.3]  1621.6 [1078.5, 2780.3]  0.031  SHBG, (nmol/L)  43 [31.7, 62.5]  25.3 [19.3, 31.3]  <0.001  36.5 [28.4, 48.2]  24.5 [15, 34.2]  <0.001  1/FI  0.047 ± 0.033  0.038 ± 0.024  0.015  0.052 ± 0.035  0.047 ± 0.039  0.31  SBP, mm Hg  120.4 ± 14.2  117.4 ± 11.8  0.061  126.8 ± 16  123.4 ± 14.1  0.19  DBP, mm Hg  77.7 ± 9.3  79.9 ± 8.9  0.034  78.2 ± 9  77.2 ± 10.1  0.50  HDL, mg/dL  44.8 ± 9.8  41.6 ± 9.3  0.004  48.2 ± 11.5  46.2 ± 10.7  0.26  LDL, mg/dL  121.6 ± 31.7  121.3 ± 26.7  0.93  133.2 ± 35.4  132.2 ± 28.4  0.86  Triglycerides, mg/dL  148.1 ± 85.4  159.5 ± 88.5  0.24  149.8 ± 86.2  152.5 ± 72.8  0.84  LDL particle density, 1/rf  3.8 [3.46, 4]  3.8 [3.46, 4.22]  0.32  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.84  Adiponectin, ng/mL  6.9 [5.3, 8.7]  6 [4.9, 7.3]  <0.001  8.4 [6.5, 11.5]  8.2 [6.3, 9.8]  0.41  High-sensitivity CRP, mg/L  0.7 ± 0.82  0.62 ± 0.55  0.39  0.65 ± 0.74  0.68 ± 0.57  0.78  Fibrinogen, μmol/L  11.2 [9.8, 13.1]  11.4 [10, 12.9]  0.85  12.1 [10.7, 13.8]  12.4 [11.2, 13.6]  0.66  tPA, ng/mL  10.9 [8.8, 13.2]  10.9 [9, 13.1]  0.56  11.1 [9.5, 13.2]  11.3 [9.7, 13.8]  0.49  Any CAC  126 (39.6%)  12 (44.4%)  0.78  72 (59.5%)  8 (66.7%)  0.86    Premenopausal Women   Postmenopausal Women     FAI, Lower 90th Percentile (n = 761)  FAI, Upper 10th Percentile (n = 85)  P  FAI, Lower 90th Percentile (n = 389)  FAI, Upper 10th Percentile (n = 44)  P  Treatment group, n (%)      0.087      0.16   Placebo  282 (37.1%)  30 (35.3%)    151 (38.8%)  14 (31.8%)     Metformin  248 (32.6%)  20 (23.5%)    106 (27.2%)  18 (40.9%)     ILS  231 (30.4%)  35 (41.2%)    132 (33.9%)  12 (27.3%)    Age group, n (%)      <0.001      0.020   25 to 44 y  405 (53.2%)  73 (85.9%)    33 (8.5%)  9 (20.5%)     45 to 59 y  356 (46.8%)  12 (14.1%)    231 (59.4%)  19 (43.2%)     ≥60 y        125 (32.1%)  16 (36.4%)    Race or ethnicity, n (%)      0.079      0.29   Non-Hispanic white  364 (47.8%)  50 (58.8%)    199 (51.2%)  22 (50%)     African American  178 (23.4%)  22 (25.9%)    125 (32.1%)  14 (31.8%)     Hispanic  157 (20.6%)  7 (8.2%)    52 (13.4%)  4 (9.1%)     Asian  25 (3.3%)  2 (2.4%)    7 (1.8%)  3 (6.8%)     American Indian  37 (4.9%)  4 (4.7%)    6 (1.5%)  1 (2.3%)    Number of pregnancies  3 [2, 4]  2 [2, 3]  0.082  3 [2, 4]  3 [2, 4]  0.92   No pregnancies  113 (14.8%)  22 (25.9%)  0.013  46 (11.8%)  7 (15.9%)  0.59  IM, n (%)  126 (16.6%)  29 (34.1%)  <0.001  64 (16.5%)  11 (25%)  0.23  Past estrogen use (birth control pills or estrogen therapy)  582 (76.5%)  65 (76.5%)  1.000  290 (74.6%)  26 (59.1%)  0.044  BMI, kg/m2  35.9 ± 7.1  37.8 ± 6.7  0.021  35.1 ± 6.7  36.8 ± 7.7  0.11  Waist circumference, cm  104.9 ± 15.6  108.7 ± 15.2  0.032  104.2 ± 14.2  108.8 ± 14.8  0.045  Fasting glucose, mg/dL  106.5 ± 8.3  107.1 ± 8.2  0.53  106.9 ± 7.9  105.7 ± 6.4  0.33  2-h glucose, mg/dL  163.8 ± 17  167.8 ± 16.4  0.038  163.5 ± 16.7  163.4 ± 17.4  0.98  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.6  0.94  6 ± 0.5  5.9 ± 0.4  0.23  DHEAS, μmol/L  2.19 [1.44, 3.22]  2.95 [1.9, 4.21]  <0.001  1.71 [1.07, 2.56]  1.62 [1.16, 2.82]  0.88  Testosterone, nmol/L  0.58 [0.45, 0.79]  1.05 [0.89, 1.39]  <0.001  0.52 [0.41, 0.66]  1.3 [0.82, 1.94]  <0.001  FAI, mean ± SD  1.5 ± 0.7  4.8 ± 1.8  <0.001  1.6 ± 0.8  9.2 ± 14.1  <0.001  Estradiol, pmol/L  210.6 [91.6, 384]  305 [139.4, 443.5]  0.006  37.3 [24.6, 74.7]  46.3 [31.8, 66.8]  0.18  Estrone sulfate, pmol/L  3281.7 [1710.2, 6223.1]  4490.1 [3193, 7224.6]  <0.001  1225.3 [619.7, 2226.3]  1621.6 [1078.5, 2780.3]  0.031  SHBG, (nmol/L)  43 [31.7, 62.5]  25.3 [19.3, 31.3]  <0.001  36.5 [28.4, 48.2]  24.5 [15, 34.2]  <0.001  1/FI  0.047 ± 0.033  0.038 ± 0.024  0.015  0.052 ± 0.035  0.047 ± 0.039  0.31  SBP, mm Hg  120.4 ± 14.2  117.4 ± 11.8  0.061  126.8 ± 16  123.4 ± 14.1  0.19  DBP, mm Hg  77.7 ± 9.3  79.9 ± 8.9  0.034  78.2 ± 9  77.2 ± 10.1  0.50  HDL, mg/dL  44.8 ± 9.8  41.6 ± 9.3  0.004  48.2 ± 11.5  46.2 ± 10.7  0.26  LDL, mg/dL  121.6 ± 31.7  121.3 ± 26.7  0.93  133.2 ± 35.4  132.2 ± 28.4  0.86  Triglycerides, mg/dL  148.1 ± 85.4  159.5 ± 88.5  0.24  149.8 ± 86.2  152.5 ± 72.8  0.84  LDL particle density, 1/rf  3.8 [3.46, 4]  3.8 [3.46, 4.22]  0.32  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.84  Adiponectin, ng/mL  6.9 [5.3, 8.7]  6 [4.9, 7.3]  <0.001  8.4 [6.5, 11.5]  8.2 [6.3, 9.8]  0.41  High-sensitivity CRP, mg/L  0.7 ± 0.82  0.62 ± 0.55  0.39  0.65 ± 0.74  0.68 ± 0.57  0.78  Fibrinogen, μmol/L  11.2 [9.8, 13.1]  11.4 [10, 12.9]  0.85  12.1 [10.7, 13.8]  12.4 [11.2, 13.6]  0.66  tPA, ng/mL  10.9 [8.8, 13.2]  10.9 [9, 13.1]  0.56  11.1 [9.5, 13.2]  11.3 [9.7, 13.8]  0.49  Any CAC  126 (39.6%)  12 (44.4%)  0.78  72 (59.5%)  8 (66.7%)  0.86  Mean ± SD or n (%) shown; sex hormone median [interquartile range] shown. Participants who did not have testosterone or SHBG levels tested were excluded. Abbreviations: rf, relative flotation; SBP, systolic blood pressure. View Large Table 2. Baseline Participant Characteristics by Menopausal Status and History of IM   Premenopausal Women   Postmenopausal Women     Regular Menses (n = 701)  IM (n = 161)  P  Regular Menses (n = 472)  IM (n = 88)  P  Treatment group, n (%)      1.0      0.87   Placebo  256 (36.5%)  59 (36.6%)    174 (36.9%)  30 (34.1%)     Metformin  222 (31.7%)  51 (31.7%)    143 (30.3%)  27 (30.7%)     ILS  223 (31.8%)  51 (31.7%)    155 (32.8%)  31 (35.2%)    Age group, n (%)      0.14      <0.001   25 to 44 y  384 (54.8%)  99 (61.5%)    31 (6.6%)  18 (20.5%)     45 to 59 y  317 (45.2%)  62 (38.5%)    256 (54.2%)  51 (58%)     ≥60 y        185 (39.2%)  19 (21.6%)    Race or ethnicity, n (%)      0.004      0.012   Non-Hispanic white  344 (49.1%)  79 (49.1%)    237 (50.2%)  41 (46.6%)     African American  174 (24.8%)  30 (18.6%)    151 (32%)  20 (22.7%)     Hispanic  134 (19.1%)  31 (19.3%)    70 (14.8%)  20 (22.7%)     Asian  16 (2.3%)  13 (8.1%)    10 (2.1%)  3 (3.4%)     American Indian  33 (4.7%)  8 (5%)    4 (0.8%)  4 (4.5%)    Number of pregnancies  3 [2, 4]  3 [2, 4]  0.47  3 [2, 4]  3 [2, 4]  0.56   No pregnancies  108 (15.4%)  28 (17.4%)  0.62  62 (13.1%)  12 (13.6%)  1.0  Past estrogen use (birth control pills or estrogen therapy)  539 (76.9%)  123 (76.4%)  0.98  337 (71.4%)  62 (70.5%)  0.96  BMI, kg/m2  35.8 ± 7  37.1 ± 7.6  0.048  34.4 ± 6.3  37.3 ± 8.1  <0.001  Waist circumference, cm  104.7 ± 15.3  107.6 ± 16.8  0.037  103.3 ± 13.6  108.7 ± 16.4  0.001  Fasting glucose, mg/dL  106.7 ± 8.4  105.8 ± 7.6  0.24  106.9 ± 7.9  106.7 ± 7.1  0.81  2-h glucose, mg/dL  163.9 ± 17  165.9 ± 16.7  0.17  164.2 ± 17  163 ± 16.9  0.54  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.5  0.58  6 ± 0.5  6 ± 0.5  0.82  DHEAS, μmol/L  2.25 [1.46, 3.28]  2.43 [1.39, 3.42]  0.81  1.58 [0.98, 2.43]  1.9 [1.04, 2.92]  0.15  Testosterone, nmol/L  0.6 [0.45, 0.84]  0.68 [0.5, 0.98]  0.007  0.51 [0.35, 0.72]  0.52 [0.35, 0.74]  0.712  Estradiol, pmol/L  226.4 [102.7, 397.8]  190.5 [90.4, 348]  0.17  33.8 [21.8, 56.4]  45.4 [23.2, 112.4]  0.059  Estrone sulfate, pmol/L  3473.4 [1780.7, 6279.8]  3547.5 [1742.1, 6038.3]  1.0  1113.8 [556.2, 2040]  1453.7 [828.4, 2435.4]  0.02  SHBG, nmol/L  40.9 [29.9, 59.9]  37.9 [26.7, 58.6]  0.17  35.4 [26, 46.3]  33.4 [23.2, 47.2]  0.30  FAI, mean ± SD  1.7 ± 1.2  2.2 ± 1.7  <0.001  2.4 ± 5.5  2.2 ± 1.8  0.74  Elevated FAI, n (%)  56 (8.1%)  29 (18.7%)  <0.001  33 (9.2%)  11 (14.7%)  0.23  1/FI  0.047 ± 0.033  0.042 ± 0.028  0.081  0.052 ± 0.033  0.045 ± 0.033  0.062  SBP, mm Hg  119.6 ± 13.8  121.9 ± 14.5  0.061  127.7 ± 16  123.8 ± 14.5  0.035  DBP, mm Hg  77.5 ± 9.1  80 ± 10  0.002  77.9 ± 9.2  77.8 ± 8.4  0.90  HDL, mg/dL  44.7 ± 9.7  43.6 ± 9.8  0.19  48.3 ± 11.5  44.7 ± 9.7  0.005  LDL, mg/dL  121.9 ± 30.9  119.3 ± 32.3  0.35  135 ± 35  131.6 ± 35.1  0.41  Triglycerides, mg/dL  147.2 ± 83.4  157.1 ± 94.5  0.19  150.3 ± 82.1  163.6 ± 101.6  0.18  LDL particle density, 1/rf  3.8 [3.46, 4.22]  3.8 [3.46, 4.05]  0.028  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.31  Adiponectin, ng/mL  6.9 [5.4, 8.7]  6.1 [4.9, 7.6]  <0.001  8.5 [6.7, 11.7]  7.8 [6.2, 9.9]  0.007  High-sensitivity CRP, mg/L  0.67 ± 0.79  0.75 ± 0.8  0.24  0.63 ± 0.72  0.67 ± 0.63  0.58  Fibrinogen, μmol/L  11.3 [9.8, 13.1]  11.2 [9.7, 12.9]  0.76  12.1 [10.7, 13.6]  12.2 [10.9, 13.5]  0.83  tPA, ng/mL  10.9 [8.8, 13]  11.1 [9.2, 13.6]  0.20  11.2 [9.4, 13.6]  11.1 [9.8, 13.1]  0.91  Any CAC  121 (40.9%)  20 (36.4%)  0.63  100 (65.4%)  17 (60.7%)  0.80    Premenopausal Women   Postmenopausal Women     Regular Menses (n = 701)  IM (n = 161)  P  Regular Menses (n = 472)  IM (n = 88)  P  Treatment group, n (%)      1.0      0.87   Placebo  256 (36.5%)  59 (36.6%)    174 (36.9%)  30 (34.1%)     Metformin  222 (31.7%)  51 (31.7%)    143 (30.3%)  27 (30.7%)     ILS  223 (31.8%)  51 (31.7%)    155 (32.8%)  31 (35.2%)    Age group, n (%)      0.14      <0.001   25 to 44 y  384 (54.8%)  99 (61.5%)    31 (6.6%)  18 (20.5%)     45 to 59 y  317 (45.2%)  62 (38.5%)    256 (54.2%)  51 (58%)     ≥60 y        185 (39.2%)  19 (21.6%)    Race or ethnicity, n (%)      0.004      0.012   Non-Hispanic white  344 (49.1%)  79 (49.1%)    237 (50.2%)  41 (46.6%)     African American  174 (24.8%)  30 (18.6%)    151 (32%)  20 (22.7%)     Hispanic  134 (19.1%)  31 (19.3%)    70 (14.8%)  20 (22.7%)     Asian  16 (2.3%)  13 (8.1%)    10 (2.1%)  3 (3.4%)     American Indian  33 (4.7%)  8 (5%)    4 (0.8%)  4 (4.5%)    Number of pregnancies  3 [2, 4]  3 [2, 4]  0.47  3 [2, 4]  3 [2, 4]  0.56   No pregnancies  108 (15.4%)  28 (17.4%)  0.62  62 (13.1%)  12 (13.6%)  1.0  Past estrogen use (birth control pills or estrogen therapy)  539 (76.9%)  123 (76.4%)  0.98  337 (71.4%)  62 (70.5%)  0.96  BMI, kg/m2  35.8 ± 7  37.1 ± 7.6  0.048  34.4 ± 6.3  37.3 ± 8.1  <0.001  Waist circumference, cm  104.7 ± 15.3  107.6 ± 16.8  0.037  103.3 ± 13.6  108.7 ± 16.4  0.001  Fasting glucose, mg/dL  106.7 ± 8.4  105.8 ± 7.6  0.24  106.9 ± 7.9  106.7 ± 7.1  0.81  2-h glucose, mg/dL  163.9 ± 17  165.9 ± 16.7  0.17  164.2 ± 17  163 ± 16.9  0.54  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.5  0.58  6 ± 0.5  6 ± 0.5  0.82  DHEAS, μmol/L  2.25 [1.46, 3.28]  2.43 [1.39, 3.42]  0.81  1.58 [0.98, 2.43]  1.9 [1.04, 2.92]  0.15  Testosterone, nmol/L  0.6 [0.45, 0.84]  0.68 [0.5, 0.98]  0.007  0.51 [0.35, 0.72]  0.52 [0.35, 0.74]  0.712  Estradiol, pmol/L  226.4 [102.7, 397.8]  190.5 [90.4, 348]  0.17  33.8 [21.8, 56.4]  45.4 [23.2, 112.4]  0.059  Estrone sulfate, pmol/L  3473.4 [1780.7, 6279.8]  3547.5 [1742.1, 6038.3]  1.0  1113.8 [556.2, 2040]  1453.7 [828.4, 2435.4]  0.02  SHBG, nmol/L  40.9 [29.9, 59.9]  37.9 [26.7, 58.6]  0.17  35.4 [26, 46.3]  33.4 [23.2, 47.2]  0.30  FAI, mean ± SD  1.7 ± 1.2  2.2 ± 1.7  <0.001  2.4 ± 5.5  2.2 ± 1.8  0.74  Elevated FAI, n (%)  56 (8.1%)  29 (18.7%)  <0.001  33 (9.2%)  11 (14.7%)  0.23  1/FI  0.047 ± 0.033  0.042 ± 0.028  0.081  0.052 ± 0.033  0.045 ± 0.033  0.062  SBP, mm Hg  119.6 ± 13.8  121.9 ± 14.5  0.061  127.7 ± 16  123.8 ± 14.5  0.035  DBP, mm Hg  77.5 ± 9.1  80 ± 10  0.002  77.9 ± 9.2  77.8 ± 8.4  0.90  HDL, mg/dL  44.7 ± 9.7  43.6 ± 9.8  0.19  48.3 ± 11.5  44.7 ± 9.7  0.005  LDL, mg/dL  121.9 ± 30.9  119.3 ± 32.3  0.35  135 ± 35  131.6 ± 35.1  0.41  Triglycerides, mg/dL  147.2 ± 83.4  157.1 ± 94.5  0.19  150.3 ± 82.1  163.6 ± 101.6  0.18  LDL particle density, 1/rf  3.8 [3.46, 4.22]  3.8 [3.46, 4.05]  0.028  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.31  Adiponectin, ng/mL  6.9 [5.4, 8.7]  6.1 [4.9, 7.6]  <0.001  8.5 [6.7, 11.7]  7.8 [6.2, 9.9]  0.007  High-sensitivity CRP, mg/L  0.67 ± 0.79  0.75 ± 0.8  0.24  0.63 ± 0.72  0.67 ± 0.63  0.58  Fibrinogen, μmol/L  11.3 [9.8, 13.1]  11.2 [9.7, 12.9]  0.76  12.1 [10.7, 13.6]  12.2 [10.9, 13.5]  0.83  tPA, ng/mL  10.9 [8.8, 13]  11.1 [9.2, 13.6]  0.20  11.2 [9.4, 13.6]  11.1 [9.8, 13.1]  0.91  Any CAC  121 (40.9%)  20 (36.4%)  0.63  100 (65.4%)  17 (60.7%)  0.80  Mean ± SD or n (%) shown; sex hormone median [interquartile range] shown. Abbreviations: rf, relative flotation; SBP, systolic blood pressure. View Large Table 3. Risk of Incident Diabetes Associated With FAI and IM During the DPP and the DPPOS   Risk During DPP, HR (95% CI)  P  Risk During DPP and DPPOS, HR (95% CI)  P  FAI (examined as a continuous measure)           Model 1: Adjusted for menopausal status  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.32   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.29   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.99 (0.95 to 1.04)  0.79  0.98 (0.94 to 1.02)  0.31   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.99 (0.94 to 1.04)  0.75  0.98 (0.93 to 1.02)  0.28   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.99 (0.93 to 1.05)  0.74  0.97 (0.93 to 1.02)  0.27  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.98 (0.73 to 1.31)  0.88  1.12 (0.92 to 1.37)  0.27   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.01 (0.75 to 1.35)  0.97  1.12 (0.92 to 1.37)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.01 (0.75 to 1.35)  0.95  1.12 (0.92 to 1.37)  0.27   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.95 (0.71 to 1.28)  0.76  1.07 (0.87 to 1.31)  0.54   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.97 (0.72 to 1.30)  0.82  1.07 (0.87 to 1.31)  0.54    Risk During DPP, HR (95% CI)  P  Risk During DPP and DPPOS, HR (95% CI)  P  FAI (examined as a continuous measure)           Model 1: Adjusted for menopausal status  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.32   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.29   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.99 (0.95 to 1.04)  0.79  0.98 (0.94 to 1.02)  0.31   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.99 (0.94 to 1.04)  0.75  0.98 (0.93 to 1.02)  0.28   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.99 (0.93 to 1.05)  0.74  0.97 (0.93 to 1.02)  0.27  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.98 (0.73 to 1.31)  0.88  1.12 (0.92 to 1.37)  0.27   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.01 (0.75 to 1.35)  0.97  1.12 (0.92 to 1.37)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.01 (0.75 to 1.35)  0.95  1.12 (0.92 to 1.37)  0.27   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.95 (0.71 to 1.28)  0.76  1.07 (0.87 to 1.31)  0.54   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.97 (0.72 to 1.30)  0.82  1.07 (0.87 to 1.31)  0.54  FAI is examined as a continuous measure, and HR >1 indicates risk associated with each unit increase in FAI. For analyses of IM, regular menses is the reference, and HR >1 indicates greater risk associated with IM. Abbreviations: CI, confidence interval; HR, hazard ratio. View Large Table 4. Associations Between FAI and IM With Odds of Any CAC at Year 10 of the DPPOS   OR (95% CI)  P  FAI (examined as a continuous measure)       Model 1: Adjusted for menopausal status  1.05 (0.92 to 1.21)  0.069   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.09 (0.95 to 1.27)  0.074   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.09 (0.95 to 1.27)  0.074   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  1.07 (0.92 to 1.24)  0.075   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  1.06 (0.92 to 1.23)  0.075  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.82 (0.51 to 1.33)  0.25   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.92 (0.56 to 1.53)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.92 (0.55 to 1.52)  0.26   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.87 (0.52 to 1.46)  0.26   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.89 (0.53 to 1.49)  0.26    OR (95% CI)  P  FAI (examined as a continuous measure)       Model 1: Adjusted for menopausal status  1.05 (0.92 to 1.21)  0.069   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.09 (0.95 to 1.27)  0.074   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.09 (0.95 to 1.27)  0.074   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  1.07 (0.92 to 1.24)  0.075   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  1.06 (0.92 to 1.23)  0.075  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.82 (0.51 to 1.33)  0.25   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.92 (0.56 to 1.53)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.92 (0.55 to 1.52)  0.26   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.87 (0.52 to 1.46)  0.26   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.89 (0.53 to 1.49)  0.26  For analyses of FAI, FAI is examined as a continuous measure, and OR >1 indicates odds associated with each unit increase in FAI. For analyses of IM, regular menses is the reference, and OR >1 indicates greater odds associated with IM. Abbreviations: CI, confidence interval; OR, odds ratio. View Large In sensitivity analyses, we attempted to characterize the risk of incident diabetes (Fig. 1) and odds of prevalent CAC (Fig. 2) among women who had both relative elevations in FAI and IM. As with diabetes risk, interaction terms for randomization arm and for menopause with FAI and IM were not significant, and thus models combined data across randomization arms as well as menopausal status; analyses presenting stratified results are presented in Supplemental Tables 2 and 3. Analyses were performed in R software version 3.4.0 (R Core Team, Vienna, Austria), and all tests were two sided, with statistical significance set at P < 0.05. No adjustments for multiple comparisons were performed, and all P values are nominal. Figure 1. View largeDownload slide Risk of incident diabetes associated with relative elevations in FAI and IM during the DPP (top panel) and the DPPOS (bottom panel) [hazard ratio (HR) and 95% confidence interval (CI)]. Reference group is women without either abnormality. Women without either abnormality are the reference group, so that an HR >1 indicates greater diabetes risk if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 1. View largeDownload slide Risk of incident diabetes associated with relative elevations in FAI and IM during the DPP (top panel) and the DPPOS (bottom panel) [hazard ratio (HR) and 95% confidence interval (CI)]. Reference group is women without either abnormality. Women without either abnormality are the reference group, so that an HR >1 indicates greater diabetes risk if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 2. View largeDownload slide Associations between relative elevations in FAI and IM with odds of any CAC at year 10 DPPOS [odds ratio (OR) and 95% confidence interval (CI)]. The reference group is women without either abnormality, so that an OR >1 indicates greater odds of CAC if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 2. View largeDownload slide Associations between relative elevations in FAI and IM with odds of any CAC at year 10 DPPOS [odds ratio (OR) and 95% confidence interval (CI)]. The reference group is women without either abnormality, so that an OR >1 indicates greater odds of CAC if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Results Table 1 shows participant characteristics at baseline by category of FAI elevation and menopausal status. Women had an average age of 48.2 ± 9.9 years; premenopausal women had an average age of 42.9 ± 5.8 years, and postmenopausal women had an average age of 56.4 ± 9.4. Among women who were premenopausal at baseline, women in the upper 10th percentile of FAI were slightly younger. Premenopausal women with higher FAI also had greater BMI, waist circumference, postchallenge glucose, DHEAS, estradiol, and estrone sulfate concentrations. Premenopausal women with higher FAI also had higher diastolic blood pressure (DBP) and lower insulin sensitivity, HDL, and adiponectin. In these unadjusted comparisons, presence of CAC was similar by FAI percentile. Among women who were postmenopausal at baseline, women with higher FAI were slightly younger and had higher waist circumference and estrone sulfate concentrations, although glucose levels and sex hormone concentrations (other than testosterone and SHBG) were similar between groups. Among postmenopausal women, no significant differences in other markers of risk were observed by FAI status. Table 2 shows participant characteristics at baseline by history of IM and menopausal status. Among women who were premenopausal at baseline, women with a history of IM were slightly less likely to be African American. Premenopausal women with histories of IM had higher BMIs and waist circumferences than women without such histories. Sex hormone profiles were similar by IM, except that women with IM had higher FAI and testosterone concentrations than women without such histories. Premenopausal women with histories of IM had lower adiponectin and higher DBP. In these unadjusted comparisons, presence of CAC was similar by FAI status. Among women who were postmenopausal at baseline, women with histories of IM were younger and more likely to be Hispanic and less likely to be African American. Postmenopausal women with histories of IM had higher BMI, waist circumference, and estradiol and estrone-sulfate levels and lower insulin sensitivity, systolic blood pressure, HDL, and adiponectin levels than women without such histories. Table 3 shows the risk of incident diabetes during the DPP and DPPOS associated with FAI and IM. None of these exposures were associated with increased risk of progression to diabetes during the DPP or the DPPOS, before and after adjustment for covariates. Table 4 shows the odds of any CAC at year 10 of the DPPOS associated with FAI and IM. None of these exposures were associated with increased odds of any CAC, before and after adjustment for covariates. We conducted several sensitivity analyses. We determined whether the subset of women with both FAI elevations and IM, two central components of PCOS, had greater risk of diabetes than women without either FAI elevations or IM: 1000 women were included in this analysis, 40 of whom had both conditions and 960 of whom had neither IM nor elevated FAI. In multivariable models that adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI, women with both conditions had a similar risk of progression to diabetes during the DPP and the DPPOS as women without either condition (Fig. 1). We also evaluated whether the subset of women with both FAI elevations and IM had higher odds of any CAC compared with women with lower FAI and regular menses. In multivariable models, these women had a similar pattern of risk compared with women with either FAI elevations or IM, that is, there was no association (Fig. 2). Finally, we performed stratified analyses by menopausal status, but we found similar patterns to when menopause was used as an adjuster (Supplemental Tables 2 and 3). Conclusions PCOS is a well-recognized risk factor for increased risk of diabetes and for subclinical and clinical cardiovascular disease. Studies have differed regarding whether the elevated androgens or IM that characterize PCOS also confer increased risk in women without PCOS. Specifically, it is unclear whether these conditions confer additional risk in women who are glucose intolerant and overweight, the primary risk factors for diabetes. Using data from the DPP and DPPOS, a large multiethnic cohort of women at high risk for diabetes, we found that neither higher androgen levels nor IM conferred additional risk for diabetes or CAC burden. Although our results do not rule out the possibility that relative androgen elevations or IM increased risk of glucose intolerance or overweight, our findings suggest that the presence of these disorders does not modify progression from impaired glucose tolerance to diabetes nor increase risk of atherosclerosis, as reflected by CAC. Whether the androgen elevation that is typical of PCOS can also be associated with increased risk of cardiometabolic disease for women without PCOS has been controversial. The majority of previous studies have examined women with a range of BMIs and degrees of glucose tolerance. A 2006 meta-analysis found cross-sectional associations, but longitudinal associations had borderline significance (29). Our findings are similar to more recent longitudinal studies (11, 12) that suggest that relative elevations in androgens do not confer risk beyond associations with initial elevations in BMI and glucose intolerance. Even though women with elevations in FAI also had additional diabetes risk factors such as lower adiponectin levels (30), it did not increase their overall risk of diabetes compared with women with lower FAI who were also obese and glucose-intolerant. Several previous reports have suggested that elevations in total testosterone are associated with greater carotid intima media thickness (15–18, 31), although associations between androgens and presence of CAC have not been significant (13, 14, 16). We found that even though relative elevations in androgens at DPP baseline were associated with poorer metabolic indices, these did not translate to greater CAC burden a decade later. It is possible that androgens have different effects on different vascular beds or that the relative burden of subclinical atherosclerosis was low and burden of CAC not particularly sensitive to androgen associations. With longer follow-up, it is possible that the association between androgen levels and CAC would be stronger. However, the lack of association with CAC is reassuring in that the absence of CAC tends to be more strongly associated with incident cardiovascular disease than other subclinical markers (32, 33). Previous reports have also been inconsistent regarding associations between IM and cardiometabolic risk among women without known PCOS. Studies have reported significant associations with incident diabetes (19, 20) or no associations (11, 23). Definitions of IM differed across studies and may have contributed to the disparate results, along with the examination of participants explicitly diagnosed with PCOS as well as those without diagnosed PCOS. In the latter population, IM may be more likely to be multifactorial. To our knowledge, only one longitudinal study has examined whether IM are associated with subclinical atherosclerosis and, similarly to our study, noted no association (11). Two studies have examined the relationship between IM and coronary disease outcomes; one noted no association with coronary heart disease mortality after adjustment for BMI (22), and the other found a significant association between IM and coronary disease (21). It is possible that these reports differed because of a stricter definition of IM in the latter report, which required ≥40 days between cycles or irregularity that did not allow estimation of cycle length (21), thus capturing women with more severe degrees of hormonal abnormalities. It is possible that the relationship between elevated androgen levels and future risk of diabetes and cardiovascular disease is mediated through obesity and associated abnormalities. Using data from the Study of Women’s Health Across the Nation, El Khoudary et al. (34) noted that a higher FAI was associated with CAC in nonobese women but not in obese women. Among premenopausal women, higher FAI was associated with higher BMI and waist circumference as well as higher postchallenge glucose, higher DBP, and lower insulin sensitivity, HDL, and adiponectin. It is also possible that PCOS increases cardiometabolic risk through pathways other than elevated androgen levels and those involving the menstrual cycle. Thus, these conditions may serve as a proxy for other cardiometabolic risk factors in the PCOS population (35); other reports have noted that male relatives of women with known PCOS have elevated cardiometabolic risk, suggesting shared genetic risk factors (36). In support of this hypothesis, one report found that women with PCOS, defined as a combination of hyperandrogenism and IM, had increased odds of CAC, but neither isolated hyperandrogenism nor IM alone predicted burden of CAC (13). We did not find that the combination of these conditions increased risk, although the number of women with both conditions in our sample was small. Currently, no single definition of biochemical hyperandrogenism in women exists because of variation in the assays used across studies and the decline in average androgen levels with age. Longitudinal studies of cohorts with PCOS have noted that androgens decreased over the reproductive lifespan and may be unremarkable in the fourth decade of life (37, 38) even as their waist circumference and degree of insulin resistance remain elevated compared with women without PCOS. We examined a cohort of women whose average age was ~50 years. Even though we examined relative elevations in androgen levels within the cohort, we may not have captured women with histories of elevated androgen levels in their earlier reproductive years. This possibility is supported by the fact that relative elevations in androgens were associated with a higher number of differences in cardiometabolic markers in premenopausal compared with postmenopausal women. Along similar lines, ~44% of women with PCOS eventually have regular menses before menopause (37). We note that the DPP requirements for overweight and glucose-intolerant women may have captured women who developed these conditions subsequent to histories of PCOS, and we were unable to determine the chronological order of their development of IM and hyperandrogenism in relation to their enrollment criteria. The strengths of this report include its use of a comparison group that was also overweight and glucose intolerant, as well as longitudinal assessment of diabetes and examination of subclinical markers of atherosclerosis and an extensive list of biomarkers. We also used sensitive mass spectrometric methods to assess sex steroid measures, which is particularly pertinent for measures of sex steroids in postmenopausal women. However, there are several limitations. As is true for other reports of hyperandrogenism in midlife women, no single cutoff point exists, and it is possible that women were misclassified, although we addressed this issue to some extent by defining androgen elevations within the population. IM was obtained by self-report, and further details of cycle duration and variability were not available. Transvaginal ultrasound was not performed, and data regarding ovarian size and polycystic ovarian morphology were not available. The measurement of other hormones to rule out other endocrine disorders that can cause IM, such as thyroid-stimulating hormone, prolactin, and 17-hydroxyprogesterone, was not performed. Finally, the burden of CAC was low, and thus our power to assess odds of CAC was limited. We conclude that compared with midlife women who are already glucose intolerant and overweight, histories of IM and relative elevations in androgens do not additionally increase risk or the burden of subclinical atherosclerosis. Although our report does not rule out that these conditions may identify women at risk in their younger years, it suggests that chronic disease burden in midlife women may be most effectively addressed through weight loss, metformin, and the other diabetes prevention measures already known to be beneficial for all women at elevated risk for diabetes. Future investigations should examine whether relative elevations in androgens or IM may further define high-risk women at younger ages. Abbreviations: BMI body mass index CAC coronary artery calcification CRP C-reactive protein CV coefficient of variation DBP diastolic blood pressure DHEAS dehydroepiandrosterone sulfate DPP Diabetes Prevention Program DPPOS Diabetes Prevention Program Outcomes Study FAI free androgen index FI fasting insulin HDL high-density lipoprotein ILS intensive lifestyle intervention IM irregular menses LDL low-density lipoprotein PCOS polycystic ovary syndrome SD standard deviation SHBG sex hormone binding globulin tPA tissue plasminogen activator. Acknowledgments The Research Group gratefully acknowledges the commitment and dedication of the participants in the Diabetes Prevention Program. This work was presented in part at the 2016 Scientific Sessions of the American Diabetes Association. Financial Support: This study was supported by Grants DK072041, DK048489, and DK53061 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health, which provided funding to the clinical centers and the Coordinating Center for the design and conduct of the study and the collection, management, analysis, and interpretation of the data. The Southwestern American Indian Centers were supported directly by NIDDK, including its Intramural Research Program, and the Indian Health Service. The General Clinical Research Center Program, National Center for Research Resources, and the Department of Veterans Affairs supported data collection at many of the clinical centers. Funding was also provided by the National Institute of Child Health and Human Development, National Institute on Aging, National Eye Institute, National Heart, Lung, and Blood Institute, Office of Research on Women’s Health, National Institute on Minority Health and Health Disparities, Centers for Disease Control and Prevention, and American Diabetes Association. Bristol-Myers Squibb and Parke-Davis provided additional funding and material support during DPP, Lipha (Merck-Sante) provided medication, and LifeScan Inc. donated materials during DPP and DPPOS. The opinions expressed herein are those of the investigators and do not necessarily reflect the views of the funding agencies. A complete list of centers, investigators, and staff can be found in the online appendix. The research group gratefully acknowledges the commitment and dedication of the participants of the DPP and DPPOS. Clinical Trial Information: ClinicalTrials.gov no. NCT00038727 (registered 4 June 2002). Author Contributions: C.K. wrote the manuscript and researched the sex hormone data and is the guarantor of the manuscript. N.Y. provided the analysis and reviewed/edited the manuscript. D.A.E., V.R.A., S.L.E., and R.B.G. reviewed and edited the manuscript, and R.B.G. researched the adiponectin and inflammatory marker assays. Disclosure Summary: The authors have nothing to disclose. References 1. Legro RS, Arslanian SA, Ehrmann DA, Hoeger KM, Murad MH, Pasquali R, Welt CK; Endocrine Society. Diagnosis and treatment of polycystic ovary syndrome: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab . 2013; 98( 12): 4565– 4592. Google Scholar CrossRef Search ADS PubMed  2. Song DK, Hong YS, Sung YA, Lee H. Insulin resistance according to β-cell function in women with polycystic ovary syndrome and normal glucose tolerance. PLoS One . 2017; 12( 5): e0178120. Google Scholar CrossRef Search ADS PubMed  3. Ciaraldi TP, Aroda V, Mudaliar S, Chang RJ, Henry RR. Polycystic ovary syndrome is associated with tissue-specific differences in insulin resistance. J Clin Endocrinol Metab . 2009; 94( 1): 157– 163. Google Scholar CrossRef Search ADS PubMed  4. Palomba S, Falbo A, Russo T, Manguso F, Tolino A, Zullo F, De Feo P, Orio F, Jr. Insulin sensitivity after metformin suspension in normal-weight women with polycystic ovary syndrome. J Clin Endocrinol Metab . 2007; 92( 8): 3128– 3135. Google Scholar CrossRef Search ADS PubMed  5. Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK, Imperial J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care . 1999; 22( 1): 141– 146. Google Scholar CrossRef Search ADS PubMed  6. Joham AE, Ranasinha S, Zoungas S, Moran L, Teede HJ. Gestational diabetes and type 2 diabetes in reproductive-aged women with polycystic ovary syndrome. J Clin Endocrinol Metab . 2014; 99( 3): E447– E452. Google Scholar CrossRef Search ADS PubMed  7. Lo JC, Feigenbaum SL, Yang J, Pressman AR, Selby JV, Go AS. Epidemiology and adverse cardiovascular risk profile of diagnosed polycystic ovary syndrome. J Clin Endocrinol Metab . 2006; 91( 4): 1357– 1363. Google Scholar CrossRef Search ADS PubMed  8. Wang ET, Calderon-Margalit R, Cedars MI, Daviglus ML, Merkin SS, Schreiner PJ, Sternfeld B, Wellons M, Schwartz SM, Lewis CE, Williams OD, Siscovick DS, Bibbins-Domingo K. Polycystic ovary syndrome and risk for long-term diabetes and dyslipidemia. Obstet Gynecol . 2011; 117( 1): 6– 13. Google Scholar CrossRef Search ADS PubMed  9. Charles MA, Pettitt DJ, McCance DR, Hanson RL, Bennett PH, Knowler WC. Gravidity, obesity, and non-insulin-dependent diabetes among Pima Indian women. Am J Med . 1994; 97( 3): 250– 255. Google Scholar CrossRef Search ADS PubMed  10. Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA . 2006; 295( 11): 1288– 1299. Google Scholar CrossRef Search ADS PubMed  11. Calderon-Margalit R, Schwartz SM, Wellons MF, Lewis CE, Daviglus ML, Schreiner PJ, Williams OD, Sternfeld B, Carr JJ, O’Leary DH, Sidney S, Friedlander Y, Siscovick DS. Prospective association of serum androgens and sex hormone-binding globulin with subclinical cardiovascular disease in young adult women: the “Coronary Artery Risk Development in Young Adults” women’s study. J Clin Endocrinol Metab . 2010; 95( 9): 4424– 4431. Google Scholar CrossRef Search ADS PubMed  12. Fenske B, Kische H, Gross S, Wallaschofski H, Völzke H, Dörr M, Nauck M, Keevil BG, Brabant G, Haring R. Endogenous androgens and sex hormone-binding globulin in women and risk of metabolic syndrome and type 2 diabetes. J Clin Endocrinol Metab . 2015; 100( 12): 4595– 4603. Google Scholar CrossRef Search ADS PubMed  13. Calderon-Margalit R, Siscovick D, Merkin SS, Wang E, Daviglus ML, Schreiner PJ, Sternfeld B, Williams OD, Lewis CE, Azziz R, Schwartz SM, Wellons MF. Prospective association of polycystic ovary syndrome with coronary artery calcification and carotid-intima-media thickness: the Coronary Artery Risk Development in Young Adults Women’s study. Arterioscler Thromb Vasc Biol . 2014; 34( 12): 2688– 2694. Google Scholar CrossRef Search ADS PubMed  14. Ouyang P, Vaidya D, Dobs A, Golden SH, Szklo M, Heckbert SR, Kopp P, Gapstur SM. Sex hormone levels and subclinical atherosclerosis in postmenopausal women: the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis . 2009; 204( 1): 255– 261. Google Scholar CrossRef Search ADS PubMed  15. Kische H, Gross S, Wallaschofski H, Völzke H, Dörr M, Nauck M, Felix SB, Haring R. Serum androgen concentrations and subclinical measures of cardiovascular disease in men and women. Atherosclerosis . 2016; 247: 193– 200. Google Scholar CrossRef Search ADS PubMed  16. Karim R, Hodis HN, Stanczyk FZ, Lobo RA, Mack WJ. Relationship between serum levels of sex hormones and progression of subclinical atherosclerosis in postmenopausal women. J Clin Endocrinol Metab . 2008; 93( 1): 131– 138. Google Scholar CrossRef Search ADS PubMed  17. Bernini GP, Sgro’ M, Moretti A, Argenio GF, Barlascini CO, Cristofani R, Salvetti A. Endogenous androgens and carotid intimal-medial thickness in women. J Clin Endocrinol Metab . 1999; 84( 6): 2008– 2012. Google Scholar CrossRef Search ADS PubMed  18. Golden SH, Maguire A, Ding J, Crouse JR, Cauley JA, Zacur H, Szklo M. Endogenous postmenopausal hormones and carotid atherosclerosis: a case-control study of the atherosclerosis risk in communities cohort. Am J Epidemiol . 2002; 155( 5): 437– 445. Google Scholar CrossRef Search ADS PubMed  19. Solomon CG, Hu FB, Dunaif A, Rich-Edwards J, Willett WC, Hunter DJ, Colditz GA, Speizer FE, Manson JE. Long or highly irregular menstrual cycles as a marker for risk of type 2 diabetes mellitus. JAMA . 2001; 286( 19): 2421– 2426. Google Scholar CrossRef Search ADS PubMed  20. Rostami Dovom M, Ramezani Tehrani F, Djalalinia S, Cheraghi L, Behboudi Gandavani S, Azizi F. Menstrual cycle irregularity and metabolic disorders: a population-based prospective study. PLoS One . 2016; 11( 12): e0168402. Google Scholar CrossRef Search ADS PubMed  21. Solomon CG, Hu FB, Dunaif A, Rich-Edwards JE, Stampfer MJ, Willett WC, Speizer FE, Manson JE. Menstrual cycle irregularity and risk for future cardiovascular disease. J Clin Endocrinol Metab . 2002; 87( 5): 2013– 2017. Google Scholar CrossRef Search ADS PubMed  22. Wang ET, Cirillo PM, Vittinghoff E, Bibbins-Domingo K, Cohn BA, Cedars MI. Menstrual irregularity and cardiovascular mortality. J Clin Endocrinol Metab . 2011; 96( 1): E114– E118. Google Scholar CrossRef Search ADS PubMed  23. LeBlanc ES, Kapphahn K, Hedlin H, Desai M, Parikh NI, Liu S, Parker DR, Anderson M, Aroda V, Sullivan S, Woods NF, Waring ME, Lewis CE, Stefanick M. Reproductive history and risk of type 2 diabetes mellitus in postmenopausal women: findings from the Women’s Health Initiative. Menopause . 2017; 24( 1): 64– 72. Google Scholar CrossRef Search ADS PubMed  24. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM; Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med . 2002; 346( 6): 393– 403. Google Scholar CrossRef Search ADS PubMed  25. Venditti EM, Bray GA, Carrion-Petersen ML, Delahanty LM, Edelstein SL, Hamman RF, Hoskin MA, Knowler WC, Ma Y; Diabetes Prevention Program Research Group. First versus repeat treatment with a lifestyle intervention program: attendance and weight loss outcomes. Int J Obes . 2008; 32( 10): 1537– 1544. Google Scholar CrossRef Search ADS   26. Goldberg RB, Aroda VR, Bluemke DA, Barrett-Connor E, Budoff M, Crandall JP, Dabelea D, Horton ES, Mather KJ, Orchard TJ, Schade D, Watson K, Temprosa M; Diabetes Prevention Program Research Group. Effect of long-term metformin and lifestyle in the Diabetes Prevention Program and its outcome study on coronary artery calcium. Circulation . 2017; 136( 1): 52– 64. Google Scholar CrossRef Search ADS PubMed  27. Hartigan J, Wong M. Algorithm AS 136: a K-means clustering algorithm. J R Stat Soc Ser C Appl Stat . 1979; 28: 100– 108. 28. Breiman L, Freidman J, Olshen RA, Stone CJ. Classification and Regression Trees.  Wadsworth Statistics/Probability. New York, NY: Chapman & Hall/CRC; 1984. 29. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, Buring JE, Gaziano JM, Liu S. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med . 2009; 361( 12): 1152– 1163. Google Scholar CrossRef Search ADS PubMed  30. Spranger J, Möhlig M, Wegewitz U, Ristow M, Pfeiffer AF, Schill T, Schlösser HW, Brabant G, Schöfl C. Adiponectin is independently associated with insulin sensitivity in women with polycystic ovary syndrome. Clin Endocrinol (Oxf) . 2004; 61( 6): 738– 746. Google Scholar CrossRef Search ADS PubMed  31. Creatsa M, Armeni E, Stamatelopoulos K, Rizos D, Georgiopoulos G, Kazani M, Alexandrou A, Dendrinos S, Augoulea A, Papamichael C, Lambrinoudaki I. Circulating androgen levels are associated with subclinical atherosclerosis and arterial stiffness in healthy recently menopausal women. Metabolism . 2012; 61( 2): 193– 201. Google Scholar CrossRef Search ADS PubMed  32. Folsom AR, Kronmal RA, Detrano RC, O’Leary DH, Bild DE, Bluemke DA, Budoff MJ, Liu K, Shea S, Szklo M, Tracy RP, Watson KE, Burke GL. Coronary artery calcification compared with carotid intima-media thickness in the prediction of cardiovascular disease incidence: the Multi-Ethnic Study of Atherosclerosis (MESA). Arch Intern Med . 2008; 168( 12): 1333– 1339. Google Scholar CrossRef Search ADS PubMed  33. Joshi PH, Patel B, Blaha MJ, Berry JD, Blankstein R, Budoff MJ, Wong N, Agatston A, Blumenthal RS, Nasir K. Coronary artery calcium predicts cardiovascular events in participants with a low lifetime risk of cardiovascular disease: the Multi-Ethnic Study of Atherosclerosis (MESA). Atherosclerosis . 2016; 246: 367– 373. Google Scholar CrossRef Search ADS PubMed  34. El Khoudary SR, Wildman RP, Matthews K, Powell L, Hollenberg SM, Edmundowicz D, Sutton-Tyrrell K. Effect modification of obesity on associations between endogenous steroid sex hormones and arterial calcification in women at midlife. Menopause . 2011; 18( 8): 906– 914. Google Scholar CrossRef Search ADS PubMed  35. Rosenfield RL, Ehrmann DA. The pathogenesis of polycystic ovary syndrome (PCOS): the hypothesis of PCOS as functional ovarian hyperandrogenism revisited. Endocr Rev . 2016; 37( 5): 467– 520. Google Scholar CrossRef Search ADS PubMed  36. Vipin VP, Dabadghao P, Shukla M, Kapoor A, Raghuvanshi AS, Ramesh V. Cardiovascular disease risk in first-degree relatives of women with polycystic ovary syndrome. Fertil Steril . 2016; 105( 5): 1338– 1344.e3. Google Scholar CrossRef Search ADS PubMed  37. Carmina E, Campagna AM, Mansuet P, Vitale G, Kort D, Lobo R. Does the level of serum antimüllerian hormone predict ovulatory function in women with polycystic ovary syndrome with aging? Fertil Steril . 2012; 98( 4): 1043– 1046. Google Scholar CrossRef Search ADS PubMed  38. Brown ZA, Louwers YV, Fong SL, Valkenburg O, Birnie E, de Jong FH, Fauser BC, Laven JS. The phenotype of polycystic ovary syndrome ameliorates with aging. Fertil Steril . 2011; 96( 5): 1259– 1265. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Clinical Endocrinology and Metabolism Oxford University Press

Androgens, Irregular Menses, and Risk of Diabetes and Coronary Artery Calcification in the Diabetes Prevention Program

Loading next page...
 
/lp/ou_press/androgens-irregular-menses-and-risk-of-diabetes-and-coronary-artery-sWYsCNCpAu
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0021-972X
eISSN
1945-7197
D.O.I.
10.1210/jc.2017-01829
Publisher site
See Article on Publisher Site

Abstract

Abstract Context It is unclear whether relative elevations in androgens or irregular menses (IM) are associated with greater cardiometabolic risk among women who are already overweight and glucose intolerant. Research Design and Methods We conducted a secondary analysis of the Diabetes Prevention Program (DPP) and the Diabetes Prevention Program Outcomes Study (DPPOS). Participants included women with sex hormone measurements who did not use exogenous estrogen (n = 1422). We examined whether free androgen index (FAI) or IM was associated with diabetes risk during the DPP/DPPOS or with coronary artery calcification (CAC) at DPPOS year 10. Models were adjusted for menopausal status, age, race or ethnicity, randomization arm, body mass index (BMI), and hemoglobin A1c. Results Women had an average age of 48.2 ± 9.9 years. Elevations in FAI and IM were associated with greater BMI, waist circumference, and blood pressure and lower adiponectin. FAI was not associated with diabetes risk during the DPP/DPPOS [hazard ratio (HR) 0.97; 95% confidence interval (CI), 0.93 to 1.02] or increased odds of CAC [odds ratio (OR) 1.06; 95% CI, 0.92 to 1.23]. IM was also not associated with diabetes risk during the DPP/DPPOS (HR 1.07; 95% CI, 0.87 to 1.31) or increased odds of CAC (OR 0.89; 95% CI, 0.53 to 1.49). Women who had both relative elevations in FAI and IM had similar diabetes risk and odds of CAC as women without these conditions. Differences by treatment arm and menopausal status were not observed. Conclusions Among midlife women who were already glucose intolerant and overweight, androgen concentrations and IM did not additionally contribute to increased risk for diabetes or CAC. Polycystic ovary syndrome (PCOS) is a common endocrinopathy that may affect up to one in five women of reproductive age (1). Existing definitions are based on the presence of two or more of the following criteria: relative elevations in androgens, irregular menses (IM), and polycystic ovary morphology (1). Women with PCOS have greater insulin resistance compared with age- and weight-matched controls (2–4) and subsequently higher risk of type 2 diabetes (5–9). However, the extent to which central PCOS components (i.e., androgens and IM) increase cardiometabolic risk among women without a diagnosis of PCOS is unclear. Specifically, it is unknown whether these characteristics increase risk beyond the presence of impaired glucose tolerance and overweight alone. Although one meta-analysis noted that women with elevated total testosterone concentrations had higher risk of diabetes in cross-sectional analyses (10), more recent longitudinal studies of women with a range of body mass indices (BMIs) and degrees of glucose tolerance have not observed associations between hyperandrogenism and diabetes (11, 12). Reports also conflict as to whether hyperandrogenism is associated with coronary artery calcification (CAC) (13, 14) or carotid intima media thickness (15–18), both of which are indicators of subclinical atherosclerosis. Similarly, some studies report an association of IM with incident diabetes (19, 20) and incident coronary disease (21) in populations with a range of BMIs and degrees of glucose tolerance, but others have noted that these associations are no longer significant after adjustment for BMI (13, 22, 23). The Diabetes Prevention Program (DPP) randomly assigned overweight, nondiabetic glucose-intolerant participants to a program of intensive lifestyle modification (ILS), metformin, or placebo (24). The follow-up observational cohort, the Diabetes Prevention Program Outcomes Study (DPPOS), has continued to ascertain incident diabetes semiannually and has also measured CAC. Thus, we were able to assess whether androgens or history of IM at study entry were associated with increased risk factors for cardiometabolic disease beyond their known associations with impaired glucose tolerance and increased weight. For the present report, we examined traditional and nontraditional cardiovascular risk factors, progression to diabetes, and presence of CAC. We hypothesized that women with relative androgen elevations or IM would have adverse risk factor profiles compared with other overweight and glucose-intolerant women without these conditions. We also hypothesized that the free androgen index (FAI), a measure of androgen levels, and IM would be associated with increased risk of incident diabetes and a higher prevalence of detectable CAC. Research Design and Methods Population and setting The design, methods, and baseline characteristics of the DPP have been previously described (24). Between 1996 and 1999, participants were recruited across 27 clinical centers located throughout the United States. Inclusion criteria were age ≥25 years, BMI ≥24 kg/m2 (≥22 kg/m2 for Asian Americans), a fasting plasma glucose level of 5.3 to 7.0 mmol/L (95 to 125 mg/dL), and a 2-hour plasma glucose level of 7.8 to 11.1 mmol/L (140 to 199 mg/dL) after an oral 75-g glucose load. Eligible participants were randomly assigned to one of three interventions: 850 mg metformin twice daily, placebo twice daily, or ILS. The goals of ILS were to achieve and maintain a weight reduction of ≥7% through consumption of a low-calorie, low-fat diet, plus moderate physical activity for ≥150 minutes per week. Each participating institution was overseen by its respective ethics review board, and written informed consent was obtained from all participants. IM, pregnancy, and menopause Standardized interviewer-administered questionnaires were used to obtain other demographic and clinical data. At the time of DPP randomization, a screening questionnaire inquired about women’s menses. Women were asked, “During most of your life, were your periods regular? That is, did they occur once a month?” (Do not include any time when you were pregnant or taking birth control pills.) Women who answered “no” or “sometimes regular, sometimes irregular” were classified as having IM. Questionnaires assessed previous use of estrogen, consisting of birth control pills or estrogen therapy, as well as menopausal status. Women were classified as postmenopausal if they reported absence of menses for a year, bilateral oophorectomy, or hysterectomy and age ≥55 years. Finally, current estrogen use was ascertained by study staff examination of pill bottles. Sex hormone measurements Ancillary investigations were conducted to ascertain associations between sex hormones and incident diabetes. Eligibility criteria included consent to participate in ancillary studies and plasma availability. Of the 2191 women originally randomly assigned, 2002 women underwent sex hormone measures on prerandomization DPP samples. For the purposes of this report, we excluded women who used estrogen therapy or oral contraceptive pills at baseline, resulting in 1422 participants. Women who were excluded were more likely to have been randomly assigned to the placebo arm and were older, more likely to be postmenopausal, white, thinner, and had lower testosterone, higher sex hormone binding globulin (SHBG), and lower FAI compared with women who were included (Supplemental Table 1). Of the 1422 women included in this report, 143 women did not have testosterone or SHBG measures; these women were included in analyses of IM but not of FAI. Blood samples for sex hormone measurements were collected fasting before 10 am. Sex hormones were measured by Endoceutics (Quebec City, Canada) using liquid chromatography–mass spectrometry. SHBG was measured with an enzyme-linked immunosorbent assay (Bioline, Taunton, MA) with interassay coefficients of variation (CVs) of 7.8% and 5.0% at 18.2 and 63.1 nmol/L, respectively. The lower limits of detection were 20 ng/dL for dehydroepiandrosterone sulfate (DHEAS) and 2, 10, 0.2, and 10 pg/mL for dihydrotestosterone, testosterone, estradiol, and estrone sulfate, respectively. The lower limits of quantification were 100 ng/mL for DHEAS and 10, 50, 1, and 50 pg/mL for dihydrotestosterone, testosterone, estradiol, and estrone sulfate, respectively. Interassay variation (CV) was 10.2%, 10.7%, 7.0%, and 12.5% for DHEAS, testosterone, estradiol, and estrone, respectively, at the lower limits of quantification level. FAI was calculated as 100 × (total testosterone/SHBG). Traditional and nontraditional cardiovascular risk factors Blood pressure and anthropometrics were measured with standardized techniques. Aside from sex hormone measures, other assays were performed at a central biochemistry laboratory (Northwest Lipid Research Laboratories, University of Washington, Seattle). For the purposes of this analysis, we used measures from DPP baseline. Plasma glucose was measured on a chemistry autoanalyzer by the glucokinase method. Insulin measurements were performed by a polyethylene glycol–accelerated double antibody radioimmunoassay method. Insulin sensitivity was assessed by the surrogate measurement 1/fasting insulin (1/FI). Lipoprotein fractions were separated with preparative ultracentrifugation by β-quantification for triglycerides of ≥4.52 mmol/L. For the high-density lipoprotein (HDL) measurement, apolipoprotein B–containing lipoproteins were removed from plasma by precipitation with dextran sulfate, and cholesterol in HDL particles was measured with Roche reagent on a Roche Modular P autoanalyzer (Roche Diagnostics, Indianapolis, IN). Plasma triglycerides were determined enzymatically with methods standardized to the Centers for Disease Control and Prevention Reference Methods. Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald equation. A nonequilibrium density gradient ultracentrifugation method was used to characterize LDL floatation rate. Total circulating adiponectin was measured with a latex particle-enhanced turbidimetric assay (Otsuka Pharmaceutical, Tokyo, Japan). The within-run and between-run CVs for this assay are 6.21% and 9.25%, respectively. Plasma high-sensitivity C-reactive protein (CRP) and fibrinogen were measured immunochemically with Dade-Behring reagent on the Behring Nephelometer autoanalyzer (Dade-Behring, Glasgow, DE), and the two CVs were 2.10% and 3.10% for CRP and 2.70% and 2.60% for fibrinogen, respectively. Tissue plasminogen activator (tPA) levels were measured in citrated plasma with an enzyme-linked immunosorbent assay (Asserachrom tPA; Diagnostica, Stago, France), which measures total tPA antigen, with CVs of 6.45% and 6.70% respectively. DPPOS incident diabetes At the conclusion of the DPP, participants had been followed for an average of 3.2 years. The placebo and metformin groups were unmasked as to their treatment assignment, and all participants were offered the ILS in a group format during a 1-year bridge period (25). The surviving consent members (n = 3149) of the three original treatment arms were invited to participate in an observational follow-up, the DPPOS. Eighty-eight percent (n = 2776) joined. Maintenance group lifestyle sessions were offered quarterly to all DPPOS participants, and metformin continued to be provided to participants originally randomly assigned to metformin who remained eligible. As in the DPP, the development of diabetes was determined with a 75-g oral glucose tolerance test performed annually and fasting plasma glucose tests every 6 months. For diagnosis of diabetes, fasting plasma glucose ≥7.0 mmol/L or 2-hour levels ≥11.1 mmol/L had to be confirmed by a repeat test within 6 weeks; to keep diagnostic criteria similar between the trial and the follow-up, hemoglobin A1c was not used to diagnose diabetes. The DPPOS is ongoing, and participants in this analysis had been followed for ~10 years. DPPOS CAC CAC was measured during year 10 of DPPOS in 2029 participants (74% of the DPPOS cohort) at all sites (26). Participants who weighed >350 lb at the time of scanning were ineligible to participate because of the inability to acquire the relevant images on conventional equipment; 201 refused consent, 57 had a stent, 32 had atrial fibrillation, and 1 was pregnant. At the DPP baseline examination, participants who later underwent CAC studies had slightly higher minority race or ethnicity representation, were slightly younger, and had modestly lower BMI and lower systolic blood pressure, and slightly fewer were smokers compared with those who did not undergo the CAC assessment. Baseline metabolic measures were not different between tested and untested groups, except for slightly higher HDL levels in tested participants. The proportion with CAC measurements did not differ between treatment groups. Sixty-five percent of the 813 women with CAC measures (n = 532) underwent sex hormone measures and were not using exogenous estrogen and thus were included for the subanalysis of CAC. Methods for CAC assessment have been previously reported (26). Briefly, chest computed tomography was performed by certified technologists at each site. Subjects were scanned twice, and measurement of CAC was calibrated against a phantom of known physical calcium concentration. A radiologist or cardiologist read all computed tomography scans at the central reading center (Los Angeles Biomedical Research Institute at Harbor–University of California, Los Angles in Torrance, California) in a manner blinded to patient characteristics and treatment assignment. Discrepancies were reviewed and agreement obtained through consensus. For each scan, a total phantom-adjusted averaged Agatston score was calculated, defined as the sum of calcium measures from the left main, left anterior descending, circumflex, and right coronary arteries. For the purposes of the current report where the burden of CAC was low, we examined the presence and absence of CAC. Statistical analyses Baseline characteristics were described as n (percentages) for categorical variables and mean [standard deviation (SD)] or median (interquartile range) for quantitative variables with normal and skewed distributions, respectively. Differences between women in the upper 10th percentile and lower 90th percentile of FAI (Table 1) and with and without histories of IM (Table 2) were tested via the χ2 test for independence for categorical variables and the t test or the nonparametric Wilcoxon test, as appropriate, for continuous variables. There is no agreement on what FAI cutoff point is consistent with hyperandrogenism, because of variability in testosterone levels and poor standardization of assays. Consequently, the upper quartile or upper 5th or 10th percentile of FAI is commonly and arbitrarily used to define androgen elevation within a particular study population. We examined whether another FAI cutoff point existed by using k-means clustering (27) and also whether there were cutoff points in FAI at which the hazard of diabetes changed markedly by using classification and regression trees (28); neither method revealed another cutoff point, so the 10th percentile was used. Analyses were stratified by menopausal status (Tables 1 and 2 ) or adjusted for menopausal status (Tables 3 and 4 ). To examine the relationship between the variables of FAI and IM with incident diabetes during the DPP and the DPPOS, we constructed Cox proportional hazards models that adjusted for baseline age, race or ethnicity, randomization arm, BMI, and A1c (Table 3). Analyses of incident diabetes throughout the DPPOS included incident diabetes in the DPP. Additional evaluations for interactions with randomization arm when randomization arm was a covariate were performed by adding interaction terms between randomization arm with IM and randomization arm with FAI; interaction terms were not significant, and thus models combined data across randomization arms. Similarly, evaluations for interactions with menopause in models examining diabetes risk were performed by adding interaction terms between FAI and menopause and IM and menopause; neither of these was significant, and thus models combined data across menopausal status. To examine the relationship between FAI and IM with the odds of any CAC at year 10 of the DPPOS, we constructed logistic regression models that adjusted for baseline age, race or ethnicity, randomization arm, BMI, and A1c (Table 4). Table 1. Baseline Participant Characteristics by Menopausal Status and FAI   Premenopausal Women   Postmenopausal Women     FAI, Lower 90th Percentile (n = 761)  FAI, Upper 10th Percentile (n = 85)  P  FAI, Lower 90th Percentile (n = 389)  FAI, Upper 10th Percentile (n = 44)  P  Treatment group, n (%)      0.087      0.16   Placebo  282 (37.1%)  30 (35.3%)    151 (38.8%)  14 (31.8%)     Metformin  248 (32.6%)  20 (23.5%)    106 (27.2%)  18 (40.9%)     ILS  231 (30.4%)  35 (41.2%)    132 (33.9%)  12 (27.3%)    Age group, n (%)      <0.001      0.020   25 to 44 y  405 (53.2%)  73 (85.9%)    33 (8.5%)  9 (20.5%)     45 to 59 y  356 (46.8%)  12 (14.1%)    231 (59.4%)  19 (43.2%)     ≥60 y        125 (32.1%)  16 (36.4%)    Race or ethnicity, n (%)      0.079      0.29   Non-Hispanic white  364 (47.8%)  50 (58.8%)    199 (51.2%)  22 (50%)     African American  178 (23.4%)  22 (25.9%)    125 (32.1%)  14 (31.8%)     Hispanic  157 (20.6%)  7 (8.2%)    52 (13.4%)  4 (9.1%)     Asian  25 (3.3%)  2 (2.4%)    7 (1.8%)  3 (6.8%)     American Indian  37 (4.9%)  4 (4.7%)    6 (1.5%)  1 (2.3%)    Number of pregnancies  3 [2, 4]  2 [2, 3]  0.082  3 [2, 4]  3 [2, 4]  0.92   No pregnancies  113 (14.8%)  22 (25.9%)  0.013  46 (11.8%)  7 (15.9%)  0.59  IM, n (%)  126 (16.6%)  29 (34.1%)  <0.001  64 (16.5%)  11 (25%)  0.23  Past estrogen use (birth control pills or estrogen therapy)  582 (76.5%)  65 (76.5%)  1.000  290 (74.6%)  26 (59.1%)  0.044  BMI, kg/m2  35.9 ± 7.1  37.8 ± 6.7  0.021  35.1 ± 6.7  36.8 ± 7.7  0.11  Waist circumference, cm  104.9 ± 15.6  108.7 ± 15.2  0.032  104.2 ± 14.2  108.8 ± 14.8  0.045  Fasting glucose, mg/dL  106.5 ± 8.3  107.1 ± 8.2  0.53  106.9 ± 7.9  105.7 ± 6.4  0.33  2-h glucose, mg/dL  163.8 ± 17  167.8 ± 16.4  0.038  163.5 ± 16.7  163.4 ± 17.4  0.98  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.6  0.94  6 ± 0.5  5.9 ± 0.4  0.23  DHEAS, μmol/L  2.19 [1.44, 3.22]  2.95 [1.9, 4.21]  <0.001  1.71 [1.07, 2.56]  1.62 [1.16, 2.82]  0.88  Testosterone, nmol/L  0.58 [0.45, 0.79]  1.05 [0.89, 1.39]  <0.001  0.52 [0.41, 0.66]  1.3 [0.82, 1.94]  <0.001  FAI, mean ± SD  1.5 ± 0.7  4.8 ± 1.8  <0.001  1.6 ± 0.8  9.2 ± 14.1  <0.001  Estradiol, pmol/L  210.6 [91.6, 384]  305 [139.4, 443.5]  0.006  37.3 [24.6, 74.7]  46.3 [31.8, 66.8]  0.18  Estrone sulfate, pmol/L  3281.7 [1710.2, 6223.1]  4490.1 [3193, 7224.6]  <0.001  1225.3 [619.7, 2226.3]  1621.6 [1078.5, 2780.3]  0.031  SHBG, (nmol/L)  43 [31.7, 62.5]  25.3 [19.3, 31.3]  <0.001  36.5 [28.4, 48.2]  24.5 [15, 34.2]  <0.001  1/FI  0.047 ± 0.033  0.038 ± 0.024  0.015  0.052 ± 0.035  0.047 ± 0.039  0.31  SBP, mm Hg  120.4 ± 14.2  117.4 ± 11.8  0.061  126.8 ± 16  123.4 ± 14.1  0.19  DBP, mm Hg  77.7 ± 9.3  79.9 ± 8.9  0.034  78.2 ± 9  77.2 ± 10.1  0.50  HDL, mg/dL  44.8 ± 9.8  41.6 ± 9.3  0.004  48.2 ± 11.5  46.2 ± 10.7  0.26  LDL, mg/dL  121.6 ± 31.7  121.3 ± 26.7  0.93  133.2 ± 35.4  132.2 ± 28.4  0.86  Triglycerides, mg/dL  148.1 ± 85.4  159.5 ± 88.5  0.24  149.8 ± 86.2  152.5 ± 72.8  0.84  LDL particle density, 1/rf  3.8 [3.46, 4]  3.8 [3.46, 4.22]  0.32  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.84  Adiponectin, ng/mL  6.9 [5.3, 8.7]  6 [4.9, 7.3]  <0.001  8.4 [6.5, 11.5]  8.2 [6.3, 9.8]  0.41  High-sensitivity CRP, mg/L  0.7 ± 0.82  0.62 ± 0.55  0.39  0.65 ± 0.74  0.68 ± 0.57  0.78  Fibrinogen, μmol/L  11.2 [9.8, 13.1]  11.4 [10, 12.9]  0.85  12.1 [10.7, 13.8]  12.4 [11.2, 13.6]  0.66  tPA, ng/mL  10.9 [8.8, 13.2]  10.9 [9, 13.1]  0.56  11.1 [9.5, 13.2]  11.3 [9.7, 13.8]  0.49  Any CAC  126 (39.6%)  12 (44.4%)  0.78  72 (59.5%)  8 (66.7%)  0.86    Premenopausal Women   Postmenopausal Women     FAI, Lower 90th Percentile (n = 761)  FAI, Upper 10th Percentile (n = 85)  P  FAI, Lower 90th Percentile (n = 389)  FAI, Upper 10th Percentile (n = 44)  P  Treatment group, n (%)      0.087      0.16   Placebo  282 (37.1%)  30 (35.3%)    151 (38.8%)  14 (31.8%)     Metformin  248 (32.6%)  20 (23.5%)    106 (27.2%)  18 (40.9%)     ILS  231 (30.4%)  35 (41.2%)    132 (33.9%)  12 (27.3%)    Age group, n (%)      <0.001      0.020   25 to 44 y  405 (53.2%)  73 (85.9%)    33 (8.5%)  9 (20.5%)     45 to 59 y  356 (46.8%)  12 (14.1%)    231 (59.4%)  19 (43.2%)     ≥60 y        125 (32.1%)  16 (36.4%)    Race or ethnicity, n (%)      0.079      0.29   Non-Hispanic white  364 (47.8%)  50 (58.8%)    199 (51.2%)  22 (50%)     African American  178 (23.4%)  22 (25.9%)    125 (32.1%)  14 (31.8%)     Hispanic  157 (20.6%)  7 (8.2%)    52 (13.4%)  4 (9.1%)     Asian  25 (3.3%)  2 (2.4%)    7 (1.8%)  3 (6.8%)     American Indian  37 (4.9%)  4 (4.7%)    6 (1.5%)  1 (2.3%)    Number of pregnancies  3 [2, 4]  2 [2, 3]  0.082  3 [2, 4]  3 [2, 4]  0.92   No pregnancies  113 (14.8%)  22 (25.9%)  0.013  46 (11.8%)  7 (15.9%)  0.59  IM, n (%)  126 (16.6%)  29 (34.1%)  <0.001  64 (16.5%)  11 (25%)  0.23  Past estrogen use (birth control pills or estrogen therapy)  582 (76.5%)  65 (76.5%)  1.000  290 (74.6%)  26 (59.1%)  0.044  BMI, kg/m2  35.9 ± 7.1  37.8 ± 6.7  0.021  35.1 ± 6.7  36.8 ± 7.7  0.11  Waist circumference, cm  104.9 ± 15.6  108.7 ± 15.2  0.032  104.2 ± 14.2  108.8 ± 14.8  0.045  Fasting glucose, mg/dL  106.5 ± 8.3  107.1 ± 8.2  0.53  106.9 ± 7.9  105.7 ± 6.4  0.33  2-h glucose, mg/dL  163.8 ± 17  167.8 ± 16.4  0.038  163.5 ± 16.7  163.4 ± 17.4  0.98  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.6  0.94  6 ± 0.5  5.9 ± 0.4  0.23  DHEAS, μmol/L  2.19 [1.44, 3.22]  2.95 [1.9, 4.21]  <0.001  1.71 [1.07, 2.56]  1.62 [1.16, 2.82]  0.88  Testosterone, nmol/L  0.58 [0.45, 0.79]  1.05 [0.89, 1.39]  <0.001  0.52 [0.41, 0.66]  1.3 [0.82, 1.94]  <0.001  FAI, mean ± SD  1.5 ± 0.7  4.8 ± 1.8  <0.001  1.6 ± 0.8  9.2 ± 14.1  <0.001  Estradiol, pmol/L  210.6 [91.6, 384]  305 [139.4, 443.5]  0.006  37.3 [24.6, 74.7]  46.3 [31.8, 66.8]  0.18  Estrone sulfate, pmol/L  3281.7 [1710.2, 6223.1]  4490.1 [3193, 7224.6]  <0.001  1225.3 [619.7, 2226.3]  1621.6 [1078.5, 2780.3]  0.031  SHBG, (nmol/L)  43 [31.7, 62.5]  25.3 [19.3, 31.3]  <0.001  36.5 [28.4, 48.2]  24.5 [15, 34.2]  <0.001  1/FI  0.047 ± 0.033  0.038 ± 0.024  0.015  0.052 ± 0.035  0.047 ± 0.039  0.31  SBP, mm Hg  120.4 ± 14.2  117.4 ± 11.8  0.061  126.8 ± 16  123.4 ± 14.1  0.19  DBP, mm Hg  77.7 ± 9.3  79.9 ± 8.9  0.034  78.2 ± 9  77.2 ± 10.1  0.50  HDL, mg/dL  44.8 ± 9.8  41.6 ± 9.3  0.004  48.2 ± 11.5  46.2 ± 10.7  0.26  LDL, mg/dL  121.6 ± 31.7  121.3 ± 26.7  0.93  133.2 ± 35.4  132.2 ± 28.4  0.86  Triglycerides, mg/dL  148.1 ± 85.4  159.5 ± 88.5  0.24  149.8 ± 86.2  152.5 ± 72.8  0.84  LDL particle density, 1/rf  3.8 [3.46, 4]  3.8 [3.46, 4.22]  0.32  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.84  Adiponectin, ng/mL  6.9 [5.3, 8.7]  6 [4.9, 7.3]  <0.001  8.4 [6.5, 11.5]  8.2 [6.3, 9.8]  0.41  High-sensitivity CRP, mg/L  0.7 ± 0.82  0.62 ± 0.55  0.39  0.65 ± 0.74  0.68 ± 0.57  0.78  Fibrinogen, μmol/L  11.2 [9.8, 13.1]  11.4 [10, 12.9]  0.85  12.1 [10.7, 13.8]  12.4 [11.2, 13.6]  0.66  tPA, ng/mL  10.9 [8.8, 13.2]  10.9 [9, 13.1]  0.56  11.1 [9.5, 13.2]  11.3 [9.7, 13.8]  0.49  Any CAC  126 (39.6%)  12 (44.4%)  0.78  72 (59.5%)  8 (66.7%)  0.86  Mean ± SD or n (%) shown; sex hormone median [interquartile range] shown. Participants who did not have testosterone or SHBG levels tested were excluded. Abbreviations: rf, relative flotation; SBP, systolic blood pressure. View Large Table 2. Baseline Participant Characteristics by Menopausal Status and History of IM   Premenopausal Women   Postmenopausal Women     Regular Menses (n = 701)  IM (n = 161)  P  Regular Menses (n = 472)  IM (n = 88)  P  Treatment group, n (%)      1.0      0.87   Placebo  256 (36.5%)  59 (36.6%)    174 (36.9%)  30 (34.1%)     Metformin  222 (31.7%)  51 (31.7%)    143 (30.3%)  27 (30.7%)     ILS  223 (31.8%)  51 (31.7%)    155 (32.8%)  31 (35.2%)    Age group, n (%)      0.14      <0.001   25 to 44 y  384 (54.8%)  99 (61.5%)    31 (6.6%)  18 (20.5%)     45 to 59 y  317 (45.2%)  62 (38.5%)    256 (54.2%)  51 (58%)     ≥60 y        185 (39.2%)  19 (21.6%)    Race or ethnicity, n (%)      0.004      0.012   Non-Hispanic white  344 (49.1%)  79 (49.1%)    237 (50.2%)  41 (46.6%)     African American  174 (24.8%)  30 (18.6%)    151 (32%)  20 (22.7%)     Hispanic  134 (19.1%)  31 (19.3%)    70 (14.8%)  20 (22.7%)     Asian  16 (2.3%)  13 (8.1%)    10 (2.1%)  3 (3.4%)     American Indian  33 (4.7%)  8 (5%)    4 (0.8%)  4 (4.5%)    Number of pregnancies  3 [2, 4]  3 [2, 4]  0.47  3 [2, 4]  3 [2, 4]  0.56   No pregnancies  108 (15.4%)  28 (17.4%)  0.62  62 (13.1%)  12 (13.6%)  1.0  Past estrogen use (birth control pills or estrogen therapy)  539 (76.9%)  123 (76.4%)  0.98  337 (71.4%)  62 (70.5%)  0.96  BMI, kg/m2  35.8 ± 7  37.1 ± 7.6  0.048  34.4 ± 6.3  37.3 ± 8.1  <0.001  Waist circumference, cm  104.7 ± 15.3  107.6 ± 16.8  0.037  103.3 ± 13.6  108.7 ± 16.4  0.001  Fasting glucose, mg/dL  106.7 ± 8.4  105.8 ± 7.6  0.24  106.9 ± 7.9  106.7 ± 7.1  0.81  2-h glucose, mg/dL  163.9 ± 17  165.9 ± 16.7  0.17  164.2 ± 17  163 ± 16.9  0.54  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.5  0.58  6 ± 0.5  6 ± 0.5  0.82  DHEAS, μmol/L  2.25 [1.46, 3.28]  2.43 [1.39, 3.42]  0.81  1.58 [0.98, 2.43]  1.9 [1.04, 2.92]  0.15  Testosterone, nmol/L  0.6 [0.45, 0.84]  0.68 [0.5, 0.98]  0.007  0.51 [0.35, 0.72]  0.52 [0.35, 0.74]  0.712  Estradiol, pmol/L  226.4 [102.7, 397.8]  190.5 [90.4, 348]  0.17  33.8 [21.8, 56.4]  45.4 [23.2, 112.4]  0.059  Estrone sulfate, pmol/L  3473.4 [1780.7, 6279.8]  3547.5 [1742.1, 6038.3]  1.0  1113.8 [556.2, 2040]  1453.7 [828.4, 2435.4]  0.02  SHBG, nmol/L  40.9 [29.9, 59.9]  37.9 [26.7, 58.6]  0.17  35.4 [26, 46.3]  33.4 [23.2, 47.2]  0.30  FAI, mean ± SD  1.7 ± 1.2  2.2 ± 1.7  <0.001  2.4 ± 5.5  2.2 ± 1.8  0.74  Elevated FAI, n (%)  56 (8.1%)  29 (18.7%)  <0.001  33 (9.2%)  11 (14.7%)  0.23  1/FI  0.047 ± 0.033  0.042 ± 0.028  0.081  0.052 ± 0.033  0.045 ± 0.033  0.062  SBP, mm Hg  119.6 ± 13.8  121.9 ± 14.5  0.061  127.7 ± 16  123.8 ± 14.5  0.035  DBP, mm Hg  77.5 ± 9.1  80 ± 10  0.002  77.9 ± 9.2  77.8 ± 8.4  0.90  HDL, mg/dL  44.7 ± 9.7  43.6 ± 9.8  0.19  48.3 ± 11.5  44.7 ± 9.7  0.005  LDL, mg/dL  121.9 ± 30.9  119.3 ± 32.3  0.35  135 ± 35  131.6 ± 35.1  0.41  Triglycerides, mg/dL  147.2 ± 83.4  157.1 ± 94.5  0.19  150.3 ± 82.1  163.6 ± 101.6  0.18  LDL particle density, 1/rf  3.8 [3.46, 4.22]  3.8 [3.46, 4.05]  0.028  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.31  Adiponectin, ng/mL  6.9 [5.4, 8.7]  6.1 [4.9, 7.6]  <0.001  8.5 [6.7, 11.7]  7.8 [6.2, 9.9]  0.007  High-sensitivity CRP, mg/L  0.67 ± 0.79  0.75 ± 0.8  0.24  0.63 ± 0.72  0.67 ± 0.63  0.58  Fibrinogen, μmol/L  11.3 [9.8, 13.1]  11.2 [9.7, 12.9]  0.76  12.1 [10.7, 13.6]  12.2 [10.9, 13.5]  0.83  tPA, ng/mL  10.9 [8.8, 13]  11.1 [9.2, 13.6]  0.20  11.2 [9.4, 13.6]  11.1 [9.8, 13.1]  0.91  Any CAC  121 (40.9%)  20 (36.4%)  0.63  100 (65.4%)  17 (60.7%)  0.80    Premenopausal Women   Postmenopausal Women     Regular Menses (n = 701)  IM (n = 161)  P  Regular Menses (n = 472)  IM (n = 88)  P  Treatment group, n (%)      1.0      0.87   Placebo  256 (36.5%)  59 (36.6%)    174 (36.9%)  30 (34.1%)     Metformin  222 (31.7%)  51 (31.7%)    143 (30.3%)  27 (30.7%)     ILS  223 (31.8%)  51 (31.7%)    155 (32.8%)  31 (35.2%)    Age group, n (%)      0.14      <0.001   25 to 44 y  384 (54.8%)  99 (61.5%)    31 (6.6%)  18 (20.5%)     45 to 59 y  317 (45.2%)  62 (38.5%)    256 (54.2%)  51 (58%)     ≥60 y        185 (39.2%)  19 (21.6%)    Race or ethnicity, n (%)      0.004      0.012   Non-Hispanic white  344 (49.1%)  79 (49.1%)    237 (50.2%)  41 (46.6%)     African American  174 (24.8%)  30 (18.6%)    151 (32%)  20 (22.7%)     Hispanic  134 (19.1%)  31 (19.3%)    70 (14.8%)  20 (22.7%)     Asian  16 (2.3%)  13 (8.1%)    10 (2.1%)  3 (3.4%)     American Indian  33 (4.7%)  8 (5%)    4 (0.8%)  4 (4.5%)    Number of pregnancies  3 [2, 4]  3 [2, 4]  0.47  3 [2, 4]  3 [2, 4]  0.56   No pregnancies  108 (15.4%)  28 (17.4%)  0.62  62 (13.1%)  12 (13.6%)  1.0  Past estrogen use (birth control pills or estrogen therapy)  539 (76.9%)  123 (76.4%)  0.98  337 (71.4%)  62 (70.5%)  0.96  BMI, kg/m2  35.8 ± 7  37.1 ± 7.6  0.048  34.4 ± 6.3  37.3 ± 8.1  <0.001  Waist circumference, cm  104.7 ± 15.3  107.6 ± 16.8  0.037  103.3 ± 13.6  108.7 ± 16.4  0.001  Fasting glucose, mg/dL  106.7 ± 8.4  105.8 ± 7.6  0.24  106.9 ± 7.9  106.7 ± 7.1  0.81  2-h glucose, mg/dL  163.9 ± 17  165.9 ± 16.7  0.17  164.2 ± 17  163 ± 16.9  0.54  Hemoglobin A1c, %  5.9 ± 0.5  5.9 ± 0.5  0.58  6 ± 0.5  6 ± 0.5  0.82  DHEAS, μmol/L  2.25 [1.46, 3.28]  2.43 [1.39, 3.42]  0.81  1.58 [0.98, 2.43]  1.9 [1.04, 2.92]  0.15  Testosterone, nmol/L  0.6 [0.45, 0.84]  0.68 [0.5, 0.98]  0.007  0.51 [0.35, 0.72]  0.52 [0.35, 0.74]  0.712  Estradiol, pmol/L  226.4 [102.7, 397.8]  190.5 [90.4, 348]  0.17  33.8 [21.8, 56.4]  45.4 [23.2, 112.4]  0.059  Estrone sulfate, pmol/L  3473.4 [1780.7, 6279.8]  3547.5 [1742.1, 6038.3]  1.0  1113.8 [556.2, 2040]  1453.7 [828.4, 2435.4]  0.02  SHBG, nmol/L  40.9 [29.9, 59.9]  37.9 [26.7, 58.6]  0.17  35.4 [26, 46.3]  33.4 [23.2, 47.2]  0.30  FAI, mean ± SD  1.7 ± 1.2  2.2 ± 1.7  <0.001  2.4 ± 5.5  2.2 ± 1.8  0.74  Elevated FAI, n (%)  56 (8.1%)  29 (18.7%)  <0.001  33 (9.2%)  11 (14.7%)  0.23  1/FI  0.047 ± 0.033  0.042 ± 0.028  0.081  0.052 ± 0.033  0.045 ± 0.033  0.062  SBP, mm Hg  119.6 ± 13.8  121.9 ± 14.5  0.061  127.7 ± 16  123.8 ± 14.5  0.035  DBP, mm Hg  77.5 ± 9.1  80 ± 10  0.002  77.9 ± 9.2  77.8 ± 8.4  0.90  HDL, mg/dL  44.7 ± 9.7  43.6 ± 9.8  0.19  48.3 ± 11.5  44.7 ± 9.7  0.005  LDL, mg/dL  121.9 ± 30.9  119.3 ± 32.3  0.35  135 ± 35  131.6 ± 35.1  0.41  Triglycerides, mg/dL  147.2 ± 83.4  157.1 ± 94.5  0.19  150.3 ± 82.1  163.6 ± 101.6  0.18  LDL particle density, 1/rf  3.8 [3.46, 4.22]  3.8 [3.46, 4.05]  0.028  3.62 [3.46, 3.8]  3.8 [3.46, 3.8]  0.31  Adiponectin, ng/mL  6.9 [5.4, 8.7]  6.1 [4.9, 7.6]  <0.001  8.5 [6.7, 11.7]  7.8 [6.2, 9.9]  0.007  High-sensitivity CRP, mg/L  0.67 ± 0.79  0.75 ± 0.8  0.24  0.63 ± 0.72  0.67 ± 0.63  0.58  Fibrinogen, μmol/L  11.3 [9.8, 13.1]  11.2 [9.7, 12.9]  0.76  12.1 [10.7, 13.6]  12.2 [10.9, 13.5]  0.83  tPA, ng/mL  10.9 [8.8, 13]  11.1 [9.2, 13.6]  0.20  11.2 [9.4, 13.6]  11.1 [9.8, 13.1]  0.91  Any CAC  121 (40.9%)  20 (36.4%)  0.63  100 (65.4%)  17 (60.7%)  0.80  Mean ± SD or n (%) shown; sex hormone median [interquartile range] shown. Abbreviations: rf, relative flotation; SBP, systolic blood pressure. View Large Table 3. Risk of Incident Diabetes Associated With FAI and IM During the DPP and the DPPOS   Risk During DPP, HR (95% CI)  P  Risk During DPP and DPPOS, HR (95% CI)  P  FAI (examined as a continuous measure)           Model 1: Adjusted for menopausal status  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.32   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.29   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.99 (0.95 to 1.04)  0.79  0.98 (0.94 to 1.02)  0.31   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.99 (0.94 to 1.04)  0.75  0.98 (0.93 to 1.02)  0.28   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.99 (0.93 to 1.05)  0.74  0.97 (0.93 to 1.02)  0.27  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.98 (0.73 to 1.31)  0.88  1.12 (0.92 to 1.37)  0.27   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.01 (0.75 to 1.35)  0.97  1.12 (0.92 to 1.37)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.01 (0.75 to 1.35)  0.95  1.12 (0.92 to 1.37)  0.27   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.95 (0.71 to 1.28)  0.76  1.07 (0.87 to 1.31)  0.54   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.97 (0.72 to 1.30)  0.82  1.07 (0.87 to 1.31)  0.54    Risk During DPP, HR (95% CI)  P  Risk During DPP and DPPOS, HR (95% CI)  P  FAI (examined as a continuous measure)           Model 1: Adjusted for menopausal status  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.32   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.99 (0.95 to 1.04)  0.71  0.98 (0.94 to 1.02)  0.29   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.99 (0.95 to 1.04)  0.79  0.98 (0.94 to 1.02)  0.31   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.99 (0.94 to 1.04)  0.75  0.98 (0.93 to 1.02)  0.28   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.99 (0.93 to 1.05)  0.74  0.97 (0.93 to 1.02)  0.27  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.98 (0.73 to 1.31)  0.88  1.12 (0.92 to 1.37)  0.27   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.01 (0.75 to 1.35)  0.97  1.12 (0.92 to 1.37)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.01 (0.75 to 1.35)  0.95  1.12 (0.92 to 1.37)  0.27   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.95 (0.71 to 1.28)  0.76  1.07 (0.87 to 1.31)  0.54   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.97 (0.72 to 1.30)  0.82  1.07 (0.87 to 1.31)  0.54  FAI is examined as a continuous measure, and HR >1 indicates risk associated with each unit increase in FAI. For analyses of IM, regular menses is the reference, and HR >1 indicates greater risk associated with IM. Abbreviations: CI, confidence interval; HR, hazard ratio. View Large Table 4. Associations Between FAI and IM With Odds of Any CAC at Year 10 of the DPPOS   OR (95% CI)  P  FAI (examined as a continuous measure)       Model 1: Adjusted for menopausal status  1.05 (0.92 to 1.21)  0.069   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.09 (0.95 to 1.27)  0.074   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.09 (0.95 to 1.27)  0.074   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  1.07 (0.92 to 1.24)  0.075   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  1.06 (0.92 to 1.23)  0.075  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.82 (0.51 to 1.33)  0.25   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.92 (0.56 to 1.53)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.92 (0.55 to 1.52)  0.26   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.87 (0.52 to 1.46)  0.26   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.89 (0.53 to 1.49)  0.26    OR (95% CI)  P  FAI (examined as a continuous measure)       Model 1: Adjusted for menopausal status  1.05 (0.92 to 1.21)  0.069   Model 2: Adjusted for menopausal status, age, and race or ethnicity  1.09 (0.95 to 1.27)  0.074   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  1.09 (0.95 to 1.27)  0.074   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  1.07 (0.92 to 1.24)  0.075   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  1.06 (0.92 to 1.23)  0.075  IM (regular menses = reference)   Model 1: Adjusted for menopausal status  0.82 (0.51 to 1.33)  0.25   Model 2: Adjusted for menopausal status, age, and race or ethnicity  0.92 (0.56 to 1.53)  0.26   Model 3: Adjusted for menopausal status, age, race or ethnicity, and randomization arm  0.92 (0.55 to 1.52)  0.26   Model 4: Adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI  0.87 (0.52 to 1.46)  0.26   Model 5: Adjusted for menopausal status, age, race or ethnicity, randomization, BMI, and hemoglobin A1c at baseline  0.89 (0.53 to 1.49)  0.26  For analyses of FAI, FAI is examined as a continuous measure, and OR >1 indicates odds associated with each unit increase in FAI. For analyses of IM, regular menses is the reference, and OR >1 indicates greater odds associated with IM. Abbreviations: CI, confidence interval; OR, odds ratio. View Large In sensitivity analyses, we attempted to characterize the risk of incident diabetes (Fig. 1) and odds of prevalent CAC (Fig. 2) among women who had both relative elevations in FAI and IM. As with diabetes risk, interaction terms for randomization arm and for menopause with FAI and IM were not significant, and thus models combined data across randomization arms as well as menopausal status; analyses presenting stratified results are presented in Supplemental Tables 2 and 3. Analyses were performed in R software version 3.4.0 (R Core Team, Vienna, Austria), and all tests were two sided, with statistical significance set at P < 0.05. No adjustments for multiple comparisons were performed, and all P values are nominal. Figure 1. View largeDownload slide Risk of incident diabetes associated with relative elevations in FAI and IM during the DPP (top panel) and the DPPOS (bottom panel) [hazard ratio (HR) and 95% confidence interval (CI)]. Reference group is women without either abnormality. Women without either abnormality are the reference group, so that an HR >1 indicates greater diabetes risk if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 1. View largeDownload slide Risk of incident diabetes associated with relative elevations in FAI and IM during the DPP (top panel) and the DPPOS (bottom panel) [hazard ratio (HR) and 95% confidence interval (CI)]. Reference group is women without either abnormality. Women without either abnormality are the reference group, so that an HR >1 indicates greater diabetes risk if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 2. View largeDownload slide Associations between relative elevations in FAI and IM with odds of any CAC at year 10 DPPOS [odds ratio (OR) and 95% confidence interval (CI)]. The reference group is women without either abnormality, so that an OR >1 indicates greater odds of CAC if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Figure 2. View largeDownload slide Associations between relative elevations in FAI and IM with odds of any CAC at year 10 DPPOS [odds ratio (OR) and 95% confidence interval (CI)]. The reference group is women without either abnormality, so that an OR >1 indicates greater odds of CAC if women have elevations in both FAI and IM. Model 1 is adjusted for menopausal status. Model 2 is adjusted for menopausal status, age, and race or ethnicity. Model 3 is adjusted for menopausal status, age, race or ethnicity, and randomization arm. Model 4 is adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI. Model 5 is adjusted for menopausal status, age, race or ethnicity, randomization arm, BMI, and baseline hemoglobin A1c. Results Table 1 shows participant characteristics at baseline by category of FAI elevation and menopausal status. Women had an average age of 48.2 ± 9.9 years; premenopausal women had an average age of 42.9 ± 5.8 years, and postmenopausal women had an average age of 56.4 ± 9.4. Among women who were premenopausal at baseline, women in the upper 10th percentile of FAI were slightly younger. Premenopausal women with higher FAI also had greater BMI, waist circumference, postchallenge glucose, DHEAS, estradiol, and estrone sulfate concentrations. Premenopausal women with higher FAI also had higher diastolic blood pressure (DBP) and lower insulin sensitivity, HDL, and adiponectin. In these unadjusted comparisons, presence of CAC was similar by FAI percentile. Among women who were postmenopausal at baseline, women with higher FAI were slightly younger and had higher waist circumference and estrone sulfate concentrations, although glucose levels and sex hormone concentrations (other than testosterone and SHBG) were similar between groups. Among postmenopausal women, no significant differences in other markers of risk were observed by FAI status. Table 2 shows participant characteristics at baseline by history of IM and menopausal status. Among women who were premenopausal at baseline, women with a history of IM were slightly less likely to be African American. Premenopausal women with histories of IM had higher BMIs and waist circumferences than women without such histories. Sex hormone profiles were similar by IM, except that women with IM had higher FAI and testosterone concentrations than women without such histories. Premenopausal women with histories of IM had lower adiponectin and higher DBP. In these unadjusted comparisons, presence of CAC was similar by FAI status. Among women who were postmenopausal at baseline, women with histories of IM were younger and more likely to be Hispanic and less likely to be African American. Postmenopausal women with histories of IM had higher BMI, waist circumference, and estradiol and estrone-sulfate levels and lower insulin sensitivity, systolic blood pressure, HDL, and adiponectin levels than women without such histories. Table 3 shows the risk of incident diabetes during the DPP and DPPOS associated with FAI and IM. None of these exposures were associated with increased risk of progression to diabetes during the DPP or the DPPOS, before and after adjustment for covariates. Table 4 shows the odds of any CAC at year 10 of the DPPOS associated with FAI and IM. None of these exposures were associated with increased odds of any CAC, before and after adjustment for covariates. We conducted several sensitivity analyses. We determined whether the subset of women with both FAI elevations and IM, two central components of PCOS, had greater risk of diabetes than women without either FAI elevations or IM: 1000 women were included in this analysis, 40 of whom had both conditions and 960 of whom had neither IM nor elevated FAI. In multivariable models that adjusted for menopausal status, age, race or ethnicity, randomization arm, and BMI, women with both conditions had a similar risk of progression to diabetes during the DPP and the DPPOS as women without either condition (Fig. 1). We also evaluated whether the subset of women with both FAI elevations and IM had higher odds of any CAC compared with women with lower FAI and regular menses. In multivariable models, these women had a similar pattern of risk compared with women with either FAI elevations or IM, that is, there was no association (Fig. 2). Finally, we performed stratified analyses by menopausal status, but we found similar patterns to when menopause was used as an adjuster (Supplemental Tables 2 and 3). Conclusions PCOS is a well-recognized risk factor for increased risk of diabetes and for subclinical and clinical cardiovascular disease. Studies have differed regarding whether the elevated androgens or IM that characterize PCOS also confer increased risk in women without PCOS. Specifically, it is unclear whether these conditions confer additional risk in women who are glucose intolerant and overweight, the primary risk factors for diabetes. Using data from the DPP and DPPOS, a large multiethnic cohort of women at high risk for diabetes, we found that neither higher androgen levels nor IM conferred additional risk for diabetes or CAC burden. Although our results do not rule out the possibility that relative androgen elevations or IM increased risk of glucose intolerance or overweight, our findings suggest that the presence of these disorders does not modify progression from impaired glucose tolerance to diabetes nor increase risk of atherosclerosis, as reflected by CAC. Whether the androgen elevation that is typical of PCOS can also be associated with increased risk of cardiometabolic disease for women without PCOS has been controversial. The majority of previous studies have examined women with a range of BMIs and degrees of glucose tolerance. A 2006 meta-analysis found cross-sectional associations, but longitudinal associations had borderline significance (29). Our findings are similar to more recent longitudinal studies (11, 12) that suggest that relative elevations in androgens do not confer risk beyond associations with initial elevations in BMI and glucose intolerance. Even though women with elevations in FAI also had additional diabetes risk factors such as lower adiponectin levels (30), it did not increase their overall risk of diabetes compared with women with lower FAI who were also obese and glucose-intolerant. Several previous reports have suggested that elevations in total testosterone are associated with greater carotid intima media thickness (15–18, 31), although associations between androgens and presence of CAC have not been significant (13, 14, 16). We found that even though relative elevations in androgens at DPP baseline were associated with poorer metabolic indices, these did not translate to greater CAC burden a decade later. It is possible that androgens have different effects on different vascular beds or that the relative burden of subclinical atherosclerosis was low and burden of CAC not particularly sensitive to androgen associations. With longer follow-up, it is possible that the association between androgen levels and CAC would be stronger. However, the lack of association with CAC is reassuring in that the absence of CAC tends to be more strongly associated with incident cardiovascular disease than other subclinical markers (32, 33). Previous reports have also been inconsistent regarding associations between IM and cardiometabolic risk among women without known PCOS. Studies have reported significant associations with incident diabetes (19, 20) or no associations (11, 23). Definitions of IM differed across studies and may have contributed to the disparate results, along with the examination of participants explicitly diagnosed with PCOS as well as those without diagnosed PCOS. In the latter population, IM may be more likely to be multifactorial. To our knowledge, only one longitudinal study has examined whether IM are associated with subclinical atherosclerosis and, similarly to our study, noted no association (11). Two studies have examined the relationship between IM and coronary disease outcomes; one noted no association with coronary heart disease mortality after adjustment for BMI (22), and the other found a significant association between IM and coronary disease (21). It is possible that these reports differed because of a stricter definition of IM in the latter report, which required ≥40 days between cycles or irregularity that did not allow estimation of cycle length (21), thus capturing women with more severe degrees of hormonal abnormalities. It is possible that the relationship between elevated androgen levels and future risk of diabetes and cardiovascular disease is mediated through obesity and associated abnormalities. Using data from the Study of Women’s Health Across the Nation, El Khoudary et al. (34) noted that a higher FAI was associated with CAC in nonobese women but not in obese women. Among premenopausal women, higher FAI was associated with higher BMI and waist circumference as well as higher postchallenge glucose, higher DBP, and lower insulin sensitivity, HDL, and adiponectin. It is also possible that PCOS increases cardiometabolic risk through pathways other than elevated androgen levels and those involving the menstrual cycle. Thus, these conditions may serve as a proxy for other cardiometabolic risk factors in the PCOS population (35); other reports have noted that male relatives of women with known PCOS have elevated cardiometabolic risk, suggesting shared genetic risk factors (36). In support of this hypothesis, one report found that women with PCOS, defined as a combination of hyperandrogenism and IM, had increased odds of CAC, but neither isolated hyperandrogenism nor IM alone predicted burden of CAC (13). We did not find that the combination of these conditions increased risk, although the number of women with both conditions in our sample was small. Currently, no single definition of biochemical hyperandrogenism in women exists because of variation in the assays used across studies and the decline in average androgen levels with age. Longitudinal studies of cohorts with PCOS have noted that androgens decreased over the reproductive lifespan and may be unremarkable in the fourth decade of life (37, 38) even as their waist circumference and degree of insulin resistance remain elevated compared with women without PCOS. We examined a cohort of women whose average age was ~50 years. Even though we examined relative elevations in androgen levels within the cohort, we may not have captured women with histories of elevated androgen levels in their earlier reproductive years. This possibility is supported by the fact that relative elevations in androgens were associated with a higher number of differences in cardiometabolic markers in premenopausal compared with postmenopausal women. Along similar lines, ~44% of women with PCOS eventually have regular menses before menopause (37). We note that the DPP requirements for overweight and glucose-intolerant women may have captured women who developed these conditions subsequent to histories of PCOS, and we were unable to determine the chronological order of their development of IM and hyperandrogenism in relation to their enrollment criteria. The strengths of this report include its use of a comparison group that was also overweight and glucose intolerant, as well as longitudinal assessment of diabetes and examination of subclinical markers of atherosclerosis and an extensive list of biomarkers. We also used sensitive mass spectrometric methods to assess sex steroid measures, which is particularly pertinent for measures of sex steroids in postmenopausal women. However, there are several limitations. As is true for other reports of hyperandrogenism in midlife women, no single cutoff point exists, and it is possible that women were misclassified, although we addressed this issue to some extent by defining androgen elevations within the population. IM was obtained by self-report, and further details of cycle duration and variability were not available. Transvaginal ultrasound was not performed, and data regarding ovarian size and polycystic ovarian morphology were not available. The measurement of other hormones to rule out other endocrine disorders that can cause IM, such as thyroid-stimulating hormone, prolactin, and 17-hydroxyprogesterone, was not performed. Finally, the burden of CAC was low, and thus our power to assess odds of CAC was limited. We conclude that compared with midlife women who are already glucose intolerant and overweight, histories of IM and relative elevations in androgens do not additionally increase risk or the burden of subclinical atherosclerosis. Although our report does not rule out that these conditions may identify women at risk in their younger years, it suggests that chronic disease burden in midlife women may be most effectively addressed through weight loss, metformin, and the other diabetes prevention measures already known to be beneficial for all women at elevated risk for diabetes. Future investigations should examine whether relative elevations in androgens or IM may further define high-risk women at younger ages. Abbreviations: BMI body mass index CAC coronary artery calcification CRP C-reactive protein CV coefficient of variation DBP diastolic blood pressure DHEAS dehydroepiandrosterone sulfate DPP Diabetes Prevention Program DPPOS Diabetes Prevention Program Outcomes Study FAI free androgen index FI fasting insulin HDL high-density lipoprotein ILS intensive lifestyle intervention IM irregular menses LDL low-density lipoprotein PCOS polycystic ovary syndrome SD standard deviation SHBG sex hormone binding globulin tPA tissue plasminogen activator. Acknowledgments The Research Group gratefully acknowledges the commitment and dedication of the participants in the Diabetes Prevention Program. This work was presented in part at the 2016 Scientific Sessions of the American Diabetes Association. Financial Support: This study was supported by Grants DK072041, DK048489, and DK53061 from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health, which provided funding to the clinical centers and the Coordinating Center for the design and conduct of the study and the collection, management, analysis, and interpretation of the data. The Southwestern American Indian Centers were supported directly by NIDDK, including its Intramural Research Program, and the Indian Health Service. The General Clinical Research Center Program, National Center for Research Resources, and the Department of Veterans Affairs supported data collection at many of the clinical centers. Funding was also provided by the National Institute of Child Health and Human Development, National Institute on Aging, National Eye Institute, National Heart, Lung, and Blood Institute, Office of Research on Women’s Health, National Institute on Minority Health and Health Disparities, Centers for Disease Control and Prevention, and American Diabetes Association. Bristol-Myers Squibb and Parke-Davis provided additional funding and material support during DPP, Lipha (Merck-Sante) provided medication, and LifeScan Inc. donated materials during DPP and DPPOS. The opinions expressed herein are those of the investigators and do not necessarily reflect the views of the funding agencies. A complete list of centers, investigators, and staff can be found in the online appendix. The research group gratefully acknowledges the commitment and dedication of the participants of the DPP and DPPOS. Clinical Trial Information: ClinicalTrials.gov no. NCT00038727 (registered 4 June 2002). Author Contributions: C.K. wrote the manuscript and researched the sex hormone data and is the guarantor of the manuscript. N.Y. provided the analysis and reviewed/edited the manuscript. D.A.E., V.R.A., S.L.E., and R.B.G. reviewed and edited the manuscript, and R.B.G. researched the adiponectin and inflammatory marker assays. Disclosure Summary: The authors have nothing to disclose. References 1. Legro RS, Arslanian SA, Ehrmann DA, Hoeger KM, Murad MH, Pasquali R, Welt CK; Endocrine Society. Diagnosis and treatment of polycystic ovary syndrome: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab . 2013; 98( 12): 4565– 4592. Google Scholar CrossRef Search ADS PubMed  2. Song DK, Hong YS, Sung YA, Lee H. Insulin resistance according to β-cell function in women with polycystic ovary syndrome and normal glucose tolerance. PLoS One . 2017; 12( 5): e0178120. Google Scholar CrossRef Search ADS PubMed  3. Ciaraldi TP, Aroda V, Mudaliar S, Chang RJ, Henry RR. Polycystic ovary syndrome is associated with tissue-specific differences in insulin resistance. J Clin Endocrinol Metab . 2009; 94( 1): 157– 163. Google Scholar CrossRef Search ADS PubMed  4. Palomba S, Falbo A, Russo T, Manguso F, Tolino A, Zullo F, De Feo P, Orio F, Jr. Insulin sensitivity after metformin suspension in normal-weight women with polycystic ovary syndrome. J Clin Endocrinol Metab . 2007; 92( 8): 3128– 3135. Google Scholar CrossRef Search ADS PubMed  5. Ehrmann DA, Barnes RB, Rosenfield RL, Cavaghan MK, Imperial J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care . 1999; 22( 1): 141– 146. Google Scholar CrossRef Search ADS PubMed  6. Joham AE, Ranasinha S, Zoungas S, Moran L, Teede HJ. Gestational diabetes and type 2 diabetes in reproductive-aged women with polycystic ovary syndrome. J Clin Endocrinol Metab . 2014; 99( 3): E447– E452. Google Scholar CrossRef Search ADS PubMed  7. Lo JC, Feigenbaum SL, Yang J, Pressman AR, Selby JV, Go AS. Epidemiology and adverse cardiovascular risk profile of diagnosed polycystic ovary syndrome. J Clin Endocrinol Metab . 2006; 91( 4): 1357– 1363. Google Scholar CrossRef Search ADS PubMed  8. Wang ET, Calderon-Margalit R, Cedars MI, Daviglus ML, Merkin SS, Schreiner PJ, Sternfeld B, Wellons M, Schwartz SM, Lewis CE, Williams OD, Siscovick DS, Bibbins-Domingo K. Polycystic ovary syndrome and risk for long-term diabetes and dyslipidemia. Obstet Gynecol . 2011; 117( 1): 6– 13. Google Scholar CrossRef Search ADS PubMed  9. Charles MA, Pettitt DJ, McCance DR, Hanson RL, Bennett PH, Knowler WC. Gravidity, obesity, and non-insulin-dependent diabetes among Pima Indian women. Am J Med . 1994; 97( 3): 250– 255. Google Scholar CrossRef Search ADS PubMed  10. Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA . 2006; 295( 11): 1288– 1299. Google Scholar CrossRef Search ADS PubMed  11. Calderon-Margalit R, Schwartz SM, Wellons MF, Lewis CE, Daviglus ML, Schreiner PJ, Williams OD, Sternfeld B, Carr JJ, O’Leary DH, Sidney S, Friedlander Y, Siscovick DS. Prospective association of serum androgens and sex hormone-binding globulin with subclinical cardiovascular disease in young adult women: the “Coronary Artery Risk Development in Young Adults” women’s study. J Clin Endocrinol Metab . 2010; 95( 9): 4424– 4431. Google Scholar CrossRef Search ADS PubMed  12. Fenske B, Kische H, Gross S, Wallaschofski H, Völzke H, Dörr M, Nauck M, Keevil BG, Brabant G, Haring R. Endogenous androgens and sex hormone-binding globulin in women and risk of metabolic syndrome and type 2 diabetes. J Clin Endocrinol Metab . 2015; 100( 12): 4595– 4603. Google Scholar CrossRef Search ADS PubMed  13. Calderon-Margalit R, Siscovick D, Merkin SS, Wang E, Daviglus ML, Schreiner PJ, Sternfeld B, Williams OD, Lewis CE, Azziz R, Schwartz SM, Wellons MF. Prospective association of polycystic ovary syndrome with coronary artery calcification and carotid-intima-media thickness: the Coronary Artery Risk Development in Young Adults Women’s study. Arterioscler Thromb Vasc Biol . 2014; 34( 12): 2688– 2694. Google Scholar CrossRef Search ADS PubMed  14. Ouyang P, Vaidya D, Dobs A, Golden SH, Szklo M, Heckbert SR, Kopp P, Gapstur SM. Sex hormone levels and subclinical atherosclerosis in postmenopausal women: the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis . 2009; 204( 1): 255– 261. Google Scholar CrossRef Search ADS PubMed  15. Kische H, Gross S, Wallaschofski H, Völzke H, Dörr M, Nauck M, Felix SB, Haring R. Serum androgen concentrations and subclinical measures of cardiovascular disease in men and women. Atherosclerosis . 2016; 247: 193– 200. Google Scholar CrossRef Search ADS PubMed  16. Karim R, Hodis HN, Stanczyk FZ, Lobo RA, Mack WJ. Relationship between serum levels of sex hormones and progression of subclinical atherosclerosis in postmenopausal women. J Clin Endocrinol Metab . 2008; 93( 1): 131– 138. Google Scholar CrossRef Search ADS PubMed  17. Bernini GP, Sgro’ M, Moretti A, Argenio GF, Barlascini CO, Cristofani R, Salvetti A. Endogenous androgens and carotid intimal-medial thickness in women. J Clin Endocrinol Metab . 1999; 84( 6): 2008– 2012. Google Scholar CrossRef Search ADS PubMed  18. Golden SH, Maguire A, Ding J, Crouse JR, Cauley JA, Zacur H, Szklo M. Endogenous postmenopausal hormones and carotid atherosclerosis: a case-control study of the atherosclerosis risk in communities cohort. Am J Epidemiol . 2002; 155( 5): 437– 445. Google Scholar CrossRef Search ADS PubMed  19. Solomon CG, Hu FB, Dunaif A, Rich-Edwards J, Willett WC, Hunter DJ, Colditz GA, Speizer FE, Manson JE. Long or highly irregular menstrual cycles as a marker for risk of type 2 diabetes mellitus. JAMA . 2001; 286( 19): 2421– 2426. Google Scholar CrossRef Search ADS PubMed  20. Rostami Dovom M, Ramezani Tehrani F, Djalalinia S, Cheraghi L, Behboudi Gandavani S, Azizi F. Menstrual cycle irregularity and metabolic disorders: a population-based prospective study. PLoS One . 2016; 11( 12): e0168402. Google Scholar CrossRef Search ADS PubMed  21. Solomon CG, Hu FB, Dunaif A, Rich-Edwards JE, Stampfer MJ, Willett WC, Speizer FE, Manson JE. Menstrual cycle irregularity and risk for future cardiovascular disease. J Clin Endocrinol Metab . 2002; 87( 5): 2013– 2017. Google Scholar CrossRef Search ADS PubMed  22. Wang ET, Cirillo PM, Vittinghoff E, Bibbins-Domingo K, Cohn BA, Cedars MI. Menstrual irregularity and cardiovascular mortality. J Clin Endocrinol Metab . 2011; 96( 1): E114– E118. Google Scholar CrossRef Search ADS PubMed  23. LeBlanc ES, Kapphahn K, Hedlin H, Desai M, Parikh NI, Liu S, Parker DR, Anderson M, Aroda V, Sullivan S, Woods NF, Waring ME, Lewis CE, Stefanick M. Reproductive history and risk of type 2 diabetes mellitus in postmenopausal women: findings from the Women’s Health Initiative. Menopause . 2017; 24( 1): 64– 72. Google Scholar CrossRef Search ADS PubMed  24. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM; Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med . 2002; 346( 6): 393– 403. Google Scholar CrossRef Search ADS PubMed  25. Venditti EM, Bray GA, Carrion-Petersen ML, Delahanty LM, Edelstein SL, Hamman RF, Hoskin MA, Knowler WC, Ma Y; Diabetes Prevention Program Research Group. First versus repeat treatment with a lifestyle intervention program: attendance and weight loss outcomes. Int J Obes . 2008; 32( 10): 1537– 1544. Google Scholar CrossRef Search ADS   26. Goldberg RB, Aroda VR, Bluemke DA, Barrett-Connor E, Budoff M, Crandall JP, Dabelea D, Horton ES, Mather KJ, Orchard TJ, Schade D, Watson K, Temprosa M; Diabetes Prevention Program Research Group. Effect of long-term metformin and lifestyle in the Diabetes Prevention Program and its outcome study on coronary artery calcium. Circulation . 2017; 136( 1): 52– 64. Google Scholar CrossRef Search ADS PubMed  27. Hartigan J, Wong M. Algorithm AS 136: a K-means clustering algorithm. J R Stat Soc Ser C Appl Stat . 1979; 28: 100– 108. 28. Breiman L, Freidman J, Olshen RA, Stone CJ. Classification and Regression Trees.  Wadsworth Statistics/Probability. New York, NY: Chapman & Hall/CRC; 1984. 29. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, Buring JE, Gaziano JM, Liu S. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med . 2009; 361( 12): 1152– 1163. Google Scholar CrossRef Search ADS PubMed  30. Spranger J, Möhlig M, Wegewitz U, Ristow M, Pfeiffer AF, Schill T, Schlösser HW, Brabant G, Schöfl C. Adiponectin is independently associated with insulin sensitivity in women with polycystic ovary syndrome. Clin Endocrinol (Oxf) . 2004; 61( 6): 738– 746. Google Scholar CrossRef Search ADS PubMed  31. Creatsa M, Armeni E, Stamatelopoulos K, Rizos D, Georgiopoulos G, Kazani M, Alexandrou A, Dendrinos S, Augoulea A, Papamichael C, Lambrinoudaki I. Circulating androgen levels are associated with subclinical atherosclerosis and arterial stiffness in healthy recently menopausal women. Metabolism . 2012; 61( 2): 193– 201. Google Scholar CrossRef Search ADS PubMed  32. Folsom AR, Kronmal RA, Detrano RC, O’Leary DH, Bild DE, Bluemke DA, Budoff MJ, Liu K, Shea S, Szklo M, Tracy RP, Watson KE, Burke GL. Coronary artery calcification compared with carotid intima-media thickness in the prediction of cardiovascular disease incidence: the Multi-Ethnic Study of Atherosclerosis (MESA). Arch Intern Med . 2008; 168( 12): 1333– 1339. Google Scholar CrossRef Search ADS PubMed  33. Joshi PH, Patel B, Blaha MJ, Berry JD, Blankstein R, Budoff MJ, Wong N, Agatston A, Blumenthal RS, Nasir K. Coronary artery calcium predicts cardiovascular events in participants with a low lifetime risk of cardiovascular disease: the Multi-Ethnic Study of Atherosclerosis (MESA). Atherosclerosis . 2016; 246: 367– 373. Google Scholar CrossRef Search ADS PubMed  34. El Khoudary SR, Wildman RP, Matthews K, Powell L, Hollenberg SM, Edmundowicz D, Sutton-Tyrrell K. Effect modification of obesity on associations between endogenous steroid sex hormones and arterial calcification in women at midlife. Menopause . 2011; 18( 8): 906– 914. Google Scholar CrossRef Search ADS PubMed  35. Rosenfield RL, Ehrmann DA. The pathogenesis of polycystic ovary syndrome (PCOS): the hypothesis of PCOS as functional ovarian hyperandrogenism revisited. Endocr Rev . 2016; 37( 5): 467– 520. Google Scholar CrossRef Search ADS PubMed  36. Vipin VP, Dabadghao P, Shukla M, Kapoor A, Raghuvanshi AS, Ramesh V. Cardiovascular disease risk in first-degree relatives of women with polycystic ovary syndrome. Fertil Steril . 2016; 105( 5): 1338– 1344.e3. Google Scholar CrossRef Search ADS PubMed  37. Carmina E, Campagna AM, Mansuet P, Vitale G, Kort D, Lobo R. Does the level of serum antimüllerian hormone predict ovulatory function in women with polycystic ovary syndrome with aging? Fertil Steril . 2012; 98( 4): 1043– 1046. Google Scholar CrossRef Search ADS PubMed  38. Brown ZA, Louwers YV, Fong SL, Valkenburg O, Birnie E, de Jong FH, Fauser BC, Laven JS. The phenotype of polycystic ovary syndrome ameliorates with aging. Fertil Steril . 2011; 96( 5): 1259– 1265. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

Journal of Clinical Endocrinology and MetabolismOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off