Identification of Plasma Lipid Metabolites Associated with Nut Consumption in US Men and Women

Identification of Plasma Lipid Metabolites Associated with Nut Consumption in US Men and Women ABSTRACT Background Intake of nuts has been inversely associated with risk of type 2 diabetes and cardiovascular disease, partly through inducing a healthy lipid profile. How nut intake may affect lipid metabolites remains unclear. Objective The aim of this study was to identify the plasma lipid metabolites associated with habitual nut consumption in US men and women. Methods We analyzed cross-sectional data from 1099 participants in the Nurses’ Health Study (NHS), NHS II, and Health Professionals Follow-up Study. Metabolic profiling was conducted on plasma by LC–mass spectrometry. Nut intake was estimated from food-frequency questionnaires. We included 144 known lipid metabolites that had CVs ≤25%. Multivariate linear regression was used to assess the associations of nut consumption with individual plasma lipid metabolites. Results We identified 17 lipid metabolites that were significantly associated with nut intake, based on a 1 serving (28 g)/d increment in multivariate models [false discovery rate (FDR) P value <0.05]. Among these species, 8 were positively associated with nut intake [C24:0 sphingomyelin (SM), C36:3 phosphatidylcholine (PC) plasmalogen-A, C36:2 PC plasmalogen, C24:0 ceramide, C36:1 PC plasmalogen, C22:0 SM, C34:1 PC plasmalogen, and C36:2 phosphatidylethanolamine plasmalogen], with changes in relative metabolite level (expressed in number of SDs on the log scale) ranging from 0.36 to 0.46 for 1 serving/d of nuts. The other 9 metabolites were inversely associated with nut intake with changes in relative metabolite level ranging from −0.34 to −0.44. In stratified analysis, 3 metabolites were positively associated with both peanuts and peanut butter (C24:0 SM, C24:0 ceramide, and C22:0 SM), whereas 6 metabolites were inversely associated with other nuts (FDR P value <0.05). Conclusions A panel of lipid metabolites was associated with intake of nuts, which may provide insight into biological mechanisms underlying associations between nuts and cardiometabolic health. Metabolites that were positively associated with intake of nuts may be helpful in identifying potential biomarkers of nut intake. nuts, peanuts, metabolites, lipidomics, lipid metabolites, metabolomics Introduction In recent decades, an extensive body of evidence has emerged linking intake of nuts to a wide range of health benefits including prevention of cardiometabolic diseases (1–3), making them a key component in dietary recommendations and goals for health promotion and disease reduction. Nuts are rich in vegetable protein and unsaturated fatty acids and provide dietary fiber, a myriad of vitamins, and other bioactive constituents such as phytosterols and phenolic compounds (4). In large prospective cohort studies, more frequent nut intake has been inversely associated with risk of type 2 diabetes, metabolic syndrome, cardiovascular disease, and total and cause-specific mortality (1, 2, 5). These findings are consistent with a primary prevention trial that found a 28% reduction in incident cardiovascular events among participants randomly assigned to a Mediterranean diet supplemented with nuts (6), and with short-term trials that have demonstrated beneficial effects of nut intake on intermediate markers of cardiovascular disease risk, including LDL cholesterol (7). However, the biological mechanisms underlying these associations remain unclear. Nutritional metabolomics is a rapidly evolving tool that can provide a comprehensive picture of overall dietary intake by measuring the full profile of small-molecule metabolites in biological samples. By doing so, metabolomics can help deepen our understanding of metabolic pathways relevant to human nutrition and also help identify novel biomarkers of dietary intake, thereby overcoming the common limitation of self-reported measurements of diet in nutritional epidemiology (8, 9). Except for chestnuts, nuts have a high total fat content, ranging from 46% of total calories in cashews and pistachios to 76% in macadamia nuts, the majority of which is from unsaturated fat (4). The high fat content of nuts may influence the plasma lipid profile of habitual consumers, which may underlie the observed cardioprotective effects of nuts and facilitate identification of candidate lipid biomarkers of nut intake. Thus, in the current study, we aimed to identify plasma lipid metabolites associated with habitual nut consumption in a large population of US men and women. We also aimed to examine associations of lipid metabolites according to nut type. Methods Study population This research is not a clinical trial and was therefore not registered. Our cross-sectional analysis was conducted in 3 ongoing prospective cohort studies: the Nurses’ Health Study (NHS), which enrolled 121,700 female nurses aged 30–55 y in 1976; the NHS II, which enrolled 116,429 female nurses aged 25–42 y in 1989; and the Health Professionals Follow-up Study (HPFS), which enrolled 51,529 male health professionals aged 40–75 y in 1986. For each cohort, mailed questionnaires were administered biennially to collect data on lifestyle factors and health, with an overall response rate of nearly 90%. Blood samples were collected from 32,826 women in the NHS from 1989 to 1990, 29,611 women in the NHS II from 1996 to 1999, and 18,225 men in the HPFS from 1993 to 1995. As previously reported, participants who provided a blood specimen were generally similar to those who did not in terms of diet and lifestyle (10). Samples were returned by overnight mail with an icepack and processed immediately upon arrival. Whole blood samples were separated into plasma, buffy coat, and erythrocytes and stored in liquid nitrogen freezers; >95% of samples arrived within 24 h of collection (11, 12). For the current study, we included participants who provided a blood sample and were previously selected either as controls for nested case-control metabolomic analyses of rheumatoid arthritis (NHS and NHS II), ovarian cancer (NHS and NHS II), Parkinson disease (NHS and HPFS), amyotrophic lateral sclerosis (NHS and HPFS), and prostate cancer (HPFS), or as participants in the Mind-Body Study (NHS II). We included participants who had ≥90% of lipid metabolites measured. We excluded participants with self-reported prevalent diabetes at blood draw or those who fasted <8 h before blood collection. After these exclusions, a total of 1099 individuals with available nut intake data (528 from the NHS, 325 from the NHS II, and 246 from the HPFS) were included in the current analysis (Supplemental Figure 1). The study protocol was approved by the Institutional Review Board of the Brigham and Women's Hospital and the Human Subjects Committee Review Board of the Harvard T.H. Chan School of Public Health. Dietary assessment Dietary intake was measured using validated FFQs administered every 4 y (13). Participants were asked to report how often, on average, they consumed a standard portion of foods and beverages, using 9 possible responses ranging from “never or less than once per month” to “6 or more times per day.” To better reflect recent nut consumption, we calculated the average of the intakes from the 2 FFQs closest to the date of blood collection for each cohort (1986 and 1990 in the NHS, 1995 and 1999 in the NHS II, and 1990 and 1994 in the HPFS). FFQ items on nut consumption included “peanuts,” “other nuts,” and “peanut butter.” Although peanuts are technically a legume, they were included in our analysis because they have a similar nutrient profile and are consumed in a similar manner to nuts. “Other nuts” was regarded as all types of tree nuts. Total nut consumption was defined as the intake of peanuts and other nuts and did not include peanut butter. One serving of nuts was equivalent to 28 g (1 oz.) of peanuts or other nuts and was equivalent to 1 tablespoon of peanut butter. A validation study of the FFQ indicated that nut intake correlated well with intakes assessed by multiple dietary records (r = 0.75) (14). The Alternate Healthy Eating Index (AHEI), a measure of dietary quality, was calculated as previously described (15). Nut intake and alcohol were not included in the calculation of the AHEI used in the current analysis (alcohol was separately adjusted in the model). Lipid metabolite profiling Profiles of lipid metabolites were obtained using LC-MS at the Broad Institute of the Massachusetts Institute of Technology and Harvard University (Cambridge, MA). A detailed description of the metabolite profiling methods has been previously published (16, 17). Briefly, plasma polar and nonpolar lipids were profiled using a Nexera X2 U-HPLC system (Shimadzu Scientific Instruments) coupled to an Exactive Plus orbitrap mass spectrometer (Thermo Fisher Scientific). Of note, this instrument identifies lipids at the sum composition level. Lipids were extracted from plasma (10 µL) using 190 µL of isopropanol containing 1,2-didodecanoyl-sn-glycero-3-phosphocholine as an internal standard (Avanti Polar Lipids). After centrifugation (10 min, 9000 × g, ambient temperature), supernatants (10 µL) were injected directly onto a 100 × 2.1 mm ACQUITY BEH C8 column (1.7 µm; Waters). The column was eluted at a flow rate of 450 µL/min isocratically for 1 min at 80% mobile phase A (95:5:0.1, by vol, 10 mM ammonium acetate:methanol:acetic acid), followed by a linear gradient to 80% mobile-phase B (99.9:0.1 vol:vol methanol:acetic acid) over 2 min, a linear gradient to 100% mobile phase B over 7 min, and then 3 min at 100% mobile-phase B. MS analyses were carried out using electrospray ionization in the positive ion mode using full scan analysis over m/z 200–1100 at 70,000 resolution and a 3-Hz data acquisition rate. Additional MS settings were: ion spray voltage, 3.0 kV; capillary temperature, 300°C; probe heater temperature, 300°C; sheath gas, 50; auxiliary gas, 15; and S-lens RF level 60. Raw data were processed using Progenesis QI software (NonLinear Dynamics) for feature alignment, nontargeted signal detection, and signal integration. Targeted processing of a subset of lipids was conducted using TraceFinder software version 3.2 (Thermo Fisher Scientific). Lipids are denoted by headgroup, total acyl carbon content, and total acyl double bond content. For the current analysis, we included 144 known lipid metabolites that demonstrated stability with delayed processing ≤24 h after blood draw. Stability was defined as an intraclass correlation ≥0.75 comparing samples processed immediately with those processed 24 h later. In addition, all metabolites had CVs ≤25%, as measured in blinded quality control samples, and detectable concentrations in ≥90% of participants. Nondietary covariates In the biennial follow-up questionnaires, we collected information on lifestyle factors and medical history, including age, body weight, smoking status, physical activity, and history of chronic diseases. For nondietary covariates in this analysis, we used the questionnaires administered closest in time to blood draw. BMI (in kg/m2) was calculated using height measured at baseline and weight measured closest to blood draw. Statistical analysis Metabolite levels were reported as measured LC-MS peak areas, which are proportional to metabolite concentration. Each metabolite peak area was log-transformed to improve the normality of its distribution. To standardize metabolite values and account for variation in sample handing and laboratory drift between batches, for each log-transformed metabolite peak area, we calculated a z score (SDs from the mean) within each batch and included the z score as the dependent variable in a linear regression model (PROC GLM in SAS version 9.2 for UNIX, SAS Institute, Cary, NC). The primary independent variable in the model was nut intake. The distribution of nut intake in each FFQ cycle was examined and nut consumption remained stable in all cohorts. We first analyzed the association of nut intake (modeled as a continuous variable) with each metabolite using multivariate linear regression. Adjustment for multiple comparisons was performed by the false discovery rate (FDR) procedure (PROC MULTTEST in SAS). Partial Spearman rank correlation coefficients were calculated between statistically significant metabolites (i.e., FDR P < 0.05). We also estimated least-squares means of metabolite z scores in categories of nut intake (never or almost never, less than once per week, once per week, 2–4 times/wk, and ≥5 times/wk). For better illustration, we presented the results using the difference of least-squares means between each higher-intake category and the referent category (i.e., never or almost never). In addition, we performed the analysis by type of nut (peanuts, other nuts), as well as peanut butter because it is a popular source of nuts in the diet. To account for other potential differences between participants that might affect metabolite concentrations, all models included the following covariates: age at blood draw (continuous), cohort (NHS, NHS II, HPFS), smoking status (current, former, never), BMI (continuous), physical activity (continuous), alcohol intake (NHS and NHS II: 0, 0.1–4.9, 5.0–14.9, ≥15 g/d; HPFS: 0, 0.1–4.9, 5.0–29.9, ≥30 g/d), total energy intake (continuous), AHEI (nut intake and alcohol were not included in the calculation; continuous), as well as menopausal status and postmenopausal hormone use (premenopausal, postmenopausal without hormone use, postmenopausal with hormone use) in NHS and NHS II. For continuous covariates, we assigned corresponding medians to the missing values. For categorical covariates, subjects with missing data were assigned to the reference group. To explore the potential predictive ability of the lipid metabolites on nut intake, we generated receiver operating characteristic curves comparing the lowest and highest nut intake categories, i.e., never and ≥2 times/wk (higher intake categories were combined in this analysis), by fitting 3 logistic regression models: a base model adjusted for age at blood draw and cohort; a multivariate model adjusted for the covariates listed above; and a multivariate model further adjusted for the 17 lipid metabolites that were significantly associated with nut intake. AUCs (95% CIs) and P values comparing the models were estimated. These analyses were performed using SAS, and 2-sided P < 0.05 was considered statistically significant. To further assess the relations between statistically significant metabolites (FDR P < 0.05), we also performed a hierarchical cluster analysis using the hclust function in R3.5.1. Results The study flow of participants is shown in Supplemental Figure 1. The age-adjusted characteristics of study participants according to their frequency of nut intake are shown in Table 1. Participants who had a higher intake of nuts tended to be older, have a lower BMI, and were more physically active than those with a lower intake. In addition, participants who had a higher intake of nuts had a higher AHEI score, indicative of better diet quality, and drank more alcohol than infrequent consumers. TABLE 1 Age-adjusted characteristics of 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study by frequency of nut consumption1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 1Values are means ± SDs unless otherwise indicated. MET-h, metabolic equivalent hours; NHS, Nurses’ Health Study. 2Not age-adjusted. 3Nut intake and alcohol were not included in the calculation. 4Values for blood concentrations of total cholesterol, LDL cholesterol, HDL cholesterol, and TGs were based on data from 362, 645, 645, and 910 participants, respectively, because only a subset of participants had data on both blood lipids (total cholesterol, LDL cholesterol, HDL cholesterol, and TGs) and blood lipid metabolites. View Large TABLE 1 Age-adjusted characteristics of 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study by frequency of nut consumption1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 1Values are means ± SDs unless otherwise indicated. MET-h, metabolic equivalent hours; NHS, Nurses’ Health Study. 2Not age-adjusted. 3Nut intake and alcohol were not included in the calculation. 4Values for blood concentrations of total cholesterol, LDL cholesterol, HDL cholesterol, and TGs were based on data from 362, 645, 645, and 910 participants, respectively, because only a subset of participants had data on both blood lipids (total cholesterol, LDL cholesterol, HDL cholesterol, and TGs) and blood lipid metabolites. View Large Of the 144 known lipid metabolites, 17 were significantly associated with nut intake (FDR P < 0.05). Details of the individual metabolites along with effect sizes, which denote the change in relative metabolite level (expressed in number of SDs on the log scale) for 1 serving (28 g)/d of nut intake, are shown in Supplemental Table 1. Among the 17 metabolites that were significantly associated with nut intake, 9 were inversely associated [C34:3 diacylglycerol (DAG), C16:1 lysophosphatidylcholine (LPC), C16:1 cholesterol ester (CE), C32:1 DAG, C22:6 lysophosphatidylethanolamine (LPE), C22:6 LPC, C18:0 sphingomyelin (SM), C50:2 TG, and C34:2 DAG] and 8 were positively associated [C24:0 SM, C36:3 phosphatidylcholine (PC) plasmalogen-A, C36:2 PC plasmalogen, C24:0 ceramide, C36:1 PC plasmalogen, C22:0 SM, C34:1 PC plasmalogen, and C36:2 phosphatidylethanolamine plasmalogen]. Similar results were observed in the categorical analysis, which shows differences (95% CIs) in metabolite concentrations by frequency of nut consumption (Table 2) (P for trend <0.01). The majority of the metabolites were positively correlated with one another (Figure 1 and Supplemental Table 2). Intakes of peanuts and peanut butter were both positively associated with 3 metabolites (C24:0 SM, C24:0 ceramide, and C22:0 SM) and intake of other nuts was inversely associated with 6 metabolites (C16:1 LPC, C16:1 CE, C34:3 DAG, C22:6 LPC, C22:6 LPE, and C32:1 DAG) (Supplemental Table 3). FIGURE 1 View largeDownload slide Partial Spearman correlations between statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. The calculation is based on the z scores of log-transformed, continuous metabolite concentrations and adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. The individual correlation coefficients between metabolites are provided in Supplemental Table 2. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 1 View largeDownload slide Partial Spearman correlations between statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. The calculation is based on the z scores of log-transformed, continuous metabolite concentrations and adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. The individual correlation coefficients between metabolites are provided in Supplemental Table 2. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. TABLE 2 Differences (95% CIs) in relative metabolite levels by nut consumption among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study1 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 1Differences in least-squares means of relative metabolite levels (expressed in number of SDs on the log scale) between each higher intake category and the referent category. The models are adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. CE, cholesterol ester; DAG, diacylglycerol; HMDB, Human Metabolome Database; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; Ref, reference; SM, sphingomyelin. 2Nut intake as a continuous variable. View Large TABLE 2 Differences (95% CIs) in relative metabolite levels by nut consumption among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study1 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 1Differences in least-squares means of relative metabolite levels (expressed in number of SDs on the log scale) between each higher intake category and the referent category. The models are adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. CE, cholesterol ester; DAG, diacylglycerol; HMDB, Human Metabolome Database; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; Ref, reference; SM, sphingomyelin. 2Nut intake as a continuous variable. View Large Among the metabolites that were significantly associated with nut intake, hierarchical clustering analysis revealed 5 clusters of metabolites (Figure 2). Clusters 2 and 5 included metabolites that were positively associated with nut intake and were primarily comprised of very-long-chain (≥C22) SMs, ceramides, PC, and phosphatidylethanolamine plasmalogens. Clusters 1, 3, and 4 included metabolites that were inversely associated with nut intake and were comprised of TGs, DAGs, LPEs, LPCs, CEs, and SMs. Of note, the metabolites associated with intake of peanuts and peanut butter were specific to cluster 2. Adding the metabolites to the multivariate model improved prediction of nut intake beyond the other factors. The AUCs were 0.81 in the multivariate model with metabolites and 0.76 in the multivariate model, with a P value of 0.002 comparing these 2 models (Figure 3). FIGURE 2 View largeDownload slide Hierarchical clustering dendrogram of statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. (A) Hierarchical cluster analysis using the “hclust” function in R. The vertical axis represents the distance or dissimilarity between clusters. (B) Metabolites were grouped into 5 clusters by further using the “cutree” function in R. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 2 View largeDownload slide Hierarchical clustering dendrogram of statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. (A) Hierarchical cluster analysis using the “hclust” function in R. The vertical axis represents the distance or dissimilarity between clusters. (B) Metabolites were grouped into 5 clusters by further using the “cutree” function in R. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 3 View largeDownload slide ROC curves comparing models to predict nut intake (never compared with ≥2 times/wk) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study (note: higher intake categories were combined in this analysis). Logistic regression models were fitted to predict nut intake (never compared with ≥2 times/wk) among 525 participants. Base model: adjusted for age at blood draw and cohort. MV model: also adjusted for smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. MV + metabolites model: also adjusted for the 17 statistically significant metabolites. P values were 0.0002 comparing the MV + metabolites model with the MV model; <0.0001 comparing the MV + metabolites model with the base model; and <0.0001 comparing the MV model with the base model. MV, multivariate; NHS, Nurses’ Health Study; ROC, receiver operating characteristic. FIGURE 3 View largeDownload slide ROC curves comparing models to predict nut intake (never compared with ≥2 times/wk) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study (note: higher intake categories were combined in this analysis). Logistic regression models were fitted to predict nut intake (never compared with ≥2 times/wk) among 525 participants. Base model: adjusted for age at blood draw and cohort. MV model: also adjusted for smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. MV + metabolites model: also adjusted for the 17 statistically significant metabolites. P values were 0.0002 comparing the MV + metabolites model with the MV model; <0.0001 comparing the MV + metabolites model with the base model; and <0.0001 comparing the MV model with the base model. MV, multivariate; NHS, Nurses’ Health Study; ROC, receiver operating characteristic. Discussion In a well-defined sample of men and women from 3 large cohorts, we identified 17 out of 144 known plasma lipid metabolites that were significantly associated with nut intake. Associations differed in direction depending on lipid structure and nut type. To our knowledge, this is the first study to specifically examine the plasma lipid metabolite profile of total and type of nut consumption in a large observational study. To date, few studies have evaluated potential biomarkers of nut intake. These include intervention studies ranging from 12 wk to 6 mo that identified conjugated fatty acids along with serotonin metabolites and microbial-derived phenolic metabolites as markers of nut intake (18–20). In the PREDIMED trial, walnut consumption was characterized by 18 urinary metabolites, including markers of fatty acid metabolism, ellagitannin-derived microbial compounds, and metabolites of the tryptophan/serotonin pathway (21). In the majority of these trials, metabolites were measured in urine and may be more reflective of short-term or acute intake of nuts at prescribed doses. In contrast, the plasma metabolites measured in our study may be more reflective of long-term habitual nut intake patterns, which may partly explain the different lipidomic profile that emerged from our analysis. Differences in the study population, type of nuts, sample size, and metabolite profiling techniques might also account for differences with previous studies. Nuts are a nutrient-dense food with a complex matrix of bioactive compounds (4). Although the total fat content in nuts is high, ranging from 46% in cashews and pistachios to 76% in macadamia nuts, the saturated fat content is low, ranging from 4% to 16% (4). Most nuts contain a high proportion of MUFA [oleic acid (18:1n–9)]; however, certain nuts, such as Brazil nuts, have similar proportions of MUFAs and PUFAs [mostly linoleic acid (18:2n–6)], whereas walnuts contain mostly PUFAs, both from linoleic acid and α-linolenic acid (18:3n–3) (22). The fatty fraction of nuts also contains large amounts of noncholesterol sterols (i.e., phytosterols) (23), which play an important structural role in membranes, where they stabilize phospholipid bilayers (24). Nuts also contain choline, which is found in SMs, PC, and LPC (25). Thus, nuts may have a unique lipidomic signature that can be used as a biomarker for intake. In our analysis, 2 SMs (C24:0 and C22:0) and a ceramide (C24:0) derived from very-long-chain SFAs (VLCSFAs) were positively associated with nut and peanut intake. Ceramides and SMs are lipid molecules with structural and signaling roles in cell membranes. However, most ceramides have been linked to increased cardiometabolic risk (26–28) and C24:0 ceramide has been found to promote insulin resistance in rodents (29). Exposure to VLCSFAs can promote ceramide formation in vitro and in animal models but how this relates to diet is not well understood (30). We previously found inverse associations between 3 plasma VLCSFAs (C20:0, C22:0, and C24:0) and risk of coronary artery disease in the NHS and HPFS (31), which may support a beneficial role of ceramides from these VLCSFAs on cardiometabolic health. This suggests that nut intake may have a beneficial effect on cardiometabolic health through a positive association with ceramide (C24:0). The other metabolites that were positively associated with nut intake in our analysis include plasmalogens, which are plasmenyl-phospholipids with a vinyl ether linkage at the sn-1 position and PUFA linkage at the sn-2 position, thought to protect mammalian cells against reactive oxygen species. Metabolites that were inversely associated with nut intake include TG, DAG, and CEs, consistent with the favorable effects of nut intake on total cholesterol, LDL cholesterol, apoB, and TGs observed in trials (7). Whether these metabolites are biologically meaningful is not clear (17). Our study has limitations. It is difficult to know whether the metabolites we identified are truly indicative of nut intake or endogenous production. In addition, because lipid data were expressed at the sum composition level, we were not able to determine concentrations of specific fatty acids and how they directly relate to nut intake. Although we adjusted for multiple potential confounders, it is not possible to rule out residual confounding from unmeasured or poorly measured factors related to diet and lifestyle. Measurement error in dietary assessment using FFQs is also inevitable and may lead to an attenuation of diet–metabolite associations (13). The strengths of our study include a large sample size and detailed diet and lifestyle information, which facilitated fine control for potential confounding. Using mean nut intake from 2 FFQs reduced within-person variability and better represented habitual intake. In analyzing the lipid metabolites, we controlled for multiple testing and ensured that all metabolites had CVs ≤25%, as measured in blinded quality control samples, and detectable concentrations in ≥90% of participants. In conclusion, we identified a set of lipid metabolites from known species that were associated with nut intake. Associations differed by type of lipid molecule and type of nut. These findings may help identify biomarkers of nut intake and provide insight into biological mechanisms underlying associations between nuts and cardiometabolic health. However, replication of our findings in additional studies of the metabolome and among diverse populations is needed. Acknowledgments The authors’ responsibilities were as follows—VSM and YB: designed the analysis, interpreted the data, wrote the manuscript, and had primary responsibility for the final content; YB: conducted the analysis; SST, EWK, KHC, AA, KMW, and LAM: provided access to the data for the analysis; MG-F, FBH, MKT, OAZ, AHE, SST, EWK, KHC, AA, KMW, LAM, ELG, and CSF: critically reviewed the manuscript for important intellectual content; and all authors: read and approved the final manuscript. Notes Supported by NIH grants UM1 CA186107, UM1 CA176726, UM1 CA167552, U01 167552, P01 CA87969, R01 AR049880, R01 CA49449, R01 CA67262, R01 CA50385, P50 CA090381, U54CA155626 (to FBH), P30DK046200, K01 HL125698 (to MKT), and KL2 TR001100; Department of Defense grants W81XWH-13-1-0493 and CA150357 (to YB); and by a grant from the International Tree Nut Council Nutrition Research & Education Foundation (to YB). Author disclosures: YB received a research grant from the International Tree Nut Council Nutrition Research & Education Foundation. VSM received research support from the Peanut Institute. MG-F, FBH, MKT, OAZ, AHE, SST, EWK, KHC, AA, KMW, LAM, ELG, and CSF, no conflicts of interest. The funders of this study had no role in its design or conduct; in the collection, management, analysis, or interpretation of the data; or in the preparation, review, or approval of the manuscript. Supplemental Figure 1 and Supplemental Tables 1–3 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/. Abbreviations used: AHEI, Alternate Healthy Eating Index; CE, cholesterol ester; DAG, diacylglycerol; FDR, false discovery rate; HPFS, Health Professionals Follow-up Study; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; SM, sphingomyelin; VLCSFA, very-long-chain SFA. References 1. Aune D , Keum N , Giovannucci E , Fadnes LT , Boffetta P , Greenwood DC , Tonstad S , Vatten LJ , Riboli E , Norat T . Nut consumption and risk of cardiovascular disease, total cancer, all-cause and cause-specific mortality: a systematic review and dose-response meta-analysis of prospective studies . BMC Med . 2016 ; 14 ( 1 ): 207 . Google Scholar Crossref Search ADS PubMed 2. Yu Z , Malik VS , Keum N , Hu FB , Giovannucci EL , Stampfer MJ , Willett WC , Fuchs CS , Bao Y . Associations between nut consumption and inflammatory biomarkers . Am J Clin Nutr . 2016 ; 104 ( 3 ): 722 – 8 . Google Scholar Crossref Search ADS PubMed 3. Guasch-Ferré M , Liu X , Malik VS , Sun Q , Willett WC , Manson JE , Rexrode KM , Li Y , Hu FB , Bhupathiraju SN . Nut consumption and risk of cardiovascular disease . J Am Coll Cardiol . 2017 ; 70 ( 20 ): 2519 – 32 . Google Scholar Crossref Search ADS PubMed 4. Ros E . Health benefits of nut consumption . Nutrients . 2010 ; 2 ( 7 ): 652 – 82 . Google Scholar Crossref Search ADS PubMed 5. Pan A , Sun Q , Manson JE , Willett WC , Hu FB . Walnut consumption is associated with lower risk of type 2 diabetes in women . J Nutr . 2013 ; 143 ( 4 ): 512 – 8 . Google Scholar Crossref Search ADS PubMed 6. Estruch R , Ros E , Salas-Salvadó J , Covas MI , Corella D , Arós F , Gómez-Gracia E , Ruiz-Gutiérrez V , Fiol M , Lapetra J et al. . Primary prevention of cardiovascular disease with a Mediterranean diet . N Engl J Med . 2013 ; 368 ( 14 ): 1279 – 90 . Google Scholar Crossref Search ADS PubMed 7. Del Gobbo LC , Falk MC , Feldman R , Lewis K , Mozaffarian D . Effects of tree nuts on blood lipids, apolipoproteins, and blood pressure: systematic review, meta-analysis, and dose-response of 61 controlled intervention trials . Am J Clin Nutr . 2015 ; 102 ( 6 ): 1347 – 56 . Google Scholar Crossref Search ADS PubMed 8. Satija A , Yu E , Willett WC , Hu FB . Understanding nutritional epidemiology and its role in policy . Adv Nutr . 2015 ; 6 ( 1 ): 5 – 18 . Google Scholar Crossref Search ADS PubMed 9. Guasch-Ferre M , Bhupathiraju SN , Hu FB . Use of metabolomics in improving assessment of dietary intake . Clin Chem . 2018 ; 64 ( 1 ): 82 – 98 . Google Scholar Crossref Search ADS PubMed 10. Hunter DJ , Hankinson SE , Hough H , Gertig DM , Garcia-Closas M , Spiegelman D , Manson JE , Colditz GA , Willett WC , Speizer FE et al. . A prospective study of NAT2 acetylation genotype, cigarette smoking, and risk of breast cancer . Carcinogenesis . 1997 ; 18 ( 11 ): 2127 – 32 . Google Scholar Crossref Search ADS PubMed 11. Hankinson SE , Willett WC , Manson JE , Colditz GA , Hunter DJ , Spiegelman D , Barbieri RL , Speizer FE . Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women . J Natl Cancer Inst . 1998 ; 90 ( 17 ): 1292 – 9 . Google Scholar Crossref Search ADS PubMed 12. Chu NF , Spiegelman D , Yu J , Rifai N , Hotamisligil GS , Rimm EB . Plasma leptin concentrations and four-year weight gain among US men . Int J Obes Relat Metab Disord . 2001 ; 25 ( 3 ): 346 – 53 . Google Scholar Crossref Search ADS PubMed 13. Willett WC , Sampson L , Stampfer MJ , Rosner B , Bain C , Witschi J , Hennekens CH , Speizer FE . Reproducibility and validity of a semiquantitative food frequency questionnaire . Am J Epidemiol . 1985 ; 122 ( 1 ): 51 – 65 . Google Scholar Crossref Search ADS PubMed 14. Salvini S , Hunter DJ , Sampson L , Stampfer MJ , Colditz GA , Rosner B , Willett WC . Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption . Int J Epidemiol . 1989 ; 18 ( 4 ): 858 – 67 . Google Scholar Crossref Search ADS PubMed 15. Chiuve SE , Fung TT , Rimm EB , Hu FB , McCullough ML , Wang M , Stampfer MJ , Willett WC . Alternative dietary indices both strongly predict risk of chronic disease . J Nutr . 2012 ; 142 ( 6 ): 1009 – 18 . Google Scholar Crossref Search ADS PubMed 16. Mayers JR , Wu C , Clish CB , Kraft P , Torrence ME , Fiske BP , Yuan C , Bao Y , Townsend MK , Tworoger SS et al. . Elevation of circulating branched-chain amino acids is an early event in human pancreatic adenocarcinoma development . Nat Med . 2014 ; 20 ( 10 ): 1193 – 8 . Google Scholar Crossref Search ADS PubMed 17. Toledo E , Wang DD , Ruiz-Canela M , Clish CB , Razquin C , Zheng Y , Guasch-Ferre M , Hruby A , Corella D , Gomez-Gracia E et al. . Plasma lipidomic profiles and cardiovascular events in a randomized intervention trial with the Mediterranean diet . Am J Clin Nutr . 2017 ; 106 ( 4 ): 973 – 83 . Google Scholar PubMed 18. Tulipani S , Llorach R , Jáuregui O , López-Uriarte P , Garcia-Aloy M , Bullo M , Salas-Salvadó J , Andrés-Lacueva C . Metabolomics unveils urinary changes in subjects with metabolic syndrome following 12-week nut consumption . J Proteome Res . 2011 ; 10 ( 11 ): 5047 – 58 . Google Scholar Crossref Search ADS PubMed 19. Andersen MB , Kristensen M , Manach C , Pujos-Guillot E , Poulsen SK , Larsen TM , Astrup A , Dragsted L . Discovery and validation of urinary exposure markers for different plant foods by untargeted metabolomics . Anal Bioanal Chem . 2014 ; 406 ( 7 ): 1829 – 44 . Google Scholar Crossref Search ADS PubMed 20. Mora-Cubillos X , Tulipani S , Garcia-Aloy M , Bulló M , Tinahones FJ , Andres-Lacueva C . Plasma metabolomic biomarkers of mixed nuts exposure inversely correlate with severity of metabolic syndrome . Mol Nutr Food Res . 2015 ; 59 ( 12 ): 2480 – 90 . Google Scholar Crossref Search ADS PubMed 21. Garcia-Aloy M , Llorach R , Urpi-Sarda M , Tulipani S , Estruch R , Martínez-González MA , Corella D , Fitó M , Ros E , Salas-Salvadó J et al. . Novel multimetabolite prediction of walnut consumption by a urinary biomarker model in a free-living population: the PREDIMED study . J Proteome Res . 2014 ; 13 ( 7 ): 3476 – 83 . Google Scholar Crossref Search ADS PubMed 22. Ros E , Mataix J . Fatty acid composition of nuts – implications for cardiovascular health . Br J Nutr . 2006 ; 96 ( Suppl 2 ): S29 – 35 . Google Scholar Crossref Search ADS PubMed 23. Segura R , Javierre C , Lizarraga MA , Ros E . Other relevant components of nuts: phytosterols, folate and minerals . Br J Nutr . 2006 ; 96 ( Suppl 2 ): S36 – 44 . Google Scholar Crossref Search ADS PubMed 24. Hartmann MA , . Plant sterols and the membrane environment . Trends Plant Sci . 1998 ; 3 : 170 – 175 . Google Scholar Crossref Search ADS 25. Li Z , Vance DE . Phosphatidylcholine and choline homeostasis . J Lipid Res . 2008 ; 49 ( 6 ): 1187 – 94 . Google Scholar Crossref Search ADS PubMed 26. Chavez JA , Summers SA . A ceramide-centric view of insulin resistance . Cell Metab . 2012 ; 15 ( 5 ): 585 – 94 . Google Scholar Crossref Search ADS PubMed 27. Holland WL , Summers SA . Sphingolipids, insulin resistance, and metabolic disease: new insights from in vivo manipulation of sphingolipid metabolism . Endocr Rev . 2008 ; 29 ( 4 ): 381 – 402 . Google Scholar Crossref Search ADS PubMed 28. Laaksonen R , Ekroos K , Sysi-Aho M , Hilvo M , Vihervaara T , Kauhanen D , Suoniemi M , Hurme R , Marz W , Scharnagl H et al. . Plasma ceramides predict cardiovascular death in patients with stable coronary artery disease and acute coronary syndromes beyond LDL-cholesterol . Eur Heart J . 2016 ; 37 ( 25 ): 1967 – 76 . Google Scholar Crossref Search ADS PubMed 29. Kirwan JP . Plasma ceramides target skeletal muscle in type 2 diabetes . Diabetes . 2013 ; 62 ( 2 ): 352 – 4 . Google Scholar Crossref Search ADS PubMed 30. Chavez JA , Summers SA . Characterizing the effects of saturated fatty acids on insulin signaling and ceramide and diacylglycerol accumulation in 3T3-L1 adipocytes and C2C12 myotubes . Arch Biochem Biophys . 2003 ; 419 ( 2 ): 101 – 9 . . Google Scholar Crossref Search ADS PubMed 31. Malik VS , Chiuve SE , Campos H , Rimm EB , Mozaffarian D , Hu FB , Sun Q . Circulating very-long-chain saturated fatty acids and incident coronary heart disease in US men and women . Circulation . 2015 ; 132 ( 4 ): 260 – 8 . Google Scholar Crossref Search ADS PubMed Copyright © American Society for Nutrition 2019. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Nutrition Oxford University Press

Loading next page...
 
/lp/oxford-university-press/identification-of-plasma-lipid-metabolites-associated-with-nut-0SLN7zl20i
Publisher
Oxford University Press
Copyright
Copyright © American Society for Nutrition 2019.
ISSN
0022-3166
eISSN
1541-6100
D.O.I.
10.1093/jn/nxz048
Publisher site
See Article on Publisher Site

Abstract

ABSTRACT Background Intake of nuts has been inversely associated with risk of type 2 diabetes and cardiovascular disease, partly through inducing a healthy lipid profile. How nut intake may affect lipid metabolites remains unclear. Objective The aim of this study was to identify the plasma lipid metabolites associated with habitual nut consumption in US men and women. Methods We analyzed cross-sectional data from 1099 participants in the Nurses’ Health Study (NHS), NHS II, and Health Professionals Follow-up Study. Metabolic profiling was conducted on plasma by LC–mass spectrometry. Nut intake was estimated from food-frequency questionnaires. We included 144 known lipid metabolites that had CVs ≤25%. Multivariate linear regression was used to assess the associations of nut consumption with individual plasma lipid metabolites. Results We identified 17 lipid metabolites that were significantly associated with nut intake, based on a 1 serving (28 g)/d increment in multivariate models [false discovery rate (FDR) P value <0.05]. Among these species, 8 were positively associated with nut intake [C24:0 sphingomyelin (SM), C36:3 phosphatidylcholine (PC) plasmalogen-A, C36:2 PC plasmalogen, C24:0 ceramide, C36:1 PC plasmalogen, C22:0 SM, C34:1 PC plasmalogen, and C36:2 phosphatidylethanolamine plasmalogen], with changes in relative metabolite level (expressed in number of SDs on the log scale) ranging from 0.36 to 0.46 for 1 serving/d of nuts. The other 9 metabolites were inversely associated with nut intake with changes in relative metabolite level ranging from −0.34 to −0.44. In stratified analysis, 3 metabolites were positively associated with both peanuts and peanut butter (C24:0 SM, C24:0 ceramide, and C22:0 SM), whereas 6 metabolites were inversely associated with other nuts (FDR P value <0.05). Conclusions A panel of lipid metabolites was associated with intake of nuts, which may provide insight into biological mechanisms underlying associations between nuts and cardiometabolic health. Metabolites that were positively associated with intake of nuts may be helpful in identifying potential biomarkers of nut intake. nuts, peanuts, metabolites, lipidomics, lipid metabolites, metabolomics Introduction In recent decades, an extensive body of evidence has emerged linking intake of nuts to a wide range of health benefits including prevention of cardiometabolic diseases (1–3), making them a key component in dietary recommendations and goals for health promotion and disease reduction. Nuts are rich in vegetable protein and unsaturated fatty acids and provide dietary fiber, a myriad of vitamins, and other bioactive constituents such as phytosterols and phenolic compounds (4). In large prospective cohort studies, more frequent nut intake has been inversely associated with risk of type 2 diabetes, metabolic syndrome, cardiovascular disease, and total and cause-specific mortality (1, 2, 5). These findings are consistent with a primary prevention trial that found a 28% reduction in incident cardiovascular events among participants randomly assigned to a Mediterranean diet supplemented with nuts (6), and with short-term trials that have demonstrated beneficial effects of nut intake on intermediate markers of cardiovascular disease risk, including LDL cholesterol (7). However, the biological mechanisms underlying these associations remain unclear. Nutritional metabolomics is a rapidly evolving tool that can provide a comprehensive picture of overall dietary intake by measuring the full profile of small-molecule metabolites in biological samples. By doing so, metabolomics can help deepen our understanding of metabolic pathways relevant to human nutrition and also help identify novel biomarkers of dietary intake, thereby overcoming the common limitation of self-reported measurements of diet in nutritional epidemiology (8, 9). Except for chestnuts, nuts have a high total fat content, ranging from 46% of total calories in cashews and pistachios to 76% in macadamia nuts, the majority of which is from unsaturated fat (4). The high fat content of nuts may influence the plasma lipid profile of habitual consumers, which may underlie the observed cardioprotective effects of nuts and facilitate identification of candidate lipid biomarkers of nut intake. Thus, in the current study, we aimed to identify plasma lipid metabolites associated with habitual nut consumption in a large population of US men and women. We also aimed to examine associations of lipid metabolites according to nut type. Methods Study population This research is not a clinical trial and was therefore not registered. Our cross-sectional analysis was conducted in 3 ongoing prospective cohort studies: the Nurses’ Health Study (NHS), which enrolled 121,700 female nurses aged 30–55 y in 1976; the NHS II, which enrolled 116,429 female nurses aged 25–42 y in 1989; and the Health Professionals Follow-up Study (HPFS), which enrolled 51,529 male health professionals aged 40–75 y in 1986. For each cohort, mailed questionnaires were administered biennially to collect data on lifestyle factors and health, with an overall response rate of nearly 90%. Blood samples were collected from 32,826 women in the NHS from 1989 to 1990, 29,611 women in the NHS II from 1996 to 1999, and 18,225 men in the HPFS from 1993 to 1995. As previously reported, participants who provided a blood specimen were generally similar to those who did not in terms of diet and lifestyle (10). Samples were returned by overnight mail with an icepack and processed immediately upon arrival. Whole blood samples were separated into plasma, buffy coat, and erythrocytes and stored in liquid nitrogen freezers; >95% of samples arrived within 24 h of collection (11, 12). For the current study, we included participants who provided a blood sample and were previously selected either as controls for nested case-control metabolomic analyses of rheumatoid arthritis (NHS and NHS II), ovarian cancer (NHS and NHS II), Parkinson disease (NHS and HPFS), amyotrophic lateral sclerosis (NHS and HPFS), and prostate cancer (HPFS), or as participants in the Mind-Body Study (NHS II). We included participants who had ≥90% of lipid metabolites measured. We excluded participants with self-reported prevalent diabetes at blood draw or those who fasted <8 h before blood collection. After these exclusions, a total of 1099 individuals with available nut intake data (528 from the NHS, 325 from the NHS II, and 246 from the HPFS) were included in the current analysis (Supplemental Figure 1). The study protocol was approved by the Institutional Review Board of the Brigham and Women's Hospital and the Human Subjects Committee Review Board of the Harvard T.H. Chan School of Public Health. Dietary assessment Dietary intake was measured using validated FFQs administered every 4 y (13). Participants were asked to report how often, on average, they consumed a standard portion of foods and beverages, using 9 possible responses ranging from “never or less than once per month” to “6 or more times per day.” To better reflect recent nut consumption, we calculated the average of the intakes from the 2 FFQs closest to the date of blood collection for each cohort (1986 and 1990 in the NHS, 1995 and 1999 in the NHS II, and 1990 and 1994 in the HPFS). FFQ items on nut consumption included “peanuts,” “other nuts,” and “peanut butter.” Although peanuts are technically a legume, they were included in our analysis because they have a similar nutrient profile and are consumed in a similar manner to nuts. “Other nuts” was regarded as all types of tree nuts. Total nut consumption was defined as the intake of peanuts and other nuts and did not include peanut butter. One serving of nuts was equivalent to 28 g (1 oz.) of peanuts or other nuts and was equivalent to 1 tablespoon of peanut butter. A validation study of the FFQ indicated that nut intake correlated well with intakes assessed by multiple dietary records (r = 0.75) (14). The Alternate Healthy Eating Index (AHEI), a measure of dietary quality, was calculated as previously described (15). Nut intake and alcohol were not included in the calculation of the AHEI used in the current analysis (alcohol was separately adjusted in the model). Lipid metabolite profiling Profiles of lipid metabolites were obtained using LC-MS at the Broad Institute of the Massachusetts Institute of Technology and Harvard University (Cambridge, MA). A detailed description of the metabolite profiling methods has been previously published (16, 17). Briefly, plasma polar and nonpolar lipids were profiled using a Nexera X2 U-HPLC system (Shimadzu Scientific Instruments) coupled to an Exactive Plus orbitrap mass spectrometer (Thermo Fisher Scientific). Of note, this instrument identifies lipids at the sum composition level. Lipids were extracted from plasma (10 µL) using 190 µL of isopropanol containing 1,2-didodecanoyl-sn-glycero-3-phosphocholine as an internal standard (Avanti Polar Lipids). After centrifugation (10 min, 9000 × g, ambient temperature), supernatants (10 µL) were injected directly onto a 100 × 2.1 mm ACQUITY BEH C8 column (1.7 µm; Waters). The column was eluted at a flow rate of 450 µL/min isocratically for 1 min at 80% mobile phase A (95:5:0.1, by vol, 10 mM ammonium acetate:methanol:acetic acid), followed by a linear gradient to 80% mobile-phase B (99.9:0.1 vol:vol methanol:acetic acid) over 2 min, a linear gradient to 100% mobile phase B over 7 min, and then 3 min at 100% mobile-phase B. MS analyses were carried out using electrospray ionization in the positive ion mode using full scan analysis over m/z 200–1100 at 70,000 resolution and a 3-Hz data acquisition rate. Additional MS settings were: ion spray voltage, 3.0 kV; capillary temperature, 300°C; probe heater temperature, 300°C; sheath gas, 50; auxiliary gas, 15; and S-lens RF level 60. Raw data were processed using Progenesis QI software (NonLinear Dynamics) for feature alignment, nontargeted signal detection, and signal integration. Targeted processing of a subset of lipids was conducted using TraceFinder software version 3.2 (Thermo Fisher Scientific). Lipids are denoted by headgroup, total acyl carbon content, and total acyl double bond content. For the current analysis, we included 144 known lipid metabolites that demonstrated stability with delayed processing ≤24 h after blood draw. Stability was defined as an intraclass correlation ≥0.75 comparing samples processed immediately with those processed 24 h later. In addition, all metabolites had CVs ≤25%, as measured in blinded quality control samples, and detectable concentrations in ≥90% of participants. Nondietary covariates In the biennial follow-up questionnaires, we collected information on lifestyle factors and medical history, including age, body weight, smoking status, physical activity, and history of chronic diseases. For nondietary covariates in this analysis, we used the questionnaires administered closest in time to blood draw. BMI (in kg/m2) was calculated using height measured at baseline and weight measured closest to blood draw. Statistical analysis Metabolite levels were reported as measured LC-MS peak areas, which are proportional to metabolite concentration. Each metabolite peak area was log-transformed to improve the normality of its distribution. To standardize metabolite values and account for variation in sample handing and laboratory drift between batches, for each log-transformed metabolite peak area, we calculated a z score (SDs from the mean) within each batch and included the z score as the dependent variable in a linear regression model (PROC GLM in SAS version 9.2 for UNIX, SAS Institute, Cary, NC). The primary independent variable in the model was nut intake. The distribution of nut intake in each FFQ cycle was examined and nut consumption remained stable in all cohorts. We first analyzed the association of nut intake (modeled as a continuous variable) with each metabolite using multivariate linear regression. Adjustment for multiple comparisons was performed by the false discovery rate (FDR) procedure (PROC MULTTEST in SAS). Partial Spearman rank correlation coefficients were calculated between statistically significant metabolites (i.e., FDR P < 0.05). We also estimated least-squares means of metabolite z scores in categories of nut intake (never or almost never, less than once per week, once per week, 2–4 times/wk, and ≥5 times/wk). For better illustration, we presented the results using the difference of least-squares means between each higher-intake category and the referent category (i.e., never or almost never). In addition, we performed the analysis by type of nut (peanuts, other nuts), as well as peanut butter because it is a popular source of nuts in the diet. To account for other potential differences between participants that might affect metabolite concentrations, all models included the following covariates: age at blood draw (continuous), cohort (NHS, NHS II, HPFS), smoking status (current, former, never), BMI (continuous), physical activity (continuous), alcohol intake (NHS and NHS II: 0, 0.1–4.9, 5.0–14.9, ≥15 g/d; HPFS: 0, 0.1–4.9, 5.0–29.9, ≥30 g/d), total energy intake (continuous), AHEI (nut intake and alcohol were not included in the calculation; continuous), as well as menopausal status and postmenopausal hormone use (premenopausal, postmenopausal without hormone use, postmenopausal with hormone use) in NHS and NHS II. For continuous covariates, we assigned corresponding medians to the missing values. For categorical covariates, subjects with missing data were assigned to the reference group. To explore the potential predictive ability of the lipid metabolites on nut intake, we generated receiver operating characteristic curves comparing the lowest and highest nut intake categories, i.e., never and ≥2 times/wk (higher intake categories were combined in this analysis), by fitting 3 logistic regression models: a base model adjusted for age at blood draw and cohort; a multivariate model adjusted for the covariates listed above; and a multivariate model further adjusted for the 17 lipid metabolites that were significantly associated with nut intake. AUCs (95% CIs) and P values comparing the models were estimated. These analyses were performed using SAS, and 2-sided P < 0.05 was considered statistically significant. To further assess the relations between statistically significant metabolites (FDR P < 0.05), we also performed a hierarchical cluster analysis using the hclust function in R3.5.1. Results The study flow of participants is shown in Supplemental Figure 1. The age-adjusted characteristics of study participants according to their frequency of nut intake are shown in Table 1. Participants who had a higher intake of nuts tended to be older, have a lower BMI, and were more physically active than those with a lower intake. In addition, participants who had a higher intake of nuts had a higher AHEI score, indicative of better diet quality, and drank more alcohol than infrequent consumers. TABLE 1 Age-adjusted characteristics of 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study by frequency of nut consumption1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 1Values are means ± SDs unless otherwise indicated. MET-h, metabolic equivalent hours; NHS, Nurses’ Health Study. 2Not age-adjusted. 3Nut intake and alcohol were not included in the calculation. 4Values for blood concentrations of total cholesterol, LDL cholesterol, HDL cholesterol, and TGs were based on data from 362, 645, 645, and 910 participants, respectively, because only a subset of participants had data on both blood lipids (total cholesterol, LDL cholesterol, HDL cholesterol, and TGs) and blood lipid metabolites. View Large TABLE 1 Age-adjusted characteristics of 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study by frequency of nut consumption1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 Frequency of nut consumption (28-g serving) Characteristics Never < Once/week Once/week 2–4 times/wk ≥5 times/wk n 302 339 235 175 48 Female, % 82.4 80.9 74.4 72.2 70.6 Age at blood draw,2 y 53.2 ± 9.8 54.2 ± 9.9 56.4 ± 10.0 57.6 ± 10.1 59.8 ± 9.8 BMI, kg/m2 25.0 ± 4.3 25.2 ± 4.2 25.4 ± 4.1 25.0 ± 4.1 24.3 ± 3.4 Physical activity, MET-h/wk 21.2 ± 41.9 21.5 ± 24.0 25.3 ± 30.7 25.1 ± 30.0 28.0 ± 28.6 Smoking status, %  Never 55.9 52.3 55.7 52.1 62.5  Former 34.6 39.5 36.2 43.8 29.1  Current 9.5 8.2 8.1 4.1 8.4 Alcohol intake, g/d 5.4 ± 10.1 7.2 ± 12.5 8.0 ± 11.4 8.3 ± 13.6 8.8 ± 14.3 Alternate Healthy Eating Index3 44.9 ± 9.8 45.1 ± 9.5 45.0 ± 9.7 45.8 ± 8.8 46.6 ± 9.3 Total cholesterol,4 mg/dL 217 ± 46.6 218 ± 37.9 218 ± 46.7 218 ± 32.9 219 ± 27.3 LDL cholesterol,4 mg/dL 132 ± 35.7 133 ± 34.1 133 ± 40.0 128 ± 30.5 132 ± 23.6 HDL cholesterol,4 mg/dL 55.1 ± 17.4 58.6 ± 16.5 58.7 ± 15.7 58.1 ± 18.2 54.1 ± 10.6 TGs,4 mg/dL 126 ± 61.5 115 ± 56.1 121 ± 59.7 124 ± 81.8 118 ± 61.1 1Values are means ± SDs unless otherwise indicated. MET-h, metabolic equivalent hours; NHS, Nurses’ Health Study. 2Not age-adjusted. 3Nut intake and alcohol were not included in the calculation. 4Values for blood concentrations of total cholesterol, LDL cholesterol, HDL cholesterol, and TGs were based on data from 362, 645, 645, and 910 participants, respectively, because only a subset of participants had data on both blood lipids (total cholesterol, LDL cholesterol, HDL cholesterol, and TGs) and blood lipid metabolites. View Large Of the 144 known lipid metabolites, 17 were significantly associated with nut intake (FDR P < 0.05). Details of the individual metabolites along with effect sizes, which denote the change in relative metabolite level (expressed in number of SDs on the log scale) for 1 serving (28 g)/d of nut intake, are shown in Supplemental Table 1. Among the 17 metabolites that were significantly associated with nut intake, 9 were inversely associated [C34:3 diacylglycerol (DAG), C16:1 lysophosphatidylcholine (LPC), C16:1 cholesterol ester (CE), C32:1 DAG, C22:6 lysophosphatidylethanolamine (LPE), C22:6 LPC, C18:0 sphingomyelin (SM), C50:2 TG, and C34:2 DAG] and 8 were positively associated [C24:0 SM, C36:3 phosphatidylcholine (PC) plasmalogen-A, C36:2 PC plasmalogen, C24:0 ceramide, C36:1 PC plasmalogen, C22:0 SM, C34:1 PC plasmalogen, and C36:2 phosphatidylethanolamine plasmalogen]. Similar results were observed in the categorical analysis, which shows differences (95% CIs) in metabolite concentrations by frequency of nut consumption (Table 2) (P for trend <0.01). The majority of the metabolites were positively correlated with one another (Figure 1 and Supplemental Table 2). Intakes of peanuts and peanut butter were both positively associated with 3 metabolites (C24:0 SM, C24:0 ceramide, and C22:0 SM) and intake of other nuts was inversely associated with 6 metabolites (C16:1 LPC, C16:1 CE, C34:3 DAG, C22:6 LPC, C22:6 LPE, and C32:1 DAG) (Supplemental Table 3). FIGURE 1 View largeDownload slide Partial Spearman correlations between statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. The calculation is based on the z scores of log-transformed, continuous metabolite concentrations and adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. The individual correlation coefficients between metabolites are provided in Supplemental Table 2. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 1 View largeDownload slide Partial Spearman correlations between statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. The calculation is based on the z scores of log-transformed, continuous metabolite concentrations and adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. The individual correlation coefficients between metabolites are provided in Supplemental Table 2. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. TABLE 2 Differences (95% CIs) in relative metabolite levels by nut consumption among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study1 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 1Differences in least-squares means of relative metabolite levels (expressed in number of SDs on the log scale) between each higher intake category and the referent category. The models are adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. CE, cholesterol ester; DAG, diacylglycerol; HMDB, Human Metabolome Database; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; Ref, reference; SM, sphingomyelin. 2Nut intake as a continuous variable. View Large TABLE 2 Differences (95% CIs) in relative metabolite levels by nut consumption among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study1 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 Metabolite chemical name Metabolite HMDB No. Frequency of nut consumption Never < Once/week Once/week 2–4 times/wk ≥5 times/wk P-trend2 n — 302 339 235 175 48 — C24:0 SM HMDB11697 Ref 0.13 (−0.02, 0.28) 0.23 (0.06, 0.40) 0.25 (0.06, 0.44) 0.48 (0.17, 0.79) 0.0002 C36:3 PC plasmalogen HMDB11244 Ref 0.04 (−0.11, 0.20) 0.12 (−0.05, 0.29) 0.16 (−0.03, 0.34) 0.45 (0.14, 0.75) 0.0003 C36:2 PC plasmalogen HMDB11243 Ref 0.04 (−0.11, 0.19) 0.15 (−0.02, 0.31) 0.19 (0.00, 0.37) 0.41 (0.11, 0.72) 0.0003 C34:3 DAG HMDB07132 Ref 0.02 (−0.13, 0.17) −0.01 (−0.18, 0.15) −0.10 (−0.29, 0.08) −0.33 (−0.63, −0.03) 0.0003 C16:1 LPC HMDB10383 Ref 0.03 (−0.12, 0.19) −0.16 (−0.33, 0.01) −0.24 (−0.43, −0.05) −0.47 (−0.78, −0.16) 0.0004 C24:0 Ceramide HMDB04956 Ref 0.11 (−0.04, 0.26) 0.17 (0.00, 0.35) 0.24 (0.05, 0.43) 0.44 (0.13, 0.75) 0.0006 C36:1 PC plasmalogen HMDB11241 Ref 0.00 (−0.15, 0.15) 0.11 (−0.06, 0.28) 0.06 (−0.12, 0.24) 0.33 (0.03, 0.63) 0.0009 C16:1 CE HMDB00658 Ref −0.05 (−0.19, 0.09) −0.20 (−0.36, −0.04) −0.24 (−0.41, −0.06) −0.33 (−0.62, −0.04) 0.0010 C32:1 DAG HMDB07099 Ref 0.00 (−0.14, 0.15) −0.12 (−0.28, 0.05) −0.12 (−0.30, 0.06) −0.30 (−0.59, −0.01) 0.0010 C22:6 LPE HMDB11526 Ref −0.11 (−0.26, 0.04) −0.15 (−0.32, 0.01) −0.29 (−0.47, −0.10) −0.44 (−0.74, −0.14) 0.0011 C22:0 SM HMDB12103 Ref 0.12 (−0.03, 0.27) 0.18 (0.01, 0.36) 0.19 (−0.00, 0.38) 0.40 (0.09, 0.71) 0.0019 C34:1 PC plasmalogen HMDB11208 Ref −0.01 (−0.16, 0.14) 0.05 (−0.12, 0.23) 0.10 (−0.09, 0.29) 0.33 (0.03, 0.64) 0.0021 C22:6 LPC HMDB10404 Ref −0.09 (−0.24, 0.06) −0.18 (−0.34, −0.01) −0.36 (−0.54, −0.17) −0.42 (−0.72, −0.12) 0.0028 C18:0 SM HMDB01348 Ref −0.02 (−0.17, 0.13) −0.03 (−0.20, 0.14) −0.26 (−0.44, −0.07) −0.44 (−0.75, −0.13) 0.0034 C50:2 TG HMDB05377 Ref −0.04 (−0.18, 0.10) −0.15 (−0.31, 0.01) −0.11 (−0.29, 0.06) −0.26 (−0.55, 0.03) 0.0034 C34:2 DAG HMDB07103 Ref −0.01 (−0.15, 0.14) −0.09 (−0.25, 0.08) −0.11 (−0.29, 0.07) −0.22 (−0.52, 0.08) 0.0038 C36:2 PE plasmalogen HMDB09082 Ref 0.10 (−0.05, 0.25) 0.17 (0.00, 0.34) 0.24 (0.05, 0.43) 0.30 (−0.01, 0.61) 0.0047 1Differences in least-squares means of relative metabolite levels (expressed in number of SDs on the log scale) between each higher intake category and the referent category. The models are adjusted for age at blood draw, cohort, smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. CE, cholesterol ester; DAG, diacylglycerol; HMDB, Human Metabolome Database; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; Ref, reference; SM, sphingomyelin. 2Nut intake as a continuous variable. View Large Among the metabolites that were significantly associated with nut intake, hierarchical clustering analysis revealed 5 clusters of metabolites (Figure 2). Clusters 2 and 5 included metabolites that were positively associated with nut intake and were primarily comprised of very-long-chain (≥C22) SMs, ceramides, PC, and phosphatidylethanolamine plasmalogens. Clusters 1, 3, and 4 included metabolites that were inversely associated with nut intake and were comprised of TGs, DAGs, LPEs, LPCs, CEs, and SMs. Of note, the metabolites associated with intake of peanuts and peanut butter were specific to cluster 2. Adding the metabolites to the multivariate model improved prediction of nut intake beyond the other factors. The AUCs were 0.81 in the multivariate model with metabolites and 0.76 in the multivariate model, with a P value of 0.002 comparing these 2 models (Figure 3). FIGURE 2 View largeDownload slide Hierarchical clustering dendrogram of statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. (A) Hierarchical cluster analysis using the “hclust” function in R. The vertical axis represents the distance or dissimilarity between clusters. (B) Metabolites were grouped into 5 clusters by further using the “cutree” function in R. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 2 View largeDownload slide Hierarchical clustering dendrogram of statistically significant metabolites (false discovery rate P < 0.05) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study. (A) Hierarchical cluster analysis using the “hclust” function in R. The vertical axis represents the distance or dissimilarity between clusters. (B) Metabolites were grouped into 5 clusters by further using the “cutree” function in R. CE, cholesterol ester; DAG, diacylglycerol; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin. FIGURE 3 View largeDownload slide ROC curves comparing models to predict nut intake (never compared with ≥2 times/wk) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study (note: higher intake categories were combined in this analysis). Logistic regression models were fitted to predict nut intake (never compared with ≥2 times/wk) among 525 participants. Base model: adjusted for age at blood draw and cohort. MV model: also adjusted for smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. MV + metabolites model: also adjusted for the 17 statistically significant metabolites. P values were 0.0002 comparing the MV + metabolites model with the MV model; <0.0001 comparing the MV + metabolites model with the base model; and <0.0001 comparing the MV model with the base model. MV, multivariate; NHS, Nurses’ Health Study; ROC, receiver operating characteristic. FIGURE 3 View largeDownload slide ROC curves comparing models to predict nut intake (never compared with ≥2 times/wk) among 1099 diabetes-free women and men from the NHS, NHS II, and Health Professionals Follow-up Study (note: higher intake categories were combined in this analysis). Logistic regression models were fitted to predict nut intake (never compared with ≥2 times/wk) among 525 participants. Base model: adjusted for age at blood draw and cohort. MV model: also adjusted for smoking status, BMI, physical activity, alcohol intake, total energy intake, Alternate Healthy Eating Index, and, in women, menopausal status and postmenopausal hormone use. MV + metabolites model: also adjusted for the 17 statistically significant metabolites. P values were 0.0002 comparing the MV + metabolites model with the MV model; <0.0001 comparing the MV + metabolites model with the base model; and <0.0001 comparing the MV model with the base model. MV, multivariate; NHS, Nurses’ Health Study; ROC, receiver operating characteristic. Discussion In a well-defined sample of men and women from 3 large cohorts, we identified 17 out of 144 known plasma lipid metabolites that were significantly associated with nut intake. Associations differed in direction depending on lipid structure and nut type. To our knowledge, this is the first study to specifically examine the plasma lipid metabolite profile of total and type of nut consumption in a large observational study. To date, few studies have evaluated potential biomarkers of nut intake. These include intervention studies ranging from 12 wk to 6 mo that identified conjugated fatty acids along with serotonin metabolites and microbial-derived phenolic metabolites as markers of nut intake (18–20). In the PREDIMED trial, walnut consumption was characterized by 18 urinary metabolites, including markers of fatty acid metabolism, ellagitannin-derived microbial compounds, and metabolites of the tryptophan/serotonin pathway (21). In the majority of these trials, metabolites were measured in urine and may be more reflective of short-term or acute intake of nuts at prescribed doses. In contrast, the plasma metabolites measured in our study may be more reflective of long-term habitual nut intake patterns, which may partly explain the different lipidomic profile that emerged from our analysis. Differences in the study population, type of nuts, sample size, and metabolite profiling techniques might also account for differences with previous studies. Nuts are a nutrient-dense food with a complex matrix of bioactive compounds (4). Although the total fat content in nuts is high, ranging from 46% in cashews and pistachios to 76% in macadamia nuts, the saturated fat content is low, ranging from 4% to 16% (4). Most nuts contain a high proportion of MUFA [oleic acid (18:1n–9)]; however, certain nuts, such as Brazil nuts, have similar proportions of MUFAs and PUFAs [mostly linoleic acid (18:2n–6)], whereas walnuts contain mostly PUFAs, both from linoleic acid and α-linolenic acid (18:3n–3) (22). The fatty fraction of nuts also contains large amounts of noncholesterol sterols (i.e., phytosterols) (23), which play an important structural role in membranes, where they stabilize phospholipid bilayers (24). Nuts also contain choline, which is found in SMs, PC, and LPC (25). Thus, nuts may have a unique lipidomic signature that can be used as a biomarker for intake. In our analysis, 2 SMs (C24:0 and C22:0) and a ceramide (C24:0) derived from very-long-chain SFAs (VLCSFAs) were positively associated with nut and peanut intake. Ceramides and SMs are lipid molecules with structural and signaling roles in cell membranes. However, most ceramides have been linked to increased cardiometabolic risk (26–28) and C24:0 ceramide has been found to promote insulin resistance in rodents (29). Exposure to VLCSFAs can promote ceramide formation in vitro and in animal models but how this relates to diet is not well understood (30). We previously found inverse associations between 3 plasma VLCSFAs (C20:0, C22:0, and C24:0) and risk of coronary artery disease in the NHS and HPFS (31), which may support a beneficial role of ceramides from these VLCSFAs on cardiometabolic health. This suggests that nut intake may have a beneficial effect on cardiometabolic health through a positive association with ceramide (C24:0). The other metabolites that were positively associated with nut intake in our analysis include plasmalogens, which are plasmenyl-phospholipids with a vinyl ether linkage at the sn-1 position and PUFA linkage at the sn-2 position, thought to protect mammalian cells against reactive oxygen species. Metabolites that were inversely associated with nut intake include TG, DAG, and CEs, consistent with the favorable effects of nut intake on total cholesterol, LDL cholesterol, apoB, and TGs observed in trials (7). Whether these metabolites are biologically meaningful is not clear (17). Our study has limitations. It is difficult to know whether the metabolites we identified are truly indicative of nut intake or endogenous production. In addition, because lipid data were expressed at the sum composition level, we were not able to determine concentrations of specific fatty acids and how they directly relate to nut intake. Although we adjusted for multiple potential confounders, it is not possible to rule out residual confounding from unmeasured or poorly measured factors related to diet and lifestyle. Measurement error in dietary assessment using FFQs is also inevitable and may lead to an attenuation of diet–metabolite associations (13). The strengths of our study include a large sample size and detailed diet and lifestyle information, which facilitated fine control for potential confounding. Using mean nut intake from 2 FFQs reduced within-person variability and better represented habitual intake. In analyzing the lipid metabolites, we controlled for multiple testing and ensured that all metabolites had CVs ≤25%, as measured in blinded quality control samples, and detectable concentrations in ≥90% of participants. In conclusion, we identified a set of lipid metabolites from known species that were associated with nut intake. Associations differed by type of lipid molecule and type of nut. These findings may help identify biomarkers of nut intake and provide insight into biological mechanisms underlying associations between nuts and cardiometabolic health. However, replication of our findings in additional studies of the metabolome and among diverse populations is needed. Acknowledgments The authors’ responsibilities were as follows—VSM and YB: designed the analysis, interpreted the data, wrote the manuscript, and had primary responsibility for the final content; YB: conducted the analysis; SST, EWK, KHC, AA, KMW, and LAM: provided access to the data for the analysis; MG-F, FBH, MKT, OAZ, AHE, SST, EWK, KHC, AA, KMW, LAM, ELG, and CSF: critically reviewed the manuscript for important intellectual content; and all authors: read and approved the final manuscript. Notes Supported by NIH grants UM1 CA186107, UM1 CA176726, UM1 CA167552, U01 167552, P01 CA87969, R01 AR049880, R01 CA49449, R01 CA67262, R01 CA50385, P50 CA090381, U54CA155626 (to FBH), P30DK046200, K01 HL125698 (to MKT), and KL2 TR001100; Department of Defense grants W81XWH-13-1-0493 and CA150357 (to YB); and by a grant from the International Tree Nut Council Nutrition Research & Education Foundation (to YB). Author disclosures: YB received a research grant from the International Tree Nut Council Nutrition Research & Education Foundation. VSM received research support from the Peanut Institute. MG-F, FBH, MKT, OAZ, AHE, SST, EWK, KHC, AA, KMW, LAM, ELG, and CSF, no conflicts of interest. The funders of this study had no role in its design or conduct; in the collection, management, analysis, or interpretation of the data; or in the preparation, review, or approval of the manuscript. Supplemental Figure 1 and Supplemental Tables 1–3 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/. Abbreviations used: AHEI, Alternate Healthy Eating Index; CE, cholesterol ester; DAG, diacylglycerol; FDR, false discovery rate; HPFS, Health Professionals Follow-up Study; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; NHS, Nurses’ Health Study; PC, phosphatidylcholine; SM, sphingomyelin; VLCSFA, very-long-chain SFA. References 1. Aune D , Keum N , Giovannucci E , Fadnes LT , Boffetta P , Greenwood DC , Tonstad S , Vatten LJ , Riboli E , Norat T . Nut consumption and risk of cardiovascular disease, total cancer, all-cause and cause-specific mortality: a systematic review and dose-response meta-analysis of prospective studies . BMC Med . 2016 ; 14 ( 1 ): 207 . Google Scholar Crossref Search ADS PubMed 2. Yu Z , Malik VS , Keum N , Hu FB , Giovannucci EL , Stampfer MJ , Willett WC , Fuchs CS , Bao Y . Associations between nut consumption and inflammatory biomarkers . Am J Clin Nutr . 2016 ; 104 ( 3 ): 722 – 8 . Google Scholar Crossref Search ADS PubMed 3. Guasch-Ferré M , Liu X , Malik VS , Sun Q , Willett WC , Manson JE , Rexrode KM , Li Y , Hu FB , Bhupathiraju SN . Nut consumption and risk of cardiovascular disease . J Am Coll Cardiol . 2017 ; 70 ( 20 ): 2519 – 32 . Google Scholar Crossref Search ADS PubMed 4. Ros E . Health benefits of nut consumption . Nutrients . 2010 ; 2 ( 7 ): 652 – 82 . Google Scholar Crossref Search ADS PubMed 5. Pan A , Sun Q , Manson JE , Willett WC , Hu FB . Walnut consumption is associated with lower risk of type 2 diabetes in women . J Nutr . 2013 ; 143 ( 4 ): 512 – 8 . Google Scholar Crossref Search ADS PubMed 6. Estruch R , Ros E , Salas-Salvadó J , Covas MI , Corella D , Arós F , Gómez-Gracia E , Ruiz-Gutiérrez V , Fiol M , Lapetra J et al. . Primary prevention of cardiovascular disease with a Mediterranean diet . N Engl J Med . 2013 ; 368 ( 14 ): 1279 – 90 . Google Scholar Crossref Search ADS PubMed 7. Del Gobbo LC , Falk MC , Feldman R , Lewis K , Mozaffarian D . Effects of tree nuts on blood lipids, apolipoproteins, and blood pressure: systematic review, meta-analysis, and dose-response of 61 controlled intervention trials . Am J Clin Nutr . 2015 ; 102 ( 6 ): 1347 – 56 . Google Scholar Crossref Search ADS PubMed 8. Satija A , Yu E , Willett WC , Hu FB . Understanding nutritional epidemiology and its role in policy . Adv Nutr . 2015 ; 6 ( 1 ): 5 – 18 . Google Scholar Crossref Search ADS PubMed 9. Guasch-Ferre M , Bhupathiraju SN , Hu FB . Use of metabolomics in improving assessment of dietary intake . Clin Chem . 2018 ; 64 ( 1 ): 82 – 98 . Google Scholar Crossref Search ADS PubMed 10. Hunter DJ , Hankinson SE , Hough H , Gertig DM , Garcia-Closas M , Spiegelman D , Manson JE , Colditz GA , Willett WC , Speizer FE et al. . A prospective study of NAT2 acetylation genotype, cigarette smoking, and risk of breast cancer . Carcinogenesis . 1997 ; 18 ( 11 ): 2127 – 32 . Google Scholar Crossref Search ADS PubMed 11. Hankinson SE , Willett WC , Manson JE , Colditz GA , Hunter DJ , Spiegelman D , Barbieri RL , Speizer FE . Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women . J Natl Cancer Inst . 1998 ; 90 ( 17 ): 1292 – 9 . Google Scholar Crossref Search ADS PubMed 12. Chu NF , Spiegelman D , Yu J , Rifai N , Hotamisligil GS , Rimm EB . Plasma leptin concentrations and four-year weight gain among US men . Int J Obes Relat Metab Disord . 2001 ; 25 ( 3 ): 346 – 53 . Google Scholar Crossref Search ADS PubMed 13. Willett WC , Sampson L , Stampfer MJ , Rosner B , Bain C , Witschi J , Hennekens CH , Speizer FE . Reproducibility and validity of a semiquantitative food frequency questionnaire . Am J Epidemiol . 1985 ; 122 ( 1 ): 51 – 65 . Google Scholar Crossref Search ADS PubMed 14. Salvini S , Hunter DJ , Sampson L , Stampfer MJ , Colditz GA , Rosner B , Willett WC . Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption . Int J Epidemiol . 1989 ; 18 ( 4 ): 858 – 67 . Google Scholar Crossref Search ADS PubMed 15. Chiuve SE , Fung TT , Rimm EB , Hu FB , McCullough ML , Wang M , Stampfer MJ , Willett WC . Alternative dietary indices both strongly predict risk of chronic disease . J Nutr . 2012 ; 142 ( 6 ): 1009 – 18 . Google Scholar Crossref Search ADS PubMed 16. Mayers JR , Wu C , Clish CB , Kraft P , Torrence ME , Fiske BP , Yuan C , Bao Y , Townsend MK , Tworoger SS et al. . Elevation of circulating branched-chain amino acids is an early event in human pancreatic adenocarcinoma development . Nat Med . 2014 ; 20 ( 10 ): 1193 – 8 . Google Scholar Crossref Search ADS PubMed 17. Toledo E , Wang DD , Ruiz-Canela M , Clish CB , Razquin C , Zheng Y , Guasch-Ferre M , Hruby A , Corella D , Gomez-Gracia E et al. . Plasma lipidomic profiles and cardiovascular events in a randomized intervention trial with the Mediterranean diet . Am J Clin Nutr . 2017 ; 106 ( 4 ): 973 – 83 . Google Scholar PubMed 18. Tulipani S , Llorach R , Jáuregui O , López-Uriarte P , Garcia-Aloy M , Bullo M , Salas-Salvadó J , Andrés-Lacueva C . Metabolomics unveils urinary changes in subjects with metabolic syndrome following 12-week nut consumption . J Proteome Res . 2011 ; 10 ( 11 ): 5047 – 58 . Google Scholar Crossref Search ADS PubMed 19. Andersen MB , Kristensen M , Manach C , Pujos-Guillot E , Poulsen SK , Larsen TM , Astrup A , Dragsted L . Discovery and validation of urinary exposure markers for different plant foods by untargeted metabolomics . Anal Bioanal Chem . 2014 ; 406 ( 7 ): 1829 – 44 . Google Scholar Crossref Search ADS PubMed 20. Mora-Cubillos X , Tulipani S , Garcia-Aloy M , Bulló M , Tinahones FJ , Andres-Lacueva C . Plasma metabolomic biomarkers of mixed nuts exposure inversely correlate with severity of metabolic syndrome . Mol Nutr Food Res . 2015 ; 59 ( 12 ): 2480 – 90 . Google Scholar Crossref Search ADS PubMed 21. Garcia-Aloy M , Llorach R , Urpi-Sarda M , Tulipani S , Estruch R , Martínez-González MA , Corella D , Fitó M , Ros E , Salas-Salvadó J et al. . Novel multimetabolite prediction of walnut consumption by a urinary biomarker model in a free-living population: the PREDIMED study . J Proteome Res . 2014 ; 13 ( 7 ): 3476 – 83 . Google Scholar Crossref Search ADS PubMed 22. Ros E , Mataix J . Fatty acid composition of nuts – implications for cardiovascular health . Br J Nutr . 2006 ; 96 ( Suppl 2 ): S29 – 35 . Google Scholar Crossref Search ADS PubMed 23. Segura R , Javierre C , Lizarraga MA , Ros E . Other relevant components of nuts: phytosterols, folate and minerals . Br J Nutr . 2006 ; 96 ( Suppl 2 ): S36 – 44 . Google Scholar Crossref Search ADS PubMed 24. Hartmann MA , . Plant sterols and the membrane environment . Trends Plant Sci . 1998 ; 3 : 170 – 175 . Google Scholar Crossref Search ADS 25. Li Z , Vance DE . Phosphatidylcholine and choline homeostasis . J Lipid Res . 2008 ; 49 ( 6 ): 1187 – 94 . Google Scholar Crossref Search ADS PubMed 26. Chavez JA , Summers SA . A ceramide-centric view of insulin resistance . Cell Metab . 2012 ; 15 ( 5 ): 585 – 94 . Google Scholar Crossref Search ADS PubMed 27. Holland WL , Summers SA . Sphingolipids, insulin resistance, and metabolic disease: new insights from in vivo manipulation of sphingolipid metabolism . Endocr Rev . 2008 ; 29 ( 4 ): 381 – 402 . Google Scholar Crossref Search ADS PubMed 28. Laaksonen R , Ekroos K , Sysi-Aho M , Hilvo M , Vihervaara T , Kauhanen D , Suoniemi M , Hurme R , Marz W , Scharnagl H et al. . Plasma ceramides predict cardiovascular death in patients with stable coronary artery disease and acute coronary syndromes beyond LDL-cholesterol . Eur Heart J . 2016 ; 37 ( 25 ): 1967 – 76 . Google Scholar Crossref Search ADS PubMed 29. Kirwan JP . Plasma ceramides target skeletal muscle in type 2 diabetes . Diabetes . 2013 ; 62 ( 2 ): 352 – 4 . Google Scholar Crossref Search ADS PubMed 30. Chavez JA , Summers SA . Characterizing the effects of saturated fatty acids on insulin signaling and ceramide and diacylglycerol accumulation in 3T3-L1 adipocytes and C2C12 myotubes . Arch Biochem Biophys . 2003 ; 419 ( 2 ): 101 – 9 . . Google Scholar Crossref Search ADS PubMed 31. Malik VS , Chiuve SE , Campos H , Rimm EB , Mozaffarian D , Hu FB , Sun Q . Circulating very-long-chain saturated fatty acids and incident coronary heart disease in US men and women . Circulation . 2015 ; 132 ( 4 ): 260 – 8 . Google Scholar Crossref Search ADS PubMed Copyright © American Society for Nutrition 2019. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Journal of NutritionOxford University Press

Published: Nov 9, 21

References

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 folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off