Abstract P2113: Metabolomic signatures of Puerto Rican dietary patterns and associations with cardiometabolic risk

Abstract

Introduction: Puerto Ricans on the US mainland have poor diet quality linked to adverse cardiometabolic risk, but underlying mechanisms are unclear. Hypothesis: Plasma metabolomic signatures reflect Puerto Rican diet patterns and are associated with cardiometabolic risk independent of diet. Methods: We used LC/MS to measure 714 plasma metabolites in 722 participants (45-75 y) [n=372 with type 2 diabetes (T2D); n=346 without T2D; n=4 with T2D missing] in the Boston Puerto Rican Health Study. Diet was assessed using an ethnic-specific validated food frequency questionnaire (FFQ). We identified diet patterns using principle component analysis (PCA) and diet-related metabolomic signatures via elastic net regression. We assessed cross-sectional and prospective associations of baseline metabolomic signatures with baseline and 5-year changes in cardiometabolic risk using linear regression, adjusting for confounders and FFQ/PCA derived diet scores. Baseline risk factors were log-transformed and back-transformed for reporting. We applied a false discovery rate (FDR<0.05) for multiple testing correction. We tested for interaction by T2D status (p<0.05). Results: We identified three distinct diet patterns: oils, rice, and beans (traditional); meat, processed meat, and French fries (meat/fries); and sweetened beverages, candy, and soft drinks (sweets). Metabolomic signatures for meat/fries (47 metabolites, r=0.39-0.43) and sweets (51 metabolites, r=0.15-0.34), but not traditional (45 metabolites, r=0.08-0.13) pattern, were significantly correlated (p<0.05) with their respective FFQ/PCA derived diet scores. Cross-sectionally, the meat/fries metabolomic signature (per SD) was associated with a 1.2% (95% CI: 0.5% to 2.0%) higher waist circumference (WC, cm) among overall participants, a 6% (3%-10%) higher homocysteine (umol/L) only among those without T2D, and a 10% (5%-16%) higher glucose (mg/dl) among those with T2D. The sweets signature (per SD) was linked to a 9% (4%-13%) higher low-density lipoprotein cholesterol (mg/dl) and a 4% (2%-6%) lower glycated hemoglobin (%) only in those with T2D. Prospectively, the meat/fries metabolomic signature (per SD) was positively associated with a mean increase in WC over time [β=0.87, 95%CI (0.37-1.38)] in those without T2D. All FDRs were <0.05. Conclusions: Metabolomic signatures for meat/fries and sweets diet patterns identified among Puerto Ricans were associated with distinct cardiometabolic risk profiles, varying by T2D status.