Abstract P057: Proteomic Signature Of Healthy Dietary Patterns And Risk Of Cardiovascular Disease In The Framingham Heart Study

Maura E Walker,Vanessa Xanthakis,Honghuang Lin,Paul Jacques,Vasan S Ramachandran

DOI: https://doi.org/10.1161/circ.145.suppl_1.p057

IF: 37.8

2022-03-01

Circulation

Abstract:Background: Identifying circulating proteins associated with healthy dietary patterns may offer novel insights into the key biological processes underlying diet-related chronic disease. As a first step toward achieving this goal, we examined the relations between a proteomic score of healthy dietary patterns and the risk of cardiovascular disease (CVD). Methods: We evaluated up to 1713 Framingham Offspring Study participants (1991-1995, mean age 54yr, 55% women) who had plasma proteins assayed using the SomaLogic platform and diet assessed using a validated food frequency questionnaire. We regressed three healthy dietary pattern indices (Alternative Healthy Eating Index, Dietary Approaches to Stop Hypertension [DASH], and Mediterranean-style diet) on 103 diet-related proteins using elastic net regression (separate model for each diet index) with a 10-fold cross-validation (CV). Protein selection was based on the minimum mean CV error. A proteomic score was created by summing (concentration*β coefficient) of selected proteins, and validated in 862 Framingham Third Generation Cohort participants (2008-2011, mean age 46yr, 56% women). We used multivariable Cox proportional hazard models to relate the proteomic score and dietary pattern indices (separate models) to CVD in participants free of CVD (n=2459). All models adjusted for age, sex, smoking, and lipid medication use. Additional models adjusted for 1) both the proteomic score and dietary pattern indices and 2) body mass index, total cholesterol to high-density lipoprotein cholesterol ratio, triglycerides, diabetes mellitus, and hypertension. Results: Across three dietary patterns, 11 proteins characterized the proteomic score. Partial correlations of the proteomic score with the healthy dietary patterns ranged from r=0.33-0.36 (adjusted for age, sex, and total energy) in the validation sample, with the strongest correlation for the DASH index. Over a mean follow-up of 20 years, 510 participants developed CVD. Adjusting for age, sex, smoking, and lipid medication, a 1-SD unit increase in the proteomic score was associated with a 22% lower risk of CVD (hazard ratio [95% CI] 0.78 [0.72, 0.84]) and a 1-SD unit increase in the DASH index was associated with a 15% lower risk of CVD (hazard ratio 0.85 [0.77, 0.93). In mutually-adjusted models, relations of the proteomic score with CVD (0.80 [0.73, 0.87]), and of the DASH index with CVD were maintained (0.89 [0.82, 0.97]). Upon further adjustment for metabolic risk factors, relations of the proteomic score and CVD were attenuated (0.96 [0.87, 1.05]), but relations of the DASH index with CVD were unchanged (0.89 [0.80, 0.98]). Conclusion: We characterized a distinct proteomic score of healthy dietary patterns in a community-based cohort of adults. A healthy diet proteomic score may represent variation in the proteome that influences CVD risk.

cardiac & cardiovascular systems,peripheral vascular disease