Abstract P524: Protein Biomarkers Mediate the Association Between Diet Quality and Incident Cardiovascular Disease and All-cause Mortality

Jiantao Ma,Sophia Lu,Jennifer E Ho,Chen Yao,Tianxiao Huan,Daniel Levy

DOI: https://doi.org/10.1161/circ.141.suppl_1.p524

IF: 37.8

2020-03-03

Circulation

Abstract:Objectives: Molecular mechanism underlying diet and cardiovascular disease (CVD) remains unclear. We aimed to examine the association between habitual diet and protein biomarkers and test whether these proteins mediate the association between diet and incident CVD and mortality. Methods: We included 6,596 participants (mean age 50 years; 54% women) of the Framingham Heart Study. A panel of 71 CVD-related circulating protein biomarkers were measured using a modified Sandwich ELISA method, multiplexed on a Luminex xMAP platform. We used three diet scores to represent overall diet quality: Alternate Health Eating Index (AHEI), a Mediterranean-style Diet Score (MDS), and the Dietary Approaches to Stop Hypertension (DASH). Primary outcomes were incident CVD (coronary heart disease, myocardial infarction, stroke, and others), CVD death, and all-cause mortality with median follow-up 14 years. Results: After adjusting for sex, age, and energy intake, 40 proteins were associated with at least one diet score at P <0.0007 (Bonferroni correction for 71 proteins) and with P <0.05 for the other diet scores. Additional adjustments for smoking status, physical activity level, and alcohol intake did not materially change these associations. Further adjustment for BMI attenuated association; however, the association remained significant for 36 of 40 proteins (P <0.001 with at least one diet score). Pathway analysis using coding genes of the 36 diet-associated proteins showed, after Bonferroni correction, enrichment of 157 Gene Ontology (GO) biological processes such as regulation of immune response (GO:0050776; P=0.003) and regulation of vasoconstriction (GO:0019229; P=0.03). All three diet scores were significantly associated with incident CVD and all-cause mortality. Among the 36 diet-associated proteins, four proteins (GDF15, UCMGP, ADM, and sICAM1) significantly mediated the association between at least one diet score and all-cause mortality at P <0.001 and with P <0.05 for other diet scores, after adjustment for covariates including BMI and cardiometabolic risk factors at baseline. GDF15 significantly mediated the association of diet scores and incident CVD (fatal and nonfatal). The proportion of mediation by GDF15 ranged from 11% to 21% for diet scores and incident CVD and from 30% to 34% for diet scores and all-cause mortality. In addition, we observed that nine proteins were nominally significant in mediation analyses for incident CVD or all-cause mortality (0.001< P <0.05). Conclusions: The present study demonstrates that diet quality is associated with 36 of 71 selected CVD-related circulating protein markers. Our study sheds light into potential pathways underlying the association of diet and CVD and mortality.

cardiac & cardiovascular systems,peripheral vascular disease