Introduction: In animal models of afterload stress, eicosapentaenoic acid (EPA) prevents interstitial myocardial fibrosis and preserves diastolic dysfunction.
Hypothesis: We hypothesized that EPA is similarly protective in humans.
Methods: In the Multi-Ethnic Study of Atherosclerosis, we tested for an effect of plasma phospholipid EPA percent of total fatty acids on primary heart failure incidence across heart failure types including, heart failure with reduced ejection fraction (HFrEF; <45% EF), and with preserved ejection fraction (HFpEF) using Cox proportional hazards modeling in 6566 subjects. Results are mean [95% CI].
Results: A total of 6566 subjects had measured baseline EPA, including 1797 black, 794 Chinese, 1444 Hispanic, and 2531 white participants; 52% were female. Over a median follow-up of 13.0 years, 293 heart failure events occurred in subjects with measured EPA: 129 had HFrEF, 110 had HFpEF, and the remaining 54 had unknown ejection fraction status. Mean EPA in HF-free subjects was 0.77% (0.76 - 0.79), and was lower in heart failure subjects 0.70% (0.65 - 0.74), p=0.002. EPA was associated with lower incidence of HF, having a hazard ratio of 0.73 (0.60 - 0.89) per unit change in percentage of total fatty acids, p=0.001. Adjusting for age, sex, race, BMI, smoking, diabetes mellitus, blood pressure, lipids and lipid-lowering drugs, and albuminuria did not change this relationship. Sensitivity analysis showed no dependence on heart failure type. Adjusting for other fatty acids with clustering did not change hazards. In animals, >2.5% EPA is required for prevention of HF, between 2.5% and 1% is marginal, and <1% EPA is insufficient. Most subjects had insufficient EPA levels (n=4794), fewer had marginal levels (n=1471), and fewer still had sufficient levels (n=301). Subjects with sufficient EPA levels were at 0.40 (0.15 - 0.81) fold risk compared to insufficient EPA (p=0.008).
Conclusion: High abundance of EPA is robustly associated with reduced risk for heart failure, independent of established risk factors and regardless of ejection fraction status.