Plasma levels of the omega-3 fatty acid docosahexaenoic acid, show an inverse threshold association with cardiovascular events

Background Blood levels of omega-3 fatty acids, specifically eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), are associated with lower risk of cardiovascular disease (CVD) mortality and coronary artery disease (CAD) events, but long-term associations with total CVD (i.e. non-fatal + fatal events across multiple arterial territories) remain unclear. We therefore assessed magnitude and shape of long-term associations of blood EPA, DHA, and the sum of EPA and DHA with total CVD. Methods We included 6,464 participants without baseline CVD from the EPIC-Norfolk prospective population study (N = 25,639, recruited between 1993-1997 at age 45-79 years), for whom EPA and DHA plasma levels (i.e. mol% of total fatty acids) were available. The endpoint was first hospitalization or death due to CAD, stroke, peripheral artery disease, or aortic aneurysm, with follow-up data until March 2024. We applied multivariate Cox proportional hazards models to estimate adjusted hazard ratios (aHR) for total CVD, comparing lowest and highest quintiles of EPA + DHA (i.e., the sum of both), EPA alone and DHA alone. Non-linearity was assessed using restricted cubic splines with four knots, with the 10th percentile as reference. Likelihood ratio tests were used to compare spline models with linear models to obtain P values (p) for non-linearity. Models were adjusted for age, sex, ethnicity, physical activity, smoking, LDL cholesterol, systolic blood pressure, BMI, diabetes status, and omega-6 fatty acid levels. Results Among 6,464 participants with known EPA and DHA plasma levels (52% women, mean age 63 years [SD 8]), 2,981 experienced a first CVD event (46% of the cohort, 27.2 first events per 1,000 person-years) during a median follow-up of 17.3 years (IQR 8.5–26.5). DHA showed an inverse association with total CVD (aHR 0.86, 95% CI 0.76–0.97) and a non-linear pattern (p for non-linearity: 0.003), characterized by a steep decline in risk at lower DHA levels and a plateau from the 37th percentile (4.4 mol%) onward. EPA and the sum of EPA + DHA both showed no association with total CVD (EPA: aHR 0.97, 95% CI 0.84–1.13, p for non-linearity: 0.985; EPA+DHA: aHR 0.89, 95% CI 0.78–1.02, p for non-linearity: 0.179) (Fig. 1). Conclusion In apparently healthy individuals, the highest quintile of plasma DHA levels is associated with a 14% lower risk of CVD, with a threshold-type, non-linear association. This pattern was absent for EPA, but the sum of EPA + DHA levels showed a trend towards lower total CVD risk. These findings suggest that plasma DHA levels may play a more prominent role than EPA in the etiology of total CVD.For image description, please refer to the figure legend and surrounding text.