Abstract MP26: Biomarkers of n-6 Polyunsaturated Fatty Acids and CVD Risk: A Global Pooling Project of 19 Cohort Studies

Abstract

Background: Despite cholesterol-lowering and other potential beneficial properties of dietary linoleic acid (LA; 18:2n-6), LA has been considered pro-inflammatory and is a precursor to arachidonic acid (AA; 20:4n-6), which has also been considered harmful for CVD. Thus, the role of n-6 fatty acids in CVD prevention remains debated. Objective: To investigate across harmonized global studies the relation of circulating/tissue biomarker levels of LA and AA with incidence of CVD, CHD, and stroke. Methods: We conducted a pooled analysis within a global consortium of prospective cohort or nested case-control studies having circulating or tissue biomarker measures of LA and AA. Each study ascertained CVD risk, including incident CVD, CHD, ischemic stroke (IS), and/or CVD mortality in adults with no prevalent CVD at baseline. Standardized individual-level analysis was conducted in each study using pre-specified models, exposures, outcomes, and covariates. Study-specific estimates were pooled using inverse variance-weighted fixed effects meta-analysis. Results: We evaluated 19 studies, including 6,569 total CVD events, 2,407 CVD deaths, 6,557 CHD events, and 2,237 IS events (values may not sum due to cohort-specific outcomes). In continuous (90th vs 10th percentile) multivariable-adjusted analyses, higher levels of both LA and AA were associated with lower risk of incident CVD, with modest between-compartment heterogeneity (I2=33-35%) (Figure). LA and AA were inversely associated with CVD mortality, with RRs (95% CI): LA 0.81 (0.73 - 0.90) and AA 0.90 (0.81 - 0.99). LA, but not AA, was associated with lower risk of incident CHD and IS, with RRs (95% CI) 0.94 (0.88 - 0.99) and 0.80 (0.69 - 0.92), respectively. Conclusions: Based on harmonized analysis of multiple studies on free-living populations, biomarker levels of the two major n-6 fatty acids are associated with lower risk of CVD incidence and mortality. Relevance of potential differences in compartment-specific associations requires further investigation.