Figure 5.

Validation in an independent cohort (the Women’s Health Study (WHS)) of significant (FDR < 0.05) associations between one-year changes in FAs, oxylipins and bioactive lipids with n-3 treatment in VITAL200. Associations between baseline levels with baseline nutritional n-3 intake in the WHS were used for validation. Only compounds with putative annotation using the LCMS assay are presented (LCMS assay—high-throughput liquid chromatography–mass spectrometry of circulating free FAs, oxylipins and bioactive lipids in plasma). Symbols with error bars are beta coefficients per 1SD with 95% confidence intervals. In VITAL200, longitudinal associations are modeled in one-way ANOVA with treatment group assignment as a predictor variable (sex and age were not significant and removed from the model). In the WHS, cross-sectional associations in linear regression were adjusted for age, sex and cardiovascular risk factors (high-density lipoprotein cholesterol (HDLC), total cholesterol, systolic blood pressure, treatment for hypertension, diabetes, current smoking). LCMS bioactive lipid features in VITAL200 were log transformed and their one-year changes were shifted to median 0 and standard deviation (SD) 1. In the WHS, batch effect-corrected baseline levels were log transformed and shifted and rescaled to median 0, SD1.