Abstract
OBJECTIVES: The objective of this study is to outline a framework for how fatty acids may be acting as mediators/suppressors and/or moderators of an individual's genetic predisposition for cardiovascular lipid biomarkers. METHODS: All UK Biobank participants with demographic and lifestyle variables, circulating cardiovascular lipids, and NMR-measured fatty acid data collected at the baseline visit (N = 229,859) were included in analyses. We fit four separate linear regression models, one for each of the following common measures of cardiovascular lipids: total cholesterol, HDL-c, LDL-c, and total triglycerides. Each model predicted cardiovascular lipids by an individual's FADS (a well-known fatty acid desaturase gene complex) haplotype, with the addition of individual ω-3 (DHA, non-DHA, and total), ω-6 (LA, non-LA, and total), or SFA factors as additive (mediation/suppression) or using an interaction term (with FADS) (moderation). All models were adjusted for a wide range of demographic and medical history variables and evaluated against a Bonferroni-adjusted significance level (p < 8.9 × 10(-4)). RESULTS: Across 56 models (four lipids × seven FAs × two conceptual models (mediation/suppression and moderation)), we found evidence of 19 moderation, 12 mediation, and 16 suppression effects of the FADS-lipid relationship. For example, adjusting for circulating DHA levels as a mediator, the association of the genotype with HDL-c substantially lessened for both minor genotypes reflecting >122% mediation of the association of FADS by DHA. Additionally, we found evidence that LDL-c is moderated, to some extent, by all fatty acid measures. CONCLUSIONS: This analysis demonstrates that an individual's fatty acid profile can act as a mediator/suppressor or moderator of the association of the FADS genotype and various cardiovascular biomarkers. Future work is necessary to expand this cross-sectional examination to determine directionality and temporality of the mediation and moderation evidence presented. This research has been conducted using the UK Biobank Resource under Application Number 85092.