The role of circulating polyunsaturated fatty acids in atrial fibrillation and atrial cardiomyopathy: a machine learning-informed structural modelling approach

AIMS: Beyond their cardioprotective effects, polyunsaturated fatty acids (PUFAs) have also been linked to arrhythmogenicity, raising concern among individuals predisposed to atrial fibrillation (AF), including patients with heart failure with preserved ejection fraction (HFpEF). Here, we examined associations of circulating PUFA profiles and PUFA supplementation with AF prevalence and incidence. METHODS AND RESULTS: We analysed 483,372 individuals from the UK Biobank (8,503 prevalent AF; 32,671 incident AF) with nuclear magnetic resonance (NMR) metabolomics, combining systematic deconfounding, an explainable supervised classifier and generalised structural equation modelling (GSEM).Anthropometric, demographic and genetic risk factors were stronger correlates of AF prevalence and incidence than lipidomic profiles. NMR analysis indicated lower plasma PUFA concentrations in AF, atrial cardiomyopathy (AtCM) and in HFpEF, most pronounced with cholesterol-lowering medication (CLM) use and partially attenuated in participants reporting PUFA supplementation. Higher ω-6 concentrations were associated with lower AF prevalence and incidence; for AF incidence, β=-0.083 (95% CI -0.097 to -0.069; p<0.001), higher ω-3 with lower AF incidence (β=-0.093, 95% CI -0.122 to -0.065; p<0.001), and higher DHA with higher incidence (β=0.058, 95% CI 0.031 to 0.085; p<0.001), an association not evident in women. Supplementation was associated with higher circulating PUFAs and lower AF prevalence (β=-0.096, 95% CI -0.145 to -0.047; p<0.001), strongest among CLM users and women. Subgroup findings in HFpEF were directionally similar. CONCLUSIONS: Circulating PUFAs were lower in AF, AtCM and HFpEF. Lipid profiles indicated favourable associations for circulating ω-3 and ω-6, with a modest adverse signal for DHA. After adjustment for established risk factors, PUFA supplementation was associated with lower AF prevalence, including in patients with CLM and those with HFpEF. Formulations that substantially increase plasma DHA levels warrant careful evaluation given the observed association with AF incidence.