LAUSR

BACKGROUND Plasma trimethylamine-N-oxide (TMAO) is associated with cardiovascular disease; however specific relationships with cardiac arrhythmias are unknown. We evaluated the association between plasma TMAO and incident atrial fibrillation (AF). METHODS Risk associations were explored among 3797 patients with suspected stable angina in the Western Norway Coronary Angiography Cohort (WECAC) and verified in 3143 elderly participants in the community-based Hordaland Health Study (HUSK). Information on endpoints was obtained from nationwide registries. RESULTS Median follow-up was 7.3 and 10.8 years in the WECAC and HUSK cohorts, respectively, and 412 (10.9%) and 484 (15.4%) subjects were registered with incident AF. The age and gender adjusted HRs were 1.16, 95% CI 1.05-1.28 and 1.10, 95% CI 1.004-1.19 per 1 SD increase in log-transformed plasma TMAO. Adjusting for hypertension, BMI, smoking, diabetes, or intake of total choline, a TMAO precursor, did not materially influence the risk associations. Among patients in WECAC, further extensive adjustment for other AF risk factors yielded similar results. Adding TMAO to traditional AF risk factors (age, gender, hypertension, BMI, smoking and diabetes) yielded a continuous net reclassification improvement of 0.108, 95% CI 0.015-0.202 and 0.139, 95% CI 0.042-0.235. CONCLUSIONS Plasma TMAO was associated with and improved reclassification of incident AF in two independent Norwegian cohorts with long-term follow-up. The relationship was independent of traditional AF risk factors, as well as of dietary choline intake. Our findings motivate further studies to explore endogenous metabolic factors influencing the relationship between TMAO and cardiovascular disease.