Trimethylamine N-oxide and major adverse cardiac and cerebrovascular events: results of umbrella review and meta-analysis

Abstract

Abstract Background Circulating trimethylamine N-oxide (TMAO) may be a causal factor or marker for major adverse cardiac and cerebrovascular events (MACE), but evidence in humans is limited to observational studies with inherent uncertainties. Methods We conducted an umbrella systematic review and meta-analysis of human studies. The search in MEDLINE, Web of Science, Cochrane Library and PROSPERO (last updated in February 2025) identified 120 publications that met our inclusion criteria. Results The pooled hazard ratio (HR) and 95% confidence intervals (95%CIs) from 34 cohort studies on TMAO and MACE = 1.20 (1.12–1.27), the prediction interval (PI) was 0.93–1.54, and I 2 = 73.3%. After correction for publication bias ( p < 0.001 Egger’s test), the pooled HR (95%CIs) was 1.10 (1.03–1.17). The pooled odds ratio (95%CIs) from 18 case-control studies after accounting for publication bias was 1.05 (0.94–1.18). As a diagnostic test, the estimated pooled area under the curve (95%CIs) for TMAO was 72.8 (68.8–77.1)%, PI = 54.7%–97.1%, I 2 = 93.8%, n = 24 studies. The mean (95%CIs) sensitivity and specificity of TMAO were 65.3 (56.1–74.5)% and 79.0 (65.4–92.6)%, respectively. The data did not suggest a threshold for elevated TMAO concentrations. Conclusions Previously reported association estimates between TMAO and MACE in human studies are systematically overestimated. The results do not support TMAO being a causal factor or marker of MACE and related conditions. Harmonisation of TMAO measurements between labs and cohort studies in healthy people that address confounding factors remain critically needed.