Trimethylamine N-oxide is associated with impaired cognitive function in patients with atrial fibrillation

Abstract Introduction Since patients with atrial fibrillation (AF) are predisposed to suffer from major adverse cerebrovascular events (MACE), they are more likely to suffer MACE linked sequelae, such as cognitive impairment. We hypothesised that the gut microbiome derivate trimethylamine N-oxide (TMAO) may amplify this pathomechanism given its hypercoagulative, proinflammatory and proatherogenic effects. Methods Patients of the Swiss-AF cohort with determined TMAO plasma levels, cognitive scores (n=2'379) and cerebral magnetic resonance imaging (cMRI) (n=1'722) collected at baseline were included. TMAO levels were measured by liquid chromatography-mass spectrometry (HPLC). Overall cognitive performance was evaluated using the Cognitive Construct (CoCo) score reflecting different cognitive functions measured by four validated neuropsychological assessments, namely the Montreal Cognitive Assessment (MoCA), Trail Making Test (TMT parts A and B), Semantic Fluency Test (SFT) and Digital Symbol Substitution Test (DSST). The scores were compared with the quartiles of patients' TMAO plasma levels (Q1: 0.6–4, Q2: 4–5.8, Q3: 5.8–9.1, Q4: 9.1–164μmol/l) in linear effect models. All models were adjusted for multiple covariates correlating with this association: For TMAO (overall meat consumption >3 times per week, physical activity, glomerular filtration rate, presence of diabetes mellitus), cognitive function (EQ-5D-5L score, geriatric depression scale, education level) and cerebral affects (white matter lesions volume and total brain volume in cMRI). The relevance of high TMAO plasma levels in different stroke groups in cMRI (i.e., clinically overt, silent, or no stroke) were analysed in a subgroup analysis. The subgroups were additionally adjusted for total brain volume to eliminate the effect of simultaneous decrease of cerebral white matter. Results After multivariable adjustment, AF patients in the highest quartile of TMAO levels performed significantly poorer in the global cognitive score (CoCo: estimate −0.11, 95% CI [−0.17, −0.05], p=0.002) compared to patients in the lowest quartile. This was observed also in the MoCA, SFT, DSST, TMT-A and TMT-B. In the subgroup analysis, an association between the highest TMAO quartile (compared with the lowest quartile) and lower CoCo score was found in the group of patients with overt strokes (−0.18, 95% CI [−0.33, −0.04], p=0.012). Weak evidence of the same association was found in the group of patients with silent strokes (−0.13, 95% CI [−0.25, 0.002], p=0.053) and patients with no strokes (−0.08, 95% CI [−0.16, 0.01], p=0.07). After adjustment for decreased brain volume, the association remained for silent strokes (−0.14, 95% CI [−0.28, −0.01], p=0.036) indicating the impact of TMAO in this subgroup. Conclusion TMAO plasma levels were associated with cognitive impairment in patients with AF. Longitudinal data will clarify dynamics and likely causality between TMAO and cognitive impairment. Funding Acknowledgement Type of funding sources: Public Institution(s). Main funding source(s): Swiss National Science Foundation