In-hospital trimethylamine N-oxide (TMAO) variation predicts major adverse cardiovascular events in myocardial infarction survivors

Abstract Funding Acknowledgements None. Background Trimethylamine N-oxide (TMAO), a gut microbiota result of the fermentation of animal foods rich in choline, L-carnitine and phosphatidylcholine (red meat, egg yolk, dairy food), is associated with alterations in lipid metabolism, increased foam cell formation, promotion of inflammation, oxidative stress, platelet aggregation and thrombosis. To date the data evaluating TMAO as a predictor of mortality and ischemic events in patients with acute myocardial infarction (AMI) based on one-time measurement of TMAO levels lead to conflicting results. TMAO levels are fluctuating and in addition to diet, genetic factors, general condition and some co-morbidities affect its circulating levels. Purpose Since, there is a need for better risk stratification of AMI patients before the hospital discharge for more personalised treatment, we evaluated in-hospital variation in TMAO levels and its impact on outcome. Methods Blood samples from 149 consecutive AMI patients were collected on admission and on discharge. The endpoint was a three-point major adverse cardiovascular events (MACE) composed as all-cause mortality, reinfarction and development of heart failure. Results The majority of AMI patients were male (77%), hospitalized for first coronary event, with mean age 64(11) years. Median circulating TMAO level was significantly higher on admission than on discharge, (7.81 [3.47 –19.98] vs 3.45 [2.3 – 4.78] μM, respectively, p<0.001). In order to evaluate the impact of TMAO variation during the hospital stay on outcome, we estimated TMAO cut-off value by continuous hazard ratio analysis; values above the 3.45 μM were associated with incident of MACE in a linear fashion (Fig. 1a). Subsequently, we stratified AMI patients in 2 groups: 1) LL/HL (low-low/high-low) that included 75 (50.3%) patients whose TMAO levels remained low or were high and decreased during hospitalization and 2) HH/LH (high-high/low-high) including 74 patients (49.7%) patients whose TMAO levels remained elevated or were initially low, but rose above the estimated threshold during hospitalisation (Fig. 1b). Thirty-two (21.5%) patients reached the combined endpoint during the median follow-up of 30 months. The Kaplan-Meier analysis demonstrated that patients from the HH/LH group had a higher likelihood of experiencing MACE (p=0.05) (Fig. 2a). The multivariate Cox analysis revealed that patients in the HH/LH group had over twice the risk of MACE compared to those in the LL/HL group. Furthermore, older age and worse left ventricular systolic function were identified as independent predictors of MACE during long-term follow-up (Fig. 2b). Conclusions AMI patients with permanently elevated or increasing levels of TMAO have a worse outcome. The novelty and the strength of our study is the fact that the direction of variation of TMAO during the index hospitalisation identifies patients at higher risk of MACE during 30 months follow-up.