A pilot and feasibility analysis of serum TMAO and choline in a randomized sample from the COSMOS trial

Objective: The metabolite trimethylamine N-oxide (TMAO) is a biomarker influenced by diet and linked to atherosclerosis, thrombosis, and cardiovascular disease (CVD). Choline is a crucial nutrient involved in maintaining cell structure, promoting liver health, and neurotransmission. However, human evidence is limited whether cocoa extract (CE) changes TMAO and choline levels over time. The COcoa Supplement and Multivitamin Outcomes Study (COSMOS) is a randomized, placebo-controlled, 2 × 2 factorial trial testing a CE supplement (containing 500 mg d^-1 flavanols) and a daily multivitamin for chronic disease prevention in 21 442 older adults. We conducted a pilot study of the impact of CE on 1 year changes in TMAO and choline in COSMOS participants to facilitate sample size calculations for larger analyses of this trial in the future. In a pilot intention to treat analyses, linear mixed effect models were used to explore serum TMAO and choline trajectories in relation to CE assignment, adjusting for age, sex, and the other randomization arm. Methods: We randomly selected 40 COSMOS participants with blood samples at baseline and year 1. TMAO and choline were measured in plasma samples by liquid chromatography with stable isotope dilution tandem mass spectrometry. We excluded 3 of 40 participants due to extreme TMAO values ≥32.9 μM. Results: The mean age at baseline was 77 ± (5.5) years, and 18 (48.7%) were female. Randomization successfully distributed baseline demographic, clinical, behavioral, and dietary characteristics by treatment group. Among the 37 COSMOS participants, those assigned to take a CE supplement, as compared to placebo, had similar TMAO and choline levels at baseline but showed trends toward lower levels of TMAO (-0.60 [95% CI, (-3.76, 2.55)]) at 1 year. However, the between-group differences in the 1 year changes in TMAO (CE minus placebo) were not statistically significant (-1.16 [95% CI, (-5.81, 3.50)]; P = 0.62). In contrast, the group assigned CE showed trends toward higher levels of choline (0.84 (-1.57, 3.25)) at year 1, but the between-group differences in the 1 year changes in choline were again not statistically significant (2.23 [95% CI, (-1.31, 5.78)]; P = 0.21). Based on this pilot study, an expanded analysis of 1500 COSMOS participants would have 84% power to detect a 1.5 μM difference in TMAO levels comparing cocoa extract versus placebo. Conclusions: We measured TMAO and choline from stored blood samples in this pilot study from the COSMOS trial. Although daily CE supplementation was not associated with statistically significant 1 year changes in TMAO or choline levels, our sample size was limited. Larger studies are needed to understand whether TMAO and/or choline contribute to the reduction in CVD death observed with CE supplementation.