The Kidney Problem: When Clearance Changes the Signal

A high TMAO level can suggest greater production, slower removal, or both. That makes kidney function a central part of interpreting the TMAO–atrial fibrillation relationship, not a technical adjustment to be added at the end of a model. This chapter uses the PREVEND cohort as a diagnostic case: it shows how a strong-looking prognostic association can weaken when renal function is considered, and why that finding illuminates AF without directly proving an AF effect.

What renal clearance does to a prognostic association

The kidney helps clear TMAO from the circulation. Estimated glomerular filtration rate, or eGFR, is a commonly used measure related to how well the kidneys filter blood. When eGFR is lower, TMAO can remain in circulation longer. A higher concentration may therefore represent a change in exposure to TMAO, a change in elimination, or a combination of the two.

This matters because impaired renal function is also associated with adverse cardiovascular outcomes. If kidney function raises circulating TMAO and independently raises the chance of illness, then a simple comparison of people with high and low TMAO can mix together at least two processes. The molecule may still have biological effects, but the measured association cannot be read as if TMAO were freely varying while every other relevant condition stayed fixed.

The PREVEND study makes the problem concrete. It followed 5,469 participants for a median of 8.3 years, during which 322 people died. TMAO was higher among participants with lower eGFR. Comparing the highest with the lowest TMAO quartile produced a crude 1.86-fold higher mortality risk. The association remained after adjustment for several clinical risk factors, but the model changed again when renal variables were included.

The result changed when renal variables were brought into the model. Adjustment for urinary albumin excretion weakened the evidence but did not initially erase the trend. After eGFR was added as well, the hazard ratio was 1.15, with a 95% confidence interval from 0.81 to 1.64, and the trend was no longer statistically significant. The study also found that the relationship between TMAO and mortality was modified by eGFR: the association was present among participants with eGFR below 90 mL/min/1.73 m², but not among those with preserved renal function.

That pattern does not have only one possible interpretation. It may mean that reduced clearance was confounding the association: kidney dysfunction raised TMAO and mortality through partly separate routes. It may mean that renal function was on a pathway connecting TMAO to mortality. It may reflect reverse or bidirectional causation, in which disease affects kidney function and clearance while TMAO or related processes also contribute to renal injury. The study’s exploratory mediation analysis estimated that eGFR accounted for about 15% of the association, but that estimate does not settle the direction of the underlying biology.