Design: Multicentre retrospective observational study
Population: 157 adults on vancomycin with measured troughs, 63 with hepatic dysfunction and albumin below 2 g/dL
Primary outcome: Deviation of 30% or more between population-model and Bayesian AUC
Effect: Hepatic dysfunction aOR 2.26 (95% CI 1.10–4.62); higher AUC and trough ratios

When the liver fails and muscle mass falls, creatinine production drops and creatinine-based estimates overstate kidney function. Vancomycin dosing software that relies on population parameters and creatinine clearance may then misjudge exposure.

This multicentre Japanese retrospective study included 157 adults on vancomycin with measured troughs, 63 of them with hepatic dysfunction and albumin below 2 g/dL. It compared the AUC estimated from population parameters with the Bayesian AUC informed by measured troughs. Hepatic dysfunction more than doubled the odds that the two differed by 30% or more (aOR 2.26, 95% CI 1.10–4.62). Patients with hepatic dysfunction had higher steady-state AUC and trough-to-dose ratios.

The study is retrospective and modest, and uses a definition of hepatic dysfunction tied to very low albumin. But the mechanism is well understood and the remedy is cheap: obtain levels early rather than relying on initial model estimates in patients whose creatinine cannot be trusted.

In patients with cirrhosis or liver failure and very low albumin, consider measuring vancomycin levels early in the course rather than relying on the initial estimate.
Treat creatinine-based clearance as an overestimate in patients with low muscle mass or liver failure.
Use Bayesian dosing with measured levels rather than population estimates alone in these patients.
Watch for accumulation: exposure tended to be higher, not lower, with hepatic dysfunction.
Monitor kidney function closely, since vancomycin toxicity may be masked by a low creatinine.

Why it matters

A normal creatinine in a patient with liver failure can hide both reduced clearance and early nephrotoxicity.

The statistics, in plain English

An adjusted odds ratio of 2.26 means the odds of a 30% or greater mismatch were about twice as high with hepatic dysfunction. The interval (1.10 to 4.62) is wide because the study was small, but it stays above 1.0.