Population pharmacokinetics and limited sampling strategy for first-line tuberculosis drugs and moxifloxacin

Abstract

Therapeutic drug monitoring (TDM) of tuberculosis (TB) drugs currently focuses on peak plasma concentrations, yet total exposure [area under the 24-h concentration–time curve (AUC0–24)] is probably most relevant to the efficacy of these drugs. We therefore assessed population AUC0–24 data for all four first-line TB drugs (rifampicin, isoniazid, pyrazinamide and ethambutol) as well as moxifloxacin and developed limited sampling strategies to estimate AUC0–24 values conveniently. AUC0–24 and other pharmacokinetic (PK) parameters were determined following intensive PK sampling in two Dutch TB referral centres. Best subset selection multiple linear regression was performed to derive limited sampling equations. Median percentage prediction error and median absolute percentage prediction error were calculated via jackknife analysis to evaluate bias and imprecision of the predictions. Geometric mean AUC0–24 values for rifampicin, isoniazid, pyrazinamide, ethambutol and moxifloxacin were 41.1, 15.2, 380, 25.5 and 33.6hmg/L, respectively. Limited sampling at various fixed sampling points enabled an accurate and precise prediction of AUC0–24 values of all drugs separately and simultaneously. In the absence of clinically validated target values for AUC0–24, average AUC0–24 values can be used as reference values in TDM. Limited sampling of AUC0–24 is feasible in many settings and allows for TDM to be performed at a larger scale.