Single Saliva Sample Linezolid Dosing for Multidrug-Resistant Tuberculosis: A Population Pharmacokinetic Modelling of Plasma and Saliva.

Abstract

Background and objective: Therapeutic drug monitoring (TDM) of linezolid for treating multidrug-resistant tuberculosis (MDR-TB) is recommended but is hindered by the invasive and logistical challenges of plasma sampling. Saliva is a promising alternative, but a saliva pharmacokinetic model to inform dosing is lacking. This study aimed to develop a saliva-based population pharmacokinetic (popPK) model and evaluate limited sampling strategies for linezolid in MDR-TB.

Methods: Plasma and saliva samples were collected at pre-dose, 2-, and 5-h post-dose from adults treated with linezolid for ≥ 7 days. A saliva-plasma popPK model was developed in NONMEM with covariate analysis. Bayesian estimation and Monte Carlo simulations were used to evaluate the predictive performance of saliva limited sampling strategies for predicting plasma AUC_0-24. A ±20% bias versus reference was considered acceptable.

Results: One-compartment plasma model incorporating a saliva hypothetical effect compartment with first-order absorption and elimination best described the data (102 paired saliva-plasma samples, 17 patients). Body weight influenced volume of distribution with an exponent of 1.1 (95% CI 1.01-1.18). The three-sample (pre-dose, 2-, and 5-h) saliva strategy adequately predicted plasma AUC_0-24 with < 5% median bias. A single saliva sample at pre-dose or 2 h post-dose provided accurate plasma AUC_0-24 predictions (< 5% median bias).

Conclusions: The validated model enables reliable estimation of plasma AUC_0-24 using limited saliva samples. A single saliva sample (pre-dose or 2 h post-dose) can accurately predict plasma AUC_0-24, supporting saliva-based TDM as a practical alternative to plasma-based monitoring of linezolid in MDR-TB.