Saliva-Based Therapeutic Drug Monitoring To Optimise Tuberculosis Treatment

Abstract

Saliva sampling presents a non-invasive alternative for therapeutic drug monitoring (TDM) in multidrug-resistant tuberculosis (MDR-TB). Despite its appeal, adoption has been limited due to insufficient data correlating saliva and plasma drug exposure. This thesis addresses this gap by investigating the mechanism of drug penetration into saliva, validating analytical methods, and evaluating clinical application of saliva-based TDM for two key anti-TB drugs: levofloxacin and linezolid. A systematic review identified physicochemical predictors of drug penetration into saliva. Based on these findings, a liquid chromatography–mass spectrometry (LC-MS) method was validated to quantify both drugs in saliva and plasma. A mobile device (NanoPhotometer®) was also evaluated against LC-MS, showing acceptable performance for potential point-of-care use. A clinical study in patients with MDR-TB demonstrated that saliva concentrations reliably identified subtherapeutic plasma exposures linked to treatment failure or toxicity. For levofloxacin, a saliva exposure threshold of 70.3 mg·h/L predicted low plasma exposure with 84% sensitivity and 78% specificity. Saliva-based identification of suboptimal and high exposures appeared to be possible for linezolid. Population pharmacokinetic modelling supported the use of limited saliva sampling to estimate plasma exposure, with simulations showing that a single sample coupled with Bayesian estimation could suffice for both drugs. Collectively, these findings validate saliva-based TDM as a feasible, patient-friendly approach to guide dosing in MDR-TB. This work supports broader implementation of saliva TDM for accessible, individualised care. Future research should assess its clinical utility and cost-effectiveness in real-world settings.