Evaluation of pharmacokinetic‐pharmacodynamic relationships and selection of drug combinations for tuberculosis

Aims Despite evidence of the efficacy of anti‐tubercular drug regimens in clinical practice, the rationale underpinning the selection of doses and companion drugs for combination therapy remains empirical. Novel methods are needed to optimise the antibacterial activity in combination therapies. A drug–disease modelling framework for rational selection of dose and drug combinations in tuberculosis is presented here. Methods A model‐based meta‐analysis was performed to assess the antibacterial activity of different combinations in infected mice. Data retrieved from the published literature were analysed using a two‐state bacterial growth dynamics model, including fast‐ and slow‐growing bacterial populations. The contribution of each drug to the overall antibacterial activity of the combination was parameterised as relative change to the potency of the backbone drug (EC 50 ‐F and/or EC 50 ‐S). Rifampicin and bedaquiline were selected as paradigm drugs to evaluate the predictive performance of the modelling approach. Results Pyrazinamide increased the potency (EC 50 ‐F and EC 50 ‐S) of rifampicin (RZ) and bedaquiline (BZ) by almost two‐fold. By contrast, pretomanid and isoniazid were found to worsen the antibacterial activity of BZ and RZ, respectively. Following extrapolation of in vivo pharmacokinetic‐pharmacodynamic relationships, the dose of rifampicin showing maximum bactericidal effect in tuberculosis patients was predicted to be 70 mg·kg −1 when given in combination with pyrazinamide. Conclusions The use of a drug–disease modelling framework may provide a more robust rationale for extrapolation and selection of dose and companion drugs in humans. Our analysis demonstrates that RZ and BZ should be considered as a backbone therapy in prospective novel combination regimens against tuberculosis.