Drug effect of clofazimine on persisting mycobacteria explain an unexpected increase in bacterial load from patients
Alan Faraj (1), Robin J Svensson (1), Andreas H Diacon (2), Ulrika SH Simonsson (1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden, (2) TASK Applied Science, Cape Town and Division of Physiology, Faculty of Medicine and Health Sciences, Stellenbosch University, Tygerberg, South Africa
Introduction
Clofazimine (CFZ) and pyrazinamide (PZA) are important components of recommended standard multi-drug treatments of TB. Paradoxically, in a Phase IIa clinical trial aiming to define the early bactericidal activity (EBA) of CFZ and PZA monotherapy over the first 14 days of treatment, no significant drug effect could be demonstrated for the two drugs using traditional statistical analysis [1]. An unexpected numerical increase in colony forming units (CFU) over time, was observed with CFZ monotherapy.
Objectives
This work aimed to establish exposure-response relationships of both drugs in monotherapy and to explain the numerical increase in CFU after CFZ monotherapy, using a model-based approach.
Methods
Pharmacokinetic (PK) and CFU data from 14 and 15 patients after CFZ or PZA in monotherapy [1], respectively was analysed using nonlinear mixed effect modelling. A population PK model was developed for CFZ. For PZA, a previously developed PK model was applied [2]. Adequate model-predicted individual PK profiles were linked to the Multistate Tuberculosis Pharmacometric (MTP) model [3,4] to explore exposure-response relationships on the killing of the multiplying, semi-dormant, persistent and/or inhibition of multiplying bacteria alone and in combination. In order to challenge the consistency of the results, a sensitivity analysis was performed where the relative ratio of the bacterial subpopulations was varied. The exposure-response relationships were subsequently re-estimated.
Results
A two-compartment model with first order absorption and elimination together with an absorption lag-time was supported by the clofazimine PK data. Inter-individual variability (IIV) was supported for apparent oral clearance (CL/F), apparent volume of distribution (V/F) and the first-order absorption (ka) parameter. Inter-occasional variability (IOV) was supported for bioavailability. No statistically significant covariate relationship was found using body weight, age and sex on CL/F or V/F. Using the MTP model [3,4], statistically significant exposure-response relationships were characterized for both drugs, with a linear concentration-dependent killing effect for CFZ on persistent tubercular bacilli and a linear concentration dependent effect for PZA on semi-dormant mycobacteria. The final model could explain the original findings of paradoxical increase in CFU with CFZ treatment as well as no effect with PZA when the analysis did not include variables for different metabolic states of mycobacteria.
Conclusions
A novel semi-mechanistic model-based analysis of individual PK and sputum CFU counts revealed significant activity of CFZ and PZA on persistent and semi-dormant mycobacteria, respectively, which remained undetected with traditional methods of quantification, of anti-tuberculosis drug effect. Further, the drug effect on persistent tubercular bacilli explained the unexpected increase in CFU after CFZ monotherapy. We propose that this quantitative approach that provides a rational framework for analysing drug effects in Phase IIa EBA studies, can accelerate anti-TB drug development.
References:
[1]. Diacon, A. H. et al. Bactericidal activity of pyrazinamide and clofazimine alone and in combinations with pretomanid and bedaquiline. Am. J. Respir. Crit. Care Med. 191, 943–953 (2015).
[2]. Wilkins, J. J. et al. Variability in the population pharmacokinetics of pyrazinamide in South African tuberculosis patients. Eur. J. Clin. Pharmacol. 62, 727–735 (2006).
[3]. Clewe, O. et al. A multistate tuberculosis pharmacometric model: a framework for studying anti-tubercular drug effects in vitro. J. Antimicrob. Chemother. 71, 964–974 (2016).
[4]. Svensson, R. & Simonsson, USH. Application of the Multistate Tuberculosis Pharmacometric Model in Patients With Rifampicin‐Treated Pulmonary Tuberculosis. CPT Pharmacometrics Syst Pharmacol 5, 264–273 (2016).
