Evaluating potential link between liver enzyme abnormalities and bedaquiline exposure in multi-drug resistant tuberculosis patients
Lénaïg Tanneau (1), Elin M Svensson (1,2), Mats O Karlsson (1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden (2) Department of Pharmacy, Radboud Institute for Health Sciences, Radboud University Medical Center, Nijmegen, the Netherlands
Objectives: Bedaquiline (BDQ) is a drug indicated in the treatment of multidrug-resistant (MDR) tuberculosis (TB) as part of combination therapy [1]. Approved in 2012, it has shown ability to increase the cure rate and is now a prioritized component in MDR-TB regimens recommended by the WHO [2,3]. Despite its general well-tolerated profile, some risks are potentially associated with BDQ intake such as QT prolongation or hepatotoxicity. Our previous work focused on the relationship between bedaquiline or its main metabolite (M2) exposure and QT prolongation [4]. The aim of this analysis was to investigate a potential link between exposure of BDQ and/or M2 and elevation of alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) levels in MDR-TB patients using the approved BDQ dose.
Methods: Data were obtained from 2 phase IIb studies (C208 [5], C209 [1]) and were pooled to include a total of 335 patients treated with BDQ and 105 patients with placebo (PLC). Patients received BDQ (400mg q.d. for 2 weeks, then 200mg t.i.w.) or placebo for 24 weeks (or 8 weeks in stage 1 of C208) in combination with a background regimen of 5-7 anti-TB drugs. ALT and AST measurements were performed at day-1 and every 2 weeks in all patients of the placebo controlled C208 study and at day -1, week 2, 8 and 24 for patients in the open-label C209 study. Individual model-predicted pharmacokinetic variables were derived from a PK-model for BDQ and M2 previously established for these trials [8]. BDQ and M2 weekly average concentrations, time after start of treatment, and presence of background TB drugs were evaluated as covariates of the base transaminase levels. The impact of treatment group (PLC vs BDQ) was tested using C208 data only and the likelihood ratio test, with calibration for type 1 error using randomization test. Furthermore, a statistical analysis was performed on C208 data considering the highest on-treatment ALT/AST value per patient, either as continuous data or categorical data (graded to a 0-4 severity score, following the study protocol standards). A Wilcoxon rank sum test was used to compare the medians of the PLC group versus the BDQ group. A proportional odds model was used to describe the categorical data, analysing treatment group and BDQ/M2 exposures (1/CLBDQ,i; 1/CLM2,i) as predictors.
Results: 5881 observations of ALT and AST levels were recorded at baseline (up to 48h before start of study treatment) and during the treatment period. ALT and AST data were fitted simultaneously with separate models but allowing correlations between random effects and residuals. A model with an effect of presence of background (BG) TB drugs and an Emax-shaped time dependency best described the data (ALT baseline of 20.1 UI/L [RSE 4%] decreased by 3.8 UI/L [RSE 19%] with BG effect and by maximally 0.74 UI/L [RSE 51%] with time, T50ALT of 0.25 weeks [RSE 101%]; AST baseline of 23.8 UI/L [RSE 2%] increased by 2 UI/L [RSE 27%] with BG effect and by maximally 2.8 UI/L [RSE 18%] with time, T50AST of 6.3 weeks [RSE 49%]).
Considering C208 data only, no significant difference between BDQ vs PLC groups (p>0.1) was found either in the expected on-treatment values of ALT and AST using the randomization test, or between the medians of maximal on-treatment values of ALT (p=0.115) and ALT (p=0842) using the Wilcoxon rank sum test, although for both analytes, the BDQ group had higher values (median of 21.5 UI/L and 34 UI/L in the PLC group vs 25 UI/L and 37 UI/L in the BDQ group for ALT and AST respectively). The difference between the BDQ and PLC groups in severity scores was not significant for AST (P=0.38) but was for ALT (P<0.014) with higher ALT severity scores in the BDQ group (4 grade (G) 1 in the PLC group vs 6 G1, 6 G2 and 1 G4 in the BDQ group). However, no significant relationship with BDQ and M2 exposures could be identified.
Conclusions: Elevated transaminases have been labelled as a potential adverse reaction of BDQ treatment. Indeed, in C208/C209 studies, higher ALT and AST levels were observed in patients treated with BDQ compared to the C208 PLC group. However, our analyses could not correlate levels of BDQ or M2 with increased ALT or AST levels (i.e. no link was found to individual BDQ or M2 exposure metrics).
Further analysis of confounding factors such as alcohol use or underlying hepatic disease may be interesting to fully describe the profile of ALT and AST levels.
References:
[1] A. S. Pym et al., ‘Bedaquiline in the treatment of multidrug- and extensively drug-resistant tuberculosis’, Eur. Respir. J., vol. 47, no. 2, pp. 564–574, Feb. 2016.
[2] G. J. Fox and D. Menzies, ‘A Review of the Evidence for Using Bedaquiline (TMC207) to Treat Multi-Drug Resistant Tuberculosis’, Infect. Dis. Ther., vol. 2, no. 2, pp. 123–144, Dec. 2013.
[3] ‘WHO | WHO updates its treatment guidelines for multidrug- and rifampicin-resistant tuberculosis’, WHO. [Online]. Available: http://www.who.int/tb/features_archive/updated-treatment-guidelines-multigrug-rifampicin-resistant-TB/en/. [Accessed: 18-Feb-2019].
[4] L. Tanneau, E. Svensson, S. Rossenu, and M. O. Karlsson, ‘Bedaquiline appears to antagonize its own main metabolite’s QTcF interval prolonging effect’, Abstr 8634, PAGE 27, 2018.
[5] A. H. Diacon et al., ‘Multidrug-Resistant Tuberculosis and Culture Conversion with Bedaquiline’, N. Engl. J. Med., vol. 371, no. 8, pp. 723–732, Aug. 2014.
[6] E. Svensson, A. Dosne, and M. Karlsson, ‘Population Pharmacokinetics of Bedaquiline and Metabolite M2 in Patients With Drug‐Resistant Tuberculosis: The Effect of Time‐Varying Weight and Albumin’, CPT Pharmacomet. Syst. Pharmacol., vol. 5, no. 12, pp. 682–691, Dec. 2016.
