Abstract
Bedaquiline, a recently approved drug for the treatment of multidrug-resistant tuberculosis (MDR-TB), is recommended for a duration of 24 weeks. There are scarce data on patients treated with this drug outside clinical trials.
All MDR-TB patients who started treatment from January 1, 2011 to December 31, 2013 and received ≥30 days of bedaquiline were included in a multicentre observational cohort.
Among 45 MDR-TB patients, 53% harboured isolates resistant to both fluoroquinolones and second-line injectables, and 38% harboured isolates resistant to one of these drug classes. Median bedaquiline treatment duration was 361 days and 33 patients (73%) received prolonged (>190 days) bedaquiline treatment. Overall, 36 patients (80%) had favourable outcome, five were lost to follow-up, three died, and one failed and acquired bedaquiline resistance. No cases of recurrence were reported. Severe and serious adverse events were recorded in 60% and 18% of patients, respectively. Values of Fridericia-corrected QT interval (QTcF) >500 ms were recorded in 11% of patients, but neither arrhythmias nor symptomatic cardiac side-effects occurred. Bedaquiline was discontinued in three patients following QTcF prolongation. No significant differences in outcomes or adverse events rates were observed between patients receiving standard and prolonged bedaquiline treatment.
Bedaquiline-containing regimens achieved favourable outcomes in a large proportion of patients. Prolonged bedaquiline treatment was overall well tolerated in this cohort.
Abstract
Treatment regimens including prolonged bedaquiline use are effective and overall well tolerated in MDR-TB patients http://ow.ly/vhtl305CYJj
Introduction
In 2014, according to the World Health Organization (WHO) [1], 480 000 people were newly diagnosed with multidrug-resistant tuberculosis (MDR-TB), defined by resistance to isoniazid and rifampicin. Among MDR-TB cases, 9% had extensively drug-resistant TB (XDR-TB) (resistance to any fluoroquinolone and any second-line injectable drug) [1]. MDR-TB treatment outcomes, although heterogeneous in different settings, are overall unsatisfactory. A meta-analysis of MDR-TB patients showed that treatment success rates were 64% with individualised regimens and 54% with standardised treatment [2]. Two large meta-analyses comprising individual patient data consistently found overall treatment success rates of 54% and 56% [3, 4]. Treatment outcomes of XDR-TB are abysmal, ranging from 27% to 40% [3, 5]. A study assessing the long-term outcomes of XDR-TB patients reported 11% of favourable outcomes and 73% mortality at 5 years of follow-up [6].
In order to face this emerging challenge, old drugs have been repurposed and new drugs have been developed. Bedaquiline and delamanid have been recently approved for the treatment of MDR-TB. A phase II trial showed that bedaquiline improves treatment outcomes when compared with placebo [7]. In an uncontrolled phase II study, bedaquiline plus optimised background regimens achieved favourable outcomes in 62.5% of patients [8]. Preliminary reports of bedaquiline compassionate-use programmes confirm these promising results [9–15]. However, interim results also underline the risk of experiencing culture reversion after the discontinuation of bedaquiline [10]. To date, the WHO recommends the use of bedaquiline for a maximum duration of 24 weeks [16] as no evidence is available supporting longer use, except for one recent case report [17]. Nausea and hepatitis are the most common side-effects associated with bedaquiline [7–9]. However, the main safety concern is cardiotoxicity. Although no serious cardiac events or arrhythmias have been reported to date, bedaquiline has been shown to prolong the QT interval and the association with other drugs (such as clofazimine or moxifloxacin) can enhance this effect [8, 9]. Due to the long terminal half-life of bedaquiline, a cumulative effect of prolonged bedaquiline administration on the QT interval could be postulated. We previously reported the interim analysis of a cohort of MDR-TB patients receiving bedaquiline-containing regimens, with 6-month culture conversion rates of 96% and no relevant safety signals [9]. The flexibility of the regulatory framework of the compassionate-use programme in France has allowed the off-label use of bedaquiline beyond 24 weeks in selected patients. The aim of this observational analysis is to complement our previous findings with the evaluation, in a bigger cohort and up to 24 months after treatment completion, of the treatment outcome and safety profile of individualised anti-TB regimens containing prolonged bedaquiline treatment.
Methods
Patients and treatment
A retrospective cohort was established including all MDR-TB patients treated with bedaquiline from January 1, 2011 to December 31, 2013 and hospitalised at three French referral TB centres (Bligny, Pitié Salpêtrière and Bichat Hospitals). Patients were followed up after the end of treatment for up to 24 months or to the censor date (March 31, 2016). All TB cases were culture-proven. Standard definitions of MDR-TB and XDR-TB were used, and treatment outcomes were assigned according to WHO definitions [16]. At the end of treatment, favourable outcomes were defined as the sum of cured and treatment completed; all other outcomes were defined as unfavourable. All patients with favourable outcomes were re-assessed at 12 and 24 months after end of treatment.
The treatment regimen was designed for each patient according to clinical features, phenotypic and genotypic drug susceptibility testing (DST) results with the advice of the MDR-TB Consilium of the French National Reference Centre, which assessed the eligibility for bedaquiline treatment and its prolongation beyond 24 weeks. The criteria that were used to identify eligible patients for bedaquiline prolongation were: delayed microbiological response, weak treatment regimens due to intolerance or drug resistance and/or individual risk factors for poor outcomes (table 1). In addition, all WHO-recommended requirements for bedaquiline use were met, including active pharmacovigilance and treatment monitoring [16].
Prerequisites and criteria used in this cohort for bedaquiline prolongation beyond 24 weeks
Bedaquiline (Sirturo) was provided in the framework of the compassionate-use programme and was administered as recommended by the manufacturer (Janssen, Beerse, Belgium). Standard bedaquiline treatment was defined as ≤190 days, representing the standard duration of 24 weeks plus a buffer period of 3 weeks needed by the Consilium to assess bedaquiline treatment duration. Prolonged bedaquiline treatment was defined as >190 days. All drugs were administered as directly observed treatment during hospitalisation. Treatment and hospitalisation were offered free of charge to all patients, including migrants and refugees. All patients were informed of the mechanism of the compassionate-use programme and the safety profile of all drugs in the treatment regimen including bedaquiline. Data were retrospectively extracted from medical records. Human research ethics approval for the study was granted by the Institutional Review Board of Bligny Hospital.
Procedures
Sputum smear and culture examinations were performed at treatment start, fortnightly up to culture conversion and monthly thereafter. Time to smear/culture conversion was measured from treatment start to the first of two consecutive negative smear/culture results. Phenotypic DST for a panel of first- and second-line anti-TB drugs was performed at the National Reference Centre using the proportion method on Löwenstein–Jensen medium [18]. Genotypic DST was obtained with commercially available line probe assays (GenoType MTBDRplus and GenoType MTBDRsl; Hain Lifescience, Nehren, Germany) or DNA sequencing. From March 2013 onward, bedaquiline DST was performed on Löwenstein–Jensen medium using the proportion method and a 64 mg·L−1 critical concentration for screening. Resistance was subsequently confirmed in TH11 medium. Bedaquiline DST was performed at baseline and repeated in case of suspicion of treatment failure. A standard 12-lead ECG was performed at baseline, at 2 weeks of treatment and monthly thereafter. QT interval correction was calculated according to Fridericia (QTcF) and Bazett (QTcB) [19]. A prolongation of the QT interval was defined as ≥60 ms increase during treatment. All adverse events were defined and graded according to severity and seriousness on the basis of the US National Institutes of Health Common Terminology Criteria for Adverse Events version 4.0 [20]. Severe adverse events were defined as any event graded as level 3, 4 or 5. Causality of adverse events was evaluated according to the WHO-Uppsala Monitoring Centre system for standardised case causality assessment [21].
Statistical analysis
Categorical variables were compared by using Fisher's exact test. Continuous variables were reported as median and interquartile range (IQR), and compared by using the two-sample Wilcoxon–Mann–Whitney test. Kaplan–Meier curves for culture conversion were estimated. The Mantel–Cox test was used to compare time to culture conversion. The association between variables and time to culture conversion was studied with a Cox proportional hazards model. Multivariable logistic regression was used to estimate the association of QT interval prolongation and explanatory variables. Variables associated in univariate analysis (p<0.20) were considered for backward multivariable analysis. p-values <0.05 were considered as significant. Statistical analysis was performed using STATA (StataCorp, College Station, TX, USA). Results are reported according to the STROBE (strobe-statement.org) guidelines for observational cohort studies.
Results
Sociodemographic and disease characteristics
Among the 102 MDR-TB patients managed in the three centres during the study period, 45 patients (44.1%) were treated with bedaquiline: 36 (80.0%) were born in Eastern Europe/Caucasus countries (table 2). Co-infection with hepatitis C virus (HCV) and HIV was present in 21 (46.7%) and two (4.4%) patients, respectively. 34 (75.6%) patients previously received TB treatment. Overall, 44 patients had pulmonary TB: 39 (88.6%) had lung cavities and 36 (81.8%) had bilateral lung involvement. More detailed baseline characteristics are reported in online supplementary table S1.
Sociodemographic characteristics, disease features and treatment regimens of the patient cohort
Resistance patterns and treatment
The majority of the patients had XDR-TB (n=24 (53.3%)). Out of the remaining patients, 11 (24.4%) had strains with additional resistance to fluoroquinolones and six (13.3%) had strains with additional resistance to any second-line injectable. Four (8.9%) had intolerance to either fluoroquinolones or second-line injectables. The strains showed phenotypical resistance to a median (IQR) of n=9 (7–11) drugs and n=5 (4–6) mutations in resistance-conferring genes. All tested strains were susceptible to bedaquiline at baseline (table 3). The most frequently prescribed companion drugs are listed in table 2. Median (IQR) treatment duration was 624 (546–730) days; injectables were administered for a median (IQR) of 341 (228–455) days. The median (IQR) duration of bedaquiline administration was 360 (31–768) days. 15 patients (33.3%) received bedaquiline for the full treatment duration. Lung surgery, mostly lobectomy, was performed in 12 (26.7%) patients after a median (IQR) of 170 (75–269) days from treatment start and after sputum culture conversion in 75% of cases.
Baseline phenotypic and genotypic resistance patterns of the isolated strains
Treatment safety profile
During treatment, 44 (97.8%) patients experienced at least one adverse event (table 4). The most frequent adverse events were gastrointestinal side-effects (n=32 (71.1%)), otovestibular impairment (n=25 (55.6%)) and peripheral neuropathy (n=18 (40.9%)). Severe and serious adverse events were recorded in 27 (60.0%) and seven (17.8%) patients, respectively. The most common severe adverse events were peripheral neuropathy (n=13) and QTc prolongation (n=8). Severe and serious adverse events are detailed in online supplementary table S2. Bedaquiline was discontinued in three (6.7%) patients due to QTc prolongation after 31, 203 and 279 days of treatment, respectively. One patient experienced uncomplicated pancreatitis a few weeks after bedaquiline discontinuation.
Treatment safety in the whole cohort and comparison between patients receiving standard (≤190 days) or prolonged (>190 days) bedaquiline treatment
With regard to the QT interval, only QTcF results will be reported, as no significant difference between QTcF and QTcB results was observed. Figure 1 shows the evolution of QTcF during treatment in the cohort. Overall, QTcF prolongation occurred in 13 (28.9%) patients. QTcF >500 ms was recorded in five (11.1%) patients, all belonging to the prolonged bedaquiline group; in three cases the QTcF prolongation occurred during the first 24 weeks of treatment. Median QTcF values remained stable during the whole treatment duration in the prolonged bedaquiline group. The median (IQR) maximum QTcF increase was 36.2 (17.9–68.5) ms. In logistic regression analysis, both QTcF >60 ms increase and QTcF >500 ms were independently associated with coadministration of moxifloxacin at 800 mg·day−1 after adjustment for age, sex and treatment with other QT-prolonging drugs. Methadone treatment was equally associated with QTcF >500 ms. No association was found with treatment with clofazimine, levofloxacin or moxifloxacin at 400 mg·day−1. The median of the maximum QTcF increase during treatment was significantly higher in patients treated with high-dose moxifloxacin (data not shown). Neither clinical arrhythmia nor any cardiac event were observed.
Box plot of the Fridericia-corrected QT (QTcF) values of the 45 patients at treatment start, at week 2 of treatment (W2), at month 1, 2, 3, 4, 5, 6, 9, 12 and 15 of treatment (M1, M2, M3, M4, M5, M6, M9, M12 and M15), and at the end of treatment. Data are presented as median and interquartile range; minimum and maximum values are shown by whiskers; outliers are indicated by circles.
Treatment outcomes
Out of 41 patients with positive sputum cultures at treatment start, 23 (56.1%) and 40 (97.6%) achieved culture conversion at 90 and 180 days, respectively. One patient achieved culture conversion at 8 months of treatment. Median (IQR) time to sputum smear and culture conversion was 90 (36–173) and 89 (45–107) days, respectively (table 5). In a multivariate Cox proportional hazard model, factors independently associated with faster time to culture conversion were HCV-negativity (hazard ratio (HR) 2.64, 95% CI 1.34–5.19; p=0.021), the absence of lung cavities (HR 4.56, 95% CI 1.41–14.75; p=0.011) and higher serum albumin levels at treatment start (HR 1.09, 95% CI 1.02–1.16; p=0.010). No association was found between prolonged bedaquiline treatment and time to culture conversion after adjustment (table 6).
Treatment outcomes of the whole cohort and comparison between patients receiving standard (≤190 days) or prolonged (>190 days) bedaquiline treatment
Univariate and multivariate Cox proportional hazards models assessing the association of factors with time to culture conversion
At the end of treatment, 36 of the 45 (80.0%) patients had favourable outcome. Nine (20.0%) had unfavourable outcome, including three deaths and one treatment failure with acquisition of resistance to bedaquiline. During post-treatment follow-up, one patient died before the 12-month end-point and one died before the 24-month end-point. No recurrences were recorded in the cohort (table 5). The characteristics of patients who died or experienced treatment failure are summarised in table 7.
Summary of the characteristics and evolution of patients who died or experienced treatment failure
With regard to causality assessment, bedaquiline was considered as unlikely related to all deaths and other serious adverse events.
Comparison of standard and prolonged bedaquiline treatment regimens
Overall, 12 (26.7%) and 33 (73.3%) patients received standard (median (IQR) 183 (168–185) days) and prolonged (median (IQR) 418 (292–665) days) bedaquiline treatment. Patients receiving prolonged bedaquiline treatment were more often previously treated for TB (p<0.001). They were more likely to have XDR-TB, bilateral lung involvement, cavitary TB and strains with resistance to a greater number of drugs, although these differences did not reach statistical significance (online supplementary table S3). No significant differences were recorded between the two groups regarding the incidence of total, severe and serious adverse events, including liver enzyme elevation. No statistical difference was found in the rate of QTcF prolongation and QTcF >500 ms nor in the maximum QTcF increase recorded during treatment (table 4). Patients in the prolonged treatment group were more frequently sputum culture-positive at treatment start (p=0.048) and had slower time to culture conversion (91 versus 71 days; p=0.021) (figure 2). Favourable and unfavourable treatment outcome rates at the end of treatment and during post-treatment follow-up were comparable between the two groups (table 5).
Kaplan–Meier analysis of time to sputum culture conversion of patients with positive sputum cultures at the beginning of treatment (n=41) according to whether they received standard (n=9) or prolonged (n=32) bedaquiline treatment. Prolonged bedaquiline treatment was defined as >190 days.
Discussion
We report successful outcomes in 80% of MDR-TB patients treated with bedaquiline-containing regimens, with a high rate of adverse events.
This rate of success is remarkable, as our cohort included a substantial number of XDR-TB patients, HCV-infected cases and intravenous drug abusers undergoing methadone treatment. Similar treatment outcomes are described in other high-resource settings, but these studies included fewer XDR-TB patients [22–24]. Previous studies including XDR-TB patients with low HIV co-infection rates as in our cohort reported lower success rates [25, 26]. Our results could be explained by multiple factors: treatment follow-up in specialised centres with comprehensive patient support and appropriate management of adverse events, free-of-charge treatment and social support for precarious populations, availability of reliable DST by a reference laboratory, and tailored treatment regimens including lung surgery [27].
The duration of bedaquiline treatment was established according to individual clinical evaluation. Interestingly, no differences in terms of efficacy and tolerance were found between standard and prolonged bedaquiline groups, although the latter group arguably contained more difficult-to-treat patients who achieved delayed sputum culture conversion. Moreover, prolonged bedaquiline treatment courses could partly explain the better outcomes in our cohort with regard to bedaquiline treatment arms of the published phase II clinical trials [7, 8]. Notably, almost all the patients in our cohort received linezolid, a drug shown to improve treatment outcomes of MDR-TB patients [28, 29]. Other repurposed drugs, such as clofazimine [30] and carbapenems [31, 32], might have played a role. Finally, more than half of the patients received a last-generation fluoroquinolone and 22% received high-dose moxifloxacin to overcome low-level fluoroquinolone resistance. A standardised 9-month treatment regimen including a fourth-generation fluoroquinolone at high dose was shown to achieve good outcomes in patients harbouring strains with low-level fluoroquinolone resistance [33].
Overall, 11% of patients were lost to follow-up during treatment and this rate markedly increased during post-therapeutic follow-up. This finding is possibly related to the precarious state of the patients, the vast majority of whom were foreign-born and often migrating to seek adequate healthcare. The rate of loss to follow-up is similar to previous studies [22–24]. In our cohort, five patients died: three during treatment and two after the end of treatment. For the two patients who died during treatment after developing neurological complications (Patients 3 and 4; table 7), causality assessment is challenging. The event was considered possibly related to linezolid because this drug has well-documented neurological toxicity [28, 29] and bedaquiline involvement was considered as unlikely, but cannot be excluded. The single treatment failure (Patient 1; table 7) was linked to bedaquiline resistance during treatment with only three likely active drugs. There is a need for efficient drug companions to protect new drugs from acquired resistance, as recently illustrated by the case of a patient who developed sequential resistance to bedaquiline and delamanid [34, 35].
Lung surgery was performed in 26.7% of our patients, who had cavitary disease, to attain sustained culture conversion. The capacity of TB drugs to penetrate into lung cavities and advanced tissue necrosis is unknown, and should lead us to consider surgical treatment [36].
Toxicity is often a major issue in the long-term treatment needed for MDR-TB: in the placebo arm of the C208 study, 98% of participants experienced adverse events [7]. Observational cohorts corroborate the high frequency of side-effects [37]. In our study, all patients except one experienced at least one adverse event and 82% had to stop one drug because of side-effects. More than half of our patients experienced peripheral neuropathy, a known complication of prolonged linezolid administration [38]. QTc prolongation in our cohort was more frequent than reported in bedaquiline-treated patients in other studies [7, 8, 10, 11], possibly because many patients in our cohort received other QT-prolonging drugs, such as moxifloxacin and clofazimine, and other non-TB drugs, such as methadone. Indeed, the administration of both high-dose moxifloxacin and methadone was associated with QTcF >500 ms in our cohort. No association was found between clofazimine treatment and QTc prolongation. However, studies with a larger sample size will be needed to confirm these results. There was no statistical difference in the number of total adverse events, serious adverse events and severe adverse events between standard and prolonged bedaquiline groups. No difference was found in the incidence of bedaquiline-associated adverse events, such as liver enzyme elevation, pancreatitis and QTc prolongation. Although all episodes of QTcF >500 ms prolongation occurred in the prolonged bedaquiline group, most patients experienced the prolongation during the first 6 months of treatment. Median QTcF values remained stable over the whole treatment duration in the prolonged bedaquiline group, irrespective of the cumulative exposure to the drug. Notably, no patient in the cohort received both bedaquiline and delamanid, so no conclusion can be drawn on the tolerability of the association of these two drugs [39, 40].
Our study has multiple limitations. First, data were collected retrospectively, possibly leading to an increase in the rate of patients lost to follow-up and to the under-reporting of adverse events. Second, the small sample size of the cohort might not have enough power to show existing differences between the standard and prolonged bedaquiline groups. Third, no control group of patients not receiving bedaquiline was analysed. Finally, no measurements of bedaquiline levels in blood or surgical specimens were performed in the study.
In conclusion, our results show promising outcomes in a cohort including mostly XDR-TB patients and a reassuring safety profile of prolonged bedaquiline administration. The prolongation of bedaquiline treatment in selected patients with multiple risk factors could have contributed to the high rate of favourable outcomes. We therefore advocate for extension of bedaquiline treatment beyond 24 weeks according to the individual patient's need and to predefined criteria (table 1), as done for other TB drugs.
The advent of new drugs offers us the opportunity to both improve outcomes and reduce the toxicity of MDR/XDR-TB treatment. A rapid increase in the evidence supporting the use of bedaquiline will hopefully provide additional treatment options for these patients.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary materials ERJ-01799-2016_Supplementary_Tables
Disclosures
Supplementary Material
M. Fréchet-Jachym ERJ-01799-2016_Fréchet-Jachym
M. Jaspard ERJ-01799-2016_Jaspard
J. Robert ERJ-01799-2016_Robert
N. Veziris ERJ-01799-2016_Veziris
Y. Yazdanpanah ERJ-01799-2016_Yazdanpanah
Acknowledgements
The members of the French MDR-TB Management Group are: C. Andrejak, A. Aubry, F. Brossier, K. Chadelat, B. Dautzenberg, N. El Helali, B. Henri, G. Petitjean, L. Raskine, B. Rivoire and G. Thouvenin.
The authors thank all patients who took part to the study. They also acknowledge all physicians and microbiologists who provided data for the study. The authors acknowledge F. Varaine (Médecins Sans Frontières, Paris, France) for providing useful comments on the manuscript.
Author contributions: L. Guglielmetti made a substantial contribution to the conception and design of the work, the acquisition, analysis and interpretation of data, performed statistical analysis, wrote the manuscript, critically revised the manuscript, and gave final approval of the current version to be published. M. Jaspard and M. Fréchet-Jachym made a substantial contribution to the conception and design of the work, the acquisition, analysis and interpretation of data, wrote the manuscript, critically revised the manuscript, and gave final approval of the current version to be published. D. Le Dû made a substantial contribution to the conception and design of the work, the acquisition, analysis and interpretation of data, critically revised the manuscript, and gave final approval of the current version to be published. M. Lachâtre, D. Marigot-Outtandy, C. Bernard, N. Veziris, J. Robert, Y. Yazdanpanah and E. Caumes made a contribution to the acquisition of the data, critically revised the manuscript, and gave final approval of the current version to be published. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Footnotes
This article has supplementary material available from erj.ersjournals.com
Conflict of interest: Disclosures can be found alongside this article at erj.ersjournals.com
Earn CME accreditation by answering questions about this article. You will find these at erj.ersjournals.com/journal/cme
- Received September 11, 2016.
- Accepted October 24, 2016.
- Copyright ©ERS 2017
References
