Abstract
Background Evaluation of novel anti-tuberculosis (TB) drugs for the treatment of multidrug-resistant (MDR)-TB continues to be of high interest on the TB research agenda. We assessed treatment outcomes in patients with pulmonary MDR-TB who received bedaquiline-containing treatment regimens in the Republic of Moldova, a high-burden MDR-TB country.
Method We systematically analysed the SIMETB national electronic TB database and performed a retrospective propensity score-matched comparison of treatment outcomes in a cohort of patients with MDR-TB who started treatment during 2016–2018 with a bedaquiline-containing regimen (bedaquiline cohort) and a cohort of patients treated without bedaquiline (non-bedaquiline cohort).
Results Following propensity score matching, 114 patients were assigned to each cohort of MDR-TB patients. Patients in the bedaquiline cohort had a higher 6-month sputum culture conversion rate than those in the non-bedaquiline cohort (66.7% versus 40.3%; p<0.001). Patients under bedaquiline-containing regimens had a higher cure rate assessed by both World Health Organization (WHO) and TBnet definitions (55.3% versus 24.6%; p=0.001 and 43.5% versus 19.6%; p=0.004, respectively), as well as a lower mortality rate (8.8% versus 20.2%; p<0.001 and 10.9% versus 25.2%; p=0.01, respectively). In patients who previously failed on MDR-TB treatment, >40% of patients achieved a cure with a bedaquiline-containing regimen.
Conclusions Bedaquiline-based MDR-TB treatment regimens result in better disease resolution when compared with bedaquiline-sparing MDR-TB treatment regimens under programmatic conditions in a country with a high burden of MDR-TB.
Abstract
Used under programmatic conditions in a high-burden MDR-TB setting, bedaquiline-based MDR-TB treatment regimens result in faster and more sustained disease resolution than bedaquiline-sparing MDR-TB treatment regimens https://bit.ly/37mSQOT
Introduction
Tuberculosis (TB) is the leading infectious cause of death worldwide [1]. Poor disease outcomes are being reported in patients with multidrug-resistant (MDR)-TB [2], defined by bacillary resistance to rifampicin and isoniazid, the two drugs considered to be the best available medicines for the treatment against TB [3]. The number of patients annually identified with MDR-TB has continuously increased during the last decades [4]. This is partially due to better diagnostic capacities, but also because of the increased spread of MDR strains of Mycobacterium tuberculosis [5, 6]. This situation generates an urgent need for the development of new TB drugs for MDR-TB treatment.
In the Republic of Moldova, an Eastern European country with one of the highest rates of MDR-TB worldwide (overall rate of MDR-TB among all TB cases ∼34%) [2], bedaquiline has been used for selected patients since the first availability in September 2016. In view of the 2019 recommendations for bedaquiline-based MDR-TB treatment regimens by the World Health Organization (WHO) and the American Thoracic Society/Centers of Diseases Control and Prevention/European Respiratory Society/Infectious Diseases Society of America, universal availability of bedaquiline for all patients with MDR-TB will be expected in the Republic of Moldova soon [7, 8]. In this context, data on the use of bedaquiline in routine clinical practice for the treatment of patients with MDR-TB continue to be a topic of high interest. Therefore, in the present study we have assessed treatment outcomes in MDR-TB patients who received bedaquiline-based treatment regimens and compared them with the outcomes in MDR-TB patients treated with bedaquiline-sparing regimens under programmatic conditions.
Methods
We performed a retrospective propensity score-matched cohort study based on data collected in the SIMETB (Sistemul Infromațional de Monitorizare și Evaluare al Tuberculozei) national electronic TB database in the Republic of Moldova. Our objective was to analyse treatment outcomes in the first cohort of patients who started a bedaquiline-based MDR-TB treatment regimen during 2016–2018 (bedaquiline cohort) and to compare them with those from a MDR-TB cohort treated without bedaquiline (non-bedaquiline cohort) from the same setting and period. In the study cohorts we assessed the rate of sputum culture conversion at 6 months of treatment, time to sputum culture conversion and final TB treatment outcomes applying WHO and TBnet definitions [3, 9].
Settings
In the initial phase of therapy, patients in the Republic of Moldova with MDR-TB are usually hospitalised at a single national medical institution, the Chiril Draganiuc Phtisiopneumology Institute, with two separate clinic sites in Chisinau and Vorniceni. Following discharge from the hospital, treatment of patients with MDR-TB is provided under direct observation at the place of residency. Anti-TB drugs and all necessary microbiological, laboratory and imaging tests required for the diagnosis of TB and treatment monitoring are provided free of charge by the government of the Republic of Moldova. Decisions regarding the initial treatment regimen and possible treatment adjustments are made by a national committee for the management of patients with MDR-TB, which meets at least once a week. All TB patients notified in the Republic of Moldova are registered in the SIMETB database. The database contains basic epidemiological, demographic, microbiological and treatment data of all TB patients in the country [10].
Treatment policy
Treatment for patients affected by MDR-TB in the Republic of Moldova is provided according to the National TB Treatment Protocol in accordance with the MDR-TB treatment guidelines of the WHO [11, 12]. During the study, all patients with MDR-TB initially started treatment with a standardised regimen of five second-line TB drugs including a fluroquinolone (levofloxacin or moxifloxacin), a second-line injectable (capreomycin or amikacin), ethionamide, cycloserine and pyrazinamide, which was adjusted, when necessary, once results of phenotypic drug susceptibility testing (DST) became available. The treatment duration is linked to the time of sputum culture conversion and consists of an initial intensive phase of 6–8 months followed by a continuation phase with a duration of 12–16 months. Usually, the MDR-TB treatment is initiated in the TB hospital where the patients is admitted at least until the sputum smear conversion is achieved and the final treatment scheme is decided. The treatment during hospital admission is supervised by a pneumologist specialised in the treatment of TB patients (phtisio-pneumologist). The average time of hospital admission is 2–3 months, after which the patient continues to receive treatment in the outpatient setting. As an outpatient, treatment is provided by the patient's family doctor or phtisio-pneumologist who has their office closest to the patient's domicile. The patient must come to the medical office daily to receive their medication. The standard treatment follow-up in MDR-TB patients includes monthly sputum culture, quarterly chest radiography and a set of blood tests on request of the treating physician [10].
In September 2016, bedaquiline became initially available for a limited number of the most severely affected patients. Since bedaquiline became more largely available, more patients with MDR-TB with resistance to later-generation fluoroquinolones or those who failed treatment on a non-bedaquiline-containing regimen have benefitted from the drug. After revision by the WHO of treatment recommendations for MDR-TB patients in 2019, indications for drug prescription became more permissive; in particular, bedaquiline started to be included in treatment regimens for newly diagnosed patients with MDR-TB also without additional resistance to second-line anti-TB drugs. All patients treated with bedaquiline received systematic ECG monitoring (on weeks 2, 4, 8 and 24 of treatment, and after that only on physician requests).
Study cohort selection
We used the SIMETB database to identify all MDR-TB patients who were diagnosed and initiated treatment during 2016–2018 in the Republic of Moldova. The selected study period was determined by the probability of MDR-TB treatment completion at the time of data extraction. Of the identified cases, adult patients (age ≥18 years) with culture-confirmed pulmonary TB who had finished their MDR-TB treatment at the time of data extraction (including patients who were lost to follow-up or died) were included in the study. MDR-TB patients without culture-confirmed TB, paediatric patients, patients with extrapulmonary TB and patients who were still on treatment at the time of data extraction were excluded from the study. MDR-TB patients who qualified according to the aforementioned inclusion and exclusion criteria were divided into bedaquiline and non-bedaquiline treatment cohorts. The resulting cohorts were matched by applying a propensity score matching algorithm [13], using age, sex, area of residence, cavitary lesion, HIV status, sputum smear positivity at baseline, fluoroquinolone resistance, past history of TB and drugs used as matching variables.
Ethical statement
The study protocol for this analysis was approved by Ethical Committee of the Chiril Draganiuc Phthisiopneumology Institute, Chisinau, Republic of Moldova (1/05.07.2019)
Statistical analysis
Statistical analysis was performed using R version 3.6.3 (www.r-project.org). Continuous variables are described as median with interquartile range (IQR). Categorical variables are presented as frequencies (percentages). Patients receiving bedaquiline were matched with patients without bedaquiline using propensity score matching. Using the R package MatchIt, the logit-based propensity score was calculated using the variables age, sex, residential area (rural/urban), presence of cavities, previous TB history, HIV, fluoroquinolone resistance, smear microscopy for TB at baseline as well as the use of drugs (fluoroquinolone, capreomycin, amikacin, linezolid, ethionamide, cycloserine, imipenem, p-aminosalicylic acid, pyrazinamide, ethambutol, isoniazid (high dose), clofazimine and delamanid) for each of the patients; on the basis of this, a comparison patient with the greatest possible overlap of the propensity was identified for each bedaquiline-receiving patient. The propensity score matching procedure was tested using the R package tableone and the resulting standardised mean differences. Continuous variables were checked for their distribution using the Shapiro–Wilk test. Statistical analysis was performed after matching using univariate logistic regressions to determine statistical significance. Afterwards, a post hoc power analysis for logistic regression was performed based on the achievement of culture conversion after 6 months for patients with and without bedaquiline with α=0.05 and a power of >0.8 was achieved. Prism version 8.0 (GraphPad, San Diego, CA, USA) was used for graphical presentation of the data.
Results
Study cohorts
2966 patients who started MDR-TB treatment during 2016–2018 were identified in the SIMETB database (including 203 cases with bedaquiline as part of the treatment regimen). The following were excluded from the study: 332 patients who were still on treatment at the time of data extraction, 96 paediatric patients (none of them received bedaquiline), 35 patients with extrapulmonary TB and 434 patients with negative or missing culture results at the time of diagnosis (figure 1). Thus, 2069 MDR-TB patients were included in the study analysis. Of them, 115 patients were attributed to the bedaquiline cohort and the other 1954 to the non-bedaquiline cohort. After propensity score matching, 114 patients were assigned to each of the two study cohorts. There were no significant differences among the cohorts by main demographic and clinical features, such as age, sex, place of residence, positive smear microscopy, radiographic cavitary lesion, resistance to fluoroquinolone and previous history of TB (table 1).
Study flowchart. MDR-TB: multidrug-resistant tuberculosis.
Cohort characteristics
Treatment regimens
Drugs included in the treatment schemes of the patients from both cohorts are presented in table 2. After propensity score matching, both cohorts were comparable by all drugs used, except for bedaquiline.
Drugs included in multidrug-resistant tuberculosis treatment regimens in the two cohorts
In all patients bedaquiline was started during hospital admission. In 21 patients bedaquiline was part of the initial treatment scheme, while in the other 93 it was used for the adjustment of the initial treatment scheme once DST results became available. 50.9% of the patients who received bedaquiline were treated after at least one MDR-TB treatment episode failure. In 50 out of 93 (53.8%) patients in whom bedaquiline was used for adjustment of the initial treatment scheme it was the single drug added to or substituted in the regimen. The median (IQR) time of bedaquiline administration was 34 (24–40) weeks. An extended duration of bedaquiline intake was decided by the committee for the management of MDR-TB cases on an individual approach base in patients with slow radiographic improvement, particularly in the case of persistent cavities, and also in those with persistent positive sputum culture results at 5 or 6 months of treatment, as well as in case of previous history of repeated MDR-TB treatment failure. In 15 patients bedaquiline administration was stopped during the first 20 weeks of drug intake (in 10 patients it was interrupted on the patient's initiative after hospital discharge, in four patients it was stopped by local TB physicians and in one patient it was due to adverse events).
Bedaquiline-associated adverse drug events
Only adverse events that lead to stopping drug intake are recorded in the SIMETB database. 55 such episodes (in 43 patients (37.8%)) were registered in the database for patients from the bedaquiline cohort and 50 episodes (in 38 patients (43.8%)) in the non-bedaquiline cohort. Of these, only one was attributed to bedaquiline (a case of dizziness that led to discontinuation of bedaquiline after 7 weeks of treatment). No critical prolongation of the QTc interval (Fridericia corrected QTc >500 ms) was observed in any of the 114 patients who received bedaquiline.
Comparison of treatment outcomes
The rate of sputum culture conversion at 6 months of treatment in the bedaquiline cohort was higher than that in the non-bedaquiline cohort (66.7% versus 40.3%; p<0.001) (figure 2 and table 3). In addition, sputum culture conversion was achieved significantly faster in the bedaquiline cohort (during the first month of treatment) than in the non-bedaquiline cohort (median (IQR) time to sputum culture conversion in the bedaquiline cohort <1 (0–1) versus 1 (0–3) months in the non-bedaquiline cohort; p<0.001) (figure 2).
a) Culture conversion and b) reversion to positive culture in patients with multidrug-resistant tuberculosis who received bedaquiline- and non-bedaquiline-containing treatment regimens.
Treatment outcomes in the two cohorts by World Health Organization (WHO) and simplified TBnet definitions
When final treatment outcomes were assessed by WHO definitions, a higher treatment success rate and a lower mortality rate were found in the bedaquiline cohort (55.3% versus 24.6%; p<0.001 and 8.8% versus 20.2%; p<0.001, respectively). At the same time, the treatment failure rate was similar (27.2% versus 29.8%; p=0.7) (figure 3a and table 3).
Treatment outcomes according to a) World Health Organization and b) TBnet criteria in patients with multidrug-resistant tuberculosis who received bedaquiline- and non-bedaquiline-containing treatment regimens.
Assessment of treatment outcomes by simplified TBnet definitions was possible in patients in whom at least 1 year had passed after treatment completion, which was the case in 92 out of 114 (80.7%) patients in the bedaquiline cohort and 107 out of 114 (88.6%) patients in the non-bedaquiline cohort. Patients from the bedaquiline cohort had a higher cure rate and a lower mortality rate (43.5% versus 19.6%; p=0.004 and 10.9% versus 25.2%; p=0.01, respectively). When applying TBnet treatment outcome definitions, the cure rate was lower compared with WHO criteria assessment, and treatment failure rates were significantly higher in both the bedaquiline and non-bedaquiline cohorts (figure 3b and table 3).
Treatment success in previous treatment failures in the bedaquiline cohort
The bedaquiline cohort included 58 (50.9%) patients with previous failure of at least one MDR-TB treatment regimen. The culture conversion rate at 6 months in these previously failing patients was 55.2% (32 out of 58), while the cure rate was 41.4% (24 out of 58) when assessed by WHO definitions and 31.4% (16 out of 51) when assessed by TBnet definitions (table 4).
Treatment outcomes in patients who failed previous tuberculosis treatment episodes in the bedaquiline cohort by World Health Organization (WHO) and simplified TBnet definitions
The 6-month culture conversion rate in patients treated with bedaquiline without a history of previous MDR-TB treatment failure (56 patients) was 82.1%, while the cure rate was 69.6% (39 out of 56) when assessed by WHO definitions and 58.5% (24 out of 41) when assessed by TBnet definitions (table 4); both were significantly better than for those in patients with previous treatment failure (p=0.003 and p=0.01, respectively).
Discussion
We evaluated treatment outcomes in patients with MDR-TB in the Republic of Moldova, a high-burden MDR-TB country in Eastern Europe, who received bedaquiline as part of their treatment regimen and compared them with treatment outcomes observed in patients with MDR-TB from the same country treated without bedaquiline. We found that patients treated with a bedaquiline-based treatment regimen had faster culture conversion and a higher 6-month sputum culture conversion rate, a higher cure rate and a lower mortality rate compared with patients treated without bedaquiline. Patients who failed previous MDR-TB treatment regimens had a higher chance of cure when they received a bedaquiline-based treatment regimen.
In previous studies, the rate of 6-month sputum culture conversion in patients treated with a bedaquiline-containing regimen varied from 64% to 100%, with treatment success rates ranging between 52% and 85%, both outcomes being higher in patients treated with bedaquiline when compared with those without bedaquiline [14–26]. Similarly, patients treated with bedaquiline in our study had better 6-month sputum culture conversion rates and higher treatment success rates than those who did not receive bedaquiline. The added value of our data consists in the comparison of primary and final treatment outcomes in bedaquiline-treated patients with propensity score-matched MDR-TB patients treated without bedaquiline-containing regimens from the same setting and comparable treatment regimens. The omission of such a comparator was reported as a limitation of previously published observational cohort studies [20, 22, 24]. Data from the present analysis support the efficacy and feasibility of the use of bedaquiline-containing regimens when applied under programmatic conditions in a high-burden MDR-TB setting.
The observed treatment success rate in the present cohort is among the lowest of those previously reported (55.3%), despite treatment regimens that also included linezolid and imipenem, drugs which are known to be associated with a higher rate of treatment success [27]. Patients from both cohorts could be characterised as exceedingly difficult to treat patients. Thus, 73% of the patients were infected with a strain of M. tuberculosis resistant to fluoroquinolones, 81% had cavitary lesions on chest radiography and 80% were retreatment cases; all of these factors are known as predictors for poor treatment outcomes in patients with MDR-TB [27]. Treatment outcomes could also have been influenced by poor patient adherence to treatment, an important challenge for TB treatment in Eastern Europe. Even though TB treatment within the National TB Program in the Republic of Moldova is provided under direct observation of medical staff, it is hard to be confident about patients swallowing the pills, particularly in an outpatient setting [28]. Poor adherence during the previous and current episodes of TB could also explain repetitive failure among previously treated cases in both study cohorts, 64% of which were retreatments after failure. Several interventions have been implemented during recent years to improve TB treatment adherence in the Republic of Moldova, but their scale-up to the level of the entire country has not yet been achieved [29, 30].
One of the factors which could also have an influence on treatment outcome is the duration of hospital admission in both study cohorts. Due to the retrospective nature of the study, data on duration of hospital admission were not available for the analysis and this represents a limitation of the results presented.
We applied WHO outcome definitions, which are currently used by the National TB Program in the Republic of Moldova, and those proposed by TBnet [3, 9]. Application of TBnet definitions allowed us to address possible disease relapse during the first year of post-treatment follow-up in both cohorts [31–34]. The higher rate of failure observed when the TBnet definitions were applied was generated by reassessment of cases lost to follow-up which had positive sputum culture when they restarted treatment after several months of interruption [34].
Bedaquiline-containing MDR-TB regimens were associated with an increased risk of death in a phase 2b clinical trial [14]. However, a later cohort analysis from South Africa reported a significantly lower mortality rate associated with bedaquiline-containing regimens compared with a standard regimen [35]. We also observed a lower mortality rate in our bedaquiline patients, which was not previously reported in bedaquiline-treated patients from an Eastern European setting.
Poor access to important TB drugs in Eastern European high-burden MDR-TB countries has been previously reported [33]. Limited use of delamanid and clofazimine in both the bedaquiline and non-bedaquiline cohorts in this study reflects the same phenomenon. Among the main causes of a slow scale-up of the approved TB drugs in countries like the Republic of Moldova are limited budgets of the public healthcare sector and subsequent high dependence on external funding, as well as a high level of bureaucracy and sometimes reticence of public authorities [36]. Of note, only in September 2016 did the first patient receive bedaquiline as part of a MDR-TB regimen in the Republic of Moldova, over 3 years since its international approval. However, release by the WHO of new treatment recommendations on an all-oral treatment regimen for patients with MDR-TB has improved the availability of anti-TB medicines. Delamanid and clofazimine have recently become available for a larger number of patients. We expect to observe an increase in treatment success rates once the new all-oral bedaquiline-based regimens become available to all patients with MDR-TB in the Republic of Moldova.
A critical finding of this study is the high rate of patients who received bedaquiline as a single additive drug for the adjustment of the MDR-TB treatment regimens. This situation is also a consequence of the reduced access to some of the MDR-TB drugs in resource-limited settings such as the Republic of Moldova. This practice may be associated with the future development of additional resistance against bedaquiline [27]. Unfortunately, we were unable to assess the acquisition of bedaquiline resistance during TB treatment in this cohort due to the study's retrospective nature and missing DST data on treatment follow-up cultures in routine practice. An additional risk for the development of bedaquiline resistance under therapy is represented by patients who discontinue bedaquiline intake after several weeks of medication following hospital discharge (14 patients (12.3%) in our cohort). Diversification of the drug supply and strengthening patient management in the outpatient setting are critically important to prevent the development of bacillary resistance to novel anti-TB drugs in the Republic of Moldova and countries with similar settings in the region.
Missing data about all adverse events limit our conclusion regarding the overall safety of the bedaquiline-containing regimens. The overall low rate of adverse events observed in both cohorts could be a consequence of underreporting, a limitation that could not be overcome in this retrospective study. However, the present data confirm the absence of critical cardiac adverse effects in patients receiving bedaquiline despite the concomitant use of other drugs also causing QTc interval prolongation. Being in concordance with previously reported data [37, 38], this finding confirms the low risk of severe adverse events of bedaquiline and proves its safety under programmatic conditions in a high-burden MDR-TB setting.
Conclusions
Bedaquiline-based MDR-TB treatment regimens result in faster and better disease resolution when compared with bedaquiline-sparing MDR-TB treatment regimens under programmatic conditions in a high-burden MDR-TB country, even in exceedingly difficult to treat patients. More than 40% of patients previously failing MDR-TB treatments achieve cure under bedaquiline-based therapy regimens. Efficient antibiotic stewardship measures should be implemented to prevent avoidable cases of newly acquired resistance to novel anti-TB drugs.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-02544-2020.Shareable
Footnotes
This article has an editorial commentary: https://doi.org/10.1183/13993003.00066-2021
Conflict of interest: D. Chesov has nothing to disclose.
Conflict of interest: J. Heyckendorf reports personal fees for lectures from Chiesi, Gilead, Janssen and Lucane, outside the submitted work.
Conflict of interest: S. Alexandru has nothing to disclose.
Conflict of interest: A. Donica has nothing to disclose.
Conflict of interest: E. Chesov has nothing to disclose.
Conflict of interest: M. Reimann has nothing to disclose.
Conflict of interest: V. Crudu has nothing to disclose.
Conflict of interest: V. Botnaru has nothing to disclose.
Conflict of interest: C. Lange reports personal fees for lectures from Chiesi, Gilead, Janssen, Lucane, Novartis, Oxoid, Berlin-Chemie and Thermo Fisher, personal fees for meeting attendance from Oxford Immunotec, outside the submitted work.
- Received July 1, 2020.
- Accepted November 23, 2020.
- Copyright ©ERS 2021