Effectiveness of the Bedaquiline, Pretomanid, and Linezolid Regimen, With or Without Moxifloxacin, for Drug-Resistant Tuberculosis in Ukraine Under Programmatic Conditions

Abstract

Bedaquiline, pretomanid, and linezolid, with or without moxifloxacin, have revolutionized treatment for drug-resistant tuberculosis. In Ukraine, we observed an 86% success rate (394 of 460) under programmatic conditions when drugs were accessible, exceeding prior estimates (58%) and aligning with trial findings (80%–90%). However, access and implementation remain a concern in conflict-affected settings.

Graphical Abstract

Graphical Abstract.

Bedaquiline, pretomanid, and linezolid (BPaL), with or without moxifloxacin—BPaL(M), have revolutionized treatment of multidrug-resistant (MDR), pre–extensively drug-resistant (pre-XDR), and extensively drug-resistant (XDR) tuberculosis [1]. The landmark trials TB-PRACTECAL, NiX-TB, and ZeNix demonstrated treatment success rates of up to 90% for MDR tuberculosis. Simultaneously, they have reduced pill burden, improved tolerability, and shortened treatment from 2 years to 6–9 months [1]. Although these findings are promising, limited data are available on the performance of BPaL(M) in real-world settings, outside controlled clinical trials, and under programmatic conditions. Data are available from only a few countries, mostly with low numbers of patients [2–7].

Ukraine is a country with a high tuberculosis incidence (112 per 100 000 population in 2023) and a high prevalence of drug resistance, with 28% of new bacteriologically confirmed pulmonary tuberculosis cases being MDR [8]. In November 2021, Ukraine became the first country in the world to adopt the BPaL regimen under operational research, as recommended by the World Health Organization (WHO). As of 19 January 2023, the Ministry of Health of Ukraine has mandated using BPaL(M) for all eligible patients (order no. 102). However, the country faces critical obstacles to tuberculosis control due to the war. Below, we describe the treatment outcomes of BPaL(M) for highly drug-resistant tuberculosis in Ukraine.

METHODS

We conducted a retrospective study of prospectively collected data stored in the multinational collaborative National Institute of Allergy and Infectious Diseases TB Portals Program database (https://www.niaid.nih.gov/research/tb-portals). The study included patients initiating BPaL treatment, with or without moxifloxacin, for bacteriologically confirmed MDR, pre-XDR, or XDR tuberculosis in 14 oblasts (regions; see Figure 1) of Ukraine in 2023. Treatment was initiated in the hospital and continued as outpatient treatment after smear or culture conversion. The decision to treat, including the duration, was made by the Regional Drug-Resistant Tuberculosis Treatment Committee, which includes at least three tuberculosis physicians. Drug susceptibility testing was performed using phenotypic and genotypic methods, and drug resistance was defined according to 2021 WHO criteria [9].

Figure 1.

Treatment success rates among patients with highly drug-resistant tuberculosis starting treatment in 2023 with bedaquiline, pretomanid, linezolid, with or without moxifloxacin, by Ukrainian oblasts: Cherkaska, 89% (95% confidence interval, 72%–96%); Dnipropetrovska, 81% (68%–90%); Ivano-Frankivska, 60% (17%–93%); Kharkivska, 82% (69%–91%); Khmelnytska, 96% (78%–100%); Kyiv City, 94% (71%–100%); Mykolaivska, 88% (72%–96%); Sumska, 80% (65%–90%); Ternopilska, 90% (67%–98%); Vinnytska, 83% (60%–94%); Volynska, 91% (74%–98%); Zakarpatska, 93% (76%–99%); Zaporizka, 83% (70%–91%); and Zhytomyrska, 85% (64%–95%). The 10 oblasts most affected by the war were identified by HeRAMS Ukraine [12]. The map was created with www.datawrapper.de (Datawrapper GmbH, Germany).

Patients were followed up until treatment completion, cure, failure, lost-to-follow-up, or death, according to WHO outcome definitions [10]. BPaL was administered for 6–9 months and BPaL with moxifloxacin (BPaLM) for 6 months. Treatment success was defined as the sum of “treatment completion” and “cured” results. Statistical analyses were performed using R software, version 4.2.3. The Kharkiv National Medical University Ethics Committee, Kharkiv, Ukraine granted study approval (5 February 2020; no. 1).

RESULTS

In total, 658 patients with MDR, pre-XDR, or XDR tuberculosis were identified. Of these, 198 were excluded: in 107 patients, BPaL(M) drugs were temporarily inaccessible, 86 did not have drug susceptibility testing results to determine drug resistance pattern, and 5 were not treated with BPaL(M). Consequently, 145 patients (32%) treated with BPaL and 315 (69%) treated with BPaLM were included. Of these, 310 (67%) had MDR, 149 (32%) had pre-XDR, and 1 (0.2%) had XDR tuberculosis. Most included patients with MDR tuberculosis received BPaLM (98% [n = 303]), while most patient with pre-XDR tuberculosis were treated with BPaL (92% [n = 137]). Most patients were male (82% [n = 376]); the median age (interquartile range [IQR]) was 44 (37–52) years and the median body mass index 20.7 (18.7–22.7) (calculated as weight in kilograms divided by height in meters squared); and 12% (n = 56) were coinfected with human immunodeficiency virus (HIV).

The median follow-up time (IQR) from treatment initiation to outcome was 181 (179–184) days. A successful treatment outcome was observed for 86% of patients (95% confidence interval [CI], 82%–89%) (n = 394), which varied by region (from 60% in Ivano-Frankivska to 94% in Kyiv City) (Figure 1). The outcomes were as follows: 338 patients (74%) were cured, 56 (12%) completed treatment, 19 (4.1%) experienced treatment failure, 19 (4.1%) were lost to follow-up, and 28 (6.1%) died during treatment. Seven patients (1.5%) experienced adverse events requiring a change in drug regimen. The median time to first sputum culture conversion (IQR) was 7.7 (4.7–10.5) weeks. Among smear-positive patients, the median time to first smear conversion (IQR) was 7.9 (4.6–12.8) weeks.

The success rates for MDR tuberculosis (84% [95% CI, 80%–88%]) and pre-XDR tuberculosis (89% [83%–94%]) did not differ significantly (Fisher exact test, P = .2). Similarly, outcomes did not differ significantly between those treated with BPaL (88% [95% CI, 82%–93%]) or BPaLM (84% [80%–88%]; Fisher exact test, P = .3).

In a sensitivity analysis including 107 patients who experienced treatment interruptions due to temporary unavailability of ≥1 BPaL(M) drug, we applied a stricter definition of treatment failure. Specifically, any modification of the initial BPaL(M) regimen resulting from drug unavailability (eg, discontinuation or substitution) was considered a treatment failure, in addition to the standard WHO-defined criteria. Using this more conservative approach, the overall treatment success rate declined to 70% (95% CI, 66%–74%). Of these 107 patients, 81 (76%) had MDR, 24 (22%) had pre-XDR, and 2 (1.9%) had XDR tuberculosis.

In a multivariable binomial regression, higher body mass index was independently associated with a lower risk of an unsuccessful outcome (relative risk, 0.91 [95% CI, .84–.98]; P = .01) while male sex showed a trend toward increased risk (1.91 [.99–4.45]; P = .08). Other factors included in the model—age, HIV status, and a history of tuberculosis—were not significantly associated with the outcome.

DISCUSSION

We observed a treatment success rate of 86% among patients initiating BPaL(M) for highly drug-resistant tuberculosis in Ukraine in 2023, a result comparable to the outcomes reported in the clinical trials on which the regimen is based. This represents a substantial improvement over our earlier estimate of a 58% (95% CI, 56%–60%) success rate among patients treated in the same regions from 2020 to 2021, as reported elsewhere [11]. Although most patients in that earlier cohort were treated before the implementation of BPaL in Ukraine, 24 patients did receive BPaL, all with successful outcomes.

Similar results have been reported from other settings. In Kyrgyzstan, a study of 50 patients treated with the 6-month BPaL regimen from 2021–2022 reported a success rate of 84% [3]. In Southeast Asia, a case series from Thailand (2021–2023) documented an 82% treatment success rate among 28 patients treated with BPaL(M) [4], while a study from Pakistan involving 116 patients reported a 95% success rate with both BPaL and BPaLM [6]. Interim data from 5 countries (Indonesia, Kyrgyzstan, Philippines, Uzbekistan, and Vietnam) showed a 91% success rate among 308 patients receiving BPaL for 6 months—comparable to clinical trial outcomes [5]. In the United States, 97% of 70 patients treated with BPaL between 2019 and 2022 completed treatment, with only 2 reported relapses [2]. A recent large-scale BPaL(M) study from Belarus and Uzbekistan reported >90% success among >600 patients with MDR or pre-XDR tuberculosis [7].

Taken together, our data align with the success rate observed in other studies, supporting the effectiveness of BPaL(M) under real-world conditions. However, when we included patients who experienced treatment interruptions due to drug unavailability, the success rate dropped significantly, to 70%. This lower rate is likely driven by ongoing war-related disruptions and limited BPaL(M) accessibility. Notably, 4 of the 6 oblasts with observed success rates <85% were among the 10 most affected regions in terms of access to critical health resources and services in 2023–2024, according to the WHO Health Resources and Services Availability Monitoring System [12]. Furthermore, 47% of patients who experienced BPaL(M) interruptions were from one of these 10 regions. Additional contributing factors—unrelated to the war—may include the high cost and limited availability of BPaL(M) drugs, which have been reported as critical obstacles to tuberculosis control across Europe [13, 14].

The current study was limited by its regional scope, as it did not encompass the entire country, although it covered a substantial part of Ukraine. Regional differences in outcomes raise the possibility of overestimation or underestimation of the regimen's effectiveness due to geographic variability. Pretomanid and, temporarily, moxifloxacin were unavailable nationwide from June to early September due to supply interruptions, preventing the inclusion of new patients during this period. Moxifloxacin was successfully redistributed across the country and was also available from local pharmacies. While BD MGIT BACTEC was the primary culture method used, Löwenstein-Jensen medium was occasionally used due to resource constraints and supply disruptions. Different culture methods may have introduced variability in case detection and outcome classification, as MGIT is more sensitive and detects growth more rapidly than Löwenstein-Jensen, which could have led to underestimation of treatment failure in some cases. Finally, while we observed an overlap between regions with lower treatment success rates and those most affected by the war, these unadjusted estimates may be influenced by many confounding factors. Therefore, the observed association does not necessarily imply causality. Nonetheless, the study included many patients treated under real-world programmatic conditions, offering insight into BPaL(M) performance outside trial settings.

Since December 2022, the WHO has recommended using the BPaL(M) regimen under programmatic conditions, marking a milestone in tuberculosis treatment. This advancement will potentially transform tuberculosis management globally. However, sustained efforts are needed to ensure equitable access, maintain uninterrupted drug supply chains, and monitor long-term outcomes, especially in regions affected by conflict. Ongoing and emerging challenges include the high cost of BPaL(M) drugs, limited accessibility, and the threat of drug resistance, all of which jeopardize the long-term success of this novel regimen. The data presented here not only highlight these concerns but also underscore the challenges of managing tuberculosis during war or other humanitarian crises.

Notes

Author contributions. V. N. D. and D. B. conceived the aim and design of the study (using already-existing data from the TB Portals Program database). V. N. D. performed data management and analyses and wrote the first draft, while O. S. P. and D. B. provided feedback for the initial revision. The other authors and Ukraine TB-Portal Study Group members contributed to data collection and provided intellectual content. All authors approved the submission of the final manuscript.

Data availability. Data can be requested from the National Institute of Allergy and Infectious Diseases TB Portals Program database (https://www.niaid.nih.gov/research/tb-portals).

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases/US Civilian Research & Development Foundation (grant DAA9-19_66199 to D. B.) and the Ministry of Health of Ukraine (grant 0123U100178 to D. B.).