Adverse events in bedaquiline‐ and pretomanid‐based regimens for drug‐resistant tuberculosis from trial, implementation and pharmacovigilance studies

1. INTRODUCTION

Tuberculosis (TB) remains a leading cause of global morbidity and mortality. The emergence of drug‐resistant TB (DR‐TB) is a major impediment to effective disease control. ¹ Recent 2023 surveillance data indicated that from 3.4 million bacteriologically confirmed pulmonary TB cases tested for rifampicin resistance (RR), 5.5% exhibited resistance. Of these, 4.7% were classified as multidrug‐resistant or rifampicin‐resistant TB (MDR/RR‐TB), and 0.85% as either pre‐extensively drug‐resistant (pre‐XDR) or extensively drug‐resistant TB (XDR‐TB). ¹

Historically, DR‐TB has been treated with regimens lasting at least 18 months, involving four or more active drugs, often including injectable agents. However, these regimens have been associated with substantial toxicity, high treatment costs and poor patient adherence, ultimately contributing to increased morbidity and mortality. ² , ³ As a result, efforts have been directed towards developing shorter, all‐oral treatments with improved tolerability and comparable efficacy. ⁴ Hence, short‐course all‐oral regimens have been increasingly recommended for the management of MDR/XDR‐TB.

Clinical trials have examined various combinations of bedaquiline and pretomanid as a base regimen, combined with other oral TB drugs, including linezolid, moxifloxacin, pyrazinamide and clofazimine, with efficacy and safety profile data showing promising outcomes for the treatment of MDR/XDR‐TB. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Subsequently in 2022, the World Health Organization (WHO) recommended an all‐oral 6‐month BPaL(M) treatment regimen consisting of bedaquiline (B), pretomanid (Pa) and linezolid (L), with moxifloxacin (M) added if the susceptibility to these drugs are proven, as the preferred alternative to conventional all‐oral 9 to 18‐month regimens, with certain exceptions such as TB involvement of the central nervous system, osteoarticular system or disseminated (miliary) disease. ¹¹ , ¹²

The inclusion of the BPaL(M) regimens in the international clinical standard has been supported by robust evidence from three pivotal clinical trials—Nix‐TB, ZeNix and TB‐PRACTECAL. ⁶ , ⁸ , ¹³ These studies consistently demonstrated non‐inferior treatment success rates and lower incidences of adverse drug reactions when compared to 9‐month short‐course regimens and longer conventional regimens lasting 18 months or more in patients with MDR/RR‐TB. ⁷ , ⁸ , ¹⁴ The Nix‐TB and Zenix‐TB studies reported a favourable outcome ranging from 84% to 93%. ⁵ , ⁶ In the modified intention‐to‐treat population of the TB‐PRACTECAL study, the BPaL group achieved a higher favourable outcome (77%) compared to the standard care regimen (9–20 months) (52%). ⁸ Additionally, the BPaL regimen had fewer early treatment discontinuations and no deaths, resulting in a lower rate of unfavourable outcomes at 23%, compared to 48% with standard care. ⁸

Despite these promising outcomes, existing safety data for BPaL(M) regimen remain limited for two major reasons. ¹¹ Firstly, most current available safety data have been derived from controlled clinical trials, which may not fully capture the heterogeneity of patients in routine practice. Secondly, the absence of long‐term follow‐up limits understanding of delayed or cumulative toxicities. ¹¹ This limited safety data is of concern in TB management, where delayed identification and management of adverse drug reactions (ADRs) may lead to decreased medication tolerability, treatment discontinuation, extended treatment duration and potentially life‐threatening complications, thereby undermining treatment outcomes. ¹⁵

A recent systematic review (2024) analysed safety data of new shorter regimens for DR‐TB exclusively from clinical trials (Nix‐TB, Zenix‐TB and TB‐PRACTECAL) in Asia and Africa. ¹⁶ Likewise, a scoping review by Lopes et al. (2025) ¹⁵ examined updated TB regimens including treatment efficacy, adverse events, safety and socio‐economic impact, but only two included studies specifically addressed BPaL adverse events and safety. ¹⁵ To date, no comprehensive review has evaluated the safety profile of BPaL regimens using both clinical trial data and evidence from real‐world implementation studies.

In light of limited BPaL safety data, an updated and comprehensive review of the safety profile from clinical trials, implementation and pharmacovigilance studies is warranted. The outcomes of this review are expected to enhance the global understanding of the safety profile of bedaquiline and pretomanid‐based regimens, thereby enabling more informed therapeutic decisions and facilitating their integration into national and international TB treatment protocols.

2. MATERIALS AND METHODS

3. RESULTS

4. DISCUSSION

Our systematic review presented data focusing on AEs for bedaquiline–pretomanid‐based regimens, which consist of BPaL, BPaLM, BPaLC and BPaMZ combinations. These regimens were used for patients with DR‐TB, including RR, MDR, pre‐XDR and XDR‐TB. In contrast to previous reviews, ¹⁵ , ¹⁶ we analysed AEs not only from clinical trials but also from real‐world implementation of these regimens, encompassing data from implementation and pharmacovigilance studies from multiple countries. The integration of data from various settings yielded comprehensive information for healthcare professionals regarding AEs associated with these regimens in clinical settings. Therefore, this systematic review reflects AEs observed in a more heterogenous ‘real‐life’ population than just clinical trials alone. This information may help healthcare professionals gain a better understanding of the new regimens and provide comprehensive, patient‐centered care. ²⁸ , ²⁹

We identified variability in methods used to detect AEs (Table 1) which may have accounted for differences in the prevalence and AEs data presented between studies. Two widely used data standards for coding AEs in clinical trials, i.e., MedDRA and Common Terminology Criteria for Adverse Events (CTCAE), were used in the clinical trials. ³⁰ MedDRA comprises 26 top‐level hierarchies known as System Organ Classes (SOC). In contrast, CTCAE groups AEs into 28 categories based on anatomy and is embedded with a severity grading system (scale 1–5). ³⁰ In addition to these general standards coding, some clinical trials used specific standards for grading severity of AEs, that is, DAIDS severity grading consisting of five scales. ³¹ Differences in AE definitions and grading criteria between MedDRA and CTCAE likely contributed to differences in AEs between clinical trials. ³⁰ Similarly, different parameters for each scale for AEs severity grading between CTCAE ³⁰ and DAIDS ³¹ that different studies used may also have contributed to heterogeneity in AE results in this systematic review.

Active AE assessment methods were used in clinical trials ⁵ , ⁶ , ⁸ , ⁹ , ¹⁰ , ²⁰ with specific standard coding, while AE detection in real‐world settings relied more on passive reporting from patients, their families and healthcare professionals. ²² , ²⁶ In clinical trials, AEs were comprehensively recorded with thorough assessment by well‐trained healthcare teams and detailed management procedures to ensure patient safety as part of good clinical practice (GCP) guidelines. ⁵ , ⁶ , ⁸ , ⁹ , ¹⁰ , ²⁰ In contrast, AE detection in non‐trial clinical settings differed from this in many ways such as the absence of a standardised method in AE detection including the format and reporting system, lack of formal healthcare professional training in monitoring and management of AEs, and less resources than clinical trials which subsequently required healthcare professionals themselves to implement the specific AE monitoring system. ³² , ³³ , ³⁴ Consequently, we found the number of recorded AEs tended to be both higher and more detailed in clinical trials compared to implementation studies (Table 1). Our findings indicated that every patient in the clinical trials experienced AEs during treatment. ⁶ However, in implementation studies, the prevalence of AEs (maximum of 73%) was lower than that reported in clinical trials, ²⁵ which could be related to underreporting of AEs due to the methodological issues in monitoring AEs in clinical settings. However, among patients with AEs, the highest rate of developing serious AE conditions in clinical trials and implementation studies was relatively similar, at 30.2% and 24.5%, respectively (Table 1).

Our findings highlight limitations in AE reporting inherent to non‐clinical trial studies, including AE underreporting and variability in reporting quality, ³⁵ a bias towards reporting predominantly SAEs, ³⁶ and the use of a narrative system to grade the severity of AEs based on clinician assessment rather than established graded scales. ²⁷ A systematic review by Lopez‐Gonzales et al. (2009) ³⁶ found that healthcare professionals' knowledge and attitude towards AEs significantly contribute to their underreporting. These factors include a lack of knowledge of the urgency of AE reporting, a misconception that only SAEs require reporting, a lack of confidence in determining whether an AE is an ADR, and the belief that all ADRs were already known when the drug was approved for marketing. ³⁵ Therefore, adequate training for all healthcare professionals and consistent implementation of good pharmacovigilance practice (GVP) are needed to improve both the quantity and quality of AE reporting in clinical settings.

The WHO Global Tuberculosis Report 2025 documented an increase in the treatment success rate for patients with MDR/RR‐TB from 68% in 2021 to 71% in 2022. This represents a substantial improvement compared to 2012, when the success rate was only 50%. ³⁷ However, AE occurring during treatment continues to pose a significant challenge to achieving a favourable outcome. ³ , ³⁸ The WHO has introduced a framework for active drug‐safety monitoring and management (aDSM) specifically for new anti‐TB drugs and novel MDR‐TB or XDR‐TB regimens, ³⁹ , ⁴⁰ which included bedaquiline–pretomanid‐based regimens in our review. aDSM refers to active safety monitoring that was designed for systematic clinical and laboratory assessment of patients during their TB treatment. ⁴⁰ Recent implementation of aDSM in detection of RR‐TB AEs in a national multi‐centre study in Taiwan resulted in more comprehensive data of AEs. ³³ Additionally, recent clinical standards for AE management during TB treatment ⁴¹ emphasised that active surveillance by healthcare professionals using aDSM is important in assessing and monitoring AEs and can lead to timely and appropriate management of these AEs in their early stages. ⁴¹ We found that ≥56.5% of the patients recovered from the AEs. ¹⁰ , ²⁶ We expect that this recovery rate could be improved through the early detection and management of AEs. By identifying AEs at an early, less severe stage, healthcare professionals can implement management strategies and therapies more effectively, leading to better outcomes.

The WHO framework (2015) and Singh et al. study (2023) recommended aDSM for all SAEs and AESIs, including visual (optic neuritis) and hearing impairment/loss, peripheral neuropathy, psychiatric disorder, myelosuppression (anaemia, thrombocytopenia, neutropenia or leukopenia), prolonged QTc interval, liver enzyme abnormalities, lactic acidosis, hypothyroidism, hypokalaemia, pancreatitis, phospholipidosis and acute kidney injury/failure. ⁴⁰ , ⁴¹ All of these AESIs were recorded in detail in clinical trials, but not in implementation studies. Six types of AEs that were detected in almost all the included studies were increased liver enzymes, anaemia, gastrointestinal disorder, prolonged QTc interval, peripheral neuropathy and ophthalmic disorder (blurred vision and optic neuritis) (Table 2).

Our review revealed that the prevalence of AE related to the neuropsychiatric system, such as depression and suicide attempts, was significantly higher in implementation studies, affecting 25% of patients, ²¹ compared to only 4% in clinical trials. ⁸ A possibility to account for this important difference may be related to the strict selection criteria for patient inclusion in clinical trials, whereas real‐world studies generally include all patients with TB disease. A systematic review in 2020 ⁴² revealed that among patients with TB, the overall prevalence of depression was 45.1%. This rate was even higher in MDR‐TB, where 52.3% of patients experienced depression compared to 43.5% in non‐MDR‐TB. ⁴² These findings are consistent with another systematic review focused on patients with DR‐TB in South Asia, which reported a depression rate of 54%. ⁴³ Depression as an AE has been linked to TB drugs, including moxifloxacin, which was part of the regimens included in our review. Additionally, several factors may have also contributed to depression among patients with DR‐TB, such as disease‐related social stigma, the perception of patients living with an untreatable disease and discrimination due to their infection. ⁴⁴ , ⁴⁵ , ⁴⁶ Moreover, anxiety in patients with MDR‐TB may already be present even prior to commencing treatment. ⁴⁴ Therefore, our review identifies the importance of screening for neuropsychiatric disease before and during treatment and integrating mental health support when necessary.

Management implemented for AEs varied depending on the severity level of AEs, from monitoring the patient's condition only, prescribing medicine to treat AE symptoms with or without drug interruption, to withdrawal of the suspected drug that caused more severe AEs. ⁵ , ⁶ , ⁹ , ¹⁰ , ²⁰ , ²² , ²³ , ²⁶ For SAEs during BPaLM treatment, drug interruption during the treatment occurred in more than 50% of patients ²⁶ while withdrawal of the suspected drug reached 13%. ¹⁰ , ²⁶ The AE impact on TB treatment is demonstrated by a systematic review for the Asia region showing that adverse drug reactions doubled the odds of treatment discontinuation (OR 2.01; 95% CI:1.20–3.34). ³⁸

Another important consideration for MDR/RR‐TB treatment is that AEs may have prolonged effects, even after treatment completion. A systematic review and meta‐analysis found that long‐term monitoring of AEs after the completion of treatment is essential to detect any sequelae of AEs caused by TB drugs. ⁴⁷ A clinical trial ¹³ with patients followed up to 72 weeks and 108 weeks after the start of treatment, it was found that approximately 30% of patients experienced AEs related to anaemia, and hepatic and cardiac systems. ¹³ With this data, we recommend establishing safety monitoring not only during treatment but also for an extended period afterwards for patients who received bedaquiline–pretomanid‐based regimen, including the BPaL(M) regimen to look for post‐TB treatment sequelae. ⁴⁸

Considering the WHO's recommendation for the widely used BPaL(M) regimen for patients with MDR/RR‐TB ⁴⁹ and noting the discrepancies between active and passive AE monitoring as previously demonstrated, we agree with others in proposing the establishment of a practical and structured tool for detecting AEs in clinical settings. ²⁹ , ³² The Global Trigger Tool, ⁵⁰ developed by the Institute for Healthcare Improvement, may be a method of implementing active methods for AE identification following modification and adaptation for TB treatments. ⁵⁰ Further research is needed to adapt these trigger tools specifically for TB management, referencing the aDSM ³⁹ its implementation in clinical settings, ⁵¹ , ⁵² and clinical guidelines for AE management. ⁴¹

5. CONCLUSION

New shorter bedaquiline–pretomanid‐based regimens for DR‐TB treatment, particularly the BPaL(M) regimen, are becoming increasingly used worldwide and are associated with a high prevalence of AEs, even after treatment completion. Our review integrated and compared data of AEs of these new shorter regimens from clinical trials, implementation and pharmacovigilance studies and identified a trend of fewer‐reported AEs in implementation and pharmacovigilance studies, likely related to underreporting from spontaneous reporting methods used in real‐world settings rather than active AE monitoring methods in clinical trials, and a bias towards reporting more severe AEs only. Despite the existence of clear guidelines such as the aDSM by WHO and international clinical standards on AE monitoring during TB treatment, healthcare professionals may still face challenges in applying these methods in routine care. Therefore, we emphasise the need to establish an active AE‐monitoring method that can be easily utilised by healthcare professionals worldwide to enhance the standard of care in DR‐TB treatment and create a uniform framework for global data collection on AEs in RR/MDR‐TB treatment.

AUTHOR CONTRIBUTIONS

All authors contributed to this study. Conception and design study developed by NM, JGC and JWA. Data collection and analyses were performed by NM, ALL, IYH, JGC and JWA. The first draft was written by NM. All authors read and approved of the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interest.

Supporting information

Maria N, Larasati AL, Hamdani IY, Cho J‐G, Alffenaar J‐WC. Adverse events in bedaquiline‐ and pretomanid‐based regimens for drug‐resistant tuberculosis from trial, implementation and pharmacovigilance studies. Br J Clin Pharmacol. 2026;92(8):2607‐2623. doi: 10.1002/bcp.70541

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.