Incidence and risk factors of adverse drug reactions in multidrug-resistant/rifampicin-resistant tuberculosis (MDR/RR-TB) regimens containing new drugs: a retrospective study of two national multicenter cohorts

Introduction

Tuberculosis remains a major global infectious threat. According to the World Health Organization (WHO) 2024 report, there were 10.8 million new tuberculosis patients globally in 2023, with an incidence rate of 134 per 100,000 population, among which approximately 400,000 cases were multidrug-resistant tuberculosis (MDR-TB) or rifampicin-resistant tuberculosis (RR-TB) (1). MDR/RR-TB treatment regimens have been characterized by high complexity, long treatment duration, significant drug toxicity, and poor patient adherence, leading to treatment interruption, increased recurrence rate, and poor prognosis (2). New drugs (bedaquiline, delamanid) have advanced MDR/RR-TB therapy—bedaquiline inhibits mycobacterial adenosine triphosphate (ATP) synthase, while delamanid targets cell wall synthesis—improving outcomes for patients.

Despite their promising efficacy, both bedaquiline and delamanid are associated with a range of adverse drug reactions (ADRs), particularly QT interval alterations (3). ADRs occur in 20–100% of MDR/RR-TB patients, involving multiple organ systems and potentially compromising treatment adherence or causing severe outcomes (4,5). Existing studies on regimens containing new drugs are limited by small sample sizes, single-center designs, or incomplete monitoring (6-8). This retrospective analysis of two national multicenter cohorts (bedaquiline and delamanid) aims to explore ADR incidence, characteristics, and risk factors, providing evidence for individualized clinical management. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2411/rc).

Methods

Study subjects

A retrospective study design was adopted, and MDR-TB cohorts containing new drugs were included between 2017 and 2022, including: China’s first bedaquiline-containing MDR-TB treatment cohort (Bedaquiline Introduction and Protection Program, hereinafter referred to as the “bedaquiline cohort”) and the first delamanid-containing MDR-TB treatment cohort (Study on the Safety and Efficacy of Delamanid-Containing Regimens for MDR-TB, hereinafter referred to as the “delamanid cohort”). The inclusion criteria for patients in both cohorts were based on WHO guidelines, bedaquiline package insert, and delamanid package insert at the time of the study. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Both cohorts in this study were approved by the Medical Ethics Committee of Beijing Chest Hospital, Capital Medical University. The bedaquiline cohort [approval No. (2019) Lin Shen No. (2018-05-04)] and the delamanid cohort [approval No. (2020) Lin Shen No. (09-01)] were separately reviewed and authorized and informed consent was taken from all the patients.

Bedaquiline cohort

This cohort included eligible tuberculosis patients from 54 hospitals in China between February 2018 and December 2019. A total of 1,546 patients were initially included, and 1,543 were finally included (3 patients were excluded due to untraceable missing key data). The inclusion criteria were as follows: (I) patients with laboratory-confirmed MDR-TB who required further treatment; (II) aged 18 years or older; (III) according to WHO MDR-TB treatment guidelines, bedaquiline was necessary to form an effective regimen, including but not limited to patients with XDR-TB, pre-XDR-TB, newly diagnosed drug-resistant TB who could not use fluoroquinolones or injectables, and MDR-TB patients with previous treatment failure; (IV) no significant arrhythmia, with an electrocardiogram (ECG) QTc ≤450 ms; (V) ability to comply with treatment and monitoring requirements, including reporting adverse reactions; (VI) signed informed consent form. The treatment regimen involved bedaquiline combined with 3–4 effective anti-tuberculosis drugs based on previous treatment history and drug sensitivity results. The treatment course included a 24-week intensive phase with bedaquiline and background regimen drugs, followed by the discontinuation of bedaquiline and continuation of the background regimen, lasting a total of 11–18 months, with negative sputum cultures at the end of the intensive phase and five consecutive negative sputum cultures at intervals of at least 4 weeks.

Delamanid cohort

This cohort included eligible tuberculosis patients from 26 hospitals nationwide (with Beijing Chest Hospital as the leading unit) between June 2020 and December 2022. A total of 611 patients were initially included, and 608 were finally included (3 patients were excluded due to untraceable missing key data). The inclusion criteria were as follows: (I) patients with laboratory-confirmed MDR/RR-TB; (II) aged 18–65 years; (III) had not started MDR-TB treatment, or had started treatment but required further intensive treatment; (IV) according to the WHO Consolidated Guidelines on Drug-Resistant Tuberculosis Treatment and the Chinese Expert Consensus on the Treatment of Multidrug-Resistant and Rifampicin-Resistant Tuberculosis (2019 version), delamanid was necessary to form an effective treatment regimen based on drug sensitivity test results and previous treatment history; no respiratory failure, no history of heart failure, and no clinically significant arrhythmia, with an ECG QTcF <450 ms; (V) ability to take medications as required by the study during treatment and follow-up, complete treatment monitoring, and promptly report adverse reactions to the attending physician; (VI) signed informed consent form. The treatment regimen involved delamanid combined with 3–4 effective anti-tuberculosis drugs based on guidelines and patient tolerance. The treatment course included a 24-week intensive phase with delamanid and background regimen drugs, followed by the discontinuation of delamanid and continuation of the background regimen, lasting a total of 11–18 months, with five consecutive negative sputum cultures at the end of the intensive phase and an interval of at least 1 month between cultures.

Study methods

Data collection and standardization

Patient information [gender, age, body mass index (BMI), treatment history, comorbidities, drug resistance type, etc.], therapeutic drug use, and ADR-related data were extracted from the electronic data capture (EDC) system. Data cleaning and standardization were performed according to unified standards to ensure data integrity and accuracy. Patients with missing key data that could not be traced were excluded from the study.

Definition and grading of ADRs

ADR types were standardized based on clinical symptoms according to the definition of the WHO-Uppsala Monitoring Centre (WHO-UMC) (9), and ADRs were classified by involved organ/system (Table 1). Meanwhile, according to the Division of AIDS (DAIDS) 2017.V2.1 classification standard (10), the severity of ADRs was graded into 1–5 levels: grade 1: mild, no intervention required; grade 2: moderate, symptomatic treatment required; grade 3: severe, drug discontinuation or hospitalization required; grade 4: life-threatening; grade 5: death. Definitions of XDR-TB and Pre-XDR-TB in this study are based on WHO definitions before 2019 (11).

Statistical analysis

Categorical data were described as “percentage (%)”, and the Chi-squared test (χ² test) was used for univariate analysis to compare the distribution differences of categorical variables between the ADR group and non-ADR group. The incidence of ADRs and their associated factors were analyzed. For continuous variables, independent samples t-test or Mann-Whitney U test was used. Potential risk factors with P<0.05 were included in multivariate analysis. Both univariate and multivariate analyses were performed for the top three most common ADRs observed during treatment. A multivariate Logistic regression model was constructed, with ADR occurrence as the dependent variable and variables with statistical significance in univariate analysis as independent variables. The backward stepwise method [Backward likelihood ratio (LR)] was used to exclude confounding factors, and independent risk factors with P<0.05 were finally retained. SPSS 27.0 statistical software was used for data analysis.

Results

General information of patients

The study included 2,151 MDR/RR-TB patients (bedaquiline cohort: n=1,543; delamanid cohort: n=608). Among them, 1,459 (67.8%) were male and 692 (32.2%) were female; 977 (45.4%) patients were under 35 years old, 1,127 (52.4%) were aged 35–<65 years, and 47 (2.2%) were 65 years old or above; 612 (28.5%) patients had a BMI <18.5 kg/m². There were 260 (12.1%) patients with comorbid diabetes, 215 (10.0%) with comorbid extrapulmonary tuberculosis, and 1,301 (60.5%) with pulmonary cavity formation. Distribution of drug-resistance types: 1,081 (50.3%) cases were RR-TB/MDR-TB, 740 (34.4%) were pre-XDR-TB, and 330 (15.3%) were XDR-TB (Table 2). This study involved a total of 21 anti-tuberculosis drugs (Table 3).

Occurrence of ADRs

Among 2,151 patients, 1,209 experienced at least one ADR (overall incidence 56.2%). Incidence was higher in the bedaquiline cohort (62.2%, 961/1,543) than the delamanid cohort (45.7%, 278/608). ADRs involved multiple organ systems. Top five by incidence: cardiovascular [29.1%, 625/2,151, primarily QTc prolongation (27.4%, 589/2,151)], hepatic (20.9%, 450/2,151), haematological (12.6%, 272/2,151), neurological (9.1%, 196/2,151), and metabolic/endocrine (6.4%, 138/2,151).

A total of 3,405 ADR episodes were recorded, among which grade 3 or above ADRs accounted for 1,827 episodes (53.7%), and grade 5 ADRs accounted for 27 episodes (0.8%). A total of 28 patients died, with causes including acute heart failure, acute renal failure, respiratory failure, severe pneumonia, gastrointestinal bleeding, severe malnutrition, shock, and suicide; the cause of death was unrecorded in 10 patients (Table 4).

Univariate analysis of ADRs

Univariate analysis showed female sex (P=0.01), age 35–<65 years (P<0.001), BMI <18.5 kg/m² (P=0.02), retreatment history (P<0.001), diabetes (P=0.009), extrapulmonary TB (P<0.001), cavitation (P<0.001), and resistance pattern (P=0.001) were significantly associated with ADRs (Table 5).

Multivariate analysis of ADRs

Multivariate logistic regression confirmed female sex [odds ratio (OR) =1.26, 95% confidence interval (CI): 1.04–1.53], age ≥35 years (OR =1.31, 95% CI: 1.12–1.53), BMI <18.5 kg/m² s(OR =1.22, 95% CI: 1.02–1.45), diabetes (OR =1.28, 95% CI: 1.04–1.58), retreatment history (OR =1.28, 95% CI: 1.08–1.52), extrapulmonary TB (OR =1.62, 95% CI: 1.30–2.02), and cavitation (OR =1.24, 95% CI: 1.04–1.48) as independent ADR risk factors. Using RR/MDR-TB as reference, pre-XDR-TB (OR =1.14, 95% CI: 1.01–1.29) and XDR-TB (OR =1.29, 95% CI: 1.06–1.57) significantly increased ADR risk (Table 6).

Analysis of specific ADRs showed QTc prolongation is associated with age, drug-resistant type, use of bedaquiline, delamanid, clofazimine, moxifloxacin and levofloxacin. Hepatotoxicity is related to extrapulmonary tuberculosis, drug-resistant type, use of linezolid and prothionamide. Adverse reactions in the hematological system reactions are related to gender, BMI, extrapulmonary tuberculosis and the use of linezolid. The related drugs have no significant risk of adverse reactions (Tables S1-S3).

Discussion

A total of 2,151 patients were enrolled, with 1,209 (56.2%) patients experiencing at least one ADR. Of these, 413 cases developed grade 3 or higher ADRs, accounting for an incidence rate of 19.2%. Specifically, the ADR incidence was 62.2% (961/1,543) in the bedaquiline cohort and 45.7% (278/608) in the delamanid cohort. ADRs manifested across multiple organ systems, with the top five most prevalent types being cardiovascular system reactions (625 cases, 29.1%—predominantly QTc interval prolongation, accounting for 589 cases, 27.4%), hepatic system reactions (450 cases, 20.9%), hematological system reactions (272 cases, 12.6%), nervous system reactions (196 cases, 9.1%), and metabolic/endocrine system reactions (138 cases, 6.4%). The total number of ADR episodes reached 3,405, among which 1,827 episodes (53.7%) were classified as grade 3 or higher severe ADRs, and 27 episodes (0.8%) were grade 5 ADRs.

A meta-analysis has reported that the incidence of ADRs during anti-tuberculosis treatment ranges from 8% to 85% (12). The 56.2% ADR incidence observed in our study aligns with the trend of high ADR rates documented in relevant Chinese studies. For instance, one retrospective analysis in China reported an ADR incidence of 41.92% (13), while another retrospective study noted a rate of 62.6% (14). Internationally, high ADR incidences have also been described: a retrospective analysis from South Korea reported an incidence of 67.8% (8), whereas a prospective study from India found that only 20% of tuberculosis patients developed ADRs—a rate lower than that observed in our cohort (15). The incidence of ADRs in this study is based on traditional long-course regimens and may therefore be higher than that in the new BPaL-M regimen (bedaquiline, pretomanid, linezolid, and moxifloxacin regimen) with a treatment duration of 6–9 months.

Notably, QTc interval prolongation (27.4%) was the most common ADR in our study, followed by hepatotoxicity (20.9%) and hematological reactions (12.6%). This contrasts with traditional studies, where hepatotoxicity typically ranks as the most frequent ADR (16). This discrepancy is likely attributable to the rigorous monitoring protocol adopted in our study: both cohorts incorporated ECG monitoring as a routine assessment, with dynamic evaluation of the QTc interval performed every 4 weeks. This approach significantly enhanced the detection of asymptomatic QTc interval prolongation. In contrast, traditional studies often rely on patient-reported symptoms or non-systematic monitoring, which tends to result in underreporting of QTc interval prolongation (17).

QTc interval prolongation—the most prevalent ADR in our study—has been well established in numerous studies to be associated with both bedaquiline and delamanid. A Chinese study investigating delamanid-containing regimens for MDR-TB reported an incidence of QTc interval prolongation of 18.8% (7). Mechanistically, research has demonstrated that QTc interval prolongation is closely linked to the level of DM-6705, the major metabolite of delamanid (6). Bedaquiline-induced QTc interval prolongation has also been widely reported; for example, a study in India found that 37.1% of patients treated with bedaquiline developed this ADR (18). Pharmacologically, bedaquiline undergoes oxidative metabolism via the CYP3A4 enzyme, generating an N-desmethyl metabolite (M2), which has been shown to be significantly associated with the occurrence of QTc interval prolongation (17). Furthermore, the concurrent administration of bedaquiline with other QTc-prolonging agents, such as fluoroquinolones, may potentiate the risk of QTc interval prolongation (19). However, conflicting evidence exists, as some studies have suggested that the co-administration of bedaquiline and delamanid does not increase this risk. Beyond bedaquiline and delamanid, clofazimine and fluoroquinolones, such as levofloxacin and moxifloxacin, have also been identified as QTc-prolonging agents (20).

Among the factors influencing ADR occurrence identified in our study, female gender emerged as a significant independent risk factor. This finding is consistent with a prospective study by Prasad, which also confirmed a strong association between female gender and ADR development (21). While several studies have reported that women face a significantly higher risk of anti-tuberculosis drug-related ADRs, others have failed to detect a significant correlation between gender and such ADRs (22-24). Age-related associations with anti-tuberculosis drug ADRs have been documented in multiple studies; our results confirm that advancing age is a risk factor for ADRs, which is in line with a study by Choi et al. that also indicated increased ADR risk with older age (8). Additionally, one study highlighted that pyrazinamide is among the most common drugs contributing to anti-tuberculosis ADRs in elderly tuberculosis patients (25).

This study has several limitations that should be acknowledged. First, its retrospective design may have led to underreporting of ADRs, as passive data collection can miss mild or asymptomatic events. Second, the absence of drug plasma concentration data precluded the establishment of a dose-toxicity relationship, which limits our ability to quantify the impact of drug exposure on ADR risk. Third, the study period [2017–2022] coincided with multiple updates to the WHO and Chinese guidelines for the treatment of drug-resistant tuberculosis, which gradually expanded the indications for new anti-tuberculosis drugs. This evolution in clinical practice may have introduced variability in patient selection and treatment regimens, potentially affecting the external validity of our results. Fourth, we did not assess the impact of ADRs on key treatment outcomes, such as treatment success rate, recurrence rate, or treatment adherence, which are critical for evaluating the clinical significance of ADRs. Future research should address these limitations by conducting prospective cohort studies that integrate therapeutic drug monitoring (TDM) to explore the dose-effect relationship of ADRs and their impact on treatment outcomes. Additionally, the small sample size of patients aged ≥65 years (n=47) in our study limited our ability to comprehensively evaluate ADR risks in the elderly population; larger subgroup analyses focusing on elderly patients are therefore warranted. It should be emphasized that this study was not designed to compare differences between the bedaquiline and delamanid cohorts but rather to conduct an aggregated analysis of data from both cohorts to characterize ADR profiles in MDR/RR-TB patients treated with new anti-tuberculosis drugs.

Conclusions

In conclusion, while new drug-containing regimens for MDR/RR-TB are associated with a relatively high ADR incidence, most ADRs are mild to moderate and clinically manageable. In the Chinese context, QT interval prolongation and hepatotoxicity are the most common ADRs of these regimens, underscoring the need for individualized management strategies for high-risk populations in clinical practice. During treatment, close monitoring of the QT interval (via regular ECG) and liver function (via biochemical markers) is imperative. When formulating treatment regimens, optimization of drug combinations—specifically, avoiding the concurrent use of multiple QT-prolonging agents—is recommended. Additionally, comprehensive supportive measures, such as nutritional intervention and comorbidity control, should be implemented to mitigate ADR risks. This study provides a critical evidence base for developing precision monitoring protocols, and future multidisciplinary collaboration will be essential to balance treatment efficacy with medication safety in the management of MDR/RR-TB.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2411/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2411/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2411/prf

Funding: This study was supported by the Special Project for the Survey of Basic Scientific and Technological Resources, Investigation on the Impact of Respiratory System Infections on Chronic Respiratory Diseases and Comorbidities (No. 2023FY100601), and the Public Health Talent Development Support Project of the National Administration of Disease Prevention and Control.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2411/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are 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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Both cohorts in this study were approved by the Medical Ethics Committee of Beijing Chest Hospital, Capital Medical University. The bedaquiline cohort [approval No. (2019) Lin Shen No. (2018-05-04)] and the delamanid cohort [approval No. (2020) Lin Shen No. (09-01)] were separately reviewed and authorized and informed consent was taken from all the patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. World Health Organization. Global tuberculosis report 2024 [R]. Geneva: WHO; 2024.
2. Chinese Society of Tuberculosis, Chinese Medical Association. Chinese expert consensus on the treatment of multidrug-resistant and rifampicin-resistant tuberculosis (2019 edition). Zhonghua Jie He He Hu Xi Za Zhi 2019;42:733-49.
3. Gao JT, Du J, Wu GH, et al. Bedaquiline-containing regimens in patients with pulmonary multidrug-resistant tuberculosis in China: focus on the safety. Infect Dis Poverty 2021;10:32. [Crossref] [PubMed]
4. Patel SV, Nimavat KB, Alpesh PB, et al. Treatment outcome among cases of multidrug-resistant tuberculosis (MDR TB) in Western India: A prospective study. J Infect Public Health 2016;9:478-84. [Crossref] [PubMed]
5. Hoa NB, Nhung NV, Khanh PH, et al. Adverse events in the treatment of MDR-TB patients within and outside the NTP in Pham Ngoc Thach hospital, Ho Chi Minh City, Vietnam. BMC Res Notes 2015;8:809. [Crossref] [PubMed]
6. Dooley KE, Rosenkranz SL, Conradie F, et al. QT effects of bedaquiline, delamanid, or both in patients with rifampicin-resistant tuberculosis: a phase 2, open-label, randomised, controlled trial. Lancet Infect Dis 2021;21:975-83. [Crossref] [PubMed]
7. Gao MQ, Gao JT, Ma XG, et al. Clinical analysis of phase adverse drug reactions in patients with multidrug-resistant and rifampicin-resistant pulmonary tuberculosis treated with delamanid-containing regimens in China. Chinese Journal of Tuberculosis and Respiratory Diseases 2024;47:638-46. [Crossref] [PubMed]
8. Choi H, Park HA, Hyun IG, et al. Incidence and outcomes of adverse drug reactions to first-line anti-tuberculosis drugs and their effects on the quality of life: A multicenter prospective cohort study. Pharmacoepidemiol Drug Saf 2022;31:1153-63. [Crossref] [PubMed]
9. World Health Organization-Uppsala Monitoring Centre (WHO-UMC). Causality assessment standards for adverse drug reactions [EB/OL]. Uppsala: WHO-UMC; 2019.
10. U.S. Department of Health and Human Services, National Institutes of Health, National Institute of Allergy and Infectious Diseases, Division of AIDS. Division of AIDS (DAIDS) table for grading the severity of adult and pediatric adverse events, Corrected Version 2.1, 2017 [2021-01-31].
11. World Health Organization. Global tuberculosis report 2018 [R]. Geneva: WHO; 2018.
12. Singh A, Prasad R, Balasubramanian V, Gupta N, Gupta P. Prevalence of adverse drug reaction with first-line drugs among patients treated for pulmonary tuberculosis. Clinical Epidemiology and Global Health 2015;3:S80-90.
13. Shi Y, Zhao JF, Hou BL, et al. Analysis of the incidence and influencing factors of anti-tuberculosis drug-related adverse drug reactions in 365 patients with pulmonary tuberculosis. Evaluation and Analysis of Drug-Use in Hospitals of China 2025;25:614-8.
14. Jiang BF, Ma CC, Chen YG, et al. Incidence of adverse drug reactions induced by anti-tuberculosis drugs and analysis of influencing factors. Chinese Journal of Disease Control and Prevention 2017;21:160-3.
15. Tewari S, Patel R, Shah K. Adverse drug reactions in drug resistant pulmonary tuberculosis patients. Indian J Tuberc 2024;71:S25-8. [Crossref] [PubMed]
16. Anand AC, Seth AK, Paul M, et al. Risk Factors of Hepatotoxicity During Anti-tuberculosis Treatment. Med J Armed Forces India 2006;62:45-9. [Crossref] [PubMed]
17. Rouan MC, Lounis N, Gevers T, et al. Pharmacokinetics and pharmacodynamics of TMC207 and its N-desmethyl metabolite in a murine model of tuberculosis. Antimicrob Agents Chemother 2012;56:1444-51. [Crossref] [PubMed]
18. Darmayani IGAAPS, Ascobat P, Instiaty I, et al. Bedaquiline Effect on QT Interval of Drugs-Resistant Tuberculosis Patients: Real World Data. Acta Med Indones 2022;54:389-96.
19. Chinese Society for Tuberculosis, Chinese Medical Association. Expert consensus on clinical application of new anti-tuberculosis drug bedaquoline (2020 update). Zhonghua Jie He He Hu Xi Za Zhi 2021;44:81-7. [Crossref] [PubMed]
20. Amankwa K, Krishnan SC, Tisdale JE. Torsades de pointes associated with fluoroquinolones: importance of concomitant risk factors. Clin Pharmacol Ther 2004;75:242-7. [Crossref] [PubMed]
21. Prasad R, Singh A, Srivastava R, et al. Frequency of adverse events observed with second-line drugs among patients treated for multidrug-resistant tuberculosis. Indian J Tuberc 2016;63:106-14. [Crossref] [PubMed]
22. Imam F, Sharma M, Khayyam KU, et al. Adverse drug reaction prevalence and mechanisms of action of first-line anti-tubercular drugs. Saudi Pharm J 2020;28:316-24. [Crossref] [PubMed]
23. Fei CM, Zainal H, Ali IAH. Evaluation of Adverse Reactions Induced by Anti-Tuberculosis Drugs in Hospital Pulau Pinang. Malays J Med Sci 2018;25:103-14. [Crossref] [PubMed]
24. Sinha DN, Palipudi KM, Jones CK, et al. Levels and trends of smokeless tobacco use among youth in countries of the World Health Organization South-East Asia Region. Indian J Cancer 2014;51:S50-3. [Crossref] [PubMed]
25. Kwon BS, Kim Y, Lee SH, et al. The high incidence of severe adverse events due to pyrazinamide in elderly patients with tuberculosis. PLoS One 2020;15:e0236109. [Crossref] [PubMed]