Highlights
-
•
Cycloserine caused neuropsychiatric side effects in 5.3 percent of MDR-TB patients studied.
-
•
Regimens containing delamanid showed the highest risk of severe neuropsychiatric events.
-
•
Being female underweight or a retreatment patient raised the risk of these side effects.
-
•
Peripheral neuropathy insomnia and headache were the most common adverse events reported.
-
•
High-risk groups need more monitoring and personalized treatment approaches.
Keywords: Tuberculosis, Neuropsychiatric, Adverse events, Bedaquiline, Delamanid
Abstract
Background
Cycloserine (Cs) is essential in bedaquiline (Bdq)/delamanid (Dlm)-based regimens for drug-resistant tuberculosis (DR-TB), but real-world evidence on neuropsychiatric/central nervous system and peripheral nervous system adverse events (AEs) during Cs-containing treatment remains limited.
Methods
A multicenter retrospective cohort study analyzed 2359 adults with multidrug-resistant or rifampicin-resistant tuberculosis (MDR/RR-TB) patients in China (2018–2023) treated with Cs alongside Bdq, Dlm, or conventional therapy. Neuropsychiatric/CNS and peripheral nervous system AEs were assessed using the Symptom Checklist-90 and graded per Division of Acquired Immunodeficiency Syndrome v2.1. Multivariable logistic regression and sensitivity analyzes were used to examine factors associated with these AEs.
Results
A total of 2359 participants were included in this study. The overall incidence of neuropsychiatric/CNS or peripheral nervous system AEs during Cs-containing treatment was 5.3%, with higher frequency in Dlm-based regimens (8.2%) compared to Bdq (4.1%) and conventional therapy (5.2%). Psychiatric/CNS events occurred in 3.4% of patients, and peripheral neuropathy (PN) occurred in 1.9%. Among the 125 events, the most common adverse events were PN at 36.0%, followed by insomnia at 16.8% and headache at 9.6%. Serious adverse events, defined as grade 3 or higher, accounted for 15.2% of all cases, with the highest proportion observed in the Dlm cohort at 23.4%. Multivariable analysis showed that the Dlm-containing regimen was associated with higher odds of overall AEs (adjusted odds ratio: 1.71, 95% confidence interval: 1.12–2.60), as were female sex (1.56, 1.07–2.27), BMI < 18.5 kg/m² (1.51, 1.03–2.23), and retreatment status (1.75, 1.01–3.05). In sensitivity analyzes, no association was observed between the Dlm-containing regimen and psychiatric/CNS events (0.95, 0.54–1.68), whereas an association was observed for PN (3.43, 1.76–6.69).
Conclusion
In this real-world observational study, neuropsychiatric/CNS and peripheral nervous system AEs were observed during Cs-containing treatment, with differences across treatment cohorts. These findings support careful clinical monitoring but cannot establish the causal effects of Cs, Dlm, or any individual treatment regimen.
Graphical abstract
1. Introduction
Drug-resistant tuberculosis (DR-TB) remains a major global health challenge, particularly in high-burden tuberculosis regions1 WHO estimates that in 2023, there will be 400,000 new cases of multidrug-resistant/rifampicin-resistant tuberculosis (MDR/RR-TB) worldwide, with only about 44% of patients successfully completing their treatment2 MDR-TB accounts for approximately one-quarter of global deaths due to antimicrobial resistance3 In recent years, the introduction of new anti-tuberculosis drugs, such as Bdq and Dlm, has significantly improved the treatment outlook for DR-TB4,5 The WHO guidelines have incorporated these drugs into core treatment regimens to increase treatment success rates and shorten treatment durations6 However, due to the complexity of DR-TB and potential drug interactions, second-line anti-tuberculosis drugs are often used in combination in clinical practice, which is associated with a higher risk of adverse events (AEs)7 These symptoms can significantly impact patients’ quality of life, complicating treatment adherence, and ultimately lowering overall treatment success rates8
Cs is a second-line anti-tuberculosis drug widely used in the treatment of drug-resistant tuberculosis9 It exerts its anti-tuberculosis effect by inhibiting key enzymes in the synthesis of the mycobacterial cell wall. Since the late 1950s, Cs has been used in tuberculosis treatment10 However, the adverse neuropsychiatric effects of Cs on patients are concerning. It can cross the blood-brain barrier and interfere with the metabolism of neurotransmitters in the central nervous system, potentially leading to neuropsychiatric side effects such as depression, anxiety, irritability, confusion, and even suicidal tendencies11 The incidence of neuropsychiatric symptoms related to Cs can range from 0% to 30% around world, and in severe cases, it can lead to treatment interruption, affecting the patient's prognosis12
China is one of the high-burden countries for tuberculosis globally, with a significant number of MDR-TB patients13 Currently, real-world data are limited on neuropsychiatric/CNS and peripheral nervous system AEs during Cs-containing regimens that include Bdq or Dlm, particularly in the Chinese population. This multicenter retrospective cohort study aimed to describe the observed frequency, clinical spectrum, and severity of these AEs and to examine factors associated with their occurrence among patients with MDR/RR-TB receiving Cs-containing regimens with Bdq, Dlm, or conventional therapy.
2. Methods
2.1. Study design
This retrospective cohort study evaluated neuropsychiatric/CNS and peripheral nervous system AEs occurring during Cs-containing treatment in patients with MDR-TB between February 2018 and June 202314, 15, 16 The study subjects were drawn from three independent real-world cohort studies in China: a national cohort of MDR-TB patients treated with Bdq from 16 hospitals across China, a national cohort of MDR-TB patients treated with Dlm from 25 hospitals in China, and a cohort of MDR-TB patients treated with conventional regimens, without the use of new drugs, from 20 hospitals in China (Table 1). These cohorts were not randomized; treatment selection was determined by clinical indications, regimen design, drug availability, and local clinical practice.
Table 1.
Geographic distribution of participating hospitals by treatment regimen in this multicenter retrospective cohort study.
| Regimen | Hosptial location (city) |
|---|---|
| Bedaquiline | Beijing, Chengdu, Changsha, Shenyang, Shenzhen, Zhengzhou, Wuhan, Shanghai, Tianjin, Kunming, Xi’an, Suzhou, Harbin, Fuzhou, Hangzhou, Jinan. |
| Delamanid | Beijing, Shanghai, Wuhan, Changsha, Chengdu, Xi’an, Harbin, Jilin, Fuzhou, Zhengzhou, Taiyuan, Suzhou, Kunming, Hangzhou, Jinan, Guangzhou, Shijiazhuang, Liuzhou, Chongqing, Guiyang, Lanzhou, Nanchang, Jiamusi. |
| Conventional | Beijing, Hefei, Changsha, Chengdu, Shenyang, Guangzhou, Jinan, Shanghai, Xi’an, Shenzhen, Hohhot, Changchun, Tianjin, Harbin, Guiyang, Zaozhuang, Ürümqi, Yinchuan. |
2.2. Patient recruitment
The Bdq and Dlm regimen cohorts include patients who are laboratory-confirmed with MDR-TB or RR-TB through molecular testing or culture-based phenotypic drug susceptibility testing. Patients must be aged 18 years or older and require the inclusion of either Bdq (for the Bdq cohort) or Dlm (for the Dlm cohort) to form an effective treatment regimen, according to the WHO guidelines and Chinese Expert Consensus for the treatment of drug-resistant tuberculosis6,17 These patients must not have pre-existing forms such as pre-extensively drug-resistant tuberculosis or extensively drug-resistant tuberculosis. Both cohorts require the patients to be able to adhere to the treatment protocols, report adverse events (AEs), and sign informed consent.
Exclusion criteria for both cohorts include patients with a history of allergy to either Bdq or Dlm, significant cardiac abnormalities (QTc interval > 450 ms for females or > 430 ms for males), severe comorbidities (e.g., heart failure, respiratory failure, liver or kidney dysfunction), and those who are pregnant or breastfeeding. Additionally, patients with a Karnofsky performance score < 50%, a history of significant cardiovascular events, or those who have participated in clinical trials for other new drugs within the past three months are excluded from both cohorts.
The inclusion criteria for the conventional regimen cohort are similar to those for the Bdq and Dlm regimens cohorts, with the key difference being that this cohort does not include Bdq or Dlm in the treatment regimen. The exclusion criteria for this cohort are largely the same as those for the Bdq and Dlm cohorts. However, the conventional regimen cohort does not specifically exclude patients with a history of QT interval prolongation or related conditions. This study includes patients who are aged 18 years or older, diagnosed with MDR-TB or RR-TB based on laboratory testing (molecular or culture-based). Eligible patients must be starting a treatment regimen that includes Cs as part of the background multidrug regimen at cohort entry.
2.3. Clinical follow-up for AEs monitoring
Patients were assessed at baseline and at weeks 2, 4, 8, 12, 16, 20, and 24, with assessments approximately every 4 weeks thereafter while treatment was ongoing. Each follow-up included microbiological assessments (sputum smear, culture, species identification, and drug susceptibility testing), laboratory tests (routine blood and urine tests, liver and kidney function, electrolytes, thyroid function), imaging (chest X-ray), and electrocardiogram monitoring. Clinical symptoms, weight changes, and auditory-visual function were also recorded. For this analysis, the AE risk window extended from initiation of the Cs-containing regimen to treatment completion or discontinuation, death, or the last available treatment record.
All AEs occurring during Cs-containing treatment were recorded, and a multidisciplinary expert panel reviewed these events to determine whether to continue treatment. Neuropsychiatric/CNS and peripheral nervous system AEs were classified based on symptoms of insomnia, anxiety, depression, mania, somnolence, confusion, tremor, dizziness, headache, PN etc. AEs were evaluated using the Division of Acquired Immunodeficiency Syndrome 2017 V2.1 grading system, with grade 3 or higher considered severe18 Psychological symptoms were assessed using the Symptom Checklist-9019 Whenever an AE occured, clinicians reported the detailed information using a standardized form that included the event grade, the responsible drug (including dosage and details of any concomitant medications), the tests performed, the measures taken, the duration, and the outcome of the event (recovered/resolved, recovered/resolved with sequelae, not recovered/resolved, death, unknown)20 This study collected comprehensive patient data from hospital medical records, including demographic characteristics, treatment history, clinical treatment regimens, diagnostic results, imaging findings, and laboratory test results.
Concomitant linezolid use was summarized for each cohort as the number and percentage of patients receiving linezolid. Among linezolid recipients for whom all recorded treatment episodes had valid start and stop dates, cumulative treatment duration was calculated by merging overlapping or contiguous treatment intervals and summing inclusive calendar days. Missing stop dates were not imputed; therefore, duration estimates are presented as complete-case descriptive summaries.
2.4. Statistical Analysis
Descriptive analysis was performed for the patients evaluated in the cohort. A multivariate logistic regression model was used to analyze the relationship between potential predictors and the occurrence of neuropsychiatric/CNS or peripheral nervous system AEs. Prior to model construction, variance inflation factors (VIFs) were used for multicollinearity diagnostics to ensure compliance with regression assumptions. VIF < 5 was considered acceptable, and all retained variables had VIF values <5 (range 1.04–3.71). Variables with a p < 0.200 in univariate analysis were included in the multivariate analysis to ensure that potentially relevant predictors were considered, whereas variables not meeting this threshold were excluded from the final model. To assess robustness to data-driven covariate selection, we additionally fitted a sensitivity model in which clinically prespecified covariates were entered irrespective of their univariable P values. These covariates included treatment cohort, age group, sex, BMI category, diabetes mellitus, treatment history, extrapulmonary tuberculosis, sputum-smear status, culture result, daily Cs dose, and dosing frequency. In the sensitivity analysis, daily Cs dose was modeled as a categorical variable, with 250 mg/day as the reference category. Psychiatric/CNS events and peripheral neuropathy were also analyzed separately. To assess the potential influence of linezolid use, the multivariable analysis of peripheral neuropathy was repeated among linezolid users only. Given the limited number of peripheral neuropathy events, model stability was assessed using 1,000 bootstrap replications. Statistically significant predictors were identified, and the adjusted odds ratios (aOR) and 95% confidence intervals (95% CI) for each predictor were estimated. A p < 0.050 was considered statistically significant. All statistical analyzes were performed using IBM SPSS Version 20.0 (SPSS, Inc., Chicago, USA).
2.5. Ethical Considerations
The study was approved by the ethics committee of all sites, led by the Institutional Review Board of Beijing Chest Hospital [Approval No. (2018-60); Approval No. (2019-2018-05-04); Approval No. (2020-09-01)]. All participants provided written informed consent before being included in the study. The study was conducted in accordance with the Declaration of Helsinki and adhered to ethical standards regarding patient confidentiality and data protection. The personal information and medical records of all participants were kept confidential, and all data were anonymized prior to analysis.
3. Results
Overall, this study included 2359 MDR-TB patients from three cohorts, among which 1321 (56.0%) received the Bdq regimen, 576 (24.4%) received the Dlm regimen, and 462 (19.6%) received the conventional regimen (Table 2). Baseline characteristics differed across treatment groups, reflecting the non-randomized real-world cohort design. Between-group comparisons showed significant differences in sex, age group, diabetes mellitus, treatment history, extrapulmonary TB, sputum smear status, culture result, presence of cavity, cumulative area of pulmonary lesions, daily Cs dose, and frequency of Cs use, whereas BMI did not differ significantly. In the overall cohort, 1600 (67.8%) patients were male, and 759 (32.2%) patients were female. The age distribution revealed that the highest proportion of patients were in the 40–59 years age group (1493 patients, 63.3%), followed by the 18–39 years group (686 patients, 29.1%) and those aged 60 years or older (180 patients, 7.6%). Among the patients, 27.4% had low body weight (BMI < 18.5), and 111 (4.7%) had diabetes mellitus. A total of 75.8% (1787 patients) were retreatment cases, and 17.9% (423 patients) had extrapulmonary tuberculosis. 37.0% (873 patients) had a positive sputum smear, and 50.9% (1200 patients) had a positive bacterial culture. Imaging examination showed that 61.8% (1459 patients) of patients had cavitary lesions, and 51.3% (1211 patients) had lung involvement less than half of a unilateral lung field.
Table 2.
Baseline characteristics of multidrug-resistant tuberculosis patients by treatment regimen.
| Characteristics | All patients (n = 2359) | Bedaquiline (n = 1321) | Delamanid (n = 576) | Conventional (n = 462) | χ2 | p |
|---|---|---|---|---|---|---|
| Sex | 18.049 | <0.001 | ||||
| Male | 1600 (67.8) | 933 (70.6) | 350 (60.8) | 317 (68.6) | ||
| Female | 759 (32.2) | 388 (29.4) | 226 (39.2) | 145 (31.4) | ||
| Age group | 51.119 | <0.001 | ||||
| 18–39 | 686 (29.1) | 320 (24.2) | 179 (31.1) | 187 (40.5) | ||
| 40–59 | 1493 (63.3) | 881 (66.7) | 367 (63.7) | 245 (53.0) | ||
| ≥60 | 180 (7.6) | 120 (9.1) | 30 (5.2) | 30 (6.5) | ||
| BMI (kg/m2) | 2.955 | 0.228 | ||||
| < 18.5 | 646 (27.4) | 369 (27.9) | 165 (28.6) | 112 (24.2) | ||
| ≥ 18.5 | 1713 (72.6) | 952 (72.1) | 411 (71.4) | 350 (75.8) | ||
| Diabetes mellitus | 118.344 | <0.001 | ||||
| No | 2248 (95.3) | 1286 (97.4) | 566 (98.3) | 396 (85.7) | ||
| Yes | 111 (4.7) | 35 (2.6) | 10 (1.7) | 66 (14.3) | ||
| History of treatment | 350.379 | <0.001 | ||||
| Initial treatment | 572 (24.2) | 330 (25.0) | 6 (1.0) | 236 (51.1) | ||
| Re-treatment | 1787 (75.8) | 991 (75.0) | 570 (99.0) | 226 (48.9) | ||
| Extrapulmonary TB | 374.983 | <0.001 | ||||
| No | 1936 (82.1) | 1181 (89.4) | 519 (90.1) | 236 (51.1) | ||
| Yes | 423 (17.9) | 140 (10.6) | 57 (9.9) | 226 (48.9) | ||
| Sputum smear positive | 7.692 | 0.021 | ||||
| No | 1486 (63.0) | 843 (63.8) | 377 (65.5) | 266 (57.6) | ||
| Yes | 873 (37.0) | 478 (36.2) | 199 (34.5) | 196 (42.4) | ||
| Culture positive | 64.275 | <0.001 | ||||
| No | 848 (35.9) | 502 (38.0) | 231 (40.1) | 115 (24.9) | ||
| Yes | 1200 (50.9) | 612 (46.3) | 280 (48.6) | 308 (66.7) | ||
| Not recorded | 311 (13.2) | 207 (15.7) | 65 (11.3) | 39 (8.4) | ||
| Presence of cavity | 161.019 | <0.001 | ||||
| No | 619 (26.2) | 296 (22.4) | 176 (30.6) | 147 (31.8) | ||
| Yes | 1459 (61.8) | 770 (58.3) | 387 (67.2) | 302 (65.4) | ||
| Not recorded | 281 (11.9) | 255 (19.3) | 13 (2.3) | 13 (2.8) | ||
| Cumulative area of pulmonary lesions | 155.407 | <0.001 | ||||
| < 1/2 | 1211 (51.3) | 606 (45.9) | 331 (57.5) | 274 (59.3) | ||
| ≥ 1/2 | 871 (36.9) | 464 (35.1) | 232 (40.3) | 175 (37.9) | ||
| Not recorded | 277 (11.7) | 251 (19.0) | 13 (2.3) | 13 (2.8) | ||
| Daily CS doses (mg) | 54.959 | <0.001 | ||||
| 250 | 177 (7.5) | 126 (9.5) | 41 (7.1) | 10 (2.2) | ||
| 500 | 1575 (66.8) | 907 (68.7) | 383 (66.5) | 285 (61.7) | ||
| 750 | 607 (25.7) | 288 (21.8) | 152 (26.4) | 167 (36.1) | ||
| Frequency of daily CS used | 57.680 | <0.001 | ||||
| Once | 140 (5.9) | 64 (4.8) | 39 (6.8) | 37 (8.0) | ||
| Twice | 1871 (79.3) | 1094 (82.8) | 467 (81.1) | 310 (67.1) | ||
| Thrice | 348 (14.8) | 163 (12.3) | 70 (12.2) | 115 (24.9) |
Note: Data are presented as n (%).
Abbreviations: BMI: Body Mass Index; Cs: cycloserine; TB: tuberculosis.
Regarding the dosing of Cs, 7.5% (177 patients) received 250 mg/day, 66.8% (1575 patients) of patients received 500 mg/day, and 25.7% (607 patients) received 750 mg/day. As for the dosing frequency, 5.9% (140 patients) received once-daily dosing, 79.3% (1871 patients) received the twice-daily regimen, and 14.8% (348 patients) received three-times-daily dosing. All patients received Cs as part of the background regimen. Linezolid was used by 1192 of 1321 patients (90.2%) in the Bdq cohort, 529 of 576 patients (91.8%) in the Dlm cohort, and 257 of 462 patients (55.6%) in the conventional-regimen cohort; clofazimine and amikacin were used by 55.4% and 48.9% of the overall cohort, respectively. Detailed information on cohort-level linezolid use and documented cumulative treatment duration is presented in Supplementary Table S1. Among participants with valid treatment start and end dates, the median documented observation duration was 505 days in the Bdq cohort, 532 days in the Dlm cohort, and 505 days in the conventional-regimen cohort (Supplementary Table S2).
The study recorded 125 neuropsychiatric/CNS or peripheral nervous system AEs occurring during Cs-containing regimens, accounting for 5.3% of the total patients. Of these, 54 cases (4.1%) occurred in the Bdq regimen group, 47 cases (8.2%) in the Dlm regimen group, and 24 cases (5.2%) in the conventional regimen group. A total of 80/2359 (3.4%) patients experienced psychiatric/CNS symptoms. Among those treated with Bdq, 39/1321 (3.0%) had psychiatric/CNS symptoms; 19/576 (3.3%) in the Dlm regimen group and 22/462 (4.8%) in the conventional regimen group were affected. A total of 45/2359 (1.9%) patients reported PN. Of those on Bdq, 15/1321 (1.1%) had PN, 28/576 (4.9%) in the Dlm regimen group, and 2/462 (0.4%) in the conventional regimen group.
The severity distribution is shown in Table 3. Grade 2 events were the most frequent (56.8%), followed by grade 1 (28.0%), grade 3 (12.8%), and grade 4 (2.4%). Among the documented AEs, the proportion classified as grade 3–4 was 23.4% in the Dlm cohort, compared with 9.3% in the Bdq cohort and 12.5% in the conventional-regimen cohort. Table 4 summarizes the serious AEs (grade 3 and above) potentially associated with Cs-containing treatment. The reported reactions included a range of neuropsychiatric/CNS and peripheral nervous system symptoms, such as emotional loss, self-injurious behavior, seizures, delirium, severe insomnia, depression with anxiety, and severe numbness, particularly in the lower limbs. Other notable symptoms included palpitations, dizziness, abnormal behavior, blank stare, mania, and anxiety, along with respiratory symptoms. Seven individuals experienced severe PN, which is suspected to be associated not only with Cs but also potentially with linezolid. All patients, except for two, reported resolution of the symptoms.
Table 3.
Observed frequency and characteristics of neuropsychiatric/CNS and peripheral nervous system adverse events during cycloserine-containing treatment.
| Variable | All Patients | Bedaquiline | Delamanid | Conventional |
|---|---|---|---|---|
| Total | 5.3% (125/2359) | 4.1% (54/1321) | 8.2% (47/576) | 5.2% (24/462) |
| Psychiatric | 3.4% (80/2359) | 3.0% (39/1321) | 3.3% (19/576) | 4.8% (22/462) |
| Peripheral neuropathy | 1.9% (45/2359) | 1.1% (15/1321) | 4.9% (28/576) | 0.4% (2/462) |
| Grade of adverse events | ||||
| 1 | 28.0% (35/125) | 33.3% (18/54) | 17.0% (8/47) | 37.5% (9/24) |
| 2 | 56.8% (71/125) | 57.4% (31/54) | 59.6% (28/47) | 50.0% (12/24) |
| 3 | 12.8% (16/125) | 7.4% (4/54) | 19.1% (9/47) | 12.5% (3/24) |
| 4 | 2.4% (3/125) | 1.9% (1/54) | 4.3% (2/47) | 0.0% (0/24) |
| Classification of adverse events | ||||
| Insomnia | 16.8% (21/125) | 22.2% (12/54) | 12.8% (6/47) | 12.5% (3/24) |
| Headache | 9.6% (12/125) | 13.0% (7/54) | 4.3% (2/47) | 12.5% (3/24) |
| Dizziness | 7.2% (9/125) | 11.1% (6/54) | 4.3% (2/47) | 4.2% (1/24) |
| Irritability and manic | 6.4% (8/125) | 7.4% (4/54) | 6.4% (3/47) | 4.2% (1/24) |
| Depression | 4.0% (5/125) | 5.6% (3/54) | 2.1% (1/47) | 4.2% (1/24) |
| Anxiety | 2.4% (3/125) | 1.9% (1/54) | 2.1% (1/47) | 4.2% (1/24) |
| Fear | 0.8% (1/125) | 0.0% (0/54) | 2.1% (1/47) | 0.0% (0/24) |
| Epilepsy | 2.4% (3/125) | 3.7% (2/54) | 2.1% (1/47) | 0.0% (0/24) |
| Low mood | 4.8% (6/125) | 1.9% (1/54) | 2.1% (1/47) | 16.7% (4/24) |
| Delirium | 8.0% (10/125) | 7.4% (4/54) | 8.5% (4/47) | 8.3% (2/24) |
| Having thoughts of suicide | 2.4% (3/125) | 1.9% (1/54) | 2.1% (1/47) | 4.2% (1/24) |
| Peripheral neuropathy | 36.0% (45/125) | 27.8% (15/54) | 59.6% (28/47) | 8.3% (2/24) |
| Tremor | 2.4% (3/125) | 1.9% (1/54) | 0.0% (0/47) | 8.3% (2/24) |
| Syncope | 0.8% (1/125) | 1.9% (1/54) | 0.0% (0/47) | 0.0% (0/24) |
| Others | 7.2% (9/125) | 5.6% (3/54) | 4.3% (2/47) | 16.7% (4/24) |
Note: Data are presented as percentages, with the number of events/denominator in parentheses. Percentages for overall, psychiatric/CNS, and peripheral neuropathy events were calculated using the total number of patients in each cohort as the denominator. Percentages for AE grade and clinical manifestations were calculated among patients with adverse events in the corresponding cohort.Patients could experience more than one clinical manifestation; therefore, the percentages for individual clinical manifestations do not necessarily sum to 100%.
Abbreviation: CNS: central nervous system.
Table 4.
Summary of the nineteen serious neuropsychiatric adverse events documented during cycloserine-containing treatment.
| Patient ID | Age (years) | Daily CS doses (mg) | Adverse reactions | Current prescribed regimen | Outcome |
|---|---|---|---|---|---|
| 1 | 35 | 500 | Emotional loss and self-injurious behavior | Bdq, Lzd, Lfx, Cfz, Cs, Amk | Resolved |
| 2 | 52 | 500 | Seizure | Mfx, Pto, Cs, Lzd, Cfz, Bdq, PAS | Resolved |
| 3 | 24 | 500 | The skin of both lower limbs was severely numb | Bdq, Z, Lfx, Lzd, Cs | Resolved |
| 4 | 28 | 750 | Severe insomnia, moderate to severe depression with anxiety | Amk, Lzd, Cfz, Pto, Bdq, PAS, Cs | Resolved |
| 5 | 59 | 500 | Delirium, drowsiness, dizziness | Bdq, Cfz, Pto, Z, Cs, Lzd | Resolved |
| 6 | 29 | 500 | Insomnia, dizziness with vomiting | Lzd, Cfz, Cs, Z, Dlm | Resolved |
| 7 | 46 | 250 | Delirium, abnormal behavior, blank stare | Lzd, Dlm, Cs, Cfz, PAS | Unknown |
| 8 | 57 | 500 | Severe numbness in both lower limbs, limited walking ability | Lfx, Lzd, Cfz, Cs, Dlm, Z | Resolved |
| 9 | 38 | 750 | Palpitations, insomnia, with brief suicidal thoughts when severe | Lfx, Cs, Lzd, Cfz, Inh, Dlm | Resolved |
| 10 | 53 | 750 | Severe numbness in both lower limbs, limited walking ability | Lfx, Lzd, Cs, Cfz, Dlm | Resolved |
| 11 | 31 | 500 | Severe numbness in both lower limbs, limited walking ability | Cs, Dlm, E, Amk, Lfx, Lzd | Unknown |
| 12 | 49 | 750 | Severe numbness in both lower limbs, limited walking ability | Lzd, Dlm, Cfz, Z, Cs | Resolved |
| 13 | 46 | 750 | Numbness in both lower limbs, insomnia, dizziness | Dlm, Lzd, Cs, Cfz, Pto, PAS | Resolved |
| 14 | 20 | 750 | Mania, abnormal behavior, with delusions | Lzd, Cfz, Cs, Z, Dlm | Resolved |
| 15 | 32 | 750 | Delirium, palpitations | Lfx, Lzd, Cfz, Cs, Dlm | Resolved |
| 16 | 28 | 500 | Anxiety, fear, triggering symptoms of dyspnea, respiratory alkalosis | Dlm, Amk, Pto, Cs, Lzd, Lfx | Resolved |
| 17 | 29 | 500 | Delirium, insomnia, anxiety | Mfx, Amk, Z, Pto, E, Cs | Resolved |
| 18 | 22 | 500 | Syncope | Amk, Mfx, Cs, Pto, Z, Lzd | Resolved |
| 19 | 56 | 500 | Severe numbness in both lower limbs, limited walking ability | Mfx, Amk, Cs, Lzd, Z, CLR | Resolved |
Note: Drugs shown in bold are those clinicians suspect may be associated with neuropsychiatric adverse reactions.
Abbreviations: Amk: amikacin; Bdq: bedaquiline; Cfz: clofazimine; CLR: Clarithromycin; Cs: cycloserine; Dlm: delamanid; E: ethambutol; Inh: Isoniazid; Lfx: levofloxacin; Lzd: linezolid; Mfx: moxifloxacin; PAS: para-aminosalicylic acid; Pto: prothionamide; Z: pyrazinamide; .
In terms of specific symptoms, the most common AEs occurring during Cs-containing regimens were PN (36.0%), insomnia (16.8%), and headache (9.6%). In the Bdq regimen group, the most common reactions were PN (27.8%), insomnia (22.2%), and headache (13.0%); in the Dlm regimen group, they were PN (59.6%), insomnia (12.8%), and delirium (8.5%); and in the conventional regimen group, the most common reactions were low mood and other symptoms (each 16.7%), followed by insomnia and headache (each 12.5%).
Candidate variables with p < 0.200 in univariate analysis were treatment regimen, sex, BMI, treatment history, culture result, daily Cs dose, and frequency of daily Cs use; age group, diabetes mellitus, extrapulmonary tuberculosis, sputum smear status, presence of cavity, and cumulative area of pulmonary lesions were excluded from the multivariate model because they did not meet this threshold (Table 5). In the multivariate model, patients treated with Dlm (1.71 [95% CI: 1.12–2.60], p = 0.012) had higher odds of overall AEs than those treated with Bdq, whereas conventional regimens did not differ significantly (1.45 [95% CI: 0.86–2.45], p = 0.160). Female sex (1.56 [95% CI: 1.07–2.27], p = 0.021), BMI < 18.5 (1.51 [95% CI: 1.03–2.23], p = 0.035), and retreatment (1.75 [95% CI: 1.01–3.05], p = 0.047) were associated with higher adjusted odds of overall AEs (Table 6).
Table 5.
Univariable analysis of factors associated with neuropsychiatric/CNS or peripheral nervous system AEs during cycloserine-containing treatment.
| Variables | Total | AE n (%) | Non-AE n (%) | χ² | p |
|---|---|---|---|---|---|
| Drug regimens | 13.27 | 0.001 | |||
| Bedaquiline | 1321 | 54 (4.1) | 1267 (95.9) | ||
| Delamanid | 576 | 47 (8.2) | 529 (91.8) | ||
| Conventional | 462 | 24 (5.2) | 438 (94.8) | ||
| Sex | 7.35 | 0.007 | |||
| Male | 1600 | 71 (4.4) | 1529 (95.6) | ||
| Female | 759 | 54 (7.1) | 705 (92.9) | ||
| Age group (years) | 0.48 | 0.786 | |||
| 18–39 | 686 | 39 (5.7) | 647 (94.3) | ||
| 40–59 | 1493 | 78(5.2) | 1415(94.8) | ||
| ≥ 60 | 180 | 8 (4.4) | 172 (95.6) | ||
| BMI (kg/m2) | 5.88 | 0.015 | |||
| ≥ 18.5 | 1713 | 79 (4.6) | 1634 (95.4) | ||
| < 18.5 | 646 | 46 (7.1) | 600 (92.9) | ||
| Diabetes mellitus | 1.57 | 0.211 | |||
| No | 2248 | 122 (5.4) | 2126 (94.6) | ||
| Yes | 111 | 3 (2.7) | 108 (97.3) | ||
| History of treatment | 6.97 | 0.008 | |||
| Initial treatment | 572 | 18 (3.1) | 554 (96.9) | ||
| Re-treatment | 1787 | 107 (6.0) | 1680 (94.0) | ||
| Extrapulmonary tuberculosis | 1.21 | 0.272 | |||
| No | 1936 | 98 (5.1) | 1838 (94.9) | ||
| Yes | 423 | 27 (6.4) | 396 (93.6) | ||
| Sputum smear positive | 0.51 | 0.476 | |||
| No | 1486 | 75 (5.0) | 1411 (95.0) | ||
| Yes | 873 | 50 (5.7) | 823 (94.3) | ||
| Culture positive | 5.32 | 0.070 | |||
| No | 848 | 48 (5.7) | 800 (94.3) | ||
| Yes | 1200 | 69 (5.8) | 1131 (94.2) | ||
| Not recorded | 311 | 8 (2.6) | 303 (97.4) | ||
| Presence of cavity | 0.31 | 0.858 | |||
| No | 619 | 34 (5.5) | 585 (94.5) | ||
| Yes | 1459 | 78 (5.3) | 1381 (94.7) | ||
| Not recorded | 281 | 13 (4.6) | 268 (95.4) | ||
| Cumulative area of pulmonary lesions | 0.62 | 0.734 | |||
| < 1/2 | 1211 | 62 (5.1) | 1149 (94.9) | ||
| ≥ 1/2 | 871 | 50 (5.7) | 821 (94.3) | ||
| Not recorded | 277 | 13 (4.7) | 264 (95.3) | ||
| Daily Cs dose (mg) | 4.51 | 0.105 | |||
| 250 | 177 | 4 (2.3) | 173 (97.7) | ||
| 500 | 1575 | 92 (5.8) | 1483 (94.2) | ||
| 750 | 607 | 29 (4.8) | 578 (95.2) | ||
| Frequency of daily cycloserine used | 6.16 | 0.046 | |||
| Once | 140 | 7 (5.0) | 133 (95.0) | ||
| Twice | 1871 | 109 (5.8) | 1762 (94.2) | ||
| Thrice | 348 | 9 (2.6) | 339 (97.4) |
Note: Data are presented as n (%).
Abbreviations: aOR: adjusted odds ratio; BMI: body mass index; CI: confidence interval; CNS, central nerve system; Cs: cycloserine.
Table 6.
Multivariate logistic regression analysis of factors associated with neuropsychiatric adverse events during cycloserine-containing treatment.
| Variable | aOR | 95% CI | p |
|---|---|---|---|
| Treatment regimen | |||
| Bedaquiline | Reference | ||
| Delamanid | 1.71 | 1.12–2.60 | 0.012 |
| Conventional regimen | 1.45 | 0.86–2.45 | 0.160 |
| Sex | |||
| Male | Reference | ||
| Female | 1.56 | 1.07–2.27 | 0.021 |
| BMI (kg/m2) | |||
| ≥ 18.5 | Reference | ||
| < 18.5 | 1.51 | 1.03–2.23 | 0.035 |
| Treatment history | |||
| Initial treatment | Reference | ||
| Re-treatment | 1.75 | 1.01–3.05 | 0.047 |
| Culture positive | |||
| No | Reference | ||
| Yes | 1.00 | 0.68–1.48 | 0.989 |
| Not recorded | 0.48 | 0.22–1.04 | 0.062 |
| Daily cycloserine dose (mg) | |||
| 250 | Reference | ||
| 500 | 2.16 | 0.78–6.03 | 0.141 |
| 750 | 2.94 | 0.99–8.75 | 0.052 |
| Frequency of daily cycloserine use | |||
| Once | Reference | ||
| Twice | 1.38 | 0.62–3.07 | 0.426 |
| Thrice | 0.54 | 0.19–1.57 | 0.261 |
Note: Data are presented as adjusted odds ratios (aORs) with 95% confidence intervals (CIs). Variables with p > 0.200 in the univariate analysis were not included in the multivariate logistic regression model.
Abbreviations: aOR: adjusted odds ratio; BMI: body mass index; CI: confidence interval.
In the sensitivity analysis including all clinically prespecified covariates, Dlm treatment (aOR: 1.73; 95% CI: 1.13–2.65) and female sex (aOR: 1.58; 95% CI: 1.08–2.32) remained significantly associated with adverse events. The associations with low BMI and re-treatment were attenuated, whereas a daily Cs dose of 750 mg was significantly associated with adverse events (aOR: 3.02; 95% CI: 1.01–9.01) (Supplementary Table S3).
In the sensitivity analysis treating daily Cs dose as a categorical variable, no clear association with peripheral neuropathy was observed (overall p = 0.821). Compared with 250 mg/d, the adjusted odds ratios were 1.00 (95% CI: 0.29–3.39; p = 0.997) for 500 mg/d and 1.29 (95% CI: 0.33–5.07; p = 0.714) for 750 mg/day (Supplementary Table S4). In sensitivity analyzes separating psychiatric/CNS events from PN, the association between Dlm and psychiatric/CNS events was attenuated (0.95 [95% CI: 0.54–1.68], p = 0.859), whereas Dlm was associated with higher adjusted odds of PN (3.43 [95% CI 1.76–6.69], p < 0.001). Conventional regimens were associated with psychiatric/CNS events (1.84 [95% CI: 1.04–3.25], p = 0.036) but not PN (0.44 [95% CI 0.10–1.98], p = 0.285) (Supplementary Table S5). In the analysis restricted to 1,978 linezolid users, the adjusted odds ratio for peripheral neuropathy was 3.37 (95% CI: 1.73–6.58; p < 0.001) for the Dlm cohort and 0.73 (95% CI: 0.16–3.29; p = 0.678) for the conventional-regimen cohort, compared with the Bdq cohort (Supplementary Table S6). In 1000 bootstrap replications of the peripheral neuropathy model, the Dlm estimate remained relatively consistent (aOR: 3.43, bootstrap 95% CI: 1.74–7.06; p = 0.001), whereas the conventional-regimen estimate was unstable (aOR: 0.44, bootstrap 95% CI: < 0.001–1.48; p = 0.166).
4. Discussion
This large-scale, multi-cohort study describes the frequency, severity, and factors associated with neuropsychiatric/CNS and peripheral nervous system AEs recorded during Cs-containing treatment for MDR/RR-TB. Overall, neuropsychiatric/CNS or peripheral nervous system AEs were documented in 5.3% of patients, including psychiatric/CNS events in 3.4% and PN in 1.9%. Female sex, low BMI, and retreatment status were associated with higher adjusted odds of the composite outcome. These associations should not be interpreted as causal risk factors.
In this study, neuropsychiatric/CNS or peripheral nervous system AEs were documented in 5.3% of patients during Cs-containing treatment, broadly consistent with previous reports. A meta-analysis estimated that the overall incidence of Cs-related psychiatric AEs was 5.7%, with central nervous system side effects occurring at a rate of 1.1%21 A recent cohort study conducted in China reported that 4.3% of MDR-TB patients receiving chemotherapy regimens developed psychiatric symptoms22 However, there is heterogeneity in the definitions of neuropsychiatric reactions related to Cs across different studies. Some studies focus exclusively on psychiatric and central nervous system symptoms, whereas our main outcome also included PN. Previous studies indicate that Cs can cause demyelination of peripheral nerves, leading to sensory or motor nerve damage, manifesting as symptoms such as numbness, tingling, and weakness23,24 It was found that Cs exposure was associated with an 8% increase in the risk of developing PN for each 100 microgram/L increase in concentration11 In the sensitivity analyzes, no association was observed between the Dlm-containing regimen and psychiatric/CNS events, whereas higher odds were observed for PN. Thus, the between-cohort difference in the composite outcome was primarily attributable, in statistical terms, to the PN component. When daily Cs dose was modeled as a categorical variable, no clear association with peripheral neuropathy was observed. However, given the limited number of peripheral neuropathy events and the relatively wide confidence intervals, these findings should be interpreted cautiously.
The observed between-cohort differences may reflect treatment selection and regimen composition rather than the effect of a single drug. Treatment allocation was not randomized and may have been influenced by clinical indications, previous treatment history, regimen design, drug availability, treatment period, and local clinical practice. Differences in concomitant medications, monitoring intensity, follow-up, and AE reporting or attribution practices may also have contributed; similar attribution challenges have been reported in observational studies of complex MDR/RR-TB multidrug regimens containing Bdq, Dlm, linezolid, and other potentially toxic agents25,26 Linezolid use was similarly common in the Bdq and Dlm cohorts (90.2% and 91.8%, respectively) but was less common in the conventional-regimen cohort (55.6%). Among recipients with complete start and stop dates, the median documented cumulative duration was 496, 408, and 500 days, respectively. At the cohort level, these data do not support a simple explanation based solely on more frequent or longer linezolid exposure in the Dlm cohort. However, linezolid-related neurotoxicity may vary according to dose and treatment duration27 These cohort-level summaries cannot account for individual dose intensity, dose reductions, interruptions, adherence, or the temporal relationship between linezolid exposure and AE onset. Residual confounding by linezolid and other concomitant neurotoxic drugs therefore cannot be excluded26 In the analysis restricted to linezolid users, the association between the Dlm cohort and peripheral neuropathy remained, suggesting that the between-cohort difference could not be fully explained by linezolid use. Nevertheless, residual confounding related to linezolid dose, treatment duration, dose modification, treatment interruption, and adherence cannot be excluded. The findings should consequently be interpreted as associations across complex real-world multidrug regimens rather than as cohort of a causal effect of Dlm, Cs, linezolid, or any individual regimen.
This study identified that female, BMI < 18.5 kg/m2, and retreatment MDR-TB patients were associated with overall AEs occurring during Cs-containing treatment. Previous studies have also reported a higher frequency of nervous system AEs among female patients during tuberculosis treatment28,29 The increased vulnerability among females may be attributed to physiological and hormonal differences. Sex-specific variations in body composition and metabolic pathways can influence drug pharmacokinetics, potentially leading to higher plasma concentrations of Cs in women due to slower hepatic metabolism, particularly via cytochrome P450 enzymes30,31 Additionally, fluctuations in oestrogen levels may modulate neurotransmitter systems, enhancing sensitivity to drug-induced neuropsychiatric disturbances.32,33
In our study, underweight status was associated with higher odds of overall AEs. This association may reflect differences in nutritional status, body composition, disease severity, or drug exposure, although these potential explanations were not directly evaluated. A study from China also found that patients with low BMI are more likely to experience neuropsychiatric AEs associated with Cs. Reduced tissue volume in underweight individuals could lead to elevated drug concentrations per unit mass34 Additionally, malnutrition, frequently associated with low BMI, may exacerbate neurological susceptibility by impairing neuronal repair mechanisms and compromising blood-brain barrier integrity, thereby enhancing drug penetration into the central nervous system35 Moreover, malnutrition-induced deficiencies in neuroprotective nutrients may further predispose underweight individuals to drug-related neurotoxicity.36,37
38 Chronic disease burden, coupled with treatment adherence challenges and psychosocial stressors, may further amplify neurological vulnerability39 These findings provide important clinical insights for identifying high-risk patients and applying individualized monitoring and intervention strategies in clinical practice.
This study has several limitations that should be considered when interpreting the results. Firstly, as a retrospective cohort study, all data were derived from three pre-existing independent real-world cohorts, which introduces the potential for selection bias. Patients were not randomly assigned to treatment regimens, and baseline characteristics differed among the Bdq, Dlm, and conventional regimen groups. Therefore, the comparative findings should be interpreted as observational evidence rather than evidence from a randomized controlled trial. Secondly, the inclusion of three different treatment cohorts (Bdq, Dlm, and standard regimens), each with varying drug combinations and protocols, could lead to treatment heterogeneity. Although this was a multicenter study, the analysis did not account for potential clustering by study center. Differences in clinical practice, monitoring intensity, and adverse-event reporting across centers may have affected the results. Thirdly, concomitant neurotoxic drugs, particularly linezolid, may have contributed to the observed AEs, especially PN. Although a sensitivity analysis was performed among linezolid users, incomplete information on linezolid dose intensity, dose modification, treatment interruption, adherence, and exposure before event onset prevented full adjustment for linezolid-related confounding. Duration estimates were restricted to patients with complete start and stop dates and may therefore have been affected by informative missingness and administrative recording practices. In addition, the peripheral neuropathy model was based on only 45 events and included 11 non-intercept parameters, with only two events occurring in the conventional-regimen cohort. Although the model converged, bootstrap analysis indicated that the conventional-regimen estimate was unstable. Therefore, the peripheral neuropathy findings should be regarded as exploratory and interpreted cautiously. Additionally, the study could not address certain important issues, such as the pharmacokinetics and pharmacodynamics of Cs. Although neuropsychiatric AEs were recorded through self-reports from physicians and patients during clinical follow-ups, the subjective nature of psychological symptoms could lead to incomplete reporting or misclassification, resulting in an underestimation of the actual incidence of these events. The exclusion criteria of the study also limit its external validity. Finally, the definition of neuropsychiatric/CNS and peripheral nervous system AEs may vary across studies and clinical settings. Differences in the classification of symptoms, such as whether peripheral nervous system symptoms are included, may complicate comparisons with other studies and affect the external validity and reproducibility of the findings.
Taken together, nonrandom treatment selection, differences in regimen composition and monitoring practices, incomplete information on concomitant-drug exposure, and other unmeasured factors may have produced residual confounding. Therefore, the observed associations cannot establish the causal effects of Cs, Dlm, linezolid, or any individual treatment regimen. Prospective studies should collect time-varying exposure information for all neurotoxic drugs and use prespecified methods for AE attribution to clarify their independent and combined associations.
5. Conclusions
Neuropsychiatric/CNS and peripheral nervous system AEs were observed during Cs-containing treatment for MDR/RR-TB, with differences across the three treatment cohorts, particularly for PN. These real-world observational findings support careful clinical monitoring, including closer observation of women, underweight patients, and patients undergoing retreatment. However, nonrandom treatment selection, differences in multidrug regimen composition, and residual confounding preclude causal attribution to Cs, Dlm, linezolid, or any individual regimen. Prospective studies with detailed time-varying drug-exposure data are needed to clarify these associations.
CRediT authorship contribution statement
Ziwen Li: Writing – review & editing, Formal analysis. Yang Liu: Writing – review & editing, Methodology. Zhili Li: Writing – review & editing, Formal analysis. Yuxian Sun: Writing – review & editing. Wei Shu: Writing – review & editing. Lijie Zhang: Writing – review & editing. Jian Du: Writing – review & editing, Funding acquisition. Yuhong Liu: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. Liang Li: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Leiwen Fu: Writing – review & editing, Writing – original draft, Funding acquisition, Formal analysis, Data curation, Conceptualization.
Informed consent
Informed consent was obtained from all participants.
Organ donation
Not applicable.
Ethical Statement
This study was approved by the Institutional Review Board of the Beijing Chest Hospital. [Approval No. (2018-60); Approval No. (2019-2018-05-04); Approval No. (2020-09-01)].
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
Animal treatment
Not applicable.
Generative AI
Generative AI was not used in the study design, data collection, data analysis, interpretation of results, or manuscript writing. Doubao (Seedream 5.0 Flash, ByteDance, Beijing, China) was used solely to assist in the creation of the graphical abstract. The graphical abstract was subsequently reviewed and edited by the authors, who take full responsibility for its content.
Funding
This study was supported by the National Natural Science Foundation of China (NO. 82504532), Young and Middle-aged Talent Development Program of Beijing Chest Hospital (ZQNGG202604), Ministry of Science and Technology (2023FY100601), and Beijing Municipal Health Commission's High Level Public Health Technical Talent Training Plan (2022-2-010 & 2024-03-11).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgment
The author would like to express gratitude to all patients for affording their treatment profiles and all health care workers who participated in this effort.
Footnotes
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.imj.2026.100292.
Contributor Information
Yuhong Liu, Email: liuyuhong0516@126.com.
Liang Li, Email: liliang69@vip.sina.com.
Leiwen Fu, Email: fuleiwen777@qq.com.
Appendix. Supplementary materials
References
- 1.World Health Organization. Tackling the drug-resistant TB crisis. Accessed March 23, 2026. https://www.who.int/activities/tackling-the-drug-resistant-tb-crisis.
- 2.Goletti D., Meintjes G., Andrade B.B., et al. Insights from the 2024 WHO Global Tuberculosis Report - more comprehensive action, innovation, and investments required for achieving WHO End TB goals. Int J Infect Dis. 2025;150 doi: 10.1016/j.ijid.2024.107325. [DOI] [PubMed] [Google Scholar]
- 3.Knight G.M., McQuaid C.F., Dodd P.J., et al. Global burden of latent multidrug-resistant tuberculosis: trends and estimates based on mathematical modelling. Lancet Infect Dis. 2019;19(8):903–912. doi: 10.1016/S1473-3099(19)30307-X. [PubMed] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fekadu G., Tolossa T., Bekele F., et al. Impact of all-oral bedaquiline-based shorter regimens in the treatment of drug-resistant tuberculosis: a systematic review and meta-analysis. BMJ Glob Health. 2025;10(4) doi: 10.1136/bmjgh-2024-018220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lopes S.R., Marçal M., Fernandes N., et al. Update in tuberculosis treatment: a scoping review of current practices. Breathe (Sheff) 2025;21(1) doi: 10.1183/20734735.0232-2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.WHO consolidated guidelines on tuberculosis: module 4: treatment and care 2025. https://www.ncbi.nlm.nih.gov/books/NBK613101/ Geneva: World Health Organization. [PubMed]
- 7.Mereškevičienė R., Danila E. The adverse effects of tuberculosis treatment: a comprehensive literature review. Medicina (Kaunas) 2025;61(5):911. doi: 10.3390/medicina61050911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dela A.I., Tank N.K.D., Singh A.P., et al. Adverse drug reactions and treatment outcome analysis of DOTS-plus therapy of MDR-TB patients at district tuberculosis centre: a four year retrospective study. Lung India. 2017;34(6):522–526. doi: 10.4103/0970-2113.217569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bruning J.B., Murillo A.C., Chacon O., et al. Structure of the mycobacterium tuberculosisd-alanine: d-alanine ligase, a target of the antituberculosis drug d-cycloserine. Antimicrob Agents Chemother. 2011;55(1):291–301. doi: 10.1128/AAC.00558-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Harris D.A., Ruger M., Reagan M.A., et al. Discovery, development, and antimicrobial properties of d-4-amino-3-isoxazolidone (oxamycin), a new antibiotic produced by Streptomyces garyphalus n. sp. Antibiot Chemother (Northfield) 1955;5(4):183–190. [PubMed] [Google Scholar]
- 11.Court R., Centner C.M., Chirehwa M., et al. Neuropsychiatric toxicity and cycloserine concentrations during treatment for multidrug-resistant tuberculosis. Int J Infect Dis. 2021;105:688–694. doi: 10.1016/j.ijid.2021.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pang Y., Liu R., Song Y., et al. High incidence of psychiatric disorders associated with cycloserine treatment of multidrug-resistant tuberculosis patients: a cohort study in Beijing, China. IDR. 2022;Volume 15:3725–3732. doi: 10.2147/IDR.S369715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zhou Y., Anthony R., Wang S., et al. The epidemic of multidrug resistant tuberculosis in China in historical and phylogenetic perspectives. J Infect. 2020;80(4):444–453. doi: 10.1016/j.jinf.2019.11.022. [DOI] [PubMed] [Google Scholar]
- 14.Sun D., Li Z.L., Qiu Y.P., et al. Cost-effectiveness evaluation of bedaquiline- and delamanid-containing regimens for multidrug-/rifampicin-resistant tuberculosis in China: a retrospective study based on three nationwide, real-world multicenter cohorts. J Thorac Dis. 2026;18(4):273. doi: 10.21037/jtd-2026-1-0163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wang Y.T., Fu L.W., Li Z.L., et al. 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. J Thorac Dis. 2026;18(2):129. doi: 10.21037/jtd-2025-aw-2411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hu X.Y., Gao M.Q., Liu Y.H., et al. Interim effectiveness and safety of prolonged bedaquiline use in comparison to standard 24-week treatment for MDR-TB: a multicenter cohort study in China. IDR. 2026;Volume 19:1–11. doi: 10.2147/idr.s572423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chinese Society for Tuberculosis Chinese Medical Association. Expert consensus on surgical treatment of multidrug-resistant and rifampicin-resistant pulmonary tuberculosis in China (2022 edition) Chin J Tuberc Respir Dis. 2023;46(2):111–120. doi: 10.3760/cma.j.cn112147-20221222-00986. [DOI] [PubMed] [Google Scholar]
- 18.Division of AIDS National Institute of allergy and infectious diseases National Institutes of health, US department of health and Human services. 2017. https://rsc.niaid.nih.gov/sites/default/files/daidsgradingcorrectedv21.pdf Division of AIDS (DAIDS) table for grading the severity of adult and pediatric adverse events: corrected version 2.1.
- 19.Tian F., Li H., Tian S., et al. Psychological symptoms of ordinary Chinese citizens based on SCL-90 during the level I emergency response to COVID-19. Psychiatry Res. 2020;288 doi: 10.1016/j.psychres.2020.112992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Borisov S., Danila E., Maryandyshev A., et al. Surveillance of adverse events in the treatment of drug-resistant tuberculosis: first global report. Eur Respir J. 2019;54(6) doi: 10.1183/13993003.01522-2019. [DOI] [PubMed] [Google Scholar]
- 21.Hwang T.J., Wares D.F., Jafarov A., et al. Safety of cycloserine and terizidone for the treatment of drug-resistant tuberculosis: a meta-analysis. Int J Tuberc Lung Dis. 2013;17(10):1257–1266. doi: 10.5588/ijtld.12.0863. [DOI] [PubMed] [Google Scholar]
- 22.Wang J., Pang Y., Jing W., et al. Efficacy and safety of cycloserine-containing regimens in the treatment of multidrug-resistant tuberculosis: a nationwide retrospective cohort study in China. IDR. 2019;Volume 12:763–770. doi: 10.2147/idr.s194484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Aronson J.K. Meyler’s Side Effects of Drugs. Elsevier; Amsterdam: 2016. Cycloserine; p. 796. [DOI] [Google Scholar]
- 24.Service USPH A pilot study of cycloserine toxicity. Am Rev Tuberc. 1956;74:196–209. [PubMed] [Google Scholar]
- 25.Hewison C., Khan U., Bastard M., et al. Safety of treatment regimens containing bedaquiline and delamanid in the endTB cohort. Clin Infect Dis. 2022;75(6):1006–1013. doi: 10.1093/cid/ciac019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lan Z., Ahmad N., Baghaei P., et al. Drug-associated adverse events in the treatment of multidrug-resistant tuberculosis: an individual patient data meta-analysis. Lancet Respir Med. 2020;8(4):383–394. doi: 10.1016/S2213-2600(20)30047-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Conradie F., Bagdasaryan T.R., Borisov S., et al. Bedaquiline-pretomanid-linezolid regimens for drug-resistant tuberculosis. N Engl J Med. 2022;387(9):810–823. doi: 10.1056/NEJMoa2119430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Djochie R.D.A., Anto B.P., Opare-Addo M.N.A. Determinants of adverse reactions to first-line antitubercular medicines: a prospective cohort study. J Pharm Policy Pract. 2023;16(1):70. doi: 10.1186/s40545-023-00577-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Chen X., Wu R., Xu J., et al. Prevalence and associated factors of psychological distress in tuberculosis patients in Northeast China: a cross-sectional study. BMC Infect Dis. 2021;21(1):563. doi: 10.1186/s12879-021-06284-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Waxman D.J., Holloway M.G. Sex differences in the expression of hepatic drug metabolizing enzymes. Mol Pharmacol. 2009;76(2):215–228. doi: 10.1124/mol.109.056705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Scandlyn M.J., Stuart E.C., Rosengren R.J. Sex-specific differences in CYP450 isoforms in humans. Expert Opin Drug Metab Toxicol. 2008;4(4):413–424. doi: 10.1517/17425255.4.4.413. [DOI] [PubMed] [Google Scholar]
- 32.Albert K.M., Newhouse P.A. Estrogen, stress, and depression: cognitive and biological interactions. Annu Rev Clin Psychol. 2019;15:399–423. doi: 10.1146/annurev-clinpsy-050718-095557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Albert K.M., Boyd B.D., Taylor W.D., et al. Differential effects of estradiol on neural and emotional stress response in postmenopausal women with remitted Major Depressive Disorder. J Affect Disord. 2021;293:355–362. doi: 10.1016/j.jad.2021.06.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Oehadian A., Bastos M.L., Centis R., et al. Occurrence and predictors of adverse events associated with Linezolid in the treatment of patients with MDR-TB. Pulmonology. 2024;30(2):184–187. doi: 10.1016/j.pulmoe.2023.09.003. [DOI] [PubMed] [Google Scholar]
- 35.Gómez-Gómez M.E., Zapico S.C. Frailty, cognitive decline, neurodegenerative diseases and nutrition interventions. Int J Mol Sci. 2019;20(11):2842. doi: 10.3390/ijms20112842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wen B., Han X., Gong J., et al. Nutrition: a non-negligible factor in the pathogenesis and treatment of Alzheimer’s disease. Alzheimers Dement. 2025;21(2) doi: 10.1002/alz.14547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Crepin S., Godet B., Chassain B., et al. Malnutrition and epilepsy: a two-way relationship. Clin Nutr. 2009;28(3):219–225. doi: 10.1016/j.clnu.2009.03.010. [DOI] [PubMed] [Google Scholar]
- 38.Vadakunnel M.J., Nehru V.J., Brammacharry U., et al. Factors associated with unfavourable treatment outcomes among patients with multidrug-resistant Tuberculosis receiving outpatients care. Sci Rep. 2025;15 doi: 10.1038/s41598-025-13227-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Alene K.A., Clements A.C.A., McBryde E.S., et al. Mental health disorders, social stressors, and health-related quality of life in patients with multidrug-resistant tuberculosis: a systematic review and meta-analysis. J Infect. 2018;77(5):357–367. doi: 10.1016/j.jinf.2018.07.007. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, upon reasonable request.