Neuropsychiatric effects of cycloserine with bedaquiline, delamanid, and conventional therapy in drug-resistant tuberculosis: a multicenter retrospective cohort study from China

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 analyses 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 analyses, 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.