Association of linezolid interruption and rechallenge with clinical outcomes in patients with rifampicin-resistant tuberculosis in South Africa, 2018-22: an observational cohort study.

Abstract

Linezolid is a key component of treatment regimens for rifampicin-resistant or multidrug-resistant tuberculosis. Toxicity-related linezolid interruption is not uncommon, but little is known about its effect on treatment-related outcomes. Furthermore, there are no data on the frequency of successful rechallenge and its associated predictors. This study aimed to ascertain the relationship between linezolid discontinuation and treatment outcome, and the likelihood of successful rechallenge in patients who discontinued linezolid. We conducted an observational cohort study to review programmatic data from a South African national database on which all patients with drug-resistant tuberculosis are required to be registered. Data were extracted between Jan 1, 2018, and March 16, 2022. Data capture occurs at primary, secondary, and tertiary centres. We included adults 18 years or older with rifampicin-resistant or multidrug-resistant tuberculosis who had received linezolid and had recorded treatment outcomes. We excluded patients with extrapulmonary tuberculosis, those who were not exposed to a bedaquiline-containing regimen, and patients who died before completing their course of linezolid. The frequency of linezolid interruption and subsequent linezolid rechallenge was evaluated from clinical data for patients in the database. The primary outcome measure was the proportion of patients with an unfavourable outcome (death, treatment failure, or loss to follow-up) in those in whom linezolid was interrupted compared with those who continued treatment. Outcomes were ascertained using logistic regression to estimate adjusted odds ratios (aORs). 10 102 (83·7%) of 12 064 patients registered on the database between Jan 1, 2018, and Dec 31, 2020, met the study inclusion criteria and were included in the analysis. 6142 (60·8%) patients were male and 3960 (39·2%) were female. Treatment with linezolid was interrupted in 1974 (19·5%) patients, and the median time to interruption was 4·6 weeks (IQR 2·0-8·6). After adjusting for potentially confounding variables, linezolid interruption was associated with greater odds of an unfavourable outcome compared with completion of treatment (aOR 4·24 [95% CI 3·79-4·75], p<0·0001). Of the 1974 patients who interrupted their course of linezolid, only 65 (3·3%) were rechallenged and, of these, 47 (72·3%) successfully completed their linezolid course. Those successfully rechallenged were more likely to have a favourable outcome than were those in whom rechallenge failed (OR 4·03 [1·08-15·0], p=0·038). A shorter time to initial linezolid interruption was associated with reduced rechallenge success (OR 0·26 [0·08-0·85], p=0·022) and a higher likelihood of unfavourable outcome (OR 1·59 [1·33-1·93]; p<0·0001). Linezolid interruption was common and associated with unfavourable clinical outcomes. Linezolid rechallenge was frequently successful and associated with better clinical outcomes. These findings inform clinical practice and have implications for the management of patients receiving oral, short-course regimens for multidrug-resistant tuberculosis. The South African Medical Research Council, the European and Developing Countries Clinical Trials Partnership, and the National Research Foundation.