Linezolid interruption in patients with fluoroquinolone-resistant tuberculosis receiving a bedaquiline-based treatment regimen.

Background: Treatment outcomes of patients with extensively drug-resistant tuberculosis (XDR-TB) are suboptimal and treatment options remain limited. Linezolid is associated with improved outcomes but also substantial toxicity, and details about the relationship between these are lacking from resource-poor HIV-endemic settings.

Methods: This was a prospective follow-up study of 63 South African XDR-TB patients (58.7% HIV-infected; median CD4 131 cells/μl) between 2014 and 2018. The frequency and severity of linezolid-associated adverse events and the impact on treatment outcomes were compared between linezolid interrupters and non-interrupters.

Results: Twenty-two patients (34.9%) discontinued or underwent dose reduction due to presumed linezolid-associated toxicity. Anaemia (77.3% vs. 7.3%; p< 0.001), peripheral neuropathy (63.6% vs. 14.6%; p= 0.003), and optic neuritis (18.2% vs. 9.8%; p= 0.34) occurred more frequently in linezolid interrupters than in non-interrupters. Anaemia, peripheral neuropathy, and optic neuritis occurred at a median of 5, 18, and 23 weeks, respectively, after treatment initiation. Linezolid interruption was not associated with unfavourable outcomes but was strongly associated with HIV co-infection (adjusted hazard ratio 4.831, 95% confidence interval 1.526-15.297; p= 0.007) and bacterial load (culture days to positivity; adjusted hazard ratio 0.824, 95% confidence interval 0.732- 0.927; p= 0.001).

Conclusions: Linezolid-related treatment interruption is common, is strongly associated with HIV co-infection, and system-specific toxicity occurs within predictable time frames. These data inform the clinical management of patients with drug-resistant TB.