Tuberculosis treatment intermittency in the continuation phase and mortality in HIV-positive persons receiving antiretroviral therapy

Abstract

Background Some tuberculosis (TB) treatment guidelines recommend daily TB treatment in both the intensive and continuation phases of treatment in HIV-positive persons to decrease the risk of relapse and acquired drug resistance. However, guidelines vary across countries, and treatment is given 7, 5, 3, or 2 days/week. The effect of TB treatment intermittency in the continuation phase on mortality in HIV-positive persons on antiretroviral therapy (ART), is not well-described. Methods We conducted an observational cohort study among HIV-positive adults treated for TB between 2000 and 2018 and after enrollment into the Caribbean, Central, and South America network for HIV epidemiology (CCASAnet; Brazil, Chile, Haiti, Honduras, Mexico and Peru). All received standard TB therapy (2-month initiation phase of daily isoniazid, rifampin or rifabutin, pyrazinamide ± ethambutol) and continuation phase of isoniazid and rifampin or rifabutin, administered concomitantly with ART. Known timing of ART and TB treatment were also inclusion criteria. Kaplan–Meier and Cox proportional hazards methods compared time to death between groups. Missing model covariates were imputed via multiple imputation. Results 2303 patients met inclusion criteria: 2003(87%) received TB treatment 5–7 days/week and 300(13%) 2–3 days/week in the continuation phase. Intermittency varied by site: 100% of patients from Brazil and Haiti received continuation phase treatment 5–7 days/week, followed by Honduras (91%), Peru (42%), Mexico (7%), and Chile (0%). The crude risk of death was lower among those receiving treatment 5–7 vs. 2–3 days/week (HR = 0.68; 95% CI = 0.51—0.91; P = 0.008). After adjusting for age, sex, CD4, ART use at TB diagnosis, site of TB disease (pulmonary vs. extrapulmonary), and year of TB diagnosis, mortality risk was lower, but not significantly, among those treated 5–7 days/week vs. 2–3 days/week (HR 0.75, 95%CI 0.55–1.01; P = 0.06). After also stratifying by study site, there was no longer a protective effect (HR 1.42, 95%CI 0.83–2.45; P = 0.20). Conclusions TB treatment 5–7 days/week was associated with a marginally decreased risk of death compared to TB treatment 2–3 days/week in the continuation phase in multivariable, unstratified analyses. However, little variation in TB treatment intermittency within country meant the results could have been driven by other differences between study sites. Therefore, randomized trials are needed, especially in heterogenous regions such as Latin America.