Relapse and Acquired Rifampin Resistance in HIV-Infected Patients with Tuberculosis Treated with Rifampin- or Rifabutin-Based Regimens in New York City, 1997–2000

Background. The relationship between rifamycin use and either relapse or treatment failure with acquired rifampin resistance (ARR) among human immunodeficiency virus (HIV)–infected patients with tuberculosis (TB) is not well understood. Methods. We conducted a retrospective cohort study of HIV-infected and HIV-uninfected persons with rifampin-susceptible TB, (1) to compare relapse rates, ARR, and treatment failure, according to HIV serostatus; and (2) to examine whether and how use of rifamycin was associated with clinical outcomes of interest among HIV-infected patients with TB. Results. HIV-infected patients were more likely to have ARR than were HIV-uninfected patients (0.9% vs. 0.1%; P = .007), and the association remained significant in multivariate analysis (adjusted odds ratio [OR], 5.5; 95% confidence interval [CI], 1.4–21.5). Among HIV-infected patients with TB, none of 57 patients treated with rifabutin-based regimens alone had ARR, and only 1 of 395 patients treated with rifabutin given in combination with a rifampin-based regimen had ARR, whereas 6 of 355 patients treated with a rifampin-based regimen alone had relapse and ARR. HIV-infected patients treated with rifampin-based regimens alone had a higher risk for relapse and development of rifampin resistance if intermittent dosing of rifampin was started during the intensive phase of treatment, compared with patients who did not receive intermittent dosing (hazard ratio [HR] for relapse, 6.7 [95% CI, 1.1–40.1]; HR for ARR, 6.4 [95% CI, 1.1–38.4]). This association remained when confined to patients with a CD4+ T lymphocyte count of