Abstract
BACKGROUND: Human immunodeficiency virus (HIV)-associated tuberculosis deaths have decreased worldwide over the past decade. We sought to evaluate the effect of HIV status on tuberculosis mortality among patients undergoing treatment for tuberculosis in Lima, Peru, a low HIV prevalence setting. METHODS: We conducted a prospective cohort study of patients treated for tuberculosis between 2005 and 2008 in two adjacent health regions in Lima, Peru (Lima Ciudad and Lima Este). We constructed a multivariate Cox proportional hazards model to evaluate the effect of HIV status on mortality during tuberculosis treatment. RESULTS: Of 1701 participants treated for tuberculosis, 136 (8.0%) died during tuberculosis treatment. HIV-positive patients constituted 11.0% of the cohort and contributed to 34.6% of all deaths. HIV-positive patients were significantly more likely to die (25.1 vs. 5.9%, P < 0.001) and less likely to be cured (28.3 vs. 39.4%, P = 0.003). On multivariate analysis, positive HIV status (hazard ratio [HR] = 6.06; 95% confidence interval [CI], 3.96-9.27), unemployment (HR = 2.24; 95% CI, 1.55-3.25), and sputum acid-fast bacilli smear positivity (HR = 1.91; 95% CI, 1.10-3.31) were significantly associated with a higher hazard of death. CONCLUSIONS: We demonstrate that positive HIV status was a strong predictor of mortality among patients treated for tuberculosis in the early years after Peru started providing free antiretroviral therapy. As HIV diagnosis and antiretroviral therapy provision are more widely implemented for tuberculosis patients in Peru, future operational research should document the changing profile of HIV-associated tuberculosis mortality.