Abstract
BACKGROUND: We describe the outcomes of a program in which antiretroviral therapy (ART) is offered to human immunodeficiency virus (HIV) infected patients in South Africa admitted with tuberculosis (TB) or other opportunistic infection (OI) as part of in-patient care. METHODS: Patients admitted with HIV and concurrent TB or other OI were initiated on early in-patient ART. The primary and secondary endpoints were respectively 24-week mortality and 24-week virologic suppression. Multivariable logistic regression modeling explored the associations between baseline (i.e., pre-hospital discharge) characteristics and mortality at 24 weeks. RESULTS: A total of 382 patients were prospectively enrolled (48% women, median age 37 years, median CD4 count 33 cells/mm(3)). Acute OIs were pulmonary TB, 39%; extra-pulmonary TB, 25%; cryptococcal meningitis (CM), 10%; and chronic diarrhea, 9%. The median time from admission to ART initiation was 14 days (range 4-32, IQR 11-18). At 24 weeks of follow-up, as-treated and intention-to-treat virologic suppression were respectively 57% and 93%. Median change in CD4 cell count was +100 cells/mm(3), overall 24-week mortality was 25% and loss to follow-up, 5%. Excess mortality was not observed among patients with CM who initiated early ART. A longer interval between admission and ART was associated with mortality (>21 days vs. <21 days after admission OR 2.1, 95%CI 1.2-4.0, P = 0.016). CONCLUSIONS: For HIV-infected in-patients with TB or an acquired immune-deficiency syndrome defining OI, we demonstrate the operational feasibility of early ART initiation in in-patients.