Use of a WHO-recommended algorithm to reduce mortality in seriously ill patients with HIV infection and smear-negative pulmonary tuberculosis in South Africa: an observational cohort study

Background In 2007, WHO released revised recommendations and an algorithm for the diagnosis and treatment of smear-negative pulmonary tuberculosis in seriously ill people living with HIV/AIDS. We aimed to assess the effect of the recommendations on clinical outcome in patients in South Africa. Methods We enrolled seriously ill patients (aged >= 15 years) with HIV infection and suspected smear-negative pulmonary tuberculosis from three hospitals in KwaZulu.Natal, South Africa. Patients were consecutively enrolled into two cohorts: the first cohort was managed according to standard practice, and the second according to the WHO-recommended algorithm. The primary endpoints were rates of continued stay in hospital at 7 days after admission and survival at 8 weeks after admission. Findings 338 patients were enrolled in the standard practice cohort between August, 2008, and February, 2009, and 187 were enrolled in the algorithm cohort between March, 2009, and December, 2009. 7 days after hospital admission, 27% (n=50) of patients in the algorithm cohort were still in hospital, compared with 38% (n=130) in the standard practice cohort (rate ratio 0.70, 95% CI 0.53-0.91; p=0.009). 8 weeks after admission, 83% (n=156) of patients in the algorithm cohort were alive, compared with 68% (n=230) in the standard practice cohort (1.23, 1.11-1.35; p=0.0001), with effect modified by hospital location. Interpretation In seriously ill patients with HIV infection and suspected smear-negative pulmonary tuberculosis, early antituberculosis treatment according to the WHO algorithm could significantly reduce mortality in South Africa.