The impact of human immunodeficiency virus on mortality of patients treated for tuberculosis in a cohort study in Zambia.

We have examined the impact of human immunodeficiency virus (HIV) on mortality of patients treated for tuberculosis in a prospective study in Lusaka, Zambia. Patients with sputum smear-positive, miliary, or meningeal tuberculosis were prescribed 2 months' daily streptomycin, thiacetazone, isoniazid, rifampicin, and pyrazinamide followed by 6 months thiacetazone and isoniazid; others, 2 months streptomycin, thiacetazone and isoniazid followed by 10 months thiacetazone and isoniazid. 239 patients (65 HIV-negative and 174 HIV-positive) were followed to 2 years from start of treatment. The crude mortality rate ratio for HIV-positive compared with HIV-negative patients over 2 years was 5.00 (95% confidence interval 2.30-10.86). Median survival for HIV-positive patients from the start of treatment was 22 months. At least 34% of HIV-positive patients for whom cause of death was known died from tuberculosis, three-quarters of these during the first month of treatment. Risk factors for death in HIV-positive patients included multi-site tuberculosis, history of prolonged diarrhoea or fever, oral thrush, splenomegaly, anergy to tuberculin, low weight, anaemia or lymphopenia, and poor compliance with regimens containing rifampicin and pyrazinamide. Tuberculosis, even treated, was a major cause of death in patients with HIV infection.