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

Abstract

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.

PIP: The impact of HIV on mortality is described in a prospective study of tuberculosis patients in Lusaka, Zambia, where more than 70% of newly diagnosed tuberculosis patients have concurrent HIV infection. Patients attending the University Teaching Hospital in Lusaka, Zambia, were recruited to a prospective cohort study from April to December 1989. Of the 239 patients included in the follow-up study, 174 (73%) were HIV-1 positive by ELISA. A higher proportion of HIV-positive patients were 25-34 years old, and they more often had a negative tuberculin response, anemia, or lymphopenia at recruitment. The probability of survival for HIV-negative and HIV-positive patients was, respectively: at 2 months 95% and 89%; at 6 months 95% and 76%; at 12 months 91% and 66%; at 18 months 87% and 55%; and at 24 months 87% and 48%. The median survival of HIV-positive patients was 22 months. The crude, 2-year mortality rate ratio for HIV-positive compared with HIV-negative patients was 5 (p 0.001). Mortality was higher for patients with both pulmonary and extrapulmonary disease than for those with either pulmonary or extrapulmonary disease alone; for individual sites, only lymph node disease was associated with a significantly higher mortality than other sites (p = 0.01). At presentation prolonged fever, prolonged diarrhea, oral Candida or splenomegaly, negative tuberculin response, anemia or lymphopenia and low weight were associated with higher mortality. Among the 39 patients seen at 2 months who had been prescribed short-course chemotherapy, subsequent mortality was lower in the group who reported receiving all 60 doses of either rifampicin or pyrazinamide or both (20 patients) than among those who had not (19 patients¿ (rate ratio 0.24, p = 0.02). 47 of the 81 patients died within 24 months of the start of treatment, 5 HIV-negative and 42 HIV-positive. 3 of 5 HIV-negative patients for whom information was available died of active tuberculosis. Among HIV-positive patients, 14 of 42 died of active tuberculosis and 2 more of complications of tuberculosis.