Objective: To measure HIV-associated adult mortality in a rural population in Tanzania. To record the signs and symptoms associated with deaths of HIV-positive adults.
Design: Prospective cohort study conducted in the context of a randomized controlled trial to evaluate the impact of a sexually transmitted disease treatment programme.
Methods: A cohort consisting of a random sample of 12 501 adults aged 15-54 years was recruited from 12 rural communities in Mwanza region, Tanzania in 1991/1992. Baseline HIV prevalence was 4.0%. The cohort was followed up after 2 years to record mortality according to baseline HIV status. A verbal autopsy questionnaire was administered for each of the deaths reported.
Results: A total of 196 deaths were recorded, of which 73 (37%) occurred in HIV-positive individuals. Mortality rates per 1000 person-years were 6.0 in HIV-negatives and 93.5 in HIV-positives. The age-adjusted mortality rate ratio was 15.68 (95% confidence interval, 11.18-21.03). The proportion of adult deaths attributed to HIV infection was 35% overall and 53% in those aged 20-29 years. Verbal autopsies showed that HIV-positive deaths were significantly associated with fever, rash, weight loss, anaemia, cough, chest pain, abdominal pain and headache, but the specificity of individual symptoms was low. The World Health Organization clinical case definition of AIDS was satisfied for only 13 deaths, of which seven were HIV-positive at baseline. Only seven respondents reported that the death was associated with HIV or AIDS.
Conclusions: This study confirms the strong association of HIV infection and mortality in rural Africa, with an annual death rate in adult seropositives of over 9%. In this rural population with a relatively low HIV prevalence of 4%, HIV has increased overall adult mortality by more than 50%. Signs and symptoms associated with HIV deaths were non-specific, and the population seemed largely unaware of the contribution of HIV to mortality, an important obstacle to prevention efforts.