Clear understanding of the natural history of HIV-1 disease is critical for planning and developing appropriate therapeutic strategies for HIV-1-infected populations in the developing world. Present knowledge about Africa is based on very limited data that largely use clinical staging as the prognostic marker; this approach has not been prospectively validated. Our objectives were to compare clinical staging and CD4+ T-cell counts as prognostic tools and to describe survival and cause of death in seroprevalent HIV-1-infected Ugandan adults by means of a prospective cohort study. Consecutive HIV-1-infected adults registering with a community HIV/AIDS clinic in Entebbe, Uganda were enrolled between October 1994 to January 1995 and observed during follow-up until the end of December 1997. Baseline CD4+ T-cell count distribution showed clear and appropriate associations with clinical stage in the 201 participants. Both provided equivalent prognostic information: median survival with CD4+ T-cell count <200 cells/microl was 9 months (95% confidence interval [CI], 7-15 months) compared with 19 and 7 months (95% CI, 10-28 and 0-8 months, respectively) in clinical stages 3 and 4, respectively; survival at 3 years with CD4+ T-cell count > or =200 cells/microl was 68% and for clinical stage 1 and 2, 80% and 60%, respectively. Clinical stage 3 and 4 were 76% sensitive and 65% specific for predicting a CD4+ T-cell count <200 cells/microl, positive predictive value of 56%, negative predictive value 78%. In all, 82 participants died (41%; mortality rate 216 of 1000 person-years) and was strongly associated with low CD4+ T-cell counts. In conclusion, clinical staging is valid and comparable with staging by CD4 T-cell counts for epidemiologic measurements. Mortality with early disease in Entebbe appears equivalent to that found in the developed world but there is poor survival with advanced disease.