BACKGROUND: Falciparum malaria and HIV-1 infection are two of the most important health problems facing sub-Saharan Africa. No convincing evidence of an association between symptomatic malaria and HIV-1 infection has been found. OBJECTIVE: To investigate the effect of HIV-associated immunosuppression on malarial fever rates. DESIGN: An observational cohort study in HIV-specific, primary healthcare clinics in Entebbe, Uganda, on 1371 HIV-1-infected adults participating in a randomized trial of 23-valent pneumococcal vaccine. METHODS: Cohort members underwent routine 6 monthly surveillance and had open clinic access when sick. Episodes of fever were assessed according to standardized protocols. Rates of malaria are described according to HIV immune status determined by CD4 T cell counts. RESULTS: Incidence rates of Plasmodium falciparum malarial fever showed a marked inverse relationship with CD4 T cell count; 140, 93 and 57 cases per 1000 pyo for CD4 T cell groups < 200, 200--499 and > 500 respectively, P < 0.001. Malarial fever definitions incorporating parasite density criteria (derived from asymptomatic surveillance) to correct for chance findings of fever and P. falciparum parasitaemia, did not affect the association of incidence rates with immunosuppression. CONCLUSION: These data support an interaction between symptomatic P. falciparum and HIV. Emphasis on mosquito avoidance measures should be an important component of education and counselling of HIV/AIDS patients in malaria-endemic areas, and suggests an additional HIV-related public health problem in Africa.