OBJECTIVES: To investigate the effects of malaria parasitaemia and clinical malaria on mortality in HIV seropositive and seronegative adults. METHODS: A cohort of adults in rural Uganda were followed from 1990 to 1998. Participants attended routine clinic visits every 3 months and also when sick (interim visits). Information was collected on HIV serostatus, history of fever, current fever and malaria parasite levels. Malaria was categorized as any parasitaemia, significant parasitaemia (>/=1.25 x 10(6) parasites/ml at routine or >/=50 parasites per 200 white blood cells at interim visits) or clinical malaria. The effect of malaria on all-cause mortality was assessed using Cox models. RESULTS: The 222 HIV seropositive participants made 2762 routine visits and 1522 interim visits. During follow-up, of the 211 participants with full records, 69% had at least one episode of parasitaemia, 51% experienced significant parasitaemia and 28% had clinical malaria. There were 90 deaths in 922 person-years of observation. There were no significant associations between numbers of visits with any parasitaemia, significant parasitaemia or clinical malaria on mortality rates. The highest mortality rates were observed in those making four or more routine visits with significant parasitaemia [adjusted mortality rate ratio (RR) 3.27 compared with those making 0 such visits; P=0.078] and those making two or more visits with clinical malaria (adjusted RR 2.23; P=0.093). There was no significant interaction between any malaria category and HIV serostatus. Conclusion We found no evidence of a strong detrimental effect of malaria on all-cause mortality in HIV seropositive adults in this setting.