BACKGROUND: Malaria and HIV are important pediatric problems in sub-Saharan Africa. It is uncertain how HIV-related immunosuppression and malaria interact in children. We aimed to describe associations among HIV status, presentation and outcome from malaria in children from Hlabisa district, KwaZulu-Natal, South Africa, a region of high HIV prevalence and unstable Plasmodium falciparum transmission. METHODS: Consecutive febrile children were screened for malaria with a rapid antigen test. After consent was given, clinical data were recorded, and blood spots were obtained for HIV antibody testing. Cases were managed according to national guidelines. RESULTS: Malaria was diagnosed in 663 children, of whom 10.1% were HIV antibody-positive. Semiquantitative asexual and sexual stage parasitemia densities were unrelated to HIV status. Overall 161 children were hospitalized; 19 (12%) were <1 year old; and 41 (25%) had severe/complicated malaria. Severe disease presented more frequently in HIV antibody-positive than in HIV-uninfected children (P = 0.05), particularly in those >1 year old with coma (P = 0.02) and hypoglycemia (P = 0.05). Receiving parenteral antibiotics was associated with severe disease (odds ratio, 3.0; 95% confidence interval, 1.3 to 6.7) whereas a low white blood cell count (<4 x 10(6)/l) was associated with nonsevere disease (odds ratio, 0.4; 95% confidence interval, 0.2 to 0.8). Seven children (4.3%) died. Coma, age <1 year and low white blood cell count were the clearest predictors of poor outcome. CONCLUSION: HIV infection was associated with severe/complicated malaria, although the magnitude of the effect may be relatively small. Given that both malaria and HIV are widespread in Africa, even small effects may generate significant morbidity and mortality and major public health consequences.