To examine the effect of iron chelation on mortality in cerebral malaria, we enrolled 352 children in a trial of deferoxamine in addition to standard quinine therapy at 2 centres in Zambia, one rural and one urban. Entrance criteria included age <6 years, Plasmodium falciparum parasitaemia, normal cerebral spinal fluid, and unrousable coma. Deferoxamine (100 mg/kg/d infused for a total of 72 h) or placebo was added to a 7 d regimen of quinine that included a loading dose. Mortality overall was 18·3% (32/175) in the deferoxamine group and 10·7% (19/177) in the placebo group (adjusted odds ratio 1·8; 95% confidence interval 0·9–3·6; P = 0·074). At the rural study site, mortality was 15·4% (18/117) with deferoxamine compared to 12·7% (15/118) with placebo (P = 0.78, adjusted for covariates). At the urban site, mortality was 24·1% (14/58) with deferoxamine and 6·8% (4/59) with placebo (P = 0·061, adjusted for covariates). Among survivors, there was a non-significant trend to faster recovery from coma in the deferoxamine group (adjusted odds ratio 1·2; 95% confidence interval 0·97–1·6; P = 0·089). Hepatomegaly was significantly associated with higher mortality, while splenomegaly was associated with lower mortality. This study did not provide evidence for a beneficial effect on mortality in children with cerebral malaria when deferoxamine was added to quinine, given in a regimen that included a loading dose.