This paper reports a randomised, double-blind, placebo-controlled clinical trial of the effect of routine vitamin A supplementation given on admission to children with severe malaria with regard to survival, recovery during hospitalisation and outcome 6 weeks after discharge. Children aged between 6 and 72 months admitted to the paediatric wards of the Central Hospital of Maputo (CHM), Mozambique with a diagnosis of severe malaria were randomly assigned either to a control group (placebo) or an experimental group (vitamin A) and were followed up 6 weeks after discharge. There were 280 children in the experimental and 290 in the placebo group. Seven (2.5%) and 13 (4.5%) children died in the experimental and the placebo groups, respectively, a relative risk of death of 0.56 (95% CI 0.23-1.38, p = 0.201). During the 1st 5 hours of admission, the relative risk of death in the vitamin A-supplemented group was 2.54 (0.50-12.96); after 5 hours of admission it was 0.19 (95% CI 0.04-0.85; p = 0.015). In the supplemented group, 4/82 (4.9%) of the children developed neurological sequelae vs 2/78 (2.6%) in the placebo group (RR = 1.90; 95% CI 0.36-10.09; p = 0.682). Although the overall reduction in the risk of death observed for all children receiving vitamin A is not statistically significant, it might be clinically important. This finding cannot, however, be accepted as a firm conclusion and requires validation by future trials.