• Coma recovery time (that is time from beginning of antimalarial treatment until patient has fully regained consciousness).
  

• Time taken to sit un supported
  
• Time to begin oral intake
  
• Duration of hospitalisation
  
• Mortality
  
• Proportion of children recovering with neurological sequelae
  

Cerebral malaria is a life-threatening complication of Plasmodium falciparum infection accounting for significant morbidity and mortality in African children despite availability of quinine, the current drug of choice. The case fatality ranges from 5 to 40% with almost 10% of survivors experiencing neurological sequelae.

Several reports have suggested that raised intracranial pressure (ICP) may be a feature of cerebral malaria. There is evidence of brain swelling on computer tomography, magnetic resonance imaging and at necropsy. It has been postulated that raised intracranial pressure can cause death by transtentorial herniation or by compromising cerebral blood flow. In fact, most children who died of cerebral malaria in a Kenyan study, had clinical signs compatible with transtentorial herniation and all those who had severe ICP (maximum ICP > 40mmHg) either died or survived with neurological sequelae.

Mannitol, an osmotic diuretic, effectively lowers ICP and is used to treat post traumatic raised intracranial pressure. There have been some case reports of reduction in mortality and morbidity in African children with cerebral malaria following administration of mannitol, but as these were not randomized controlled trials it is difficult to evaluate their significance. Currently the WHO contends that there is insufficient evidence for using mannitol as adjunct therapy for cerebral malaria.

A recent Cochrane review found no randomized or quasi-randomized controlled trial to support or refute the use of mannitol as adjunct therapy for cerebral malaria.

Hypothesis: A single dose of intravenous mannitol (1g/kg) given to children with cerebral malaria will reduce mean coma recovery time from 22.5 to 13.1 hours.

We calculated a sample size of 78 patients in each group for 90% power and 95% confidence. In the calculation, we assumed that the children receiving intravenous mannitol would have a mean coma recovery time of 13.1 (SD 18.5) hours and those receiving placebo would have a mean coma recovery time of 22.5 (SD 18.5) hours (42.3% effect size), according to a recent study by Aceng, Byarugaba and Tumwine in the same hospital.