La evaluación y la práctica pedagogica en la formación por competencia

In addition to hypoxia, there are several other factors that may trigger seizure activity in children with cerebral malaria. Fever is an obvious candidate, yet the majority of seizures in both this and other studies (Waruiru 1996) occurred at temperatures of below 38®C. In addition, many seizures were partial motor, multiple, or prolonged (lasting for more than 30 minutes), unlike the short-lived generalised seizures characteristic of “febrile seizures” commonly observed in young children. Metabolic disturbances (hypoglycaemia, hyponatraemia, hypocalcaemia, hypomagnesaemia) can also precipitate seizures but, with the exception of two children admitted with hypematraemic dehydration (Na >150mmol/l, urea >2 0mmol/l) and status epilepticus, no association between metabolic derangement and subsequent seizure activity was found in the studies presented here. At

the time that these studies were carried out, chloroquine was widely used throughout Kenya for the treatment of uncomplicated malaria. A number of case reports (Fish 1988; Jaffe 1988; Luijckx 1992) have suggested an association between seizures and the ingestion of chloroquine yet, in children recruited to the placebo arm o f the phenobarbitone prophylaxis study, no association was found between blood chloroquine concentrations on admission and subsequent seizures. Similarly, there was no obvious relationship between the occurrence of seizures and cerebrospinal fluid (CSF) concentrations of quinolinic acid, an endogenous excitotoxin. Compared to a reference UK population, however, children with cerebral malaria had CSF concentrations of quinolinic acid that were significantly raised, and CSF concentrations of nitrate and nitrite, stable breakdown products of nitric oxide, that were significantly reduced. In addition, there was a graded increment in the concentration of quinolinic acid across outcome groups of increasing severity. These studies need to be repeated on a larger number of children, and should include a control group derived from the local population. The mechanism underlying the unexpected reduction in CSF concentrations of nitric oxide metabolites could be explored further by measuring CSF concentrations of arginine (a substrate for nitric oxide synthase) and citrulline (a product of the same enzyme). Quinolinic acid, an N-methyl-D-aspartate (NMDA) receptor agonist, induces the influx of both sodium and calcium into neurones, leading to reversible neuronal swelling and subsequent neuronal disintegration. Nitric oxide is an NMDA receptor blocker, and may provide cerebral protection. The increased CSF concentrations of quinolinic acid and reduced concentrations of nitric oxide metabolites observed in these patients with cerebral malaria therefore lends support to the hypothesis that some of the neurological manifestations of cerebral malaria may be explained by an excitotoxic mechanism. In addition, the advent of NMDA-receptor antagonists, such as magnesium sulphate (Duley

Intervention studies

The best way of proving a causative link between seizures and adverse outcome in children with cerebral malaria is to demonstrate, by means of a randomised, controlled intervention study, that effective anticonvulsant therapy reduces the subsequent incidence of neurological sequelae and death. Phenobarbitone is a highly effective anticonvulsant drug that has been used for years in the treatment of both partial and generalised seizures. It is cheap, widely available throughout Africa, and may be given by intramuscular injection. In children with cerebral malaria, a single intramuscular dose of phenobarbitone 2 0mg/kg produced blood concentrations above the therapeutic level of 15pg/ml within 2 hours of administration. These levels were subsequently maintained for a minimum of 48 hours. An initial clinical tolerance study showed that, at this dose, phenobarbitone had no adverse effects on pulse, respiratory rate, oxygen saturation, venous blood gas or lactate, when compared to placebo. Phenobarbitone halved the proportion of children with prolonged or multiple seizures, and reduced from 2 1% to 13% the proportion of children with neurological sequelae at discharge. In many hospitals throughout Africa, prolonged seizures may go untreated, due to shortages of both staff and equipment. The morbidity and mortahty of seizures under such “real world” conditions may therefore be higher than that observed in this study. The unexpected and disturbing finding o f this study was, however, that the mortality of children treated Avith phenobarbitone 2 0mg/kg was double that of the children given placebo. Post hoc analysis demonstrated that mortality was highest among children treated with both phenobarbitone and multiple (3 or more) doses of diazepam, raising the possibility that phenobarbitone had increased mortality by depressing respiratory drive. It is also possible that, if mortality in the phenobarbitone group had been less, the incidence o f neurological sequelae among this group might have been higher. These findings pose a dilemma for clinicians treating children with cerebral malaria in Africa. Throughout the continent, phenobarbitone remains the single most widely available prophylactic anticonvulsant drug. In children with cerebral malaria, a dose of 20mg/kg is clearly contraindicated. But is there a lower dose of phenobarbitone that is both effective and

priority? There is an urgent need for studies that address these questions but, at this stage, it would seem pragmatic to treat children with recurrent, prolonged convulsions with a single dose of phenobarbitone 10-15mg/kg. Under these circumstances, it would also seem advisable to limit the administration of diazepam to a maximum of two doses. Paraldehyde, another cheap and effective drug for the rapid termination of seizures, could be used as an alternative.