Neurotoxicity and artemisinin compounds do the observations in animals justify limitation of clinical use?

High parenteral doses of certain artemisinin derivatives can produce a limited and unique, selective brain stem neuronopathy in laboratory animals. There is necrosis of a small number of nerve cells in certain brain stem nuclei and more extensive chromatolysis of neurons in the same nuclei a few days after intramuscular or intravenous injection of dihydroqinghaosu, artemether and arteether in doses exceeding about 6mg/kg/d intramuscular or intravenous for about 3-5 days (in an oily solvent) in the dog, or after a single parenteral dose exceeding about 100 mg/kg. The limited information available about the monkey suggests that it is only affected after doses several times higher. The probable order of sensitivity of species is dog > rat > monkey, but this is based on only few results. No lesions have been reported after various intramuscular and oral dosages of artesunate and artelinate. The limited reports of clinical observations have not suggested any specific pattern of abnormalities. The lesion is unique in its distribution, in the small number of neurons that become necrotic and the occurrence in nearby cells of chromatolysis. The latter is almost certainly reversible because more prolonged or higher dose studies have not shown more extensive neuronal damage. As the pathogenesis of this toxic response is not known, evaluation of the risk to man must be based on conventional assessment of active doses in animals versus those employed in the treatment of cerebral malaria. It is argued that there is no reason to anticipate a particular risk of conventional regimes employing up to artemether 3-6mg/kg/d intramuscular or other regimes involving artesunate per rectum for a few days.