Dexamethasone has recently been shown to block the production of cachectin (implicated in the pathogenesis of cerebral malaria) if administered prior to endotoxin induction of mouse macrophages. Using the hamster cheek pouch-cerebral malaria model, we tested the hypothesis that dexamethasone is effective as a therapeutic agent in severe malaria if given before some yet undefined trigger point in the disease. Infected hamsters were treated with dexamethasone (0.7 mg/kg) daily on days 7-12, 4-12, or 1-12 post-challenge. When treatment was started on day 1, whole body oxygen consumption (used as a measure of erythrocyte transport to sites of diffusion) on day 12 was greater than (P less than 0.05) that of infected control animals, though the degree of anemia was no different in treated and untreated groups. Furthermore, treatment produced a reduction in monocyte accumulation, capillary malfunction, and monocyte/red blood cell aggregate formation observable in the cheek pouch in vivo and a similar reduction in monocyte presence, capillary pathologic change, and multifocal hemorrhage in the brain on postmortem. These data suggest that mediator(s), whose production can be blocked by pretreatment with dexamethasone, are involved in the pathogenesis of disease leading to death of the Plasmodium berghei infected hamster.