Ischemic brain injury was produced in rats by the intracarotid injection of microspheres. During a 72-hour period of study, large doses of dexamethasone did not favorably affect survival, severity of brain edema, brain electrolyte shifts, or alterations in cerebral red blood cell, albumin, and pertechnetate spaces. The role of steroid therapy in cerebral infarction requires critical clinical appraisal. (27:209-212, 1972) G, riucocorticosteroids have been used in the therapy of cerebral edema since the recognition, in 1945, that adrenal corti¬ cal extract prevented swelling of the cat brain after exposure to air.1 Clinical studies demonstrated the utility of steroids in ede¬ ma due to primary and metastatic brain neoplasm, head trauma, neurosurgical pro¬ cedures, and other causes.2·3 Experimental studies, using various animal models, have largely confirmed a beneficial effect.3Brain edema is a frequent accompaniment of cerebral infarction. This has been demon¬ strated by pathologic studies which suggest that edema is a significant contributory fac¬ tor to the rapidly fatal outcome in patients dying within seven days of acute infarction.4·5 In a retrospective clinical study, Plum showed that brain swelling con¬ tributed to deterioration or death in 22 of 106 patients with acute hemispheric infarcts.6 The use of steroids to treat stroke remains a controversial issue. Initial uncontrolled studies suggested that such therapy might be beneficial.7·8 The results of controlled studies have been discrepant; Rubinstein9 reported favorably, whereas Dyken and White10 found a slightly, but statistically insignificant, higher mortality in the corti¬ sone-treated group. In both groups, rela¬ tively small numbers of patients were studied.Very few experimental studies of this prob¬ lem have been performed, probably due, in part, to the absence of a suitable experimen¬ tal model. Plum and co-workers11 found an adverse effect on mortality following a dose of 0.55 mg/kg of body weight of dexamethasone every 12 hours in rats subjected to hypoxia alone, or unilateral carotid ligation plus hypoxia. Surviving treated animals were unimproved with respect to clinical status, severity of edema and brain electro¬ lyte changes, and altered vascular permea¬ bility to trypan blue. Their results have been criticized because the observations did not extend beyond 24 hours. Pappius and McCann12 demonstrated that studies must be continued for longer periods of time before an effect becomes apparent. The same objec¬ tion applies to the results of Cantu and Ames,13 who found that methylprednisolone, administered 45 minutes to two hours before aortic ligation, did not relieve the cerebro-