These studies were undertaken to evaluate the optimal effects of corticosteroids in an experimental septic shock model, with regard to time of therapy institution and the dose of corticosteroids (dexamethasone and methylprednisolone). Septic shock was induced by intraperitoneal injection of 5 x 10(8) live E coli bacteria/100 g body weight. Hematocrit was measured before and 4 h after injection of bacteria. Survival time was recorded continuously for 24 h and then daily for 7 days. Mortality rate correlated with hematocrit increase at 4 h in untreated control animals. Surviving rats did not increase their hematocrit, but nonsurvivors did. Survival time in untreated rats was 9.2 +/- 0.2 h (SEM). Corticosteroids significantly increased survival time when given prophylactically. This effect, however, decreased linearly with time when dexamethasone was given up to 8 h after bacteria injection. Dexamethasone (3-96 mg/kg) and methylprednisolone (19-225 mg/kg) significantly prolonged survival time. Twenty-two percent (73/329) of corticosteroid-treated animals lived longer than 24 h compared to 14% (19/135) of the controls. This difference was of borderline significance (P = 0.05).