ObjectiveTo investigate the effects of stress doses of hydrocortisone on the duration of vasopressor therapy in human septic shock.DesignProspective, randomized, double-blind, single-center study.SettingTwenty-bed multidiscipllnary intensive care unit in a 1400-bed university hospital.Patients4 L/min/m2or=to6 [micro sign]g/kg/min). Secondary study end points were the evolution of hemodynamics and the multiple organ dysfunction syndrome (MODS). The severity of illness at recruitment was graded using the Acute Physiology and Chronic Health Evaluation II and the Simplified Acute Physiology Score II scoring systems. MODS was described by the Sepsis-related Organ Failure Assessment score.Interventions155 mmol/L, the hydrocortisone infusion was tapered in steps of 24 mg/day. Physiologic saline solution was the placebo.Measurements and Main ResultsHemodynamic and oxygen-derived variables were measured at previously defined time points over a study period of 5 days. Relevant clinical and laboratory measurements were registered for a study period of 14 days to assess the evolution of organ dysfunction. Baseline data at recruitment did not differ between the two groups. Shock reversal was achieved in 18 of the 20 patients treated with hydrocortisone vs. 16 of the 20 patients treated with placebo. Hydrocortisone significantly reduced the time to cessation of vasopressor support. The median time of vasopressor support was 2 days (1stand 3rdQuartiles, 1 and 6 days) in the hydrocortisone-treated group and 7 days (1stand 3rdQuartiles, 3 and 19 days) in the placebo group (p = .005 Breslow test). There was a trend to earlier resolution of the organ dysfunction syndrome in the hydrocortisone group.ConclusionsInfusion of stress doses of hydrocortisone reduced the time to cessation of vasopressor therapy in human septic shock. This was associated with a trend to earlier resolution of sepsis-induced organ dysfunctions. Overall shock reversal and mortality were not significantly different between the groups in this low-sized single-center study. (Crit Care Med 1999; 27:723-732)