The postresuscitation phase after out-of-hospital circulatory arrest shares similarities with severe sepsis. Corticosteroid replacement is beneficial in patients with septic shock and adrenal dysfunction. The goal of this study was to assess baseline cortisol and adrenal reserve of out-of-hospital circulatory arrest patients after recovery of spontaneous circulation. Thirty-three consecutive patients successfully resuscitated after cardiac arrest were prospectively included between March 2002 and June 2003. A serum cortisol assay and a corticotropin test (250 microg i.v.) were done 6 to 36 h after circulatory arrest. A cortisol increase smaller than 9 microg/dL after corticotropin (nonresponders) defined adrenal reserve insufficiency. Response status was compared in the three outcome groups: survival with full neurologic recovery (n = 4), early death from refractory shock (n = 10), or later death from neurologic dysfunction (n = 19). Patients who died of early refractory shock had lower baseline cortisol levels than patients who died of neurologic dysfunction (27 microg/dL [15-47] vs. 52 microg/dL [28-73], respectively; P < 0.01), suggesting an inadequate adrenal response to severe systemic inflammation. Corticotropin response status was not associated with standard severity markers and seemed uninfluenced by therapeutic hypothermia. In conclusion, patients who die of early refractory shock after cardiopulmonary resuscitation may have an inadequate adrenal response to the stress associated with this condition. Thresholds for cortisol levels at baseline and after corticotropin need to be determined in this clinical setting.