We recently published a comparison of two hydrocortisone dosage regimens in patients with septic shock. We compare the results conferred by the two regimens as a function of the response to cosyntropin stimulation test (CST). Patients with septic shock were treated by one of two hydrocortisone regimens: either a 50-mg intravenous bolus every 6 h during 7 days (200 mg group; n = 49), or a 100-mg initial bolus followed by a continuous infusion of 300 mg daily for 5 days (300 mg group; n = 50). Nonresponders was defined as a CST response of 9 μg/dL or less. Nonresponders had more severe septic shock, greater fluid resuscitation needs, and greater vasopressor dependence than responders. When analyzed only as a function of CST results, there was no difference in survival between responders and nonresponders. However, analyses crossing CST results and the treatment regimens showed that patients who were responders and in the 300 mg group had significantly less intensive care unit mortality compared with responders in the 200 mg group (respective mortality of 24% vs. 55% [relative risk 0.43, 95% confidence interval, 0.20 to 0.94, P = 0.018]). Multivariate analysis identified baseline blood cortisol as an independent prognostic factor for 28-day mortality in all groups (hazard ratio 1.002, 95% confidence interval, 1.001 to 1.002, P ≤ 0.0001). The results suggest that in patients who respond to CST, hydrocortisone can provide a dose-dependent benefit. In contrast, nonresponse may indicate corticosteroid resistance. This heterogeneity of response to hydrocortisone may explain the difficulties encountered when trying to demonstrate its benefit in septic shock.