There is increasing evidence that the hypercortisolemia in inflammatory diseases suppresses the elaboration of proinflammatory cytokines, thus protecting the host from its own defence reactions. In severe sepsis and septic shock cortisol levels are usually elevated, but some patients may have relative adrenal insufficiency. This may contribute to the overwhelming systemic inflammatory response syndrome. We evaluated the impact of low-dose hydrocortisone infusion (10 mg/h) on the course of the systemic inflammatory response syndrome. This dose corresponds to a maximum secretory rate of cortisol achieved in corticotropin-stimulated healthy humans. In a prospective observational study 57 surgical patients with severe sepsis or septic shock were studied, of which in addition to the conventional treatment 12 patients were infused with low-dose hydrocortisone, and 45 were treated without any corticosteroid. In the longitudinal analysis the systemic inflammatory response--as judged by body temperature, cardiovascular response, and kinetics of inflammatory mediators such as phospholipase A2, C-reactive protein, and neutrophil elastase--started to differ in favor of the hydrocortisone-treated patients after 2 days of treatment (P < 0.05, Mann-Whitney U test). The difference disappeared after withdrawal of exogenous cortisol. Shock reversal was achieved in all patients treated with low-dose hydrocortisone. The data provide evidence that low-dose hydrocortisone infusion attenuates the systemic inflammatory response in human septic shock. From an immunological point of view a relative cortisol deficiency may contribute to the amplified immune response in systemic inflammatory diseases. A randomized clinical trial must clarify the impact of low-dose hydrocortisone infusion on the clinical course and outcome of septic shock patients.