ObjectiveTo identify the incidence of secondary adrenal insufficiency in severe sepsis.DesignProspective clinical trial testing 100 patients with a 250-&mgr;g adrenocorticotropic hormone (ACTH) stimulation test.SettingCounty-university teaching hospital.PatientsOne hundred patients with sepsis and septic shock. Forty patients had bacteremia and 17% shock.InterventionsACTH, cortisol, aldosterone, and electrolyte concentrations were measured at baseline. Cortisol and aldosterone were measured 30 and 60 mins after ACTH (250 &mgr;g).Measurements and Main ResultsNine of the 100 patients (9%) failed the ACTH stimulation test (all serum cortisol <20 &mgr;g/dL). The 91 patients with sepsis began with a serum cortisol at 29.3 ± 2.5, and it increased to 40.1 ± 2.6 and 46.9 ± 2.7 &mgr;g/dL at times 30 and 60 mins, respectively. Serum cortisol in nine septic patients who failed the ACTH stimulation test had an initial concentration of 11.3 ± 1.8 &mgr;g/dL, and it increased at time 30 mins to 14.0 ± 1.9 &mgr;g/dL and at 60 mins to 15.7 ± 1.8 &mgr;g/dL. Four of the nine patients had secondary adrenal insufficiency as determined by a normal aldosterone response to ACTH. The remaining five patients had an absent aldosterone response to ACTH and baseline ACTH concentrations that were not elevated, suggesting adrenal dysfunction. Serum sodium (128 ± 4 vs. 138 ± 1 mmol/L,p< .05) and glucose concentrations (121 ± 20 vs. 163 ± 11 mg/dL,p< .05) were reduced in the nine patients. Of the four patients with secondary adrenal insufficiency, two had a history of amenorrhea after birth of their children many years earlier.ConclusionsThese data demonstrate that 9% of adults with sepsis fail the ACTH stimulation test due to a mixture of etiologies. A reduced sodium or glucose concentration may be helpful in identifying glucocorticoid (adrenal) insufficiency in patients with sepsis.