The basal cortisol level and cortisol response to ACTH stimulation test were assessed in patients with sepsis, the results being compared to a control group of 30 healthy persons. The study group included 49 patients with sepsis and 30 healthy subjects as a control group. The mean age in the study group was 42.6 +/- 18.7 years and 41.4 +/- 12.1 years in the control group. Fifteen of the 49 (30.6%) patients had hospital-acquired and 34 (69.4%) patients community-acquired sepsis. Etiological agent was isolated in 35 (71.4%) patients (57.1% gram negative bacteria and 34.3% gram positive bacteria, plus 8.6% polymicrobial) Fourteen of 49 (28.6%) patients died. Mean basal cortisol level was 597.1 +/- 304.6 nmol/l (range 217.8-1667.9) in the study group and 460.2 +/- 180.8 nmol/l (range 253.6-988.9) in the control group. Mean basal cortisol level in the study group was significantly higher than that of the control group (p < 0.05). Mean basal cortisol level was found to be 725.5 +/- 448.9 nmol/l in the patients who died and 545.8 +/- 210.9 nmol/l in the patients who recovered. The difference between the two groups was found to be significant (p < 0.05). ACTH stimulation test was performed in 43 of the patients and 30 healthy subjects. Cortisol response was significantly lower (mean 277.7 +/- 216.9 nmol/l) in the patients than that detected in the control group (mean 519.6 +/- 279.2) (p < 0.001). Mean cortisol response in the patients who died was 227.2 +/- 224.5 nmol/l and 302.1 +/- 212.7 nmol/l in the patients who recovered (p > 0.05). Adrenocortical insufficiency was detected in 16.3% of the patients and 42.9% of these patients died. In conclusion, sepsis is characterized by high basal cortisol level which may show a poor prognosis and a blunted cortisol response to ACTH stimulation. A small percentage of patients with sepsis may develop adrenocortical insufficiency.