Background It is known that the hypothalamic-pituitary-adrenal (HPA) axis is highlighted by stimulation, such as sepsis,trauma, etc, when corticortropin increases and plasma cortisol levels enhance. Relative adrenal insufficiency is not uncommon in critically ill patients and may occur in severe sepsis patients with high plasma cortisol levels. It has been demonstrated that a short corticotropin test has a good prognostic value and is helpful in identifying patients with septic shock at high risk for death, but it has not been established for all severe sepsis patients, especially in China. The aim of this study is to explore the relationship between adrenal function and prognosis in patients with severe sepsis.Methods This prospective study was conducted between July and December 2004 in 6 teaching hospitals. Two hundred and forty patients with severe sepsis were enrolled in this study. A short corticotropin stimulation test was performed in all patients by intravenous injection of 250 μg of corticotropin. Blood samples were taken immediately before the test (T0), 30 (T30) and 60 (T60) minutes afterward, and the plasma cortisol concentration was measured by radio-immunoassay. At the onset of severe sepsis, the following parameters were recorded: age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) Ⅱ, heart rate, mean arterial pressure (MAP), arterial partial pressure of oxygen (PaO2)/fraction of inspired oxygen (FiO2), peripheral blood of hemoglobin, platelets and leukocyte concentration and the number of organ failure. Patients were designated into two groups (survival and non-survival groups) according to the 28-day mortality. Relative adrenal insufficiency was defined as the difference between T0 and the highest value of T30 or T60 (△Tmax) ≤9 μg/dl.Results (1) Two hundred and forty patients with severe sepsis were included in this study, with 134 patients in the survival group and 106 in the non-survival group. The 28-day mortality was 44.2%. (2) Between the survival group and non-survival group age, APACHE Ⅱ, peripheral blood of platelets, the number of organ failures, T0 and △Tmax showed significant differences. T0 was (23±10) μg/dl and (36±18) μg/dl in the survival group and nonsurvival group respectively.△Tmax was (18±9) and (10±8) μg/dl in the survival group and non-survival group respectively. The areas under the ROC curve for T0 and △Tmax were both 0.72, and the area under the ROC curve for APACHE Ⅱ was 0.70. By multivariate analysis age, T0, the number of organ failures and relative adrenal insufficiency (△Tmax ≤9 μg/dl) were independent predictors of death. (3) The incidence of relative adrenal insufficiency was 38.3% in total, 19.4% in the survival group and 62.3% in the non-survival group (P＜0.001). The 28-day mortality was 71.7% among the relative adrenal insufficiency patients but 27.0% among normal adrenal function patients.Conclusions The prevalence of relative adrenal insufficiency is high in severe sepsis. Relative adrenal insufficiency has a good prognostic value for severe sepsis.