The clinical significance and even existence of critical illness-related corticosteroid insufficiency is controversial. Here, hypothalamic-pituitary-adrenal (HPA) function was characterized in severe canine Staphylococcus aureus pneumonia. Animals received antibiotics and titrated life-supportive measures. Treatment with dexamethasone, a glucocorticoid, but not desoxycorticosterone, a mineralocorticoid, improves outcome in this model. Total and free cortisol, adrenocorticotropic hormone (ACTH). and aldosterone levels, as well as responses to exogenous ACTH were measured serially. At 10 h after the onset of infection, the acute HPA axis stress response, as measured by cortisol levels, exceeded that seen with high-dose ACTH stimulation but was not predictive of outcome. In contrast to cortisol, aldosterone was largely autonomous from HPA axis control, elevated longer, and more closely associated with survival in early septic shock. Importantly, dexamethasone suppressed cortisol and ACTH levels and restored ACTH responsiveness in survivors. Differing strikingly, nonsurvivors, sepsis-induced hypercortisolemia, and high ACTH levels as well as ACTH hyporesponsiveness were not influenced by dexamethasone. During septic shock, only serial measurements and provocative testing over a well-defined timeline were able to demonstrate a strong relationship between HPA axis function and prognosis. HPA axis unresponsiveness and high aldosterone levels identify a septic shock subpopulation with poor outcomes that may have the greatest potential to benefit from new therapies. septic shock; sepsis; corticosteroids; adrenal insufficiency; cortisol; adrenocorticotropic hormone; aldosterone "RELATIVE ADRENAL DEFICIENCY" in critically ill patients with otherwise normal hypothalamic-pituitary-adrenal (HPA) function was recently renamed "critical illness-related corticosteroid insufficiency" (CIRCI) (33). However, there are no widely accepted criteria for the diagnosis and treatment of this hypothetical condition (40,41,57,58). No test of HPA function reliably predicts outcome, and the relative importance of endogenous glucocorticoid (cortisol) vs. mineralocorticoid (aldosterone) activity to survival is not known. Importantly, patient selection for corticosteroid therapy in sepsis has varied in clinical practice and across randomized trials, in which benefit has not been reproducible (4,40,41,57).To date, HPA testing in patients with septic shock has focused primarily on criteria for inadequate cortisol production using both random cortisol levels and adrenocorticotropic hormone (ACTH)-stimulated adrenal responsiveness. However, both low and high cortisol levels have been associated with worst outcomes (5, 23, 39, 52). Total cortisol increases of less than 9 g/dl in response to high-dose ACTH (250 g) (5) or a random total cortisol below an arbitrary cutoff (e.g., Ͻ10 to Ͻ25 g/dl) (6,20,23,(33)(34)(35) have been proposed to define CIRCI (33). However, neither high-nor low-dose (1 g) ACTH stimulation testing has been shown to identify sept...