To investigate the endogenous adrenocortical response to sepsis, plasma cortisol concentrations were measured in 37 patients (53 ± 3 yr of age) with spetic shock. Patients were studied 11 ± 2 h after shock commenced. Vasopressor therapy was required in 35 of 37 patients (median dopamine infusion rate of 11 μg/kg · min, range 3 to 74). Plasma cortisol concentrations were increased markedly (median 50.7 μg/dl, range 15.6 to 400) above normal values (10 to 20 μg/dl) in patients with septic shock. Neither patients who reversed their shock nor those who survived to hospital discharge had significantly different plasma cortisol concentrations from those who did not. Patients with Gram-positive infections had increased cortisol levels compared with those who had Gram-negative infections (median 83 μg/dl, range 32 to 400 vs. median 44 μg/dl, range 16 to 81, respectively; p < .05). The source of infection, amount of vasopressor infused, and severity of shock were not associated with differences in cortisol concentrations. The length of time in shock before collection of the blood sample for measurements of cortisol and mean arterial pressure at the time of blood collection had significant but weak negative correlations with cortisol concentrations (p < .05, r(s) = -.37 and p < .05, r(s) = -.40, respectively). We conclude that plasma cortisol concentrations are increased in patients with septic shock, but that the degree of increase is variable. This variability may, in part, be releated to type of infection, length of time in shock, and BP at the time of blood sampling. There is no indication that plasma cortisol concentrations may be used in the prediction of patients outcome. Based on this study, adrenal insufficiency as judged by absolute cortisol concentrations is probably rare in septic shock.
