A prospective study on adrenal cortex responses and outcome prediction in acute critical illness: results from a large cohort of 203 mixed ICU patients.

OBJECTIVE: To assess whether adrenal cortex hormones predict ICU mortality in acute, mixed, critically ill patients. DESIGN AND
SETTING: Prospective study in consecutive intensive care patients in the general ICU of a teaching hospital. PATIENTS: 203 severely ill patients with multiple trauma (n = 93), medical (n = 57), or surgical (n = 53) critical states. MEASUREMENTS AND
RESULTS: Within 24 h of admission in the ICU a morning blood sample was obtained to measure baseline cortisol, corticotropin (ACTH), and dehydropiandrosterone sulfate (DHEAS). Subsequently a low-dose (1 microg) ACTH test was performed to determine stimulated cortisol. The incremental rise in cortisol was defined as stimulated minus baseline cortisol. Overall, 149 patients survived and 54 died. Nonsurvivors were older and in a more severe critical state, as reflected by higher SOFA and APACHE II scores. Nonsurvivors had a lower incremental rise in cortisol (5.0 vs. 8.3 microg/dl and lower DHEAS (1065 vs. 1642 ng/ml) than survivors. The two groups had similar baseline and stimulated cortisol. Multivariate logistic regression analysis revealed that age (odds ratio 1.02), SOFA score (1.36), and the incremental rise in cortisol (0.88) were independent predictors for poor outcome.
CONCLUSIONS: In general ICU patients a blunted cortisol response to ACTH within 24 h of admission is an independent predictor for poor outcome. In contrast, baseline cortisol or adrenal androgens are not of prognostic significance.