Objective: To refine the value of baseline and adrenocorticotropin hormone (ACTH)-stimulated cortisol levels in relation to mortality from severe sepsis or septic shock.
Design: Retrospective multicenter cohort study.
Setting: Twenty European intensive care units.
Patients: Patients included 477 patients with severe sepsis and septic shock who had undergone an ACTH stimulation test on the day of the onset of severe sepsis.
Measurements and main results: Compared with survivors, nonsurvivors had higher baseline cortisol levels (29.5 +/- 33.5 vs. 24.3 +/- 16.5 microg/dL, p = .03) but similar peak cortisol values (37.6 +/- 40.2 vs. 35.2 +/- 22.9 microg/dL, p = .42). Thus, nonsurvivors had lower Deltamax (i.e., peak cortisol minus baseline cortisol) (6.4 +/- 22.6 vs. 10.9 +/- 12.9 microg/dL, p = .006). Patients with either baseline cortisol levels <15 microg/dL or a Deltamax <or=9 microg/dL had a likelihood ratio of dying of 1.26 (95% confidence interval, 1.11-1.44), a longer duration of shock, and a shorter survival time. Patients with a Deltamax <or=9 microg/dL but any baseline cortisol value had a likelihood ratio of dying of 1.38 (95% confidence interval, 1.18-1.61).
Conclusions: Although delta cortisol and not basal cortisol level was associated with clinical outcome, further studies are still needed to optimize the diagnosis of adrenal insufficiency in critical illness. Etomidate influenced ACTH test results and was associated with a worse outcome.