Diagnosing Adrenal Insufficiency in Sepsis

Abstract & Commentary

By David J. Pierson, MD, Professor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington. Dr. Pierson reports no financial relationships relevant to this field of study.
This article originally appeared in the February 2007 issue of Critical Care Alert. It was peer reviewed by William Thompson, MD. Dr. Thompson is Staff Pulmonologist, VA Medical Center; Associate Professor of Medicine, University of Washington. Dr. Thompson reports no financial relationships relevant to this field of study.

Synopsis: Using the metyrapone test as a gold standard, the authors confirmed that a resting serum cortisol level < 10 micrograms/dL, or an increase of < 9 micrograms/dL after cosyntropin stimulation, accurately diagnoses adrenal insufficiency in ICU patients with severe sepsis or septic shock.

Source: Annane D. Diagnosis of adrenal insufficiency in severe sepsis and septic shock. Am J Respir Crit Care Med. 2006;174:1319-1326.

In an attempt to clarify criteria for diagnosing adrenal insufficiency in patients with severe sepsis and septic shock, this group of French investigators serially performed both the cosyntropin stimulation test and the metyrapone stimulation test in 2 cohorts of patients, and compared the findings with those from concurrent ICU patients without sepsis, as well as a group of healthy volunteers. Annane and colleagues used the overnight metyrapone test (no longer available in the United States) as a gold standard for establishing corresponding diagnostic criteria using the cosyntropin test.

The 2 consecutive cohorts of septic patients (61 and 40 patients, respectively) met consensus, clinical, diagnostic criteria of the American College of Chest Physicians and Society of Critical Care Medicine for severe sepsis or septic shock. These septic cohorts did not differ from each other by any of the criteria examined. ICU patients without sepsis (n = 44, including patients with drug overdoses, cardiogenic pulmonary edema, acute inhalation injury, status epilepticus, and other presumably non-infectious processes) were younger and less severely ill, had fewer comorbidities, and required less support in the form of mechanical ventilation and vasopressors. The 32 healthy volunteers were age-matched to the septic patients. On ICU admission, patients had blood drawn before and 60 minutes after intravenous administration of 250 micrograms of cosyntropin. At least 8 hours later, they received metyrapone, 30 mg/kg, via nasogastric tube, with blood samples drawn before and 8 hrs after metyrapone administration. The outcomes of interest were serum cortisol and free cortisol at the 2 baselines and after stimulation.

Criteria for adrenal insufficiency by the metyrapone test (an increment from baseline in 11-beta-deoxycortisol concentration of < 7 micrograms/dL with fall in cortisol level to < 8 micrograms/dL) were met in 31/61 and 24/40 of the septic patients in the 2 cohorts, respectively, or approximately 60% in each group. Corresponding results were 3/44 (7%) in the ICU patients without sepsis, and none in the healthy volunteers. The best predictor of adrenal insufficiency as documented by metyrapone stimulation was the presence of a baseline serum cortisol level of < 10 micrograms/dL or a post-cosyntropin stimulation increase of < 9 micrograms/dL: these cut-offs had a specificity of 96%, a positive predictive value of 94%, and a positive likelihood ratio of > 10. Despite varying albumin levels in the septic and non-septic patients and normal volunteers, free cortisol performed no better than total cortisol for any diagnostic variable.


Patients with septic shock are particularly prone to development of relative adrenal insufficiency, and patients who manifest this phenomenon do less well clinically than those who do not. The mechanism is not known, but sepsis appears to predispose patients to relative adrenal insufficiency more than other critical illnesses of equivalent severity. Annane and colleagues, who documented this phenomenon in previous studies that have led to widespread testing and treatment for relative adrenal insufficiency in patients with sepsis, have now clarified several areas of uncertainty with respect to its diagnosis. In patients with severe sepsis or septic shock, adrenal insufficiency is very likely when baseline cortisol levels are less than 10 micrograms/dL or the level increases by less than 9 micrograms/dL 60 minutes after cosyntropin administration. This diagnosis is unlikely when the baseline cortisol level is 44 micrograms/dL or more, or when the increase after cosyntropin exceeds 16 micrograms/dL. Serum cortisol measurements are sufficient: even in hypoalbuminemic patients, there is no need for the more expensive measurement of free cortisol.