[Skip to Content]
[Skip to Content Landing]
Citations 0
July 19, 2000

Prognostic Value of Cortisol Response in Septic Shock

Author Affiliations

Stephen J.LurieMD, PhD, Contributing EditorIndividualAuthor

JAMA. 2000;284(3):308-309. doi:10.1001/jama.284.3.303

To the Editor: Dr Annane and colleagues1 found that in patients with septic shock, a high baseline serum cortisol level (>34 µg/dL), or a maximum increase in stimulated cortisol concentration of ≤9 µg/dL (following high-dose cosyntropin testing), or both predicts a higher mortality. We believe that many of the hypotheses on which this study is based may be incorrect.

The degree of activation of the sympathetic nervous system and hypothalamic-pituitary-adrenal (HPA) axis is related to the severity of the stressor. In animal models, hypotension and sepsis are second only to decapitation as the most intense stressors.2 Therefore, serum cortisol levels must be interpreted in the context of the clinical state. In patients with severe stress (eg, hypotension, sepsis, trauma, surgery), most studies find random cortisol levels of greater than 25 µg/dL.3 The change in cortisol level following cosyntropin stimulation is a measure of adrenal reserve and not of adrenal function, and therefore must be interpreted in the context of the baseline cortisol level.

We believe that the study by Annane et al failed to identify those patients with adrenal insufficiency (AI) who would have benefited from treatment with corticosteroids. Recent data suggest that systemic inflammatory states such as sepsis can be associated with reversible AI due to HPA axis suppression by cytokines and other inflammatory mediators.4 Traditionally, adrenal function has been tested using 250 µg of cosyntropin. A plasma level of 100 to 300 pg/mL of corticotropin should produce a maximal cortisol response. However, plasma levels of corticotropin after 250 µg of cosyntropin are generally greater than 10,000 pg/mL. Thus, 250 µg of corticotropin is supraphysiologic. We and others prefer to use a more physiologic dose of 1 µg of corticotropin, which produces a corticotropin level of approximately 300 pg/mL.5,6 Using the 1-µg dose, we have found that many critically ill patients in shock with baseline stress cortisol levels of less than 20 µg/dL who fail to respond to the 1-µg dose can generate cortisol levels of higher than 20 µg/dL following administration of 250 µg of corticotropin.5 These patients typically improve clinically with hydrocortisone treatment.

We believe, therefore, that critically ill patients can be stratified into 3 groups according to their baseline serum cortisol levels. The first group of patients should have baseline serum cortisol levels of less than 20 µg/dL. This group is likely to have AI and high mortality. Treatment with corticosteroids may reduce mortality in these patients. Patients with high baseline cortisol levels (>45 µg/dL) are those with severe stress, who are likely to have high mortality as a result of overwhelming illness. Patients whose baseline serum cortisol levels are between 20 µg/dL and 45 µg/dL are likely to have the best prognosis.

Annane  DSébille  VTroché  GRaphaël  JCGajdos  PBellissant  E A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin.  JAMA. 2000;283:1038-1045.Google Scholar
Hart  BBStanford  GGZiegler  MGLake  CRChernow  B Catecholamines: study of interspecies variation.  Crit Care Med. 1989;17:1203-1218.Google Scholar
Streeten  DHAnderson Jr  GHBonaventura  MM The potential for serious consequences from misinterpreting normal responses to the rapid adrenocorticotropin test.  J Clin Endocrinol Metab. 1996;81:285-290.Google Scholar
Jaattela  MIlvesmaki  VVoutilainen  RStenman  UHSaksela  E Tumor necrosis factor as a potent inhibitor of adrenocorticotropin-induced cortisol production and steroidogenic P450 enzyme gene expression in cultured human fetal adrenal cells.  Endocrinology. 1991;128:623-629.Google Scholar
Marik  PEKiminyo  KOlexo  SZaloga  GP Occult adrenal insufficiency in critically ill patients: an underdiagnosed entity.  Crit Care Med. 1999;27(suppl):A141.Google Scholar
Henzen  CSuter  ALerch  EUrbinelli  RSchomo  XHBriner  VA Suppression and recovery of adrenal response after short-term high-dose glucocorticoid treatment.  Lancet. 2000;355:542-545.Google Scholar