The adrenocorticotropic hormone-induced cortisol response in acute pancreatitis

Abstract

The evidence that severe acute pancreatitis can result in critical illness-related corticosteroid insufficiency following impaired adrenal secretion is accumulating. The study by Peng and coworkers in Critical Care certainly contributes to that idea, even though the question whether relative adrenal insufficiency should prompt for treatment by substitution doses of corticosteroids remains unresolved. The study is discussed in terms of the risk factors, circumstances and significance of impaired corticosteroid secretion by adrenals in severe acute pancreatitis.

In a recent issue of Critical Care, Peng and coworkers [1] elegantly document a relatively low circulating cortisol response to adrenocorticotropic hormone (ACTH) stimulation in non-survivors (versus a normal response in survivors) of acute biliary pancreatitis. This was associated with more frequent bacteremia and severe disease. In fact, the ACTH-induced cortisol response inversely related to disease severity measures, as noted before [2]. None of the patients were treated by adrenal substitution doses of corticosteroids, even though often suffering from septic shock, as can be deduced from the data, likely associated with infection of pancreatic necrosis and subsequent bacteremia, as commonly occurs in severe cases [3]. Hence, it was suggested that insufficient adrenal cortisol secretion (reserve), relative to severity of disease, contributed to a poor outcome, reflecting critical illness-related corticosteroid insufficiency (CIRCI) even though baseline cortisol values did not differ between outcome groups.

Even though the recent Corticosteroid Therapy of Septic Shock (CORTICUS) study [4] does not suggest either the presence of CIRCI (as defined above) and a benefit of corticosteroid treatment in (relatively mild) septic shock, the concept can still not be refuted in view of a number of other studies documenting the issues in more severe septic shock [2, 5, 6]. Nevertheless, the question remains whether (septic) shock originating from pancreatitis would manifest differently than septic shock from other origins, and, thus, whether shock originating from pancreatitis would similarly benefit from corticosteroid treatment. To prove the latter, studies including large numbers of subjects are probably needed, so a definitive answer is unlikely to be provided in the near future. Therefore, current practise recommendations can only be based on the literature on 'general' septic shock, and substitution doses of corticosteroids are still considered useful in the early treatment of the syndrome according to a recent metaanalysis [6], particularly when associated with relatively low ACTH-induced cortisol responses [2] and in spite of the negative CORTICUS trial [4, 5].

A current study on pancreatitis [1] is also in line with the observed adrenal enlargement, probably caused by increased blood flow and haemorrhage/necrosis following intense stress-induced stimulation during acute pancreatitis, the exhausted secretion reserve upon ACTH stimulation noted before in pancreatic necrosis/infection associated with severe disease and non-survival, and the positive haemo-dynamic effects of substitution doses of corticosteroids, even in the absence of documented infection, reported by others [7–11]. Indeed, a direct role for CIRCI in contributing to pancreatic necrosis has been suggested [8, 10]. Finally, there is experimental evidence for adrenal changes during pancreatitis, a role for phospholipase A2 in this condition and the benefit of low-dose cortisol administration in the development and ultimate outcome of necrosis [1, 10, 12, 13]. However, we can not determine the cause and effect relationships from the associations reported by Peng and coworkers [9], since the independent predictive value for mortality of ACTH-induced cortisol responses was not studied. Such predictive value, independent of other markers of disease severity, would indeed argue in favour of a contributory role of CIRCI in outcome and treatment by substitution doses of corticosteroids. Moreover, optimal cutoff levels in ACTH-induced cortisol responses with the highest predictive values for outcome are missing, and the increase of <250 nmol/l (9 mg/dL), found to discriminate in other studies [4, 6, 10], was used as the cutoff to denote CIRCI, even though it is dependent, almost by definition, on the severity of disease [2].

Notwithstanding some limitations, the study in Critical Care by Peng and coworkers [1] supports the idea that acute (necrotic/infected) pancreatitis can result in CIRCI following relative adrenal insufficiency, that the ACTH test can help to recognise it, and that the condition likely warrants treatment with substitution doses of corticosteroids. More data are welcome that can be included in a systematic analysis to confirm these ideas.