Adrenal function in cirrhosis: The pendulum swings

Abstract

Adrenal insufficiency (AI) in critically ill patients is a well recognized phenomenon.1 AI may be associated with structural damage to adrenal gland or hypothalamus or resistance to the biological effects of circulating cortisol. The pathophysiological mechanisms of AI in critically ill patients are endotoxemia, increased levels of resultant proinflammatory mediators, decreased levels of apolipoprotein A and high density lipoprotein.1 The same pathophysiological mechanisms are prevalent in patients with liver cirrhosis; hence the term “hepatoadrenal syndrome” was coined in 2005 for AI in patients with liver cirrhosis observed in a liver transplant intensive care unit.2 In a cohort study of 340 patients with liver disease (patients with fulminant liver failure [FHF], chronic liver disease patients who had undergone transplant in the past) AI was diagnosed in 72%. Since the first report, there have been multiple publications on AI in liver diseases, and a recently published systematic review emphasized some major limitations of this entity.3 First, there is no consensus on appropriate tests and normal values to assess the adrenal function in patients with liver disease. Second (and partly as a result of the first point), there is not yet a consensus definition of AI for patients with liver disease. Third, the frequency of AI may vary between different categories of patients, those with fulminant hepatic failure (FHF), critically ill patients with cirrhosis, patients with stable cirrhosis, and those who have received liver transplant. Thus, reported incidence rates of AI in patients with FHF are 33–62%, critically ill patients with cirrhosis 51–66%, patients with stable cirrhosis 9–64% and patients with liver transplant 61–92%. Unfortunately these studies did not adopt a uniform definition of AI. In addition, four studies reported the use of hydrocortisone to treat AI in FHF, acute on chronic liver failure and cirrhotic patients with septic shock. In none of the studies were survival benefits demonstrated.3 In this issue of the Journal, Thevenot et al.4 report high serum levels of free cortisol (SFC) in patients of cirrhosis who were hemodynamically stable. Those with Child–Pugh Class C had higher basal SFC levels than less severe Child–Pugh Class A and B cases. One year non-transplant mortality was higher in patients with high SFC (more than 79 nmol/L). AI based on serum total cortisol (STC) ranged between 7.4% (T0 STC > 139 nmol/L and 49% (change in STC < 250 nmol/L). The authors conclude that their findings are contrary to the “hepatoadrenal syndrome” theory. When we take a close look all the publications on AI in patients with critical illness and liver cirrhosis, these findings are not surprising. Annane et al.12 reported a three level prognostic classification in septic shock based on serum cortisol levels and cortisol response to corticotrophin. High serum total cortisol level was associated with high mortality rate similar to untreated patients with low cortisol levels. Marik et al.2 also reported high mortality rates in patients with high STC levels similar to patients with adrenal failure; they suggested that STC be performed as a marker of disease severity and poor prognosis. Similarly, STC levels and free cortisol levels were shown to vary widely in stable patients with cirrhosis by Tan et al.13 and Fede et al.14 This suggests patients with stable cirrhosis are in a proinflammatory state leading to increased SFC levels, while during further decompensation there may be exhaustion of the adrenal response to stress leading to AI. In summary, adrenal function in patients with cirrhosis needs to be evaluated with appropriate tests, and consensus on normal values and most appropriate diagnostic indices of adrenal function in cirrhosis are now needed. Further studies on adrenal function and insufficiency during the various stages and complications of liver cirrhosis, including detailed analyses of serial changes during longitudinal follow up would be a valuable initiative in this field. Only when these issues are fully canvassed can we know whether AI is a factor contributing to mortality in cirrhosis, and whether therapeutic intervention with corticosteroid replacement has any role in treatment of patients with advanced cirrhosis. Meanwhile, the pendulum will keep swinging!