Relative adrenal insufficiency in the non-critically ill patient with cirrhosis: A systematic review and meta-analysis.

Background & Aims: Characterization of relative adrenal insufficiency (RAI) in cirrhosis is heterogeneous with regard to studied patient populations and diagnostic methodology. We aimed to describe the prevalence and prognostic importance of RAI in non-critically ill patients with cirrhosis.

Methods: A systematic review and meta-analysis was performed using MeSH terms and Boolean operators to search five large databases (Ovid-MEDLINE, ScienceDirect, Web of Science, Cochrane Library and ClinicalTrials.gov). The population of interest was patients with cirrhosis and without critical illness. The primary outcome was the pooled prevalence of RAI as defined by a peak total cortisol level <18 μg/dl, delta total cortisol <9 μg/dl or composite of the two thresholds in response either a standard-dose or low-dose short synacthen test. Odds ratios and standardized mean differences from random-effects models estimated important clinical outcomes and patient characteristics by adrenal functional status.

Results: Twenty-two studies were included in final analysis, comprising 1991 patients with cirrhosis. The pooled prevalence of RAI was 37% (95% CI 33-42%). The prevalence of RAI varied by Child-Pugh classification, type of stimulation test used, specific diagnostic threshold and by severity of illness. Ninety-day mortality was significantly higher in patients with RAI (OR 2.88, 95% CI 1.69-4.92, I2 = 15%, p < 0.001).

Conclusions: Relative adrenal insufficiency is highly prevalent in non-critically ill patients with cirrhosis and associated with increased mortality. Despite the proposed multifactorial pathogenesis, no studies to date have investigated therapeutic interventions in this specific population.