Declining mortality in critically ill patients with cirrhosis in Australia and New Zealand between 2000 and 2015.

BACKGROUND & AIMS: Few studies have described the outcomes of patients with cirrhosis receiving intensive care unit (ICU) admission at a population level. We aimed to describe trends in the mortality of such patients in Australia and New Zealand (ANZ), and to investigate the relationship with associated organ failures.

METHODS: We studied patients admitted to 172 ICUs on a non-elective basis, with and without cirrhosis between January 1st 2000 and December 31st 2015, as recorded by the ANZ Intensive Care Society Centre for Outcome and Resource Evaluation Adult Patient Database. We assessed severity of illness on admission using organ failure models and acute physiology scores. The primary outcome was hospital mortality.

RESULTS: Patients with cirrhosis accounted for 17,044 of 776,873 non-elective ICU admissions (2.2%). Cirrhosis hospital mortality was 32.4% compared to 16.9% in the non-cirrhotic group (p<0.0001). After adjustment for key confounders, cirrhosis had an independent effect on mortality with an odds ratio (OR) of 1.10 (1.06-1.15). There was no difference in the adjusted annual decline in mortality between patients with or without cirrhosis (OR 0.96 [0.95-0.97] vs. 0.96 [0.96-0.96], p=0.67). No difference was seen in the adjusted decline in mortality of patients with cirrhosis when stratified by mechanical ventilation (p=0.92), liver transplant centre status (p=0.27) or presence of sepsis (p=0.09). Mortality increased with number of organ failures, however, the presence of cirrhosis was not found to affect this relationship (p=0.33).

CONCLUSIONS: The mortality of patients with cirrhosis admitted to ICU on a non-elective basis has declined significantly over time, comparable to patients without cirrhosis, and is predominantly governed by the number of organ failures. Outcomes are similar between non-liver transplant ICUs and liver transplant centres.

LAY SUMMARY: The outcomes of patients with liver cirrhosis admitted to intensive care units (ICUs) have been previously regarded as poor. We have demonstrated that in Australia and New Zealand, annual in-hospital death rates following ICU admission in this patient group are lower than previously reported, have improved over 16years to 29% and are at a rate similar to patients without cirrhosis. Our data justify recommendations that advocate better access to intensive care for patients with cirrhosis.