The Prediction of in-Hospital Mortality in Decompensated Cirrhosis with Acute-on-Chronic Liver Failure

Link to article at PubMed

Liver Transpl. 2021 Sep 26. doi: 10.1002/lt.26311. Online ahead of print.


Acute-on-chronic liver failure (ACLF) is a condition in cirrhosis associated with organ failure and high short-term mortality. Both the European (EASL-CLIF) and North American (NACSELD) ACLF definitions have been shown to predict ACLF prognosis. Aim To compare the ability of the EASL-CLIF versus NACSELD systems over baseline clinical and laboratory parameters in the prediction of in-hospital mortality in admitted patients with decompensated cirrhosis.

METHODS: Five NACSELD centers prospectively collected data to calculate EASL-CLIF and NACSELD ACLF scores for admitted patients with cirrhosis who were followed for the development of organ failure (OF), hospital course, and survival. Both the number of OFs and the ACLF grade or presence were used to determine the impact of NACSELD versus EASL-CLIF definitions of ACLF above baseline parameters on in-hospital mortality.

RESULTS: 1031 patients with decompensated cirrhosis (age: 57±11years; male: 66%; Child-Pugh-Turcotte score: 10±2, MELD: 20±8) were enrolled. Renal failure prevalence (28% vs 9% p<0.001) was more common using EASL-CLIF versus NACSELD definition, but those for brain, circulatory and respiratory failures were similar. Baseline parameters including age, admission white cell count, and MELD reasonably predicted in-hospital mortality (area under the curve: 0.76). The addition of number of OFs according to either system did not improve on the predictive power of the baseline parameters for in-hospital mortality, but NACSELD ACLF presence did. However, neither system was better than baseline parameters in the prediction of 30-day or 90-day outcomes.

CONCLUSIONS: The presence of NACSELD ACLF is equally effective as the EASL-CLIF ACLF grade, and better than baseline parameters in the prediction of in-hospital mortality in patients with cirrhosis, but not superior in the prediction of longer-term 30-day or 90-day outcomes.

PMID:34564944 | DOI:10.1002/lt.26311

Leave a Reply

Your email address will not be published.