doiSerbia

Evaluation of treatment outcome in patients with acute-on-chronic liver failure using clinical scores

Milovanović Tamara (University of Belgrade, Faculty of Medicine, Serbia + Clinical Center of Serbia, Clinic for Gastroenterology and Hepatology, Belgrade, Serbia)
Stojković-Lalošević Milica (Clinical Center of Serbia, Clinic for Gastroenterology and Hepatology, Belgrade, Serbia)
Dragašević Sanja (University of Belgrade, Faculty of Medicine, Serbia + Clinical Center of Serbia, Clinic for Gastroenterology and Hepatology, Belgrade, Serbia)
Jocić Nevena (Clinical Center of Serbia, Clinic for Gastroenterology and Hepatology, Belgrade, Serbia)
Baralić Marko (Clinical Center of Serbia, Clinic of Nephrology, Belgrade, Serbia)
Dumić Igor (Mayo Clinic Health System, Department of Medicine, Eau Claire, Wisconsin, USA + Mayo Clinic School of Medicine and Sciences, Rochester, Minnesota, USA)
Pavlović-Marković Aleksandra (University of Belgrade, Faculty of Medicine, Serbia + Clinical Center of Serbia, Clinic for Gastroenterology and Hepatology, Belgrade, Serbia)

Introduction/Objective. Due to a very high mortality risk, acute-on-chronic liver failure (ACLF) patients require early identification and intensive treatment. Precise prediction is crucial for determining the urgency degree and therapy appropriateness, considering high mortality and multitude of clinical resources. The aim of our study was to determine the exact cut-off values of various prognostic scores in the prediction of morality of ACLF. Methods. This prospective study includes chronic liver disease (CLD) patients, admitted due to decompensation, that were subsequently diagnosed with ACLF at the Emergency unit. All patients were evaluated based on various prognostic scores, including Child–Pugh, MELD Na, MELD, SOFA, APACHE II, and CLIF C, which were calculated on admission. Results. Alcoholic liver disease (ALD) was the most common underlying CLD cause (77.9%), followed by viral (8.6%), autoimmune (7.7%), and other causes (5.8%). A total of 37.5% of the patients died at the end of the first month of treatment. Average values of Child–Pugh, MELD Na, MELD, SOFA, APACHE II, and CLIF C scores were significantly higher in patients who died compared to survivors (p < 0.05). CLIF C score showed the best performance with a cut-off value of 50.5, with a sensitivity of 94.9% and specificity of 40%. Conclusion. ACLF remains a condition with a high short-term mortality. Of all of the scores examined in our study, CLIF C proved to be the best scoring system for predicting short term and end of treatment mortality in patients with ACLF.

Keywords: liver failure, ACLF, prognosis, mortality, scores