IDDF2019-ABS-0257 Factors and models for predicting posthepatectomy liver failure in patients undergoing hepatic resection for hepatocellular carcinoma

Abstract

Background Hepatic resection has been the main treatment option for patients with hepatocellular carcinoma (HCC). Several previous studies attempted to establish reliable criteria to predict mortality or post-hepatectomy liver failure (PHLF) after liver resection Methods A total of 1,481 consecutive patients underwent hepatic resection between January 2016 and December 2017. Laboratory tests were assessed before, and 1,2,3,5,7, and 10 days after surgery. Post-hepatectomy hepatic dysfunction was defined as an increase in the total bilirubin ≥2.9 mg/dL (50 μmol/L) or an increase INR on or after postoperative day 5, compared with the values of the previous day. PHLF includes post-hepatectomy hepatic dysfunction, ICU stay, development of ascites requiring diuretics or drainage procedure, and mortality from any cause. Patients were randomly divided into train set (n=1,111) and validation set (n=370). Results Of 1,565 patients with available data, the mean age was 58.3 years and male comprised 80.6% of patients. 403 (32.1%) patients underwent major hepatic resection and 481(38.3%) revealed cirrhosis at pathologic specimens. The mortality at 30 and 90 days were 0.3% and 0.9%, respectively. PHLF developed in 117(9.3%) patients: posthepatectomy hepatic dysfunction in 62 (4.9%) patients, ICU stay in 18(1.4%) patients, ascites in 42(3.3%) patients, and death 11 (0.9%) patients. In multivariable logistic regression, age over 70 years (adjusted odds ratio [AOR]: 2.89, 95% confidence interval [95% CI]: 1.65–4.97), cirrhosis (AOR:2.60, 95% CI:1.67–4.08), albumin less than 3.5 (AOR:2.17, 95% CI:1.56–3.01), major hepatic resection compared with minor resection (AOR:2.26, 95% CI:1.41–3.62), and high indocyanine green (ICG) index at 15 minutes over 20% (AOR:2.53, 95% CI:1.21–5.46) were significantly associated with a high risk of PHLF. Integer values were assigned to each factor to develop a model that predicted PHLF, which presented an area under the curve of 0.780. Conclusions Cirrhosis, age over 70 years, major resection, lower albumin and high ICG level were associated independently with PHLF. A composite integer-based risk scoring model could accurately predict PHLF in patients undergoing hepatic resection for HCC