MELD score is the better predictor for 30-day mortality in patients with ruptured hepatocellular carcinoma treated by trans-arterial embolization

Background and Aims: Spontaneous hepatocellular carcinoma (HCC) rupture is a catastrophic life-threatening complication that could be rescued by trans-arterial embolization (TAE). However, deteriorated liver function with total bilirubin more than 3 mg/dL was deemed as a relative contraindication. This study was aimed to reevaluate this relative contraindication. Methods: Patients with ruptured HCC and treated by TAE between February 2005 and December 2016 in Chang Gung Memorial Hospital, Linkou branch were recruited. Pre-TAE characteristics including age, gender, etiology, liver biochemistry, Child-Pugh classification, Model for End-Stage Liver Disease (MELD) score, the presence of shock, tumor staging and post TAE liver function were compared between patients with and without post-TAE 30-day mortality. Results: A total of 186 patients were enrolled. The successful hemostatic rate after embolization was 91.4% and the median overall survival was 224 days. The 30-day cumulative mortality rate is 20.4%. By multivariate logistic regression analysis, male [aOR: 0.25, P=0.034] MELD score [aOR: 13.61, P<0.001], tumor size [aOR: 1.21, P=0.023] are the independent predictors for 30-day mortality. MELD score has better predictability of post-TAE 30-day mortality than total bilirubin level (AUROC: 0.818 vs. 0.668). The cut-off points of MELD score 13 has higher negative predictive value of 95% for post-TAE 30-day mortality. Conclusion: TAE is effective for the initial hemostasis in patients with HCC rupture. MELD score >= 13 rather than only total bilirubin level >3 mg/dL be more predictive of post TAE 30-day mortality.