A new score to predict outcome after liver transplantation

Background: An ideal liver allocation system should reduce waitlist mortality and also improve post-transplant survival. Aim: To identify a new scoring system that predicts recipient survival at 3 months following liver transplantation (LT) in the Romanian program. Methods: We included into analysis 242 adult patients (183 patients within the training set and 59 in the validation cohort) with liver cirrhosis consecutively transplanted between January 2012 and June 2014. Results: Post-transplant overall survival was 84.2% at 3 months. Independent risk factors for survival following LT were: recipient age >53 years (p=0.01), serum albumin <2.7g/dl (p=0.02), diabetes mellitus (p=0.14), hyponatremia <130mmol/L, presence of non-malignant portal vein thrombosis (p=0.01), retransplantation (p=0.0005) and donor resuscitation following cardiac arrest (p=0.03). AUROC of RoSOFT score is 0.86 in both training and validation set. Diagnostic accuracy of RoSOFT for predicting 3 months mortality is 89.6%. Recipients with HCC outside Milan criteria had a significantly lower MELD score at LT compared to patients inside Milan (p=0.008) and received a higher proportion of marginal organs (p=0.005), but survival did not differ (p=0.47). Conclusions: Combined recipient and donor risk factors can accurately predict 3-months survival following LT in our National LT Program and can be used to improve donorrecipient matching.