The new liver allocation score for transplantation is validated and improved transplant survival benefit in Germany but not in the United Kingdom

Abstract

Prognostic models for the prediction of 90-day mortality after transplantation with pretransplant donor and recipient variables are needed to calculate transplant benefit. Transplants in adult recipients in Germany (Hannover, n = 770; Kiel, n = 234) and the United Kingdom (Birmingham, n = 829) were used for prognostic model design and validation in separate training and validation cohorts. The survival benefit of transplantation was estimated by subtracting the observed posttransplant 90-day mortality from the expected 90-day mortality without transplantation determined by the Model for End-Stage Liver Disease (MELD) score. A prognostic model called the liver allocation score (LivAS) was derived using a randomized sample from Hannover using pretransplant donor and recipient variables. This model could be validated in the German training and validation cohorts (area under the receiver operating characteristic curve [AUROC] > 0.70) but not in the English cohort (AUROC, 0.58). Although 90-day mortality rates after transplantation were 13.7% in Hannover, 12.1% in Kiel, and 8.3% in Birmingham, the calculated 90-day survival benefits of transplantation were 6.8% in Hannover, 7.8% in Kiel, and 2.8% in Birmingham. Deployment of the LivAS for limiting allocation to donor and recipient combinations with likely 90-day survival as indicated by pretransplant LivAS values below the cutoff value would have increased the survival benefit to 12.9% in the German cohorts, whereas this would have decreased the benefit in England to 1.3%. The English and German cohorts revealed significant differences in 21 of 28 pretransplant variables. In conclusion, the LivAS could be validated in Germany and may improve German allocation policies leading to greater survival benefits, whereas validation failed in England due to profound differences in the selection criteria for liver transplantation. This study suggests the need for national prognostic models. Even though the German centers had higher rates of 90-day mortality, estimated survival benefits were greater. Liver Transplantation 22 743-756 2016 AASLD.