The AKI-Predict-Score

Abstract

Background Postoperative acute kidney injury (AKI) is a frequent complication after liver transplantation and associated with impaired long-term survival rates, chronic kidney disease and higher costs. Although multiple donor, surgical and recipient risk factors have been previously identified, the cumulative impact on the development of AKI remains unknown. Our aim was therefore to design a new model to predict the frequency of severe AKI after liver transplantation. Methods A risk analysis was performed in all adult patients undergoing primary liver transplantation for end-stage liver disease in two centers (2007-2015; n=1230). AKI was defined following KDIGO criteria. A new risk score to predict severe AKI (stage 2 and 3 including renal replacement therapy [RRT]) was calculated based on weight of the factors in a multivariable regression analysis according to the Framingham scheme. Results Overall, 34% of the recipients developed severe AKI, including 18% requiring postoperative RRT. Five factors were identified as strongest predictors for severe AKI: donor BMI, use of a DCD graft, recipient BMI, requirement of fresh frozen plasma during transplantation, and graft implantation time (recipient warm ischemia), leading to 0-25 score points with an AUC of 0.7 in the new AKI-Predict-Score (Figure 1). Three risk classes were identified: low-risk (0-10 points), intermediate-risk (11-20 points) and high-risk group with >20 points. In addition, a score of >20 points correlated with impaired long-term graft survival and more postoperative complications assessed with the Comprehensive Complication Index. Conclusion The AKI-Predict-Score is a new and reliable instrument to identify recipients at risk for severe post-transplant AKI. This score is readily available at end of the transplant procedure. This model offers therefore a great potential to decide which renal protective strategies might be implemented right after liver transplant, i.e. fluid management and modifications in immunosuppression.