Donor and Recipient Factors Predicting Time to Graft Failure Following Orthotopic Liver Transplantation: A Transplant Risk Index

• Published in: Transplantation proceedings 2013-07, Vol.45 (6), p.2077-2082
• Main Authors: Stey, A.M, Doucette, J, Florman, S, Emre, S
• Format: Article
• Language: English
• Subjects: Area Under Curve; Decision Support Techniques; Graft Survival; Humans; Kaplan-Meier Estimate; Liver Transplantation - adverse effects; Middle Aged; Multivariate Analysis; Odds Ratio; Predictive Value of Tests; Proportional Hazards Models; Reproducibility of Results; Retrospective Studies; Risk Assessment; Risk Factors; ROC Curve; Surgery; Time Factors; Tissue Donors; Treatment Failure
• ISSN: 0041-1345; 1873-2623
• DOI: 10.1016/j.transproceed.2013.06.001

Abstract Introduction Much of the success of a transplant depends on appropriate matching of donor to recipient. Methods We validated the Donor Risk Index formula (DRI) using Mount Sinai Medical Center's 10-year cohort of over 1000 transplants. Results The DRI was significantly associated to graft failure with a relative risk (RR) of 1.32. Our cohort had an average DRI of 2.6 with survival of 83% at 3 months, 79% at 1 year, and 70% at 3 years. The low rate of graft failure at a high DRI implies that there are other factors important in transplant pairing that are not considered in the DRI model. To determine these variables and quantify their importance, we constructed a Transplant Risk Index by identifying recipient and donor variables not captured in United Network for Organ Sharing (UNOS) that significantly correlated to graft failure. The most significant independent predictors of time to graft failure were donor age, weight, and peak serum sodium; ischemia time; and recipient creatinine, international normalized ratio, and hepatitis C infection. The coefficients for each of these factors were compiled into a Transplant Risk Index formula. A cutoff of 5 resulted in a graft survival rate of 86% at 3 months, 76% at 1 year, and 62% at 3 years using Kaplan-Meier analysis. The predictive ability of the Transplant Risk Index was greater than the DRI or DMELD (the product of donor age and preoperative MELD) as assessed by the area under the receiver operating curve and the positive and negative predictive values. Conclusion The Transplant Risk Index captures recipient factors and donor factors not captured in UNOS. Including these factors may improve the ability to predict good donor–recipient pairing.