Validation of Predictive Nomograms for Delayed Graft Function in Kidney Transplant Recipients
Multi-Organ Transplant Program, Toronto General Hospital, Toronto, ON, Canada.
Meeting: 2015 American Transplant Congress
Abstract number: A33
Keywords: Kidney transplantation, Methodology
Session Information
Session Name: Poster Session A: Delayed Function/Acute Injury/Outcomes/Glomerulonephritis
Session Type: Poster Session
Date: Saturday, May 2, 2015
Session Time: 5:30pm-7:30pm
Presentation Time: 5:30pm-7:30pm
Location: Exhibit Hall E
Background: Delayed graft function (DGF), defined by the need for dialysis within 7 days after transplantation, impacts short and long term graft outcomes. Predictive nomograms for DGF allow for the potential to reduce unfavourable outcomes and expenses associated with kidney transplantation. This study aims to validate the predictive accuracies of four nomograms (Irish 2003, Jeldres 2009, Irish 2011, Chapal 2014), in a contemporary Canadian kidney transplant population.
Methods: This retrospective cohort study examined 1,038 deceased donor kidney transplant recipients between 1 Jan 2000 and 31 Dec 2013. Multivariable logistic regression models were fitted to examine the association between various donor and recipient factors pre-transplant and the occurrence of DGF. Nomogram scores were calculated and accuracy was quantified by the area under the Receiver Operator Characteristics (ROC) curve. Two hundred bootstrap resamples were used for internal validation.
Results: A total of 317 cases of DGF were observed in this population. The Irish 2003 model yielded a mean nomogram risk of 0.551 (95% CI: 0.531, 0.558) among DGF patients, 0.394 (95% CI: 0.385, 0.395) among non-DGF patients, and area-under-curve (AUC) of 0.691. The Jeldres 2009 model yielded a mean probability of 0.322 (95% CI: 0.294, 0.332) among DGF patients, 0.263 (95% C.I. = 0.248, 0.277) among non-DGF patients, and AUC of 0.575. The Irish 2011 model yielded a mean risk of 0.471 (95% CI: 0.454, 0.487) among DGF patients, 0.393 (95% CI: 0.382, 0.404) among non-DGF patients, and AUC of 0.642. The Chapal 2014 yielded a mean risk of 0.379 (95% CI: 0.342, 0.407) among DGF patients, 0.263 (95% CI: 0.239, 0.288) among non-DGF patients, and AUC of 0.595.
Conclusions: Our results indicate that the Irish 2003 DGF nomogram has the best prediction performance in a Canadian cohort of deceased donor kidney transplant recipients. These results have implications for using these nomograms in populations outside the United States.
To cite this abstract in AMA style:
Chau L, Famure O, Kim J. Validation of Predictive Nomograms for Delayed Graft Function in Kidney Transplant Recipients [abstract]. Am J Transplant. 2015; 15 (suppl 3). https://atcmeetingabstracts.com/abstract/validation-of-predictive-nomograms-for-delayed-graft-function-in-kidney-transplant-recipients/. Accessed November 21, 2024.« Back to 2015 American Transplant Congress