Development of an index to predict posttransplant diabetes mellitus

Abstract

The objectives of this study were to (1) identify risk factors and develop a prediction index for posttransplant diabetes mellitus (PTDM), (2) validate the risk factors and assess the index's predictive ability, and (3) modify the index to enhance its predictive accuracy. Preoperative variables included gender, race, age at transplantation, donor source, number of previous transplants, percentage of ideal body weight, family history of diabetes, and HLA type. Postoperative variables were glucose intolerance on postoperative days 0‐3 and 4‐7. Age, family history, glucose intolerance during postoperative d 4‐7, and specific HLA types were significant independent risk factors for PTDM. No independent effects of race, gender, or donor source were detected. The four independent risk factors produced an equation that accurately predicted PTDM in 77% of the patients. Specificity and negative predictive values reached 75% and 97%, respectively, for the population of men.