An easily calculable and highly predictive risk index for postoperative renal failure after heart transplantation

Objectives: This study derived and validated a risk index for postoperative renal failure after orthotopic heart transplantation. Methods: Adult orthotopic heart transplantations performed between 2000 and 2010 were identified in the United Network for Organ Sharing database. Patients were randomly divided 4: 1 into derivation and validation cohorts. The primary outcome was new-onset postoperative renal failure requiring dialysis. A multivariable model was created incorporating variables associated with renal failure in univariate analysis, with significant risk factors assigned points based on odds ratios. A simple guide was generated to demonstrate ranges of risk scores associated with clinically meaningful renal failure rates. Results: A total of 14,635 orthotopic heart transplantation recipients were included. New-onset postoperative renal failure occurred in 1128 patients (7.7%). A 100-point risk score was generated using 13 significant risk factors. There was a high degree of correlation between actual renal failure rates in the validation cohort and predicted rates in the derivation cohort based on risk scores (r = 0.91, P < .001). Renal failure risk categories were generated on the basis of probabilities determined in the derivation cohort: low (< 5% risk, score 0-15), average (5%-10% risk, score 16-26), above average (10%-20% risk, score 27-39), and high (>20% risk, score >= 40). The actual renal failure rates in the validation cohort for these risk score ranges corresponded to the risk category they were assigned to: score 0 to 15 (4.1% rate), score 16 to 26 (6.8% rate), score 27 to 39 (13.2% rate), and score 40 or more (20.2% rate). Conclusions: This 100-point risk index incorporating 13 risk factors is highly predictive of new-onset postoperative renal failure after orthotopic heart transplantation. Prospective assessment of orthotopic heart transplant recipients using the risk categories that were generated on the basis of score ranges may help in tailoring perioperative management.