Abstract 3038: Risk Prediction for Mortality Prior to Hospital Discharge for Children Undergoing Heart Transplantation in the Current Era
Background: Although risk factors for early post-transplant mortality in children undergoing heart transplant (HT) are well known, quantitative risk-prediction for early post-HT mortality based on these variables has not been well developed.
Methods: We developed a risk-prediction model for in-hospital early post-transplant mortality in children undergoing HT using data from Scientific Registry of Transplant Recipients for all children< 18 years of age undergoing HT between 1999 and 2009. Logistic regression analysis was used to develop a prediction model for post-HT mortality prior to hospital discharge.
Results: Of 2745 children in the study, the model was developed using a random sample of two-thirds (N=1830 children) and validated in the remaining one-third (N=915). The predictive model had 5 dichotomous variables (CHD, OR, 2.8; 95%CI 1.9, 4.1), ECMO, OR 6.6; 95%CI 4.0, 10.7), ventilator, OR 2.9; 95%CI 1.9, 4.5), dialysis, OR 3.1; 95%CI 1.5, 6.8), creatinine clearance <40 ml/min/m2, OR 2.7; 95%CI 1.6, 4.4). The C-statistic (=0.76) and the Hosmer-Lemeshow goodness-of-fit (P=0.52) suggested moderately strong internal prediction for mortality prior to discharge.
Conclusion: This risk prediction model using 5 baseline clinical factors at the time of transplant has moderate predictive value for in-hospital post-transplant mortality. The model may be useful for guiding decision-making around transplant listing and timing of mechanical support, and informing a broader dialogue about defining heart transplant futility in children.