Abstract 17863: Prospective Validation of a Model to Predict In-Hospital Mortality After Pediatric Heart Transplantation
Background: We previously developed and cross-validated a risk-prediction model to predict in-hospital mortality for children undergoing heart transplant (HT). This model has not been independently validated in a prospective cohort of patients. We sought to prospectively validate this risk-prediction model in an independent cohort of children transplanted during the year since the model was developed.
Methods: We applied a previously developed risk prediction model to all children <18 years undergoing HT between January 2009 and February 2010 identified through OPTN data. The risk prediction model was originally developed and cross-validated (C-statistic 0.78, Hosmer-Lemeshow P 0.89) using a cohort of 2700 children transplanted between 1999 and 2009 using 4 clinical covariates: cardiac diagnosis, support type, renal dysfunction, liver dysfunction. Model performance was assessed using the C-statistic and Hosmer-Lemeshow goodness-of-fit test.
Results: Of 338 children in the prospective validation cohort, the median age was 3.7 yrs (IQR 0.7 to 12.7), 45% had CHD, 4% were on ECMO, 19% were on VAD support, and 4% had severe renal dysfunction. Overall, 6.2% died prior to hospital discharge with the risk of in-hospital mortality ranging from <1% for patients with cardiomyopathy supported on medical therapy with preserved end-organ function to >60% for patients with CHD supported on ECMO with severe end-organ dysfunction. The C-statistic (0.835) and the Hosmer-Lemeshow goodness-of-fit (P=0.54) suggested excellent prediction of in-hospital mortality.
Conclusions: When applied to an independent cohort of patients, this risk prediction model using 4 baseline clinical factors available at transplant had excellent prediction characteristics for in-hospital mortality after HT. This model may be useful to inform decision-making around patient eligibility for transplant, timing of mechanical support, and potentially organ allocation policy.
- © 2010 by American Heart Association, Inc.