Introduction:
U.S. heart allocation policies are currently under review. However, investigations into prioritization schemes for patients waiting for transplantation are hampered by the paucity of post-listing event data and complicated by widespread Mechanical Circulatory Support (MCS) usage.
Hypothesis:
A dynamic score can predict mortality and inform prioritization of patients awaiting heart transplant or considered for listing at time of MCS implant using clinical data, time-varying exposures, and laboratory values during listing.
Methods:
From 1/2008 to 6/2013, patients listed for transplant (n=314, 76%) or implanted with a MCS device with intention to list (n=101, 24%) were identified. Factors for a predictive, time-varying hazard model were selected from 74 pre-listing variables, 8 events and complications, and routine laboratory results after listing.
Results:
Seventy-eight deaths and 381 adverse events occurred during listing. A score based on 17 variables was developed, including black race, actual listing (vs. MCS), pulmonary artery pressure, history of coronary artery disease, diabetes, and ventricular tachycardia at listing, MCS implant and complications, dialysis, respiratory events, and neurologic events on the waitlist, and time-varying serum total bilirubin and creatinine. The figure shows a sample of predicted mortality scores varying over time as events and changes in biomarkers occur.
Conclusions:
We demonstrated a dynamic, updatable mortality risk score that predicts urgency among patients waiting for hearts independent of most demographics, heart failure etiology, symptoms, and cardiac function. This score tracks dynamic changes with time as biomarkers change and adverse events occur. MCS-related complications, neurologic events, and kidney failure significantly affect mortality. Future changes to U.S. heart transplant policies could use this newly developed tool as part of a larger allocation system.