Instrumental Variable Estimation of the Effect of Increased Donor Heart Ischemic Time on Post-Transplant Survival

Abstract

<h3>Purpose</h3> As donor hearts are shared more widely under the new U.S. heart allocation policy, more hearts with long cold ischemic time are used for transplant. Ischemic time > 4 hours has previously been associated with lower post-transplant survival. However, the association may have been biased. Transplant programs are more likely to accept organs with longer ischemic time for more urgent candidates with higher risks of mortality post-transplant. Using the policy revision as a natural experiment, we assessed the isolated impact of ischemic time on post-transplant survival. <h3>Methods</h3> Using the Scientific Registry of Transplant Recipients, we examined heart transplants occurring prior to (11/01/2016 - 10/31/2017) and after the policy change (11/01/2018 - 10/31/2019). Graft with extended (>4 hours) cold ischemic time was our exposure of interest. An instrumental variable (IV) model was constructed to address bias from unobserved confounders in the relationship between the receipt of ischemic organ and mortality post-transplant, while controlling for observed candidate and donor characteristics. <h3>Results</h3> An additional 10% of heart recipients post-policy change (n=3418) received an organ with > 4 hours ischemic time compared to the pre-policy period (n=3278) (15% pre vs. 25% post, F-value = 116). Without addressing unobserved confounders, 1-year post-transplant survival was significantly worse among recipients of extendedly ischemic heart (adjusted HR: 1.43, in green, P < 0.001). By using policy period as an instrument to address unobserved confounders, we found no difference in survival between recipients of graft with < 4 hours ischemic time and those who received an extendedly ischemic organ solely due to the policy change (adjusted HR: 0.689, in blue, P = 0.73). <h3>Conclusion</h3> For heart transplant recipients who were "randomized" a graft with > 4 hours of ischemic time by changes in heart allocation policy, we found no effect on survival within the first-year of transplantation.