Abstract
OBJECTIVE: Size-matching in heart transplantation is performed using Predicted Heart Mass (PHM). It is common to regard a Donor-Recipient PHM Ratio smaller than 0.86 to be undersized. Novel applications of nonlinear models to estimate the dose-response effect of heart size mismatching can better calibrate postoperative risk. METHODS: Patients receiving an isolated heart transplant ages 18 and older between 2000 and 2022 were included. Donor-Recipient PHM Ratio was continuously weighted using Tübbicke's entropy balance method. A Cox proportional hazards model was applied to the weighted sample. We fit penalized smoothing splines to estimate the continuous effect size of Donor-Recipient PHM Ratio on survival. Significance of the nonlinear spline terms was assessed and confidence bands for effect size were constructed. RESULTS: A total of 47,942 patients met inclusion criteria. The nonlinear spline terms for Donor-Recipient PHM Ratio were significantly associated with survival (p = 0.004). The effect of Donor-Recipient PHM Ratio was approximately flat between 0.9 and 1.2. Hazard grew approximately linearly below 0.9. At 0.86, the effect was estimated to be HR = 1.05 (95% CI: 1.02, 1.09). At 0.8, the effect was estimated to be HR = 1.13 (95% CI: 1.06, 1.20). CONCLUSIONS: Accurately modeling the effect of donor-recipient heart size matching is necessary to inform the decision of whether to accept or reject a donor organ. We found that heart undersizing confers an increased risk of mortality. We fail to find large changes in the slope of the effect at any particular PHM ratio cutoff.