Alexandre Loupy

Background: In heart transplantation, antibody-mediated rejection (AMR) is a major contributor to patient morbidity and mortality. Multiple routine endomyocardial biopsies (EMB) remain the gold standard to detect AMR, but this invasive procedure suffers from many limitations. We aimed to develop and validate an AMR risk model to improve individual risk stratification of AMR. Methods: Heart recipients from 2 referral transplant centers, Cedars-Sinai (US) and Pitié-Salpêtrière (France), were included from 2012 to 2019. Database included detailed clinical, immunologic, imaging, and histological parameters. The US cohort was randomly distributed in a derivation (2/3) and in a test set (1/3). The primary end point was biopsy-proven AMR. A mixed effect logistic regression model with a random intercept was applied to identify variables independently associated with AMR. Simulation analyzes were performed. Results: The US and French cohorts comprised a total of 1341 patients, representing 12 864 EMB. Overall, 490 AMR episodes were diagnosed (3.8% of EMB). Among the 26 potential determinants of AMR, 5 variables showed independent association: time post-transplant ( P <0.001), pretransplant sensitizing event ( P =0.001), circulating donor-specific anti–human leukocyte antigen antibody ( P =0.001), graft dysfunction ( P =0.004), and prior history of definite AMR ( P <0.001). In the US test set, the calibration and the discrimination of the model were accurate (area under the curve, 0.79 [95% CI, 0.78–0.81]). Those results were confirmed in the external validation cohort (area under the curve, 0.78 [95% CI, 0.77–0.79]) and reinforced by various sensitivity analyses. The model also showed good performance to predict overall cause of rejection. Simulation models revealed that the AMR risk model could safely reduce the number of EMB. Conclusions: Our results support the use of the AMR risk model as a clinical decision tool to minimize the number of routine EMB after heart transplantation.