Predicting 2-Year Early Allograft Failure After Kidney Transplant: A Perioperative Risk Model from the MOVER Cohort

Background: Allograft failure after kidney transplantation remains common despite improving long-term outcomes and persistent organ scarcity. We aimed to develop and internally validate a perioperative risk model for kidney allograft failure within 2 years of transplantation. Methods: We conducted a single-center retrospective cohort study using the Medical Informatics Operating Room Vitals and Events Repository. Adult patients (≥18 years) undergoing kidney transplantation between January 2018 and July 2023 with at least 2 years of follow-up were included. The primary outcome was allograft failure within 2 years, defined as return to dialysis, pre-emptive re-transplantation, or death. Candidate predictors included demographic characteristics, comorbidities, preoperative laboratory values, and intraoperative variables. After univariate screening and variable selection with LASSO-penalized regression, we estimated relative risks using modified Poisson regression and assessed internal validity with 200 bootstrap resamples. Results: Among 319 recipients, 53 (16.6%) experienced early allograft failure. In the final multivariable model, obesity (relative risk [RR] 4.76; bootstrap 95% CI 2.88–9.31) and thrombocytopenia (RR 1.96; bootstrap 95% CI 1.18–3.38) were independently associated with increased risk. Anemia (RR 0.22; bootstrap 95% CI 0.13–0.37), preoperative clonidine use (RR 0.33; bootstrap 95% CI 0.00–0.85), and female sex (RR 0.55; bootstrap 95% CI 0.26–0.83) were associated with reduced risk. Model performance was modest (pseudo-R2 0.21) but identified clinically distinct risk strata. Conclusions: A five-variable perioperative model based on obesity, thrombocytopenia, anemia, preoperative clonidine use, and female sex identified kidney transplant recipients at differing risk of allograft failure within 2 years. These associations highlight potentially modifiable targets that warrant further study and external validation before clinical use.