Discharge AI-ECG-Derived Age and Incidence of Cardiac Allograft Vasculopathy at 3 Years: A Pilot Study

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Cardiac allograft vasculopathy (CAV) remains a leading cause of graft failure after heart transplantation, yet an early, non-invasive risk stratification tool is not available. Artificial intelligence electrocardiography (AI-ECG) can estimate biological heart age, and it has been associated with cardiovascular risk. Its deviation with chronological age, called delta age, and its utility in the prediction of CAV has never been explored. We conducted a retrospective single-center study for heart transplant recipients bridged with Impella 5.5 (Abiomed, Danvers, MA, USA) at Mayo Clinic Florida (2020–2023). ΔAge was calculated at discharge and 1-year post-transplant. The primary outcome was confirmed CAV at 1 year. Univariate and multivariable logistic regression models were used to assess the independent association between the ΔAge and the development of CAV, with discrimination evaluated by AUC and internally validated via bootstrap resampling (2000 iterations). Fifty patients were reviewed. CAV developed in 13 (26%) patients in 3 years. CAV+ patients had a significantly higher ΔAge at discharge compared to CAV- patients (median −1.0 vs. −11.0 years, p = 0.0023) and at 1 year (median +7.0 vs. −4.0 years, p = 0.0080). On univariate analysis, ΔAge at discharge was a significant predictor of the development of CAV (OR 1.084 per year, AUC 0.788). On multivariable adjustment, ΔAge at discharge remained an independent predictor (adjusted OR 1.138, 95% CI 1.023–1.265, p = 0.018, and model AUC 0.859). ΔAge at discharge is a novel and potential predictor of the development of CAV within 3 years. These findings support the integration of AI-ECG into post-transplant surveillance as a scalable, low-cost tool.

This publication has 20 references indexed in Scilit: