Early prediction of cardiac allograft vasculopathy and heart transplant failure.

Early risk-prediction is essential to prevent cardiac allograft vasculopathy (CAV) and graft failure in heart transplant patients. We developed multivariate models to identify patients likely to experience CAV, severe CAV, and failure due to CAV, at 1, 5 and 10 years. A cohort of 172 patients was followed prospectively for 6.7 ± 3.9 years. Logistic regression models were developed and cross-validated using bootstrap resampling. Predictive markers of atherothrombosis (myocardial fibrin deposition, and loss of vascular antithrombin and tissue plasminogen activator) and arterial endothelial activation (intercellular adhesion molecule-1 expression) were measured in serial biopsies obtained within 3 months posttransplant. Most markers were univariately associated with outcome. Multivariate models showed that loss of tissue plasminogen activator was the dominant and, in most cases, only predictor of long-term CAV (p < 0.001), severe CAV (p < 0.001), and graft failure due to CAV (p < 0.001). The models discriminated patients having adverse outcomes, had particularly high negative predictive values (graft failure due to CAV: 99%, 99% and 95% at 1, 5 and 10 years) and predicted event incidence and time to event. Early absence of atherothrombotic risk identifies a patient subgroup that rarely develops CAV or graft failure, implying that this low-risk subgroup could possibly be followed with fewer invasive procedures.