Abstract
Cardiovascular disease (CVD) is a leading cause of late morbidity and mortality in hematopoietic cell transplantation (HCT) survivors. HCT-specific CVD risk prediction models are needed to facilitate early screening and prevention. In the current study, patients who underwent HCT at City of Hope (COH) and survived 1-year free of clinically evident CVD (N = 1828) were observed for the development of heart failure (HF) or coronary artery disease (CAD) by 10-years from index date (1 year from HCT). CVD occurred in 135 individuals (92 HF, 43 CAD). Risk prediction models were developed for overall CVD (HF and/or CAD) using COH-derived integer risk scores. Risk scores based on selected variables (age, anthracycline dose, chest radiation, hypertension, diabetes, smoking) achieved an area under the curve (AUC) and concordance (C) statistic of 0.74 and 0.72 for CVD; these varied from 0.70 to 0.82 according to CVD subtype (HF or CAD). A Fred Hutchinson Cancer Research Center case cohort (N = 580) was used to validate the COH models. Validation cohort AUCs ranged from 0.66 to 0.75. Risk scores were collapsed to form statistically distinct low-, intermediate-, and high-risk groups, corresponding to 10-year cumulative incidences of CVD of 3.7%, 9.9%, and 26.2%, respectively. Individuals in the high- and intermediate-risk groups were at 7.8-fold (95% confidence interval, 5.0-12.2) and 2.9-fold (95% confidence interval, 1.9-4.6) risk of developing CVD (referent group: low risk). These validated models provide a framework on which to modify current screening recommendations and for the development of targeted interventions to reduce the risk of CVD after HCT.