Abstract
BACKGROUND: Hypertrophic cardiomyopathy (HCM) often causes major adverse cardiovascular events (MACE) and worsening heart failure (HF). Endothelial cell (EC) dysfunction is known to be involved in the pathogenesis of HCM. However, the prognostic value of proteins related to EC function (EC-related proteins) in HCM is unknown. METHODS: In this prospective cohort study of patients with HCM, we measured plasma levels of 90 EC-related proteins upon enrollment. The primary outcome measure was MACE. The secondary outcome measure was worsening HF. We developed machine learning models based on EC-related proteins to predict MACE or worsening HF using data from one institution (training set). We tested the predictive ability in independent samples from the other institution (test set) and performed time-to-event analyses. RESULTS: The study included 722 patients (n=458 in the training set and n=264 in the test set). Using our EC-related protein-based model, the area under the receiver-operating-characteristic curve to predict MACE was 0.71 (95% CI, 0.64-0.77) and that for worsening HF was 0.71 (95% CI, 0.63-0.79). When we divided the test set into low- and high-risk groups according to the predicted probabilities derived from the training set, the high-risk groups had significantly higher risks of developing MACE and worsening HF compared with the low-risk groups (both P(logrank)<0.001). CONCLUSIONS: The present prospective study demonstrated that EC-related proteins in plasma predict MACE and worsening HF in patients with HCM. These EC-related proteins have a potential to become novel biomarkers for risk stratification in HCM to improve current prediction models.