Risk scores for predicting incident heart failure admission in patients with chronic coronary syndromes: Validation in a prospective, monocentric, long-term, cohort study.

Abstract

Heart failure (HF) admission in chronic coronary syndrome (CCS) patients has a prognostic impact. Stratification schemes have been described for predicting this endpoint, but none of them has been externally validated. Our aim was to develop point scores for predicting incident HF admission with data from previous studies, to perform an external validation in an independent prospective cohort and to compare their discriminative ability for this event. Independent predictive variables of HF admission in CCS patients without baseline HF were selected from four previous prospective studies (CARE, PEACE, CORONOR and CLARIFY), generating scores based on the relative magnitude of the coefficients of Cox of each variable. Finally, the scores were validated and compared in a monocentric prospective cohort. The validation cohort included 1212 patients followed for up to 17 years, with 171 patients suffering at least one HF admission in the follow-up. Discriminative ability for predicting HF admission was statistically significant for all, and paired comparisons among them were all nonsignificant except for CORONOR score was superior to CLARIFY score (C-statistic 0.73, 95%CI 0.69-0.76 vs. 0.69, 95% CI 0.65-0.73; p= 0.03). All tested scores showed significant discriminative ability for predicting incident HF admission in this independent validation study. Their discriminative ability was similar, with significant differences only between the two scores with higher and lower performance.