Validation of the SCORE2 risk prediction algorithm in a Portuguese population: A new model to estimate 10-year cardiovascular disease incidence in Europe

Abstract

Introduction and Objectives

Subjects without cardiovascular (CV) disease (CVD) may suffer from subclinical atherosclerosis, and are at increased risk for atherosclerotic CV events (ASCVE). The ESC/EAS risk SCORE was updated by SCORE2, which estimates 10-year risk of fatal and non-fatal CVD in European populations aged 40–69 years without established CVD or diabetes. Our aim was to compare the two ESC/EAS risk scores and to validate SCORE2 in our population.

Methods

A total of 1071 individuals (age 57.2±6.1 years; 75.2% male) without CVD or diabetes, from GENEMACOR study controls, were analyzed over 5.4±3.9 years. The population was stratified into risk categories according to the two scores, and the area under the ROC curve (AUC) and Harrell's C-index assessed the scores’ performance. Calibration was performed using the goodness-of-fit test, and occurrence of the first event assessed by Cox regression. Kaplan–Meier analysis estimated SCORE2 survival.

Results

SCORE stratified subjects into four risk categories: low (7.4%), moderate (46.5%), high (25.3%) and very high (20.8%), and SCORE2 into three: low-to-moderate (24.7%), high (59.0%) and very high (16.2%). SCORE presented good discrimination for CV mortality (AUC=0.838; C-index=0.834, 95% CI: 0.728–0.940), as did SCORE2 for total CV events (AUC=0.744; C-index=0.728, 95% CI: 0.648–0.808). Calibration did not show a disparity between observed and expected ASCVE. The probability of ASCVE was eight times higher in very-high-risk SCORE2 (p=0.001), and three times in the high-risk group (p=0.049). Event-free survival was 99%, 90% and 72% in the low-to-moderate, high and very-high-risk categories, respectively (p<0.0001).

Conclusions

SCORE2 improved population stratification by identifying higher-risk patients, enabling early preventive measures. It showed good discriminative ability for all ASCVE.