Multinational validation of the PREVENT and SCORE2 cardiovascular risk equations across 6.4 million individuals.
Neuen BL., Major RW., Grams ME., Sang Y., Coresh J., Ballew SH., Surapaneni A., de Pinho NA., Arnlov J., Arnott C., Bell S., Berry J., Brenner H., Ciemins E., Chang AR., Cushman M., de Lemos JA., Diaz-Tocados JM., Farjat AE., Fletcher RA., Gansevoort RT., Heerspink HJL., Herrington WG., Hong C., Horwitz EJ., Hwang S-J., Jassal SK., Kalra PA., Katz R., Khan SS., Kovesdy CP., Kronenberg F., Lees JS., Lloyd-Jones DM., Mill JG., Naimark DMJ., Ndumele C., Polkinghorne KR., Psaty BM., Schloemer P., Shlipak M., Shrestha P., Song Z., Staplin N., Steubl D., Vart P., Visseren FLJ., Woodward M., Yamagishi K., Young BA., Sabanayagam C., Valdivielso JM.
The American Heart Association's PREVENT equations estimate risk of total cardiovascular disease (CVD), atherosclerotic CVD (ASCVD), and heart failure (HF) to guide lipid and blood pressure-lowering therapy in people ages 30 to 79 years in the United States. The SCORE2 risk algorithm is used to estimate CVD risk for similar purposes in people ages 40 and older in Europe. Neither set of equations has been comprehensively validated in global observational cohorts and randomized trials. Here, in 44 observational cohorts and 18 randomized trials, we assessed discrimination and calibration of the two risk algorithms across geographical regions (North America, Europe, Asia/other, multi-region trials). We also created scaling factors for risk prediction over 1-9 years using the PREVENT equations, enabling shorter-term risk prediction for research purposes or to facilitate clinical trial enrolment. Over 5.1 years of mean follow-up, 293,737 PREVENT total CVD events (fatal and non-fatal ASCVD or HF) and 258,086 SCORE2 CVD events (myocardial infarction, stroke, or cardiovascular death) were observed among 6,422,714 and 5,437,384 individuals, respectively. Despite differences in CVD outcome definitions, target populations and predictor variables, overall discrimination and calibration were similar for both equations, with generally good performance across regions, including in multi-regional randomized trials. These findings lend support for adoption of PREVENT or SCORE2 for cardiovascular risk stratification across diverse settings.