In Switzerland, two distinct algorithms are recommended for cardiovascular prevention: (a) Arbeitsgruppe Lipide und Atherosklerose (AGLA); and (b) European Society of Cardiology (ESC). We validated and determined which algorithm better predicts incident atherosclerotic cardiovascular disease and assessed statin eligibility in Switzerland.
A prospective population-based cohort.
We employed longitudinal data of the CoLaus study involving 6733 individuals, aged 35-75 years, with a 10-year follow-up. Using discrimination and calibration, we evaluated the predictive performance of the AGLA and ESC algorithms for the prediction of atherosclerotic cardiovascular disease.
From the 6733 initial participants, 5529 were analysed with complete baseline and follow-up data. Mean age (SD) was 52.4 (10.6) years and 54% were women. During an average follow-up (SD) of 10.2 years (1.7), 370 (6.7%) participants developed an incident atherosclerotic cardiovascular disease. The sensitivity of AGLA and ESC algorithms to predict atherosclerotic cardiovascular disease was 51.6% (95% confidence interval (CI) 46.4-56.8) and 58.6% (53.4-63.7), respectively. Discrimination and calibration were similar between the AGLA and ESC algorithms, with area under the receiver operating characteristic curve values of 0.78 (95% CI 0.76-0.80) and 0.79 (0.76-0.81), and Brier scores of 0.059 and 0.041, respectively. Among 370 individuals developing incident atherosclerotic cardiovascular disease, only 278 (75%) were eligible for statin therapy at baseline, including 210 (57%) according to both algorithms, 4 (1%) to AGLA only and 64 (17%) to ESC only.
AGLA and ESC algorithms presented similar accuracy to predict atherosclerotic cardiovascular disease in Switzerland. A quarter of adults developing atherosclerotic cardiovascular disease were not identified by preventive algorithms to be eligible for statin therapy.