Cardiovascular disease risk score derivation and validation in Abu Dhabi United Arab Emirates.Retrospective Cohort Study

AbstractCardiovascular disease (CVD) risk assessment is key to rational decision-making in primary prevention. The CVD risk depends on dynamic factors requiring continuous equation updates. DesignThe Abu Dhabi Risk Study (ADRS) is the first and longest-duration retrospective cohort study in Abu Dhabi and the United Arab Emirates (UAE), aiming to develop 10-year risk prediction equations for CAD, stroke, and ASCVD (Atherosclerotic Cardiovascular Disease, combining CAD and stroke) and validate international risk equations. MethodThe included 8699 subjects are participants of the national cardiovascular screening program of 2011-2013 with an average follow-up of 9.2 years. They were assessed retrospectively in 2023 for health outcomes. The validation cohort, 2554 subjects, is another community-based screening program done in Abu Dhabi in the period from 2016 and 2017. With an average follow-up of 6.67 years. ResultsOf 8504 who were ASCVD-free, 250 had new coronary artery events. Identified risk factors for ASCVD in this population were the conventional risk factors such as age, gender, smoking, high cholesterol/HDL ratio, and diabetes diagnosis, in addition to low vitamin D level, and low glomerular filtration rate (GFR) levels. Three ADRS prediction models were derived from Cox regression. The ADRS-CAD had a C- statistic of 0.899 compared to 0.828 of FRS (Framingham score) in the same population. ADRS-Stroke had a c-statistic of 0.904. The ADRS-ASCVD had a c-statistic of 0.898 compared to 0.891 of PCE (pooled cohort equations) and 0.825 of FRS-CVD. Applying the developed formulas to the validation cohort showed good predictability of CAD and ASCVD events with an ASCVD c-statistic of 0.825, for CAD the c-statistic was 0.799, and for stroke, it was 0.761. The PCE showed similar performance in this cohort with a c-statistic for ASCVD of 0.824. ConclusionThis study demonstrated the value of tailoring risk assessments to local populations and healthcare contexts.