Performance of the Framingham coronary heart disease risk score for predicting 10-year cardiac risk in adult United Arab Emirates nationals without diabetes: a retrospective cohort study

Abstract

BackgroundPrimary prevention guidelines recommend the use of the Framingham risk score (FRS) to estimate the 10-year coronary heart disease (CHD) risk in patients without diabetes for statin eligibility. However, the FRS model has never been validated in an Arab population. Therefore, this study aimed to examine the clinical performance of the FRS model for predicting 10-year CHD risk in adult United Arab Emirates (UAE) nationals without diabetes.MethodsThis 10-year retrospective cohort study included patients from the primary care clinics and outpatient specialty departments of a large tertiary care hospital in Al-Ain, UAE. They were aged 30–79 without a baseline history of cardiovascular disease and diabetes. The FRS for each subject was calculated. Follow-up data on hard CHD (hCHD) events (myocardial infarction or coronary death) for each participant were collected from the baseline visit in 2008 until December 31, 2019. The area under the time-dependent receiver operating characteristic (ROC) curve (AUROC) was used to assess the FRS model discrimination. Calibration was measured by using the Hosmer-Lemeshow χ2 test and the calibration curve. The optimal cutoff-point for hCHD risk prediction was determined by ROC curve analysis.ResultsA total of 554 participants were included. The mean age was 48.0 ± 12.8 years and 45% were men. The mean predicted FRS of the study cohort was 5.2% and approximately 7% were classified as high-risk (≥ 20% threshold) by the FRS model. During a median follow-up of 10.2 years (interquartile range, 7.8–11.0 years), 26 hCHD events occurred. The FRS model displayed reasonably good discrimination (time-dependent AUROC value: 0.83) and calibration in predicting hCHD (Hosmer-Lemeshow χ2 statistic 11.2, P = 0.191). Applying the 20% high-risk threshold, the FRS model had a sensitivity of only 37% in identifying patients at high-risk for an hCHD event over 10 years. While a 7.5% optimal cutoff-point improved the sensitivity to 74%.ConclusionsThe FRS can be used in the prediction of coronary risk among UAE nationals without diabetes, however, the recommended hCHD risk threshold for statin eligibility may be too high. Lowering the cutoff-point to 7.5% could improve the identification of patients for preventive treatment.