Abstract 1004: Optimal Use of Framingham Risk Scores to Identify Individuals for Intensive Medical Risk Factor Modification
BACKGROUND: There is a strong positive association between Framingham Risk Scores (FRS) in a population and incidence of hard coronary heart disease (hCHD) events. Under current guidelines, individuals with FRS that indicate >= 20% 10-year risk of hCHD are recommended to receive intensive medical risk factor modification. We sought to assess the performance of FRS as a predictive tool when used as in current guidelines.
DESIGN: Retrospective analysis of a prospective cohort study.
SETTING: The Atherosclerosis Risk in Communities (ARIC) study, 1987–2001.
PARTICIPANTS: Data from 13,007 White and African American (24%) men and women (56.5%), aged 45– 64 years, without known CHD or diabetes at baseline, with available FRS variables.
MEASUREMENTS: The FRS was computed according to standard algorithm. Risk was stratified into 3 FRS categories: low (0 –5%), intermediate (6 –19%), and high (>=20%). The main outcome was hCHD event defined as MI or coronary death. Using Receiver operating characteristics curves, sensitivity, specificity, and Youden’s Index were computed to determine predictive accuracy of FRS at various thresholds.
RESULTS: At baseline, 7,758 individuals (59.6%) were at low FRS, 4,407 (33.9%) at intermediate, and 842 (6.5%) at high FRS. During 10-year follow-up, 792 hCHD events occurred. FRS was significantly associated with the outcome with a C statistics of 0.72 (p-value <0.0001). However, at standard “high risk” cut-off (>=20%), the sensitivity of FRS was only 20% and Youden’s Index 0.15; 80% of hCHD events occurred in individuals without “high risk” FRS. Lowering the threshold to >=15% resulted in sensitivity of 32%, and a cut-off of 10% had 53% sensitivity. At cut-off >=6% sensitivity improved to 77%, with a specificity of 55%. The Youden’s index improved to 0.33 indicating greater accuracy at this cut point.
CONCLUSION: The FRS is highly correlated statistically with subsequent hCHD events. However, when used dichotomously as in current guidelines, it is insensitive and only 20% of those who will experience hCHD events would be identified a priori for intensive risk factor modification. Given the low morbidity of treatment and the high morbidity and mortality of events, the lower threshold would seem to be preferable to that in current guidelines.