Estimating Cardiometabolic Risk Using Existing Scoring Systems

Abstract

While useful for prediction of cardiovascular disease (CVD), risk estimators such as the Framingham Risk Score (FRS) and the Atherosclerotic CVD (ASCVD) calculator have not been evaluated for prediction of incident type 2 diabetes (T2D) and progression to CVD, or for how changes in scores over time predict future risk. We evaluated these scores alongside a metabolic syndrome (MetS) severity Z-score (http://mets.health-outcomes-policy.ufl.edu), using data from 8,026 participants in the Atherosclerosis Risk in Communities (ARIC) study without T2D or CVD by Visit 2. We characterized individuals by incident T2D by V4 (mean years from V1 = 8.9), and whether they had an adjudicated CVD event by V4 ("early" CVD) or a "late" CVD event (after V4; mean total follow-up = 20.7 years). For each disease category, we estimated mean scores at V1 and changes in score, V2-V1 (Figure 1). MetS severity distinguished those developing T2D vs. not and future CVD vs. not. FRS and ASCVD distinguished CVD development in each incident T2D group, but trends were not as discernible across T2D status. We then standardized FRS and ASCVD scores to allow for comparisons of OR’s, using logistic regression to model odds of each disease category by each score, including V1 and the (V2-V1) change (OR’s in Figure 1). While CVD scores predict future CVD, MetS severity, based both on baseline levels and change in score, better predicts individuals who go on to develop T2D and further progress to CVD. Disclosure M.J. Gurka: None. S.L. Filipp: None. T.A. Pearson: Consultant; Self; Bayer AG. M.D. DeBoer: None.