Published in:
Open Access 01-12-2018 | Research
Assessing the added predictive ability of a metabolic syndrome severity score in predicting incident cardiovascular disease and type 2 diabetes: the Atherosclerosis Risk in Communities Study and Jackson Heart Study
Published in: Diabetology & Metabolic Syndrome | Issue 1/2018
Login to get accessAbstract
Background
The severity of the metabolic syndrome (MetS) predicts future coronary heart disease (CHD) and diabetes independent of the individual MetS components. Our aim was to evaluate whether MetS severity conferred additional discrimination to existing scoring systems for cardiovascular disease (CVD) and diabetes risk.
Methods
We assessed Cox proportional hazard models of CHD- and diabetes risk among 13,141 participants of the Atherosclerosis Risk in Communities Study and the Jackson Heart Study, using the Framingham Risk Calculator, the American Heart Association’s Atherosclerotic CVD calculator, the American Diabetes Association diabetes risk score and an additional diabetes risk score derived from ARIC data. We then added a MetS-severity Z-score to these models and assessed for added risk discrimination by assessing Akaike information criterion, c-statistic, integrated discrimination improvement (IDI) and continuous net reclassification improvement (NRI).
Results
The MetS severity score appears to add to the predictive ability of individual CHD and diabetes risk scores. Using the IDI, MetS improved risk prediction for diabetes but not CHD risk. In all 4 scoring systems, MetS severity had a significant non-event NRI, improving the ability to exclude individuals without events. Assessing interactions between risk scores and MetS severity revealed that MetS severity was more highly associated with disease risk among those in the lowest quintiles of risk score, suggesting that MetS was particularly able to identify risk among individuals judged to be of low risk by existing algorithms.
Conclusions
Mets severity improved prediction of diabetes more so than CHD. Incorporation of multiple risk predictors into electronic health records may help in better identifying those at high disease risk, who can then be placed earlier on preventative therapy.