Background The association between vitamin D (VitD) and cardiometabolic health has been described however the ability of VitD to predict metabolic syndrome (MetSyn) beyond individual demographics and common lifestyle factors is unclear.
Methods 1799 adults from the Canadian Health Measures Survey, a nationally representative survey were examined. MetSyn was defined based on the National Cholesterol Education Program's Adult Treatment Panel III criteria. The comparative value of smoking status, energy expenditure from physical activity and plasma 25 (OH)D VitD to predict the presence of metabolic syndrome was evaluated using receiver operating characteristic curves (ROC) from fully adjusted logistic regression models (age, sex, ethnicity, education, annual fruit intake and month of interview). Somer's D and c statistics were used to compare across models. Sampling weights were applied to all models.
Results Inclusion of plasma VitD, energy expenditure and smoking status all lead to increases in the c statistic and Somer's D suggesting increased predictive ability compared to the baseline model. ROC contrasts, however, showed that only inclusion of VitD significantly increased the predictive ability of the model (p=0.001).
Discussion VitD may be a useful predictive tool for cardiometabolic risk prediction in addition to conventional factors. Prospective analyses may provide more insight into VitD as a potential population-based intervention target for reduction of chronic disease burden related to MetSyn.