A new metabolic syndrome prediction model for self-evaluation as a primary screening tool in an apparently MetS-free population

Background: Metabolic syndrome (MetS) is a growing global public health concern associated with increased morbidity and mortality. The study aimed to establish a simple self-evaluated prediction model to identify MetS.

Methods: A cross-sectional study based on the American National Health and Nutrition Examination Survey database was performed. Participants aged ≥20 in the 2009-2018 surveys, with no current pregnancy or major morbidities, were included. The model was built with data from 2009 to 2016 and validated using 2017-2018 data. MetS was defined according to AHA/NHLBI guidelines. Multivariable logistic regression was applied to build a prediction model. The area under the receiver operating characteristic curve (AUC) was used to assess the discrimination ability and the maximal Youden's index was used to identify the optimal cut-off value.

Results: The study included 4245 individuals (median age 37, IQR 28-49, 51.8% females) in the training group and 911 individuals (median age 37, IQR 28-52, 52.5% females) in the validation group. Older age, male gender, non-Black race, no postsecondary education, and higher BMI were significantly associated with increased risk of MetS. The final model included age, gender, race, education, and BMI, and showed good discrimination ability (AUC = 0.810, 95% CI 0.793-0.827, sensitivity 80.4%, specificity 66.2%, positive likelihood ratio 2.379, negative likelihood ratio 0.296, PPV 59.6% and NPV 84.5%).

Conclusion: A new model for self-evaluation may serve as a primary, easy-to-use screening tool to identify MetS in an apparently MetS-free population. A simple application may serve for primary and secondary prevention, thus enabling risk reduction in the development of cardiovascular morbidity and health expenditure.