Performance of different definitions of metabolic syndrome for children and adolescents in a 6-year follow-up: Tehran Lipid and Glucose Study (TLGS).

AIM

To assess the utility of different definitions of the metabolic syndrome (MetS) in predicting adulthood MetS among Tehranian adolescents in a 6-year follow-up.

METHODS

A random sample of 2645 adolescents was selected from the Tehran Lipid and Glucose Study (TLGS) in 1999-2001; MetS was assessed based on five definitions for adolescents. In 2005-2007, 572 age 18 years and over were assessed for MetS by NCEP ATP III adults' criteria. Sensitivity, specificity, and area under receiver operating curve for attaining adulthood MetS, obesity and overweight were calculated for each MetS definitions.

RESULTS

The prevalence of adolescence MetS varied from 0.7+/-0.2 to 15.1+/-0.8% by different definitions in adolescents and 4.0+/-0.9% among adults. The highest area under curve for prediction of adulthood MetS pertained to the de Ferranti's definition (0.723), and for prediction of adulthood obesity and overweight were 0.723 and 0.606, respectively, for this definition. Kappa for agreement between these definitions was fair (0.195).

CONCLUSIONS

Definition of MetS for adolescents showed a fair agreement with adult definition. Yet, a diagnosis of MetS in adolescents seems to be a good predictor of developing adulthood MetS within a short time.