This file was dowloaded from the institutional repository Brage NIH -brage.bibsys.no/nih McMurray, R. G., Andersen, L. B. (2010). The influence of exercise on metabolic syndrome in youth : a review. American Journal of Lifestyle Medicine, 4(2), 176-186.Dette er siste tekst-versjon av artikkelen, og den kan inneholde ubetydelige forskjeller fra forlagets pdf-versjon. Forlagets pdf-versjon finner du på sagepub.com: http://dx.doi.org/10.1177/1559827609351234This is the final text version of the article, and it may contain insignificant differences from the journal's pdf version. used growth curve modeling to develop age and sex-specific criteria for MetS in youth based on the accepted criteria for adults. This approach is logical because the risk factors have a propensity to track from childhood through adulthood and the adult risk factor cut-points are clearly associated with the development of cardiovascular disease. 13 The most biologically meaningful approach may be to select the children where risk factors were not independently distributed, as done by Andersen et al. in 2003. 14 Another approach used in publication from the European Youth Heart Study (EYHS) is to develop an age and sex-specific metabolic risk score based on z-scores for each of the MetS characteristics. 15--17 The authors argue that since there is no accepted definition of MetS in children, the z-score provide a continuous score which may be more appropriate for investigating associations. Dichotomization of each variable causes reduction in the information and therefore the diagnostic value. One system used a standardized score developed as a sum each of seven criteria (insulin, glucose, triglyceride, HDL-cholesterol, skinfolds, diastolic and systolic blood pressure). The standardized score used the subjects value minus sample mean, divided by standard deviation to produce a continuous overall MetS risk score. 17 The lower the standard score the more favorable the overall profile. However, this score has the problem because it lacks a cut-point above which to identify MetS. An identifying cut-point was found through another analysis where upper quartile in each risk factor was defined to be at risk, and the number of risk factors was then summed in each child. By using this approach the authors were able to calculate when risk factors were not independently distributed. An independent distribution of the risk factors would follow a binominal distribution, and the authors calculated for each number of risk factors if there were more children than expected from the binominal distribution. This approach selected between 10-15% of the children to have clustered risk, and the cut point in the continuous z-score could now be defined. 19 Since this approach was studyspecific comparisons with other studies could be difficult. However, data were available from three different countries, and it turned out that almost the same z-score defined clustered risk in the different populations, which may indicate that cut points may not be population ...