Diagnostic tests for adiposity and metabolic risk factors in adolescence : results from the Stockholm weight development study (SWEDES)

Background: Despite growing concern about the development of pediatric obesity, there is still controversy about its classification. International classification systems have been proposed and national systems are in use in many countries, often in parallel. Most proposed classification systems are distribution-based and not anchored to either fatness, metabolic risk or risk of adult obesity.

Aims: To assess the diagnostic accuracy of BMI, WC and W-ER for fatness, the characteristics of commonly used classification systems, and to estimate the relationship between diagnostic tests for overweight, insulin resistance and cholesterol profile.

Design: SWEDES is a prospective cohort study of 481 children followed from birth until age 17y, and their mothers. BMIdevelopment was assessed retrospectively from healthcare records, and clinical measurements at age 17y were performed. Measurements included height, weight, WC, HC, body composition (FM, FFM), and blood samples. Proposed classification systems for childhood and adolescent obesity were evaluated by use of ROC analyses. Through regression analyses, the association between different screening measures and insulin resistance (HOMA-IR) was also investigated.

Results: The IOTF/Cole classification system was found to be highly specific (0.95-1.00), but insensitive for fatness, especially in females (0.22-0.25). The outcome was shown to be fairly insensitive to the choice of gold standard to define fatness. The same characteristics remained in longitudinal analyses when using BMI at different ages as diagnostic tests for both BMI-based and FMIbased overweight in late adolescence. Using alternative measures as indicators of total fatness revealed that WC and BMI display similar diagnostic accuracy, while WHR performs significantly worse, also when evaluated against blood variables. In boys, the associations between simple anthropometric measures and insulin resistance were weaker than for detailed body composition measures, while in girls no significant associations were observed. The associations between BMI in childhood and adolescence and blood variables were non-existent or weak, implying that anchorage of classification against such outcomes may be complicated.

Conclusion: Currently proposed BMI-based classification systems for adolescent overweight are highly specific, but less sensitive for fatness as well as for metabolic risk. The awareness about the trade-offs involved in applying such classification systems should be raised, since it may affect efficiency in the usage of healthcare and public health resources. A classification system designed for international monitoring cannot be optimal for the different needs of risk group stratification in such different settings as specialized care, public health and targeted prevention. Misuse is likely to result in wasted resources as well as foregone health improvements.

List of scientific papers

I. Neovius MG, Linne YM, Barkeling BS, Rossner SO (2004). "Sensitivity and specificity of classification systems for fatness in adolescents. " Am J Clin Nutr 80(3): 597-603
https://pubmed.ncbi.nlm.nih.gov/15321798

II. Neovius M, Linne Y, Rossner S (2005). "BMI, waist-circumference and waist-hip-ratio as diagnostic tests for fatness in adolescents. " Int J Obes Relat Metab Disord 29(2): 163-9
https://pubmed.ncbi.nlm.nih.gov/15570312

III. Neovius M, Linne Y, Vagstrand K, Ekelund U, Poortvliet E, Sjostrom M, Hoffstedt J, Rossner S (2005). "Insulin resistance, anthropometry and body composition in adloescents and their mothers." (Submitted)

IV. Neovius M, Linne Y, Ekelund U, Poortvliet E, Sjostrom M, Hoffstedt J, Rossner S (2005). "BMI-development and its relation to fat mass, insulin resistance and cholesterol profile in late adolescence." (Submitted)