OBJECTIVES: 1) Examine associations between visceral adipose tissue (VAT), subcutaneous abdominal adipose tissue (SAAT), and anthropometric and demographic variables; 2) generate and cross-validate prediction equations for estimating VAT and SAAT in overweight Latino children. STUDY DESIGN: Cross-sectional. PARTICIPANTS: 196 overweight 8-13-year-old Latino youth. Two-thirds (n = 131) were randomly assigned to a development group to generate prediction equations for VAT and SAAT; one-third (n = 65) was used as a cross-validation group. METHODS AND PROCEDURES: Anthropometric measurements (height, weight, skinfold thicknesses, and circumferences) were performed. VAT and SAAT were measured using magnetic resonance imaging (MRI). RESULTS: The strongest univariate correlate for VAT was waist circumference (WC) (r = 0.65, p < 0.01) while the strongest correlate for SAAT was hip circumference (r = 0.88, p < 0.001). Regression analyses showed approximately 50% of the variance in VAT was explained by WC (43.8%), Tanner stage (4.2%) and calf skinfold (1.7%). Variance in the SAAT model was explained by WC (77.8%), triceps skinfold (4.2%) and gender (2.3%). Residual analyses showed no bias in either equation. Though mean differences between measured and predicted VAT and SAAT were small, there was a large degree of variability at the individual level especially for VAT. CONCLUSIONS: Both VAT and SAAT prediction equations performed well at the group level, but the relatively high degree of variability suggests limited clinical utility of the VAT equation. MRI is currently required to derive an accurate measure of VAT at the individual level.
RCT Entities:
OBJECTIVES: 1) Examine associations between visceral adipose tissue (VAT), subcutaneous abdominal adipose tissue (SAAT), and anthropometric and demographic variables; 2) generate and cross-validate prediction equations for estimating VAT and SAAT in overweight Latino children. STUDY DESIGN: Cross-sectional. PARTICIPANTS: 196 overweight 8-13-year-old Latino youth. Two-thirds (n = 131) were randomly assigned to a development group to generate prediction equations for VAT and SAAT; one-third (n = 65) was used as a cross-validation group. METHODS AND PROCEDURES: Anthropometric measurements (height, weight, skinfold thicknesses, and circumferences) were performed. VAT and SAAT were measured using magnetic resonance imaging (MRI). RESULTS: The strongest univariate correlate for VAT was waist circumference (WC) (r = 0.65, p < 0.01) while the strongest correlate for SAAT was hip circumference (r = 0.88, p < 0.001). Regression analyses showed approximately 50% of the variance in VAT was explained by WC (43.8%), Tanner stage (4.2%) and calf skinfold (1.7%). Variance in the SAAT model was explained by WC (77.8%), triceps skinfold (4.2%) and gender (2.3%). Residual analyses showed no bias in either equation. Though mean differences between measured and predicted VAT and SAAT were small, there was a large degree of variability at the individual level especially for VAT. CONCLUSIONS: Both VAT and SAAT prediction equations performed well at the group level, but the relatively high degree of variability suggests limited clinical utility of the VAT equation. MRI is currently required to derive an accurate measure of VAT at the individual level.
Authors: Tiago V Barreira; Stephanie T Broyles; Alok K Gupta; Peter T Katzmarzyk Journal: Obesity (Silver Spring) Date: 2014-02-19 Impact factor: 5.002