Validation of new prediction equations using skin-fold thickness in estimating body fat percentage in children and adolescents

Abstract

Objective: Prediction equations can be used as a useful, inexpensive, and highly accurate tool for estimating Percent Body Fat (PBF) in many countries all over the world. We conducted this study to develop new predictive equations for estimating PBF in children and adolescents aged 6 to 18 years in Ho Chi Minh City (HCMC), Vietnam. Methods: This cross-sectional study included 149 students in HCMC. The new PBF prediction equations based on skinfold thickness and other anthropometric factors were evaluated using a multivariate linear regression model. The agreement between PBF estimated by equations and values derived from DXA measurements was assessed using the Bland-Altman method and Lin’s concordance correlation coefficient (CCC). The performance of the new equations for the classification of nutritional status was studied using the ROC Curve and indicators such as AUROC, accuracy, sensitivity, specificity, and Youden-Index. Results: Weight, height, and the sum of three skinfold thicknesses were added into the new equation for males (CCC = 0.87; Radj 2 = 0.76; RMSE = 4.06). Age, puberty status, the sum of three skinfold thickness measurements, and the sum square of subscapular and calf skinfold thicknesses were incorporated in the equations for females (CCC = 0.78; Radj 2 = 0.62; RMSE = 3.84). In comparison to BMI and WHtR, our equations proved to be more accurate in identifying individuals with an excess of body fat. Conclusions: New prediction equations based on studied participants’ skinfold thickness and anthropometric factors could produce predicted PBF estimates that are more consistent and accurate than other computed techniques utilizing DXA-derived PBF. DOI: 10.59715/pntjmp.3.2.15