Assessment of body composition in Sri Lankan children: validation of a skin fold thickness equation

• Published in: Ceylon medical journal Vol. 53; no. 3; pp. 83 - 88
• Main Authors: Wickramasinghe, V P, Lamabadusuriya, S P, Cleghorn, G J, Davies, P S W
• Format: Journal Article
• Language: English
• Published: Sri Lanka 01.09.2008
• Subjects: Adolescent; Body Composition; Body Fat Distribution - methods; Child; Child, Preschool; Cross-Sectional Studies; Female; Humans; Male; Morbidity; Obesity - diagnosis; Obesity - epidemiology; Predictive Value of Tests; Regression Analysis; Skinfold Thickness; Sri Lanka - epidemiology; Sri Lanka
• ISSN: 0009-0875; 2386-1274
• DOI: 10.4038/cmj.v53i3.247

Morbidity associated with obesity is related to the fat mass (FM) of the body. The direct estimation of FM is difficult. Skin fold thickness (SFT) is a simple and cheap alternative for accurate assessment of FM, and population specific equations are necessary for accurate assessment of FM using SFT. To develop a SFT prediction equation to estimate FM of Sri Lankan children. DESIGN, SETTING AND METHOD: Across-sectional descriptive study was done at the University Paediatric Unit of Lady Ridgeway Hospital, Colombo. Data were collected from 5 to 15 year old healthy children. Triceps, biceps, supra-iliac and subscapular SFT were measured using Harpendens skin fold caliper. Total body water was assessed using an isotope dilution method (D2O), and fat free mass calculated. FM was assessed based on 2 compartment body composition model. Multiple regression analysis was used to develop prediction equation and validated using PRESS (prediction of sum of squares) statistical technique. Independent variables were age, triceps SFT, subscapular SFT and sex. Prediction equation for FM [(0.68 x age) + (0.246 x triceps SFT) + (0.383 x subscapular SFT) - (1.61 x sex code) -3.45] was able to predict 76.4% of variance with a root mean squared error (RMSE) of 3.4 kg. PRESS statistics was 3.4 kg with press residuals of 1.56 kg. Bland-Altman technique showed that the majority of the residuals were within mean bias +/-1.96 SD. Results of this study provide an SFT equation for the prediction of FM in Sri Lankan children.