• The validity of two compartment model methods of body composition as compared to magnetic resonance imaging in Asian Indian versus Caucasian males

      Davies, Ben Rhys (University of Bedfordshire, 2010-11)
      Background: The two-compartment (2C) model is a relatively accessible, inexpensive and time efficient method for body composition measurement. There is very little validated research on the 2C model in Asian Indians: a high risk population in terms of obesity and related disorders. This highlights the need for valid estimates of body composition from the 2C model. Purpose: The goal was to compare 2C model (predictor) estimates of body composition to those from magnetic resonance imaging (MRI) (criterion), an established gold standard measure of total adiposity in order to determine the validity of the 2C model in the Asian Indian population. From this data it is hoped that a correction equation may be determined for more accurate prediction of Asian Indian body composition using 2C model methods. Methods: 21 males (10 Asian Indian and 11 Caucasian, aged 18-55 yrs) had estimates of percent body fat from 2C methods (sum of four skinfolds and anthropometry, bioelectrical impedance analysis [Bodystat 1500 and Tanita segmental impedance analyser], air displacement plethysmography [Bod Pod] and hydrostatic weighing) compared to MRI measured body composition values. Agreement was assessed using multiple linear regression analysis and Bland-Altman plots. Differences were assessed using repeated measures analysis of variance. Results: Regression analysis showed air displacement plethysmography predicts MRI body composition in Caucasian males (adjusted r2 = 0.74; SEE =3.27 ). In Asian Indians, tricep skinfold thickness and hydrostatic weighing predicted MRI body composition with a low prediction error (adjusted r2 = 0.90; SEE =1.75). Despite strong correlations and no significant difference between mean differences of the 2C methods, used in the prediction model, and MRI, BlandAltman plots revealed no acceptable limits of agreement between the methods. Asian Indian body composition was underestimated by all two compartment devices compared to MRI. Conclusion: There appears to be potential for the use of tricep skinfold thickness and hydrostatic weighing to predict an established reference measure (MRI) in the high risk Asian Indian population. The 2C model underestimated Asian Indian body composition, this suggests that un-validated, the 2C model may misidentify obesity and in turn health risk. However the small sample tested, has implications for the interpretation of the findings. Further investigation is required with a greater sample size to validate the 2C model against an established reference measure such as MRI as there is currently little published validation data in this ethnic group.