Evaluation of selected field and laboratory measures of body composition

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
University of Alabama Libraries

The four-compartment (4C) model is a laboratory method that is a valid method for the assessment of body composition. Three experiments were performed to 1) determine the validity of selected bioimpedance equations with the 4C model, 2) determine the impact of predicted vs. simultaneous residual lung volume (RLV) during underwater weighing (UWW) on the 4C model, and 3) develop a new body fat prediction equation. In the first study, subjects had body fat percentage (BF%) and fat-free mass (FFM) predicted from four bioelectrical impedance equations and compared to the 4C model. Three equations produced a significant mean difference, while another was non-significant. However, all four equations had a small standard error of the estimate (SEE) and fairly narrow limits of agreement. In a second study, RLV was measured simultaneously and predicted when determining UWW and 4C model BF%. The mean differences for UWW BF% was significant when comparing predicted vs. simultaneous RLV, but non-significant when the body density values derived via UWW were incorporated in the 4C model. The error was lower when using RLV prediction equations for determining BF% via the 4C model than UWW. In a third study the variables sex, body mass index (BMI), sum of combined handgrip strength, and vigorous physical activity was utilized in a regression equation to predict 4C model BF%. The new BF% equation, previous BMI-based BF% equations and skinfolds were compared to the 4C model. The new equation and BMI-based equations had significantly different BF% values and provided large 95% limits of agreement. Similarly, skinfolds had a significant mean difference, but the SEE was 3.7%, leading to the recommendation of skinfolds over the new equation and BMI-based equations. In conclusion, BIA equations can be used in the field, but practitioners should consider the tendency of the equations to over-predict BF% and under-predict FFM. Furthermore, the prediction of RLV can be used for the UWW procedure when determining 4C model BF%, but should not be used for UWW BF% alone. Lastly, the new BF% equation and BMI-based equations did not compare favorably with 4C model and the use of skinfolds is recommended.

Electronic Thesis or Dissertation