Validity of body composition assessment methods for older men with cardiac disease.

Abstract

This study was designed to determine which of several body composition assessment methods was most accurate for patients with cardiac disease for the purpose of outcome measurement. Six body composition assessment methods were administered to each of 24 men with cardiac disease. Methods included circumference measurement, skinfold measurement, near-infrared interactance via the Futrex-5000, bioelectrical impedance via the BioAnalogics ElectroLipoGraph and Tanita TBF-150, and hydrostatic weighing, the criterion measure. A repeated measures analysis of variance indicated no significant (P > .05) difference between circumference and skinfold measurements compared to hydrostatic weighing. Near-infrared interactance presented the best standard error of estimates (3.5%) and the best correlation (r = .84) with hydrostatic weighing; however, the constant error was 3.76%. Bioelectrical impedance measured by the ElectroLipoGraph and TBF-150 instruments significantly underestimated percent body fat by 8.81% and 4.8%, respectively. In this study of middle-aged to older men with cardiac disease, the best method for determining body fat was circumferences. This technique was accurate, easy to administer, inexpensive, and had a lower error potential than the other techniques. Skinfold measurements were also closely related to hydrostatic weighing, but should be performed only by experienced practitioners because there is a greater potential for tester error in certain patients. In the future, near-infrared interactance measurements may be a viable technique for body composition assessment in patients with cardiac disease. However, algorithms specific to the population of patients with cardiac disease being tested must be developed before this technique can be routinely recommended for body composition assessment. Bioelectrical impedance assessment by either method is not recommended for patients with cardiac disease, as it consistently underestimated percent body fat when compared to hydrostatic weighing in this population.