Comparison of Body Composition Assessment Techniques in Older Adults

Body composition is an important measurement for health assessments in older adults. The purpose of this study was to evaluate percent body fat (%BF) estimations in older adults by the Tanita, a two-compartment model (Siri 2-C) and dual energy x-ray absorptiometry (DXA) compared to a three-compartment model (Lohman 3-C). Fifty-two females and fifty men between the ages of 54 and 75y volunteered for the study. The estimate of % BF from the Tanita was measured using the ‘adult’ mode. The Bod Pod was used to determine body density (D_b) for the 2-C and 3-C. DXA was used to obtain an estimate of %BF, and determine BMC for use in the 3-C. Compared to the 3-C estimate of %BF (26.1 ± 1.0%), %BF in males was significantly underestimated by Tanita (22.1 ± 0.8%), and overestimated by 2-C (27.6 ± 09%) and DXA (28.6 ± 0.9%) (P< 0.001). The bias and limits of agreement for all methods in males were: Tanita, -3.93 ± 10.3; 2-C, 1.53 ± 3.38; DXA, 2.51 ± 6.36. Compared to the 3-C estimate of %BF (39.2%), %BF in females was significantly underestimated by Tanita (36.4 ± 1.2%; P < 0.001) and overestimated by DXA (40.5 ± 1.2%; P = 0.013). There was no significant difference in females between the 3-C and 2-C estimates of %BF (39.2 ± 1.2% vs. 39.3 ± 1.2% respectively, P > 0.05). The bias and limits of agreement for all methods in females were: Tanita, -2.85 ± 8.66; 2-C, 0.05 ± 3.24; DXA 1.27 ± 5.64. This study suggests that the techniques used in this investigation are not interchangeable when estimating %BF in older adults.