Accuracy Of Bioimpendance Analysis For Measuring Body Compositionand Resting Metabolism In Youth.

PURPOSE: This study assessed the accuracy of bioimpedance analysis (BIA) for measuring body composition and resting metabolic rate (RMR) against criterion measures in youth. METHODS: Youth of varying sex, age (child, 7-12 y; teen, 13-17 y), weight status (healthy weight, BMI < 85th %ile; overweight, BMI ≥ 85th %ile; obese, BMI ≥ 95th %ile), and pubertal status (years from peak height velocity) were recruited from Athens and surrounding counties of Southeast Ohio. Using paired t-tests, fat mass (FM; kg), fat free mass (FFM; kg), visceral adipose tissue (VAT; cm2) and RMR (kcal/day) measures from a multi-frequency, tetrapolar, eight-point BIA device were compared to dual-energy X-ray absorptiometry (DXA) and a portable metabolic analyzer, respectively. Relationships among all measures were assessed using correlation analyses. Accuracy was calculated as the root mean squared error (RMSE) and bias between methods. RESULTS: In total, 130 youth (11.9 ± 2.9 y; 68 female, 85 healthy weight, 77 children, 61 pre-pubertal) participated in the study. BIA measures of FM was greater in girls (16.0 ± 1.2 kg), and obese (27.6 ± 2.5 kg) compared to boys (13.2 ± 0.9 kg) and healthy weight youth (10.9 ± 0.4 kg). FFM was higher in boys (41.8 ± 1.8 kg), teens (49.2 ± 1.4 kg), and post-pubertal youth (52.0 ± 9.6 kg) compared to their counterparts (35.3 ± 1.2, 30.9 ± 0.9, and 28.5 ± 5.4 kg, respectively). BIA-estimated RMR was higher in boys (1276 ± 40 kcal/day), teens (1434 ± 29 kcal/day), and post-pubertal youth (1493 ± 207 kcal/day) compared to their counterparts (1133 ± 25, 1038 ± 19, 985 ± 116 kcal/day, respectively). Assessing the entire sample, BIA reported lower FFM (38.4 ± 1.1 kg) and higher FM (11.8 ± 0.8 kg) and VAT (50.4 ± 4.1 cm2) compared to DXA (32.6 ± 1.0 kg, 14.7 ± 0.8 kg, 45.8 ± 2.7 cm2), whereas estimated and measured RMR values were similar (1199 ± 24 vs. 1145 ± 254 kcal/day; p > 0.05). Correlations between methods ranged from r = 0.71 to 1.00 (p < 0.05). BIA overestimated FM by 2.8 kg and underestimated FFM by 5.7 kg. CONCLUSIONS: While measures between methods were statistically different, the magnitude of these differences were physiologically insignificant and strong relationships between BIA and criterion measures were observed. Future research should assess the sensitivity of BIA to body composition and RMR changes across growth trajectories.