BACKGROUND & AIMS: Severe obesity imposes physical limitations to body composition assessment. Our aim was to compare body fat (BF) estimations of severely obese patients obtained by bioelectrical impedance (BIA) and air displacement plethysmography (ADP) for development of new equations for BF prediction. METHODS: Severely obese subjects (83 female/36 male, mean age=41.6+/-11.6 years) had BF estimated by BIA and ADP. The agreement of the data was evaluated using Bland-Altman's graphic and concordance correlation coefficient (CCC). A multivariate regression analysis was performed to develop and validate new predictive equations. RESULTS: BF estimations from BIA (64.8+/-15 kg) and ADP (65.6+/-16.4 kg) did not differ (p>0.05, with good accuracy, precision, and CCC), but the Bland- Altman graphic showed a wide limit of agreement (-10.4; 8.8). The standard BIA equation overestimated BF in women (-1.3 kg) and underestimated BF in men (5.6 kg; p<0.05). Two BF new predictive equations were generated after BIA measurement, which predicted BF with higher accuracy, precision, CCC, and limits of agreement than the standard BIA equation. CONCLUSIONS: Standard BIA equations were inadequate for estimating BF in severely obese patients. Equations developed especially for this population provide more accurate BF assessment.
BACKGROUND & AIMS: Severe obesity imposes physical limitations to body composition assessment. Our aim was to compare body fat (BF) estimations of severely obesepatients obtained by bioelectrical impedance (BIA) and air displacement plethysmography (ADP) for development of new equations for BF prediction. METHODS: Severely obese subjects (83 female/36 male, mean age=41.6+/-11.6 years) had BF estimated by BIA and ADP. The agreement of the data was evaluated using Bland-Altman's graphic and concordance correlation coefficient (CCC). A multivariate regression analysis was performed to develop and validate new predictive equations. RESULTS: BF estimations from BIA (64.8+/-15 kg) and ADP (65.6+/-16.4 kg) did not differ (p>0.05, with good accuracy, precision, and CCC), but the Bland- Altman graphic showed a wide limit of agreement (-10.4; 8.8). The standard BIA equation overestimated BF in women (-1.3 kg) and underestimated BF in men (5.6 kg; p<0.05). Two BF new predictive equations were generated after BIA measurement, which predicted BF with higher accuracy, precision, CCC, and limits of agreement than the standard BIA equation. CONCLUSIONS: Standard BIA equations were inadequate for estimating BF in severely obesepatients. Equations developed especially for this population provide more accurate BF assessment.
Authors: Mônica Fernandez; Rosana Farah Toimil; Zied Rasslan; Elias Jirjoss Ilias; Ana Lúcia Torloni Gradinar; Carlos Alberto Malheiros Journal: Arq Bras Cir Dig Date: 2016