Validity of leg-to-leg bioelectrical impedance analysis to estimate body fat in obesity.

BACKGROUND: Bioelectrical impedance analysis (BIA) is a safe and easy method of assessing body composition. Its accuracy to predict fat mass (FM) in obesity and the change in FM following weight loss is questioned. Our objective was to compare leg-to-leg BIA to dual-energy X-ray absorptiometry (DXA) in the assessment of FM in a large population, the changes in FM after Roux-en-Y gastric bypass (RYGB) and to estimate between-method differences (bias) and limits of agreement.
METHODS: BIA (Tanita BC-420MA) and DXA (Hologic Discovery W) were used in 5,740 consecutive patients (mean BMI, 37.7 ± 8.2 kg/m(2)) examined in a clinical nutrition department and in 72 women undergoing RYGB (BMI, 47.2 ± 7.2 kg/m(2)). Analyses included correlations between methods and Bland Altman analysis.
RESULTS: In the entire population, BIA significantly overestimated FM in comparison with DXA (1.1 ± 6.1 kg, 0.8 ± 5.6%). FM estimates by each method were significantly correlated in absolute value (kg; r(2) = 0.9 in the whole population), and in percentage (r(2) = 0.6). However, wide limits of agreement were observed. In surgery patients, BIA significantly overestimated FM both before and 12 months after bypass. BIA significantly overestimated changes in FM after RYGB at 3 months (2.9 ± 5.0 kg) and at 12 months (1.9 ± 3.9 kg) but not at 6 months (0.9 ± 5.0 kg; p = 0.08). Estimates of changes in FM by each method were significantly correlated (r (2) = 0.4, 0.6, and 0.9, respectively).
CONCLUSION: According to the wide limits of agreement, BIA seems more interesting for epidemiological rather than individual use to evaluate body FM and FM changes in obese women undergoing RYGB.