Single frequency bioelectrical impedance is a poor method for determining fat mass in moderately obese women

Background: The primary aim of weight loss intervention in obesity is the loss of fat mass (FM). Hence, determinations of changes in FM and fat free mass (FFM) during weight loss are of clinical value. The authors compared the clinical utility of SkinFold Thickness (SKF), tetrapolar bioelectrical impedance analysis (BIA) and a body mass index (BMI) based calculation, in determining changes in percentage of fat mass (Delta%FM). Methods: Using dual X-ray absorptiometry (DEXA) measurements of %FM as a standard, BIA, SKF and BMI were compared in 41 moderately obese women (BMI 30-35) before and after significant weight lost (-13.9 +/- 5.8 kg). Results: When measuring fat mass loss, SKF was precise and accurate with a bias of +0.86 +/- 6.16 %, while the BMI-based estimation had a systematic bias of +6.36 +/- 6.04 % (r(2) = 0.791, P<0.001). BIA using the Lukaski formula had a bias of +5.22 % and limits of agreement that approached the magnitude of the measurement (+/- 20.82 %), thus providing no information. In contrast, BIA using the Segal formula had a systematic bias of +7.81 % (r(2) = 0.636, P<0.001) and gave narrower limits of agreement (+/- 8.34 %). Conclusion: For measuring changes in %FM with weight loss, BIA has no clinical value using the Lukaski formula, and using the Segal formula BIA provided no additional information to that given by BMI. We show that BIA instrument variables confound the estimates of %FM achieved by the BMI component of the Lukaski and Segal formulas.