Methods of assessing body composition suitable for use in clinical trials should be accurate, reliable, and easy to perform. One such technique routinely implemented is handheld bioelectrical impedance analysis (BIA). The validity of this method, however, in body composition assessment of overweight women is not known. The aim of this study was to validate the hand-held BIA technique with magnetic resonance imaging (MRI) for the assessment of body composition in overweight women. Fat mass, percent fat mass, fat-free mass, and percent fat-free mass values estimated by hand-held BIA were compared to those measured by MRI. Thirty-one Caucasian women (50.1 +/- 8.2 years, body mass index of 26.9 +/- 3.1 kg/m(2)) participated in the study. BIA measurements were highly reproducible (technical error (TE) was 0.06 +/- 0.07 kg for fat mass and 0.08 +/- 0.11% for percent fat mass), but were significantly different (P < 0.0001) for each body composition parameter when compared to MRI. BIA underestimated fat mass by 2.3 +/- 3.3 kg and percent fat mass by 5.6 +/- 3.9%. Likewise, BIA overestimated fat free mass by 7.4 +/- 2.7 kg and percent fat free mass by 5.6 +/- 3.2%. No relationship between the bias and the mean of the two measurements was noted suggesting that bias is not related to measurement size. Although hand-held BIA gives reproducible findings, the bias noted for all body composition parameters puts into question the validity of this regional impedance device for use in clinical trials in overweight women.