Lack of agreement of in vivo raw bioimpedance measurements obtained from two single and multi-frequency bioelectrical impedance devices

Background It is important for highly active individuals to accurately assess their hydration level. Bioelectrical impedance (BIA) can potentially meet these needs but its validity in active individuals is not well established. Methods We compared whole-body bioimpedance measurements obtained from multi-frequency bioelectrical impedance spectroscopy (BIS, Xitron 4200) at a 50 kHz frequency with those determined by a phase-sensitive single-frequency device (SF-BIA, BIA-101, RJL/Akern Systems) in two populations: active adults and elite athletes. Results One hundred twenty-six participants, including active males involved in recreational sports (N = 25, 20-39 yr) and elite athletes (females: N = 26, 18-35 yr; males: N = 75, 18-38 yr) participated in this study. Reactance (Xc), Resistance (R), Impedance (Z), and phase angle (PhA) were obtained by BIS and SF-BIA. Small but significant differences (R: -9.91 +/- 15.09 Omega; Xc: -0.97 +/- 2.56 Omega; Z: -9.96 +/- 15.18 Omega; PhA: 0.12 +/- 0.2 degrees) were observed between the bioimpedance equipment in all measured variables (p < 0.05) though differences were within the devices' technical error of measurements. Device-specific values were highly (p < 0.0001) correlated [R-2 ranged from 0.881 (Xc) to 0.833 (R)], but slopes and intercepts were different (p < 0.0001) from 1 and 0, respectively. Relatively large limits of agreement were observed for R (-40 to 21 Omega), Xc (-6 to 4 Omega), PhA (-0.4 to 0.5 degrees), and impedance (-40 to 20 Omega). Conclusion: Bioimpedance measurements from the current single- and multi-frequency devices should not be used interchangeably. The of lack of agreement between devices was observed in determining individual values of R, Xc, Z and PhA of highly active populations possibly due to methodological and biological factors.