Reliability and variability of bioimpedance measures in normal adults: effects of age, gender, and body mass.

This study aimed to analyze the reliability and evaluate the causes of variability of bioimpedance parameters. Direct measures were analyzed because they are not affected by inappropriate prediction models. Resistance (R), reactance (Xc), and phase angle (PA) were determined at three fixed frequencies (5, 50, and 100 kHz) in 653 normal Germans (244 males and 409 females), aged 20-90 years, using a phase-sensitive whole-body tetrapolar bioimpedance analyzer (BIA 2000-M, Data Input, Germany). From these values, six bioimpedance ratios were calculated (R(5)/R(50), R(5)/R(100), Xc(5)/Xc(50), Xc(5)/Xc(100), PA(5)/PA(50), and PA(5)/PA(100)). Reliability of duplicate measurements, as determined by technical error, is high. ANOVA for repeated measurements yields a significant frequency main effect (within-subjects factor) and significant effects of age and gender (between-subject factors) on variation of resistance, reactance, and phase angle. Multiple regression analyses indicate independent effects of age, gender, and body mass index on variability of resistance, reactance, and phase angle at the three frequencies. Gender primarily influences variation in resistance (smaller values in males), whereas age mainly affects variations in reactance and phase angle (smaller values in older adults). Obesity is associated with smaller resistance (at all frequencies) and smaller reactance (high frequencies), but larger phase-angle values (low frequency). The study shows that variability of direct bioimpedance measures depends on age, gender, and body mass characteristics of the study population. The potential benefit for using both low and high frequencies in R measures is to differentiate between extra- and intracellular fluid spaces, which may be altered during human growth and aging. Copyright 2003 Wiley-Liss, Inc.