Objectives: Bioimpedance analysis (BIA) methods have potential to predict appendicular skeletal muscle mass (SM), although available 50 kHz prediction models include, in addition to impedance (Z), an independent age term. An age term in models is undesirable as it reflects incomplete understanding of underlying conduction physiology. This study tested the hypothesis, based on fluid distribution models related to aging, that appendicular SM bioimpedance analysis (BIA) prediction models would no longer include an independent age term, after first controlling for stature-adjusted appendicular impedance (height2/Z), at injected frequencies greater than 50 kHz.
Design: Cross-sectional evaluation of adults who had segmental Z and phase angle (phi) measured with multiple frequency BIA, and arm and leg SM with dual-energy X-ray absorptiometry (DXA). Skeletal muscle prediction models were developed with appendicular SM as dependent variable and height2/Z, gender, age and phi as potential independent variables.
Results: Examination of hypothesis in 49 subjects indicated: both arm and leg SM were highly correlated with height2/segmental Z at frequencies ranging from 1-300 kHz; gender was significant covariate in prediction models only at 1 kHz; age remained a significant covariate after controlling for height2/segmental Z at all frequencies; phi did not add significantly to models; and SM prediction models gave maximum R2 at 50 kHz for arm but R2 continued to rise up to 300 kHz for leg.
Conclusion: Although multifrequency BIA did not eliminate SM prediction model age term, our findings suggest injected frequencies up to 300 kHz may have advantages for evaluating leg SM over conventional 50 kHz method.