Validation of bioelectrical impedance analysis (BIA) equations to predict appendicular skeletal muscle and fat mass in older adults with sarcopenia

Rationale: No BIA equations exist to predict appendicular skeletal muscle mass (ASMM) and appendicular fat mass (AFM) in older adults with sarcopenia. NHANES public reference values for body composition are applicable only to Hologic DXA systems. Therefore we 1) developed and cross-validated appendicular soft tissue BIA equations with both Hologic and GE Lunar systems as their reference and 2) compared the new ASMMHOLOGIC equation to two previously published BIA equations.
Methods: Participants with sarcopenia class I or II were recruited from 18 European study centers for the PROVIDE intervention study (NTR2329). In total, 291 cases having both an evaluable BIA (50kHz) and DXA measurement were included (mean age 77.6±6.9yrs; mean %fat 37.2±5.9%). BIA equations, using DXA-derived ASMM and AFM as the dependent variable, and age, gender, weight, height²/resistance and reactance as independent variables, were developed using multiple linear regressions.
Results: Cross-validation resulted in 4 equations using the whole sample.
Hologic (n=104): ASMM (kg) = 4.957 + (0.196*Height²/Resistance) + (0.060*Weight) – (2.554*Sex) (R²=0.90; SEE=1.28)
AFM (kg) = -4.716 + (0.142*Height²/Resistance) + (0.316*Weight) + (4.453*Sex) – (0.040*Reactance) (R²=0.73; SEE=1.54)
GE Lunar (n=187): ASMM (kg) = 1.821 + (0.168*Height²/Resistance) + (0.132*Weight) –(1.931*Sex) + (0.017*Reactance) (R²=0.86; SEE=1.37)
AFM (kg) = -6.553 + (0.093*Height²/Resistance) + (0.272*Weight) + (4.295*Sex) (R²=0.70; SEE=1.53) Sex: women=1, men=0
Previously published BIA-equations overestimated significantly ASMMHOLOGIC with mean biases of -0.36 kg and -1.05 kg.
Conclusion: We propose new BIA equations for the accurate estimation of limb soft tissue masses in sarcopenic older adults. Users of both Hologic and GE Lunar may benefit from these equations in field research.