Rationale: Diagnosis of sarcopenia in older adults is essential for early treatment in clinical practice. Bio-electrical impedance
analysis (BIA) may be a valid means to assess appendicular lean mass (ALM) in older adults, but limited evidence is available.
Therefore, we aim to evaluate the validity of BIA to assess ALM in older adults.
Methods: In 215 community dwelling older adults (age ≥ 55 years), ALM was measured by BIA (Tanita MC-780; 8-points) and
compared with dual-energy X-ray absorptiometry (DXA, Hologic Discovery A) as reference. Validity for assessing absolute values of
ALM was evaluated by: 1) bias (mean difference), 2) percentage of accurate predictions (within 5% of DXA values), 3) individual
error (root mean squared error (RMSE), mean absolute deviation), 4) limits of agreement (Bland-Altman analysis). For diagnosis of
low ALM, the lowest quintile of ALM by DXA was used (below 21.4 kg for males and 15.4 for females). Sensitivity and specificity
of detecting low ALM by BIA were assessed.
Results: Mean age of the subjects was 71.9 ± 6.4, with a BMI of 25.8 ± 4.2 kg/m2, and 70% were females. BIA slightly
underestimated ALM compared to DXA with a mean bias of -0.6 ± 0.2 kg. The percentage accurate predictions was 54% with
RMSE 1.6 kg and limits of agreements -3.0 – +1.8 kg. Sensitivity was 79%, indicating that 79% of subjects with low ALM
according to DXA also had low ALM with the BIA. Specificity was 90%, indicating that 90% of subjects with ‘no low’ ALM
according to DXA also had ‘no low’ ALM with the BIA.
Conclusions: This comparison showed a poor validity of BIA to assess absolute values of ALM, but a reasonable sensitivity and
specificity to diagnose a low level of ALM in community-dwelling older adults in clinical practice.