In women with type 2 diabetes mellitus (T2DM), sarcopenia is common but often underdiagnosed because reference imaging techniques are not widely available. In Diabetes Research and Clinical Practice, a cross-sectional analysis evaluated a consumer-grade bioelectrical impedance analyzer (BIA) compared with dual-energy X-ray absorptiometry (DXA) for detecting sarcopenia.

The analysis included 103 women with T2DM who underwent body composition assessment using a BIA and DXA. Sarcopenia was defined using fat-free mass index (FFMI), skeletal muscle index (SMI), and established DXA-based criteria. Statistical approaches included Spearman correlation, Bland–Altman plots, Cohen’s kappa, receiver operating characteristic curves, and regression-based bias correction.

FFMI measured by BIA correlated strongly with FFMI measured by DXA (r = 0.881) and with SMI (r = 0.854). Receiver operating characteristic analyses yielded area under the curve values of 0.878 and 0.873 using optimized FFMI-BIA thresholds, with sensitivity and specificity of 87.8% and 79.3%, and 85.1% and 82.8%, respectively. The bias-correction model explained 78.3% of the variance in FFMI-DXA and reduced mean bias to −0.0009 kg/m², with limits of agreement from −2.07 to 2.07 kg/m². Moderate agreement was observed for sarcopenia classification (κ = 0.575).

These findings show that a consumer-grade BIA demonstrated strong diagnostic performance with minimal systematic bias after calibration.