Nutrition in Clinical Practice · Oct 20, 2026

Validity and agreement of bioelectrical impedance equations against dual energy X-ray absorptiometry for muscle mass assessment in patients on hemodialysis: A cross-sectional study

By Borges S, Martins TF, Machado FVC et al.

BACKGROUND: Low muscle mass is common in patients on hemodialysis (HD), and bioelectrical impedance analysis (BIA) predictive equations have been proposed as an alternative for estimating muscle mass. This study aimed to analyze the validity and agreement of different BIA-based prediction equations against dual energy X-ray absorptiometry (DXA) for estimating skeletal muscle mass (SMM) and appendicular skeletal muscle mass (ASMM) in patients on HD. METHODS: A cross-sectional study enrolled patients on HD. Clinical and laboratory data and body composition by BIA and DXA were collected. The accuracy of six prediction equations (Janssen, Barbosa-Silva, Kyle, Lin, Macdonald, and Sergi) for estimating SMM and ASMM was tested against DXA as the reference standard. Agreement was assessed using Lin's Concordance Correlation Coefficient (CCC), Bland-Altman analysis, Kappa statistics, Receiver Operating Characteristic (ROC) curves, and Passing-Bablok regression. RESULTS: The sample consisted of 57 participants, age 52 ± 13 years, 32 (56%) male, median dialysis vintage 14 (4-40) months, 17 (30%) with diabetes, and 48 (84%) with hypertension. The best agreement with DXA was observed for the Lin (CCC = 0.911), Sergi (CCC = 0.902) and Kyle (CCC = 0.902) equations. For low muscle mass diagnosis, Kappa values ranged from moderate to substantial, with the Lin equation showing the highest agreement (κ = 0.734). Additionally, the Lin equation demonstrated good accuracy (AUC = 0.794, 95% CI 0.68-0.91), with 68,4% sensitivity and 89.5% specificity. CONCLUSION: The Lin equation provided the most accurate and reproducible estimates of muscle mass compared with DXA and showed good diagnostic performance for low muscle mass in patients on HD.

doi.org/10.1002/ncp.70192Open accessPubMed 42902512Biomarkers