Related Experiment Video
Updated: Jul 12, 2026
Segmentation and Linear Measurement for Body Composition Analysis using Slice-O-Matic and Horos
Published on: March 21, 2021
Agreement and reliability analysis of sarcopenia measures based on muscle function, body composition, and imaging in
Alexa Semonche1, Ann A Lazar2, Justin K Scheer3
11Department of Neurological Surgery, University of California, San Francisco.
Objective:
There is no gold standard for assessing sarcopenia. The authors aimed to quantify agreement and reliability among common sarcopenia measures based on muscle function, body composition, and imaging characteristics in a cohort of adult spinal deformity (ASD) patients.
Methods:
This was a cross-sectional study. Preoperative ASD patients at a single tertiary-care center underwent the following sarcopenia assessments: 2 functional muscle assessments (grip strength and gait speed), 3 bioelectrical impedance (BIA) measures (skeletal muscle index [SMI], phase angle, and extracellular water [ECW]/total body water [TBW] ratio), and 2 imaging-based assessments (psoas muscle index [PMI] and total psoas area [TPA]/vertebral body area [VBA] ratio). Spearman's correlation analysis was used to test for associations. Measurements were standardized into sex-specific z-scores. Bland-Altman analysis was used to quantify agreement and Cronbach's alpha analysis was used to quantify reliability between sarcopenia measures.
Results:
Between 2017 and 2025, 272 ASD patients were enrolled in the study. Of these, 88 ASD patients (59.1% female, median age 67.7 years) completed all sarcopenia assessments and thus were included in the analysis. Most sarcopenia measures were weakly correlated, except for PMI and TPA/VBA, which were moderately correlated (ρ = 0.84, p < 0.001), and phase angle and ECW/TBW (ρ = -0.82, p < 0.001). Bland-Altman analysis demonstrated poor agreement between all pairs of sarcopenia measures. Cronbach's alpha analysis showed poor reliability (α < 0.7) between all pairs of sarcopenia measures except for PMI and TPA/VBA (α = 0.89). These results did not significantly change when all 272 patients with imputed missing data were included.
Conclusions:
The authors found low agreement and reliability between 7 commonly used sarcopenia measures, except for good reliability between PMI and TPA/VBA. These sarcopenia measures are not interchangeable and may not be measuring the same underlying clinical entity. This is the first study to quantify agreement and reliability between sarcopenia measures. Future studies are needed to determine which sarcopenia measures best predict clinical outcomes in ASD patients.
