Ai-Chun Huang ¹ Hsueh-Kuan Lu ² Chien-Wei Liang ³ Yi-Sung Tsai ⁴ Kuen-Chang Hsieh ^5,6 Chung-Liang Lai ^7,8*

• ¹ Physical Education and Health Center, National Kaohsiung University of Hospitality and Tourism, Kaohsiung City, Taiwan
• ² General Education Center, National Taiwan University of Sport, Taichung, Taichung City, Taiwan
• ³ Office and Physical Education and Sport, National Chung Hsing University, Taichung City, Taiwan
• ⁴ Department of Physical Medicine and Rehabiliation, Taichung Hospital, Ministry of Health and Welfare, Taichung City, Taiwan
• ⁵ Department of Research and Development, Starbia Meditek Co., Ltd., Taichung City, Taiwan
• ⁶ Big Data Center, National Chung-Hsing University, Taichung City, Taiwan
• ⁷ Department of Occupational Therapy, Asia University, Taichung City, Taiwan
• ⁸ Department of Physical Medicine and Rehabilitation, Puzi Hospital, Ministry of Health and Welfare, Chiayi, Taiwan

Objective: Lower limb muscle mass (LLMM) accounts for more than 50% of the total body skeletal muscle mass. Assessing leg muscle mass in middle-aged and elderly individuals is crucial for the prevention and diagnosis of sarcopenia. Current bioelectrical impedance analysis (BIA) devices are capable of measuring LLMM, but validation studies are limited. This study compares the accuracy of BIA devices with different frequencies for measuring LLMM in middle-aged and elderly populations.Methods: LLMM measurements were obtained using the following devices: foot-tofoot dual-frequency (StarBIA201, 5, 50 KHz), multi-segment single-frequency (Tanita BC418, 50 KHz), dual-frequency (InBody270, 20, 100 KHz), triple-frequency (Tanita MC780MA, 5, 50, 250 KHz), and six-frequency (InBody770, 1, 5, 50, 250, 500, 1000 KHz). Dual-energy X-ray absorptiometry (DXA) served as the reference standard. Comparisons were conducted using the following metrics: (1) mean difference (bias), (2) limits of agreement (LOA), (3) Pearson correlation coefficients, and (4) ordinary least product (OLP) regression analysis.Results: A total of 153 community-dwelling individuals aged over 55 years (102 females, 51 males) were recruited. The average age of participants was 67.5 ± 8.9 years, with a BMI of 23.9 ± 3.9 kg/m² and a body fat percentage of 35.8 ± 6.5%. The correlation coefficients of StarBIA201, BC418, InBody270, MC780, and InBody770 with DXA were 0.902, 0.903, 0.917, 0.925, and 0.928, respectively. Their mean differences were -0.141, -2.731, -0.587, -1.613, and -0.625 kg, with LOAs of 4.3, 5.7, 4.0, 5.1, and 3.8 kg, respectively. StarBIA201 and InBody270 showed no fixed or proportional biases.This study demonstrates that the four-electrode foot-to-foot BIA method shows significant practicality and potential in assessing LLMM. Compared to multifrequency BIA and DXA, this method is simpler to operate and more convenient, making it particularly suitable for preliminary screening and assessment of sarcopenia in clinical and community settings.