**Background:** Loss of skeletal muscle mass is a key driver of frailty and sarcopenia in older adults. The psoas major muscle, as the only muscle connecting the lumbar spine to the lower extremities, plays a critical role in trunk stability and motor function. Its cross-sectional area (CSA) has been used to predict whole-body muscle mass, cardiorespiratory fitness, sarcopenia, and surgical prognosis. While CT and MRI provide accurate measurements, they are costly, time-consuming, and involve radiation exposure. Bioelectrical impedance analysis (BIA) is a non-invasive, rapid, and portable alternative, but its ability to estimate individual deep trunk muscles like the psoas major had not been established. This study aimed to develop and cross-validate a BIA-based equation for estimating psoas major CSA at L3-L4 in older adults.
**Methods:** Ninety-two community-dwelling older adults (≥60 years) in central Taiwan were enrolled (47 females, 45 males; mean age females 70.2±6.8 years, males 75.2±8.4 years). Exclusion criteria included chronic diseases (cancer, renal failure, hepatitis, chronic lung disease). Participants were randomly divided into a modeling group (MG, n=62) and a validation group (VG, n=30). Whole-body and segmental impedance were measured using a standing octopolar single-frequency (50 kHz, 0.55 mA) BIA device (Tanita BC418MA). Whole-body impedance index (WBI) was calculated as height² divided by whole-body impedance (Z_whole). CT scans (Somatom Sensation 64, Siemens) at L3-L4 were used as the reference standard, with psoas major CSA (PMM_CT) manually segmented using Slice-O-Matic software (CoV 0.43%). Stepwise multiple linear regression in the MG selected predictors from height, weight, age, gender, WBI, trunk impedance index (TBI), waist circumference, hip circumference, and BMI. The final equation was cross-validated in the VG using correlation and Bland-Altman analysis.
**Key Results:** The final BIA estimation equation for psoas major CSA was: PMM_BIA = 0.183 × (h²/Z_whole) – 0.223 × age + 4.443 × gender (female=0, male=1) + 5.727, with r²=0.702, SEE=2.432 cm², p<0.001 (n=62). Variance inflation factors were all <4, indicating no multicollinearity. In the MG, the correlation between PMM_BIA and PMM_CT was r=0.708, with Bland-Altman bias ±1.96 SD of -4.30 to 4.30 cm². In the VG (n=30), the correlation was r=0.846, SEE=2.45 cm², and limits of agreement ranged from -4.55 to 4.75 cm². PMM_CT was significantly higher in males (19.1±4.7 cm²) than females (12.6±2.6 cm², p<0.001). WBI showed the strongest univariate correlation with PMM_CT (r=0.77), while TBI showed negligible correlation (r=0.03). Within-day CoV for whole-body impedance was 1.0%–1.9%; between-days CoV was 1.4%–2.1%.
**Clinical Implications:** This study provides preliminary evidence that standing single-frequency BIA, using whole-body impedance index, age, and gender, can estimate psoas major CSA with moderate accuracy (r²=0.702) and acceptable agreement with CT in healthy older adults. The method is fast, non-invasive, and radiation-free, making it suitable for routine clinical use and large-scale screening for sarcopenia and frailty. However, the SEE of 2.432 cm² and LOA up to ~4.75 cm² indicate that the BIA equation cannot yet replace CT for individual diagnostic decisions. The study is limited by its small sample size, single-ethnicity (Taiwanese), exclusion of sarcopenic or chronically ill patients, and use of a single-frequency BIA device. The authors note that the SEE% was 19.3% for women and 12.7% for men, leaving room for improvement. Future work should incorporate multi-frequency BIA or bioimpedance spectroscopy, segmental trunk impedance, and 3D body scanning to improve accuracy, and must validate the equation in other populations and clinical settings before it can be used for sarcopenia diagnosis.