Hip muscle radiodensity measured using computed tomography is associated with femoral density and proximal femoral fracture status in an elderly population: a retrospective case–control study
Preventing proximal femoral fractures in older adults is of paramount importance. In clinical practice, bone mineral density measured by dual-energy X-ray absorptiometry is the reference standard for assessing skeletal fragility, whereas muscle is rarely quantified. Skeletal muscles are also crucial for maintaining movement stability, and computed tomography (CT) has become a reliable and commonly used imaging modality. However, existing studies are limited by small sample sizes and a lack of clinically applicable reference thresholds. This study investigated whether CT-derived hip muscle and femoral parameters are associated with proximal femoral fracture status in older adults. In this retrospective case–control study, 354 patients aged ≥ 60 years who underwent hip CT at The Second Affiliated Hospital of Xi’an Jiaotong University from October 2023 to October 2025 were included and divided into a fracture group ( n = 125) and a non-fracture group ( n = 229) according to clinical diagnosis. CT images were used to evaluate the following regions of interest: gluteus maximus muscle (G max), gluteus medius muscle (G med), iliopsoas muscle, femoral head (FH), femoral neck (FN), femoral intertrochanteric (FIT), and the cortical thickness index (CTI) of the femur and calcar. Skeletal muscle cross-sectional area (CSA), Hounsfield unit (HU) radiodensity, and bone parameters were used to assess the hip. Compared with the non-fracture group, the CSA and radiodensity of several muscle and femoral parameters were lower in the fracture group ( P < 0.001). In a fully adjusted hierarchical logistic regression model, G med radiodensity (odds ratio [OR] 0.758, 95% confidence interval [CI] 0.588–0.977), Iliopsoas radiodensity (OR 0.776, 95% CI 0.617–0.976), FH radiodensity (OR 0.972, 95% CI 0.950–0.994), and femoral cortical radiodensity (OR 0.993, 95% CI 0.987–0.999) were independently associated with fracture status after adjustment for age, sex, and BMI. The model discriminated fracture status with an area under the curve (AUC) of 0.846 (95% confidence interval [CI] 0.798–0.894; sensitivity 79.4%, specificity 77.8%) in the derivation dataset and a mean AUC of 0.834 ± 0.077 in five-fold internal cross-validation; no external validation was performed. Correlation analysis showed consistent associations between muscle and femoral parameters. In this retrospective case–control cohort, CT-derived hip muscle radiodensity, particularly that of the G med and Iliopsoas muscle, was associated with proximal femoral fracture status independently of age, sex, and BMI. When reduced muscle radiodensity or a thin calcar is identified on CT obtained for another indication, clinicians may consider further fracture risk assessment. Prospective and external validation are required before clinical use.