Abstract
BACKGROUND: The purpose of this study was to analyze differences in geometrical properties of the proximal femur and predict the occurrence of basicervical fractures through a comparative study of femoral neck and basicervical fractures in patients undergoing hip structural analysis (HSA). METHODS: All patients with hip fractures who were at least 65 years old and admitted to our hospital between March 2017 and December 2019 were eligible for this study. During the study period, 149 femur neck fractures (FNF) and basicervical fractures (intertrochanteric fractures of A31.2) were included in this study. Fifty-nine patients were included in the final analysis. Factors considered to be important confounders affecting the occurrence of basicervical hip fractures were chosen for propensity-score analysis. A logistic model with basicervical hip fracture as the outcome and age, sex, weight, spinal T-score, hip T-score, and vitamin D levels as confounders was used to estimate the propensity score. RESULTS: The cross-sectional moment of inertia (CSMI) of the intertrochanter was significantly lower in patients with basicervical hip fracture (HF) than in patients with FNF (p = 0.045). However, there was no significant differences in any other HSA variable between the two groups. Receiver operating characteristic (ROC) analysis showed that cutoff point for HSA was 100 for hip axis length (HAL) (AUC = 0.659, p < 0.001) and 5.712 for CSMI of the intertrochanter (AUC = 0.676, p < 0.001). ROC analysis showed that cutoff points of HAL, CSMI of intertrochanter, and handgrip strength were 104.8, 8.75, and 16.9, respectively (AUC = 0.726, p < 0.001). CONCLUSIONS: Proximal femoral geometric analysis using HSA is a useful method for predicting the type of hip fracture. Additionally, a lower CSMI, a shorter HAL, and a lower grip strength are major predictors of basicervical fractures.