Anatomical factors associated with femoral neck fractures of elderly Beijing women.

Analyses using newly developed structural measures of minimal model (aBMD, W, Sigma, Delta) in a Chinese female age-matched cohort of femoral neck fracture patients and non-fracture community controls resulted in improved fracture prediction, suggesting the usefulness of new variables, extending the value of widely available DXA technology. INTRODUCTION We have developed a new approach to evaluate 2D femoral neck (FN) structure, the minimal model (MM). This model includes FNaBMD and FNWidth with two new internal structural measures, the standard deviation of normalized mineral mass projection profile distribution (FNSigma) and the displacement between center of mineral (CoM) mass and geometric center of mineral mass projection profile (FNDelta). METHODS Differences in these four structural measurements together with age, weight, and height were compared in the contralateral hip of 285 FN fracture Chinese female patients and 261 age-matched community controls. Structural variables were calculated from DXA equivalent 2D images obtained from QCT scans analyzed using Mindways Software. RESULTS Review of FN scanned profiles of fracture patients and controls identified substantial reduction in mineral mass in the superior segment of FN. Fracture participants were taller, weighed less, and had lower FNaBMD and larger FNDelta, due to a larger inferior displacement in the CoM, consistent with greater reduction in superior segment bone. Logistic regression identified increased height, reduced FNaBMD, increased FNDelta, and reduced FNSigma per SD as significant independent contributors to differentiating fracture from non-fracture. Area under ROC analysis identified significant improvement in discrimination with the addition of FNDelta and FNSigma to the model including age, height, weight, and FNaBMD (C statistic 0.87 and 0.84 respectively). CONCLUSIONS These data extend previous data that identified the benefit of 2D FN internal structural information in discriminating those at increased future fracture risk from recently fractured individuals. These data support continuing investigation of MM analysis as a straightforward analytical approach adding value to DXA hip aBMD in predicting fracture risk.