Poster Presentation ANZBMS-MEPSA-ANZORS 2022

Assessment of measures of bone distribution within DXA image cross sections improves prediction of femoral neck fracture independent of aBMD (#216)

Richard Prince 1 , Ben Khoo 2 , Joshua Lewis 3 , Marc Sim 3 , Keenan Brown 4
  1. University of Western Australia, Crawley, WA, Australia
  2. Medical Technology and Physics, Sir Charles Gairdner Hospital, Nedlands, Western Australia, Australia
  3. School of Medical and Health Sciences, Edith Cowan University, Joondalup, Western Australia, Australia
  4. Mindways Software, , Austin, , Texas,, USA.

We have previously published on additional benefits of assessing geometric variables of bone distribution within cross sections of DXA images for hip fracture prediction in Chinese populations using Sigma, a measure of mineral mass distribution within the cross section and Delta, the inferior displacement of center of mass from the geometric center. We now report on role of these variables in the Perth Longitudinal Study of Aging in Women of 1,151 participants mean (SD) age 75(3) years with 15-years of follow up.

Baseline DXA structure in 1023 women who did not fracture were compared to 69 who sustained a femoral neck fracture (FNF), a median of 9.8 years after baseline.

The analytical approach focused on using Sigma and Delta in addition to common hip fracture prediction variables, age and Total Hip aBMD (THaBMD). In Cox analysis for FNF in which age and SD THaBMD were forced and SD Sigma and Delta entered in forward regression, the reported HRs (95% CI) were SD THaBMD 0.79 (0.61 – 1.03), age 1.13 (1.04–1.24) and SD Delta 1.70 (1.36-2.12). Interestingly in this model THaBMD was not significant. Effects of using tertiles of Delta in addition to age and Total hip aBMD are shown in Figure 1. Compared to lowest tertile of Delta, Tertile 2 HR was 2.21 (1.05–4.67), and Tertile 3 HR was 3.80 (1.88–7.69). The C statistic for age and THaBMD was 0.62, (0.55–0.69), the C statistic for FN Delta, age and THaBMD was improved at 0.69, (0.63–0.75) P for difference=0.01).

These data identify FN Delta in addition to aBMD as an important determinant of FNF prediction years before actual fracture. The larger the FN Delta the greater bone mass loss in the superior segment of FN compared to the inferior, related to fracture in buckling the biomechanical mechanism of FN fracture.

 

62999fc77b74d-Fig+1_ANZBMS.jpg