Mandibular trabecular bone reflects skeletal bone condition and may improve fracture prediction with FRAX: a longitudinal study

Talk Code: 
EP1D.01
Presenter: 
Grethe Jonasson
Co-authors: 
Valter Sundh, Dominique Hange, Margareta Ahlqwist, Magnus Hakeberg, Lauren Lissner.
Author institutions: 
Institute of Odontology, Institute of Medicine University of Gothenburg, Sweden. 

Problem

Fragility fractures are a major health problem. Fracture risk is highest in osteoporotic women but bone mineral density measurements (BMD) is not recommended for screening because of high costs and low sensitivity. To evaluate fracture risk based on simple clinical variables, a Web-tool called FRAX, has been developed by WHO. It can be used without BMD. Perfect prediction is unattainable because fractures are primarily caused by falls but improvement in fracture risk assessment is needed. . Most adults in Western countries visit their dentists regularly and dental radiographs are performed showing the trabecular structure. We have earlier shown that a sparse trabeculation in the mandible is associated with increased fracture risk and hypothesize that this risk factor may improve FRAX prediction.

Approach

In a longitudinal study, 413 women, born 1930, 1922 or 1914, from the Prospective Population Study of Women in Gothenburg examined 1992/93 were included. The FRAX questionnaire assesses 10-year fracture risk based on following variables: age, BMI, previous fracture, current smoking, glucocorticoids, rheumatoid arthritis, and alcohol. It was calculated using the link http://www.shef.ac.uk/FRAX/charts/Chart_SWE_ost_wom_bmi.pdf. The radiographic trabecular bone was evaluated with a visual index as dense, mixed dense plus sparse, or sparse. Self-reported fracture were available both at start and end of the 10-year risk period (1992-2002). Odds ratios, relative risk, Chi-square, sensitivity, specificity, PPV, NPV and Area under Curve (AUC) were calculated.

Findings

In the ten-year period, 67 (16.2%) women suffered an incident fracture. 199 (48.1%) women had sparse trabeculation, 26% sustained an incident fracture compared to 7.0% in the non-sparse group (p<0.001). Mean FRAX value was 8.3+9.76 (6.5-56). 197 women (47.7%) had FRAX score >15 and among them 27% sustained a fracture, compared to 6.5% in the group <=15 (p<0.001). Odds ratios for sparse trabeculation and FRAX >15 as fracture predictors were associated with a more than four-fold risk for future fracture (4.7 and 5.3). If at least one of FRAX>15 and sparse trabeculation were present as exposure, 23% had a fracture versus 1.5% in those lacking both, giving a 16-fold increase in risk (RR = 15.8, 95% CI= 3.9 - 63.5); 65 of 67 fractures were identified. If the group without one of these two risk factors was compared with the group with both risk factors, the relative risk for the latter was 34.1. Using only FRAX>15 as fracture predictor AUC was 0.69 (0.63-0.74), and when sparse trabeculation was included AUC increased to 0.75 (0.70-0.81).

Consequences

Mandibular sparse trabeculation is easily assessed in dental clinics and ameliorates fracture prediction with FRAX. Collaboration between dental and medical profession may be useful for identifying women before the first fracture occur.

Submitted by: 
Grethe Jonasson
Funding acknowledgement: 
The Health &amp; Medical Care Committee of the Region Västra Götaland funded the study.