Comparison of calcaneal quantitative ultrasound and bone densitometry parameters as fracture risk predictors in type 2 diabetes mellitus

Aim To investigate the utility of calcaneal quantitative ultrasound compared with bone densitometry (DXA) in predicting incident low-trauma fracture in type 2 diabetes. Methods This retrospective cohort study included a subset of participants in the Dubbo Osteoporosis Epidemiology Study who had concurrent calcaneal quantitative ultrasound and DXA measurement, comprising 809 people without type 2 diabetes and 96 with type 2 diabetes. Fracture data had been collected prospectively. Cox proportional hazard models and receiver operating curves (ROC) were used to compare calcaneal quantitative ultrasound and DXA parameters as predictors for any low-trauma fracture. Results The median age of participants was 71 years (IQR 68-76, 50% men) for those without type 2 diabetes and 70 years (IQR 68-76, 55% men) for those with type 2 diabetes. There was no difference in low-trauma fracture incidence between groups when stratified by sex. In those without type 2 diabetes, the hazard ratio for fracture per 1 sd decrease in broadband ultrasound attenuation and femoral neck bone mineral density (BMD) was 1.47 [95% confidence interval (CI) 1.26-1.71] and 1.39 (95% CI 1.17-1.64), respectively. The corresponding figures in type 2 diabetes were 1.81 (95% CI 1.03-3.19) for broadband ultrasound attenuation and 2.55 (95% CI 1.28-5.08) for femoral neck BMD. Conclusion Broadband ultrasound attenuation is comparable with femoral neck BMD as a predictor for low trauma incident fracture in type 2 diabetes. Calcaneal quantitative ultrasound offers several advantages over DXA and should be considered in further studies of bone health screening or in clinical practice where DXA is unavailable.