Calcaneus Ultrasonometry and Dual-Energy X-Ray Absorptiometry for the Evaluation of Vertebral Fracture Risk

Abstract. The aim of this retrospective, cross-sectional, controlled, non-population-based study was to evaluate the specificity and sensitivity of quantitative ultrasonometry (QUS) of the heel and of dual-energy X-ray absorptiometry (DXA) in the prediction of morphometric vertebral fracture in postmenopausal women and to establish whether the combination of the two devices could improve the capacity to identify the presence of vertebral fracture. Also, we tried to identify the best Tscore threshold for high risk of vertebral fracture for both QUS and DXA, highlighting the discrepancies between the two methodologies and between the various sites examined with DXA. From 6,300 patients examined by DXA (total body, lumbar spine, total femur, femoral neck), QUS and DXA vertebral morphometry (MXA), we selected 764 postmenopausal women with nontraumatic vertebral fractures; 770 postmenopausal women with normal morphometry were chosen as a control group. Logistic regression analysis yielded odds ratios (ORs) for bone mineral density (BMD) measurements and QUS that were comparable: BMD-total body 4.16, BMD-lumbar spine 4.80, BMD-total femur 3.77, BMD-femoral neck 3.86, and QUS 4.41, without statistical differences even after correction for different confounding variables (menopausal years, weight, height, body mass index, and age). The ORs obtained from different combinations of QUS and DXA results did not show statistically significant differences compared to those from a single method alone. The sensitivity and specificity of all measurements were determined by area using the receiver operating characteristic curve; these were 0.94 for total body, 0.95 for lumbar spine, 0.86 for total femur, 0.89 for femoral neck, and 0.93 for QUS, without statistical difference. The areas under the curve obtained from the combination of QUS and DXA were higher but without statistical significance compared to QUS alone. In conclusion, both QUS and DXA were able to discriminate women with fracture from women without fracture and independently contributed to determining the association with fracture. The combination of QUS and BMD did not improve the diagnostic ability of either individual technique. We found different diagnostic thresholds for QUS and DXA