RESUMEN
The purpose of this study was to assess the ultrasound (US) parameters measured by Contact Ultrasound Bone Analyser (CUBA) clinical system for discriminating the subjects with vertebral fractures from those without vertebral fractures. The subjects consisted of 114 postmenopausal women over age 50 (mean +/- SD: 72.2 +/- 8.7). Seventy-three had vertebral fractures (fracture group) and 41 had no vertebral fractures (control group). Values of all US parameters and bone mineral density (BMD) in the fracture group were significantly lower than those in the control group, even after adjusting for age. Areas under the receiver operating characteristic (ROC) curve was 0.768 +/- 0.056 (mean +/- SE) for broadband ultrasound attenuation (BUA) 0.828 +/- 0.045 for velocity of sound (VOS), 0.707 +/- 0.058 for lumbar spine, 0.872 +/- 0.050 for femur neck, 0.790 +/- 0.050 for trochanter, and 0.695 +/- 0.060 for Ward's triangle. There were no significant differences among the areas under the ROC curves in BUA, VOS, lumbar spine, femur neck, trochanter, and Ward's triangle. US parameters (BUA and VOS) had the same discriminatory power as spine and hip BMD for evaluating the vertebral fracture risk. As far as disadvantages of the use of BMD measurement against US measurement, US measurement is potentially useful for screening of vertebral fractures. However, the bias concerning the propotion of the number of patients with or without vertebral fractures could not be neglected in this study.