[en] Absorptiometry, Photon ; Adult ; Aged ; Analysis of Variance ; Bone Density ; Female ; Fingers/ultrasonography ; Humans ; Logistic Models ; Lumbar Vertebrae/radiography ; Male ; Metacarpus/ultrasonography ; Middle Aged ; Osteoporosis, Postmenopausal/ultrasonography ; ROC Curve ; Reproducibility of Results ; Sensitivity and Specificity
[en] OBJECTIVE: The present study was designed to assess the reproducibility and the diagnostic sensitivity of the amplitude-dependent speed of sound (SoS) at the distal metaphysis of the proximal phalanges. METHOD: Fourteen presumably healthy volunteers were repeatedly measured every 6 weeks for approximately 6 months in order to assess the reproducibility of the SoS of the phalanges. We recruited 91 post-menopausal women, aged 55-75 years, who were divided in three groups according to their lumbar bone mineral density (BMD) and the existence of prevalent vertebral fractures. The objective was to evaluate the diagnostic sensitivity of SoS measurements. We used DBM Sonic 1200 equipment, and assessed the velocity at which US cross the phalanx in a lateral-medial direction. In post-menopausal women, BMD was measured by dual energy X-ray absorptiometry (DXA) at the level of the lumbar spine, the total zone of the non-dominant hip and the femoral neck zone of the non-dominant hip. RESULTS: The precision of the SoS measurements was 0.71+/-0.05% (mean+/-S.E.M) whereas the reproducibility was 0.95+/-0.06%. Subjects with low BMD or prevalent fractures had significantly lower values of SoS (P < 0.001) than the controls. ROC curve analysis applied to the study population confirmed that SoS was able to discriminate between the controls and osteoporotic subjects (area under the ROC curves were 0.82 (low bone mineral density) and 0.85 (prevalent fractures), respectively). Hip BMD was found to be the most significant variable when comparing the controls and the low density patients by stepwise discrimination and SoS significantly improved the discrimination between the groups when added to the hip BMD. The hip BMD was again the most discriminant variable when applying the same techniques to controls and patients with prevalent fractures, followed by SoS and lumbar BMD. A cut-off value of 1881 m/s is defined for SoS by logistic discrimination and likelihood ratio function. With this value, the sensitivity and the specificity for SoS used in the diagnosis of established osteoporosis were, 81.5% and 79.3%, respectively. Sensitivity and specificity were significantly improved when combining ultrasonometry and densitometry. CONCLUSION: Measurement of ultrasound velocity at the phalanges appears to be a precise and reproducible technique. SoS discriminates between normal post-menopausal women and patients with either low lumbar BMD or prevalent fractures to the same extent as BMD measurements.