Cut-off values determined for vertebral fracture by peripheral quantitative computed tomography in Japanese women.

Abstract

In spite of the benefits of bone mass measurement by dual-energy X-ray absorptiometry (DXA), the use of DXA has limitations. It is unable to assess a true bone density, and cannot discriminate between the trabecular and cortical bone compartments. Ultradistal radius bone density was measured using peripheral quantitative computed tomography (pQCT) to determine reference values for total bone density (BD), trabecular bone density (TBD), polar strength strain index (pSSI), total bone mineral content (BC), trabecular bone mineral content (TBC), cortical bone density (CBD), cortical bone mineral content (CBC) and polar cross-sectional moment of inertia (pCSMI) in the Japanese female population, and to ascertain the cut-off values of the measured indicators that could most efficiently discriminate osteoporotic subjects with vertebral fractures. A total of 5266 healthy Japanese women aged 20-89 years were included in this study to determine Japanese reference values. Additionally, 621 who had undergone radiographic examination of the thoracic and lumbar spine at the time of pQCT measurement were selected to determine the cut-off values of BD, TBD, pSSI and other indicators for vertebral fractures. All the healthy subjects were divided into 5 year age groups. The BD showed nonsignificant changes from the 20-24 year age group to the 45-49 year age group, and fell significantly thereafter. The TBD maintained a plateau until the 40-44 year group, which corresponds to the young adult mean (YAM) values of the lumbar spine, femoral neck and radius BMDs measured using DXA. The TBD decreased significantly thereafter. The pSSI did not change significantly from the 20-24 year age group to the 45-49 year age group, and decreased slightly in the 50-54 year age group and markedly after 55-59 years. The cut-off values for the discrimination of vertebral fractures were obtained by the calculation of sensitivities, specificities and the area under the curves obtained using age-adjusted receiver operating characteristics (ROC) analysis. Odds ratios and 95% confidence limits (CL) were calculated using age-adjusted logistic analysis. The cut-off values for vertebral fractures, the area under the ROC curves (AUC) and odds ratios were 270.1 mg/cm3 (-2.2 SD, 66.6% of YAM), 0.689+/-0.025, 2.10 (1.63, 2.70) for BD, 104.8 mg/cm3 (-2.2 SD, 53.5% of YAM), 0.699+/-0.023, 2.17 (1.69, 2.77) for TBD and 192.8 mm3 (-1.9 SD, 59.8% of YAM), 0.631+/-0.028, 1.72 (1.34, 2.21) for pSSI, respectively. These findings suggest that ultradistal radius BMD measured using pQCT can be used to discriminate women with vertebral fractures.