Department of Physical Therapy, Faculty of Health and Science, Kio University, Nara, Japan.
J Clin Densitom. 2012 Jul-Sep;15(3):343-50. doi: 10.1016/j.jocd.2012.02.001. Epub 2012 Jun 6.
Screening for low bone mass is important to prevent fragility fractures in men as well as women, although men show a much lower prevalence of osteoporosis than women. The purpose of this study was to establish a screening model for low bone mineral density (BMD) using a quantitative ultrasound parameter and easily obtained objective indices for elderly Japanese men. We examined 1633 men (65-84 yr old) who were subjects of the Fujiwara-Kyo Study. Speed of sound (SOS) at the calcaneus was determined, and BMD was measured by dual-energy X-ray absorptiometry at the lumbar spine (LS), total hip (TH), and femoral neck (FN). Low BMD was defined as >1 standard deviation below the young adult mean, in accordance with World Health Organization criteria. We performed receiver operating characteristic (ROC) analysis to identify a better screening model incorporating SOS and determined the optimal cutoff value using Youden index. Prevalences of low BMD at the 3 skeletal sites were 27.8% (LS), 33.5% (TH), 48.6% (FN), and 43.3% at either LS or TH. The greatest area under the ROC curve (0.806, 95% confidence interval: 0.785-0.828) and smallest Akaike's information criterion were obtained in the multivariate model incorporating SOS, age, height, and weight for predicting low BMD at all skeletal sites. This model predicted low BMD at TH with the sensitivity of 0.726 and specificity of 0.739, whereas a similar model predicted low BMD at LS with much lower validity. We conclude that the multivariate model for TH could be used to screen for low BMD in elderly Japanese men.
对男性和女性来说,筛查低骨量以预防脆性骨折都很重要,尽管男性骨质疏松症的患病率明显低于女性。本研究的目的是建立一个使用定量超声参数和易于获得的客观指标来筛查老年日本男性低骨密度(BMD)的模型。我们检查了 1633 名(65-84 岁)富士wara-kyo 研究的男性受试者。在跟骨处测定声速(SOS),并通过双能 X 射线吸收法测量腰椎(LS)、全髋(TH)和股骨颈(FN)的 BMD。根据世界卫生组织的标准,将 BMD 定义为低于年轻成年人平均值 1 个标准差以上。我们进行了接收者操作特征(ROC)分析,以确定更好的包含 SOS 的筛查模型,并使用 Youden 指数确定最佳截断值。3 个骨骼部位的低 BMD 患病率分别为 27.8%(LS)、33.5%(TH)、48.6%(FN)和 43.3%(LS 或 TH)。在预测所有骨骼部位低 BMD 的多变量模型中,ROC 曲线下面积最大(0.806,95%置信区间:0.785-0.828),Akaike 信息准则最小,该模型包含 SOS、年龄、身高和体重。该模型预测 TH 低 BMD 的灵敏度为 0.726,特异性为 0.739,而类似的模型预测 LS 低 BMD 的准确性要低得多。我们得出结论,TH 的多变量模型可用于筛查老年日本男性的低 BMD。