女性人群骨密度分布描述及其预测模型的研究

目的通过了解农村地区成年女性人群骨矿物质密度随年龄的变化趋势及年龄、身高、体重、绝经年限对骨密度的独立作用,建立骨密度的预测模型。方法根据统一的问卷收集了3008对农村女性双生子的人口学特征及环境资料,利用美国Norland公司生产的外周双能X光吸收骨密度仪(pDXA)测定腰椎、髋部及前臂近端桡骨骨密度。采用SAS6·12软件包及SPLUS进行数据分析。结果研究的结局变量为腰椎、股骨颈、股骨三角区、大转子、手臂骨及全身骨6个部位的骨密度。不同部位的骨密度值在年龄上的分布趋势类似,但它们达到骨峰值的年龄及骨丢失上存在着区别:以小梁骨为主要成份的骨密度达骨峰值年龄较早,并先出现骨丢失,骨丢失率较高;年龄、绝经年限、身高、体重在整个年龄段中对骨密度的独立作用不呈线性关系;将人群分成3个年龄段,建立年龄、绝经年限、身高及体重对各部位的预测模型,经检验,这些预测模型所获得的估计值与实际值较为接近。结论预测模型能较好地利用一些易测量的指标来预测个体的骨密度值,为中国农村地区女性人群的骨质疏松症诊断提供了简便的手段。

Description of bone mineral density distribution in women and the research of epidemiology forecast model of bone mineral density in women

ObjectiveTo explore the distribution of bone mineral density(BMD) with age and to depict independent effects of age, year since menopause, height and weight on BMD and then to develop predictive model based on such determinants. Methods 3008 pair of female twins were enrolled and their sociodemographic and envionmental data were collected with a standard questionnaire. Dual-energy X ray absorptionmetry (DPX) was used to measure subjects' BMD of all bones. Statistic analysis were done by applying SAS 6.12 and SPLUS software. Results The outcome variables studied here included BMDs of second to forth lumbar spine, total spine, femoral neck, wards'triangle, trochanter, arm and total body. Our results showed that BMDs of all sites changed with age similarly, but both the age when peak bone mass reached and the bone mass loss varied among different site. Those made of trabecular mainly reached peak bone mass earlier and began to bone mass loss earlier,and the loss rates were higher.BMD values was affected by age, years since menopause, height and weight, but the independent effect of each variables was nonlinear over the whole age range studied here; All subjects were classfied into three groups according to their age, and predictive models for each bone site were developed. These models were tested and found that the predictive values of bone mass density, obtained by these predictive model, were similar to their actual values.Conclusion Our result showed that BMD values can be predicted by some easily measure variables based on the predictive models provided in this study, which provides a simple method to diagnose Osteoporosis for rural female.