【作者】 邓静；

【导师】 孙振球；

【作者基本信息】 中南大学 ， 社会医学与卫生事业管理， 2010， 博士

【摘要】 研究目的：制订长沙市区围绝经期妇女钙摄入膳食频率调查表(Food Frequency Questionnaire, FFQ)并进行信度和效度评价,为更好地研究钙摄入和绝经后骨质疏松之间的关系提供工具。研究方法：采用3天24小时回顾的方法对157名围绝经期妇女进行基线膳食调查和膳食营养分析,筛选出主要的含钙食物,加上各种季节性变异食物和钙补充剂构建钙摄入FFQ。选择基线调查中依从性好的自愿者对钙摄入FFQ的信度和效度进行评价,采用基线3天24小时回顾的结果作为参考,4周后进行重测信度调查。采用Wilcoxon符号等级检验、中位数差别比对钙摄入量进行比较,采用相关分析对问卷的重测信度和效标效度进行分析,采用四分位一致率评价FFQ和实际钙摄入分组的一致性。结果：本研究制订的钙摄入FFQ包含基线膳食调查筛选出主要含钙食物64种和钙补充剂12种共76种食物条目。相关分析显示：两次FFQ估计钙摄入Spearman’s相关系数分别为0.868(包含补充剂)和0.548(不包含补充剂),季节系数0.5调整后,两次FFQ钙摄入的Spearman’s相关系数分别为0.844(包含补充剂)和0.552(不包含补充剂)；3天24小时实际摄入钙与FFQ估计钙摄入的Spearman’s相关系数分别为0.730(包含补充剂)和0.528(不包含补充剂),季节系数0.5调整后,两者Spearman’s相关系数分别为0.764(包博士学位论文中文摘要含补充剂)和0.601(不包含补充剂),上述相关系数的P值均小于0.001。个体实际摄入钙量和本研究制订的FFQ钙摄入量之间呈直线相关关系,FFQ摄入钙(包含补充剂)与实际膳食钙摄入的差与实际膳食摄入钙呈线形关系,对于实际钙摄入量过低的的个体,本研究制订的钙摄入FFQ可能会高估个体钙摄入量。包含钙补充剂,进行季节系数为0.5调整,实际摄入钙和FFQ摄入钙(包括补充剂)分类结果的粗一致率为60.42%,Kappa系数为0.472(P=0.064)。按照季节系数0.5进行调整得到的FFQ钙摄入量(包含钙补充剂)进行四分类,其分类的标准分别为FFQ钙摄入小于505.6mg/d,大于等于505.6mg/d且652.2 mg／d,大于等于652.2 mg/d且d小于851.7mg/d,大于等于851.7mg/d。结论：1.围绝经期妇女钙摄入FFQ(包含钙补充剂)具有较好的信度和效度。2.个体实际摄入钙量和FFQ估计的钙摄入量之间呈直线相关关系,实际钙摄入量过低的的个体,本研究制订的钙摄入FFQ可能会高估个体钙摄入量。2.实际摄入钙和FFQ摄入钙(包括补充剂)分类结果的粗一致率为60.42%,Kappa系数为0.472。具有较好的区分度。4.利用本FFQ问卷对钙摄入等级进行4分类,其分类的标准分别为FFQ钙摄入小于505.6 mg/d,大于等于505.6 mg／d且652.2 mg／d,大于等于652.2 mg／d且d小于851.7mg/d,大于等于851.7mg／d。研究目的明确钙摄入、体力活动以及两者的交互作用在围绝经妇女骨丢失中的作用,为提高现有的各种防治绝经后骨质疏松症措施的效果和效益提供依据。研究方法：以湖南省妇幼保健院围绝经期门诊常规体检的妇女作为研究对象,采用美国GE公司生产的LUNAR DPX-NT型双能X射线骨密度仪测量受试者左侧髋部股骨颈,Ward三角区,大转子,髋部总体和腰部正位(L2-L4)骨密度,同时利用前期制订的钙摄入FFQ和WHO推荐的体力活动问卷对研究对象的钙摄入和日常体力活动的大小进行准确的测量和分级,10月至13月后进行骨密度重测,根据月经频率的改变情况将研究对象分成绝经过渡和未绝经组。计算骨丢失速度,采用秩和检验比较不同钙摄入等级、不同体力活动等级和月经状态之间骨丢失速度的差异。进一步采用多变量方差分析、重复测量方差分析和混合线性模型对各因素和骨丢失之间的主效应和交互关系进行分析。结果：1、189名调查对象随访期间各部位骨密度均出现不同程度的丢失,单因素分析结果显示,月经状况与各部位骨丢失速度相关(P均小于0.001)。FFQ钙摄入等级与股骨颈、腰椎和髋部总体的骨丢失速度有关(P<0.05),体力活动等级对不同部位骨丢失速度无关。2、逐步回归、多变量方差分析和重复测量分析结果均提示FFQ钙摄入等级、体力活动等级以及二者之间的交互作用与绝经过渡期骨丢失速度有关,但是,各种方法筛选的变量不完全一致,混合线性模型能同时分离出个体的随机效应和重复测量的相关,检验效能高。3、混合线性模型结果显示：股骨颈、Ward三角、大转子、髋部总体和腰椎骨密度丢失的总变异中,分别有61.90%、64.51%、55.0%、54.16%和60.71%的变异是因为个体间变异导致。4、绝经过渡期妇女各部位骨丢失速度比绝经前增加,股骨颈、Ward三角,大转子,髋部总体,腰椎骨丢失速度分别增加1.39%/yr,0.37%/yr,0.58%／yr,0.59%／yr和0.76%／yr。5、钙摄入的高低和股骨颈、Ward三角和腰椎骨丢失密切相关,与钙摄入充足(FFQ4)相比,钙摄入极低(FFQ1)组股骨颈骨丢失额外增加1.22%,同时Ward三角和腰椎骨丢失速度额外增加0.19%和0.36%。6、妇女体力活动等级与股骨颈骨密度有关,与Ward三角骨丢失速度有关,完全静坐组股骨颈骨密度比体力活动活跃组平均低0.0655,而Ward三角的骨丢失比对照额外增加0.21%。7、钙摄入等级和体力活动等级之间的交互作用主要对Ward三角、大转子和髋部总体骨丢失产生影响,在钙摄入极低的妇女中,活动不足者与活动活跃相比Wards骨丢失额外增加0.64%,髋部总骨密度骨丢失额外增加0.84%；在钙摄入不足的妇女中,活动不足者与活动活跃相比Wards骨丢失额外增加0.79%；在钙摄入略低的妇女中,静坐者与活动活跃相比Torch骨丢失速度额外增加1.67%,髋部总体骨丢失额外增加1.88%。结论：1长沙市绝经过渡期妇女各部位骨丢失速度比绝经前增加。个体间变异是导致骨密度丢失差异的主要原因。2 FFQ钙摄入等级主要和长沙围绝经期妇女股骨颈、Ward三角和腰椎骨丢失密切相关,分别导致骨丢失额外增加0.19%-0.36%；体力活动等级与股骨颈骨密度有关,且导致Ward三角骨丢失额外增加0.21%；钙摄入等级和体力活动等级对骨丢失的影响存在交互作用,主要对Ward三角、大转子和髋部总体骨丢失产生影响,其中,极低钙摄入和静坐方式对骨丢失的影响最大。3利用混合线性模型处理骨密度重测数据,通过对时间和个体效应的估计,可以改善模型对数据的拟合效果,主效应的估计更精确,参数的标准误下降,提高了分析的效率,值得在医学纵向观察数据中加以应用。更多还原

【Abstract】 Objecive:To develope Food Frequency Questionnaire (FFQ) to estimate calcium intake in perimenopausal women in Changsha City and to provide tools for further research of the relationship between dietary calcium and postmenopause osteoporosis.Methods:3 days 24-hrs recall were used in baseline dietary survey to estimate the nutritional status among 157 recruited perimenopausal women living in Changsha urban areas, calcium-rich food, a variety of seasonal variation food, and calcium supplements were selected to compose of dietary calcium FFQ. Reproducibility validation study were conducted among those seleted volunteers with good compliance in baseline dietary survey. Reproducibility of the calcium intake FFQ was assessed four weeks later and validity was assessed by comparing the FFQ result to that 3-day 24-hrs recall. Wilcoxon signed rank test, percentage of median differences, quartile agreement rate and correlation analysis, Pearson correlation coefficient and Spearman rank correlation coefficient were used to assess the reproducibility and validity of FFQ.Results:64 calcium-rich food item were selected, take 12 calcium supplements into account, a total of 76 of food items were selected to compose of dietary calcium FFQ. Result of correlation analysis showed that the Spearman’s correlation coefficient of test-retest FFQ calcium intake is 0.868 (with supplements) and 0.548 (without supplements) (P<0.001), take seasonal coefficient of 0.5 into account, the coefficient is 0.764 to 0.601 respectively(P<0.001). The Spearman’s correlation coefficient of individual 3 days 24-recall and FFQ calcium intake is 0.730 (with supplements) and 0.528(without supplements) (P<0.001), take seasonal coefficient of 0.5 into account, the coefficient is 0.764 to 0.601 respectively(P<0.001).The individual actual calcium intake and FFQ estimated calcium intake showed a linear correlation, for the lowest calcium intake individual, FFQ estimated calcium intake is overestimated.Take seasonal coefficient of 0.5 into account, the adjusted quartile agreement rate (With or without supplements) of individual actual diary intake of calcium and FFQ intake is 60.42% and Kappa is 0.472(P=0.008).Compare to 3 days 24 hours recall, the validation of calcium intake FFQ(with supplements) with the seasonal adjustment factor of 0.5 is highest, and the quartile classification criteria is, low than 505.6 mg/d(quertilel), between 505.6 mg/d～652.2 mg/d (quertile2), between 652.2 mg/d～851.7mg/d(quertile 3) and great than 851.7mg/d (quertile 4). Conclusion:1. The FFQ developed to estimate calcium intake in perimenopausal women had good reproducibility and validity.2. The individual actual calcium intake and FFQ estimated calcium intake showed a linear correlation, for the lowest calcium intake individual, FFQ estimated calcium intake is overestimated.3. The FFQ has good consistency actual diary intake of calcium, with adjusted quartile agreement rate is 60.42% and Kappa is 0.472.4. The quartile classification criteria of the calcium intake FFQ is, low than 505.6 mg/d(quertilel), between 505.6 mg/d～652.2 mg/d (quertile2), between 652.2 mg/d～851.7mg/d(quertile 3) and great than 851.7mg/d(quertile 4).OBJECTIVE:To determine the relationship between dietary calcium, physical activity and their interaction with bone loss in perimenopausal women, and to provide evidence to enhance the effects and benefits of existing postmenopausal osteoporosis prevention measures.Methods:Perimenopausal women who go to Maternal and Child Health Hospital of Hunan Province perimenopausal clinic for routine physical examination were selected as subject. The bone mineral density at Femoral Neck, Ward’s triangle, Trochanter, Total hip and Lumber 2-4 were measured using a LUNAR DPX-NT dual-energy X-ray absorptiometry bone densitometers produce by GE of the United States, the Daily calcium intake were estimated by the prior calcium FFQ and physical activity were estimated by WHO recommended physical activity questionnaire(IPAQ short form). The BMD were retest 10～13 months later. The subject were divide into premenopausal group and transition into menopause group (transition group) according to the situation of their menstrual frequency. Bone loss rate were calculated and compared between different calcium intake levels, different levels of physical activity and different menstrual status using the non-parametric test. Futher analysis of main effect of calcium intake, physical activity and their interaction on bone loss were analyzed by Multivariate analysis of variance, Repeated measures analysis of variance amd Mixed linear model.Result:1.189 subjects suffered from bone loss in all sites. Univariate analysis showed that menopausal status is related to bone loss rate of all sites(P<0.001). FFQ calcium intake levels is related to bone loss rate of Femoral neck, lumbar spine and total hip(P<0.001), the physical activity levels have no relationship with bone loss of all sites. 2. Results of stepwise regression, multivariate analysis of variance and repeated measures analysis showed FFQ calcium intake levels, physical activity levels and their interaction is related to bone loss rate of the menopausal transition period, but the determinants selected by each methods is not the same, Mixed linear model can isolate the random effects of subject and repeated time, and is a high-performance test.3. Mixed linear model results showed among the total variance of the femoral neck, Ward triangle, trochanter, total hip and lumbar spine bone mineral density loss, the variance attibuted to the individual variation is 61.90%,64.51%,55.0%,54.16% and 60.71% in respectively. The bone loss rate of all site in transition period women (transition group) is increase than in pre-menopausal, femoral neck, Ward triangle, trochanter, total hip, lumbar spine bone loss rate increased by 1.39%/yr,0.37%/yr, 0.58%/yr,0.59%/yr and 0.76%/yr.Calcium intake levels are closely related to the bone loss of femoral neck, Ward triangle and lumbar spine, compared to the adequate dietary calcium intake(FFQ4), low dietary calcium intake(FFQ1) leading to an additional 1.22% bone loss of the femoral neck, while leading to an additional 0.19% and 0.36% bone loss of Ward triangle and lumbar spine.Physical activity levels are closely related to the BMD of the femoral neck and related to bone loss of Ward triangle. The average BMD femoral neck in full sit-active group(PAO) is 0.0655 lower than that in high active physical activity group(PA3), while leading to an additional 0.21% bone loss of Ward triangle compared with the control.The interaction between calcium intake and physical activity level impact the bone loss of Ward triangle, trochanter and total hip. Very low calcium intake in women,Among women with very low dietary calcium intake(FFQ1), insufficient activity (PA1) leading to an additional 0.64% bone loss of the Wards triangle and an additional 0.84% bone loss of the total hip compared to the high active physical activity group(PA3). Among women with insufficient dietary calcium intake(FFQ2), insufficient activity (PA1) leading to an additional 0.79% bone loss of the Wards triangle compared to the high active physical activity group(PA3). Among women with slightly lower dietary calcium intake(FFQ3), full sit-active group(PAO) leading to an additional 1.67% bone loss of the Torchanter and an additional 1.88% bone loss rate of the totol hip.Conclusion:1 The bone loss rate of all site in transition period women (transition group) is increase than in pre-menopausal in Changsha. Variation between individuals is the main source leading to differences in bone mineral density loss.2 Calcium intake levels are closely related to the bone loss of femoral neck, Ward triangle and lumbar spine, leading to an additional 1.22%, 0.19% and 0.36% bone loss respectively, physical activity levels are closely related to bone loss of Ward triangle and leading to an additional 0.21% bone loss.The interaction between calcium intake and physical activity level impact the bone loss of Ward triangle, trochanter and total hip, the very low dietary calcium intake(FFQ1) and full sit-active (PAO) is of the greatest impact.3 Using mixed linear model to analyze re-measurement bone mineral density data, by estimate the random effect of individual and time, the parameter estimate of the main effect is more accurate and standard errors decreased. it can improving the model fit and improving the efficiency of the analysis, it is worthy be applied in the medical longitudinal data analysis.更多还原