中国成人腹型肥胖与糖尿病发病关系的前瞻性队列研究

【作者】 薛海峰；

【导师】 顾东风；

【作者基本信息】 北京协和医学院 ， 流行病与卫生统计学， 2014， 博士

【摘要】 第一部分中国成人糖尿病发病率及腹型肥胖所致糖尿病发病人数背景和目的腹型肥胖和糖尿病已经成为世界上主要的公共卫生问题。本研究拟估计中国男性和女性成人糖尿病发病率和腹型肥胖所致糖尿病发病人数。研究对象和方法本研究在25025名35-74岁无糖尿病的中国成人中进行。1998年和2000-2001年基线调查时,按照统一方法测量腰围(Waist circumference, WC)等人体测量指标,并收集社会人口统计学、疾病史和生活方式信息。腹型肥胖定义为：男性腰围≥90cm,女性腰围≥80cm。2007-2008年随访期间,收集糖尿病患病、使用胰岛素和口服降糖药以及死亡信息。对于未诊断为糖尿病的对象,以空腹血糖水平≥7.0mmol/L,确定是否患有糖尿病。采用Cox回归模型定量估计腹型肥胖所致糖尿病发病的风险比(Hazard ratio, HR)和95%置信区间(Confidence interval, CI)。利用多因素调整的人群归因危险度(Population attributable risk)、糖尿病发病率和2010年中国人口普查数据,估计腹型肥胖所致糖尿病发病人数。结果平均随访8.0年,35-74岁中国成人的年龄标化糖尿病发病率男性与女性分别为9.6/1000人年和9.2/1000人年。腰围每增加1cm时,男性与女性多因素调整的糖尿病发病的HR(95%CI)均为1.06(1.05,1.07)。男性以腰围<90cm组为参照,腰围90-94cm、95-99cm和≥100cm组多因素调整的糖尿病发病的HR(95%CI)分别为：2.13(1.68,2.70)、2.71(2.01,3.65)和3.33(2.41,4.59)(P趋势<0.001)；女性以腰围<80cm组为参照,腰围80-84cm、85-89cm和≥90cm组多因素调整的糖尿病发病的HR(95%CI)分别为：1.77(1.38,2.26)、2.60(2.04,3.31)和3.36(2.66,4.25)(P趋势<0.001)。在男性和女性中腹型肥胖分别解释28.1%(PAR)(95%CI：14.8%,40.5%)和41.2%(95%CI：28.3%,52.6%)的新发糖尿病。估计2010年,腹型肥胖所致糖尿病发病人数(95%CI)为240(150,320)万。其中,男性为100(50,140)万,女性为140(100,180)万。结论腹型肥胖是糖尿病的主要危险因素。加强不良生活方式干预,倡导预防和控制腹型肥胖,应该是减轻中国如此巨大糖尿病负担的首选国家策略。第二部分中国成人不同腹型肥胖指标预测糖尿病发病的作用比较背景和目的肥胖,尤其是腹型肥胖是心血管疾病和糖尿病的重要危险因素。然而,不同腹型肥胖指标——腰围、腰臀比和腰高比——预测糖尿病发病的作用,仍有争议。因此,本研究旨在通过大样本队列人群资料,在中国成人中进行了腰围、腰臀比和腰高比预测糖尿病发病的作用比较。研究对象和方法本研究在25025名35-74岁无糖尿病的中国成人中进行。1998年和2000-2001年进行基线调查,测量身高、腰围和臀围等人体测量指标,并收集社会人口统计学、疾病史和生活方式信息。腰臀比和腰高比分别计算为腰围(cm)/臀围(cm)和腰围(cm)/身高(cm)。2007-2008年对基线人群进行随访研究,收集糖尿病患病、使用胰岛素和口服降糖药以及死亡信息。对于未诊断为糖尿病的对象,以空腹血糖水平≥7.0mmol/L,确定是否患有糖尿病。采用Cox回归模型定量估计腹型肥胖所致糖尿病发病的风险比(Hazard ratio, HR)和95%置信区间(Confidence interval,CI)。利用受试者工作特征(Receiver operating characteristic, ROC)曲线比较不同腹型肥胖指标预测糖尿病发病的作用,并估计各腹型肥胖指标预测糖尿病发病的适宜切点值及其相应灵敏度和特异度。结果平均随访8.0年,共1101例新发糖尿病患者(男性537例,女性564例)。不论男性和女性,腰围、腰臀比和腰高比均能有效预测糖尿病发病风险,并与糖尿病发病风险呈剂量反应关系(P趋势<0.001)。腰围、腰臀比和腰高比每增加一个标准差时,男性多因素调整的糖尿病发病的HR(95%CI)分别为：1.36(1.20,1.54)1.24(1.14,1.33)和1.45(1.24,1.69)；女性多因素调整的糖尿病发病的HR(95%CI)分别为：1.30(1.17,1.45)、1.17(1.07,1.27)和1.27(1.14,1.42)。男性腰围、腰臀比和腰高比预测糖尿病发病的ROC曲线下面积(95%CI)分别为：0.69(0.67,0.72)、0.68(0.66,0.70)和0.70(0.67,0.72)；女性的ROC曲线下面积(95%CI)分别为：0.72(0.70,0.74)、0.67(0.64,0.69)和0.72(0.70,0.74)。男性腰围、腰臀比和腰高比的ROC曲线下面积相似,而女性腰围和腰高比的ROC曲线下面积均大于腰臀比(P<0.001)。腰围、腰臀比和腰高比预测糖尿病发病的最佳切点值、灵敏度和特异度,男性分别为：84.9cm、61.3%和69.3%,0.87、73.2%和55.5%,0.50、67.6%和63.3%；女性分别为：77.0cm、75.4%和57.8%,0.84、56.9%和67.9%,0.51、67.7%和64.9%。结论三个腹型肥胖指标——腰围、腰臀比和腰高比——都是糖尿病发病风险的强预测指标。在女性中腰围和腰高比预测糖尿病发病的作用优于腰臀比。腰围预测糖尿病发病的最佳切点值：男性为84.9cm,女性为77.0cm。腰围可能更适用于临床和公共卫生实践中用来指导防治腹型肥胖和糖尿病。更多还原

【Abstract】 Part One:Incidence of diabetes and association between abdominal obesity and risk of diabetes among Chinese adults:a prospective cohort studyBackground and ObjectiveAbdominal obesity and diabetes have become major public health problems around the world. We aimed to estimate the incidence of diabetes and the number of incident cases attributable to abdominal obesity among Chinese men and women.Design and MethodsWe conducted a prospective cohort study in a sample of25025Chinese adults free of diabetes, aged35-74years. Waist circumference (WC) and other anthropometric measurements were measured using the same method, and information on sociodemography, medical history and life style was collected at baseline in1998and2000-2001. Abdominal obesity was defined as WC>90/80cm (men/women). Follow-up examination was carried out in2007-2008. To ascertain newly developed diabetes, information on diabetes status, use of insulin and oral antidiabetic drugs was collected, and fasting blood was drawn for glucose measurement. Cox regression was used to estimate association between abdominal obesity and risk of diabetes. The number of diabetes cases attributed to abdominal obesity was evaluated by using multivariable-adjusted population-attributable risk, incidence of diabetes and the population size of China in2010.ResultsDuring a mean follow-up of8.0years, age-standardized incidence of diabetes was9.6and9.2per1000person-years in Chinese men and women aged35-74years, respectively. Multivariate-adjusted HR (95%CI) for diabetes per centimeter increase in WC was1.06(1.05,1.07) in both men and women. With the group of WC<90cm as reference group, HR (95%CI) of the group of WC:90-94cm,95-99cm and≥100cm for diabetes was2.13(1.68,2.70),2.71(2.01,3.65) and3.33(2.41,4.59) in men, respectively (P for trend <0.001); With the group of WC<80cm as reference group, HR (95%CI) of the group of WC:80-84cm,85-89cm and≥90cm for diabetes was1.77(1.38,2.26),2.60(2.04,3.31) and3.36(2.66,4.25) in women, respectively (P for trend<0.001). Abdominal obesity accounted for28.1%(95%confidence interval [CI]:14.8%,40.5%) of incident diabetes among men and41.2%(28.3%,52.6%) among women. It was estimated that in2010,2.4(1.5,3.2) million incident diabetes were attributable to abdominal obesity:1.0(0.5,1.4) million in men and1.4(1.0,1.8) million in women.ConclusionsAbdominal obesity is a major risk factor for diabetes. Strengthening programs and initiatives for preventing obesity focusing lifestyle changes should be a priority in the national strategy to reduce such huge diabetes burden in China. Part Two:Comparison between different measures for abdominal obestiy in predicting risk of diabetes among Chinese adultsBackground and ObjectiveObesity, especially abdominal obesity, is a strong risk factor for diabetes and cardiovascular diseases. However, it remains controversial which can better predict diabetes risk among waist circumference (WC), waist-hip ratio (WHR) or waist-height ratio (WHtR). The aim of this nanlysis is to compare and examine WC, WHR and WHtR in predicting diabetes in Chinese adults.Design and MethodsWe conducted a prospective cohort study in a sample of25025Chinese adults free of diabetes, aged35-74years. WC, hip circumference, height and other anthropometric measures were measured, and information on sociodemography, medical history and life style was also collected at baseline in1998and2000-2001. WHR and WHtR was calculated as WC (cm)/hip circumference (cm) and WC (cm)/height (cm), respectively. Follow-up examination was carried out in2007-2008. To ascertain newly developed diabetes, information on diabetes status, use of insulin and oral antidiabetic drugs was collected, and fasting blood was drawn for glucose measurement. Cox regression was used to estimate association between abdominal obesity and risk of diabetes. Predictive power of different abdominal obesity for diabetes was compared by receiver operating characteristic (ROC) curve. We also evaluated appropriate cut-off values of different abdominal obesity and their sensitivity and specificity.ResultsDuring a mean follow-up of8.0years,1101cases of diabetes (537in men and564in women) occurred. WC, WHR and WHtR all predicted future diabetes and there was a dose-response association between WC, WHR and WHtR and diabetes risk (P for trend<0.001). The multivariable-adjusted HR (95%CI) for diabetes per one standard deviation increase in WC, WHR and WHtR was1.36(1.20,1.54),1.24(1.14,1.33) and1.45(1.24,1.69) in men; and1.30(1.17,1.45),1.17(1.07,1.27) and1.27(1.14,1.42) in women, respectively. In men, there was no significant differences between area under ROC curve (AUC) of WC, WHR, and WHtR in predicting DM. In women, there was statistical evidence that WC (AUC:0.72;95%CI:0.70,0.74) and WHtR (AUC:0.72;95%CI:0.70,0.74) were stronger predictors than WHR (AUC:0.67;95%CI:0.64,0.69)(P<0.001). The optimal cutoff value, sensitivity and specificity of WC were84.9cm,61.3%and69.3%,0.87,73.2%and55.5%for WHR, and0.50,67.6%and63.3%for WHtR in men, respectively; and77.0cm,75.4%and57.8%for WC,0.84,56.9%and67.9%for WHR,, and0.51,67.7%and64.9%for WHtR in women, respectively.ConclusionWC, WHR and WHtR all are strong predictors of abdominal obesity for diabetes risk in Chinese adluts. In women, WC and WHtR are stronger predictors than WHR. The appropriate cutoff value of WC is84.9and77.0cm in men and women, respectively. WC is likely a better predictor of diabetes in clinical and public health practice to control and prevent abdominal obesity and diabetes.更多还原