【作者】 张兴；

【导师】 武阳丰；

【作者基本信息】 中国协和医科大学 ， 流行病与卫生统计学， 2007， 硕士

【摘要】 背景：房颤是一种常见的心律失常，长期发作对健康危害严重，是引起脑卒中，心衰和致残的主要原因之一。房颤由于目前尚无有效药物治疗措施且手术治疗费用昂贵治疗后尚有较高的复发率已经成为本世纪心血管疾病研究领域的热点。因此寻找可改变的房颤危险因素突显其重要意义。近来国外研究表明肥胖与房颤存在关联。然而几乎没有来自亚洲人群的报道。在人群整体BMI水平较低的我国人群是否仍有上述关联呢?本次研究进行了初步的探索。目的：初步探索中国中老年人群肥胖与房颤的关系，为房颤病因研究和预防策略制定提供依据。方法：利用国家十五科技攻关课题(非瓣膜性房颤预防血栓栓塞的研究，项目编号：2001BA703B14)2004年秋季在全国10个地区(北京石景山农民、北京市居民、山西盂县农民、江苏金坛农民、浙江舟山渔民、广西武鸣农民、陕西汉中农民、四川德阳居民、广州番禺农民及黑龙江牡丹江居民)进行的35岁以上人群心血管疾病危险因素调查的横断面资料，进行肥胖与房颤之间的关系及影响因素分析。排除体重指数不足18.5的人群后，共有18115人资料完整，进入最后分析，其中男性8029人，女性10086人。建立Logistic回归模型评价不同肥胖水平的房颤风险。房颤主要通过问卷调查和现场心电图确定，任一者为阳性者定义为房颤。房颤患者进一步接受了详细的病史调查和超声心动检查。非瓣膜性房颤定义为超声诊断显示没有瓣膜病变的房颤患者，分析时进一步剔除了手术后房颤和甲亢引起的房颤及超声和房颤病史检查缺失的患者。肥胖的测量指标有二，分别为体重指数(BMI)和腰围(WC)。体重、身高和腰围由调查员采用标准化手段测量。BMI=身高(cm)／体重(Kg)~2，BMI＜24 cm／Kg~2定义为体重正常，24 cm／Kg~2≤BMI＜28 cm／Kg~2定义为超重，28 cm／Kg~2≤BMI定义为肥胖。WC(cm)作为分组变量时分为正常、轻度升高和重度升高三组，切点在男性为85cm和95cm，在女性为80cm和90cm。结果：共发现房颤患者193人，男性89人，女性104人。房颤患病总体率呈现男性高于女性，城市高于农村的趋势，60岁以上房颤患病率明显增加。人群超重肥胖率分别为34.2％、13.1％；腰围轻度升高和重度升高率为32.6、16.0％。房颤患者的年龄、BMI、心肌梗死患病率、左室肥厚患病率、高血压患病率、糖尿病患病率、饮酒率等均显著高于非房颤患者。以往研究表明瓣膜性疾病、手术、甲亢等亦为房颤的危险因素，但是十五攻关课题仅对房颤患者进行了超声检查和详细的房颤病史调查，因此无法在模型中控制上述因素，因此考虑剔除瓣膜性房颤、手术后房颤及甲亢引起的房颤后进一步分析非瓣膜性房颤与肥胖之间的关系。男性不同病因的房颤构成比为非瓣膜性房颤79％、瓣膜性房颤~*21％；女性不同病因房颤构成比为非瓣膜性房颤65％、瓣膜性房颤~*35％。一、肥胖与房颤的关系1、BMI与房颤的定性定量关系体重正常、超重、肥胖者的房颤患病率分别为0.9％、1.1％和1.7％。房颤患病率随BMI升高而升高，体重正常组年龄调整房颤患病率为0.67，肥胖组年龄调整房颤患病率高达1.02％。男性女性年龄调整房颤患病率随BMI升高而升高，正常和超重组男性患病率高于女性。将BMI作为连续变量带入Logistic回归方程，调整年龄、地区、心梗等危险因素后，BMI值与房颤患病率显著正相关，OR值为1.09(CI：1.05-1.14)／单位。将BMI作为分类变量引入方程，以体重正常人群为参照，超重、肥胖组的OR值分别为1.21(CI：0.86-1.69)、1.88(CI：1.26-2.81)，趋势检验P均小于0.01。男性、城市及老年肥胖人群与房颤的关系更为密切。2、WC与房颤的定性定量关系腰围正常、轻度升高、重度升高者的房颤患病率分别为0.8％、1.1％、1.9％。男性腰围随年龄变化不明显，女性腰围随年龄升高。房颤患病率随腰围增加而增加，男性患病率高于女性。将腰围作为连续变量带入Logistic回归方程，调整年龄、地区、心梗等危险因素后，腰围值与房颤患病率显著正相关，男性女性OR值分别为1.03(CI：1.01-1.06)／cm和1.03(CI：1.01-1.05)。将腰围作为分类变量引入方程，以腰围正常人群为参照，调整其他因素后，男性女性腰围重度升高组的OR值分别为2.53(CI：1.41-4.56)、1.57(CI：0.95-2.59)。二、肥胖与非瓣膜性房颤~+的关系1、BMI与非瓣膜性房颤患病率的定性定量关系剔除瓣膜性房颤、手术后房颤及甲亢引起的房颤后，进一步分析BMI与非瓣膜性房颤的关系。如所预期，BMI与该类房颤关联更强。采用Logistic回归方程调整其他因素后发现，BMI每增加一单位，房颤患病风险升高12％(CI：6％-18％)。调整其他因素后，与正常体重组相比，超重和肥胖组的OR值分别为1.53(CI：0.99-2.35)、2.39(CI：1.44-3.98)。其中，女性BMI与非瓣膜性房颤的关系更明显，超重和肥胖组的OR值分别为1.86(CI：0.99-3.47)和2.53(CI：1.25-5.15)。2、WC与非瓣膜性房颤的定性定量关系剔除瓣膜性房颤、手术后房颤及甲亢引起的房颤后，WC与非瓣膜性房颤的关联也明显加强。尤其是在女性中，WC每增加1cm，非瓣膜性房颤风险升高6％(CI：3％-9％)，WC轻度升高和WC重度升高与WC正常女性相比，OR值分别为2.35(CI：1.16-4.76)和3.37(CI：1.61-7.08)。3、BMI与WC对非瓣膜性房颤的联合作用进一步分析BMI与WC对非瓣膜性房颤的联合作用。分别按BMI切点24kg／m~2、28kg／m~2；腰围切点：男性85cm、95cm，女性80cm、90cm将人群分为3*3共9组(BMI三组，WC三组)。以体重和腰围均正常组为参照，调整其他危险因素后，85／80cm≤WC＜95／90cm(男性／女性)，BMI＜24 cm／Kg~2、24 cm／Kg~2≤BMI＜28 cm／Kg~2及BMI≥28 cm／Kg~2组的OR值分别为1.12(CI：0.54-2.29)、1.83(CI：1.08-3.11)和2.28(CI：0.87-5.98)；WC＞95／90cm(男／女)，BMI＜24 cm／Kg~2、24 cm／Kg~2≤BMI＜28 cm／Kg~2及BMI≥28 cm／Kg~2组的OR值分别为3.43(CI：1.00-11.78)、2.29(CI：1.18-4.43)和2.85(CI，1.60-5.09)。而WC＜85／80cm(男／女)时，房颤患病风险与BMI的关系不明显。由上述结果可看出WC＞95／90cm(男／女)，BMI＜24 cm／Kg~2组房颤患病风险最高，其余各组基本随BMI和腰围的升高而升高。结论：在我国中老年人群中，肥胖与房颤密切相关。此结果提示肥胖可能是房颤的一个重要危险因素，干预肥胖可能起到预防房颤的作用。使用腰围和BMI联合评估肥胖与房颤的关系更为充分。更多还原

【Abstract】 BACKGROUNDRecently, results from several population-based prospective studies validate theassumption that obesity is closely associated with the risk of atrial fibrillation （AF）.Most of the studies, however, were conducted among white populations in westerncountries. Reports about Asian were limited. Furthermore, many developing countrieslike China have experienced rapid economic growth within recent decades. As anaftermath of changes in lifestyle and environment, the prevalence of obesity isincreasing at an unprecedented pace. Under such circumstances, this nationwidesurvey was conducted in fall 2004.OBJECTIVETo explore the relationship between obesity and AF, and to provide foundation foretiological studies and AF management.METHODSThe data from National 10th Five-year Plan Project（Non-valve atrail fibrillationpreventing thromboembolism, No.2001BA703B14） was used to analyze theassociation between obesity and atrial fibrillation（AF）, which was directed to explorethe association between the AF and risk factors in fall 2004. The participants aged 35and above was recruited based on China Multicenter Collaborative Study ofCardiovascular Epidemiology （China MUCA Study） from 10 populations （Beijingrural, Beijing urban, Yuxian Shanxi Province, Jintan Jiangsu Province, Zhoushan,Zhejiang Province, Wuming Guangxi Province, Hanzhong Shaanxi Province, Deyang,Sichuan Province, Panyu Guangzhou Province and Mudanjiang HeilongjiangProvince） located in various parts of China. After excluding the people whose BMIwas less than 18.5 Kg/m², a total of 18115 participants with complete informationwere valid to analyze, t test and chi square test were used to compare continuous variables and categorical variables.Multivariable Logistic regression was used toevaluate the relationship between obesity and AF.The AF was diagnosed by risk factor questionnaire or field ECG. It was defined as AFwhen either was positive. And then the AF patients performed the echocardiogram testand were ask to fill out a specific AF history questionnaire.Non-valve AF（NVAF） was defined as AF without valvular heart disease（VHD）, and inthis study we further excluded 1 Af occurred within 30 days after surgeries and 4 AFsoccurred within 12 months after incidence of hyperthyroidism, when NVAF wasanalyzed.BMI was calculated by equation weight（Kg）/height（m）2. Underweight participants（BMI＜18.5） were excluded to reduce the possibility of including individuals withcachexia. Normal-weight, overweight, and obesity were defined respectively asBMI＜24, 24≤BMI＜28, and BMI＞28; the cut-off points of WC were 85/80 cm,95/90cm for men/women, based on the latest Chinese guideline for the obesityprevention and controll 11,12. WC＜85/80cm（men/women） was defined as normal WC,85/80cm≤WC＜95/90cm modestly increased WC, 95/90cm≤WC heavily increasedWC.RESULTSThere were 193 individuals with AF out of total 18115 participants, 89 men and 104women. The prevalence of AF was higher in men than in women and higher in theurban than the rural, the prevalence of AF significantly increased in participants agedover 60 years. The prevalence of overweight and obesity were 34.2% and 13.1%; theprevalence of modestly increased WC and heavily increased WC were 32.6% and16.0%. It was higher for AF patients in age, BMI, prevalence of myocardialinfarction（MI）, the prevalence of left ventricular hypertrophy（LVH）, the prevalence ofhypertension, the prevalence of diaetes mellitus and the prevalence of use alcohol etc..But the prevalence of hypercholesteremia was lower than non-AF patients.The prior studies had shown that the VHD, surgery and hyperthyroidism were the riskfactors of AF, but the project only had the AF patient to perform the echo test and to fill out the AF history questionnaire. Thus, these factors can not be controlled inLogistic regression model. We further analyzed the assiociation between AF andobesity after excluding these factors. The cause-specific ratio of AF was NVAF 71％,VAF^?29％for men and NVAF 65％, VAF35％for women.1. The relationship between obesity and AF1) The relationship between BMI and AFThe prevalences of AF across three BMI groups were 0.9％, 1.1％and 1.7％.Age-adjusted prevalence was 1.02％in obese group. Prevalence of AF in normalweight group and overweight group were higher in men than in women. Theprevalence of AF were higher in the urban than in the rural. The prevalence of AF wasincreased sharply after 60 years. The presence rates of city, MI, hypercholesteremeia,hypertension, and diabetes increased across 3 BMI categories, while presence rate ofelectrocardiographic LVH decreased. When BMI was used as continuous variable inLogistic Regression model, it was significantly associated with risk of AF, OR was1.09（CI: 1.05-1.14）. When BMI was used as categorical variable in LogisticRegression model, ORs in overweight and obese groups were 1.21（CI: 0.86-1.69） and1.88（CI: 1.26-2.81） compared to normal weight group, P_trend＜0.01. Association inmale, urban and elder was much closer.2) The relationship between WC and AF.The prevalence of AF across three WC groups were 0.8％, 1.1％and 1.9％. There wasno association between age and WC in men, while WC was increased with age inwomen. Prevalence of AF was increased with WC, and it higher in men than inwomen. When WC was used as continuous variable in Logistic Regression model,WC was significantly associated with risk of AF. ORs in men and women were1.03（CI: 1.01-1.06） and 1.03（CI: 1.01-1.05）. When WC was used as categoricalvariable in Logistic Regression model, the ORs in male and female WC heavilyincreased groups were 2.53（CI: 1.41-4.56） and 1.57（CI: 0.95-2.59）. 2. The relationship between obesity and NVAF^§1) The relationship between BMI and NVAFAs we expected, BMI was associated with this kind of AF much more closely. Whenthis kind of AF was used as independent variable, the risk of AF was increased12％（CI: 6％-18％） with 1 unit increase of BMI after adjusted other factors. WhenBMI was used as categorical variable, the ORs in overweight and obese groups were1.53（CI: 0.99-2.35） and 2.39（CI: 1.44-3.98）. Being different from total AF, therelationship between BMI and this kind of AF in women was more significant, whoseORs were 1.86（CI: 0.99-3.47） and 2.53（CI: 1.25-5.15） in overweight and obesegroups. It was probably because prevalence of valvar heart diseases was higher inwomen, the association between obesity and AF was enhanced after excluding effectof valvar heart diseases.2) The relationship between WC and NVAFThe association between WC and this kind of AF was closer too, especially in women.The risk of this kind of AF was increased 6％（CI: 3％-9％） with increase of 1cm WC.The ORs in WC modestly increased group and WC heavily increased group were2.35（CI: 1.16-4.76） and 3.37（1.61-7.08） compared with the normal WC in women. Itprobably resulted from WC was associated with cardiovascular diseases much closer,especially in these populations with low BMI, such as Chinese. BMI does not accountfor the wide variation in body fat distribution in this kind of populations. Meanwhile,BMI may be less useful indicator of adiposity among the elderly, who tend to have ashift of fat form peripheral to central sites but no increase in BMI. WC, however,compensates for this limitation of BMI, by bringing regional fat into consideration. Sowe further analyzed the combining effect of BMI and WC.3) Combining effect of BMI and WCThe combining effect of BMI and WC was further analyzed after excluding valvar AF,postoperative AF and hyperthyroidism AF. The entire population was subdivided in to18 groups, considering 2 sexes, 3 BMI groups and 3 WC groups. There was highestAF risk in WC heavily increased and obese group in men after adjusted for other risk factors, which was 3.28（CI: 1.45-7.43）; there was highest risk of AF in WC modestlyincreased and obese group in women after adjusted for other risk factors, which was3.70（CI: 1.17-11.74）; the AF risk of total population was generally increased withBMI and WC the OR of WC heavily increased and obese group was 2.85（CI:1.60-5.09） compared with the group with normal BMI and WC. It possibly suggestedthat bringing WC into consideration was more meaningful, when evaluating therelationship between BMI and AF.CONCLUSIONObesity is closely associated with risk of AF in Chinese middle aged and elderlypopulatins. The results indicated that BMI and WC are likely the independent riskfactors of AF; bringing WC into consideration may be more appropriate whenevaluating the relationship between BMI and AF. It is probably to prevent AF bycontrolling obesity.更多还原