【作者】 陈畅；

【导师】 刘映峰；

【作者基本信息】 南方医科大学 ， 临床医学（专业学位）， 2014， 博士

【摘要】 冠状动脉粥样硬化性心脏病简称冠心病,是严重危害人类健康的心血管疾病。其发病率和病死率在伞球范围内均居首位。虽然在近20多年来,随着社会投入的加大和治疗技术的发展,冠心病的发病率和病死率在西方工业化同家有明显下降。但随着我国经济发展和人民生活方式的改变,我国冠心病发病率却是呈上升趋势。目前冠心病被认为足遗传因素和环境因素共同作用下导致的复杂疾病。从上世纪开始,人们就开始研究ABO血型与心血管疾病的相关性。多项研究表明ABO血型与心血管疾病发病相关,非O血型人群心血管疾病发病率明显升高。随后很多流行病学研究揭示冠心病与血型之间存在着一定的联系,越来越多的证据证实非O型血的人群较O型血人群易于罹忠冠心病。但具体机制未明。低密度脂蛋白存动脉粥样硬化、冠心病的发生发展扮演着重要的角色,血浆中高胆固醇,尤其是低密度脂蛋白是冠心病危险因素。基础研究和大规模临床流行病学资料证实血胆固醇及低密度脂蛋白升高是冠心病最主要的独立危险因素之一。近年来随着他汀类药物在临床的使用从而减少了冠心病事件发生。ABO血型与总胆固醇、低密度脂蛋白的关联长期以来一直为人们所关注,早至50多年前的流行病学研究已经表明了ABO与血液中胆固醇水平相关,非O型者胆固醇水平更高。近期尤其是对ABO血型基因的研究使得他们的联系得到进一步的证实。最新的全基因组学和荟萃分析结果都支持ABO基因犁在调控血液中总胆固醇、低密度脂蛋白以及植物固醇等水平起的作用,是动脉硬化性心脏病的危险因素。虽然流行病学和基因学方面表明ABO血型和血液中脂蛋白和胆固醇水平相关,但是仍需要进一步明确相关的机制和潜在的临床治疗意义。非O型血人群,具有较高的冠心病风险,且该人群中血总胆固醇及低密度脂蛋白水平明显升高。有研究显示非O型血人群具有较高的胆固醇吸收率,而胆固醇吸收率又与冠心病的发生发展正相关。因此,有理由得岀假设：不同血型对于冠心病的发生影响有部分可能是通过影响血脂水平来发生的。但是,目前对于这一影响到底有多大程度仍不清楚。因此,我们设计了本研究课题：假设不同血型是通过影响血脂水平,从而促进冠心病的发病。通过中介效应分析方法对这一假设进行验证,并试图去估计这一中介效应的大小。本研究主要目的有：(1)明确不同血型与冠心病、心肌梗死关系；(2)不同血型间血脂的差异；(3)冠心病与血脂的关系；(4)探讨血脂作为中介因素对ABO血型致冠心病的影响,并应用中介效应分析方法对其中介作用进行量化；本研究的意义：(1)首次针对中国南方人群调查了不同血型间冠心病、心肌梗死的发病率；不同血型间血脂的的差异；进一步证实了冠心病与血脂的紧密联系；(2)综合本研究内容,结合了大量研究成果,大胆假设了非O血型人群通过升高血脂水平增加冠心病发病风险,并首次通过中介效应分析方法对这一假设进行了论证,为研究冠心病发病机制提供了一个新思路；(3)冠心病是常见病和多发病,对其发病机制和干预措施的研究哪怕是取得很小的进展,都会使很多患者受益,有巨大的社会效益。方法我们共收集了6476例患者,病例分别来自汕头大学第一附属医院,广州医科大学第一、第二附属医院,珠江医院,所有患者均进行了冠脉造影或是经皮冠状动脉成形术,患者均为中国人,本研究获得相关医院论理委员会的批准。我们收集了所有的相关资料,包括年龄、性别、ABO血型,血浆总胆固醇水平、低密度脂蛋白水平、高密度脂蛋白水平及甘油三脂水平,冠脉造影的资料,是否合并有糖尿病,测量血压水平。是否有心肌梗死的病史(包括了既往及新发的心肌梗死)。冠状动脉造影结果判断：由两名副主任以上医师进行判读,采用目测直径法,病变狭窄处直径较其近心端和远心端正常管径减少的百分率,以血管直径狭窄大于等于50%为确定冠心病的判断标准,并按照冠脉血管病变数分为0,1,2,3支血管病变,如果二者意见不一致,则南第三名副主任以上医师进行决断。统计学分析ABO血型与冠心病、心肌梗死之间的关系用回归方程进行分析,并对年龄、性别作为协变量进行校正。冠心病与非冠心病组间、心肌梗死组及非心肌梗死组、O型血及非O型血间的血脂水平区别用秩和检验。并对年龄、性别作为协变量进行校正。然后按性别区分对以上结果再次分析,并对年龄作为协变量进行校正。所有的P值均为双侧。以上方法均应用SPSS18.0软件进行分析。本研究应用了中介分析来研究血型是如何通过血脂作为一中介变量来影响冠心病的发生。在本研究中,血型被认为处理变量,总胆固醇水平、低密度脂蛋白水平或是非高密度脂蛋白水平做为中介变量。本研究的这一中介效应方法通过由整合到SPSS中的PROCESS脚本来实现。结果(一)一般情况：在6541例患者中,有效病例为6242,平均年龄为64.11,男性占了66%,AB、A、B、O血型分布为7.2%,27.1%,25.8%,39.9%,有4377患者冠脉造影提示为冠心病患者(至少一支血管直径狭窄超过50%),占总病例数的70.1%,其中有2083例患者有心肌梗死病史(包括既往及新发),占总例数的31.8。糖尿病病患者共占了22.4%。(二)血型与冠心病、心肌梗死间关系我们发现,在对年龄、性别、血脂进行较正情况下,非O型血与冠心病的发病率显著相关,OR值为1.253(95%CI=1.109-1.416,p<0.001)。我们分别将病例按性别进行分析,可得到类似的结果,不管是男性还是女性,非O型血与冠心病的发病率显著相关,男性组OR值1.269,(95%CI=1.087-1.482,p=0.003),女性组OR值1.227,(95%CI=1.004-1.500,p=0.046)。在观察血型与心肌梗死的关系中,发现非O型血患者与心肌梗死的发生史显著相关,OR值为1.221,(95%,CI=1.081-1.362,p=0.003),将病例按性别进行分析,在男性患者中存在同样的差异OR=1.223(95%,CI=1.067-1.401,p=0.003),而在女性患者组间虽有差异的倾向,但无统计学上差异,OR=1.115(95%,CI=0.882-1.401,p=0.362);(三)各血型间胆固醇、低密度脂蛋白、高密度脂蛋白、非高密度脂蛋白及甘油三脂差异；1、O型血的胆固醇水平较非O型血患者的胆固醇水平低(4.66±1.64mmol/L vs4.83±1.55mmol/L, p<0.001)。2、低密度脂蛋白也是同样的结果,O型血的低密度脂蛋白较非O型血患者的低密度脂蛋白(2.82±1.32mmol/L vs2.92±1.274mmol/L,p<0.001)。3、高密度脂蛋白水平非O型血人群较O型血人群高(1..08±0.35mmol/L vs1.09±0.39mmol/L, p=0.100).4、非高密度脂蛋白在非O型血人群中要高于O型血人群患者水平(3.54±1.48mmol/L vs3.68±1.43mmol/L,p<0.001)5、不同血型间的甘油三脂之间无差异(1..28±0.90mmol/L vs1.31±0.93mmol/L,p=0.097).6、男性亚组中各血型间胆固醇、低密度脂蛋白、高密度脂蛋白、非高密度脂蛋白及甘油三脂差异(详见表6)(1)、O型血的胆固醇水平较非O型血患者的胆固醇水平低(4.69±1.28mmol/L vs4.80±1.21mmol/L, p=0.001)。(2)、低密度脂蛋白也是同样的结果,O型血的低密度脂蛋白较非O型血患者的低密度脂蛋白低(2.88±1.06mmol/L vs2.99±1.04mmol/L,p<0.001)。(3)、不同血型间高密度脂蛋白水平无差异。(1.08±0.31mmol/L vs1.09±0.32mmol/L, p=0.609)。(4)、非高密度脂蛋白在非O型血人群中要高于O型血人群患者水平(3.72±1.13mmol/L vs3.60±1.19mmol/L, p<0.001)(5)、不同血型间的甘油三脂之间无差异(1.58±1.17mmol/L vs1.54±1.19mmol/L, p=0.056)。7、女性亚组中各血型间胆固醇、低密度脂蛋白、高密度脂蛋白、非高密度脂蛋白及甘油三脂差异(详见表7)(1)、O型血的胆固醇水平较非O型血患者的胆固醇水平低(4.96±1.20mmol/L vs5.17±1.25mmol/L, p=0.001)。(2)、O型血的低密度脂蛋白较非O型血患者的低密度脂蛋白低(2.96±0.96mmol/L vs3.15±1.04mmol/L, p<0.001)。(3)、高密度脂蛋白水平非O型血人群较O型血人群高(1.20±0.35mmol/Lvs1.17±0.32mmol/L, p=0.035)。(4)、非高密度脂蛋白在非O型血人群中要高于O型血人群患者水平(3.97±1.16mmol/L vs3.807±1.10mmol/L, p=0.003)(5)、不同血型间的甘油三脂之间无差异(1.72±1.56mmol/L vs1.72±1.56mmol/L, p=0.904)。(四)冠心病组、心肌梗死组与冠脉正常组间胆固醇、低密度脂蛋白、非高密度脂蛋白及甘油三脂的差异。1、在冠心病组患者中,其总胆固醇、低密度脂蛋白、非高密度脂蛋白水平较非冠心病组患者为高(4.92±1.29mmol/L vs4.78±1.15mmol/L, p=0.002,3.04±1.06mmol/L vs2.90±0.96mmol/L, p<0.001,3.81±1.20mmol/L vs3.63±1.06mmol/L, p<0.001)。甘油三脂两组间无差异(1.63±1.30vs1.59±1.30,p=0.144)。冠心病组中高密度脂蛋白水平较非冠心病组为低(1.11±0.32vs1.16±0.35,p<0.001)2、心肌梗死组患者中的总胆固醇、低密度脂蛋白、非高密度脂蛋白水平较非心肌梗死组患者为高(4.92±1.30mmol/L vs4.85±1.23mmol/L, p=0.031,3.08±1.06mmol/L vs2.95±1.02mmol/L, p<0.001,3.81±1.19mmol/L vs3.73±1.14mmol／L,p=0.046)。甘油三脂,高密度脂蛋白两组间无差异(1.59±1.34vs1.63±1.27,p=0.062,1.12±0.32vs1.13±0.33,p>0862)。3、男性亚组中：冠心病组、心肌梗死组与冠脉正常组间胆固醇、低密度脂蛋白、非高密度脂蛋白及甘油三脂的差异(表9)。(1)在冠心病组患者中,其总胆固醇、低密度脂蛋白、非高密度脂蛋白水平较非冠心病组患者为高(4.81±1.27mmol/L vs4.60±1.13mmol/L, p<0.001,2.99±1.06mmol/L vs2.79±0.97mmol/L, p<0.001,3.74±1.19mmol/L vs3.50±1.05mmol/L, p<0.001)。甘油三脂,高密度脂蛋白水平两组间无差异(1.56±1.15vs1.58±1.25, p=0.949;1.08±0.31vs1.11±0.33, p=0.130)(2)、心肌梗死组患者中的总胆固醇、低密度脂蛋白、非高密度脂蛋白水平较非心肌梗死组患者为高(4.88±.28mmol/L vs4.69±1.213mmol/L, p<0.001,3.07±1.06mmol/L vs2.87±1.02mmol/L, p<0.001,3.77±1.18mmol/L vs3.62±1.14mmol/L, p<0.001)。甘油三脂两组间无差异(1.53±1.16vs1.59±1.18,p=0.112)。心肌梗死组中高密度脂蛋白水平较非心肌梗死组为高(1.10±0.32vs1.08±0.31,p=0.002)。3、女性亚组中：冠心病组、心肌梗死组与冠脉止常组间胆固醇、低密度脂蛋白、非高密度脂蛋白及甘油三脂、高密度脂蛋白的差异(表10)。(1)在冠心病组患者中,其总胆固醇、低密度脂蛋白、非高密度脂蛋白水平较非冠心病组患者为高(5.18±1.31mmol/L vs4.99±1.13mmol/L, p=0.004,3.14±1.06mmol/L vs3.02±0.94mmol/L, p=0.037,4.00±1.21mmol/L vs3.77±1.04mmol/L, p<0.001)。心肌梗死组中高密度脂蛋白水平较非心肌梗死组为低(1.18±0.33vs1.22±0.36mmol/L,p=0.007)。(2)、心肌梗死组患者中的总胆固醇、低密度脂蛋白、非高密度脂蛋白水平与非心肌梗死组患者对比有增高趋势,但无统计学上差异(5.11±1.36mmol/L vs5.09±1.21mmol/L, p=0.919,3.10±1.07mmol/L vs3.08±1.00mmol/L, p=0.856,3.93±1.22mmol/L vs3.92±113mmol/L, p=0.970)。(五)血胆固醇、低密度脂蛋白、非高密度脂蛋白作为ABO血型与冠心病之间的中介变量的中介效应分析：1.在冠心病组中,冠心病作为因变量,非O型血为自变量,总胆固醇水平为中介变量,年龄、性别为协变量进行校正,结果显示非O型血较O型血人群更容易得冠心病可能,这一区别有10.09%的作用是通过升胆固醇来达到的。在同一模型中,我们将低密度脂蛋白、非高密度脂蛋白代替胆固醇进行分析,显示低密度脂蛋白、非高密度脂蛋白作为中介效应所起的作用分别为10.69%和11.63%。1.1将以上的分析按性别分别进行类似的分析,在男性组别中,总胆固醇水平为中介变量,年龄、性别为协变量进行校正,结果显示非O型血较O型血人群更容易得冠心病可能,这一区别有7.82%的作用足通过升胆固醇来达到的。在同一模型中,我们将低密度脂蛋白、非高密度脂蛋白代替胆固醇进行分析,显示低密度脂蛋白、非高密度脂蛋白作为中介效应分别为9.40%和9.48%。1.2将以上的分析按性别分别进行类似的分析,在女性组别中,总胆吲醇水平为中介变量,年龄、性别为协变量进行校正,结果显示非O型血较O型血人群更容易得冠心病可能,这一区别有13.1%%的作用是通过升胆固醇来达到的。在同一模型中,我们将低密度脂蛋白、非高密度脂蛋白代替胆固醇进行分析,显示低密度脂蛋白、非高密度脂蛋白作为中介效应分别为11.13%和13.8%。2、在心肌梗死组中,心肌梗死作为因变量,非O型血为自变量,总胆固醇水平为中介变量,年龄、性别为协变量进行校正,结果显示非O型血较O型血人群更容易得冠心病可能,这一区别有6.48%的作用是通过升胆固醇来达到的。将低密度脂蛋白、非高密度脂蛋白代替总胆固醇进行分析,显示低密度脂蛋白、非高密度脂蛋白作为中介效应所起的作用分别为10.31%和5.95%。2.1将以上的分析按性别分别进行类似的分析,在男性组别中,总胆固醇水平为中介变量,年龄、性别为协变量进行校正,结果显示非O型血较O型血人群更容易得心肌梗死可能,这一区别有5.67%的作用是通过升胆固醇来达到的。在同一模型中,我们将低密度脂蛋白、非高密度脂蛋白代替胆固醇进行分析,显示低密度脂蛋白、非高密度脂蛋白作为中介效应分别为9.21%和5.44%。2.2将以上的分析按性别分别进行类似的分析,在女性组别中,血型与心肌梗死相关不显著,不存在中介效应可能。结论非O型血人群较O型血人群具有较高的冠心病发病率；非O型血人群较O型血人群具有较高的心肌梗死发病率；非O型血人群较O型血人群有较高的胆固醇、低密度脂蛋白及非高密度脂蛋白水平；非O型血的患者较O型血患者更易得冠心病及心肌梗死的可能机制为非O型血人群由于其较高的血脂水平从而导致冠心病及心肌梗死的发生率升高,胆固醇作为中介因素起到了约10.09%,6.48%的作用,低密度脂蛋白作为中介因素起到了约10.69%,10.31%的作用,非高密度脂蛋白作为中介因素起到了约11.63%,5.95%的作用。更多还原

【Abstract】 Coronary atherosclerotic heart disease also known as coronary artery disease, is the leading cause of disability and death worldwide. During the recent20years, the mortality and morbidity declined apparently in the western countries. But with the development of social economy and changing of lifestyle, the blood pressure, serum cholesterol level and body weight index are on the rising tendency, it indicates the prevalence of CHD will gradually increasing in China.Now it is believed that coronary heart disease is caused by genetic and environment factor. The association between ABO blood type and cardiovascular disease was started since last century. Many researches elucidate the association between ABO blood type and cardiovascular disease, especially that the non-O blood group has higher risk.Many epidemiology studies revealed the association between coronary heart disease and ABO blood type, more and more evidence points that compared with O blood group, non-O blood group has higher risk of CHD. However the mechanisms for this remain unclear, and further studies are needed.Low density lipoprotein plays an important role in the atherosclerosis and coronary heart disease. High level of cholesterol especially LDL-c is an independent high risk for coronary heart disease. Large amounts of researches and epidemiology data confirm that serum cholesterol and low density lipoprotein level are independent risk factors for coronary heart disease.It has been acknowledged that ABO blood type might affect serum total cholesterol and low density lipoprotein levels. The epidemiology studies in the early1960had suggested that non-O blood group had higher levels of serum cholesterol.An association of ABO group with total cholesterol (TC) and LDL-c levels has also long been recognized, and has recently been reinforced in genetic studies of variation at the ABO locus. The recent genomics study and meta-analysis supports that ABO genotype became risk for atherosclerosis heart disease via regulation of serum TC, LDL-c and phytosterol level. Epidemiology and genetic studies have suggested the association between ABO group with serum lipoprotein and cholesterol level, however further studies are required to investigate the mechanisms and underlying clinical therapeutic significance.Specifically, the non-0types, which confer an increased risk of CHD, are associated with increased TC and LDLc levels. Furthermore, studies have shown that non-O type individuals have higher cholesterol absorption rates, and that the cholesterol absorption rate is positively associated with cardiovascular risk. Therefore we can conclude that ABO on CHD susceptibility is part mediated by its influence on cholesterol levels.In this study, we try to estimate the mediation effect size. We recruited6476individuals undergoing diagnostic or interventional coronary angiography. We first verified the associations between ABO group, cholesterol levels and CHD phenotypes, and then performed mediation analyses in a model with cholesterol levels as a mediator for the effect of ABO group on CHD phenotypes.MethodsWe studied6476consecutive patients undergoing diagnostic or interventional coronaryangiography in the First Affiliated Hospital of Shantou University Medical College, the FirstAffiliated Hospital of Guangzhou Medical College, and the Second Affiliated Hospital ofGuangzhou Medical College, in Southern China. All subjects were Chinese. The study was approved by theappropriate research ethics committee.We collected demographic and clinical data including age, gender, ABO blood type plasma levels of TC, LDL-c, high density lipoprotein cholesterol and triglyceride, coronary angiographic findings, blood pressure, and the presence or absence of diabetes mellitus. And history of myocardial infarction (including past and new onset).The coronary angiography results were read by two least association chief physicians. The criteria for CAD was diameter of stenosis>50%than the normal diameter of the proximal and distal end. And the number of involved artery was recorded as0,1,2,3. If there were disagreement about the angiography results, it would be evaluated by the third least association chief physician.Statistical analysisLogistic regression analysis was used to ascertain association of ABO blood type with CAD and MI, with adjustment for age, gender as covariates. Rank test was used to examine difference in lipid levels between CAD and non-CAD groups, between MI and non-MI group, and between O and non-0type, with adjustment for age, gender as covariates. The above analysis were then performed in male subjects and female subjects separately, with age as covariate. All P value were two side. And all above were processed with SPSS18.0software.We used mediation analysis to evaluate the mediation effect of lipid in CAD, with blood type as treatment variable, and total cholesterol, LDL-c, and NHDL-c as mediation variable. The mediation analysis was integrated to the PROCESS script in the SPSS software.Result1.Basic characteristicsOf the6542cases,6242cases were effective with mean age64.21and about66%male. The frequency of AB, A, B, O type was7.2%,27.1%,25.8%and39.9%. Of the6242subjects,4377subjects (accounts70.1%) had significant angiographically documented CAD as having>50%diameter stenosis in at least one major epicaridal coronary artery. And a total of2083(accounts31.8%) had prevalent or incident MI with22.4%diabetes mellitus patients.2.ABO blood type association with CHD and MIWe observed an association between non-O type and coronary heart disease with odd ratio for significant CAD was OR=1.253(95%CI=1.109-1.416,p<0.001).The result was similar in male OR=1.269,(95%CI=1.087-1.482,/?=0.003) and female OR=1.227,(95%CI=1.004-1.500,p=0.046)We observed an association between non-O type and prevalent or incident MI as a whole OR=1.221,(95%, CI=1.081-1.362,p=0.003) and in male OR=1.223(95%, CI=1.067-1.401,p=0.003),with a similar but non-significant trend in female OR=1.115(95%, CI=0.882-1.401, p=0.362).3.the difference of TC,LDL-c,HDL-c and non-HDL-c between ABO blood type3.1. TC levels were marginally lower in the O type individuals than non-O type individuals(4.77±1.26mmol/L vs4.93±1.23mmol/L,p<0.001).3.2. So was the case for LDL-c (2.01±1.02mmol/L in O-type vs3.04±1.04mmol/L in non-O type,p<0.001) 3.3. HDL-c levels were marginally higher(rather than lower) in the non-0type than in O-type subjects(1.13±0.33mmol/L vs1.11±0.32mmol/L,p=0.001).3.4. The levels of non-high-density lipoprotein(NHDL-c) was higher in non-O type individuals than O type individuals(3.80±1.15mmol/L vs3.67±1.17mmol/L, p<0.001).3.5. There was no significant association between ABO group and triglycerides levels (1.63±1.29mmol/L vs1.60±.33mmol/L,p=0.097)3.6. The association of ABO blood group and cholesterol, LDL-c, HDL-c, NHDL-c and TG in male subgroup.3.6.1. The O-type had lower cholesterol level than the non-0type(4.69±1.28mmol/L vs4.80±1.21mmol/L, p=0.001).3.6.2.Compared with non-O type subjects, the O type subjects had lower LDL-c level (2.88±1.06mmol/L vs2.99±1.04mmol/L,p<0.001)3.6.3. There was no association between ABO blood type and HDL-c level(1.08±0.31mmol/L vs1.09±0.32mmol/L,p=0.609).3.6.4The NHDL-c level was higher in non-0type subjects than the O-type individuals(3.72±1.13mmol/L vs3.60±1.19mmol/L,p<0.001).3.6.5There was no association between ABO blood group and TG level (1.58±1.17mmol/L vs1.54±1.19mmol/L, p=0.056)3.7. The association of ABO blood group and cholesterol, LDL-c, HDL-c, NHDL-c and TG in female subgroup.3.7.1The O-type had lower cholesterol level than the non-0type(4.96±1.20mmol/L vs5.17±1.25mmol/L,p=0.001).3.7.2Compared with non-0type subjects, the O type subjects had lower LDL-c level (2.96±0.96mmol/L vs3.15±1.04mmol/L,p<0.001)3.7.3The HDL-c level was higher in non-0type subjects than the O type subjects(1.20±0.35mmol/L vs1.17±0.32mmol/L, p=0.035).3.7.4The NHDL-c level was higher in non-O type subjects than the O-type individuals(3.97±1.16mmol/L vs3.807±1.10mmol/L,p=0.003).3.7.5There was no association between ABO blood group and TG level (1.72±1.56mmol/L vs1.72±1.56mmol/L,p=0.904)4.The difference of cholesterol, LDL-c, NHDL-c, and TG among CAD group, MI group and normal coronary group.4.1. Compared with non-CAD(coronary artery disease)group, the total cholesterol, LDL-c, NHDL-c level were higher in the CAD group (4.92±1.29mmol/L vs4.78±1.15mmol/L,p=0.002,3.04±1.06mmol/L vs2.90±0.96mmol/L,p<0.001,3.81±1.20mmol/L vs3.63±1.06mmol/L,p<0.001), but no difference on TG level (1.63±1.30vs1.59±1.30,p=0.144).And the CAD group had lower level of HDL-c (1.11±0.32vs1.16±0.35,p<0.001)4.2. Compared with the non-MI(myocardial infarction)group, the total cholesterol, LDL-c, NHDL-c level were higher in the MI group (4.92±1.30mmol/L vs4.85±1.23mmol/L, p=0.031,3.08±1.06mmol/L vs2.95±1.02mmol/L,p<0.001,3.81±1.19mmol/L vs3.73±1.14mmol/L,p=0.046).But no difference on TG and HDL-c level (1.59±1.34vs1.63±1.27,p=0.062,1.12±0.32vs1.13±0.33,p=0.862).4.3.In the male subgroup:difference of cholesterol, LDL-c, NHDL-c and TG among CAD,MI and normal coronary group,4.3.1Compared with non-CAD(coronary artery disease)group, the total cholesterol, LDL-c, NHDL-c level were all higher in the CAD group (4.81±1.27mmol/L vs4.60±1.13mmol/L, p<0.001,2.99±1.06mmol/L vs 2.79±0.97mmol/L,p<0.001,3.74±1.19mmol/L vs3.50±1.05mmol/L, p<0.001) But no difference on TG and HDL-c level (1.56±1.15vs1.58±1.25, p=0.949;1.08±0.31vs1011±0.33,=0.130)4.3.2Compared with non-MI(myocardial infarction), the MI group had higher level of total cholesterol, LDL-c and NHDL-c (4.88±1.28mmol/L vs4.69±1.213mmol/L,p<0.001,3.07±1.06mmol/L vs2.87±1.02mmol/L, p<0.001,3.77±1.18mmol/L vs3.62±1.14mmol/L, p<0.001).But no difference on TG level (1.53±1.16vs1.59±1.18, p=0.112).However, the HDL-c was higher in the MI (1.10±0.32vs1.08±0.31, p=0.002)4.4In the female subgroup:difference of cholesterol, LDL-c, NHDL-c and TG among CAD, MI and normal coronary group4.4.1. Compared with non-CAD group, the CAD group had higher levelof total cholesterol, LDL-c, NHDL-c (5.18±1.31mmol/L vs4.99±1.13mmol/L, p=0.004,3.14±1.06mmol/L vs3.02±0.94mmol/L, p=0.037,4.00±1.21mmol/L vs3.77±1.04mmol/L, p<0.001). And the MI group had lower HDL-c level than the non-MI group.4.4.2. Compared with non-MI, the MI group had higher level of total cholesterol, LDL-c, NHDL-c, but without stastiscal significance (5.11±1.36mmol/L vs5.09±1.21mmol/L,/7=0.919,3.10±1.07mmol/L vs3.08±1.00mmol/L, p=0.856,3.93±1.22mmol/L vs3.92±113mmol/L,p=0.970)5.A mediation analysis the association of ABO blood group with CAD, with serum cholesterol, LDL-c, NHDL-c level as the mediator variable5.1In the a model with CAD as the dependent variable, non-0versus O type as the independent variable, total cholesterol as the mediator variable, age and gender as covariates. And a mediation analysis estimated that10.09%of the difference was mediated by increased cholesterol level.In the same model, replacing cholesterol with LDL-c NHDL-c, and the mediation effect size was10.69%and11.63%respectively.5.2In a model with MI as the dependent variable, non-O versus O type as the independent variable, total cholesterol as the mediator variable, with age, gender as covariate, and a mediation analysis showed that non-0type was susceptible to coronary artery disease. And about6.48%of the difference was mediated by cholesterol.Substituting cholesterol with LDL-c or NHDL-c as the mediator variable in this model showed mediation effect size about10.31%and5.95%respecitively.5.3.When the mediation analysis were performed in males and females separately, in the male group the mediation effect size of LDL-c, total cholesterol, NHDL-c for CAD or MI were similar. But not in female group.conclusionNon-0type individuals have higher incidence of CHD than O-typeNon-O type individuals have higher mortality of myocardial infarction than O typeCompared with O type individuals, the non-O type have higher level of serum cholesterol, LDL-c and non-HDL-c.The underlying mechanism for non-O types have higher susceptibility is might attributed to cholesterollevels as mediators with about10.08%of the effect of non-0type on CAD and MI susceptibility.更多还原