【作者】 周波；

【导师】 王斌；

【作者基本信息】 南京医科大学 ， 药理学， 2008， 硕士

【摘要】 对于发达国家及发展中国家来说,癌症均是导致死亡的首要因素之一。在美国年龄小于85岁人群中,癌症已经超过心脏疾病成为首要致死因素。我们前期的研究发现随着中国的人群越来越多的采用西方的饮食习惯以及高压力的生活方式,许多慢性疾病包括癌症的发病率正在逐步上升。根据国家卫生统计部门统计,在城市癌症是首要的致死因素而在农村则位居第三。癌症也给患者及他们的家庭乃至整个社会带来巨大的经济负担。许多发达国家已经开展了大量的关于癌症所带来的直接和间接医疗支出的研究。但是在中国这方面的研究还非常的缺乏。癌症是基因环境相互作用,复杂的,多因素多步骤的结果。目前许多流行病学研究以及临床研究都对癌症的危险因素进行了探索,但是由于方法学等一些因素的不同导致结果往往有很大的差异,容易使人产生困惑或误解。在这种情况下,meta分析这种定量的文献分析方法可综合这些局部、分散的研究结果从而得到更加接近真实情况的综合分析结果。第—部分:社会医疗保险对住院肿瘤患者药费的影响1950到1980年间,城市地区的卫生保健制度主要包括两方面:一个是国家保险计划(government insurance scheme,GIS)又称为公费医疗,该计划主要涵盖了公共及政府部门的工作人员同时还包括了大学及学院的学生;另一个是劳动保险计划(labor insurancescheme,LIS),主要涵盖了在国家所有制以及集体所有制的企业工作的工人。但是,由于GIS和LIS的收益人在支付医疗费用方面只需要分担很少的费用,这样又导致中国医疗费用急剧的增加。自1978年改革开放以来,中国原有的计划经济被市场经济所替代。为了在不减少医保覆盖率的基础上控制医疗卫生支出增长过快,中国进行了一系列的医疗保险制度的改革。1998年,中国政府作出了一项重大的政策决定,为城镇职工建立社会医疗保险(social healthinsurance,SHI)来替代之前的GIS和LIS,并实行社会统筹和个人帐户相结合的制度。这次改革主要目的就是使中国人民能更加有效及平等的获得医疗服务。中国政府还建立新的价格体系。这个价格体系里,绝大多数的服务项目的价格较低。但是,医院可以在药物以及检查费用上获取利润。据我们所知,在中国医疗保险对于癌症患者药费的影响的研究还非常少。这次研究的主要目的就是评价SHI对癌症住院患者药物费用的影响以及探讨影响药物费用的因素。1631名癌症患者纳入我们的研究中。癌症患者药费的中位数为RMB 8,069。在七种癌症中,药费最高的为食管癌(中位数=RMB11,028),最低的为乳腺癌(中位数=RMB 4,309)。SHI组患者用药费用(中位数=RMB 8,933)明显高于(P＜0.001)非SHI组(中位数=RMB 7,616)。根据癌症类型,在乳腺癌(P＜0.001),肺癌(P=0.018),肝癌(P=0.046),胰腺癌(P=0.040),食管癌(P=0.048)以及直肠结肠癌(P=0.004)中SHI组患者的药费均高于非SHI组。胃癌患者的药费在两组间未发现显著性差异(P=0.400)。在根据年龄中位数(59岁)、性别和入院方式分层中SHI组的药费均高于非SHI组。根据LOS分层,SHI组在住院天数≤20(P＜0.001)和20-40(P=0.011)亚组中的用药费用高于非SHI组。但是在LOS≥41(P=0.568)亚组中未发现两组间存在显著性差异。我们并没有发现SHI组和非SHI组患者在每天用药费用方面存在显著性差异(P=0.06)。经分层分析我们还发现只有在女性(P=0.016)、年轻者(P=0.023)以及住院天数在21-40亚组(P=0.003)中两组患者每天用药费用存在显著性差异。多元线性回归分析结果显示年龄较大,SHI,长LOS,严重入院方式,女性均是引起药费增高的独立因素。在这些因素中,LOS的影响最大。回归分析结果还显示年龄是影响非SHI组患者药费高低最主要的因素,而LOS则是影响SHI组患者药费高低最主要的因素。第二部分:饮茶与卵巢癌发病风险相关性的meta分析茶叶是摘自茶树的顶端叶子,在世界范围内它是仅次于水的第二大饮料。目前许多实验性研究在探索茶对癌症的作用以及作用机制。迄今为止,大多数关于茶的研究关注的都是绿茶的效应和机制。绿茶主要含有儿茶素,有效的成分为EGCG(epigallocatechingallate),ECG(epicatechin gallate),EGC(epigallocatechin)和EC(epicatechin)儿茶素。在这些成分中,EGCG受到普遍关注,目前已发现它参与诱导凋亡,抑制细胞周期,抗氧化,吞噬自由基因,抑制肿瘤入侵和转移等过程。一些体外实验和动物实验发现红茶中富含的茶黄素以及茶玉红精也有抗癌活性,如可以调节H-ras,c-Myc,p53和Bcl-2等基因蛋白的表达水平从而诱导凋亡等。最近的30多年里,关于饮茶与卵巢癌的发病风险相关性的流行病学研究有许多。但是,他们的结果并不完全一致的。而且据我们所知,到目前为止还没有一篇关于饮茶和卵巢癌发病风险的meta分析。因此,我们开展了这项meta分析,旨在进一步明确饮茶与卵巢癌发病风险的相关性。两篇前瞻性研究和7篇病例对照研究纳入meta分析。合并的风险值没有表明饮茶与患卵巢癌的风险显著相关(RR=0.84,95%CI=0.66-1.07)。敏感性分析证实了合并结果的稳定性。前瞻性研究(RR=0.69,95%CI=0.45-1.06)以及病例对照研究的合并结果(RR=0.87,95%CI=0.67-1.13)均未发现饮茶与卵巢癌的风险显著相关。将来源于人群以及来源于医院的病例对照研究分开合并,我们并没有发现来源于人群(RR=1.06,95%CI=0.88-1.27)以及来源于医院(RR=0.73,95%CI=0.46-1.15)的病例对照研究的合并结果有明显的差异。来自西方国家的8项研究的合并结果(RR=0.98,95%CI=0.84-1.15)以及来自西方国家的6项病例对照研究的合并结果都没发现饮茶与卵巢癌之间存在相关性(RR=1.04,95%CI=0.93-1.15)。西方国家的研究结果与来自中国的研究结果差异显著(P＜0.001)。异质性检验结果表明9项研究间存在显著的异质性(P＜0.001),7项病例对照研究间也存在显著的异质性(P＜0.001)。但是,前瞻性研究间并没发现异质性(P=0.181)。敏感性分析表明异质性主要来源于Zhang等人在中国开展的这项研究。这篇文章排除后,8项研究(P=0.085)以及7项病例对照研究间(P=0.193)并没发现显著的异质性。直观的漏斗图以及Egger检验(P=0.191)中均未发现发表性偏见。第三部分:激素替代治疗与卵巢癌发病风险相关性的meta分析激素替代治疗(hormone replacement treatment,HRT)是用于替代女性体内一种或多种激素,以雌激素和孕激素常见。HRT原本用于缓解绝经期综合症,后来发现它还具有心脏保护以及减少骨质疏松引起的骨折等作用从而得到了广泛的应用。但是,许多研究发现HRT会增加癌症的发病率尤其增加女性生殖系统肿瘤的发生。现有研究已证实HRT使用者中乳腺癌及子宫内膜癌的发病率比未使用HRT人群要高,而且随着使用时间的递增风险增加。但是,HRT和卵巢癌发病风险的相关性还存在很大的争议,之前的meta分析结果差异也很大。meta分析结果大多都显示HRT与卵巢癌的发病风险无关或并不显著相关,而且这些分析的结果主要根据病例对照研究。近来,有许多大样本且随访时间较长的前瞻性研究探讨HRT和卵巢癌的相关性。而且其中有研究发现HRT可显著增加卵巢癌的发病风险。因此,我们开展了这项meta分析,旨在进一步明确HRT与卵巢癌发病风险的相关性。8篇前瞻性文章合并结果显示使用HRT者患卵巢癌的风险增加了24%(95%CI=1.15-1.34),并且没有发现研究间存在异质性。此外,从直观的漏斗图以及Egger检验(P=0.496)中都没有发现发表性偏见。与未使用HRT者相比,现在使用HRT者患卵巢癌的风险显著上升(RR=1.28,95%CI=1.15-1.42)。合并结果显示曾经使用HRT与卵巢癌的发病风险未发现有显著相关性(RR=1.03,95%CI=0.95-1.13)。4篇前瞻性文章研究了雌激素替代治疗(estrogenreplacement therapy,ERT)以及雌孕激素替代治疗(estrogen plusprogestin replacement therapy,EPRT)与卵巢癌的关系。合并结果显示使用ERT者患卵巢癌的风险(RR=1.51,95%CI=1.21-1.88)要高于使用EPRT者(RR=1.24,95%CI=1.00-1.54)。19篇病例对照研究合并结果显示与未使用HRT者相比,使用HRT者患卵巢癌的风险显著上升(RR=1.19,95%CI=1.02-1.40),但是研究间存在异质性。通过逐步排除的方法,我们发现Hartge et al和Tavani et al所做的这两个研究是引起异质性的主要来源。当这两个研究被排除,其余的17个研究间不存在异质性(P=0.120),合并的风险值为1.19(95%CI=1.07-1.32)。此外,从直观的漏斗图以及Egger检验(P=0.064)中都发现存在发表性偏见。19个病例对照研究中,8个是采用人群来源的病例对照研究的合并风险值为1.17(95%CI=1.01-1.35),11个是采用医院来源的病例对照研究的合并风险值为1.21(95%CI=0.92-1.60),没有达到统计学意义。6个病例对照文章研究了不同种类HRT与卵巢癌的关系。结果显示使用ERT者患卵巢癌的风险(RR=1.19,95%CI=1.01-1.40)要高于使用EPRT者(RR=1.01,95%CI=0.83-1.22)。第四部分:吸烟与子宫内膜癌发病风险相关性的meta分析关于子宫内膜癌致病的生物学机制,长期以来高雌激素水平合并不充足的孕激素这一假设得到了许多研究者的肯定。研究发现吸烟可通过其抗雌激素效应降低子宫内膜癌的患病风险。而吸烟的抗雌激素作用可能是通过增加2-羟基雌二醇的水平来实现的,因为2-羟基雌二醇能减少细胞增生降低子宫内膜癌的患病风险。此外,吸烟还能够增加循环中雄激素水平,而雄激素能够降低雌激素所引起的细胞增生从而起到保护作用。在过去三十年中,还有许多流行病学文章研究了吸烟和子宫内膜癌发病风险的关系,但是结果至今无人进行合并。所以我们开展了这项meta分析。共10个前瞻性研究和24个病例对照研究纳入。前瞻性研究(RR=0.81,95%CI=0.74-0.88)以及病例对照研究的合并结果(RR=0.72,95%CI=0.66-0.79)均表明吸烟可显著降低患子宫内膜癌的风险。异质性检验结果发现前瞻性研究之间不存在异质性(P=0.338)而病例对照研究之间发现有异质性(P=0.021)。通过逐步排除的方法,我们发现Shields et al和Weiss et al所做的这两个研究是引起异质性的主要来源。当这两个研究被排除,其余22个病例对照研究间不存在异质性(P=0.180),合并的风险值为0.73(95%CI=0.67-0.82)。从直观的漏斗图以及Egger检验中均未发现前瞻性研究(P=0.460)以及病例对照研究(P=0.892)存在发表性偏见。前瞻性研究的合并结果(RR=0.74,95%CI=0.64-0.84)以及病例对照研究的合并结果(RR=0.63,95%CI=0.55-0.72)均发现现在吸烟能显著降低患子宫内膜癌的风险。前瞻性研究的合并结果(RR=0.88,95%CI=0.78-0.99)以及病例对照研究的合并结果(RR=0.80,95%CI=0.72-0.88)显示曾经吸烟也可显著降低患子宫内膜癌的风险。6项前瞻性研究以及6项病例对照研究纳入剂量-效应研究中。前瞻性研究的合并结果(RR=0.84,95%CI=0.71-0.99)以及病例对照研究的合并结果(RR=0.73,95%CI=0.60-0.89)都表明每天增加吸烟20只可显著降低患子宫内膜癌的风险。一些文章还研究了吸烟对绝经前以及绝经后女性患子宫内膜癌的影响。合并结果发现吸烟可显著降低绝经后女性患子宫内膜癌的风险(RR=0.71,95%CI=0.65-0.78),但是对于绝经前女性未发现显著影响(RR=1.06,95%CI=0.88-1.28)。两者的风险值差异显著(P＜0.001)。一篇前瞻性文章以及5篇病例对照文章还研究了HRT使用对于吸烟和子宫内膜癌相关性的影响。根据这些文章所提供的结果,HRT使用者中吸烟者患子宫内膜癌的风险(RR=0.45,95%CI=0.29-0.70)比未使用HRT者中的吸烟者(RR=0.65,95%CI=0.51-0.84)要低。第五部分:他莫昔芬与子宫内膜癌发病风险相关性的meta分析他莫昔芬是一种非类固醇类合成雌激素的衍生物。从20世纪70年代开始,他莫昔芬就已经广泛用于乳腺癌的治疗同时还用于单边乳腺癌患者预防对侧乳房癌变。此外,初级预防研究还发现他莫昔芬可能能够作为高风险女性预防乳腺癌的药物。使用他莫昔芬最严重的不良反应就是它具有潜在的致癌性。在女性中,他莫昔芬最常导致的就是子宫内膜癌。一些研究中发现采用他莫昔芬治疗的女性患子宫内膜癌的几率显著增加。但是每个研究的结果之间差异很大,风险增加的幅度为两到七倍不等。据我们所知,到目前为止还没有一篇关于他莫昔芬和子宫内膜癌发病风险的meta分析。因此,我们开展了这项meta分析,旨在进一步明确他莫昔芬与子宫内膜癌发病风险的相关性。14个随机对照研究(randomized controlled trials,RCTs)纳入meta分析。合并结果显示他莫昔芬可显著增加患子宫内膜癌的风险(RR=3.0,95%CI=2.30-3.91),并且没有发现研究间存在异质性(P=0.74)。此外,直观的漏斗图以及Egger检验(P=0.21)中均未发现发表性偏见。将绝经后女性单独进行分析时我们发现他莫昔芬组患子宫内膜癌的风险显著的增加(RR=5.19,95%CI=2.58-10.47)。将预防的随机试验单独进行分析时我们发现他莫昔芬能显著增加患子宫内膜癌的风险(RR=2.61,95%CI=1.89-3.60)。9个病例对照研究以及5个前瞻性研究也纳入meta分析。合并观察性文章的结果也显示他莫昔芬可显著增加患子宫内膜癌的风险(RR=1.90,95%CI=1.40-2.47),但是研究间存在异质性(P＜0.001)。直观的漏斗图以及Egger检验(P=0.88)中都没有发现发表性偏见。分析结果发现在前瞻性研究(RR=2.18,95%CI=1.48-3.21)以及在病例对照研究(RR=1.36,95%CI=1.10-1.67)中他莫昔芬都能显著增加患子宫内膜癌的风险。异质性检验结果发现前瞻性研究间不存在异质性(P=0.71)而病例对照研究间发现有异质性(P＜0.001)。根据7篇文章所提供的结果进行不同使用时间对子宫内膜癌的影响的分析。服用他莫昔芬时间少于5年者患子宫内膜癌的风险呈中等程度的增加(RR=1.75,95%CI=1.29-2.37)。服用他莫昔芬5年以上者患子宫内膜癌的风险显著的增加(RR=5.86,95%CI=3.79-9.06)。比较RCT以及观察性研究所得出的合并风险值发现两者之间存在显著性差异(P=0.02)。将RCT以及观察性研究的合并分析,结果显示他莫昔芬可显著增加患子宫内膜癌的风险(RR=2.02,95%CI=1.82-2.25)。更多还原

【Abstract】 Cancer is one of the leading causes of death in both developed and developing countries.In the United States,when deaths are aggregated by age,cancer has surpassed heart disease as the leading cause of death for those younger than age 85.As indicated previously by us,the populations in developing countries adopt Western diets and a high-stress lifestyle,the change of geographic distribution of many chronic diseases incidence has emerged,with a marked elevation in developing countries.According to national health statistics,cancer ranks as the first major cause of death among urban population and third among rural population in China.Cancer is also associated with an immense economic burden on patients,families and society.Estimation of charges and costs for cancer has been performed in many developed countries.However,little has been known about the medical expenditures for cancer in China.Cancer is considered to be a result of gene-environment interactions, which is a complex,multi-step and multi-factorial process.Many epidemiological studies and randomized controlled trials are aimed to assess the potential risk factors of cancer.However,due to the differences in the methodology,the results varied.Meta-analysis is a statistical technique for summarising,and reviewing the finding from independent studies to derive an overall estimate.By using meta-analysis,a wide variety of questions can be investigated.PartⅠ:Social health insurance and drug expenditures among cancer inpatients in ChinaBetween 1950 and 1980,in the urban areas the health care system in China traditionally comprised two major schemes:the government insurance scheme(GIS)covered workers in the public sector and government agencies as well as students in colleges and universities;and the labor insurance scheme(LIS)covered the workers in state-owned and collective-owned enterprises.However,GIS and LIS are third-party insurance,providing comprehensive benefits with minimal cost sharing to constrain beneficiaries,which contributed to China’s rapid health care cost inflation.Since 1978 when the economic reform was launched,the planned economy of China has been replaced by the market economy.Meanwhile, China announced series of health insurance system reform to control health care costs without reducing or dropping coverage.In 1998,the government made a major policy decision to establish a social health insurance(SHI)program for urban workers which replaced the existing GIS and LIS.The purposes of the reform were to ensure that the new SHI can be more efficient and more equitable in accessing health care. Furthermore,a new pricing structure was introduced.Most of the established services were priced below cost,while profits can be made from drugs and technology.And to our knowledge,little is known about the association of health insurance coverage with drug expenditures among cancer patients in China.The present study is to assess the differences in drug expenditures among cancer inpatients according to SHI status and to explore the factors that influence drug costs.One thousand six hundred and thirty one cancer inpatients were included in our study.The median drug cost was approximately RMB 8,069 per inpatient.Among these seven cancers,the median of drug cost was highest for esophageal cancer(median=RMB 11,028)and lowest for breast cancer(median=RMB 4,309).The drug cost for SHI inpatients (median=RMB 8,933)was significantly higher(P＜0.001)than that for non-SHI inpatients(median=RMB 7,616).Statistics tests for each type of cancer also showed that the differences were significant as regards breast (P＜0.001),lung(P=0.018),liver(P=0.046),pancreatic(P=0.040), esophageal(P=0.048)and colorectal(P=0.004)cancer,whereas the drug cost for gastric cancer showed no measurable difference(P=0.400).We also performed stratified analyses by the median age of all patients (59 years),LOS,sex,type of admission.In statistical analyses stratified by age,sex and type of admission,the difference between SHI and non-SHI group were significant.When stratified by LOS,the significant differences according to SHI status remained in≤20(P＜0.001)and 20-40 LOS subgroups(P=0.011),but was lost in≥41 LOS subgroup(P=0.568).We compared the drug cost per day between insured and uninsured inpatients.We did not observe a significant difference in drug cost per day between SHI and non-SHI patients(P=0.06).And the drug cost per day for female insured patients was significantly higher than that for female uninsured patients(P=0.016).The differences were also significant in younger patients(P=0.023)and in patients with 21-40 LOS subgroup(P=0.003).The results of multiple linear regression analysis demonstrated that older age,SHI coverage,longer LOS,higher severity of admission,and female sex were associated positively and independently with higher drug costs.Among these factors,LOS made the largest contribution. Furthermore,the results of multiple linear regression analysis also demonstrated that age made the most effects on higher drug costs for non-SHI patients,while LOS was the most influential factor among SHI patients.PartⅡ:The association of tea consumption with ovarian cancer risk:a meta-analysisTea,derived from the top leaves of the plant Camellia sinensis,is second only to water in terms of worldwide popularity as a beverage. Several laboratory studies found that tea may have anticancer activity. Most of tea researches to date have focused on the effect and mechanism of green tea.The main catechins,effective component in green tea,are epigallocatechin gallate(EGCG)and epicatechin gallate(ECG), epigallocatechin(EGC),and epicatechin(EC)catechins.In these components,EGCG has been the focus of a great deal of attention.It has been found to participate in induction of apoptosis and cell cycle arrest, antioxidation and scavenging the free radicals,inhibition of tumor invasion and metastasis.Extracts of black tea such as theaflavins and thearubigins also have been proved by some in vitro and animal studies to have anticarcinogenic properties,such as modulation protein expression of H-ras,c-Myc,p53,and Bcl-2 genes and induction of apoptosis.Moreover,a number of epidemiologic studies also have been conducted to investigate the association of tea consumption with cancer risk including ovarian cancer risk.However,results of these studies were not entirely consistent.And to our knowledge,there is no meta-analysis concerning association between tea consumption and ovarian cancer risk.Therefore,we undertook this meta-analysis to further clarify the association of tea consumption with ovarian cancer risk.Two cohorts and seven case-control studies were included.The summary RR of all studies,did not show that tea consumption was associated with decreased risk of ovarian cancer(RR=0.84,95% CI=0.66-1.07).We also performed a sensitivity analysis which confirmed the stability of our results.The summary RRs neither from cohort studies(RR=0.69,95%CI=0.45-1.06)nor from all case-control studies(RR=0.87,95%CI=0.67-1.13)showed that tea intake was related to decreased ovarian cancer risk.When we separated the population-based case-control studies from their hospital-based case-control studies,we found no apparent difference between hospital-based case-control studies(RR=0.73,95%CI=0.46-1.15)and population-based case-control studies(RR=1.06,95%CI=0.88-1.27).The summary RR of eight studies in all Western countries(RR=0.98, 95%CI=0.84=1.15)and the summary RR of six case-control studies in Western countries were not substantially altered(RR=1.04,95%CI= 0.93-1.15).The test for difference between study from China and studies from Western countries was statistically significant(P＜0.001).There was statistically significant heterogeneity in results across the nine studies(P＜0.001)and seven case-control studies(P＜0.001). However,results from cohort studies alone did not reject the homogeneity hypothesis(P=0.181).The sensitivity analysis showed that the study in China as contributing most to the results and heterogeneity. When this study was excluded,the test of heterogeneity was not statistically significant in results across remain eight studies(P=0.085) and six case-control studies(P=0.193).In addition,no indication of publication bias was found from both visualization of the funnel plot and the Egger’s test(P=0.191).PartⅢ:The association of hormone replacement therapy with ovarian cancer risk:a meta-analysisHormone replacement treatment(HRT)is a form of treatment designed to replace one or more female hormones,commonly estrogen and progesterone.In addition to treating the menopausal symptoms such as hot flashes and sleep disorders,HRT gained popularity because it was thought to help reduce the risk of heart disease and bone fractures caused by osteoporosis.However,a number of studies have reported that HRT could increase the risk of cancer especially cancer of female reproductive organs.The available evidence indicated that the risk of breast or endometrial cancer was increased among HRT users,the risk increasing with increasing duration of use.However,knowledge about the relation between HRT use and ovarian cancer risk is less unequivocal and results of previous meta-analyses varied.There was either no suggestion of or a weak if any association between HRT and ovarian cancer and results were mainly restricted to finding of case-control studies.Several prospective cohort studies with large sample sizes and long term follow-up were published recently, providing evidence for an increased risk of ovarian cancer in HRT user. Therefore,we undertook this meta-analysis to further clarify the association of hormone replacement therapy with ovarian cancer risk.Eight prospective studies met the inclusion criteria.The ever use of HRT was associated with a statistically significant 24%increased risk of ovarian cancer(95%CI=1.15-1.34);there was no significant heterogeneity among studies.In addition,no evidence of publication bias was observed with a nonsignificant Egger test(P=0.496).Compared with individuals who had never used HRT,those with current hormone use had a significant higher risk of ovarian cancer(RR=1.28,95% CI=1.15-1.42).Furthermore,Risk was higher among current users of more than 5 years(RR=1.47,95%CI=1.12-1.92)than among those of less than 5 years(RR=1.04,95%CI=0.91-1.20).However,for those with former HRT use,the association was not statistically significant (RR=1.03,95%CI=0.95-1.13).And we also did not find risk estimates significantly related to the duration of use for women who had former HRT use for less than 5 years(RR=1.00,95%CI=0.88-1.13),and for duration of 5 years or more(RR=1.10,95%CI=0.76-1.61).Four cohort studies addressed the association of ERT and EPRT with ovarian cancer, respectively.And a stronger association of ovarian cancer risk was noted in ERT users(RR=1.51,95%CI=1.21-1.88)than in EPRT users (RR=1.24,95%CI=1.00-1.54). Nineteen case-control studies were included.Compared with women who had never used HRT,the pooled estimate of ovarian cancer was 1.19 (95%CI=1.02-1.40)for those who ever used HRT;but there was some indication of heterogeneity among studies.By using a stepwise process, we determined that most of the heterogeneity in the ever-use summary RR was accounted for two studies by Hartge et al.and by Tavani et al. When these two studies were excluded,the remaining studies were homogenous(P=0.120),and the summary estimate for ever use of HRT and ovarian cancer incidence was 1.19(95%CI=1.07-1.32).In additional, we did obtain evidence of publication bias as shown by a statistically significant P value of 0.064 from the Egger’s test.The summary estimate was 1.17(95%CI=1.01-1.35)when we pooled eight case-control studies with population-based controls.When we combined 11 case-control studies with hospital-based controls,the association was not statistically significant(RR=1.21,95%CI=0.92-1.60).Summary RRs for the development of ovarian cancer with different type of hormone use were based on data abstract from six studies.And a stronger association of ovarian cancer risk was still noted in ERT users(RR=1.19,95%CI 1.01-1.40)than in EPRT users(RR=1.01,95%CI 0.83-1.22).PartⅣ:The association of cigarette smoking with endometrial cancer risk:a meta-analysisEndometrial cancer originates in the endometrial lining of the uterus. The major etiologic hypothesis for the development of endometrial cancer is exposure to high levels of estrogen in conjunction with inadequate progesterone.Several laboratory studies found that smoking could decrease the risk of endometrial cancer through its anti-estrogenic effect.The antiestrogenic effects of cigarette smoking might through increased level of 2-hydroxylation of estradiol which could decrease cellular proliferation and endometrial cancer risk.In addition,cigarette smoking could increase circulating androgens which might decrease estrogen-induced cellular proliferation in endometrial glands and protect against the development of endometrial cancer.During last three decades, many epidemiological studies have evaluated the association between cigarette smoking and endometrial cancer risk,and results of these studies have not yet been summarized.Ten prospective and 24 case-control studies were included.We found that ever smoking was statistically significantly associated with reduced risk of endometrial cancer among prospective(RR=0.81,95% CI=0.74-0.88)and among case-control studies(RR=0.72,95% CI=0.66-0.79);there was statistically significant heterogeneity among case-control(P=0.021)but not among prospective studies(P=0.338).By using a stepwise process,we determined that most of the heterogeneity was accounted for two studies by Shields et al.and by Weiss et al.When these two studies were excluded,the summary estimate was unchanged (RR=0.73,95%CI=0.67-0.81),but a concomitant shift in heterogeneity was measured(from P=0.02 to P=0.180).Egger’s test suggested no statistically significant asymmetry of the funnel plot for prospective (P=0.460)or cancer-control(P=0.892)studies.The RR for current smoking were 0.74(95%CI=0.64-0.84)for prospective studies and 0.63(95%CI=0.55-0.72)for case-control studies. The RR for former smoking were 0.88(95%CI=0.78-0.99)for prospective studies and 0.80(95%CI=0.72-0.88)for case-control studies.Six prospective and six case-control studies were included in this quantitative analysis.Overall,we found that an increment of 20 cigarettes per day was significantly inversely associated with endometrial cancer risk in prospective studies(RR=0.84,95%CI=0.71-0.99)or in case-control studies(RR=0.73,95%CI=0.60-0.89).Several studies have examined this association according to menopausal status.Pooled results found that there was a statistically significant reduction in endometrial cancer risk for cigarette smoking among postmenopausal women(RR=0.71,95%CI=0.65-0.78)but not among premenopausal women(RR=1.06,95%CI=0.88-1.28).The difference between two estimates was statistically significant(P＜0.01). One prospective study and five case-control studies have examined this association according to HRT status.The risk reduction appeared to be stronger among HRT users(RR=0.45,95%CI=0.29-0.70)than among none users(RR=0.65,95%CI=0.51-0.84).PartⅤ:The association of tamoxifen with endometrial cancer risk:a meta-analysisTamoxifen is a nonsteroidal triphenylethylene derivative that has been widely used to treat breast cancer since the early 1970s and that also reduces the risk of a contralateral malignancy in patients with unilateral breast cancer.Tamoxifen is being investigated in trials as a possible prophylactic agent in women at high risk of breast cancer.The most serious adverse effect of tamoxifen is its potential tumor-promoting activity.Treatment with tamoxifen has found to be associated with an increased risk of endometrial cancer,in accord with the selective uptake of tamoxifen by endometrial tissue,with its agonist effects on the endometrium,and with laboratory results.However,the magnitude of the increased risk varies substantially between studies.In these studies,the effect of tamoxifen on endometrial cancer has been reported to increase the incidence rate from two-fold to seven-fold.And to our knowledge, there is no meta-analysis concerning this association.Therefore,we undertook this meta-analysis.Fourteen randomized controlled trials(RCTs)were included. Tamoxifen use was found to be statistically significantly associated with higher risk of endometrial cancer(RR=3.0,95%CI 2.30-3.91).The Cochran’s Q test resulted in a P=0.74,indicating that the results of the RCT studies were homogeneous.The funnel plot had the expected funnel shape.The P values for the Egger test were P=0.21,suggesting a low probability of publication bias.When women who were postmenopausal or in treatment trials were considered separately,risk increases were greater(RR=5.19,95%CI 2.58-10.47).When women in risk reduction trials were considered separately,their risk increase was still statistically significant(RR=2.61,95%CI=1.89-3.60).Nine case-control studies and five cohort studies were included. Tamoxifen use was modestly significantly related to the risk of endometrial cancer among observational studies(RR=1.90,95%CI 1.40-2.47).The Cochran’s Q test resulted in a p＜0.01,indicating that the heterogeneity among the studies were high.The funnel plot also had the expected funnel shape and the P values for the Egger test were p=0.88. We found that the modest increase of endometrial cancer risk was statistically significant among the case-control studies(RR=2.18,95% CI=1.48-3.21)or among the cohort studies(RR=1.36,95% CI=1.10-1.67).We also conducted analysis based on data abstracted from seven studies to evaluate the association of endometrial cancer risk with different durations of tamoxifen use.The summary risk estimate for less than five years and were 1.75(95%CI=1.29-2.37).More than five years use of tamoxifen was found to be strongly associated with increased risk of endometrial cancer(RR=5.86,95%CI=3.79-9.06). And the difference between these two estimates was statistically significant(P＜0.01).We compared the pooled RR estimates derived from the two separate analyses.The difference between estimates was statistically significant (P=0.02).In addition,we performed a combined analysis of RCTs and observational studies.Tamoxifen use was significantly related to the risk of endometrial cancer(RR=2.02,95%CI 1.82-2.25).更多还原