酒精、咖啡、绿茶和乳制品的摄入量与前列腺癌的关联：两样本孟德尔随机化研究.pdf

现代泌尿外科杂志前列腺癌专题酒精、咖啡、绿茶和乳制品的摄人量与前列腺癌的关联：两样本孟德尔随机化研究柴红强，邵晋凯，吴明，赵克洁，庞磊（山西医科大学第五临床医学院，山西省人民医院泌尿外科，山西太原0 30 0 12）Associations of alcohol,coffee,green tea and dairy products with prostate cancer:atwo-sample Mendelian randomization studyCHAI Hongqiang,SHAO Jinkai,WU Ming,ZHAO Kejie,PANG Lei(Department of Urology,The Fifth Clinical Medical College of Shanxi Medical University,Shanxi ProvincialPeoples Hospital,Taiyuan 030012,China)ABSTRACT:Objective A two-sample Mendelian randomization method was used to explore whether there is a causalrelationship between the intake of alcohol,coffee,green tea and dairy products and the incidence of prostate cancer(PCa),inorder to clarify the risk factors for the incidence of PCa and find a prevention pathway for PCa.Methods Data of alcohol,coffee,green tea,dairy products and prostate cancer were collected with genome-wide association study(GWAS).The causalrelationship between their intake and the risk of PCa was analyzed with two-sample Mendelian randomization(2SMR).MRanalysis was conducted with inverse-variance weighting(IVW).Sensitivity analysis was performed with weighted median,MR-Egger regression,Mendelian randomization pleiotropy residual sum and outlier(MR-PRESSO)tests.Results Coffee intake(OR:0.994,95%CI:0.990-0.999,P=0.014)and green tea intake(OR:0.999,95%CI:0.998-0.999,P=0.036)werenegatively correlated with the risk of PCa.Alcohol intake(OR:0.997,95%CI:0.990-1.004,P=0.392)and dairy intake(OR:1.025,95%CI:0.983-1.069,P=0.256)were not associated with the risk of PCa.In weighted median,MR-Eggerregression,and retention one method analyses,the results were robust without heterogeneity or pleiotropy.ConclusionThere was a causal association between coffee intake and green tea intake and the onset of PCa,but no causal associationbetween alcohol intake and dairy intake and PCa onset.KEY WORDS:Mendelian randomization;prostate cancer;causal inference;genome-wide association study;alcohol;coffee;green tea;dairy products摘要：目的采用两样本孟德尔随机化方法探究酒精、咖啡、绿茶和乳制品的摄入量与前列腺癌（PCa)发病之间是否存在因果关系，以期帮助明确PCa的发病风险因素,从而帮助寻求 PCa的预防途径。方法采用全基因组关联研究(GWAS)收集酒精、咖啡、绿茶、乳制品和PCa的数据，采用两样本孟德尔随机化（2 SMR)方法探讨上述饮品摄入量与PCa发病风险的因果关联。逆方差加权法（IVW）是本次孟德尔随机化（MR)分析的主要方法。敏感性分析采用加权中位数、MR-Egger回归、MR多效性残差和离群值（MR-PRESSO）检验。结果咖啡摄人量（OR：0.9 9 4，9 5%CI:0.9 9 0 0.9 9 9，P=0.0 14）、绿茶摄人量（OR：0.9 9 9，95%CI:0.9980.999,P=0.036)与PCa发病呈负向因果效应。酒精摄人量（OR：0.9 9 7,9 5%CI:0.9 9 0 1.0 0 4,P=0.39 2）和乳制品摄入量（OR：1.0 2 5,9 5%C I:0.9 8 3 1.0 6 9,P=0.2 56）与PCa发病风险无因果关联。在加权中位数、MR-Egger回归和留一法分析中,结果均稳健，无异质性和多效性。结论咖啡摄入量和绿茶摄入量与PCa发病存在因果关联，而酒精摄人量与乳制品摄入量和PCa发病则没有因果关联。关键词：孟德尔随机化法；前列腺癌；因果推断；全基因组关联研究；酒精；咖啡；绿茶；乳制品中图分类号：R679根据世界卫生组织2 0 2 0 年全球癌症统计报告，2020 年全球新发前列腺癌（prostate cancer，PC a）病例约14142 59 例，是男性癌症死亡的第2 大常见类型收稿日期：2 0 2 3-0 7-31基金项目：山西省卫生健康委科研课题项目（No.2022073）通信作者：庞磊，副主任医师。E-mail：p a n g l e i _3511 s i n a.c o m作者简介：柴红强，硕士研究生。研究方向：泌尿系肿瘤。E-mail:2024年3月第2 9 卷第3期文献标志码：A修回日期：2 0 2 3-12-0 4http:/jmurology.xjtu.edu,cn;zgmnwk.cug.top253D0l:10.3969/j.issn.1009-8291.2024.03.011和第5大主要原因。与西方国家相比，亚洲国家PCa的发病率较低2-3。在中国2 0 2 0 年PCa的发病率为10.2/10 万，而北欧的发病率为8 3.4/10 万1。在死亡率方面，中国在全球排名也较低。2 0 2 0 年中国的死亡率为4.6%，而加勒比地区的死亡率最高为27.9%1。尽管中国的发病率和死亡率都较低，但由于人口众多，中国占全球PCa新发病例的8.2%和死254亡病例的13.6%。因此，中国应对PCa给予更多的关注，应加大公共卫生方面的管控，预防PCa的发生。明确PCa的病因、可改变的危险因素对于实施预防干预措施至关重要。流行病学研究已经调查了常见饮品（酒精-6、咖啡7-1、绿茶10-11、乳制品1-141）与PCa 发病风险之间的因果关联，研究结果却并不一致，由于混杂和反向因果关系的限制，无法确定其因果关系。随机对照试验可以得出较可靠的因果关系推断，然而，癌症在暴露于危险因素和疾病的临床诊断之间可能有很长的滞后时间，随机对照试验可能昂贵、耗时且不可行。孟德尔随机化（mendelian randomization，M R）分析近年来已被广泛用于评估各种疾病的潜在因果关系。由于遗传变异在减数分裂中是随机分配的，并且基因型是不变的,因此MR分析可以克服混杂因素和反向因果关联对因果推断的影响15。全基因组关联研究（genome-wide associationstudy,GWAS)是一种研究基因突变和表型之间相关X独立性假设J Mod Urol,Vol.29 No.3 Mar.2024性的方法C16。G W A S数据库汇总了来自GWAS的基因型和表型关联。本研究利用公开的数据，使用两样本孟德尔随机化（two-sampleMendelianrandomization，2 SM R)分析评估了酒精、咖啡、绿茶和乳制品等常见饮品摄人量与PCa发病风险之间的潜在因果关联,以期明确PCa的发病风险因素,降低PCa 的发病率。1资料与方法1.1研究设计及数据来源本研究采用2 SMR分析评估酒精、咖啡、绿茶和乳制品4种常见饮品摄人量与 PCa发病风险之间的因果关系。本研究需满足以下3个假设：相关性假设：所选的工具变量与暴露之间强相关；独立性假设：所选的工具变量与暴露和结局之间的任何混杂因素均不相关；排除限制假设：工具变量不会直接影响结果，只能通过暴露影响结局(图 1)。-混杂因素1工具变量相关性假设暴露酒精摄入量咖啡摄入量绿茶摄入量乳制品摄入量前列腺癌风险结局排除限制假设图1孟德尔随机化假设示意图本文从IEU数据库中（https:/gwas.mrcieu.ac.uk/)获得了最新和最大样本量的GWAS暴露汇总数据，包括酒精摄人量（GWASID：ie u-b-48 34）、咖啡摄人量（GWASID：u k b-b-52 37）、绿茶摄入量（G W A SI D：u k b-b-40 7 8）、乳制品摄人量（GWASID：u k b-b-2 39 3），所选的研究人群均为欧洲人。GWAS结局数据集同样也来自于IEU数据库,PCa发病风险的GWAS数据包括30 9 45例患者和368725例对照组，研究人群也为欧洲人。1.2工具变量的选择工具变量的筛选标准如下：具有全基因组显著性(P 0.0 5（表2），因此，酒精、咖啡、绿茶和乳制品的摄人量与PCa发病风险之间的因果关联不存在水平多效性。本研究中基于IVW法和MR-Egger法的CochranesQ的P0.05（表2),说明上述因果研究均不存在异质性。酒精、咖啡、绿茶和乳制品的漏斗图也表明工150-140-110-100-200-J Mod Urol,Vol.29 No.3 Mar.2024具变量的因果效应估计值分布基本对称，整体不存在异质性(图3)。此外，本研究还采用留一法对SNPs做逐一剔除检验，各个SNPs的效应估计差异无统计学意义，证明了本研究结果的稳健性（图4)。逆方差加权法逆方差加权法MR-Egger方法MR-Egger方法16-AS/14-12-10-8-0.02-0.01PIv逆方差加权法MR-Egger方法0-0.11160-0Prv逆方差加权法MR-Egger方法0.1100-1100-50-rs115616389rs142896482rs77089642rs142080835rs116917149rs606308SSrsTSrsrs4410790rs142All-0.012图4酒精、乳制品、咖啡及绿茶摄入量与前列腺癌发病风险关联的孟德尔随机化分析留一法图http:/;zgmnwk.cug.top-0.04图3酒精、乳制品、咖啡及绿茶摄入量与前列腺癌发病风险关联的孟德尔随机化分析漏斗图-0.010-0.005-0.008-0.02rv00-0.004A：酒精；B：乳制品；C：咖啡；D：绿茶。0.02A：酒精；B：乳制品；C：咖啡；D.绿茶。0.0050-0.002rs9273508Ts35225200rs35929758rs41279350rs11678980Ts9272263rs7503353Ts4557666rs7903146rs838133rs1652376Ts7325293rs853866rs937301Ts13340461rs116869426rs1065852TS6902789rs4631704rs28529403Ts142733003rs10760201rs58231973TS34647936Ts13074214rs9825208Is78186330All-0.025rs117077082rs644205rs12144868rs12958992rs11976995rs78547201rs115952340rs62059726rs79638269rs113322644rs142373582All-0.001Bv！0-1e-0300.0250.050-5e04-0.0010.075B0现代泌尿外科杂志3讨 论许多体内和体外研究表明，绿茶及其成分特别是表没食子儿茶素没食子酸酯（epigallocatechingallate，EGCG），可以通过促进细胞调亡、防止侵袭和转移等作用影响PCa 的发生和进展2 3-2 4。但是绿茶的抗癌作用在临床试验中尚存在争议。一些流行病学研究表明，绿茶的摄人对PCa有保护作用2 5，而其他研究则显示两者之间并无相关性，甚至显示出PCa风险增加的趋势2 6。2 0 16 年一项荟萃分析评估了绿茶摄人量与PCa风险的关系，结果表明：饮用绿茶会降低PCa的发病率,每天摄人量超过 7 杯显著降低了PCa的风险10 1。EGCG的抗癌作用机制已被广泛研究，很多研究表明EGCG通过抑制雄激素和胰岛素样生长因子(insulin like growth factor,IGF)途径降低了PCa 细胞在体内外的增殖2 7-2 8。REN等2 7 发现EGCG作用于雄激素受体启动子，降低LNCaP细胞中雄激素受体的表达,抑制细胞生长。EGCG通过抑制关键酶5-还原酶，导致前列腺肿瘤的生长抑制2 7。此外,EGCG还降低了有助于PCa转移的基质金属蛋白酶活性2 9。咖啡是由10 0 0 多种生物活性成分组成的复杂混合物，包括咖啡因、绿原酸和各种矿物质等。其中许多化合物已被证明具有抗癌潜力30。但是咖啡是否能降低患PCa的发病风险仍存在争议。此前已经有一项MR分析调查了咖啡和PCa风险之间的相关性31，作者使用了rs4410790和rs2472297作为工具变量，通过分析46 0 0 0 例男性PCa的汇总数据，发现遗传风险评分与PCa发病风险并不相关，这与我们的研究结果不一致。这可能是由于本文在分析中使用了更多的与咖啡摄人量相关的SNPs。在过去的几十年中，已经进行了几项综述和荟萃分析以检验PCa与饮酒之间的相关性32-35。LONG-NECKER等32 的早期综述得出PCa与饮酒之间没有关系的结论。BRESLOW等33回顾了32 项研究，其中2 4项研究报告了PCa风险与饮酒之间没有显著的关联。CHOI等34 的荟萃分析也认为饮酒与PCa发病风险没有因果关系。乳制品与PCa的关联近年来也被激烈讨论，2019年的一项荟萃分析显示，没有证据表明乳制品或牛奶消费与PCa风险之间存在关联35。2 0 18 年世界癌症研究基金会/美国癌症研究所在其2 0 18 年持续更新项目专家报告中（饮食、营养、身体活动和PCa)得出结论，乳制品消费量增加PCa风险的证据有限36 ,这也与本研究的结论一致。2024年3月第2 9 卷第3期http:/;zgmnwk.cug.top259本研究存在一些局限性：首先，纳人研究的数据库仅包括欧洲人群，研究结果可能不能直接适用于其他人群；其次，本文没有对乳制品、酒精的种类进行分类分析，这也需要在未来进行更广泛的研究。总之，在本研究中发现咖啡、绿茶摄入量与PCa发病存在反向因果关联，没有发现酒精、乳制品摄人量与PCa发病风险之间的因果关系，可能需要更大样本量来验证它们之间的因果关系。参考文献：1 SUNG H,FERLAY J,SIEGEL RL,et al.Global cancer statistics2020:GLOBOCAN estimates of incidence and mortality world-wide for 36 cancers in 185 countriesJJ.CA Cancer J Clin,2021,71(3):209-249.2 CENTER MM,JEMAL A,LORTET-TIEULENT J,et al.Inter-national variation in prostate cancer incidence and mortality ratesJJ.EurUrol,2012,61(6):1079-1092.3J ITO K.Prostate cancer in Asian menJJ.Nat Rev Urol,2014,11(4):197-212.4J ROTA M,SCOTTI L,TURATI F,et al.Alcohol consumptionand prostate cancer risk:a Meta-analysis of the dose-risk relationJJ.Eur JCancer Prev,2011,21(4):350-359.5J HONG S,KHIL H,LEE DH,et al.Alcohol consumption and therisk of prostate cancer:a dose-response meta-analysisJJ.Nutri-ents,2020,12(8):2188-2205.6 ZIOUZIOU I,TOUZANI AM,LAHLOU L,et al.Association ofprostate cancer with nuts,seeds,alcohol and processed meats:aworldwide population-based study J.Nutrition,Nutr Cancer,2021,73(11-12):2538-2545.7J SEN A,PAPADIMITRIOU N,LAGIOU P,et al.Coffee and teaconsumption and risk of prostate cancer in the European prospec-tive investigation into cancer and nutritionJ.Int J Cancer,2019,144(2):240-250.8 CHEN X,ZHAO Y,TAO Z,et al.Coffee consumption and riskof prostate cancer:a systematic review and meta-analysis J.BMJ open,2020,11(2):e038902.9 IMATOH T,SAWADA N,YAMAJI T,et al.Association be-tween coffee consumption and risk of prostate cancer in Japanesemen:a population-based cohort study in JapanJJ.Cancer Epide-miol Biomarkers Prev,2022,31(2):471-478.1oJ GUO Y,ZHI F,CHEN P,et al.Green tea and the risk of pros-tate cancer:a systematic review and Meta-analysisJ.Medicine(Baltimore),2017,96(13):e6426.11J BEYNON RA,RICHMOND RC,SANTOS FDL,et al.Investi-gating the effects of lycopene and green tea on the metabolome ofmen at risk of prostate cancer:the prodiet randomised controlledtrialJJ.Int J Cancer,2019,144(8):1918-1928.12 AUNE D,NAVARRO RDA,CHAN DS,et al.Dairy products,calcium,and prostate cancer risk:a systematic review and Meta-260analysis of cohort studiesJJ.Am J Clin Nutr,2015,101(1):87-117,13J LOPEZ-PLAZA B,BERMEJO LM,SANTURINO C,et al.Milkand dairy product consumption and prostate cancer risk and mor-tality:an overview of systematic reviews and Meta-analysesJ.AdvNutr,2019,10(suppl_2):S212-S223.14J PREBLE I,ZHANG Z,KOPP R,et al.Dairy product consump-tion and prostate cancer risk in the United StatesLJJ.Nutrients,2019,11(7):1615-1627.15J SMITH GD,TIMPSON N,EBRAHIM S.Strengthening causalinference in cardiovascular epidemiology through mendelianrandomizationJl.Ann Med,2008,40(7):524-541.16 TEARE MD.Methods in molecular biology:genetic epidemiologyvolume(Chapter 4)MJ.NY:Springer science:37-49.17J HARTWIG FP,DAVIES NM,HEMANI G.Two-sample Men-delian randomization:avoiding the downsides of a powerful,widely applicable but potentially fallible techniqueJ.Int J Epi-demiol,2016,45(6):1717-1726.18J BURGESS S,FOLEY CN,ZUBER V,et al.Inferring causal rela-tionships between risk factors and outcomes from genome-wideassociation study dataLJJ.Annu Rev of Genomics,2018,19:303-327.19J BOWDEN J,DAVEY SMITH G,BURGESS S.Mendelianrandomization with invalid instruments:effect estimation andbias detection through Egger regression JJ.Int J Epidemiol,2015,44(2):512-525.2oJ BOWDEN J,DAVEY SMITH G,HAYCOCK PC,et al.Consist-ent estimation in Mendelian randomization with some invalid in-struments using a weighted median estimatorJJ.Genet Epide-miol,2016,40(4):304-314.21 HIGGINS JP,THOMPSON SG,DEEKS JJ,et al.Measuringinconsistency in Meta-analysesLJJ.BMJ,2003(7414):557-560.22J VERBANCK M,CHEN CY,NEALE B,et al.Detection of wide-spread horizontal pleiotropy in causal relationships inferred fromMendelian randomization between complex traits and diseasesJ.Nat Genet,2018,50(5):693-698.23J SARTOR L,PEZZATO E,DONA M,et al.Prostate carcinomaand green tea:(-)epigallocatechin-3-gallate inhibits inflammation-triggered MMP-2 activation and invasion in murine TRAMPmodelJJ.Int J Cancer,2004,112(5):823-829.24J CHUU CP,CHEN RY,KOKONTIS JM,et al.Suppression of androgen receptor signaling and prostate specific antigen expres-sion by(-)-epigallocatechin-3-gallate in different progressionstages of LNCaP prostate cancer cellsLJJ.Cancer Lett,2009,275J Mod Urol,Vol.29 No.3 Mar.2024(1):86-92.25 KURAHASHI N,SASAZUKI S,IWASAKI M.Green tea con-sumption and prostate cancer risk in japanese men:a prospectivestudyJJ.Am JEpidemiol,2008,168(1):71-77.26 MONTAGUE JA,BUTLER LM,WU AH,et al.Green andblack tea intake in relation to prostate cancer risk among Singa-pore ChineseJJ.Cancer Causes Control,2012,23(10):1635-1641.27 REN F,ZHANG S,MITHELL SH,et al.Tea polyphenols down-regulate the expression of the androgen receptor in LNCaP pros-tate cancer cellsJJ.Oncogene,2000,19(15):1924-1932.28 HARPER CE,PATEL BB,WANG J,et al.Epigallocatechin-3-Gallate suppresses early stage,but not late stage prostate cancerin TRAMP mice:mechanisms of actionJ.Prostate,2007,67(14):1576-1589.29 PEZZATO E,SARTOR L,DELLAICA I,et al.Prostate carci-noma and green tea:PSA-triggered basement membrane degra-dation and MMP-2 activation are inhibited by(-)epigallocate-chin-3-gallateJJ.Int J Cancer,2004,112(5):787-792.30 POOLE R,KENNEDY OJ,RODERRICK P,et al.Coffee con-sumption and health:umbrella review of meta-analyses of multi-ple health outcomesJJ.BMJ,2017,359:j5024-j5042.31J WANG M,JIAN Z,YUAN C,et al.Coffee consumption andprostate cancer risk;results from national health and nutritionexamination survey 1999-2010 and mendelian randomizationanalysesJJ.Nutrients,2021,13(7):2317-2326.32 LONGNECKER MP.Alcohol consumption and risk of cancer inhumans:an overviewJJ.Alcohol,1995,12(2):87-96.33J BRESLOW RA,WEED DL.Review of epidemiologic studies ofalcohol and prostate cancer:1971-1996JJ.Nutr Cancer,1998,30(1):1-13.34 CHOI YJ,MYUNG SK,LEE JH.Light alcohol drinking and riskof cancer:a Meta-analysis of cohort studies JJ.Cancer ResTreat,2018,50(2):474-487.35 LOPEZ-PLAZA B,BERMEJO LM,SANTURINO C,et al.Milkand dairy product consumption and prostate cancer risk and mor-:tality:an overview of systematic reviews and meta-analysesJJ.Adv Nutr,2019,10(suppl_2):S212-S223.36J SHAMS-WHITE MM,BROCKTON NT,MITROU P,et al.Operationalizing the 2018 World Cancer Research Fund/Ameri-can Institute for Cancer Research（W CRF/A ICR)c a n c e r p r e-vention recommendations:a standardized scoring systemJJ.Nu-trients,2019,11(7):1572-1585.（编辑闫玉梅)http:/jmurology.xjtu.edu,cn;zgmnwk.cug.top