Research paper Daily coffee consumption and prevalence of nonmelanoma skin canc - PDF document

446ResearchpaperDailycoffeeconsumptionandprevalenceofnonmelanomaskincancerinCaucasianwomenErnestL.Abel,SusanO.Hendrix,S.GeneMcNeeleyKarenC.Johnson,CarolA.Rosenberg,YasminMossavar-RahmaniMaraVitolinsandMichaelKrugerThepurposeofthisstudywastoassesstherelationshipbetweendailycoffeeconsumptionandnonmelanomaskin eligibilitycriteriaandstudymethodshavebeenpre-viouslydescribedindetail[TheWomen’sHealthInitiative(WHI)StudyGroup,1998],aswellasbaselinecharacteristicsofparticipants(Langeretal.,2003).Briefly,womenofdiverseethnicbackgrounds,represent-ingthemajorethnicgroupsintheUnitedStates,whovolunteeredtoparticipate,wereeligibleifpost-menopausalattimeofenrollment.Attimeofenrollment,womencompletedscreeningandenrollmentquestion-nairesthatincludedquestionsaboutdemographicandlifestylevariables,coffeeconsumptionandhistoryofskincancer.Thequestionnairecategorizedcoffeeandteaconsumptioneachday,asnone,one,twotothree,fourtofive,orsixormorecupsperday.Respondentswerealsoaskedwhethertheyconsumedcaffeinatedcoffee,decaffeinatedcoffee,orteaforthepast3months.DataanalysisDataanalyseswerebasedoninformationobtainedatthebaselinevisit.PrevioushistoryofNMSCwasbasedonself-report.ApriorivariablesincludedintheanalyseswerethoseknownorsuspectedasbeingassociatedwithNMSC(Schottenfeld,1996;Rosenbergetal.,2004)usingthesamecriteriaorcutpointsasRosenbergetal.(2004).Theseincludedageatscreening,ethnicity(nativeAmerican,Asian,Black/African-American,Hispa-nic,White/Caucasian,other),latitudeoftheclinicwherepatientswereenrolled(Southern35N,MiddleN;Northern&#x-174;&#x.200;40N);bodymassindex(BMI)(25or25kg/m),historyofsmoking(current,past,nonsmoker),alcoholuse(nondrinker,pastdrinker,7or7drinksperweek),menopausalhormonetherapy(currentusers,pastusers,definedasthosereceivingtherapyfor&#x-168;&#x.500;3monthsbutnotreceivingtherapyattimeofinterviewandnonusers),education(highestgradecompleted:12years,highschool,somecollege,collegedegreeorpostgraduateschool),householdincome($20000,$20000–49000,&#x-168;&#x.500;$50000),percentagedailycaloriesfromfat(30.0,30.0–35.0,35.1–40.0,&#x-166;&#x.600;40.0),physicalactivity(METs/week)andintake(g);thelattertwowerecategorizedintofourlevelsbasedonthe25thpercentilecut-points(0–25,26–50,51–75,&#x-166;&#x.600;75),havingacurrenthealthcareproviderandmaritalstatus(nevermarried,divorcedorseparated,presentlymarried).AsintheRosenbergetal.(2004)study,weconsideredwomenreportingdailyenergyintakesof600or&#x-168;&#x.500;5000kcal/dayasunreliableandexcludedthemfromouranalyses.Demographicandlife-stylevariables,includingdailycoffee(regularanddecaf-feinated)andteaconsumption,wereevaluatedfortheirassociationwithNMSCbytests.Samplesizeforeachanalysisdifferedbecauseofmissingdatavalues.Othersourcesofcaffeineintake(e.g.colas)werenotincludedintheregressionmodelbecauserelevantinformationwasnotavailable.Stepwiselogisticregressionanalysiswasusedtoconstructadose–responsemodelforcaffeinatedcoffee’sassociationwithNMSC.Inordertoavoidthespuriouschoiceofvariablesinthemodelasaconsequenceofthelargesamplesize,weusedacriteriaofasignificantFvaluetoenterandanotherFvaluetoremovefromtheanalysesof0.01and0.05,respectively.ThedatawereanalyzedusingtheSPSS(version11.0)statisticalpackageforWindows(SPSSInc.,Chicago,Illinois,USA).Atotalof93676womenofdiverseethnicbackgrounds,16.4%ofwhomwereminority,providedbaselinedata.Overallprevalenceofself-reportedNMSCamongallwomeninthestudywas8.3%(=7775).AsthemajorityofwomeninthisstudywereCaucasian(83.6%),andnearlyallcases(97.8%)ofself-reportedNMSCoccurredamongthesewomen,andbecauseethnicitywassig-nificantlyrelatedtotheoccurrenceofNMSC=1265.3,d.f.=5,0.001),wefocusedonthissubgroupof78013women.Ofthese78013women,informationwithrespecttoself-reportedNMSCwasmissingfor638women.Afurther467(0.6%)hadmissingdatawithrespecttodailycoffeedrinking.Dailycoffeeconsumptionfortheremaining77373womenisshowninTable1;40.1%(=31048)ofthesewomenonlydrankcaffeinatedcoffee,15.0%(=11597)onlydrankdecaffeinatedcoffee,16.5%(=12759)drankbothcaffeinatedanddecaffeinatedcoffee,and28.4%=21969)reporteddrinkingnocoffee.Therelationshipbetweendemographic/lifestylevari-ablesandNMSCisshowninTable2.LatitudeofresidencewassignificantlyassociatedwithNMSC;thehighestprevalenceoccurredamongwomenlivinginthesouthernmostlatitude(12.6%)andthelowestinthenorthernmostlatitude(7.6%).HigherBMI25)wasalsosignificantlyassociatedwithNMSC.Thehighertheeducationandincomelevels,thehighertheprevalenceofNMSC.SmokingandalcoholconsumptionwerebothsignificantlyassociatedwithNMSC.PastorcurrentusersofmenopausalhormonereplacementtherapyhadahigherprevalenceofNMSCthannonusers,butonlythedifferencesbetweenpastusersorcurrentusers,andneverusersweresignificant(0.001foreach).PhysicalactivitywaspositivelyassociatedwiththeprevalenceofNMSCwhereaspercentageofdailydietarycaloriesfromfatwasnegativelyrelated.Beta-caroteneintakewassignificantlyrelatedtoNMSCbuttheassociationwithNMSCwasnotlinear.Therelationshipbetweendailycaffeinatedcoffeecon-sumptionandNMSCisshowninTable3.Drinkingcaffeinatedcoffeewasassociatedwithadecreasedprevalenceofself-reportedNMSCcomparedwithnon-drinkers(9.1vs.10.2%,respectively).ThisassociationCoffeeandcanceretal.447 caffeinatedcoffeeconsumptionandtheriskofNMSCwasstatisticallysignificant(0.001).AschroniccumulativeexposuretosolarultravioletradiationisthemostimportantfactorassociatedwithNMSC(TaylorandSober,1996),wealsotestedforaninteractionbetweencupsofcaffeinatedcoffeeandlatitude,forriskofNMSC.Theinteractionwasnotstatisticallysignificant.WealsoconductedasubgroupanalysisoftherelationshipbetweencoffeeconsumptionandNMSCstratifyingforhistoriesofothercancers.TherelationshipbetweencaffeinatedcoffeeconsumptionandNMSCwasnotsignificantforwomenwithahistoryofcancerotherthanNMSC.Asignificantdose–responserelationship,however,wasstillevidentforthosewomenwithoutahistoryofothercancers.Theoddsratioforsixormoreforthelattercupswas0.70(confidenceinterval:0.56–0.89,0.001). Table2Demographic/lifestylevariablesandriskofnonmelanomaskincancerCrudeORUnadjustedORVariableCategoryOR95%CIOR95%CIAge(years)50–59————60–691.05(1.01–1.10)1.05(1.02–1.09)70–791.45(1.38–1.52)1.44(1.39–1.49)AlcoholNondrinkerPastdrinker1.09(0.94–1.27)1.03(0.93–1.14)7drinksperweek1.24(1.09–1.42)1.08(0.99–1.18)&#x-171;&#x.200;7drinksperweek1.38(1.19–1.60)1.20(1.08–1.33)Education12yearsHighschooldegree1.31(1.03–1.65)1.13(0.95–1.34)Afterhighschool1.59(1.27–2.00)1.43(1.22–1.68)Collegedegree1.90(1.52–2.39)1.74(1.48–2.04)HormonereplacementtherapystatusNeverusedPastuser1.17(1.07–1.28)1.20(1.12–1.29)Currentuser1.04(0.97–1.11)1.04(0.99–1.09)RegionSouthernMiddle0.99(0.95–1.04)1.00(0.97–1.04)Northern0.76(0.73–0.80)0.76(0.73–0.78)Bodymassindex25250.89(0.83–0.94)0.81(0.77–0.85)Caloriesfromfat(%)3030–351.03(0.94–1.13)1.00(0.94–1.06)35.1–401.05(0.96–1.15)0.94(0.88–1.01)&#x-171;&#x.300;401.19(1.08–1.30)0.85(0.79–0.91)OR,oddsratio;CI,confidenceinterval. Table3DailycaffeinatedcoffeedrinkingandriskofnonmelanomaskincanceramongCaucasianwomen(77373)PercentagePercentagewithskincancerOddsratio95%ConfidenceintervalUnadjustedoddsratio3356843.410.31.0——11258416.310.20.96(0.89–1.03)0.99(0.92–1.05)2–32313329.99.10.87(0.81–0.92)0.87(0.82–0.92)4–563528.28.10.83(0.75–0.92)0.76(0.69–0.84)6ormore17382.26.70.70(0.60–0.88)0.63(0.52–0.76)Total773759.7Oddsratiovariablesincludedinstepwiseregressionmodel:ageatscreening,alcoholconsumption,smoking,income,regionofresidencebylatitude,education,menopausalhormonetherapy,bodymassindex,and-caroteneintake.Unadjustedoddsratiobasedoncupsofregularcoffeeastheonlypredictorinthemodel.Drinkseithernocoffeeordecaffeinatedcoffeeonly. Table4Dailyintakeofdecaffeinatedcoffeeandteaandunadjustedoddsratio(OR)ofnonmelanomaskincancer(NMSC)riskDecaffeinatedcoffeeTeaPercentagePercentageNMSCOR95%CICupsPercentagePercentageNMSCOR95%CI02977754.39.1——05639173.99.6——11153721.010.01.121.04–1.201965312.610.11.060.99–1.142–31080319.710.11.131.05–1.222–3810510.611.01.050.97–1.144–522464.18.80.960.83–1.124–516832.29.10.950.80–1.12�64830.99.71.080.80–1.46�64880.68.20.840.61–1.17Total548469.5Total763209.7CI,confidenceinterval.Coffeeandcanceretal.449 manywomenintheyearspriortoassessment,thedietaryinformationrecordedatenrollmentmaynotreflectgeneralpatternsofdietoverthelifespan.Denialandunder-reportingalsohastobeconsideredinthecontextofsmokingandalcoholuse(Abel,1998).We,however,arenotawareofanydatasuggestingdifferentialunder-reportingofeithersmokingoralcoholuseinwomenwithandwithoutNMSC.Informationaboutcoffeeandteausemayalsobeproblematicaswomenwereaskedtorecalltheirlevelsofdailyconsumptionatthetimeofenrollmentandconsumptionlevelsmayhavenotbeensimilaratthetimewomenwerediagnosedwithskincancer.Despitethesecaveats,thedecreasedprevalenceinNMSCassociatedwithconsumptionofdailyconsump-tionofcaffeinatedcoffeethatweobservedwasdose-relatedandconsistentwithotherstudies.Amongthepossibleexplanationsforcaffeine’sprotectiveeffectonNMSCthathavebeensuggestedisanantioxidanteffect(Kuo,1997;TrevisanatoandKim,2000)and/orinhibitionofDNAsynthesisandcelldivision,makingcellslesssusceptibletocarcinogenesis(Timson,1997).Weconcludethatdailycaffeinatedcoffeeconsumptionisassociatedwithadose-relateddecreasedprevalenceofNMSCinCaucasianwomen.Ideally,theobservationsreportedhereandourconclusionshouldbeexploredinfutureprospectivestudies.AcknowledgementsTheWHIprogramisfundedbytheNationalHeart,LungandBloodInstitute,USDepartmentofHealthandHumanServices.AbelEL(1998).Fetalalcoholabusesyndrome.NewYork:PlenumPress.AubryF,MacGibbonB(1985).Riskfactorsofsquamouscellcarcinomaoftheskin:acase-controlstudyintheMontrealregion.:907–911.BlackHS(1998).Influenceofdietaryfactorsonactinically-inducedskincancer.MutatRes:185–190.CoronaR,DogliottiE,D’ErricoM,SeraF,LevaroneI,BalivaG,etal.(2001).RiskfactorsforbasalcellcarcinomainaMediterraneanpopulation:roleofrecreationalsunexposureearlyinlife.ArchDermatol:1162–1168.DeHertogSAE,WensveenCAH,BastiaensMT,KiehichCJ,BerkhoutMJP,WestendorpRGJ,etal.(2001).Relationbetweensmokingandskincancer.JClinOncol:231–238.FraryCF,JohnsonRK,WangMQ(2005).FoodsourcesandintakesofcaffeineinthedietsofpersonsintheUnitedStates.JAmDietAssocGrodsteinF,SpeizerFE,HunterDJ(1995).Aprospectivestudyofincidentsquamouscellcarcinomaoftheskininthenurses’healthstudy.JNatlCancerInst:1061–1066.HuangM-T,XieJ-G,WangZY,HoCT,LouYR(1997).CarcinogenesisinSKH-1mice:demonstrationofcaffeineasabiologicallyimportantconstituentoftea.CancerRes:2623–2629.JacobsenBK,BjelkeR,KvaleG,HeuchI(1986).Coffeedrinking,mortality,andcancerincidence:resultsfromaNorwegianprospectivestudy.JNatlCancer:823–834.JohnsonML,JohnsonKG,EngelA(1984).Prevalence,morbidity,andcostofdermatologicdiseases.JAmAcadDermatol:930–936.KaaksR,LukanovaA,KurzerMS(2002).Obesity,endogenoushormones,andendometrialcancerrisk.Asyntheticreview.CancerEpidemiolBiomarkers:1531–1543.KuoSM(1997).Dietaryflavonoidandcancerprevention:evidenceandpotentialCritRevOncol:47–69.LangerRD,WhiteE,LewisCE,KotchenJ,HendrixS,TrevisanM(2003).TheWomen’sHealthInitiativeObservationalStudy:baselinecharacteristicsofparticipantsandreliabilityofbaselinemeasures.AnnEpidemiolS107–S121.MarksR(1996).Squamouscellcarcinoma.347:735–738.MichelsKB,HolmbergL,BergkvistL,WolkA(2002).Coffee,tea,andcaffeineconsumptionandbreastcancerincidenceinacohortofSwedishwomen.AnnEpidemiolMingME,RossR,LevyRM,OleJ,HoffstadMA,FilipJ,etal.(2004).Validityofpatientselfreportedhistoryofskincancer.ArchDermatol140:730–735.RiboliE(2001).TheEuropeanprospectiveinvestigationintocancerandnutrition(EPIC):plansandprogress.JNutr131:170S–175S.RosenbergCA,GreenlandP,KhandekarJ,LoarA,AscensaoJ,LopezAM(2004).Associationofnonmelanomaskincancerwithsecondmalignancy.TheWomen’sHealthInitiativeObservationalStudy.100:130–138.SchottenfeldD(1996).Basal-cellcarcinomaoftheskin:aharbingerofcutanenousandnoncutaneousmultipleprimarycancer(editorial).AnnIntern:852–854.StromSS,YamamuraY(1997).Epidemiologyofnonmelanomaskincancer.PlastSurg:627–636.TavaniA,LaVecchiaC(2000).Coffeeandcancer:areviewofepidemiologicalstudies,1990–1999.EurJCancerPrevent:241–256.TaylorCR,SoberAJ(1996).Sunexposureandskindisease.AnnRevMedTheWomen’sHealthInitiative(WHI)StudyGroup(1998).DesignoftheWomen’sHealthInitiativeclinicaltrialandobservationalstudy.ControlClinTrials:61–109.TimsonJ(1997).Caffeine.MutatResTrevisanatoS,KimYI(2000).Teaandhealth.NutrRevVeierodMB,ThelleDS,LaakeP(1997).Dietandriskofcutaneousmalignantmelanoma:aprospectivestudyof50757Norwegianmenandwomen.IntJ:600–604.VonDomarusH,StevensPJ(1984).Metastaticbasalcellcarcinoma:reportoffivecasesandreviewof170casesintheliterature.JAmAcadDermatol:1043–1060.WangZY,AgarwalR,BickersDR,MukhtarH(1991).ProtectionagainstultravioletBradiation-inducedphotocarcinogenesisinhairlessmicebygreenteapolyphenols.WangZY,HuangM-T,FerraroT,WongCQ,LouYR,ReuhlK,etal.(1992).Inhibitoryeffectofgreenteaindrinkingwaterontumorigenesisbyultravioletlightand12--tetradecanolylphorbol-13-acetateintheskinofSKH-1mice.CancerRes:1162–1170.WangZY,HuangMT,LouY-R,XieJG,ReuhlKR,NewmarkHL,etal.(1994).Inhibitoryeffectsofblacktea,greentea,decaffeinatedblacktea,anddecaffeinatedgreenteaonultravioletBlight-inducedskincarcinogenesisin7,12-dimethylbenz(a)anthracene-initiatedSKH-1mice.CancerRes3428–3435.WebbAR,KlineL,HolickMF(1988).InfluenceofseasonandlatitudeonthecutaneoussynthesisofvitaminD3:exposuretowintersunlightinBostonandEdmontonwillnotpromotevitaminD3synthesisinhumanskin.JClinEndocrinolMetab667:373–378.AppendixShortlistofWomen’sHealthInitiativeinvestigatorsProgramofficeNationalHeart,Lung,andBloodInstitute,Bethesda,Maryland:BarbaraAlving,JacquesRossouw,LindaPottern.ClinicalcoordinatingcenterFredHutchinsonCancerResearchCenter,Seattle,Washington:RossPrentice,GarnetAnderson,AndreaLaCroix,CharlesL.Kooperberg,RuthE.Patterson,AnneMcTiernan;WakeForestUniversitySchoolofMedicine,Winston-Salem,NorthCarolina:SallyShumaker;MedicalResearchLaboratories,HighlandHeights,Kentucky:EvanStein;UniversityofCaliforniaatSanFrancisco,SanFrancisco,California:StevenCummings.Coffeeandcanceretal.