sidney, stephen et al_marijuana use and cancer incidence

8/14/2019 Sidney, Stephen Et Al_Marijuana Use and Cancer Incidence

1/7

CancerCauses ndControl, 997,8, pp. 722-728

Marijuana u s e a n d c a n c e r incidence(California, U n i t e d S t a t e s )

Stephen Sidney, Charles P. Quesenberry, Jr., Gary D. Friedman,and Irene S. Tekawa(Received 16 December 1996; accepted in revised form 12 May 1997)

The purpose of this retrospective cohort study was to examine the relationship of marijuana use to cancer incidence.The study population consisted of 64,855 examinees in the Kaiser Permanente multiphasic health checkup in SanFrancisco and Oakland (California, United States), between 1979-85, aged 15 to 49 years, who completed self-ad-ministered questionnaires about smoking habits, including marijuana use. Follow-up for cancer incidence wasconducted through 1993 (mean length 8.6 years). Compared with nonusers/experimenters (lifetime use of less thanseven times), ever- and current use of marijuana were not associated with increased risk of cancer of all sites (relativerisk [RR] = 0.9, 95 percent confidence interval [CI] = 0.7-1.2 for ever-use in men; RR = 1.0, CI = 0.8-1.1 in women)in analyses adjusted for sociodemographic factors, cigarette smoking, and alcohol use. Marijuana use also was notassociated with tobacco-related cancers or with cancer of the following sites: colorectal, lung, melanoma, prostate,breast, cervix. Among nonsmokers of tobacco cigarettes, ever having used marijuana was associated with increasedrisk of prostate cancer (RR = 3.1, CI = 1.0-9.5) and nearly significantly increased risk of cervical cancer (RR = 1.4,CI = 1.0-2.1). We conclude that, in this relatively young study cohort, marijuana use and cancer were not associatedin overall analyses, but that associations in nonsmokers of tobacco cigarettes suggested that marijuana use mightaffect certain site-specific cancer risks. Cancer Causes and Control 1997, 8, 722-728Key words: Cancer incidence, marijuana, retrospective cohort study, United States.

IntroductionMarijuanause is by far more prevalentthan the use ofany other illegal drug in the United States.It has beenestimated that nearly one-third of the US populationaged12and olderhas usedmarijuana.t hemost commonroute of administration is smoking. Since marijuanasmoke contains many of the same chemicalsas tobaccosmoke, in addition to containing 60 cannabinoidcom-pounds, it is plausible to hypothesize that it may becarcinogenic.2However, marijuanahas not been estab-lishedasa riskfactor for any cancer,although t mayplay

a role in the development of lung cancer -5 and of headand neck cancers.59 The uncertainty in establishing mari-juana as a cancer risk factor may be due, in part, to thedifficulty in separating the effects of marijuana from theeffects of tobacco cigarettes, since in many instances themarijuana smoker also smokes tobacco cigarettes. Eventhough a marijuana cigarette may be smoked in a moreintense manner (greater puff volume and length of inha-lation)10 than a tobacco cigarette, the total quantity ofmarijuana smoked by a user during his/her lifetime is

Authors are with the Division of Research, Kaiser Permanente Medical Care Program, Oakland, California, USA. Address correspondenceto Dr Sidney, Kaiser Permanente Division of Research, 3505 Broadway, Oakland, California 94611-5714, USA. This research wassupported by Grant no. R01 DA06609 from the US National Institute on Drug Abuse. Dr Friedman is supported in part by Grant no.R35 CA49761 from the US National Cancer Institute. Data collection regarding alcohol use was supported by a grant from TheAlcoholic Beverage Medical Research Foundation (Baltimore, MD, USA).722 CancerCauses and Control. Vol 8. 1997 ( 1997RapidSciencePublishers


2/7


3/7

S. Sidneyet al

Statistical nalyseswereperformedusingSASprograms.8Age-adjusted ncidencerateswerecalculatedby the directmethod using 10-year age groupingswith the 1990 USpopulation as the standard. Cox proportional hazardsmodels were used to examine the joint effect of so-ciodemographic characteristics,marijuanause, tobacco,and alcoholuse on theriskof cancer,rom whichestimatesof relative isks(RR)andassociated95percentconfidenceintervals CI) were obtained.'9Two-way interactionsbe-tween marijuana ndtobacco use and betweenmarijuanaand alcoholweretestedby includingcross-product ermsin theproportionalhazardsmodels.The interaction ermswere statisticallysignificant n a few of the models (to-bacco and ever-use of marijuanaor colorectalcancer nwomen and melanomain men; alcohol and both ever-and currentuse of marijuanaor prostatecancer n men;alcohol and both ever- and currentuse of marijuanaorbreast cancer in women). However, the RR estimatesassociatedwith tobacco or alcohol use within any givenlevel of marijuana se in these models had wide CIs andlack of dose-response,so that only the main effects arereportedin this paper.To obtain estimatesof the overallrisk associatedwith marijuanause, adjusted or tobaccocigaretteuse, tobacco cigaretteuse was categorizedas athree-level variable (i.e., nonuser, former user, currentuser).

ResultsSociodemographiccharacteristics Table 1)Current users, former users, experimenters, and nonusers,respectively, comprised 22 percent, 20 percent, 20 percent,and 38 percent of the cohort. The percentage of ever-userswas highest in the 20 to 29 year age group. Ever-use ofmarijuanawas more common in men than in women, andmore common in Whites followed by Blacks, Hispanics,and Asians.Tobacco cigarette useThe distribution of marijuana use (ever cf nonusers/experimenter) and tobacco cigarette smoking (ever cfnonsmoker) are shown in Table 2. The prevalence of ahistory of tobacco-cigarette smoking was 54 percenthigher in ever-users of marijuana (59.9 percent) than innonusers/experimenters (38.8 percent). A substantial pro-portion of marijuana users were nonsmokers of tobaccocigarettes (40.1 percent of ever-users of marijuana[10,710of 26,733]).Distribution of cancer sitesOverall, there were 379 incident cancers in men, and 1,042incident cancers in women. We performed analyses forcancers of all sites, tobacco-related cancers, and individual

Table 1. Sociodemographic characteristics of a cohort of Kaiser Permanente Medical Care Program members (n = 64,855) bymarijuanause status, June 1979 throughDecember 1985Status at multiphasic health Never Experimenter Former CurrentcheckupAge (yrs)15-19 902 558 508 757

20-29 5,489 3,864 4,734 5,97030-39 8,929 5,320 5,751 5,75940-49 9,647 3,413 1,707 1,547GenderMale 9,072 5,271 6,075 7,502Female 15,895 7,884 6,625 6,531RaceWhite 9,894 7,316 8,372 8,743Black 7,654 3,770 2,838 3,974Asian 5,085 1,062 649 399Hispanic 1,525 615 476 528Other/unknown 809 392 365 389

Education< Highschool 5,837 2,420 1,887 2,518Technical/businessschool/ 8,706 4,760 4,458 5,435some collegeCollege graduate/post graduate 9,941 5,703 6,034 5,850Unknown 483 272 321 230

Total 24,967 13,155 12,700 14,033724 Cancer Causes and Control. Vol 8. 1997


4/7

Marijuanaand cancer

sites with 30 or more gender-specific incident cancers(colorectal, lung, and melanoma for both genders;prostate for men, and breast and cervix for women). Thecounts for all these groupings and sites by marijuana andtobacco-cigarette use status are shown in Tables 3 and 4.

Table 2. Distributionof study populationby gender and byever-use of marijuanaand tobacco cigarettes, Kaiser Perma-nente MedicalCare ProgramMarijuana/cigarette Men Womensmoking history

Marijuana Tobacco No. No.use cigarette useNever Never 8,170 15,360Never Ever 6,173 8,419Ever Never 5,593 5,117Ever Ever 7,984 8,039Total 27,920 36,935

Marijuana and tobacco cigarette use in relation tocancer incidenceAs expected, tobacco cigarette use, regardless of marijuanause status, was associated with increased risks of tobacco-related cancer ranging from 2.7 to 3.8 relative to nonuseof marijuana and tobacco cigarettes in models adjustedfor age, race/ethnicity, education, and alcohol use (Tables3 and 4); RRs for lung cancer associated with tobaccocigarette use ranged from 9.2 to 11.2. Ever-use of mari-juana by nonsmokers of tobacco cigarettes was associatedwith increased risk in men of prostate cancer (RR = 3.1,CI = 1.0-9.5) and with a nearly significant increased riskin women of cervical cancer (RR = 1.4, CI = 1.0-2.1) rela-tive to nonuse of marijuana and tobacco cigarettes.Additional adjustment for marital status did not appre-ciably alter the estimates of RRs for cancer of these sites,nor was it statistically significantly associated with thesecancer sites. We also examined the association of mari-juana use with invasive cervical cancer, comprising 11percent (33 of 302) of all cervical cancers. Ever-use ofmarijuana in nonsmokers of tobacco cigarettes was asso-

Table 3. Counts, age-adjusted rates, and relative risks (RR) of all cancers, tobacco-related cancers, and cancers of specificsites by marijuanaand cigarette smoking status in men, Kaiser Permanente Medical Care ProgramSite of cancer Marijuana, cigarette Number of cases Ratea RRb (CI)csmoking historyAllsites Never, never 109 149 1.0

Never, ever 138 169 1.1 (0.9-1.5)Ever,never 36 186 0.8 (0.5-1.2)Ever,ever 96 165 1.1 (0.8-1.5)Tobacco-related Never, never 12 19 1.0Never, ever 56 68 3.1 (1.6-5.8)Ever,never 3 7 0.8 (0.2-2.9)Ever,ever 28 39 2.8 (1.4-5.6)Colorectal Never, never 16 16 1.0Never, ever 15 18 0.8 (0.4-1.6)Ever, never 4 7 0.7 (0.2-2.1)Ever,ever 10 29 0.9 (0.4-1.9)

Lung Never, never 2 3 1.0Never, ever 32 39 10.3 (2.4-43.7)Ever, never 0 -Ever,ever 14 26 9.2 (2.0-42.0)Melanoma Never, never 23 27 1.0Never, ever 10 9 0.5 (0.2-1.0)Ever,never 6 11 0.5 (0.2-1.3)Ever,ever 19 25 1.1 (0.6-2.0)Prostate Never, never 9 16 1.0Never, ever 21 28 2.0 (0.9-4.5)Ever,never 5 95 3.1 (1.0-9.5)Ever,ever 7 31 1.7 (0.6-4.9)

a Per 100,000 person-years, age adjusted to 1990 US population.b Adjusted for age, race, education, and alcohol use.c Cl = 95% confidence interval.

Cancer Causes and Control. Vol 8. 1997 725


5/7

S. Sidneyet al

Table 4. Counts, age-adjusted rates, and relative risks (RR) of all cancers, tobacco-relatedcancers, and cancers of specificsites by marijuanaand cigarette smoking status in women, Kaiser Permanente Medical Care ProgramSite of cancer Marijuana,cigarette Number of cases Ratea RRb (CI)csmoking historyAll sites Never,never 408 260 1.0

Never, ever 353 393 1.3 (1.1-1.5)Ever,never 93 403 1.1 (0.8-1.3)Ever,ever 188 408 1.2 (1.0-1.4)Tobacco-related Never,never 18 12 1.0Never, ever 55 54 3.8 (2.1-6.7)Ever, never 0-Ever,ever 10 47 2.7 (1.2-6.3)Colorectal Never, never 27 18 1.0Never, ever 28 26 1.4 (0.8-2.4)Ever,never 1 2 0.3 (0.0-2.5)Ever,ever 6 8 0.9 (0.4-2.3)

Lung Never,never 5 4 1.0Never,ever 36 35 9.8 (3.6-26.5)Ever, never 0-Ever,ever 8 44 11.2 (3.3-37.7)Melanoma Never,never 23 11 1.0Never,ever 18 18 1.1 (0.6-2.0)Ever,never 8 71 1.0 (0.4-2.3)Ever,ever 16 19 1.2 (0.6-2.5)Breast Never,never 171 101 1.0Never, ever 113 117 0.9 (0.7-1.2)Ever,never 22 115 0.8 (0.5-1.3)Ever,ever 54 138 1.0 (0.8-1.5)Cervix Never, never 92 62 1.0Never, ever 80 118 1.7 (1.2-2,3)Ever, never 48 119 1.4 (1.0-2.1)Ever,ever 82 125 1.6 (1.2-2.2)

abc

Per 100,000 person-years, age adjusted to 1990 US population.Adjustedfor age, race, education, and alcohol use.Cl= 95% confidence interval.

ciated with a nonsignificant increased risk of invasivecervical cancer (RR = 2.4, CI = 0.8-6.7). The overall risksof cancer associated with ever-use of marijuanaare shownin Table 5, adjusted for tobacco cigarette smoking status(non-, former, current) as well as the other variables notedearlier. Ever use of marijuana was not associated in menor in women with increased or decreased risk of anycancer grouping or individual site. The results of analysesof cancer incidence excluding the 222 (15.6 percent oftotal cancers) which developed within the first two yearsof follow-up was similar to the overall results, suggestingthat the overall results were not affected by symptomscaused by cancer prior to its diagnosis that affected thedecision to use marijuana.

The findings in regard to current marijuana use weresimilar to those for ever-use of marijuana. Current mari-juana use also was associated with increased risk ofprostate cancer (RR = 4.7, CI = 1.4-15.5) and a nearly726 Cancer Causes and Control. Vol 8. 1997

Table 5. Risk of cancer in ever-users of marijuanarelative tononusers/experimentersadjustedforcigarette smokingstatusand other variables,a Kaiser Permanente Medical CareProgramSite of cancer Men Women

RRb (Cl)c RRb (CI)cAll sitesTobacco-relatedColorectalLungMelanomaProstateBreastCervix

0.9 (0.7-1.2) 1.00.9 (0.6-1.4) 0.70.9 (0.5-1.8) 0.60.9 (0.5-1.7) 1.11.2 (0.7-2.1) 1.11.3 (0.6-2.6) -1-.0~- -- ~ 1.1

(0.8-1.1)(0.3-1.4)(0.2-1.3)(0.5-2.6)(0.6-1.9)(0.8-1.3)(0.9-1.5)

a Adjustedfor age, race, education, alcohol use, andtobacco cigarette smoking.b RR = relative risk.c CI = 95% confidence interval.


6/7

Marijuanaand cancer

significant ncreasedriskof cervicalcancer(RR = 1.6, CI= 1.0-2.5).We examined the associationof duration(years)andfrequencyof marijuana se for cancer-sitegroupingsandfor individualsites with 10 or more gender-specificnci-dent cancersamongever-usersof marijuana.Duration ofuse was not associatedwith the riskof any cancersite orgroupingof cancers.For prostatecancer,a tendencyto-ward an association of frequencyof marijuanause withincreasedrisk was evident,with usageone or more timesperweek (56.3percentof maleever-users) ssociatedwitha nonsignificanttwofold increase n the risk of prostatecancerrelativeto nonusers/experimenters.However, theaddition of frequencyof use did not improvethe fit ofany of the models for prostatecanceror for cancerof anyother site or groupingof cancers.

DiscussionThe main findingswere the lack of associationof mari-juana use with overall cancer incidence and with theincidence of all tobacco-relatedcancerscombined. Mari-juanause was associatedwith anincreased isk of prostatecancer in men who were nonsmokers of tobacco ciga-rettes, and a nearly significantrisk of cervical cancer inwomen who were currentmarijuana mokers and non-smokersof tobacco cigarettes.The association between marijuanause and prostatecancer ound in this studyraises he questionof potentialhormonal mediation. There is a significantbody of evi-dencesuggesting hatandrogensareimportant nprostatecarcinogenesis, ncludingthe following:(i) androgensarerequiredfor the normalgrowth and maintenanceof theprostate;0 ii)testosteronehas beenshown to inducepros-tate cancer n laboratory odents;21nd(iii)prostatecancerrarely occurs in eunuchs, and antiandrogen herapy orablation results n regressionof prostatecancer.22tudiesexamining herelationshipof serumtestosterone o pros-tate cancer in humans have not shown consistentresults.2324n early study showing thatplasma estoster-one levels in men were decreased n chronic marijuanausers25 as not been supportedby most of the numeroussubsequentstudies regarding his issue.26 exualactivitymeasureshave been inconsistentlyassociatedwith pros-tate cancer,27although they are associated withpsychoactive drug use.28 It is notable that the prostatecancercasesrepresentedn thisstudyarerelativelyyoungcomparedwith thoseoccurring nthe generalpopulation,since the maximumage reachedby this cohort duringfollow-up was 63 years.

Cigarettesmoking has been shown to be a risk factorfor cervical cancer in many studies.'5The associationofmarijuanause with cervicalcancer n nonsmokersmightreflect confounding by sexual activity,since numberof

sexual partners is a risk factor for cervical cancer29and isassociated with psychoactive drug use.28The major limitations of this study include reliance onself-report for ascertainment of marijuana-use status;inability to study change in marijuana-use status duringfollow-up; and lack of lengthy follow-up into the geriatricage range, as noted earlier. We do not believe that under-reporting of marijuana use is likely to have been asignificant problem because data collection on marijuanause was completed before the acceleration of the 'War onDrugs' in the late 1980s, during a period of time whenthe environment surrounding marijuana use was moreaccepting. Additionally, marijuana use rates in the studycohort were similar to those reported during the sametime period by the National Household Survey on DrugAbuse, the most authoritative source of statistics on illegaldrug use in the US.30Lack of longitudinal data regardingan exposure is a common feature of many cohort studies.We believe that it is unlikely that ever-use of marijuanawould have increased substantially over time, because arelatively small proportion of the cohort are likely to haveinitiated marijuanause during follow-up which occurredin aperiod marked by a significant decline in self-reportedmarijuana use in the US.'In conclusion, this study did not show any significantassociations between marijuana use and cancer in overallanalyses, but showed associations suggesting increasedrisk of prostate and cervical cancer in marijuana userswho were nonsmokers of tobacco cigarettes. We do notconsider any of the findings to be conclusive, and as withall epidemiologic work, we strongly suggest that mari-juana use and cancer be studied in other settings.Acknowledgements- The authors acknowledgeSteve Wilson and Anne Zishka for computer program-ming, Leo Hurley for consultation regarding the useof the Kaiser Permanente HIV/AIDS database, andDiana Holt for technical assistance in preparing themanuscript.

References1. SubstanceAbuse and MentalHealthServicesAdministra-tion. Preliminary Estimates from the 1994 NationalHouseholdSurvey.AdvanceReportNo. 10.Rockville,MD(USA): SAMHSA, 1995.2. Hoffman D, Brunnemann KD, Gori BG, Wynder EL. Onthe carcinogenicity of marijuana smoke. Recent Adv Phy-tochem 1975; 9: 63-81.3. Gong H Jr, Fligiel S, Tashkin DP, Barbers RG. Tracheo-bronchial changes in habitual, heavy smokers of marijuanawith and without tobacco. Am Rev Respir Dis 1987; 136:142-9.4. Taylor FM. Marijuana as a potential respiratory tract car-cinogen: A retrospective analysis of a community hospitalpopulation. South MedJ 1988; 81: 1213-6.

Cancer Causes and Control. Vol 8. 1997 727


7/7

S. Sidneyet al

5. Tashkin DP. Is frequentmarijuana moking harmful tohealth[Letter].WestJ Med 1993;158: 635-7.6. Caplan GA, BrighamBA. Marijuana moking and carci-noma of the tongue.Is there an association?Cancer1989;66: 1005-6.7. Donald PJ.Advancedmalignancyn the young marijuanasmoker.In: FriedmanH, SpecterS, KleinTW,et al, eds.Drugs of Abuse, Immunity,and Immunodeficiency.NewYork,NY (USA): PlenumPress, 1991:33-46.8. Donald PJ.Marijuana ndupper aerodigestive ractmalig-nancy in young patients.In: Nahas GG, LatourC, eds.CannabisPhysiopathology,Epidemiology,Detection. BocaRaton, FL (USA): CRC Press,1993: 165-83.9. NahasG, LatourC. Thehuman oxicityof marijuana.MedJ Aust 1992;156:495-7.10. WuTC, TashkinDP, DjahedB, Rose JE.Pulmonaryhaz-ards of smoking marijuana s comparedwith tobacco.NEnglJ Med 1988;318:347-51.11. SidneyS,BeckJE,TekawaIS,QuesenberryCP,FriedmanGD. Marijuana se andmortality.AmJ PublicHealth 1997;87:585-90.12. WorldHealthOrganization. nternationalClassification fDiseases,NinthRevision.Geneva,Switzerland:WHO, 1977.13. InternationalAgency for Researchon Cancer. TobaccoSmoking. Lyon, France: ARC, 1986. IARC MonogrEvalCarcinogRisk ChemHum, Vol. 38.14. FriedmanGD. Cigarettesmoking,leukemia,andmultiplemyeloma.Ann Epidemiol1993;3: 425-8.15. WinkelsteinW Jr. Smokingand cervicalcancer- currentstatus:A review.AmJ Epidemiol1990;131: 945-57.16. ArellanoM, PetersonG, Petitti DB, SmithR. The Cali-fornia automatedmortality linkagesystem. Am J PublicHealth 1984;74: 1324-30.17. FriedmanG, Sidney S, Polen M. Smokinghabits amongmultiphasic xaminees,1979to 1984.West]Med 1986;145:

651-6.18. SAS Institute, Inc. SAS User'sGuide. Cary,NC (USA):SASInstitute,Inc., 1990.19. Cox DR, Oakes D. Analysisof Survival Data. New York,

NY (USA): ChapmanandHall, 1984.20. HendersonBE, Ross RK, Pike MC, Casagrande T.En-dogenous hormones as a majorfactor in human cancer.CancerRes 1982;42: 3232-9.21. Noble RL. The developmentof prostaticadenocarcinomain Nb ratsfollowing prolongedsex hormone administra-tion. CancerRes 1977;37: 1929-33.22. HugginsC, Hodges CV.Studieson prostatecancer:Effectof castration,of estrogen,and of androgeninjection onserumphosphatasen metastatic arcinoma f theprostate.CancerRes 1941;1: 293-7.23. Gann PH, Hennekens CH, MaJ, Longcope C, StampferMJ.Prospectivestudy of sex hormone levels and risk ofprostatecancer. NCI 1996;88: 1118-26.24. Hsing AW.Hormones andprostatecancer:Where do wego from here?[Editorial] NCI 1996;88: 1093-5.25. Kolodny RC, MastersWH, Kolodner RM, Toro G. De-pression of plasma testosterone levels after chronicintensivemarihuanause. N EnglJ Med 1974;290: 872-4.26. Block RI, FarinpourR, SchlechteJA. Effects of chronicmarijuanase on testosterone,uteinizinghormone, olliclestimulatinghormone, prolactinand cortisol in men andwomen. Drug AlcoholDepend 1991;28: 121-8.27. Nomura AMY, Kolonel LN. Prostate cancer: a currentperspective.In: ArmenianHK, Gordis L, Levine MM,ThackerSB,eds. EpidemiolRev 1991;13: 200-27.28. Newcomb MD, Bentler PM. Consequences f AdolescentDrug Use. Newbury Park,CA (USA): SagePublications,1988:21.

29. Brinton LA. Epidemiologyof cervicalcancer- an over-view. In: Mufioz N, Bosch FX, ShahKV,Meheus A, eds.TheEpidemiology f Human Papillomavirus nd CervicalCancer.Lyon,France: nternationalAgency for Researchon Cancer,1992:119;4-6.30. National Institute on Drug Abuse. National HouseholdSurvey on Drug Abuse; Population Estimates 1985.Rockville, MD (USA): NIDA, 1987; DHHS Pub. No.(ADM)87-1539.

728 Cancer Causes and Control. Vol 8. 1997

sidney, stephen et al_marijuana use and cancer incidence

Documents