Tobacco Documents Online

Re.search. 1. The Analysis of Case-conuol Studie.s, pp. 192-244. Lyon, France: IARC, 1980. 14. Zlrang, Z. F., Morgenstem. H_ Spitv., M. R., Yu, 0. P.. Marshall. J. M.. Tashkin. D., Hsu, T. C., and Schantz S. P. Marijuana smoking and increased risk of squamous cell carcinoma of the head and neck. Cancer Epidemiol. Biomark. Prev., 8- 1071-1078. 1999. 15. Levois, M., and Swituq P. D,fferenlial eeposure misclassificatiun In caa- ontrol studies of environmenul tobacco smoke and lung cancer. J. Clin. Epide- miol.. 51: 37-54. 1998. 16. Nilsson, R. Ervirunmental tobaccu smoke xnd lung cancen a reappraisal. Eromaicnl. Environ. Saf, 34 2-1"!. 1996. 1]. Riboli, B., Preston-Manm. S.. Saracei. R. Haley. N. 1., Tdchopnulos, D_ Bcchcr. H., Burch. J. D.. Fantham, F.. T_ Gao, Y. T., and Jindal, S. K. Exposurc of nonsmoking women lo envlronmental tobacco smoke: a 10-eountry collaborative swdy. Cancer Causes Control, I: 243-252. 1990. 18, Cummings, K. M., Markello, S. J_ Mahoney, M., Bhargava, A. K, MeElroy, P. D., and Marshall, 1. R. Measuremem of current exposure to environmenlal tobacco smoke. Arch. Environ. Health. 45 r 74-29, 1990. 19. Jarvis, M. 1. Uptake of environmenral tobacco smoke. IARC Sci. Publ.. 81 : 43-58,198]. 20. Trichopoulos. D. Risk of lung eancer and passive smoking. Importznr Adv. Oncol., ]]-85r 1995. Cancer F.pidemiology, Biomarkers & Prerentiurr J6J9 21. Caporaso, N. Grneilcs nf smoking-relaled cancer and muwagen s.n•ISivily. 1. Nad. Cancer Insl . 91: IIYJ2-1098. 1999. 22 2. Cloos, 1., Nieuwenhuis. E. J.. Boomsmx, D. 1.. Kuik D. 1.. van der Sterre, M L., Arwen, F.. Snow, tJ. B_ and Braakhuls. R. 1. Inherited susceptibiliiy to hlenmycin-indaced chromaud breakv In culmred penphernl blood /vmphocyiec 1. Nml-Cancer Inar, 9/: 1125-1130, 1999- I 3. Inxba. S., Kunsu. Y_ Nagau. C.. Tokatsuka. N_ Kawakaml, N, and Sbin,izu. H- Associmions uf indlriJuals' health-relaied behavior with Iheir own or their spouses' smoking s~ams. 1. Epidembl_ 8 42-46. 1998. 24. Schanrz, S. P., Zhang, Z. F_ Spa. M. S.. Sun. M., and Hsu. T. C. Genctic suscepfiblllry to head and neck cancer: i aaion belween nutrition anA rnmagen sensiuvity~yngoxope, 107 265-78Le199"I. 25. SpLLz, M. R., Fueger. 1. J_ Halabr, S., Schan¢, S. P.. Sample. D.. and Hsu. T. C. Mutagen nshlvi~y i upper zerodigestive vact cancer: a ra.e-wnvol analysis. CaneerEpidemloL'Biomark. Prev.. 2: 329333, 1993. 29. Spia. M. R-, Hsu. L C. and Schantz. S- P- f netic and e emal interachons zs ri.sk, for aenvllgenive cancers. Adv. Exp. Med. BioLn320n31-34. 1992. 27 . Cloos, J_ Spitz, M. R.. Schanlz, S. P., Hsu. T C., Zhang. Z F., Toht, H.. Braakhuis, B. 1., and Snow. C. B. Genetic susceptibiliry lo head and neck squamous cell cercinonm. J. Nall Cancer Insi_ 88r 530-535, 1996

16,46 Environmental Tobacco Smoking and Head and Neck Cancer 2'ablu 2 Etimnted dfeeu of EIS (OR and 95%CI) on the risk of head and neck canar, by covanams ,eleemd for adjuslmenl Foo of ~ - . No. of onvols No corarlnm, - (crud.y - Packycarz of makine Packs~,ears of smokmg plus" ETS NeveP IU 27 10 1 fl 1 0 Eve 145 139 2.8(1.3~) .Ul '_I((19-5111 ?4(0.".8) Degree of ETS cxposure Ncvcr' 10 27 1.0 10 10 Modcrare 99 115 2.3(1 1-50I I-8(07-1 2) 2.1(076.) Heavy` > 24 5.3 Q.2_I2'ry 1,9 f I?_9 "II 36(l ..I-1 1.5) P for Irend OOrUI rkIX139 00249 ETS al home Nev :17 60 1.0 I.U LII (Rcasianxlly 41 46 14(O8-261 1.4(Oi=7) 1.6(0.8-33) Regularly 77 60 2.I(13-391 13(0.9--'.8) 1](0.8-33) P for bend 0 OOh9 (1 1596 01574 tifS at work Never 36 55 10 10 10 Occasionally 46 50 1.4(0N-25) 12il1Cr231 10(05-2.1) Regularly 23 61 1.8(I1-3.11 14(0.8-22) 1.0(0.5-2.1) P for trend 000288 11 2380 09240 Spnuse smoking No 80 1Uf 10 I 0 1.0 Yes 33 22 2.0(LI-3.6) I610-8-3.1) IJ(08-3]) ° Also adjusted for age (confinuous varlable), gerMer (male, 0: female. I): race (white. 0: non-whilra I): etlucauon 1=h-h hool. 0sulle¢_e. 2: >rollege. 2); heavy sleohol se (< I00/month. 0, a100/monlA, q: an<I nwdjuana use (no, 0: yes. I). ^ Never exposed to ETS at both honre and wurA. ` Regularly expused m ETS at both home and work. Table 3 Estimated rrudc effecls (OR and 95% Ct) of Er5 on IAe risk of head and neck cancer. bv ucGve elgarette emoking Never smoking Ever smoking Cases Contrvls ETS Never Ever Degree of ETS Never Moderate Heavy P fortrend ETS at home Never Occasbnal ly Rcgularly P for trend ETS ar work Never /kcasianally Regularly P for trend Spousc smoking No Yes 3 1 1 23 46 3 13 17 4(1 6 6 8 28 9 10 9 21 7 24 I I 17 8 I S II 36 2 7 ijuana only) + 2.6 (ETS only) - I= 5.1 < ORe, (95% CI) r'.r,es Controls OR~l (95% Cp L0 6 w 1 0 Z 2(0-".4) 121 92 11 tl.l-8.3) 1,0 6 14 1.0 L8f0.5-7.3) 80 74 2S((L9-b -9) 43(0.8-235) il 18 5.3(18-16.1) 0.0082 0.0016 1.0 28 32 1 0 32 (1 0-10 4) 92 35 1.1 (05-2.1) t5(0.5-0.5) 62 39 2.0(1.0-32) 0.4483 0.0264 1.0 2I }0 1.0 2.2 (0."l~.9) ~5 33 1.2(0.6-24) 15005 .-5.0) 65 43 L"l(09-31) 0.4670 0.U997 10 68 67 1 0 09(0.2-9.2) 31 15 2.0(1.0-sE0) 7.1 (both ezpo- sures) < 3.5 x 2.6 = 9.1. In each case, however, the power for testing each null hypothesis (effccts are additive or multiplica- tive) and for comparing the fits of additive versus multiplicative models was low. Discussion We report here an effect of exposure to ETS on the risk of head and neck cancer. Not only did we find an elevated cancer risk among individuals ecposed to ETS, but we also observed dosc-response associations for the degree of ETS exposure, adjusting for several potential confounders. These associations and the dosc-response relationships were still present when the analysis was restricted to non-active-smok- ing cases and controls- This studv has several possible limitations. One limitation is potential selection bias, which might have resulted in an overestimate or underesnmate of the ETS effect (bias away from or toward null). Using controls from the blood hank may result in potential selection bias because blood donors might be

4'"r V. IUJI-IUJY. V~ «~Le, _>(XM) a,,r. Ilnini:,rkcr~ .\ I're. enliall l(IJI Environmental Tobacco Smoking, Mutagen Sensitivity, and Head and Neck Squamous Cell Carcinoma' Zuo-Feng Zhang," Hal Morgenstern, Margaret R. Spitz, Donald P. Tashkin, Guo-Pei Yu, Tao C. Hsu, and Stimson P. Schantz Depanmenl of Epidemiology, Unirersity of Califomia tus Angeles School of Public Heslth, and Jonsson Comprehensive Cancer Center, tnz Angeles, California 90095-1772 [Z-F. Z., H. M]; Department of Epidemiology. The University of Texas M. D. And<rsbn Cancer Ccnter. Housmn. Texas 77030 [M. R. S.]: tbpunnent of Medicine. University of California Lot Angeles School of MeAicine, Ins Angeles. Califomia 9t1095 [D. P. T.I: New York Eye and Ear Infinnary, New York, New York 10003 [G-P. Y., S. P. S.I; and Department of Cell Biology. The University of Texas M. D. Anderson Cancer Cemm,. Houston, Texas 77030 (f. C. H.I Abstract Although active tobacco smoking has been considered a major risk factor for head and neck cancer, few studies have evaluated environmental tobacco smoke (ETS) and its interaction with mutagen sensitivity on the risk of head 'and neck cancer. We investigated the relationship between ETS and head and neck cancer in a case-control study of 173 previously untreated cases with pathologically confirmed diagnoses of squamous cell carcinoma of the head and neck and 176 cancer-free controls at Memorial Sloan-Kettering Cancer Center between 1992 and 1994. A structured questionnaire was used to collect ETS exposure and other covariates including a history of active tobacco smoking and alcohol use- ETS measures include a history of ETS exposure at home and at workplace. The associations between passive smoking and head and neck cancer were analyzed by Mantel-Haenszel methods and logistic.regression models. Additive and multiplicative models were used to evaluate effect modiffcations between ETS and mutagen sensitivity. The crude odds ratio (OR) for ETS exposure was 2.8 [95% confidence intervals (CI), 1-3-6.0]. Controlling for age, sex, race, education, alcohol consumption, pack-years of cigarette smoking, and marijuana use, the risk of squamous ce6 carcinoma of the head and neck was increased with ETS (adjusted OR, 2.4; 95% CI, 0.9-6.8). Dose-response t-elationships were observed for the degree of ETS exposure; the adjusted Received 9/8/99; revised )n/00: acapled 2!L/00. "rTe costs of publication of this article were defrayed in part by the paymem of page charga. This anicle musl therefore be hereby mukad adveMaemenr In accordanee with Ig U.S.C. Secllon 1734 solely to indicate this fact. ' Supponed in pan by Granr FS0d)18 from the National Institute of Environ- mcntal Heahh Services; Grants CA-51845, and CA16042-24 fron, Ihe Naional Cancer Instimte; Granl DA/CA113g6 from the Natlonal Inaliwta on Dmg Abuse. NIH. Department of Health and Humaa Services: by a serd grant by UCLA lonsson Cancer Center Foundation; and by the Weisman Fund. z To whom requests for reprinls should be addressed, at Depanmem of Epidemiology. Universily of Callfornia tns Angeles School of Public Heallh,'H-225 CHS, Box 951772, Los Angeles. Califomia 90095-rR2. Plmne: (310) 825-8418: Fa. (310) 206-6039: E-mail. zfzhang@ucla cdu ORs were 2,1 (95% Cl, 0.7-6.1) for those with moderate exposure and 3.6 (95% CI, LI-11.5) for individuals with heavy exposure (P for trend = 0.025), in comparison with those who never had ETS exposures. These associations and the dose-response relationships were still present when the analysis was restricted to nonactive smoking cases and controls (crude OR, 2_2; 95% CI, 0.6-8.4). Crude odds ratios were 1.8 for those with moderate ETS exposure and 4.3 for individuals with heavy ETS exposure among nonsmoking cases and controls (P for trend = 0.008). More than multiplicative interaction was suggested between passive smoking and mutagcn sensitivity. This study suggests that ETS exposure may increase the risk of head and neck cancer with a dose- response pattern. Our analysis indicated that passive smoking may interact with mutagen sensitivity and othcr risk factors to increase the risk of head and neck cancer. Our results need to be interpreted with caution because of potential residual confounding effects of active tobacco smoking and other methodological limitations. Future large-scale studies are warranted to confirm our findings. Introduction ETS' is generated by exhalations of smokers, plus emissions from cigarettes, cigars, or pipes between puffs. Approximately 80% of ETS comes from SS and 20% from exhaled MS. Carcinogens in SS are much higher in ponccntration than in MS, such as N-nitrosamines (20 fold), 4-antinobiphenyls (30- fold), 2-naphthylamine (30-fold), aniline (30-fold), benzene (5-fold), benzo(a)pyrene (3-fold), and 4-(methylnitrosoamino)- l-(3-pyridyl)-I-butanone (2-fold). Strong evidence linking ETS to carcinoma of the lung cancer has been documented by a series of well-designed and conducted epidemiological studies, as summarized by Dockery and Trichopoulos (I). ETS was classified as a human carcinogen (group A) by the United States Environmental protection Agency in 1993 (2)- A recent large- scale multicenter epidemio]ogical investigation by Boffetta er at. from the Intemational Agency for Research on Cancer suggested a moderately elevated risk of lung cancer among nonsmokers exposed to ETS as adults at home or in the work- place, with risk tending to rise with the amount of ETS expo- sure (3, 4). The association belwcen ETS and lun_e cancer and the similariUes between the carcinogenic pmperties of passively and actively inhaled tobacco smoking suggest the possible role of environmental tobacco smoke or passive smoking on the development of squamous cell carcinoma of the head and neck. Although tobacco smoking and alcohol drinking havc been identified as major risk factors for head and neck cancer, few 1 llm abbreviaunn. n+ed arr F.TS. e ral ioba¢ o svunke SS. +Aesneam +mnke MS_ mainaurczm .mokc. OR Vodds ratio_ CL ronGdumc inie~.al

statistical adjustment for pack-years of cigarette smoking (con- tinuous variable); (c) statistical adjustment for pack-years of cigarette smoking plus age (continuous variable), sex (male versus female), race (white versus non-white), education ('high school, college education, postgraduate education), his- tory of marijuana use (yes versus no), and hcavy alcohol drink- ing (?100 drinks/month versus <100 drinks/month;). Strati- fied analysis was used to assess departures from additive or multiplicative effects between ETS and other known risk fac- tors for head and neck cancer, including cigarette smoking, alcohol drinking, and mutagen sensitivity. Results The overall prevalence of ETS exposure was 83.7%% in controls and 93.6% in cases. Higher ETS exposure was found in 95.7% of patients who had laryngeal cancer, 95.7% of pharyngeal cancer, and 92.7% of carcinoma of the oral cavity. The distri- butions of ETS exposure, stratified by demographic character- istics, active cigarette smoking, alcohol drinking, mutagen sen- sitivity, and marijuana use, among cases and controls and the distributions of ETS exposure in cases stratified by sites of head and neck tumor are shown in Table I. No obvious differences for passive smoking were found in terms of age, gender, and race. Education was strongly associated with ETS exposure; high prevalence of the ETS exposures was found in people with lower education (less than or equal to high school) in controls, but not in cases. Cigarette smoking was generally independent of ETS exposure in both cases and controls, although the prevalence of ETS exposure was elevated with increased pack- years of smoking, cigarettes smoked/day, and years of smoking in conuols. Heavy alcohol drinking and mutagen hypersensi- tivity were not related to ETS exposure in cases or controls. Marijuana use was associated with passive smoking in cases but not in controls. The estimated crude OR for the effect of lifetime ETS exposure (ever versus never) on the risk of head and neck cancer was 2.8 (95% CI, t.3-6.0). Adjusting for age, gender, race, education, heavy alcohol drinking, pack-years of cigarette smoking, and marijuana use, the OR was 2.4 (95% CI, 0.9-6.8; Table 2). Strong dose-response relationships were observed for the effects of the degree of ETS exposure. The adjusted ORs were 2.1 for those who had moderate ETS exposure and 3.6 for those who had heavy ETS exposure (P for trend test = 0.025). When ETS exposure at home, ETS exposure at work, and spouse smoking were analyzed separately and after controlling for potential confounding effects, the ETS effect for each vari- able was weaker in comparison to results of crude analysis. The observed associations between ETS exposure and head and neck cancer in never-cigarette-smokers were similar to those in cigarette smokers (Table 3). The cmde OR for the effect of lifetime ETS exposure was 2.2 (95% CI, 0.6-8.4) for never-smokers and 3.1 (95% Cl, LI-8.3) for ever-smokers. Strong dose-response relationships were also observed for the effects of the degree of ETS exposure among never-smokers and ever-smokers. The crude ORs were 1.8 for those who had moderate ETS exposure and 4.3 for those who had heavy ETS exposure (P for trend test = 0.0082) among never-smokers. The crude ORs were 2.5 and 5.3 for moderate and heavy ETS exposure, respectively, among ever-smokers (p for trend = 0.0016). Table 4 shows the combined effects of ETS exposure (ever versus never) and each of four potential effect modifiers: to- bacco cigarette smoking, alcohol use, marijuana use, and mu- tagen sensitivity. For these analyses, we used > 1.0 breaks/cell Cancer Epidemiology, Hiomarkers & Prevention I415 Table !'1'he prevalence of El'S ,n casc. nnd connols, by cxagory nf seleaed deono¢mphm racnors. smoking. alcnhnl. znd mutagen tenxirlvily Cammt. Czses Ycx ou ` Towl YETS,r) 9 Totol Tolnl 119 S3 7 166 145 93.6 155 Age <60 68 840 81 64 928 69 60{r9 14 83.0 53 4? 956 45 ?70 -17 . 84 3 30 38 92.5 41 Genda:r Male 90 84 9 106 92 949 97 Female 49 817 60 53 91.4 58 Rzce While 12' 94 7 150 129 949 136 Non-whilc 11 7511 16 16 84 2 19 kAUCanon sHigh school a~ 97 7 44 89 93 7 95 Collage 65 "/5.6 86 41 976 42 Postgmdume 31 86.1 36 14 87.5 16 Cigarc[m smoking Never 46 78 0 59 23 88.5 26 Ev<r 9? 86.8 106 121 953 127 Quit 60 85 7 70 36 97.3 37 CurrenY' ]2 889 36 85 944 90 Paek-Years 0 +6 78 0 59 23 885 26 I-22A 40 8i3 48 20 95.2 21 22.5~4.9 26 929 28 36 9] 3 3"/ ?45 1] 944 Itl 57 93.4 61 Marijuana me No 126 846 149 126 9f,? 131 Yes I l 76.5 I l 19 "I9.2 24 Alcohol use (drinks/monlhs) <100 Iltl 81 1 1 142 N9 93.7 95 1100 14 933 I5 51 927 55 Mwagen ser.shiviry (Meaks/celb <l 7S 82.1 95 29 90.6 32 ?I 25 862 29 46 939 49 Tamor sires (1cD9r Lip(140) , 1 500 - Tongue (141) 43 91.5 47 Gum (143) 10 90.9 1 1 Floor of mauni (144) 14 1000 14 Other parrt of muuth (145) 9 90.0 10 Oropaarynx(146) 10 90.9 1 1 Nzsopharynx (14]) 2 100.0 3 Hypupharynx (148) 10 IOQO 10 Other oral caviry (149) I 100.0 I Esophagus (I50) I 1n4n I I.arynx (161) 44 95.7 46 "Clrrenr smoking category included these whu were slill smoking and people whn had quit smoking for <5 years. as the cutoff value to define mutagen hypersensitivity. We categorized subjects as heavy alcohol drinkers if they had 100 or more drinks/per month and tobacco cigarette smoking into never smokers/ever smokers. The effects of ETS and mutagen hypersensitivity were more than multiplicative; the adjusted OR for the joint category of ETS exposure and mutagen sen- sitivity was greater than the product of the two component effects for those two factors, i.e.. 17.5 > 2.6 x 2.0 = 5.2. The effects of ETS and alcohol consumption appeared more than additive but less than multiplicative, i.e., 4.9 (alcohol only) + 2.5 (ETS only) minus l= 6.4 < 10.2 (both exposures) < 4.9 x 2.5 = 12.2i The effects of E15 and marijuana use appeared more than additive but less than multiplicative, i-a. 35 (mar-

Cancer F,pidemioloKf• aiumarkers & Prevention I(W) Tnhle 4 Fstimated combined effects (OR and 95% Cl) of ETS (ever verrus neveq and each of four polcmial niodifier, fugxrene smoking, heavy alcohol use. marijuana srnoking, and mtagen hypersensitivity) on the risk of head and neck canrvrz, by c r.arla¢s selected for adjuslmanl Passive smoking Potential modifier No. of cases Nao of controls No covariales Icrude) Pack-ycars of smoking Cigareue smoking No Never 3 13 1.0 1 0 I I) No Ever 6 14 1-9(04 -90) 09103-4 91 04 I0 1 271 Yes Never 23 4(1 2.2(It6-8.4) 22(0-6-B4I 1.5/03-6 51 Yes Ever 121 92 53 (1 6-20 -6) 19(t1.5-0.4) 11(U361) Alcohol (tlrinksimonth) No <Ilx) 6 24 1-0 1D 10 No a1(q 4 I 16.0 (1.5-170) 6.4 (0.5-83 5) 49(0 ?-75-81 Yes <100 89 11g 3.0(1.2-9]) 1 -0(0.8-54) 2.51n.ft-761 Yes ~1W 51 14 14-6 (5 0.4261 7.8 12.5-24.11 1 0 2(2 7 37.8) Mutagen sensitivity (brcakslcelp No <1 3 I) 1A 1 0 1 0 No al 3 4 4.3(0b-29.5) 2.9(0.3-30.31 2b(0-1-71-I) Yes <I 29 78 2.1(0 6-22) 1.6(03-8-8) 20002-17.71 Yes ? 1 46 25 104(2.8-39.1) II-6(2.1-630) 1 7 .5(19-I(121 Marijuana use No No 5 23 1-0 1 0 1.0 No Yes 5 4 5.8(1.I-29.fi 3-6(06-21.2) 35 (0~214 4) Yes No 126 126 4-6(U/-125) 3-1(Lo-9.8) 2.6((1J-9.0) Yea Yes 19 13 6.9(20-22.3) 5-8(15-22.8) 9-111 1-345-i "Also adjusted for age (continuous vadable): gender (male, 0; female, 1); race (white, 0: non-white, I): education (=high schouL 0: college. 2: >college. 2): hea.y alcohol use (<100/month 0: ?I00/month. I): and marijuana use (no, 0: yes, 1)) more health oriented. To assess the potential selection bias, we have compared selected demographic and potential risk factors between blood donor controls and non-cancer controls from the Surgical Day Hospital during the same study period at Memo- rlal Sloan-Kettering Cancer Center. Generally, the two groups of controls were similar in terms of those selected factors; contrary to expectation, however, blood donor controls had a slightly higher proportion of cigarette smokers and a markedly higher proportion of alcohol drinkers than the other control group, potentially producing an underestimation of the associ- a[ion between cigarette smoking, alcohol drinking, and head and neck cancer and may lead to incomplete control of these potential confounders. Because we did not collect information on passive smoking in non-cancer controls from the Surgical day Hospital, we did not know whether the prevalence of passive smoking in blood donors would also be higher than non-cancer hospital controls so that the selection bias could not be clearly ruled out. On the other hand, because the majority of blood donors were relatives or friends of cancer patients hos- pitalized at Memorial Sloan-Kettering Cancer Center, they might have a slightly higher chance of exposure to ETS. We believe that the selection of blood donors as controls would probably underestimate the association between ETS and head and neck cancer under study. When we evaluated the interaction between passive smok- ing and mutagen sensitivity (Table 4), a possible selection bias might exist because those with blood samples for mutagen assay may be different from individuals without blood samples. A total of 26.1% of controls and 46.8% of cases refused to provide a blood sample for the bleomycin sensitivity test in this study. We have compared the differences between those with and without blood samples on selected variables. This attempt is crucial to show whether there is selection bias because of missing samples that may threaten the validity of the interaction between passive smoking and mutagen sensitivity. The propor- tjon of passive smoking was comparable in those with and without blood specimens for both cases and controls. No ob- vious difference was found between those with and without blood samples in terms of age, gender, education, and alcohol drinking in both cases and controls. Significant differences were found for cigarette smoking and race in cases; cases with blood specimens had a higher proportion of smokers and non- white than cases without blood specimens. Those differenccs indicate that the subjects with blood samples might not be a representative group for smoking habits and non-whites froln the original study population, which might lead to a stronger confounder effects on the association between passive smoking and head and neck cancer. The second limitation is differential misclatsification of ETS, which may bias the estimated effect under study (15, 16). The degree of overreporting may be greater for cases than controls because cases might want to rationalizetheir disease. Thus, the estimates of ETS effects could be positively biased. Self-reported ETS in the recent past has been validated in several studies and believed to be apparently valid (17-19). The confirmation that dose-response relationships.exist between urinary cotinine concentrations and self-reported passive smok- ing partially validates questionnaire measures of the degree of environmental smoke exposure (19). The results of the analysis of self-reported recent exposure to ETS from any source in relation to urinary concentrations of cotinine indicated that duration of exposure and number of cigarettes to which the subject reported being exposed were strongly related to urinary cotinine (17). However, questionnaire-based infolmation on long-term exposure to ETS is difficult to integrate over time and almost impossible to validate. Nevertheless, the validity of self-reported exposure to ETS in the recent past supports the validity of self-reported of long-term exposure to ETS (20). Thus, we believe that differential misclassification of past ex- posure to ETS is probably not sufficient to explain the positive findings in this study. The possible differential misclassifica- tion of using mutagen sensitivity assays in case-control study was discussed by Caporaso (21). Cultured cells obtained from patients with cancer or control subjects in a hospital setting can differ for abnormal nutrition, secondary metabolic alterations of neoplastic disease, and effect of treatment, hospitalization, in-

1048 Envirunmenwl Tubacm SnrokinR and Ilead and Neck rlancur activity, or stress, which will allow bias because of differential misclassification. However, a recent paper by Cloos et at. (22) reported a high heritability estimate of the susceptibility to bleomycin-induced chromatid breaks, which indicates that a clear genetic basis for mutagen sensitivity-related cancer sus- ceptibilily may exist in the general population. If the mutagcn sensitivity is highly inherited, Ihe differential misclassification bias for Ihis assay might be minimal. The third limitation is the small sample size_ We only have 10 cases and 27 controls who had no ETS exposure. If we further stratified by cigarctte smoking or mut:ation sensitivity, the number becomes much smaller. The relatively small sample size may lead to the low power of ttte study and a poor precision of the measurement, which would limit our ability to estimate the ETS effect effectively and precisely. Confounding by active cigarette smoking and alcohol con- sumption on the association betueen ETS and head and neck cancer was apparent Although we attempted to adjust for active cigarette smoking and alcohol drinking in our analyses, residual confounding might still exist because ETS may be closely related to active cigarette smoking and alcohol con- sumption. The observed association between exposures to ETS and head and neck cancer is relatively weak, s[milar to the observed association between ETS expxure and lung cancer. In com- parison with the ETS effect in lung cancer, the OR of ETS for head and neck cancer is slightly higher. It may be caused by the small sample size of this study. On the other hand, considering the upper aerodigestive tracts as a first entrance for the ETS exposure, the degree of exposure might be higher than that in lung. In addition, the mechanism of the ETS carcinogenic action may be different in upper aerodigestive tract cancers from that in lung cancer. Our results are supporte(1 by compelling biological evi- dence. This evidence includes the higher concentration of car- cinogens in SS than MS, the strong causal link between active smoking and both lung and head and neck cancers, and the convincing evidence of the association between ETS exposure and lung cancer (I)_ Although non-cigarette smokers are major potential vic- tims of the health consequences of ETS, active smokers might also have a greater opportunity to be exposed to S5, in addition lo MS- Therefore, if ETS is associated with certain cancers in nonsmokers, it would be reasonable to assume that ETS would have a similar or even stronger impact on the risk of tobacco- related cancers in smokers when active cigarette smoking and other potential confounding effects are controlled for. The hypothesis is supported by the following: (a) cigarette smokers are exposed not only to MS but also to SS from their own cigarettes; (6) smokers tend to socialize with other smokers, thereby increasing their exposure to other smokers' SS; (c) smokers are more likely than nonsmokers to have a smoking spouse or partner, thus further increasing their exposure to ETS (23). Because most published studies of ETS and lung cancer are limited to nonsmoking wamen, we may not know the full impact of ETS on the risk of lung cancer. The effects of ETS need to be further studied in nonsmoking men, as well as in active smokers. In the present study, we found slightly different effects of ETS on the risk of head and neck cancer between smokers and nonsmokers, In never-smokers with 26 cases and 59 controls, the crude OR was 2.2 for ETS exposure. The dose-response relationship was apparent with ORs of 1.8 for intermediate ETS exposure and 4.3 for heavy ETS exposure (P for trend = 0.0082). Interestingly, the adjusted ORs for BTS exposure in the whole studv population, including both active smokers and never-active smokers, were very similar to those in the subanal,vsis with never-active smokers: 2.4 for ETS exposure, 2.1 for moderate ETS exposure, and 3.6 for heavy ETS expozure, respectively (P for trend = 0.0249). These observations arc cortsistcnt with our assumption. Pnssible interaction effects were suggested between ETS and mutagen sensitivity and other risk factors for head and neck cancer, The interplay between carcinogens and intrinsic host susceptibility is an important factor in environmental carcino- genests. Mulagen Itvpersensitivity, an indirect marker for DNA repair, intemcts with tobacco smoking in head and neck cancer risk (24-27). Svnergy between mutagen hypersensitivity and ETS was suggested in this study because the effects were much more than multiplicative, which suggests that Ihe development of the uppet aerodigestive cancers may be affected by gene- environment interaction_ Because of the low power for testing Ihese interactions, however, the present findings will need to be replicated in future large studies. [n summary, we found that ETS is associated with a dose-depcn(lent increased risk of head and neck cancer. This association is supported by other evidence that provides a biologically plausible basis for the hypothesis that ETS is a risk factor for human head and neck cancer. Our results need to be examined with caution because of potential residual confound- ing effects of active tobacco smoking and small sample size. Further large-scale epidemiological studies are needed to rep- licate our results, to examine the relationships between ETS and increased risk of cancer, and to assess potential interactions between ETS and olher risk factors. References I. t)oekery, D. W., and Tdchopoulos. D. Risk of Inng cancer from mvirnnmemal cxposuree lo iohacco vmoke. Cancer Causes Control, 8r 333-345, 199]. 2. Unired Snmcs Enrironmemnl Protection Agency_ Respiratory hcalrh effecns of passive moking: lung cancer and orher disorders. Washingmn, Dc: Envlrom menlal Prmecnon Agency, 1992. 3. Bofferta, P , agudo. A_ Ahrens, W, Benhamou. E.. Benhamou. S., D6arby. S. C_ Feno. Q. Fortes. C_ Gonzalez, C. A., Jockeh K H., Krauss, M., Kreien- brnck, L, Kreuzer. M. Mendes. A., Merleni, F.. Nyberg. E, Pershagea, G., Pohlabcln. H.. Riboli, F., S_hmid, G., Simonato, L., Tredaniel. 1., Whitley, E.. W,chmaru,, H_ G., and Saraccl, R. Mulricenl<r case-comrol study of exposure to environmemal mba¢u nmoke and lung cancer in Europe.). Na0_ Cancer Insr, 90: 1440-1450.199g 4. Illor W. J, and Mcl aughlln.l. K. Passive smoking and lung cancer ri.sk: what is the story now? 1. NaU. Cancnr InsL. 90r 1416-141"/, 1998. 5, Srrome. M_ W ard- P.. Jolmson. 1., and Goepfert. H. Carcinoma of Ne mnsil. Head Ncck,/5 a65-46N,1993. 6. Tan. E. H_ Adelsiein. D. 1., Dtoughmn, M. L, Van Kirk, M. A.,and Lzvenu, P. Seryamuu. cell Lezd and oeck cancer in nonsmokers. Am. J. Clin. Oncol_ 20: 146-150,1999. 7 . Spltz. M R.. and Hsu, l. C- Mumgen sensitivi, as a markcr of cancer nsk. Cancer Derea. Prev, J8S 299-303, 1994. 8. Oldor, K. Muragen hypersensirieiry as a biunarker of genetic predisposition to carclnogenesvs.l. Natl Cancer Inst_. 86: 1660-1661, 1994. 9 Hw, T. C'_ Johnsmn. D. A., Chetry. L. M.. Ramkissoon. D.. Schanlz, S. P., Jessup. I M_ R9nn. R. 1. Shirley, L., and Furlung, C. Sens[riviry to genaoxic. effac¢ of bleomycin in bumzns: possible relationshlp lo enCuonmental carcinu- genesis. Inr 1. Cancer, 4 t: 403-409, 1989_ In. Cherry, L. M.. ano Hvr T. C. nteomycin-induced chromosome damage in lymphocyms of inedullarti ~,yroid casinoma pmieme and their family members. AnGCZncer Kes., j: 361-:i 2 1983_ I L Hsu,'1'. (:.. Cheny. L M., and Sanraaq N. A. Differential mmagen suscep- tibilily in culeur.d lymphocyres of normal individuals and cancer parients _ Cancer Genel. Cylogenet., !]: 30?-313. 1985. 12. Hsu, T. C. Genetir predisposition ro cancer wlrh special referenra to mmzgen sensiliv,ry. in viva Cell Dev. Biol.. 23. 591-6(li 1987. 1]. Breslow N. E, ard Day, N. E. Unconditional logistic regression 6x large strata JnN F. Bresln.w and N. E. Day reds.J, Searisucal Melhods in Cancer

1041 F.nvironmenUl Tobaam Snraking nnd Iluad antl Neck fanrer clinical or epidemiological studies have been conducted to examine the possible effects of passive smoking on the dsk of head and neck cancer- One of those studies was a case report involving a 69-year-old nonsmoking woman with squamous carcinoma of the left tonsil. Her illnes.s was suspected to be related to her husband's 40 ycars of smoking (5). In another study of the possible role of passive smoking in the develop- ment of squamous cell head and neck cancer (SCHNC), a group of 59 non-smoking cases with squamous cell head and neck cancer were retrospectively studied at the Cleveland Clinic Foundation (6)- When compared with the individuals without ETS exposure, an elevated risk of head and neck cancer was noted for those exposed to ETS both in the home and in the workplace. The authors concluded that a cause and effect re- lationship between passive smoking and head and neck cancer cannot be clearly established on the basis of their study because of the small sample size, the limited data collection, and inad- equate adjustment for potential confounding effects in their analysis. ' The aim of the present analv.eis was to examine the asso- ciation between ETS and head amd neck cancers, controlling for other known risk factors for the disease, including cigarette smoking and alcohol drinking. We will also explore the possi- ble interactions between ETS and mutagen sensitivity, and with other known risk factors for head and neck cancer. Mutagen sensitivity is considered a predisposition marker of cancer risk (7, 8). Defects in one or more steps of the DNA repair process may play a significant role in environmental carcinogene,sis, and the extent of such defects mav be partially responsible for susceptibility or resistance to environmental mutagens (9). Mu- tagen sensitivity tests are indirect indicators of DNA repair competence. Bleomycin, a radiomimetic agent, was used as the test mutagen, an assay developed to evaluate the rates of in- duced chromosome breakage as a crude indicator of the re- sponse to a genotoxic agent (10, I I L Patients and Methods Subject Selection. Untreated new patients with a histologi- cally confirmed diagnosis of first primary squamous cell car- cinoma of the head and neck, seen at Memorial Sloan-Kettering Cancer Center from 1992 to 1994. were considered as cases in this study. We have approached 192 patients, and 173 agreed to participate- Eleven tumor sites were classified by the American Joint Committee on Cancer criteria. Age and sex frequency- matched controls were identified for this study. Controls were without a history of cancer and were identified from the Blood Bank Center of Memorial Sloan-Kettering Cancer Center dur- ing the same period. We approached 196 blood donors, and 176 agreed to participate the study. Data Collection and Variables. "I'he study was approved by the Institutional Research Board on Human Subjects of Memo- rial Sloan-Kettering Cancer Center. All cases and controls were asked to sign an informed consent form if Ihey agreed to participate in the study, to completc a structured questionnaire. and to donate a sample of blooai The questionnaire requested infonnation on the following variables: age, gender, race, year and place of birth, religion, family income, and education; average number of tobacco cigarcues smoked/day, years of smoking, age at initiation of smoking; exposure to ETS (at home and at work); alcohol consumption, types and frequency of alcohol consumption; occupational and environmental ex- posures; family history of cancer, sexual history; medical his- tory; and oral hygienic history. In addition, all subjects were asked if they had ever used mariryana. If they responded ves, suhjects were asked the avcrage number of times they smakcd per day and the nurnber of years of marijuana use- For ETS exposure, all suhjects were asked: (a) "Have you ever been regularly exposed to other people's cigarette sntokc at home7" and (6) "Have you c•er been regularly exposed to odter people's cigareue smoke at work?" There were three choices for each questton: neva- occasionally, and regularly- In addition, a question regarding partner's or spouse's smoking was asked: "Does yuur current panner/spouse smoke? (yes/no)"- A total of 155 cases (89,64b) and 166 controls (94-3%) had complete data on passive smoking at home and at work. Individuals with missing data on passive smoking were excluded from analysis- Those rcported either occasional or regular exposure to ET'S at home or work uere categorized as ever exposed to ETS, and those who rcported no exposure to ETS both at home and at work were defined as never exposed to ETS- The degree of ETS exposure was defined according to history of passive smoking both at home and at work. "Never exposed to ETS" was defined as those individuals who were exposed to ETS neither at home nor at work, "heavily exposed to ETS" was defined as those individuals who were exposed to ETS both at home and at work, and "moderatelv exposed to ETS" included those who were either exposed ro ETS at home or at work. Mutagen Sensitivily 4ssay. A total of 91 patients and 131 controls provided a blood specimen for the assessment of mutagen sensitivity. "Die mutagen sensitivity assay used in this study has been descrihc:d in derail previously (12). A peripheral blood sample (!0 nd or less) was collected from cases and controls in a heparinized tube prior to initiation of lymphocyte culture.'[he standard lymphocyte culture procedure used RPMI 1640, supplemented with 15% FCS and phytohemagglutinin, in a ratio of hla)d:medium of 1:9. At 67 h of incubation, one set of cultures was treated with bleomycin (0.03 units/ml) for 5 h. Colcemid (0.tM mg/m]) was added in the last hour to induce mitotic arrest prior Io harvesting. A conventional cell harvest- ing procedure followed. The cells were treated with hypotonic KCl (0.975 M KCI) solution for 15-20 min, fixed, washed with a freshly prepared imxture of methanol and acetic acid (3:1), and air-drled on wet slides. The slides were stained with Giemsa solution without banding. Fifty well-Spread metaphases were examined from coded slides. Chromatid aberrations re- corded were frank chromatid breaks or exchanges. Bleomycin tends to induce few chromatid exchanges (which, if present, are considered as two breaks). Chromatid gaps or attenuated re- gions were disregarded_ The frequency of breakage was ex- pressed as hrcaks/celL The reliability of cytogenetic scoring has been evaluated previously by comparing four separate blood sarnples from a respective donor with a minimum interval between samples of I week (9). Statistical Analysis. "lhe effects of ETS on the risk of head and neck cancer were estimated with ORs and their 95% C[s, derived from logistic regression analysis (13). Dummy vari- ables were used in logistic regression analysis to estimate ORs for each category of exposure. Trend tests for ordered variables were performed by assigning the score j to the jth exposure level of a caicgorical variable (where j= 1, 2, ..-) and treating the categorical variable as an interval predictor in unconditional logistic regression. We have selected several potential con- founders or effect modifiers in our analysis. In addition to active tobacco smoking and alcohol drinking, we considered marijuana smoking as a possible confounder or a effect mod- ificr in our analysis because marijuana smoking was associated with head and neck cancer (t4). Three models were used to assess ETS effeetr ta) no cuvariazes (crode analysis): (b)