REFERENCES (1) Christen, A.G., Swanson, B.B., Glover, E.D., and Henderson, A.H. Smokeless tobacco: The folkIore and social history of snuffing, sneez- ing, dipping, and chewing. J. Am. Dent. Assoc. 105: 821-829,1982. (2) Schuman, L.M. Patterns of smoking behavior. In: M.E. Jarvick, J.W. Cullen, E.R. Critz, T.M. Vogt, and L.J. West teds.). Research on smok- ing behavior (NIDA Research Monograph 17). U.S. Government Print- ing Office (Stock No. 017-024-00694-7), 1977. (3) Shelton, A. Smokeless sales continue to climb. Tobacco Reports, p. 42. August 1982. (4) Maxwell, J.C., Jr. Chewing, snuff is growth segment. Tobacco Reports, p. 31, September 1980. (5) Winn, D.M. Tobacco chewing and snuff dipping: An association with human cancer. In: I.K. O'Neil, R.C. Borstel, C.T. Miller, J. Long, and H. Bartsch (eds.). N-Nitroso Compounds Occurrence, Biological Effects and Relevancy to Human Cancers. IARC Scientific Publication No. 57, Oxford University Press, 1985. (6) Rosenfeld, L., and Cahaway, J. Snuff dipper's cancer. Am. J. Surg. 106: 840-844, 1963. (7) Agricultural Market Service. Tobacco in the United States. U.S. Department of Agriculture (Miscellaneous Publication No. 867). 1979. (8) Davis, D.L. Smokeless tobacco products and their production in the United States. Presented at the NIH Consensus Development Confer- ence on the Health Implications of Smokeless Tobacco Use, Bethesda, Maryland, January 13-15, 1986, and personal correspondence. (9) Bantle, L.F. Smokeless tobacco-A trend to watch. Ihbak J. Int. 4: 344-346, 1980. (10) Rizio, D. Smokeless tobacco. Thbak J. Int. 2: 183-184, 1984. (ll/ U.S. Department of Agriculture, Economic Research Service. Tobacco: Outlook and Situation Report. Washington, D.C., 1985. (12) U.S. Department of Agriculture, Agricultural Marketing Service. Annual report on tobacco statistics, 1973. (Statistical Bulletin No. 528). Washington, D.C., 1974. (13) U.S. Department of Agriculture, Agricultural Marketing Service. Annual report on tobacco statistics, 1976. (Statistical Bulletin No. 570). Washington, D.C., 1977. (14) U.S. Department of Agriculture, Agricultural Marketing Service. Annual report on tobacco statistics, 1981. (Statistical Bulletin No. 685). Washington, D.C., 1982. (15) U.S. Department of Agriculture, Agricultural Marketing Service. Tobacco stocks, as of January 1, 1983. Washington, D.C., 1983. (16) U.S. Department of Agriculture, Agricultural Marketing Service. Tobacco stocks, as of January 1, 1984. Washington, D.C., 1984. (17) U.S. Department of Agriculture, Agricultural Marketing Service. Tobacco stocks, as of January 1, 1986. Washington, D.C., 1986. 26 (18) National Clearinghouse for Smoking and Health. Use of tobacco: Prac- tices, attitudes, knowledge, and beliefs. U.S. Fail 1964 and Spring 1966, July 1967. (19) U.S. Department of Health, Education, and Welfare. Smoking and health-A report to the Surgeon General. Office on Smoking and Health (DHEW Publication No. PHS 79-50066). 1979. (20) National Clearinghouse to3 Smoking and Health. Adult use of tobacco, 1975. U.S. Department of Health, Education, and Welfare, Public Health Service, 1976. (21) National Center for Health Statistics. National Health Interview Survey, 1970 (unpublished). (22) Simmons Market Research Bureau, Inc. Study of Media and Markets, 1980-1985. (23) Simmons Market Research Bureau, Inc. Simmons National College Study, 1983 and 1985. (24) Office on Smoking and Health. Current Population Survey, 1985 (unpublished). (25) Rouse, B.A. National prevalence of smokeless tobacco use. Presented at the NIH Consensus Development Conference on the Health ImpIica- tions of Smokeless Tobacco Use, Bethesda, Maryland, January 13-15, 1986. (26) U.S. Department of Commerce, Bureau of the Census. (CP Series P25, No. 922), 1982. (27) Hartge, P., Hoover, R., and Kantor, A. Bladder cancer risk and pipes, cigars, and smokeless tobacco. Cancer 55: 901-906, 1985. (2%) Glover. E.D., O'Brien, K., and Holbert, D. Prevalence of smokeless tobacco use in Pitt County, North Carolina. Int. J. Addict. (in press). (29) Gritz, E.R., Ksir, C.. and McCarthy, W.J. Smokeless tobacco use in the United States: Past and future trends. Ann. Behav. Med. 7: 24-27,1985. (30) Glover, E.D., Johnson, R., Laflin, M., and Christen, A. Smokeless tobacco trends in the United States. World Smoking and Health (in press). (31) Marty, P.J., McDermott, R.J., Young, M., and Guyton, R. Prevalence and psychosocial correlates of dipping and chewing in a group of mraI high school students. Health Educ. (in press). (32) Marty, P.J., McDermott, R.J., and Wihiams, T. Patterns of smokeless tobacco use in a population of high school students. Am. J. Public Health 76: 190-192,1986. CW Newman, I.M.. and Duryea. E.J. Adolescent cigarette smoking and tobacco chewing in Nebraska. Nebr. Med. J. 66: 243-244, 1981. (34) Bonaguro, J.A., Pugy, M., and Bonaguro, E.W. Multivariate analysis of smokeless tobacco use by adolescents in grades four through twelve. Health Educ. (in press). (35) Lichtenstein, E., Severson, H.H., Friedman, L.S., and Ary, D.V. Chew- ing tobacco use by adolescents: Prevalence and relation to cigarette smoking. Addict. Behav. (in press). 27 (36) Severson, H., Lichtenstein, E., and Gallison. C. A pinch or a pouch in- stead of a puff? Implications of chewing tobacco for addictive pro- cesses. Bulletin of Psychologists in Addictive Behaviors 4: 85-92,1985. 137) Jones, R.B. Smokeless tobacco: A challenge for the 80's. Journal of the Wisconsin Dental Association 10: 71'7-721, 1985. (38) Chassin. L., Presson, C.C., and Sherman, S.J. Stepping backward in order to step forward: An acquisition oriented approach to primary prevention. J. Consult. Clin. PsychoL (in press). (39) Chassin, L., Presson. C.C., Sherman, S.J., McLaughlin, L., and Gioia, D. Psychological correlates of adolescent smokeless tobacco use. Addict. Behav. (in press). (40) Greer, R.O., and Poulson, T.C. Oral tissue alterations associated with the use of smokeless tobacco by teenagers, I. Clinical findings. Oral Surg. 56: 275-284, 1983. (41) Paulson, T.C., Lindenmuth, J.E., and Greer, R.O. A comparison of the use of smokeless tobacco in rural and urban teenagers. CA 34: 248-261, 1984. (42) Hunter, S.M., Croft, J.B., Burke, G.L., Parker, F.C., Webber, L.S., and Berenson, G.S. Longitudinal patterns of cigarette smoking and smoke less tobacco use in youth. Am. J. Public Health 76: 193-195,1986. (43) Guggenheirner, J., Zullo, T.G., Krupee, D.C., and Verbin, R.S. Chang- ing trends of tobacco use in a teenage population in western Pennsyl- vania. Am J. Public Health 76: 196-197, 1986. (44) Schaefer, S.D., Henderson, A.H., Glover, E.D., and Christen, A.G. Pat- terns of use and incidence of smokeless tobacco consumption in school- age children. Arch. Otolaryngol. (in press). (4.5) Young, M., and Williamson, D. Correlates of use and expected use of smokeless tobacco among kindergarten children. Psychol. Rep. 56: 63-66,1985. (46) Boyd, G.M.. et al. Use of smokeless tobacco among children and adolescents in the United States. Prev. Med. (in press). (47) Evans, R.I., Hansen, W.G., and Mittehnark. M.B. Increasing the valid- ity of self-reports of smoking behavior in children. J. Appl. Psychol. 62: 521-523. 1977. (48J Murray, D.M.. O'Connell, C.M., Schmiel. L.A., and Perry, C.P. The validity of smoking self-reports by adolescents: A reexamination of the logic and pipeline procedure. Addict. Behav. (in press). (49) Bauman, K.E., and Dent, C.W. Influence of an objective measure on self-reports of behavior. J. Appl. Psychol. 67: 623-638, 1982. (50) Office of the Inspector General. Youth use of smokeless tobacco: More than a pinch of trouble. U.S. Department of Health and Human Ser- vices, January 1986. 28 Chapter 2. CARCINOGENESIS ASSOCIATED WTTH SMOKELESS TOBACCO USE CONTENTS Introduction.. ................................... ..3 3 Epidemiologic Studies and Case Reports of Oral CancerinReIationtoSmokelessTbbaccoUse ............. .33 DataFromNorthAmericaandEurope .............. .33 DataFromAsia ................................ .42 summary......................................4 4 References ................................... ..4 4 Epidemiologic Studies of Other Cancers inRelationtoSmokeless'I%accoUse ................... .47 NasalCancer.................................... 8 EsophagealCancer............................... 8 LaryngealCancer................................ 0 StomachCancer ............................... ..5 1 Urinary'It-actCancer ............................ .52 Other Cancers .................................. .54 summary.. .................................. ..5 5 References ................................... ..5 6 Chemical Constituents, Including Carcinogens, of Smokeless Tobacco ............................... .58 Chemical Composition of Smokeless Tobacco .......... .58 Carcinogens in Smokeless `Ibbacco .................. .58 Summary......................................6 5 References ................................... ..6 7 Metabolism of Constituents of Smokeless Tobacco .... .70 Metabolism of NNK ........................ .71 Metabolism of NNN ........................ .73 Metabolism of NMOR ...................... .74 summary ................................ .75 References ............................... .75 Experimental Studies Involving Exposing Laboratory Animals to Smokeless `Ibbacco or Its Constituents ......... .79 Bioassays With Chewing Tobacco .................. .79 Bioassays With Snuff. ........................... .83 Bioassays With Constituents of Smokeless Tobacco .... .86 Mutagenicity Assays and Other Shoti'Ihrrn lb&s ...... .88 Summary.. .................................. ..8 9 References ................................... ..8 9 cOnclusions........................................92 ResearchNeeds.....................................93 31 This chapter presents the results of a systematic review of the world's medical literature describing experimental and human evidence perti- nent to the evaluation of smokeless tobacco as a potential cause of cancer. Five categories of research relevant to assessing the role of smokeless tobacco in cancer causation were defined: 1. Epidemiologic studies and case reports of oral cancer in relation to smokeless tobacco use. 2. Epidemiologic studies of other cancers in relation to smokeless toba~0 use. 3. Chemical constituents of smokeless tobacco. 4. Metabolism of constituents of smokeless tobacco. 5. Experimental studies involving exposing laboratory animals to smokeless tobacco or its constituents. consensus summari es of the literature in each of these categories were prepared and form the basis of this report. In addition, recommen- dations for future i-search to clarify suggestive findings or fi.lI gaps in knowledge are made. EPIDEMIOLOGIC STUDIES AND CASE REPORTS OF ORAL CANCER IN RELATION TO SMOKELESS TOBACCO USE Because smokeless tobacco products used in different regions of the world vary considerably in composition and usage patterns, this section will consider North American and European data separately from Asian data. Citations to the literature from India and other Asian coun- tries where quids containing tobacco and other ingredients are com- monly used orally focus on articles that attempt to distinguish tobacco from other ingredients in the quids as possible determinants of cancer risk. Data From North America and Europe Although about a dozen informative epidemiologic studies of smoke less tobacco use and oral cancer in North America or Europe have been reported, only a few were specifically designed to examine this relation. There are two major reasons for the relative paucity of studies. Apart from the recent increased prevalence in use of smokeless tobacco, the habit has not been widely practiced in America during this century, ex- cept in localized areas such as parts of the rural South (1,2). Further- more, cancer of the mouth is uncommon in the Western Hemisphere, 33 exacerbating the difficulty of conducting epidemiologic investigations, particularly cohort studies, into the relation between smokeless tobacco and oral cancer. The ageadjusted incidence rate for cancers of the buc- cal cavity and pharynx in the United States is approximately 11 cases per 100,000 population per year, with these tumors accounting for about 3 percent of all cancer deaths (3). Nevertheless, sufficient informa- tion is available to evaluate whether the use of smokeless tobacco increases the risk of oral cancer. case studies In their review of 566 oral cancer patients treated in two hospitals in Nashville, Rosenfeld and Callaway (4,5) noted that the proportion of women (61 percent) with buccal and gingival carcinoma was higher than the proportion of men (36 percent). Approximately 90 percent of women with buccal and gingival carcinoma used snuff for 30 to 60 years; in con- trast, 22 percent of women with cancers in other oral cavity subsites used snuff. Many of these women began practicing "snuff dipping," namely, the placement of tobacco snuff in the gingivobuccal sulcus, be tween the ages of 10 to 20 years. These reports are typical of numerous and sometimes large series of cases from the South, which reported that high percentages of patients with gingivobuccal cancers were snuff dip- pers or tobacco chewers (613). The articles describing these case series generally did not use comparison (control) groups, but the authors con- sistently commented on an apparently high prevalence of the use of snuff by the cancer patients. Clinicians also noted that the usual male predominance for epidermoid carcinomas of the oral cavity diminished or disappeared for the subgroup of gingivobuccal carcinomas occurr@ in geographic areas where there was relatively common use of snuff and chewing tobacco. Ahblom reported in the 1930's on a possible association between smokeless tobacco and cancer in Sweden (14). Among male patients with cancers of various sites seen at the Ftadiumhemmet (Stockholm), the use of snuff or chewing tobacco was reported in 70 percent with buc- cal, gingival, and "mandibuIar" cancers as compared to 26 to 37 percent with cancers in other oral subsites, the larynx, pharynx, and esophagus. Ax611 et al. reviewed medical records of male patients with squamous cell carcinoma in the oral cavity diagnosed between 1962 and 1971 and recorded in the Register of the Swedish Board of Health and Welfare (Is). The authors were only able to determine a history for the pattern of use of snuff in 25 percent of eligible patients but commented that two thirds of patients who were verified snuff users had oral cancers in regions where the snuff was generally placed. Reports of a single or a few cases, usually among male tobacco chewers, in the Northern United States and Canada also described buccal carcinomas that were often located precisely in the area where the tobacco was retained in the mouth (1619) 34 In the early 1940's, Friedell and Rosenthal associated the use of snuff or chewing tobacco with an exophytic, verrucous type of squamous car- cinoma of the oral cavity (16). Ackerman described in detail the morpho logic and clinical features of verrucous carcinoma of the oral cavity (20). Where the lesions originated in the buccal mucosa, a history of chronic use of chewing tobacco was elicited in 60 percent of the patients. The morphologic description was that of a well-differentiated, locally inva- sive, papillary squamous carcinoma, often in association with leuko plakia. In more than half of these patients, there was poor oral hygiene and carious and missing teeth. In summary, clinical and pathological reports published during the past four decades in the United States and elsewhere have commented on the use of smokeless tobacco by oral cancer patients and have described the entity known as snuffdipper's carcinoma (4,7,11), providing the basis for the hypothesis that the prolonged use of snuff or chewing tobacco is associated with an increased risk of low-grade, verrucal or squamous cell carcinoma of the buccal mucosa and gingivobuccal s&us. case control studies Most of the epidemiologic evidence comes from several case-control studies of oral cancer. The low prevalence of smokeless tobacco use in most North American populations contributes to a low statistical effi- ciency in most of these studies. Good information has been obtained, however, from studies that were either very large, conducted in an area of high prevalence of smokeless tobacco use, or analyzed according to site within the oral cavity (since the tissue affected by snuff use appears to be highly localized). One study, by Winn et al., with these characteris- tics consequently provides the most informative body of data on the carcinogenicity of smokeless tobacco in North America (21). The major concern for validity in the epidemiologic studies of smoke less tobacco and oral cancer is uncontrolled confounding. A smalI num- ber of subjects in crucial categories prevented efficient adjustment for confounding by stratification in many of these studies. Many of the studies were conducted before the advent of sophisticated epidemic logic analyses and make no attempt to control confounding. The two primary confounding factors of concern are alcohol consumption and smoking (22). Alcohol consumption is a strong risk factor for oral can- cer. It is not clear on a priori grounds, however, to what extent alcohol consumption would be correlated with smokeless tobacco use. The rela- tion between smoking, also a strong risk factor for oral cancer (2), and smokeless tobacco use may be complex. Users of smokeless tobacco may be more likely to have been smokers at some time. On the other hand, heavy users of smokeless tobacco typically cannot be heavy users of cigarettes, so that smoking is p resumably negatively correlated with smokeless tobacco use. Failure to control confounding by smoking would therefore lead to underestimates of the effect of smokeless tobacco. 35 TABLE l.-Smokeless Tobacco and Mouth Cancer, Case-Control Data From Moore et al. (23,24) Smokeless Tobacco Mouth Cancer Cases controls Users 26 12 Nonusers 14 26 Totals 40 38 Crude RR = 4.0 95%.Confidence Interval: 1.6-10 Chronologically, the first casecontrol study of smokeless tobacco was conducted by Moore et al. in Minnesota (23,24). Patients at the University of Minnesota `lhrnor Clinic with a diagnosis of cancer of the mouth were interviewed about tobacco use as part of a general inter- view procedure for clinic patients. Surgical outpatients who received the same interviews served as controls. Prom the data that were reported by these authors, one can calculate a crude relative risk estimate for mouth cancer among smokeless tobacco users of 4.0 with a 95percent confidence interval of 1.6-10 (table 1). An oddity was an ap- parent lack of effect for other forms of tobacco use. A partial explana- tion might be negative confounding between smokeless and smoked tobacco; indeed, 26 of the 40 cases of mouth cancer chewed tobacco. Still, the extent of disparity in crude effect estimates for smokeless tobacco (relative risk estimate 4.0) and smoked tobacco (all relative risk estimates < 1.0) is surprising. Wynder et al. reported on a case-control study of squamous cell cancers of the upper alimentary and respiratory tract that was con- ducted at Sweden's Radiumhemmet in 1952-55, including 33 tongue cancer patients, 14 lip cancer patients, 19 gingival cancer patients, and 8 patients with cancer of the buccal mucosa, among others (25). Con- trols were patients with cancers of the skin, head, and neck other than squamous cell carcinoma, stomach cancer, lymphoma, salivary-gland tumors, leukemia, sarcoma, cancers of the colon and rectum, and cancers of the female genital tract. A variety of risk factors was exam- ined, including the use of chewing tobacco. The authors state that the data suggested that an increased risk is associated with the duration of chewing tobacco for cancers of the gingiva and oral cavity but not for cancers of the tongue, lip, hypopharynx, esophagus, or larynx, but the data as presented do not permit an estimation of risk. In addition, data were not adjusted for other potential confounders, including cigarette smoking. Wynder and colleagues also reported in 1957 data from a similar hospital-based casecontrol study of mouth cancer conducted in New York (26). ?bbacco chewing was found to be more common among men with oral cavity cancers than among controls; but it was noted that almost all of these patients also drank alcoholic beverages and smoked and no further analyses were attempted. 36 TABLE 2.-Smokeless Tobacco and Mouth Cancer, CaseControl Data From Peacock et al. (27) Age Smokeless Tobacco User Nonuser Total 40-49 59-59 60-69 case controls case controls Case Controls 0 16 7 13 18 20 5 14 6 16 9 37 5 60 13 29 27 57 RR=0 RR = 1.4 RR = 3.7 RRMH = 2.0 95%.Confidence Interval: 1.0-4.2 Peacock et al. studied 56 cases of mouth cancer, including malignan- cies of the buccal mucosa, alveolar ridge, and floor of the mouth, and compared their tobacco histories with those of two control groups: 146 hospitalized controls with diagnoses other than cancer and 217 outpa- tients (27). Agespecific results using the hospitalized controls are sum- marized in table 2. The overall relative risk was estimated to be 2.0 (95percent confidence interval 1.0-4.2); the relative risk seemed to in- crease with age with an estimate of 3.7 for the 60 to 69 age group. The data were not reported in sufficient detail to control for confounding by smoking, which presumably led to underestimates of the relative risk. There was also insufficient detail reported to evaluate the relation be tween the risk of mouth cancer and the amount or duration of smokeless tobacco use. In Atlanta, patients with oral, pharynx, and larynx cancer were com- pared to three control groups having other mouth diseases, other can- cers, or no cancer (28. Among urban women, 40 percent of the cases used snuff compared to 3 percent or less of the controls (table 3). Among rural women, 75 percent dipped snuff compared to 20 percent or less among controls. Cigarette smoking was common in urban women and not specifically controlled for. Few rural female cases smoked cigarettes (7 percent) so confounding by smoking was minimal. The association between snuff dipping and oral, pharynx, and larynx cancer in women was generally evident in most age groups. Among the cases, the propor- tion of snuff dippers was highest among oral cancer patients: 53/72 were dippers compared to 2/18 pharynx and larynx cancer patients. Among men, insufficient information was provided to obtain precise epidemic logic estimates of the effect of chewing tobacco, although date from one of the bar charts presented indicate that urban cases were more likely to be users of smokeless tobacco than controls, that rural men with oral, pharynx, and larynx cancer or mouth disease were more likely to chew than controls, and that oral cancer patientu were more likely to chew 37 TABLE 3.-Estimated Relative Risks Associated With Snuff Use for Cancers of the Oral Cavity, Pharynx, and Larynx, Case.C!ontrol Data From Vogler et aL (2B), Females Only oral/ Other PlLSryIlXl Mouth Larynx Disease Other No ClUlCer Cancer Urban User Nonuser Crude Relative Risk Estimate RlUal User Nonuser Crude Relative Risk Estimate 15 1 5 4 23 56 165 373 60.8 1.7 2.8 1.0* 41 4 26 17 14 33 103 133 22.9 0.9 2.0 1.0* TABLE 4.-Smokeless Tobacco and Head and Neck Cancer by Anatomic Site, CaseControl Data From Vincent and Marchetta (29), Males Only Smokeless Tobacco Use User Nonuser Total Relative Risk Estimate 95%~Confidence Interval Control Larynx 5 2 95 21 100 23 1.8 0.3-9.8 oral All Head Phw Cavity and Neck 3 9 14 30 24 75 33 33 89 1.9 7.1 3.5 0.4-8.3 2.4-21 1.3-9.8 than the pharynx and larynx cancer cases. Among men, confounding by smoking could not be ruled out. Vincent and Marchetta reported the results of a case-control study of head and neck cancer according to anatomic site. Table 4 summarizes the findings for males (29). The oral cavity seems to be the anatomic site where the bulk of the effect is noted; only mild increases in risk were estimated for the larynx and pharynx, whereas users of smokeless tobacco were estimated to have a sevenfold greater risk for cancer of the oral cavity. These estimates are imprecise because of the small number of subjects and are uncontrolled for age and smoking. 38 TABLE 5.-Estimated Relative Risk for Cancer of the Head and Neck From Smokeless `lbbacco Use by Anatomic Site, Third National Cancer Survey (311, Males Only Relative Risk Estimate Anatomic Site Low Exposure High Exposure Gum-Mouth 5.6 3.9 Pharynx 0.6 - Lip-Tongue 0.3 1.1 LarYm 2.0 1.7 Martinez reported on a case-control study in Puerto Rico of risk fac- tors for cancers of the mouth, pharynx, and esophagus (XI). This population-based study included 400 cases of epidermoid carcinomas of those sites and 1,200 controls matched on age ( + 5 years) and sex to the cases. One control per case was drawn from the same hospital or clinic and two from the same community. There were 153 cases of mouth cancer (115 male and 38 female) and 68 cases of pharyngeal cancer (55 male and 13 female). The authors concluded that "Patients with cancer of the mouth did not often use chewing tobacco disproportionately. " However, calculation of the relative risks of mouth cancer that are asso ciated with chewing tobacco based on comparing the use of chewing tobacco only with no tobacco use suggests a strong effect for oral and pharyngeal cancer in males (data from table 13 in the paper). The esti- mated relative risks were 11.9 (95percent confidence interval 2.5-56.4) for oral cancer and 8.7 (95percent confidence interval 1.4-54.5) for pharyngeal cancer among chewers. These numbers do not include the experience of the many study subjects whose use of tobacco was mixed" (that is, those who used any combination of cigarette, cigar, and pipe smoking and chewing tobacco), and these calculations were based on unmatched data. Further evidence for the site specificity arose from a case-control analysis of multiple cancers using data from the Third National Cancer Survey (31). There were few female users of smokeless tobacco and scanty data by site within the head and neck region even for males; the findings do seem to indicate that the effect is greater for the site that is labeled gum-mouth as opposed to other head and neck sites (table 5). Browne et al. conducted interviews with 75 oral cancer patients, or (usually) their next of kin, and 150 living sex-, neighborhood-, and occupation-matched controls in the West Midlands area of the United Kingdom where oral cancer mortality rates were high and tobacco chewing was common among miners (32). Controls on average were born about 10 years earlier than the cases. The proportion of tobacco chewers was approximately the same among the 16 cases and 43 con- 39 trols who were miners, although data on this variable were missing for onefourth of the cases, and the authors apparently assumed that all cases with missing information were nonchewers. If the proportion of tobacco chewers among the cases with missing information was similar to those miners with known information, then the data would have shown a positive association between chewing tobacco and oral cancer. All of the miners with oral cancer who chewed tobacco also smoked pipes, further complicating interpretation of this study. Additional evidence that a carcinogenic effect of smokeless tobacco may be greatest at the anatomic site of exposure came from Westbrook et al. who compared the medical records of 55 female patients with cancers of the alveolar ridge or buccal mucosa who were treated at the University of Arkansas with those of 55 randomly selected female hospital controls (33). Fifty of the cases, but only one control, were snuff dippers, with the tumors among the cases typically appearing at the site where the snuff was usually placed. No reliable estimates of risk can be derived from this study because of the strong possibility that them was not comparable elicitation of exposum information for cases and controls. `Iwo large case-control studies were not reported in a way that enables a meaningful quantitative assessment of the effect of smokeless to. bacco in chewers and dippers compared to tobacco abstainers (34,35). The first study found that 10 percent, and the second 9 percent, of male oral cancer cases had ever chewed tobacco, while the corresponding fig- ure for controls was 9 percent. These studies, like many of the others cited here, were not undertaken specifically to evaluate the carcino genicity of smokeless tobacco. Although the data seem to indicate a weak relation, if any, between smokeless tobacco and cancer of the oral cavity, the findings are uncontrolled for age, race, geography, and smoking. The recent casecontrol study of Winn et al. is by far the most infor- mative study on the carcinogenicity of smokeless tobacco (21). The case series comprised 255 women with oral and pharyngeal cancer who were living in 67 counties in a high-risk (for oral cancer) region of North Carolina. ?tyo female controls were obtained for all but a few cases and were individually matched for age, race, source of ascertainment (hospital or death certificate), and county of residence. There was a four- fold increased risk of oral-pharyngeal cancer among nonsmoking white women who dipped snuff. The association could not be explained by smoking or alcoholic beverage consumption (21), denture wearing or poor dentition (36), diet (37), or mouthwash use (38). The data provided evidence for a strong relation between the duration of snuff use and risk for cancer, as well as a striking localization of the carcinogen&y to the gum and buccal mucosa (table 6). For long-term chronic users of snuff, there was nearly a fiftyfold increase in risk for cancers of the gum and buccal mucosa. Indeed, almost all of the patients with cheek and gum cancers had dipped snuff. 40 TABLE 6.--Estimated Relative R&k of tipharyngeal Cancer According to Duration of Snuff Use and Anatomic Site, Winn et al. (21) Anatomic Duration of Site Snuff use (yrb Relative Risk Estimate 95%. Confidence IUtelVal Gum and Buccal Mucosa Other Mouth and Pharynx 0 1 - 24 25 - 49 1 50 0 1 - 24 25 - 49 2 50 1.0 - 13.8 1.9 - 98 12.6 2.7 - 53 48.0 9.1 - 250 1.0 - 1.7 0.4 - 7.2 3.8 1.5 - 9.6 1.3 0.5 - 3.2 Although some of the exposure information came from interviews with next of kin, when the analysis was restricted to interviews with study subjects, the association between snuff and oral cancer was even stronger @9). Matched conditional logistic analysis yielded similar results (35). Based on calculations of attributable risk, the authors estimated that 87 percent of these cancers were due to the patients' snuff-dipping habits. The authors also provided data that demonstrated the negative confounding by tobacco smoking in the population, raising the possibility of a serious validity problem with the other studies that did not control for smoking. If the negative correla- tion between the use of smokeless and smoked tobacco holds in other populations, estimates of the carcinogenic effect of smokeless tobacco in studies without the control of smoking may be underestimates. The quantitative information that was provided by the Winn et al. study led its authors to conclude that the long-standing use of smokeless tobacco by Southern women was the principal cause of the elevated mortality from oral cancer among women in the Southern United States. Cohort Studies Few cohort studies of smokeless tobacco have been undertaken because of the rarity of both the exposure (smokeless tobacco use) and the outcome (oral cancer) of most interest. Bjelke and Schuman (40) reported on cancer mortality in cohorts of 12,945 Norwegian men and 16,930 American men and found increases in the risk of death for can- cers of the buccal cavity, pharynx, and esophagus (relative risk estimates ranged from 2.6 to 3.1(41); no further detail was given). They noted a negative association between smoking and chewing tobacco, confirming the pattern that was observed from the case-control research. In a 16.year followup of U.S. veterans, Winn et al. reported no deaths from oral or pharyngeal cancer among 951 smokeless tobacco 41 users who did not use other forms of tobacco (about 0.5 deaths were ex- pected) but a significant increase in both oral and pharyngeal cancers among smokeless tobacco users who were light smokers (42) These data, as well as those from Bjelke and Schuman (40). were reported only as abstracts in scientific journals or proceedings, with little or no detail as to the methods used, hindering interpretation of the results. Smith and colleagues followed a group of about 1,500 patients with changes in the oral mucosa to evaluate the effects of smokeless tobacco use (43,44). No oral cavity cancers were found in about 16,000 person- years of followup. Based on the results of other studies, two or three should have been detected over the study period. Smith gave little docu- mentation of the methods that were employed for followup; however, 12 percent of the original group (201 subjects) were lost without any data on outcome, and there was apparently no effort to trace them. It seems likely that persons who died and persons who developed cancer, includ- ing some with tumors of the oral cavity, may have been lost to followup. In fact, no deaths among cohort members were reported, whereas perhaps as many as 100 or more would have been expected among such a cohort of middleaged adults, making Smith's data uninterpretable. Data From Asia The highest rate-e of oral cancer among the more than 100 that are listed from population-based registries around the world that report stan- dadized cancer incidence statistics are found in India (45). In many areas of Asia, hospital statistics suggest that oral cancer is extremely common and often accounts for 25 or more percent of all cancers (4649), propor- tions that are far greater than in most areas of the United States where oral cancers typically comprise only 3 percent of all malignancies (3). It has long been thought that the chewing of quids that contain tobacco and other substances is the cause of the increased risk of oral cancer in these areas (50). The smokeless tobacco products that are commonly used include to bacco with betel leaf, areca nut, and lime mixtures (often referred to as "pan"); Khaini (powdered tobacco and slaked lime paste); mishri (pow- dered partially burnt black tobacco); nass (tobacco, ash, and cotton or sesame oil; lime is used in Iran and certain Soviet Republics); and various preparations that vary locally throughout the Southeast Asia region. The inclusion of lime, areca nut, and other ingredients in many of the smokeless tobacco-containing quids hinders the evaluation of the con- tribution of tobacco per se to the increased risk of oral tumors. From five investigations, however, relative risks of oral cancer among chewers of betel quids with versus without tobacco can be calculated. Data from these case-control studies, which were conducted in Cal- cutta, Madras, Karachi, Bombay, and several parts of India and Sri Lanka (4',51+%), reveal considerably higher risks of oral cancer for the 42 TABLE 7.-Relative Risk of Oral Canax From Betel Quid With and Without `Ibbacco (With 95Percent Confidence Limit) Study Location (References) Betel Betel Quid . No With W%zut Chewing Tobacco Tobacco Habit Remarks Calcutta, India (5054) Madras, India (5154) Karachi. Pakistan (5&w Bombay, India (53) India and SI-iLanka (47) Cases 138 46 Controls 61 70 Relative risk 4.3 1.2 estimates (3.0-6.1) 10.8-1.9) Cases 219 33 Controls 35 144 Relative risk 25 0.91 estimates 115-41) 10.4-1.6) 135 256 25 99 Smokers not included in these data. Only buccal mucosa can- cers considered. Smokers not included in these data. Only buccaI mucosa and tongue cancer cases included. Numbers reconstructed from percentages and totals. Cases Controls ReIative risk estimates Cases Controls Relative risk wtimates cases Controls Relative risk estimates 339 474 (11~~7, 238 513 120 63 (7.Ot532, 40 216 3.6 (2.4-5.2) 44 152 3.0 (2.1-4.3) 3 8 2.9 (0.6-14) 88 Smokers not included 1.690 in these data. 129 1.340 6 47 Separate analyses indicate that ele vated risks of oral cancer associated with tobacco chew- ing are found among nonsmokers as well as smokers. Smokers not included in these data. Only buccaI mucosa cancer considered. use of tobacco-containing compared to nontobacco-containing quids (table 7). The findings thus suggest that the addition of tobacco con- tributes substantially to the elevated cancer risk among chewers, although other differences between those who use versus those who do not use tobacco-containing quids could influence the differences. Smok- ing, however, is not such a difference, since most of the investigations referred to in table 7 demonstrated high relative risks of oral cancer (with excesses among tobacco chewers often exceeding tenfold com- pared to nonquid users) among chewers who did not smoke, ruling out confounding by cigarette smoking. The studies also generally found that the large majority of oral cancer patients had been tobacco chewers and suggest that the habit of quid chewing accounts for most of the oral cancers in the diverse populations studied (5.5%). 43 Summary Numerous case reports, especially in the South, have described oral cancers among smokeless tobacco users. The tumors often arose at ana- tomic locations where the tobacco was routinely placed. The number of epidemiologic investigations evaluating the relation between smokeless tobacco and oral cancer is not large, and several studies have method- ologic limitations. The pattern of increased oral cancer risk among smokeless tobacco users, however, is generally consistent across studies, with evidence of an increasing risk with increasing duration of exposure, and with excess risks tending to be greatest for those ana- tomic sites where tobacco exposures are greatest. The best designed study was drawn from a female population in the Southern United States where exposure rates are high and potentially confounding vari- ables could be taken into account. This study showed that chronic snuff users were at substantially increased risk of oral cancers and that nearly all tumors of the cheek and gum were due to snuff use. Evidence from parts of Asia, where the prevalence of smokeless tobacco use is high and oral cancer is the most common tumor, indicates a strong asso ciation between the chewing of quids and oral cancer. Users of quids that contain tobacco have much higher oral cancer rates than users of quids that do not, and the association is not confounded by cigarette smoking, raising the possib&ty that tobacco per se contributes to the elevated oral cancer risk in this part of the world. In summary, users of smokeless tobacco face a strongly increased risk of oral cancer, particu- larly for the tissues that come in contact with the tobacco. References (1) Office on Smoking and Health. Smoking and Health: A Report of the Surgeon General. U.S. Department of Health, Education, and Welfare. Washington, D.C., U.S. Government Printing Office, 1979. (2) Blot, W.J., and Fraumeni, J.F. Geographic patterns of oral cancer in the United States: Etiologic implications. J. Chron. Dis. 30: 745-757.1977. (3) Young, J.L., Percy, C.L., and Asire, A.J. Surveillance, epidemiology, and end results: Incidence and mortality data, 1973-77. NC1 Monogr. 57, 1981. (4) Rosenfeld, L., and Callaway, J. Snuff dipper's cancer. Am. J. Surg. l&C 840-844, 1963. (5) Rosenfeld, L., and Callaway, J. Squamous cell carcinoma of the oral cavity. South. Med. J. 56: 1394-1399, 1963. (6) Landy, J.L., and White, H.J. Buccogingival carcinoma of snuff dippers. Arm Surg.27: 442-447, 1961. (7) Wilkins, S.A., and Vogler, W.R. Cancer of the gingiva. Surg. Gynecol. Obstet. 10.5: 145-152, 1957. (8) Brown, R.L., Sun, J.M., Scarborough, J.E., Wilkins, S.A., and Smith, R.R. Snuff dippers intraoral cancer: Clinical characteristics and response to therapy. Cancer 18: `2-13, 1965. 44 (9) Coleman, C.C. Surgical treatment of extensive cancers of the mouth and pharynx. Ann. Surg. 161: 634-644, 1965. (10) Fonts, E.A., Greenlaw, R.H., Rush, B.F., and Rovin, S. Verrucous squamous cell carcinoma of the oral cavity. Cancer 23: 152-160, 1969. (11) HartseIIe, M.L. Oral carcinoma as related to the use of tobacco. Ala. J. Med. Sci. 14: 188-194, 1977. (12) McGuirt, W.F. Snuff dipper's carcinoma. Arch. OtoIaryngol. 109: 757-760, 1983. (13) McGuirt, W.F. Head and neck cancer in women-a changing profile. Laryngoscope 93: 106-107, 1983. (14) Ahblom, H.E. Predisposing factors for epitheliomas of the oral cavity, larynx, pharynx, and esophagus. Acta Radial. 18: 163-185, 1937 (in Swedish). (15) Ax&II, T., et al. Snuff dipping and oral cancer-a retrospective study. Tandlakartidningen 75: 2224-2226, 1978. (16) FriedeII, H.L., and Rosenthal, L.M. The etiologic role of chewing to bacco in cancer of the mouth. JAMA 116: 2130-2135, 1941. (17/ Moertel, C.G., and Foss, E.L. Multicentric carcinomas of the oral cav- ity. Surg. Gynecol. Obstet. IOf? 652654, 1958. (18) Sorger, K., and Myrden, J.A. Verrucous carcinoma of the buccaI mucosa in tobaccochewers. Can. Med. Assoc. J. 83: 1413-1417, 1960. (19) Stecker, R.H., Devine, K.D., and Harrison, E.G., Jr. Verrucose "snuff dippers" carcinoma of the oral cavity. JAMA 189: 838-840, 1964. (20) Ackerman L.V. Verrucous carcinoma of the oral cavity. Surgery 23: 670-678,1948. (21) Winn, D.M., Blot, W.J.. Shy, C.M., et aI. Snuff dipping and oral cancer among women in the Southern United States. N. Engl. J. Med. 304: 745-749, 1981. (22) Rothman, K., and Keller, E. The effect of joint exposure to alcohol and tobacco on risk of cancer of the mouth and pharynx. J. Chron. Dis. 25: 711-716, 1972. (23) Moore, G.E., Bissmger, L.L., and ProehI, E.C. Tobacco and intraoral cancer. Surg. Forum 3: 685688, 1952. (ZX#) Moore, G.E., Bissinger, L.L., and Pro&I, EC. Intraoral cancer and the use of chewing tobacco. J. Am. Geriatr. Sot. 1: 497-506,1953. (25) Wynder, E.L., H&berg, S., Jacobsen, F., and Bruss, I.J. Environmental factors in cancer of the upper alimentary tract. Cancer 10: 470-487,1957. (26) Wynder, E.L., Bruss, I.J., and Feldman R.M. A study of the etiologicaI factors in cancer of the mouth. Cancer 10: 1300-1323,1957. (27) Peacock, E.E., Greenberg, B.G., and Brawley. B.W. The effect of snuff and tobacco on the production of oral carcinoma. Am. Surg. 151: 542-550, 1960. (28 VogIer, W.R., Lloyd. J.W., and Mihnore, B.K. A retrospective study of etiologicaI factors in cancer of the mouth, pharynx, and Iarynx. Science 15: 246-258.1962. 45 1291 Vincent, R.G., and Marchetta F. The rehrtionship of the use of tobacco and alcohol to cancer of the oral cavity, pharynx, or Iarynx. Am. J. h-g. 106: 501-505, 1963. (30) Martinez, I. Factors associated with cancer of the esophagus, mouth, and pharynx in Puerto Rico. J. Natl. Cancer Inst. 42: 1069-1094,1969. (311 WiIIiams, R.R., and Harm, J.W. Association of cancer sites with to bacco and alcohol consumption and socioeconomic status of patients: Interview study from the Third National Cancer Survey. J. Natl. Cancer Inst. 58: 525-547, 1977. (32) Browne, R.M., Camsey, M.C., Waterhouse, J.A.H., and Manning, G.L. EtiologicaI factors in oral squamous ceII carcinoma. Community Dent. Oral Epidemiol. 5: 301-306, 1977. (33 Westbrook, K.C., Sven, J.Y., Hawkins, J.M., andMcKinney. D.C. Snuff dipper's carcinoma: Fact or fiction? In: H.E. Nieburg (ed.). Prevention and Detection of Cancer. New York, Marcel Dekker, 1980. pp. 1367-1371. (34) Wynder, E.L., and Stelhnan, S.D. Comparative epidemiology of tobacco-related cancers. Cancer Res. 37: 46084622, 1977. (3.5) Wynder, E.L., Kabat, G., Rosenberg, G., and Levenstein, M. Oral can- cer and mouthwash use. JNCI 70: 255-260, 1983. (3s Winn, D.M., Blot, W.J., and Fraumeni, J.F. Snuff dipping and oral cancer. N. Engl. J. Med. 305: 230-231,198l. (37) Winn, D.M., Ziegler, R.G., Pickle, L.W., GridIey, G., Blot, W.J., and Hoover, R.N. Diet in the etiology of oral and pharyngeaI cancers among women from the Southern United States. Cancer Res. 44: 1216-1222, 1984. 138) Blot, W.J., Winn, D.M., and Fraumeni, J.F. Oral cancer and mouthwash. JNCI 70: 251-253, 1983. (39) Winn, D.M. Smokeless tobacco and oral-pharynx cancer: The role of cofactors. Banbury Report (in press). (40) BjeIke, E., and Schuman. L.M. Chewing tobacco and use of snuff: ReIa- tionships to cancer of the pancreas and other sites in two prospective studies. Proceedings of the 13th International Congress on Cancer, 1982. p. 207. (41) International Agency for Research on Cancer. Tobacco habits other than smoking: Betelquid and areca-nut chewing and some related nitrosamines. IARC Monogr. EvaI. Carcinog. Risk Chem. Hum. 37: 103-104, 1985. (4.2) Winn, D.M., Wahath, J., Blot, W., and Rogot, E. Chewing tobacco and snuff in relation to cause of death in a large prospective cohort (Abstract). Am. J. Epidemiol. 116: 567, 1982. 1431 Smith, J.F., Mincer, H.A., Hopkins, K.P., and Bell, J. Snuff-dippers lesion: A cytologicaI and pathological study in a large population. Arch GtohuyngoI 92: 450-456, 1970. 1441 Smith, J.F. Snuff dippers lesion, a ten-year follow-up. Arch. OtoIaryn- gol.l01:2767-277, 1975. 46 (45) Waterhouse, J., Muir, C., Shanmugamnam, K., and Powell, J. Cancer incidence in five continents, Vol. IV. Lyon, France, International Agency for Research on Cancer, 1982. (46) Pindborg, J.J. Epidemiologic studies of oral cancer. Int. Dent. J. 27: 172-178, 1977. (47) Wahi, P.N. The epidemiology of oral and oropharyngeal cancer. Bull. WHO 38: 495-521.1968. (48) Hirayama, T. An epidemiologic study of oral and pharyngeal cancer in Central and Southeast Asia. BulL WHO 34: 41-69, 1966. (49) Paymaster, J.C. Cancer and its distribution in India. Cancer 17: 1026, 1964. (So) Orr, I.M. Oral cancer in betel nut chewers in Travancore. Lancet 2: 575-580, 1933. (51) Chandra, A. Different habits and this relation with cheek cancer. Bull. Cancer Hosp. Natl. Cancer Res. Center 1: 33, 1962. (52) Shanta, V., and Krishnamurthi, S. A study of aetiological factors in oral squamous cell carcinoma. Br. J. Cancer 13: 381, 1959. (537 Jafary, N.A., and Zaidi, S.H. Carcinoma of the oral cavity in Karachi, Pakistan: An appraisal. Trap. Doct. 6: 63, 1976. (54) Jus~awdh, D.J.. and Deshpande, V.A. Evaluation of cancer risk in tobacco chewers and smokers: An epidemiologic assessment. Cancer 28: 244-252, 1971. (55) Gupta, P.C., Pindborg, J.J., and Mehta, F.S. Comparison of carcine genicity of betel quid with and without tobacco. An epidemiological review. EcoL Dis. 1: 213-219, 1982. (56) Jayant. K.. Balakrishnan, V.. S anghvi, L.D., and Jussawalla, D.J. Quantification of the role of smoking and chewing tobacco in oral, pharyngeal, and esophageal cancers. Br. J. Cancer 35: 232-235, 1977. EPIDEMIOLOGIC STUDIES OF OTHER CANCERS IN RELATION TO SMOKELESS TOBACCO USE The epidemiologic studies reported in the preceding section that show an association between the use of smokeless tobacco and oral cancers, particularly malignancies of the cheek and gum, indicate that the topical exposure of tissues to tobacco can cause cancers at the site of the exposure. In the United States, the tissues in direct prolonged con- tact with the tobacco are generally those of the oral cavity. Smokeless tobacco may occasionally come in contact with other tissues. One case has been reported of squamous cell carcinoma that developed in the ear of an individual in Minnesota who habitually placed snuff in his ear for 42 years at the site where the neoplasm developed (II. Although but a single report, this highly unusual observation raises the possibility of a carcinogenic potential of smokeless tobacco at other anatomic sites when exposure is direct and prolonged. 47 Nasal Cancer In some areas of the world snuff is inhaled, so that tissues of the nasal cavity come in contact with the tobacco powder. The earliest report that links any form of tobacco to cancer was published over two centuries ago when what were probably nasal cancers were described in several patients in England who were heavy inhalers of snuff (2). There have been no systematic evaluations of snuff inhalation and nasal cancer in the United States, United Kingdom, or other European countries, most likely because both the sniffing habit and nasal cancer are uncommon, Sniffing snuff has been reported, however, to be a frequent habit among Bantu men, whose rates of nasal cancer have been reported to be high (31. In case-control studies of nasal sinus cancer reported in 1955,80 per- cent of patients with tumors of the maxillary antrum were prolonged and heavy snuff users, in contrast to about onethird of Bantu men with other cancers (4,s). The snuff used by the Bantu is thought to contain aloe plant ash, trace elements such as nickel and chromium, and other ingredients in addition to tobacco (6). Snuff use (presumably by inhala- tion) was reported not to account for the high rates of nasal adenocarci- noma among furniture makers in studies in England and Denmark, but evaluations of snuff itself as a risk factor were not undertaken (7,s). One casecontrol study of cancers of the nasal cavity and paranasal sinuses in the United States addressed the issue of smokeless tobacco (91. A total of 193 cases were identified in four hospitals in Virginia and North Carolina over a lo-year period. No association between sinonasal cancers and chewing tobacco was found (relative risk 0.7,95-percent con- fidence interval 0.4-1.5). However, a relative risk of 1.5 was observed for users of snuff (g&percent confidence interval 082.8). Risk was increased in snuff users for both adenocarcinomas (relative risk 3.1) and squamous cell carcinomas (relative risk 1.9) but not for other histologic types (relative risk 0.6) and was found for both sexes. The implications of the findings are not clear since the snuff used by the cases and controls was oral snuff not coming in contact with nasal tissues. Animal experiments, however, suggest that tumors distant to the site of exposme may l-ealllt from exposure to constituents of snuff (see the section on animal studies). An apparent excess of posterior nasal space tumors was reported among certain tribes in Kenya, and 6 or 12 cases interviewed were found to be chronic "liquid snuff" users (10). Multiple subsites of the respiratory tract were considered however, increasing the likelihood of a chance association No increased risk of nasopharyngeal cancer associ- ated with snuff use was noted in a casecontrol study in Singapore (11). Esophageal Cancer Other tissues that come in contact with constituents of smokeless tobacco in more dilute concentrations include the linings of the esopha- gus, larynx (supraglotic portion), and stomach. The results of studies of 48 7I!ABLE l.-Relative Risks of Ektphageal Cancer in Pemom Expoeed to Chewing `lbbacco and Snuff: Summa@ of Four CaseControl Studies controle Fht zz f3e.x % IReldw AUthtK No. Erpud No. Exjmd Risk* Wynder Chewing Any M 150 20 150 10 2.3 (4 < 1oyrs. 14 4 3.9 2 iop. 6 6 1.2 Williams chewing Level1 M 38 5.2 1,788 5.4 0.9 W or Snuff Level2 0 0 - WYgr cl=+% Any M 183 10.9 2,560 9.0 1.2 snuff Any M 4.4 2.7 1.7 Martinez t%ewingt Any M 120 2.5 360 3.6 1.2 (14) F 59 `11.9 177 7.3 2.7 *caldatedfrompubliabedrepultifuutpruvidedby~uthur. tRgtectedtonouamdtea. cancers of these three sites in relation to smokeless tobacco are inam- elusive. The studies are gtmemlly of limited power to detect small in- cxeases in risk, and many did not control for relevant, potentially con- founding variables. However, some studies of these three cancera do show au increaseinriskinrelationtotheuseofsmokeless~~.As shown in table 1, elevated relative risks of esophageal cancer up to twofold or higher were found in two hospital-based casecontrol studies in the United States involving 150 and 183 cancer patients (1213) and one in Puerto Rico (described in the previous section) with 179 casea (I#. One of the studies by Wynder and colleagues, however, found no evidence of an incmase in risk with duration of exposure, and all chewers were also smokers (12). The effect of smoking was not adjusted for in the other study (13). Another casecontrol study involving 120 black male cases of esophageal cancer was conducted in Washington, D.C. (15). Few of the cases or controls had used either chewing tobacco or snuff, suggesting that it did not contribute to the high rates of esophagealcancerobservedin~eareaFinally,datafromaprospective (cohort) study of U.S. veterans were analyzed to determine whether mortalityratesof~cdiseaseswereincreasedinusersofsmokeless tobacco (16,l. In the absence of smoking, the standardize mortality ratio for esophageal cancer was found to be 228, but this value was based on only one death In a cohort study of 12,945 Norwegian and 16,930 American men followed over 10 years, the risk of esophageal cancer was reported to be significantly increased among men who used 49 chewing tobacco or snuff, after controlling for age, residence, and smok- ing habits (17,18). Unfortunately, the results of both cohort studies have been published only as abstracts, so additional details are not available. Some evidence that the chewing of quids may increase the risk of esophageal cancer arises from studies in Southeast Asia. In a series of 237 cases of esophageal cancers in Sri Lanka, interview information from 111 revealed that 90 (81 percent) habitually used betel containing tobacco leaf (19). This percentage was considerably higher than the fre quency of betel chewing in the general population (30 percent). Betel chewing was more common among women. Esophageal cancer also was more common among women, an unusual observation since this cancer occurs more frequently among men in almost all areas of the world that report standardized cancer statistics @Q). Since few women were reported to smoke or use alcohol, the possibility of an etiologic role of chewing is increased. However, the potential effects of tobacco as op posed to other ingredients in the quids cannot be distinguished. In a case-control investigation in Bombay involving interviews with 305 esophageal cancer patients and nearly 2,000 population controls of age, sex, and religions similar to all head and neck cancer cases, a 2.5-fold in- creased risk of esophageal malignancy was observed (p < .Ol) among nonsmokers who chewed pan, a mixture usually consisting of tobacco, betel, lime, and other ingredients (21). The excess was higher, however, among those chewing quids without tobacco (relative risk 3.5) than with tobacco (relative risk 2.1). A more recent analysis (22) in Bombay based on 649 patients with esophageal cancer and 649 controls yielded similar qualitative findings, but the excess among users of pan without tobacco (relative risk 12.1) was accentuated compared to users of tobacco containing chews (relative risk 2.81. On the other hand, in an earlier case control investigation in southern India of several upper digestive tract tumors, including 93 esophageal cancers, increases in esophageal cancer risk were much greater among men who used betel with tobacco (c&u- lated relative risk 11) than without tobacco (calculated relative risk 2) (23,). The chewing of nass was not associated with esophageal cancer risk in a case-control study conducted in an area of Iran with among the world's highest rates for this cancer (24). Of 638 identified cases of esophageal cancer, interviews were completed with 344 and with 2 neighborhood controls matched to each case. The relative risk associ- ated with ever using nass was 0.9, with an upper limit of the 95-percent confidence interval of 1.5, suggesting that any major effect of nass on the origins of this cancer could be excluded. Laryngeal Cancer In a case-control analysis of the interview data from the Third Na- tional Cancer Survey (TNCS), Williams and Horm compared the prior use of smokeless tobacco products (in the aggregate) in persons with a 50