Document 91eqQ075MBoJ5pDGoRnOm9YL7

[CANCER RESEARCH 49, 5459-5468, October 1, 1989) A Case-Control Study of Occupational and Dietary Factors in Colorectal Cancer in Young Men by Subsite*1 Ruth K. Peters,2 David H. Garabrant, Mimi C. Yu, and Thomas M. Mack Department ofPreventive Medicine, University ofSouthern California School ofMedicine, Los Angeles, California 90033 [R. K. P., M. C. Y., T. M. M.], and School of Public Health, University ofMichigan, Ann Arbor, Michigan 48109 [D. H. G.J ABSTRACT have evaluated each risk separately for three subdivisions of the colon. With the hope that exposures responsible for colorectal cancer might be especially obvious among those in whom it develops early, 147 men with colorectal carcinomas first diagnosed between the ages of 25 and 44 years were compared to neighborhood controls. Physical activity on the job was protective for tumors located in the transverse and descending portions of the colon. Rectal cancer and to a lesser extent sigmoid cancers were associated with jobs in which dusts or fumes were inhaled, especially if those jobs were held for long periods in young adulthood. While risk for rectal cancer did not seem to be limited to any particular type of dust or fume, the excess risk was strongest for wood and metal dusts. Con sumption of fruits and vegetables and a preference for whole grain breads were protective for colon but not rectal cancers, while consumption of deep fried foods and barbecued/smoked meats increased risk at specific subsites. Beef intake, alcohol consumption, and cigarette smoking ap peared to play little or no role at any subsite. MATERIALS AND METHODS The cases were consecutively diagnosed, histologically confirmed incident cases ofadenocarcinoma of the colon or rectum which occurred in white male residents of Los Angeles County who were under 45 years of age at diagnosis. Patients were identified through the Los Angeles County Cancer Surveillance Program (18), a population-based tumor registry covering the entire county for all cases of cancer that are microscopically verified or mentioned on a death certificate. Efforts were made to recruit all cases diagnosed between July 1, 1974, and February 28, 1982. When the case died before an interview could be arranged, a proxy interview was conducted with the closest relative familiar with the habits and occupational history of the case. A control was individually matched to each case on race, sex, date INTRODUCTION of birth (within 5 years), and neighborhood of residence. He was identified by an algorithm that uses the house of the index case as a reference point and proceeds in a systematic and invariable sequence Like most malignant epithelial tumors, carcinomas of the colon and rectum increase in frequency with increasing age and rarely occur before the fifth decade of life. In Los Angeles County, fewer than 5% of these tumors are diagnosed in persons under 45 years of age. With the hope that environmental or life-style exposures responsible for these diseases might be especially striking among those in whom they develop early in life, we conducted a case-control study in young men. Environmental factors play a major role in the etiology of colorectal cancers since international differences in incidence until up to 80 residential units have been canvassed. Efforts were made to choose the first eligible resident in this sequence as the control. If the first eligible subject refused to participate, the second eligible subject in the sequence was asked. At the close of case recruitment, 232 eligible patients with colorectal cancer had been identified. The attending physician refused permission for 22 patients, 30 cases could not be located in the area, and 4 cases could not speak a language for which a translator was available. Of the remaining 176 patients, case (or surrogate) interviews were conducted for 151; the remaining 25 patients (or surrogates) did not wish to participate. Finally, an eligible neighborhood control could not be are at least 10-fold in magnitude and migrants acquire the levels of their new neighbors within two to three decades after migra tion (1). The most commonly cited environmental hypotheses are dietary, but case-control studies have produced inconclusive or conflicting results regarding most foods and nutritional factors (2). For many years occupational studies have tentatively linked colorectal cancer to asbestos (3, 4) and to work with metal and textiles (5). More recent investigators have observed excess cancer of the large bowel among woodworkers (6-8) and men with sedentary jobs (9-12). Colorectal cancer in adoles cents has also been linked to agricultural herbicide and pesticide exposure in childhood (13,14). Colon and rectal cancers are often combined in etiological studies, but the social class gradients for the two sites tend to be in opposite directions (5), and many investigations in which they have been examined separately resulted in the conclusion that different etiological factors prevail at the two sites and even at the various subsites within the colon (15-17). We have examined colon and rectum cancer separately and together and located for 4 interviewed cases, leaving a total of 147 case-control pairs for study. A comparison of these 147 cases with other colorectal cases of the same age collected by the Los Angeles County Cancer Surveil lance Program showed them to be representative of all eligible patients; i.e., there were no statistically significant differences in their distribu tions by marital status, religion, birthplace, social class, or subsite. The 147 controls were found after screening 2146 residential units, or an average of 14.6 units/control. Most of the controls were the first (n - 112) or second (n = 27) eligible neighbors. No match resided in 80.7% of the residential units canvassed, eligible but unwilling individ uals lived in 2.4%, no census could be completed for 10.1%, and the 147 interviewed controls resided in the remaining 6.8%. All interviews were conducted between April 1, 1975, and January 31,1984, by the same interviewer and most took place in the homes of the respondents (95% of cases and 90% of controls). The remaining interviews were conducted by telephone at the request of the respond ents. The median time between diagnosis and interview of the cases was 7 months; the median time between the interviews of case and matching control was 4 months. All interviews used the same structured questionnaire, designed to elicit a lifetime occupational history, voca tional and avocational exposures to specific substances and industrial Received 3/10/89; revised 6/19/89; accepted 7/6/89. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by USPHS Grants R01CA19476, K04CA00884, K07CA01291, and R01CA36501 from the Division of Extramural Activities, National Cancer Institute. 2 To whom requests for reprints should be addressed, at USC School of Medicine, 1420 San Pablo St., PMB-B305, Los Angeles, CA 90033. processes, use of tobacco and alcohol, usual consumption of foods grouped into a few broad categories, prior medical conditions, and family history of selected diseases. All events occurring after the cancer diagnosis were ignored for the case and his matching control. Proxy interviews were conducted with the spouse for 18 cases, and a parent, a sibling, and a child responded for 4, 2, and 1 cases, respec tively. An early attempt to obtain proxy interviews from the analogous member of the control family proved to be impractical; therefore only 5459 BOWEL CANCER IN YOUNG MEN BY SUBSITE 8 proxy interviews were conducted among the matched controls of the 25 proxy case interviews. Analyses were performed both with and without the pairs and no important differences were found; all tables subsets of cases by comparing them with all (n = 147) controls, using cross-classification and unconditional logistic regression methods (25). Results of the stratified analysis were similar to those obtained from therefore include the proxy interviews except for Tables 6 and 7 when it was believed a priori that only the subject himself was likely to be aware of the specific job exposures. Some forms of colon cancer have an hereditary component and/or predisposing medical condition, especially those occurring early in life the matched analysis. All subsite analyses reported in this paper were based on the unconditional scheme since they provide more stable risk estimates. Subjects who failed to answer any particular question were eliminated from that analysis, as was the pair in a matched analysis when either case or control failed to answer. All reported P values are (19, 20). Twelve cases (and one control) in this series reported at least two-sided. 1 first degree relative with large bowel cancer, 2 of these had multiple polyposis as well. Another 6 cases (and no control) gave a history of chronic colon disease: 2 with ulcerative colitis; 3 with nonulcerative RESULTS colitis; and 1 with adenomatous polyps. Analyses were performed with and without these subjects and no appreciable differences were found; the reported results include these 18 cases. Conversion factors were used to compute total ethanol intake from reported amounts of beer, wine, and spirits: 1 can of beer = 12.96 g ethanol; 1 glass of dinner wine = 10.10 g ethanol; 1 glass of sweet wine = 15.76 g ethanol; and one jigger of spirits = 14.03 g ethanol (21). No question in the questionnaire addressed physical activity directly, since physical activity was not a hypothesis when the study was initiated. Job activity, however, was assessed in two ways, (a) Activity levels were inferred from job titles. All job titles were coded using the 1970 United States Bureau of Census Index of Industries and Occupations (22). Each job title in this Index had been assigned previously to one of three activity levels (9): "mainly sedentary" when physical activity is required less than 20% of the time spent on the job; "moderately active" when physical activity is required between 20 and 80% of the time; and "very active" when physical activity is required more than 80% of the time. (b) For colon pairs only, one of us (D. G.) blindly (with respect to case or control status) examined each questionnaire and, for each job, estimated the hours per day spent sitting, standing, or moving about; being moderately active; and being vigorously active. These judgments were based on a combination of job title, industry, employer, dates of employment, major responsibilities, and reported exposures to chemi cals, dusts, and other occupational hazards. A weighted job activity score was formed by multiplying the proportion of time spent in category i (i = 1, 2, 3, and 4) by weight <*, and then summing the four cross-product terms. These weights were chosen to be roughly propor tional to the ratio of kilocalories normally expended at the four levels of activity (23). Since the 3-level classification system based on job titles conservatively classifies over 70% of all subjects as "moderately active," the middle level was used as the baseline category in computing ORs3 for both methods of assessing job activity. Occupational exposures to specific substances were examined in three ways, (a) Exposures were inferred from job titles alone. Each job title in the United States Bureau of Census 1970 Index had been assigned previously to one of 10 categories of particulate exposure reflecting the type(s) of dust exposure likely to occur in jobs with that title (24). (b) For each job in their job history, subjects were asked whether they inhaled "chemical solvents, dusts, or other fumes" and whether they routinely "got chemicals, oils, dust, etc. on their skin or clothes." When responses to either question were positive, the specific The cases ranged in age between 24 and 44 years, but 74% were between 35 and 44 years of age at diagnosis. Only 4 of the tumors were diagnosed at the preinvasive stage; roughly onefifth were classified as mucinous adenocarcinoma. By primary subsite, the cases were divided as follows: 1 in the ileocecum; 19 in the cecum; 18 in the ascending colon; 6 at the hepatic flexure; 18 in the transverse colon; 6 at the splenic flexure; 13 in the descending colon; 25 in the sigmoid; 13 at the rectosig moid; and 28 in the lower parts of the rectum. Controls were similar to cases in their distributions by age, height, marital status, religion, birthplace, and ethnicity (Latino versus non-Latino). Cases tended to be less educated (P = 0.02, t test) and slightly more obese (P = 0.02) (Table 1). Subjects were asked if, over most of their adult lives, they had eaten each of several foods or types of foods "once a week or less, 2 to 4 times a week, or 5 times a week or more." For tumors located on the right (ascending) side of the colon, elevated risks were associated with heavy consumption of deep fried foods (OR = 3.9, P = 0.008), fried bacon or ham (OR = 2.6, P = 0.08), and barbecued or smoked meats (OR = 2.9, P = 0.02) (Table 2). These associations were not seen for the other subsites of the colon. Consumption of fresh fruits and raw vegetables was inversely related to risk in the colon (P for trend = 0.006) but not the rectum; this protective effect was strongest in the transverse through descending colon (P for trend = 0.004). A preference for whole grain bread was also protective in the colon (P for trend = 0.02) but not the rectum. The only item associated with rectal cancer was deep fried food (P for trend = 0.01). Beef was not associated with risk at any of the four subsites, and no consistent pattern of risk was discernible for consumption of milk. Adjustment for Quetelet's index did not alter any of these results. Cigarette smoking and alcohol consumption were unrelated to cancer at any of the four subsites in these young men (Table 3). Cases and controls were similarly distributed among the never smokers, exsmokers, and current smokers; neither aver age daily ethanol intake nor cumulative ethanol drink-years substances inhaled or contacted were elicited and coded. Finally, sub jects were asked if they had ever held a job in which they were exposed to each of 23 specified substances or processes. The pathology report for each case was examined to ascertain the subsite location of the primary lesion. Subsite analyses were based on the following groups of subsites: (<z) right-sided (ileocecum, cecum, and ascending colon); (b) transverse/descending colon (hepatic flexure, transverse colon, splenic flexure, and descending colon); (c) sigmoid (only sigmoid); and (d) rectum (rectosigmoid, and other rectum). Standard matched-pair methods (25) were used to compare interview responses between cases and their controls. Study variables were ex amined individually and simultaneously. The exact binomial test was used on individual dichotomous variables, and multivariate logistic regression was used for variables with more than two levels as well as for the multivariate analyses. Due to the limited number of cases at each subsite, we also examined case-control differences in various were associated with significantly increased risk at any subsite. Specific types of alcoholic beverages consumed were also ex amined separately, and none, including beer, was associated with excess risk. Table 4 shows risk by job activity. When activity levels were based on the job title of the longest held job, sedentary jobs were associated with significant excess risk in the transverse/ descending colon (OR = 3.0, P < 0.05), but not at the other colorectal subsites. The same patterns of risk prevailed when activity levels were based on the title of the most recent job, the number of years in a sedentary job, and the proportion of job history spent in a sedentary job. When job activity levels were based on tertiles of the weighted job activity scores for all jobs held between the ages of 18 and 28 years, at the time when these tumors are presumed to have been forming, we again 3 The abbreviation used is: OR, odds ratio. observed an increased risk associated with job inactivity which 5460 BOWEL CANCER IN YOUNG MEN BY SUBSITE Table 1 Demographic characteristics ofsubjects: colorectal cancer, Los Angeles County Cases Characteristic Controls [All (n = 147)] All (11=147) Age at diagnosis (yr) Yr of education Height (cm) Wt 1 yr ago (kg) Quetelet's index* Married (%) Hispanic white (%) Catholic (%) Protestant (%) 38.2 5.7 14.5 3.2 180.0 7.3 81.3 12.4 25.0 3.4 77.6 6.8 27.9 40.1 38.5 5.2 13.7 2.7* 179.0 7.8 83.5 13.4 26.0 3.7* 76.2 10.9 32.7 42.2 Mean SD. * P < 0.0S (t test for comparison with all controls). e Weight (kg) 1 year ago, divided by height (meters) squared. Right sided (n = 38) 38.5 4.8 13.9 2.7 178.6 10.0 83.3 13.4 26.2 4.5 71.1 7.9 34.2 47.4 Transverse/ descending ( = 43) 38.8 4.9 13.8 3.1 179.2 7.6 82.0 14.5 25.4 3.5 79.1 9.3 25.6 41.9 Sigmoid ( = 25) 38.2 5.3 13.1 2.6* 178.9 6.4 85.0 13.1 26.5 3.3 72.0 16.0 32.0 44.0 Rectum (n = 41) 38.5 5.8 13.8 2.4 179.2 6.6 84.2 12.7 26.1 3.2 80.5 12.2 39.0 36.6 Table 2 Risk by subsitefor usual consumption ofspecifiedfoods (over most ofadult life): colorectal cancer, Los Angeles County Times eaten/wk Cases/ controls OR (95% Cl)* [All cases (/i = 147)] ORc (95% Cl) --------------------------------------------------------------------------------------------- Right Transverse/ side descending Sigmoid Rectum (n = 38) (* = 43) (#! = 25) (n = 41) Deepfriedfoods <1 71/81 1.0 1.0 1.0 2-4 51/55 1.0 0.8 0.8 (0.6, 1.8) (0.3, 1.9) (0.4, 1.7) >5 25/11 2.1 3.9 1.2 (1.0,4.6) (1.4, 10.7) (0.4,4.2) Fried bacon or ham <1 2-4 >5 89/85 45/51 13/11 1.0 0.8 (0.5,1.3) 1.0 (0.4, 2.4) 1.0 1.0 (0.5, 2.3) 2.6 (0.9, 7.9) 1.0 0.8 (0.4,1.6) 0.2 (0.02, 1.7) Barbecued or smoked meats <1 125/130 1.0 1.0 1.0 2-4 20/161 1.3 2.9 1.1 >5 2/1 f (0.6, 2.7) (1.2, 7.3) (0.4, 3.1) Beef <1 2-4 3/7 1 59/54 J 1.0 1.0 1.0 >5 85/86 1.0 1.0 0.9 (0.6, 1.6) (0.5, 2.0) (0.4, 1.7) Freshfruits or raw vegetables (e.g., salads)d <1 15/10 1.7 1.6 4.0 (0.7,4.1) (0.4, 6.7) (1.3, 12.3) 2-4 46/37 1.8 2.0 2.7 (1.0, 3.1) (0.9, 4.7) (1.2, 6.1) >5 75/99 1.0 1.0 1.0 Bread preference White None Whole grain 78/64 30/32 39/51 1.0 0.8 (0.4, 1.4) 0.6 (0.4,1.1) 1.0 0.4 (0.1, 1.2) 0.6 (0.3, 1.4) 1.0 0.7 (0.3, 1.7) 0.5 (0.2, 1.2) Milk <1 34/40 1.0 1.0 1.0 2-4 27/18 1.6 1.3 2.9 (0.8, 3.3) (0.4, 4.0) (1.0, 8.2) >5 86/89 1.0 0.8 1.0 (0.6, 1.8) (0.3, 1.8) (0.4, 2.4) * Adjusted for education, age was a matching variable. * Cl, confidence interval. c Adjusted for age and education in analyses using all controls. d Numbers do not sum to 147 since this item was added after 11 cases and 1 control had already been interviewed. 5461 1.0 1.2 (0.5, 3.0 1.1 (0.2, 5.6) 1.0 0.5 (0.2, 1.4) 0.0 (-) 1.0 1.0 (0.3, 3.8) 1.0 1.1 (0.4, 2.6) 1.7 (0.4, 7.9) 1.6 (0.6, 4.4) 1.0 1.0 1.1 (0.4, 3.2) 0.7 (0.2, 1.9) 1.0 0.4 (0.1, 2.4) 0.7 (0.3, 1.9) 1.0 1.4 (0.6, 3.0) 4.3 (1.5, 12.1) 1.0 0.9 (0.4, 2.0) 1.5 (0.4, 4.8) 1.0 0.8 (0.3, 2.7) 1.0 1.1 (0.6, 2.3) 0.5 (0.1, 2.5) 0.6 (0.3, 1.6) 1.0 1.0 1.5 (0.6, 3.6) 1.0 (0.4, 2.3) 1.0 2.1 (0.6, 7.4) 2.1 (0.8, 5.4) BOWEL CANCER IN YOUNG MEN BY SUBSITE Table 3 Risk by subsitefor cigarette smoking and alcohol intake: colorectal cancer, Los Angeles County Cases/ controls OR* (95% Cl)* [all cases (n = 147)] ORf (95% Cl) Colon (n = 106) Rectum (n = 41) Cigarette smoking Never Exsmoker l pack/day 2+ packs/day 48/43 34/38 23/27 38/30 1.0 1.0 1.0 0.7 (0.4, 1.4) 0.6 (0.3, 1.3) 0.7 (0.3, 1.8) 0.7 (0.3, 1.4) 0.4 (0.2, 1.0) 0.9 (0.3, 2.5) 0.9 (0.4, 1.8) 1.1 (0.5, 2.1) 0.5 (0.2,1.5) Daily ethanol intake (g) 0-9 10-39 40-69 70+ 61/63 39/38 25/33 20/12 1.0 1.0 1.0 1.0 (0.5, 1.8) 1.0 (0.5, 1.9) 1.2 (0.5, 2.7) 0.7 (0.4, 1.3) 0.8 (0.4, 1.5) 0.6 (0.2, 1.8) 1.6 (0.7, 3.7) 1.6 (0.6, 3.7) 1.4 (0.4,4.5) Years drank beer at least once/wk 0 1-9 10-19 20 56/59 18/21 37/35 34/32 1.0 1.0 1.0 0.9 (0.4, 1.8) 0.9 (0.4, 2.0) 0.7 (0.2, 2.4) 1.1 (0.6, 2.1) 1.0 (0.5, 1.8) 1.5 (0.6, 3.5) 0.9 (0.5,1.8) 1.0 (0.5, 2.0) 0.9 (0.3, 2.3) a Adjusted for education; age was a matching variable. * Cl, confidence interval. c Adjusted for age and education in analyses using all controls. was specific for the transverse/descending colon. Similar results were obtained when the analysis was based on the weighted job activity scores of the most recent job, the longest held job, and the entire job history. Table 5 shows risk by exposure to any job judged to be dusty on the basis of its job title. Throughout the colorectum, having ever held a dusty job was associated with a modest excess risk (OR = 1.7; P = 0.07); there were significant dose-response relationships with both the number of years in a dusty job and the interval since the first dusty job. The excess risk was strongest in the rectum (OR = 2.1; P = 0.09), followed closely by the sigmoid (OR = 1.9; P = 0.25); there was no excess risk in the midsection of the colon and a small excess in the cecum/ ascending colon (OR = 1.6; P = 0.26). This same pattern of risk persisted for exposure to most of the specific types of dust, but the greatest excess risks were associated with organic and metal dusts, for which there were 3-fold excess risks in the rectum (P = 0.03 and 0.03 for organic and metal dusts, respec tively). Table 6 shows the risk associated with several job character istics (excluding proxy interviews). There was no excess risk associated with getting chemicals on one's skin/clothes, but for tumors of the rectum, there was an almost 4-fold excess risk (P = 0.04) for ever inhaling dusts or fumes on a job. Also for rectal cancer, there were significant dose-response relationships with both the years on jobs in which dust or fumes were inhaled (P = 0.04) and the interval since the first such job (P = 0.045). Over 90% of these dust/fume-exposed tumors occurred at least 10 years after this first exposure. There was also a 2-fold elevation in risk (P = 0.28) of sigmoid cancer after 10 or more years of inhaling dusts/fumes. Tumors located above the sig moid were not associated with inhaling dusts/fumes, except, possibly, 20 years after the first exposure. Substances described by subjects as inhaled on a job were grouped first as particulates (dusts and metal fumes) and non particulates (vapors and gases) and then by type of substance within each category. In the rectum, moderate but nonsignifi cant excess risks were associated with both particulates (OR = 3.4, P = 0.06) and nonparticulates (OR = 3.0, P = 0.09). Examination of specific types of dust indicated that the risk of rectal cancer was highest for wood dust (OR = 9.4, P = 0.005) (Table 6). Other organic dusts showed no clear association with rectal cancer, but there were nonsignificant elevations in risk for metal (OR = 2.4, P = 0.26) and mineral dusts (OR = 3.0, P = 0.14). No specific type of nonparticulates explained the elevated risk for rectal cancer associated with inhaling vapors and gases. In the sigmoid colon, the pattern of risk associated with selfdescribed particulates closely resembled that in the adjacent rectum, but the risk estimates in the sigmoid were not as high and none was statistically significant (Table 6). Moderately elevated risks were present for both wood (OR = 3.6, P = 0.16) and metal dusts (OR = 2.0, P = 0.38). The only exposure for which the pattern of risk was not similar between the rectum and sigmoid was to nonwood organic dusts (OR = 3.2 in the sigmoid and 1.2 in the rectum); these included textile dust, soil, and nonspecific household dust. In the right, transverse, and descending colon there was no significant association with any inhaled dust or gas, but the risks for all organic dusts were somewhat elevated. When subjects were asked directly about 23 specific sub stances and industrial processes, only nonferrous metal expo sure was significantly associated with sigmoid cancer (OR = 3.0, P = 0.03) (Table 7). Review of the metals involved did not implicate a specific metal; aluminum, lead, beryllium, copper, tin, zinc, titanium, and molybdenum were all represented. Ex- Activity level of longest held job Mainly sedentary Moderately active Very active Table 4 Risk by subsiteforjob activity: colorectal cancer, Los Angeles County Cases/ controls ORa (95% Cl)* [all cases (n = 147)] Right side (n = 38) ORc (95% ci) Transverse/ descending <* = 43) Sidmoid <* = 25) 28/21 101/109 18/17 1.7 (0.9, 3.4) 1.0 1.1 (0.5, 2.3) 1.0 (0.3, 3.1) 1.0 1.0 (0.3, 3.2) 3.0 (1.2, 7.2) 1.0 0.8 (0.2, 2.7) 1.7 (0.5, 6.0) 1.0 1.6 (0.5, 5.1) Tertiles of weighted job activity score for ages 18 through 28 yr* 1st (least active) 2nd 3rd (most active) 30/35 42/35 34/36 0.8 (0.4, 1.6) 1.0 0.8 (0.4, 1.5) 0.7 (0.3, 1.7) 1.0 0.5 (0.2, 1.2) * Adjusted for education; age was a matching variable. * Cl, confidence interval. c Adjusted for age and education in analyses using all controls. 4 Score was not computed for cases of rectum cancer and their matched controls. 5462 1.6 (0.7,4.1) 1.0 0.9 (0.4, 2.3) 0.2 (0.04, 1.1) 1.0 1.0 (0.4, 2.7) Rectum (* = 41) 1.5 (0.6,4.0) 1.0 0.7 (0.2, 2.5) BOWEL CANCER IN YOUNG MEN BY SUBSITE Ever held a dusty job No Yes Table 5 Risk by subsitefor exposure to dustyjobs (as inferredfrom job titles): colorectal cancer, Los Angeles County Cases/ controls OR (95% Cl)* [all cases (,, = 147)] OR' (95% Cl) ------------------------------------------------------------------------------- Right Transverse/ side descending Sidmoid (* = 38) (* = 43) (* = 25) 39/57 108/90 1.0 1.7 (1.0, 2.9) 1.0 1.6 (0.7, 3.8) 1.0 1.0 (0.5, 2.1) 1.0 1.9 (0.6, 5.8) Years held a dusty job 1-9 10+ Test for trend {P) 55/51 53/39 1.6 (0.9, 2.8) 2.0 (1.0, 4.0) 0.05 1.4 (0.6, 3.6) 2.1 (0.7, 5.8) 0.16 1.1 (0.5, 2.4) 0.8 (0.3, 2.3) >0.50 1.9 (0.6, 6.1) 2.0 (0.5, 7.2) 0.32 Years since first held a dustyjob 1-9 10-19 20+ Test for trend (P) 5/10 40/30 63/50 0.3 (0.1, 1.8) 1.7 (0.9, 3.4) 1.9 (1.0, 3.6) 0.03 0.8 (0.1, 4.8) 0.9 (0.3, 3.0) 2.3 (0.9, 6.0) 0.10 0.0 (-) 1.4 (0.5, 3.7) 0.9 (0.4, 2.2) >0.50 0.8 (0.1, 9.7) 2.5 (0.7, 9.4) 1.8 (0.5, 6.1) 0.26 Type of dust on dusty job Organic dust Metal dust Mineral dust Smoke or exhaust Unknown dust 39/30 47/34 27/26 18/21 70/60 1.8 (0.9, 3.8) 1.9 (1.0, 3.9) 1.4 (0.7, 3.0) 1.2 (0.5, 2.6) 1.6 (0.9, 2.9) 1.7 (0.6, 5.0) 1.1 (0.4, 3.5) 1.4 (0.4,4.3) 1.4 (0.4, 4.8) 1.8 (0.8, 4.4) 0.5 (0.2, 1.7) 0.9 (0.3, 2.5) 0.8 (0.2,2.3) 0.4 (0.1,1.7) 1.0 (0.4, 2.2) a Adjusted for education using "never dusty job" as reference group; age was a matching variable. * Cl, confidence interval. c Adjusted for age and education in analyses using all controls and "never dusty job" as reference group. 2.1 (0.6, 8.0) 2.7 (0.8, 9.5) 1.5 (0.4, 6.3) 1.5 (0.3, 6.8) 1.5 (0.4, 4.9) Rectum (* = 41) 1.0 2.1 (0.9, 5.0) 2.0 (0.8, 4.9) 2.4 (0.9, 6.7) 0.09 1.2 (0.2, 7.8) 2.4 (0.8, 6.9) 2.1 (0.8, 5.4) 0.09 3.0 (1.1, 8.6) 3.0 (1.1, 8.0) 1.7 (0.6, 5.4) 1.3 (0.4,4.8) 2.1 (0.8, 5.2) posure to nonferrous metals was also somewhat associated with rectal cancer, but this OR was not statistically significant (OR = 2.0, P = 0.08). Interestingly, exposure to ferrous metals showed no association with cancer at any site in the colon or rectum. Wood dust when asked as a specific item was associated with sigmoid and rectum cancer, but not significantly, and not as strongly as when subjects were asked to name the substances they inhaled on specific jobs. Direct questions were also asked about leisure time exposure to 17 specified substances. No significantly elevated risks were found in association with any of these substances. Multivariate analyses demonstrated that the findings for the dietary factors (deep fried foods, fresh fruits or raw vegetables, and bread preference), for physical activity, and for occupational exposure to dusts/fumes were not confounded by each other; i.e.9 the ORs for each of these variables were not substantively altered after adjustment for the other variables. Similarly, ad justment of these same variables for Quetelet's index did not substantively alter any of these risk estimates. barbecued/cured meats increased risk only in the right colon. Occupational physical activity was protective in the midsections of the colon (transverse/descending subsites) but not in the more proximal or distal segments of the colorectum. Tumors in the rectum and to a lesser degree the sigmoid were associated with dusty jobs. This latter effect was somewhat nonspecific; /.., all types of dusts and fumes increased risk to some degree. However, wood and metal dust conveyed the greatest risk in both the rectum and the adjacent sigmoid. Beef, milk, alcohol, and cigarette smoking appeared to play little or no role in the etiology of these tumors at any subsite. Perhaps more impor tantly, however, this case-control study of colorectal cancer in young men does not support the hypothesis that different subsites of the colorectum share the same environmental risk factors. Because this study was limited to young cases, the findings may not be generalizable to bowel cancers arising in older subjects. Tumors occurring early in life may be more likely to have a hereditary component and/or a predisposing medical DISCUSSION condition and may represent a different spectrum of disease than those which occur later in life. Nonetheless, we found no Analytical studies of colorectal cancer have revealed associ ations with exposures which have been diverse, often inconsist ent, and usually of relatively small magnitude. In this study of colorectal cancer in young men, we observed several rather strong associations, particularly when restricting the focus to specific segments of the large bowel. Consumption of fresh fruits and raw vegetables and a preference for whole grain breads were protective throughout the colon but not in the rectum; fatty foods increased risk at both proximal and distal important differences in levels of risk when our analyses ex cluded cases with a family history of colorectal cancer and/or chronic colitis. Furthermore, all exposures associated with bowel cancer in this study have been reported in previous studies which included youthful cases as a small fraction of the subjects if at all. On the other hand, a study limited to young cases does have special advantages. The interval between exposure and disease onset is comparatively short, and thus recall of past experience ends of the colorectum but not at the intervening subsites; and is relatively accurate. Such a short interval may also reflect a 5463 BOWEL CANCER IN YOUNG MEN BY SUBSITE Table 6 Risk by subsiteforjob characteristics andjob exposures described by subjects (self-respondents only): colorectal cancer, Los Angeles County Cases/ controls OR" (95% Cl)b [all cases (n = 147)] ORc (95% Cl) ------------------------------------------------------------------------------------------- Right Transverse/ side descending Sigmoid Rectum (n = 38) (n = 43) (* = 25) (n = 41) Every held job in which "chemicals, oils. or dusts got on skin or clothes** No Yes 45/52 77/87 1.0 1.0 (0.6, 1.7) 1.0 0.8 (0.3, 1.7) 1.0 0.7 (0.3, 1.6) 1.0 1.2 (0.4, 3.9) 1.0 1.0 (0.5, 2.3) Ever held job in which "chemicals, dust, or fumes were inhaled** No Yes 20/35 102/104 1.0 1.5 (0.8, 3.0) 1.0 1.1 (0.4, 2.8) 1.0 1.1 (0.4, 2.9) 1.0 1.4 (0.4, 5.4) 1.0 3.8 (1.1, 13.6) Years in a job in which chemicals, dust, or fumes were inhaled 0 1-9 10+ Test for trend (P) 20/35 51/50 51/54 1.0 1.6 (0.8, 3.2) 1.5 (0.7, 3.2) 0.38 1.0 1.4 (0.5, 3.8) 0.8 (0.3, 2.4) >0.50 1.0 1.5 (0.6, 4.1) 0.7 (0.2, 2.2) 0.43 1.0 0.8 (0.2, 3.8) 2.2 (0.5, 9.4) 0.17 1.0 3.4 (0.9, 12.9) 4.3 (1.1, 16.5) 0.04 Years since first job in which chemicals, dust, or fumes were inhaled 0 1-9 10-19 20+ Test for trend (P) 20/35 6/16 41/38 55/50 1.0 0.3 (0.1, 1.3) 1.7 (0.7, 4.0) 1.7 (0.8, 3.8) 0.06 1.0 0.3 (0.1, 2.1) 0.8 (0.2, 2.5) 1.8 (0.6, 5.5) 0.28 1.0 0.2 (0.02, 2.0) 1.1 (0.4, 3.4) 1.5 (0.5, 4.4) 0.33 1.0 0.0 (-) 1.6 (0.4, 7.1) 2.4 (0.5, 12.3) 0.16 1.0 1.8 (0.3, 11.1) 4.9 (1.3, 18.4) 3.4 (0.9, 13.7) 0.045 Substances described as inhaled on a job Dusts or metal fumes Organic dusts Wood dust Other organic dust Metal dusts Mineral dusts Vapors or gases 71/73 32/19 21/10 11/9 24/27 26/30 89/100 a Adjusted for education; age was a matching variable. b Cl, confidence interval. e Adjusted for age and education in analyses using all controls. 1.6 (0.8, 3.3) 2.6 (1.1, 6.3) 3.6 (1.2, 11.0) 1.9 (0.6, 5.5) 1.2 (0.5, 2.9) 1.4 (0.6, 3.2) 1.3 (0.6, 2.6) 0.9 (0.3, 2.6) 2.1 (0.6, 6.6) 2.1 (0.5, 8.5) 2.1 (0.5, 8.7) 0.8 (0.2, 2.8) 1.0 (0.3, 3.4) 1.0 (0.4, 2.5) 1.1 (0.4,2 .9) 1.5 (0.4, 5.2) 1.5 (0.3, 6.6) 1.6 (0.3, 7.4) 1.0 (0.3, 3.3) 0.9 (0.3, 3.1) 1.0 (0.4, 2.7) 1.7 (0.4,6.9) 3.4 (0.7, 15.4) 3.6 (0.6, 20.5) 3.2 (0.5, 19.3) 2.0 (0.4,9.5) 1.3 (0.3,6.7) 1.3 (0.3, 5.2) 3.4 (0.9, 12.6) 5.2 (1.2, 22.1) 9.4 (2.0, 44.7) 1.2 (0.1, 13.5) 2.4 (0.5, 10.8) 3.0 (0.7, 12.6) 3.0 (0.8,10.9) single unusually intense exposure, and therefore perhaps a cantly associated with rectal cancer using this method of as predominant causal element. signing exposure. When subjects were asked to name the spe Recall bias is not likely to have produced the positive findings. cific exposures associated with their jobs, organic and metal The interviewer followed the same structured questionnaire and dusts were again associated with rectal cancer, although only used the same probes and follow-up questions with all subjects the former association was statistically significant. Exposure to [and indeed with subjects in several other cancer case-control wood dust was primarily responsible for this association be studies conducted simultaneously (26-30)]. If cases had system tween organic dust and rectal cancer. As it turned out, almost atically recalled more exposures than controls, findings would 70% of the subjects were classified the same way by job title not have been so site specific or exposure specific. Laymen and self-report, and the strengths of the observed risks were generally fail to distinguish between cancers at different sites, enhanced when the analysis was limited to this subset. The much less between cancers at subsites within the large bowel. third method of assessing exposure, i.e., asking subjects to Some exposures conventionally linked with cancer, namely, review lists of specific exposures, is designed to minimize the smoking, beef, alcohol consumption, and direct chemical con recall bias inevitably associated with volunteered reports but tact, were not associated with elevated risk at any subsite. tends to overestimate less salient exposures. This it is not Finally, a number of specific exposures were, if anything, in surprising that the associations with both organic and metal versely associated with risk. dusts were weakest when exposures were assessed by this third The positive occupational findings were generally consistent method. over the different methods of assessing exposure. Inferring We observed a small association with obesity, and this asso exposure from job titles offers objectivity and freedom from ciation was relatively consistent over the subsites of the colo- bias, but it may result in substantial misclassification. Nonethe rectum. Other case-control studies have reported cases to be less, exposures to both organic and metal dusts were signifi both more (12, 31, 32) and less (33, 34) obese than controls. 5464 BOWEL CANCER IN YOUNG MEN BY SUBSITE Table 7 Risk by subsitefor occupational exposure to prespecified substances and processes (self-respondents only): colorectal cancer, Los Angeles County Cases/ controls OR" (95% Cl)* [all cases (n = 147)] ORr (95% Cl) ---------------------------------------------------------------------------------------------------- Right Transverse/ side descending Sigmoid Rectum (n = 38) (n = 43) (n 25) (n = 41) Wood dust Grain dust Any metal dusts or fumes Ferrous metals Nonferrous metals Cutting, cooling, lubricating oils Asbestos Fibrous glass or glass wool Pesticides Wood preservatives Paints and lacquers Spray paints Petroleum products Organic solvents Coal tar, soot, pitch Arsenic Dyestuffs Plastics processing Rubber processing 28/26 10/7 42/43 21/25 38/34 40/42 14/20 19/36 12/22 6/12 37/43 27/44 45/51 67/68 13/19 6/6 17/15 15/16 6/7 1.2 (0.6, 2.3) 1.9 (0.6, 5.6) 1.1 (0.6, 1.9) 0.8 (0.4, 1.7) 1.3 (0.8, 2.4) 1.1 (0.6, 2.0) 0.8 (0.4, 1.7) 0.4 (0.2, 0.8) 0.6 (0.3, 1.3) 0.5 (0.2, 1.4) 1.0 (0.6, 1.8) 0.6 (0.3, 1.1) 1.0 (0.6, 1.7) 1.3 (0.7, 2.2) 0.8 (0.4, 1.7) 1.4 (0.4, 5.0) 1.5 (0.7, 3.4) 1.1 (0.5, 2.6) 1.1 (0.3, 4.0) 1.1 (0.4, 2.9) 2.5 (0.7, 9.2) 0.6 (0.2, 1.4) 0.6 (0.2, 1.8) 0.7 (0.3, 1.8) 0.9 (0.4, 2.2) 0.4 (0.1, 1.6) 0.4 (0.2, 1.3) 0.5 (0.1, 1.9) 0.0 (-) 0.7 (0.3, 1.6) 0.3 (0.1, 1.0) 0.7 (0.3, 1.7) 0.9 (0.4, 2.1) 0.6 (0.2, 2.3) 1.6 (0.3, 8.2) 1.5 (0.5, 4.4) 1.5 (0.5,4.4) 1.3 (0.3, 6.7) 0.6 (0.2, 1.8) 0.5 (0.1, 4.0) 0.7 (0.3, 1.7) 0.7 (0.2, 2.0) 0.9 (0.4, 2.2) 0.9 (0.4, 2.2) 0.5 (0.1, 1.8) 0.5 (0.2, 1.3) 0.3 (0.1, 1.4) 0.6 (0.1, 2.9) 1.1 (0.5, 2.4) 0.4 (0.1, 1.1) 0.9 (0.4, 2.1) 1.3 (0.6, 2.9) 1.0 (0.3, 2.9) 0.0 (-) 1.0 (0.3, 3.3) 1.1 (0.3, 3.4) 0.0 (-) 1.9 (0.6, 5.7) 3.2 (0.7, 14.4) 2.0 (0.7, 5.5) 1.1 (0.3, 3.7) 3.0 (1.1, 8.2) 0.8 (0.3, 2.6) 1.5 (0.4, 5.0) 0.5 (0.1, 1.8) 0.6 (0.1, 2.9) 1.3 (0.3, 6.5) 0.7 (0.2, 2.3) 0.7 (0.2, 2.1) 1.0 (0.3, 2.7) 1.4 (0.5, 4.0) 0.3 (0.04, 2.8) 3.1 (0.5, 16.9) 3.9 (1.2, 12.1) 0.5 (0.1, 3.9) 2.7 (0.5,14.4) 1.4 (0.6, 3.4) 0.9 (0.2, 4.9) 1.7 (0.8, 3.7) 1.1 (0.5, 2.8) 2.0 (0.9, 4.2) 1.1 (0.5, 2.4) 0.8 (0.3, 2.4) 0.4 (0.1, 1.1) 0.8 (0.3, 2.3) 0.6 (0.1, 2.7) 1.0 (0.4, 2.1) 0.8 (0.4, 1.8) 1.1 (0.5, 2.2) 1.1 (0.5, 2.3) 0.7 (0.2, 2.3) 1.3 (0.3, 6.8) 0.4 (0.1, 2.0) 1.2 (0.4, 3.4) 1.1 (0.2, 5.7) a Adjusted for education; age was a matching variable. * Cl, confidence interval. e Adjusted for age and education in analyses using all controls. The observed protection against colon cancer associated with County (9); this finding has since been reproduced in several dietary fruits, raw vegetables, and whole grain bread is consist other settings using different study designs (10-12, 52-54). ent with previous case-control studies (35-38). The risk con While three-fourths of the cases in the present case-control veyed by deep fried foods is consistent with elevated risks from study comprise 23% of the cases in the initial report who were fat intake (17, 32-34, 39), and the elevated risk of cecum and under age 45 years at diagnosis, it should be noted that the ascending colon cancer after cured/barbecued meat consump initial descriptive study found the protective effect for job tion is consistent with the hypothesis that /V-nitroso compound activity to be strongest in men over the age of 45 years; in men formation increases risk (36,40). under age 45 years, the protective effect was present only among The association between colorectal cancer and consumption those residing in lower socioeconomic neighborhoods. The of alcohol, especially beer, has been widely studied, and while present trend of increasing risk for cancers of the middle bowel some investigators have found a moderate excess risk for heavy with decreasing activity and the absolute magnitude of risk at alcohol consumption (12, 35, 41, 42), especially for rectal that subsite are consistent with both the original and subsequent cancer after heavy consumption of beer (43-45), others have reports. found none (46-49). While we found no significant excess risk Wood dust is a known cause of nasal cancer (55-57) and has associated with any specific beverage, with derived indices of been associated with lung, stomach, and bladder cancers in daily ethanol intake, or with cumulative drink-years, the level some but not all studies (55,58,59). A higher risk for colorectal of association between rectal cancer risk and heavy alcohol cancer was observed in three independent cohort studies of consumption was at least consistent with a small excess risk, woodworkers in the automotive industry (6-8); a screening even though heavy or long term beer consumption did not survey in the same industry found a higher prevalence of colo explain it. If the previously observed relationship is real, the rectal polyps among pattern and model makers compared to subjects in this study may have been too young for the cumu other workers (60). While an excess of rectal but not colon lative effects of beer or alcohol to become manifest. cancer was linked to the lumber and wood products industry in The absence of a smoking effect for either colon or rectal the Third National Cancer Survey Interview (16), cohort studies cancer is consistent with the findings of most other studies (36, of carpenters and furniture workers (56, 61) have failed to find 41,46, 50, 51). an excess risk for colorectal cancer; and the large American The initial report of a protective effect for colon cancer from Cancer Society (ACS) cohort study found a reduced incidence physical activity on the job was based on cases from Los Angeles ofcolorectal cancer among woodworkers, defined as carpenters, 5465 BOWEL CANCER IN YOUNG MEN BY SUBSITE sawmill operators, furniture makers, or one of 30 other wood- movements (83, 84), related to physical activity (85), deliver related occupations (58). The ACS investigators did observe stool toward the sigmoid. The lower colon and rectum are significant excess risks for lung, stomach, and bladder cancer capable of distension and serve a reservoir function between as well as nasal cancer, while excess risk was found only for powerful defecatory contractions (86). Other subsite variations nasal cancer in the cohort studies of carpenters and furniture which must bear on the degree of contract between stool ele workers. This study differs from previous studies not only in ments and mucosal cells include luminal geometry, variations the youth of the cases and in the case-control design but also in the impact of gravity, and the subsite-specific composition in that exposure to wood dust was not based on job title alone. of protective mucus (87). The strongest excess risk was found for the rectum and when Our findings are consistent with the existence of subsite- subjects volunteered wood dust as a substance inhaled on a job. specific mechanisms of carcinogenesis which parallel these The comparable exposure derived only from job title provided physiological differences. Excess risk in the upper colon was a similar statistically significant pattern of excess risk for or associated with animal fat, consistent with the hypothesis that ganic dust. 3-ketosteroids are carcinogenic, and the increased mucosal con For many years investigators have been reporting excess tact and absorption in that segment. Physical activity exerts a colorectal cancer in metal workers, e.g., machinists, millwrights, more profound effect upon transit time, as well as upon risk, in sheetmetal workers, grinders, coppersmelters, tool and die mak the middle colon; the reduction in risk there associated with ers, metal polishers and platers, nickel refiners, and metal frame fruit and vegetable consumption may have a related explana makers (5,16,62-71). More recent studies have reported these tion, based on the effect of increased dietary fiber. Air-borne associations for rectal but not colon cancer (72-74). In this dust particles are swallowed in mucus from the respiratory tract study, excess risk for metal dust was observed using all three and may therefore contact the gastrointestinal mucosa at any methods of assessing exposure and appeared to be restricted to site. The duration of contact with any particular cell in the the rectum and sigmoid. Most previous studies have attempted upper colon is probably short, even though transit time through to explain these associations on the basis of the cutting oils the cecum is relatively long. In the lower bowel, the other site and/or abrasives used in metal work rather than the metal dust of long transit time, there is little churning and mixing of the itself. We found no excess risk associated with self-reported relatively solid stool, and prolonged contact is a distinct possi exposure to "cutting, cooling, or lubricating oils" but we did bility, particularly if particulates are nonabsorbable. It is also not ask specifically about exposure to abrasives. Such sub possible that the transit of particulates of high density, such as stances, however, were rarely mentioned among the dusts which metal particles, is unusually slow through portions of the gas were inhaled on specific jobs. trointestinal tract where flow is reduced generally by anatomic We found no elevated risk after exposure to asbestos. This or physiological means (88). This hypothesis is offered because does not support the 2- to 3-fold excess risk observed by Selikoff in concurrent studies of carcinoma at other gastrointestinal and others (3, 4, 75, 76) but is in accord with the conclusions sites, using the same instrument, we have observed increased of recent reviewers who concluded that the bulk of the evidence risk in relation to metal dust in the lower esophagus (30) and from both epidemiological and animal studies does not support in the pylorus.4 an association between asbestos exposure and colorectal cancer It is quite possible that subsite-specific variation in environ (77, 78). However, few of our subjects held jobs in which mental carcinogenicity is responsible for some inconsistency asbestos exposure would have been heavy, and this study had between the findings of past observational studies, especially low power to detect such an association. since measurement errors in the exposure indices are likely to We also found no association between exposure to pesticides be large. Future studies of large bowel cancer should certainly and either colon or rectal cancer and thus fail to confirm the examine results by subsite as a routine procedure. association suggested by investigators in the Mississippi River Valley where adolescent patients with this disease were noted to have had heavy childhood exposures to agricultural herbi ACKNOWLEDGMENTS cides and pesticides (13, 14, 79, 80). However, our cases were urban and we did not ask about exposures during early child hood. We thank Mohammed S. Hoseyni, Kazuko Arakawa, and Jerzy Lysikowski for their assistance with the data analysis. Abundant evidence of subsite-specific differences in bowel cancer risk has long been available, but the presumption of a REFERENCES common etiology is convenient and has been justified by a wish for parsimony in the face of an incomplete understanding of carcinogenesis. Still, as more evidence of physiological varia tion within the large bowel accumulates (81), the proposition that bowel cancer etiology is uniform becomes less plausible. The cecum and the ascending and proximal transverse colon derive from the embryonic midgut; the distal transverse, de scending, and sigmoid colon is derived from the hindgut; and the rectum and the anal canal from the cloaca. The three subdivisions have correspondingly different blood supplies, and their different patterns of motility suggest different physiolog ical functions. Retrograde peristalsis in the upper colon churn and mix the liquid stool (81), and it is in that region that the 1. Doll, R., and Peto, R. The Causes of Cancer: Quantitative Estimates of Avoidable Risks of Cancer in the United States Today. Oxford, England: Oxford University Press, 1981. 2. Riboli, E., and Sasco, A. J. Current hypotheses on the etiology of colorectal cancer critical review of the epidemiological evidence. Soz. Praventivmed., 31:78-80, 1986. 3. Selikoff, 1. J., Hammond, E. C., and Seidman, H. Mortality experience of insulation workers in the United States and Canada, 1943-76. Ann. NY Acad. Sci., 330:91-116, 1979. 4. Newhouse, M. L., and Berry, G. Patterns of mortality in asbestos factory workers in London. Ann. NY Acad. Sci., 330:53-60, 1979. 5. Berg, J. W., and Howell, M. A. Occupation and bowel cancer. J. Toxicol. Environ. Health, 1:75-89,1975. 6. Schottenfeld, D., Warshauer, M. E., and Zauber, A. G. Study of Cancer Mortality and Incidence in Wood Shop Workers of the General Motors Corporation. New York: Memorial Sloan-Kettering Cancer Center, 1980. 7. Robinson, C., Waxweller, R. J., and McCammon, C. S. Pattern and model traffic of materials into and out of the lumen is greatest. In the middle colon, rings of contraction separate the stool into roughly equivalent aliquots (82), and periodic mass propulsive 4 A. H. Wu-Williams, M. C. Yu, and T. M. Mack. Risk factors for cardia and other stomach cancers in young men in Los Angeles County, submitted for publication. 5466 BOWEL CANCER IN YOUNG MEN BY SUBSITE makers: proportionate mortality 1972-1978. Am. J. Indust. Med., 1: 159 165,1980. 8. Swanson, G. M., and Belle, S. H. Cancer morbidity among woodworkers in the U. S. automotive industry. J. Occup. Med., 24:315-319, 1982. 9. Garabrant, D. H., Peters, J. M., Mack, T. M., et al. Job activity and colon cancer risk. Am. J. Epidemiol., 119:1005-1014,1984. 10. Vena, J. E., Graham, S., Zielezny, M., et al. Lifetime occupational exercise and colon cancer. Am. J. Epidemiol., 122:357-365,1985. 11. Gerhardsson, M., Norell, S. E., Kiviranta, H., et al. Sedentary jobs and colon cancer. Am. J. Epidemiol., 123:775-780,1986. 12. Wu, A. H., Paganini-Hill, A., Ross, R. K., et al. Alcohol, physical activity, and other risk factors for colorectal cancer a prospective study. Br. J. Cancer, 55:687-694, 1987. 13. Pratt, C. B., Rivera, G., Shanks, E., et al. Colorectal carcinoma in adolescents: implications regarding etiology. Cancer (Phila.), 40:2464-2472,1977. 14. Pratt, C. B., George, S. L., O'Connor, D., et al. Adolescent colorectal cancer and dioxin exposure. Lancet, 2:803, 1987. 15. Weisburger, J. H., Reddy, B. S., Spingarn, N. E., et al. Current views on the mechanisms involved in the etiology of colorectal cancer. In: S. Winawer, D. Schottenfeld, and P. Sherlock (eds.). Colorectal Cancer Prevention, Epide miology, and Screening, pp. 19-41. New York: Raven Press, 1980. 16. Williams, R. R., Stegens, N. L., and Goldsmith, J. R. Associations of cancer site and type with occupation and industry from the Third National Cancer Survey Interview. J. Natl. Cancer Inst., 59:1147-1185,1977. 17. Miller, A. B., Howe, G. R., Jain, M., et al. Food items and food groups as risk factors in a case-control study of diet and colo-rectal cancer. Int. J. Cancer, 32:115-161,1983. 18. Mack, T. M. Cancer surveillance program in Los Angeles County. Natl. Cancer Inst. Monogr., 47:99-101, 1977. 19. Peltokallio, P., and Peltokallio, V. Relationship of familial factors to carci noma of the colon. Dis. Colon Rectum, 9:367-370,1966. 20. Jarvinen, H. J., and Turunen, M. J. Colorectal carcinoma before 40 years of age: prognosis and predisposing conditions. Scand. J. Gastroenterol., 19: 634-638,1984. 21. Adams, D. F. Nutritive Value of American Foods in Common Units. Agri culture Handbook No. 456. DHEW Publication (United States Department of Agriculture), p. 31. Washington, DC: Government Printing Office, 1975. 22. United States Bureau of the Census. 1970 Census of Population Alphabetical Index ofIndustries and Occupations. Washington, DC: Government Printing Office, 1971. 23. Astrand, P., and Rodahl, K. Textbook of Work Physiology: Physiological Bases of Exercise, pp. 501-512, New York: McGraw-Hill Book Co., 1986. 24. Wright, W. E., Bernstein, L., Peters, J. M., et al. Adenocarcinoma of the stomach and exposure to occupational dust. Am. J. Epidemiol., 128:64-73, 1988. 25. Breslow, N. E., and Day, N. E. Statistical Methods in Cancer Research, 1ARC Scientific Publications No. 32. Lyon, France: International Agency for Research on Cancer, 1980. 26. Yu, M. C., Mack, T., Hanisch, R., et al. Hepatitis, alcohol consumption, cigarette smoking, and hepatocellular carcinoma in Los Angeles. Cancer Res., 43:6077-6079,1983. 27. Mack, T. M., Peters, J. M., Yu, M. C., et al. Pancreas cancer is unrelated to the workplace in Los Angeles. Am J Indust. Med., 7: 253-266, 1985. 28. Mack, T. M., Yu, M. C., Hanisch, R., et al. Pancreas cancer and smoking, beverage consumption, and past medical history. Natl. Cancer Inst., 76:49 60,1986. 29. Yu, M. C., Mack, T. M., Hanisch, R., et al. Cigarette smoking, obesity, diuretic use, and coffee consumption as risk factors for renal cell carcinoma. J. Natl. Cancer Inst., 77; 351-356,1986. 30. Yu, M. C., Garabrant, D. H., Peters, J. M., et al. Tobacco, alcohol, diet, occupation, and carcinoma of the esophagus. Cancer Res., 48: 3843-3848, 1988. 31. Phillips, R. L., and Snowdon, D. A. Dietary relationship with fatal colorectal cancer among Seventh-Day Adventists. J. Natl. Cancer Inst., 74: 307-317, 1985. 32. Graham, S., Marshall, J., Haughey, B., et al. Dietary epidemiology of cancer of the colon in western New York. Am. J. Epidemiol., 128:490-503, 1988. 33. Jain, M., Cook, G. M., Davis, F. G., et al. A case-control study of diet and colorectal cancer. Int. J. Cancer, 26:757-768, 1980. 34. Bristol, J. B., Emmett, P. M., Heaton, K. W., et al. Sugar, fat, and the risk of colorectal cancer. Br. Med. J., 291:1467-1470,1985. 35. Modan, B., Bareli, V., Lubin, F., et al. Low-fiber intake as an etiologic factor in cancer of the colon. J. Natl. Cancer Inst., 55:15-18,1975. 36. Dales, L. G., Friedman, G. D., Ury, H. K., et al. A case-control study of relationships of diet and other traits to colorectal cancer in American Blacks. Am. J. Epidemiol., 109:132-144,1978. 37. Kune, S., Kune, G. A., and Watson, L. F. Case-control study of dietary etiological factors: the Melbourne colorectal cancer study. Nutr. Cancer, 9: 21-42, 1987. 38. Slattery, M. L., Sorenson, A. W., and Ford, M. H. Dietary calcium intake as a mitigating factor in colon cancer. Am. J. Epidemiol., 128: 504-514, 1988. 39. Martinez, I., Torres, R., Frias, Z., et al. Factors associated with adenocarci nomas of the large bowel in Puerto Rico. Adv. Med. Oncol., 3:45-52,1979. 40. Fraser, P., Chilvers, C., and Beral, V. Nitrate and human cancer a review of the evidence. Int. J. Epidemiol., 9:73-78,1980. 41. Williams, R. R., and Horm, J. W. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: interview study from the Third National Cancer Survey. J. Natl. Cancer Inst., 58:525 547, 1977. 42. Potter, J. D., and McMichael, A. J. Diet and cancer of the colon and rectum: a case-control study. J. Natl. Cancer Inst., 76:557-569,1986. 43. Dean, G., MacLennan, R., McLoughlin, H., et al. Causes of death of blue- collar workers at a Dublin brewery, 1954-73. Br. J. Cancer, 40: 581-589, 1979. 44. Kabat, G. C., Howson, C. P., and Wynder, E. L. Beer consumption and rectal cancer. Int. J. Epidemiol., 15:494-501,1986. 45. Kune, S., Kune, G. A., and Watson, L. F. Case-control study of alcoholic beverages as etiological factors: the Melbourne colorectal cancer study. Nutr. Cancer, 9:43-56, 1987. 46. Graham, S., Dayal, H., Swanson, M., et al. Diet in the epidemiology of cancer of the colon and rectum. J. Natl. Cancer Inst., 61:709-714, 1978. 47. Jensen, M. Cancer morbidity and causes of death among Danish brewery workers. Int. J. Cancer, 23:454-463,1979. 48. Manousos, O., Day, N. E., Trichopoulos, D., et al. Diet and colorectal cancer: a case-control study in Greece. Int. J. Cancer, 32:1-5, 1983. 49. Vobecky, J., Caro, J., and Devroede, G. A case-control study of risk factors for large bowel carcinoma. Cancer (Phila.), 51:196-201,1983. 50. Haenszel, W., Berg, J. W., Segi, M., et al. Large-bowel cancer in Hawaiian Japanese. J. Natl. Cancer Inst., 51:1765-1779,1973. 51. Haenszel, W., Locke, F. B., and Segi, M. A case-control study of large bowel cancer in Japan. J. Natl. Cancer Inst, 64:17-22, 1980. 52. Vena, J. E., Graham, S., Zielezny, M., et al. Occupational exercise and risk of cancer. Am. J. Clin. Nutr., 45:318-327,1987. 53. Slattery, M. L., Schumacher, M. C., Smith, K. R., et al. Physical activity, diet, and risk of colon cancer in Utah. Am. J. Epidemiol., 128: 989-999, 1988. 54. Gerhardsson, M., Floderus, B., and Norell, S. E. Physical activity and colon cancer risk. Int. J. Epidemiol., 17:743-746,1988. 55. Evaluation of the Carcinogenic Risk of Chemicals to Humans, IARC Scien tific Publications No. 25. Lyon, France: International Agency for Research in Cancer, 1981. 56. Rang, E., and Acheson, E. Cancer in furniture workers. Int. J. Epidemiol., 10:253-261, 1981. 57. Acheson, E. E., Pippard, E. C., and Winter, P. D. Mortality of English furniture makers. Scand. J. Work Environ. Health, 10: 211-217,1984. 58. Stellman, S. D., and Garfinkel, L. Cancer mortality among woodworkers. Am. J. Indust. Med., 5:343-357,1984. 59. Siemiatycki, J., Richardson, L., Gerin, M., et al. Associations between several sites of cancer and nine organic dusts: results from an hypothesis-generating case-control study in Montreal, 1979-1983. Am. J. Epidemiol., 123: 235 249,1986. 60. Demers, R. Y., Demers, P., Hoar, S. K., et al. Prevalence of colorectal polyps among Michigan pattern and model makers. J. Occup. Med., 27: 809-812, 1985. 61. Olsen, J., and Sabroe, S. A follow-up study of non-retired and retired members of the Danish carpenter/cabinet maker's trade union. Int. J. Epi demiol., 8:375-382,1979. 62. Milham, S. Cancer mortality patterns associated with exposure to metals. Ann. NY Acad. Sci., 271: 243-249, 1976. 63. Tokudome, S., and Kuratsune, M. A cohort study on mortality from cancer and other causes among workers at a metal refinery. Int. J. Cancer, 17:310 317,1976. 64. Decoufle, P. Further analysis of cancer mortality patterns among workers exposed to cutting oil mists. J. Natl. Cancer Inst., 61:1025-1030,1978. 65. Enterline, P. E., and Marsh, G. M. Mortality studies of smelter workers. Am. J. Indust. Med., 1:251-259,1980. 66. Blair, A. Mortality among workers in the metal polishing and plating indus try, 1951-1969. J. Occup. Med., 22:158-162, 1980. 67. Jarvholm, B., Lillienberg, L., Sallsten, G., et al. Cancer morbidity among men exposed to oil mist in the metal industry. J. Occup. Med., 23:333-337, 1981. 68. Enterline, P. E., and Marsh, G. M. Mortality among workers in a nickel refinery and alloy manufacturing plant in West Virginia. J. Natl. Cancer Inst., 68:925-933,1982. 69. Wang, J., Wegman, D. H., and Smith, T. J. Cancer risks in the optical manufacturing industry. Br. J. Indust. Med., 40:177-181,1983. 70. Dubrow, R., and Wegman, D. H. Cancer and occupation in Massachusetts. Am. J. Indust. Med., 6:207-230,1984. 71. Svensson, B. G., Englander, V., Akesson, B., et al. Deaths and tumors among workers grinding stainless steel. Am. J. Indust. Med., 15: 51-59,1989. 72. Gallagher, R. P., and Threlfall, W. J. Cancer mortality in metal workers. Can. Med. Assoc. J., 129:1191-1194,1983. 73. Vena, J., Sultz, H. A., Fiedler, R. C., et al. Mortality of workers in an automobile engine and parts manufacturing complex. Br. J. Indust. Med., 42:85-93, 1985. 74. Park, R. M., Wegman, D. H., and Silverstein, M. A. Causes of death among workers in a bearing manufacturing plant. Am. J. Indust. Med., 13: 569 580, 1988. 75. Selikoff, I. J., Churg, J., and Hammond, E. C. Asebetos exposure and neoplasia. JAMA, 188: 22-26,1964. 76. Selikoff, I. J., and Hammond, E. C. Multiple risk factors in environmental cancer. In: J. F. Fraumeni (ed.). Persons at High Risk of Cancer, pp. 467 483. New York: Academic Press, 1975. 77. Doll, R., and Peto, J. Effects on Health of Exposure to Asbestos. London: Health and Safety Commission, Her Majesty's Stationery Office, 1985. 5467 BOWEL CANCER IN YOUNG MEN BY SUBSITE 78. Edelman, D. A. Exposure to asbestos and the risk of gastrointestinal cancer a reassessment. Br. J. Indust. Med., 45:75-82, 1988. 79. Pratt, C. B., and George, S. L. Epidemic colon cancer in children and adolescents? In: B. Correa, and W. Haenszel (eds.). Epidemiology of Cancer of the Digestive Tract, pp. 127-146. Boston: Martinus Nijhoff, 1982. 80. Caldwell, G. G., Cannon, S. B., Pratt, D. B., et al. Serum pesticide levels in patients with childhood colorectal carcinoma. Cancer (Phila.), 48:774-778, 1981. 81. Johnson, L. R., Christensen, J., Jackson, M. J., et al. Physiology of the Gastrointestinal Tract. New York: Raven Press, 1987. 82. Ritchie, J. A. Movement of segmental constrictions in the human colon. Gut, 72:350-355, 1971. 83. Hertz, A. F. The passage of food along the human alimentary canal. Guy's Hosp. Rep., 61: 389-427, 1907. 84. Hertz, A. F., and Newton, A. The normal movements of the colon in man. J. Physiol. (Lond.), 47: 57-65,1913. 85. Holdstock, D. J., Misiewicz, J. J., Smith, T., et al. Propulsion (mass move ment) in the human colon and its relationship to meals and somatic activity, Gut, 11:91-99, 1970. 86. Schuster, M. M., Hendrix, T. R., and Mendeloff, A. I. The internal and sphincter response: manometric studies on its normal physiology, neural pathways, and alteration in bowel disorders. J. Clin. Invest., 42: 196-207, 1963. 87. Neutra, M. R., and Forstner, J. F. Gastrointestinal mucus: synthesis, secre tion, and function. In: L. R. Johnson (ed.). Physiology of the Gastrointestinal Tract, pp. 975-1003. New York: Raven Press, 1987. 88. Amidon, G. L. Fluid mechanics and intestinal transit. Gastroenterology, 88: 858, 1985. 5468