Document NG1gzOymQV3M1XeeGKa405aQw

Colon Cancer, Physical Activity, and Occupational Exposures A Case-Control Study MATS FREDRIKSSON, BSC,* NILS-OLOF BENGTSSON, MD,t LENNART HARDELL, MD,t AND OLAV AXELSON, MD,*$ A case-control study on colon cancer was conducted encompassing 329 cases and 658 controls. Occupations and various exposures were assessed by questionnaires. A decreased risk was found in persons with physically active occupations. This effect was most pronounced in colon descendens and sigmoideum with an odds ratio (OR) of 0.49 whereas no reduced risk was found for right-sided colon cancer. Regarding specific jobs, reduced ORs were found for agricultural, forestry, and saw mill workers and increased OR for railway employees. High-grade exposure to asbestos or to organic solvents gave a two-fold increased risk. Regarding exposure to trichloroethylene in general, a slightly increased risk was found whereas such exposure among dry cleaners gave a seven-fold increase of the risk. Cancer 63:1838-1842, 1989. C OLON CANCER was the third most common cancer among Swedish males in 1984representing 7.3%of all cancer cases. In females colon cancer was the second most common malignant disease with 8.1% of all cases. The incidence increased slowly between 1965 and 1984 with an annual change of 0.6% in males and 0.5% in females. The age standardized annual incidence rate per 100.000was 32.9 among males and 26.9 among females in 1984. It is known since before that the incidence is lower in the northern part of the country. The population underlying this case-control study lives in the catchment area of UmeB University Hospital corresponding to the three most northern counties of Sweden where the age adjusted annual incidence rate per 100.000 was 27.1 among males and 25.4 among females in 1984. The etiology of colon cancer is not well understood. Correlation studies of the incidence and mortality rates of colon cancer and food habits indicate that dietary factors may play an etiologicrole. Investigations of migrants support that concept. Dietary fat increases the secretion of bile acids and neutral steroids which may act as promoters. Another mechanism might involve carcinogens From the *Department of Occupational Medicine,UniversityHospital, Linkoping, Sweden, and ?Department of Oncology, University Hospital, Umei, Sweden. Supported by grants from the Swedish Cancer Society. 4 Professor. Address for reprints: Lennart Hardell, MD, Department of Oncology, University Hospital, S-901 85 Umea, Sweden. Accepted for publication November 15, 1988. formed by bacterial degradation. High consumption of meat may increase the risk through the formation of carcinogenic amino acid degradation products in the large bowel or carcinogens produced by frying or boiling the meat.3.4Mutagenic substancessuch as fecapentaenes and ketosteroids have been identified in extracts of the feces of approximately one third of individuals on Western diet.5,6Fibers, indoles, beta-carotene,vitamin C, selenium, and calcium in the diet have been discussed as possibly preventive factors for colon ~ a n c e r . ~ Although colorectal cancer has not generally been considered an occupational disease, heavy asbestos exposure has been suggested to increase the risk.8An increased risk for colorectal cancer has been reported in metal workers, but also among workers in yarn or textile industries and leather goods manufacture.' An association between physical activity and colon cancer was suggested by Jones and Godding." This relationship was also discussed in a study on colon cancer by Hardell," where it was noted that use of motor saws was twice as common among the controls than among the cases indicating a possible preventive effect of physical activity. In a more recently published study, men with sedentary jobs were found to have a colon cancer risk of 1.6 times that of men with jobs requiring a high level of activity. * The complementary preventive effect of phys- ical activity was strongest in the descending colon. No correlationwas seen between job activityand risk for rectal cancer. In another study,I3the overall relative risk (RR) of colon cancer in sedentary occupations was estimated 1838 No. 9 COLONCANCERA, CTIVITYAND OCCUPATION Fredriksson et al. 1839 to 1.3 with the highest risk for the transverse colon (RR = 1.6). This report presents the results of analyses relating to occupational factors only. The effects of diet, drugs, etc., will be presented in another report, but these factorshave been considered to some extent from the point of view of confounding, however. Materials and Methods Cases The cases of the study consisted of all persons aged 30 to 75 years with large bowel cancer adenocarcinoma (ICD-8code 153.01-153.89) who had been diagnosed between 1980 and 1983 as reported to the Swedish Cancer Registry, and who were residents of the admission region of the Department of Oncology in UmeH. Only persons who were alive at the time of the data collection for the study (1984-1986) were included. In total 402 cases (206 women, 196 men) were originally identified accordingto the criteria described,but 58 cases died and were therefore excluded. One case had been used as a control in a previouscase-control study and was excluded for ethical reasons, ie.,for avoiding another contact. Of the remaining 344 cases another eight were excluded due to inadequate histopathologic data and one case turned out to have a rectal adenocarcinoma having been miscoded and was therefore also excluded. Six additional cases were unable to participate for medical reasons such as older age and dementia. Eventually, the study therefore included 329 cases; 165 men and 164 women. Controls For each case two persons from the same county, of the same sex and closest in age were drawn as controls from the National Population Register. Thirty-five controls were unable to answer the questionnairedue to medical reasons. For these as well as for 24 controls who recently died who were still in the population register, the next closest control in age was used instead, so that the number of the controls was twice that of the cases or 658. Assessment of Exposure An 18-page questionnaire was mailed to all included cases and controls. In the first part of the questionnaire occupationalhistories until retirement were mapped. Details on specificoccupationsand exposureswere also collected. Questions on food and drinking habits were included in the second part of the questionnaire, as well as previous diseases and drug intake. If necessary, data were supplemented over the phone, i.e., when the questions TABLE1. Anatomical Localizations of the 312 Cases of Colon Cancer (ICD 8) Localization ICD 8 No. (%) Appendix Cecum Ascendens/ileocecum Transversum/flexura Descendens Sigmoideum Multiple 153.01 153.02 153.03 153.11 153.21 153.31 153.87 9 (2.9) 77 (24.7) 37 (11.9) 57 (18.3) 25 (8.0) 105 (33.6) 2 (0.6) were not completely answered or apparently misunderstood. Jobs were scored for physical activity, I11 for high level of activity,I for sedentaryjobs, and I1 for the intermediate group. Housewives were included in category I1 and pretime retired persons in category I. Employment years in the three categories were assessed for cases and controls (Table 2). In a further analysis the number of the employment years in each categoryfor each individualwere multiplied by a score taken as 1,2, and 3 so as to obtain an exposure index reflecting the duration and intensity of physical activity. Then, in the calculations of odds ratios (OR) the individuals were grouped according to their indices into the categories0 to 66,67 to 132, and 133to 198,labelled A, B, and C, respectively (Table 3). Occupational exposures to asbestos, impregnating agents, and organic solvents were independently classified as high grade or low grade by three persons well experienced in occupation health: two physicians and one hygienist. Statistical Methods The statistical analyses of the data were based on Mantel-Haenszel methods.l4 The confidence intervals were obtained with the approximative procedure as described by Miettinen.'' Years of employment were also compared between cases and controlsin terms of a formal OR (Table 2) not to be confused with the traditional epidemiologic OR concept based on numbers of cases and controls with and without exposureand as used elsewhere in this report. Results Of the 329 finally included cases, 17 (5.2%)refused to participate leaving 156 male and 156 female cases in the study. Of the 658 controls included, 35 (5.3%)were lost for the same reason, whereas 306 male and 317 female controls answered the questionnaire. The mean age of cases as well as controls was 63 years. The anatomical localizations are presented in Table 1. 1840 CANCERMay 1 1989 Vol. 63 Cases Controls OR X2 TABL2E. Employment Years and Their Odds Ratios in the Three Categories of Physical Activity I11 I1 M F Total M F Total M 8670 19848 0.82 27.3 1671 4480 0.68 141.4 10341 24328 0.78 186.2 6197 11613 1.oo 0.0 1 4599 8842 0.95 4.9 10796 20455 0.97 3.6 1134 2118 ( 1-.O) I F 651 1 11869 (1-.0) Total 7645 13987 (1.0) - M: males; F females; OR: odds ratios. 111: physically active jobs; I: sedentary jobs; 11: intermediate group; category I taken as the reference. The distribution of employment years in the three categories of physical activity are presented in Table 2. A significantlydecreased odds ratio in physically activejobs (category111)was found both for males and females, i.e., an OR of 0.82 and 0.68 with a chi-square value of 27.3 and 141.4, respectively. However, by using the score index categories for the total material of cases and controls, only a relatively weak preventive effect was found forjob-related physical activity. Consideringthe anatomical localizationsof the cancers (Table 3) a right-sided to left-sided decrease of the risk was noted among physically active persons with an OR for cancer in colon descendens and sigmoideum of 0.49 with a 95% confidence interval (C19J of 0.25 to 0.93. A corresponding but weaker effect was also seen in the intermediate group of physical activity (OR = 0.73, C195 = 0.48-1.2). Some further analyses considered occupation and colon cancer. Since ulcerative colitis is a well-defined risk factor for colon cancer and an aggregation of such cases might be suspected to occur in some occupations, the ten cases and two controls who had reported that disease in the questionnaire were excluded in the further analyses. The material was then stratified for sex, age, and score index categories for physical activity. Table 4 presents OR for various occupations. An increased OR was seen for rail- TABL3E. Odds ratios and Confidence Intervals for Different Localizations of Colon Cancer by Categories of Physical Activity in the Whole Material OR (Cbs) AB C 153.01, 02, 03 153.11 153.21, 31 (1.0) (1.0) (1.0) 0.92 (0.58-1.5) 0.82 (0.45-1.5) 0.73 (0.48-1.2) 1.22 (0.63-2.4) 0.66 (0.28-1.6) 0.49 (0.25-0.93) A: sedentary jobs; C physical active jobs; B: intermediate group; 153.01: appendix; 153.02: cecum; 153.03: ascendens; 153.11: flexura, transversum; 153.21: descendens; 153.31: sigmoideum; OR: odds ratios; CI9,: 95% confidence intervals. way employees, in men 2.4, (C195 = 1.1-5.1) and in women 2.3, (CIg5= 0.4-1 1.6). For male painters a risk of 3.0 was obtained (C195= 0.9-9.2). No clearly elevated and/or consistent risks were found for other occupations (Table4). For agricultural,forestry, and saw mill workers there were reduced OR in men, i.e.,0.7 (CIg5= 0.4-1.0), TABLE4. Number of Exposed Male and Female Cases and Controls in Various Occupations When Exceeding Five* Exposed Occupation M/F Case/Control MHOR c195 Cabinet maker Chemical workert C1eaner Construction worker Dress maker Dry cleaner Farmer Gardener Ladies' hairdresser Lumberer Mechanic Metal worker Miner Motor car driver Motor mechanic Needle worker Nurse Painter Paper mill worker Pulp worker Railway worker Road worker Saw mill worker M M F F M F F M F M F F M M M M M M F F M M M M F M M 4/10 8/23 11/12 29/50 26/64 5/15 515 501131 47/112 619 717 315 641160 6/16 10/23 6/17 29/47 11/17 614 316 715 216 10126 16/14 313 18/48 19/68 0.8 0.3-2.6 0.7 0.3-1.5 2.0 0.9-4.7 I .2 0.7-2.1 0.8 0.5-1.3 0.6 0.2- 1.8 2.0 0.5-7. I 0.7 0.4-1.0 0.8 0.5-1.2 1.3 0.5-3.5 2.5 0.8-7.9 1.1 0.3-4.6 0.7 0.4- I .O 0.8 0.3-2.0 0.9 0.4-1.9 0.7 0.3-1.9 1.2 0.7-2.1 1.3 0.6-2.8 3.0 0.9-10.6 1.o 0.5-2.1 3.0 0.9-9.2 0.5 0.1-2.8 0.7 0.3-1.6 2.4 1.1-5.1 2.3 0.4-1 1.6 0.7 0.4- 1.3 0.5 0.3-0.9 * Mantel-Haenszel odds ratios (MHOR) and 95% confidence intervals (CIgS)for colon cancer by stratification on age and physical activity are given for each sex. t Unspecified. -No. 9 COLONCANCERA, CTIVITYAND OCCUPATION Fredriksson et al. 1841 0.7 (C195= 0.4-1.0), and 0.5 (CIg5= 0.3-0.9), respectively. For women there was a decreased OR for farmers (0.8, c195 = 0.5-1.2). Risk ratios with regard to specific exposures are listed in Table 5. No increased risk was found for exposure to pesticides such as phenoxy acetic acids, dichloro-diphenyltrichloroethane (DDT), or impregnating agents. Highgrade exposure to organic solvents gave an OR of 2.1 (C195= 0.8-5.8). Also for high-grade exposure to asbestos a somewhat increased OR was found (2.1, cI95 = 0.85.8). The OR for exposure to man-made fibers, that is glasswool and rockwool, were less than unity. A slightly increased risk was found for exposure to trichloroethylene in general, whereas such exposure among dry cleaners gave an OR of 7.4 (cI95 = 1.1-47.0). Discussion This investigation showed a decreased risk for colon cancer in jobs classified as physically active. There was a right-sided to left-sided trend with the highest prevention for cancer in the descendingcolon and sigmoideum. This is in agreement with some of the previous findings. After stratification for physical activity, farming, forest work, and saw mill work were still associated with a reduced risk for colon cancer in men, thereby indicating the possibility of other preventive factors in these individuals. On the other hand, physically activejobs with high-grade exposure to agents like organic solvents or asbestos might not show a reduced risk because of some effect of such agents. It has been postulated by Burkitt et a1.I6 that increased bowel activity is preventive for colon cancer. This may be the reason for the decreased risk in persons with physically active jobs, but there could also be other preventive factors associated with physical activity that exerts the preventive effect. A significantlyincreased risk for colon cancer was found among persons employed by the Swedish railways. This is a somewhat surprising finding for which there is no particular explanation and it should be noted that the elevated risk for this employment category persists also after stratification for degree of physical activity. In view of the current interest for electromagnetic fields as a risk factor for cancer, it may be mentioned that the Swedish railroads are electrified to a great extent, presumably causing considerable exposure to low frequency electromagnetic fields for many of the employees. Exposure to organic solvents has previously not been described as a risk factor for colon cancer. An especially high risk was found in this study among dry cleaners exposed to trichloroethyelene. Although this finding was surprising, this substance is known to be mutagenic." There is also a suggestion of pneumocystosis of the colon TABLE5. Odds Ratios and 95% Confidence Intervals for Exposure to Various Agents* Agents OR cI95 Asbestos High grade Low grade Asphalt? Glasswoolt Hexachlorophene Impregnating agents High grade Low grade Organic solvents High grade Low grade Thinners Trichloroethylene Trichloroethylene among dry cleaners White spirit Pesticides DDT Phenoxyacetic acids Rockwoolt Smoker (including exsmokers) Snuff 2.1 1.2 1.7 0.6 0.6 0.9 0.7 2. I I .3 1.2 1.5 1.4 1.1 1.6 1.3 1.3 0.7 1.o 1.5 0.8-5.8 0.6-2.4 0.3-1 1.7 0.3-1.1 0.1-2.0 0.2-3.5 0.3-1.3 0.8-5.8 0.8-2.0 0.6-2.4 0.4-5.7 1.1-47.0 0.5-2.3 0.8-3.0 0.8-2.1 0.7-2.4 0.4-1.3 0.7- 1.3 0.9-2.5 OR: odds ratios; CIg5:95% confidence intervals; DDT: dichloro-diphenyl-trichloroethane. * Cases and controls were stratified for age, sex, and scoring index of physical activity. t Only men. as possibly relating to trichloroethylene exposure, which to some extent could suggest the colon to be a potential target organ for effects of this chlorinated ethylene.18Further studies are required before any definitiveconclusions can be made. Exposure to asbestos has in some, but not all, investigations been associated with an increased risk for colorectal cancer.8 In this study a somewhat increased risk was seen for persons with high-grade exposure although with a confidence interval including unity. Since colon cancer is more common in higher age groups the controls were originally drawn as similar to the cases with regard to.age. This allowed for an equal number of employment years within various occupations for cases and controls and age was taken care of by stratification in the analyses. Furthermore, only alive subjects were included in order to avoid information bias due to second hand information from relatives. This is of special importance regarding food habits, previous diseases, and drug intake as also assessed in the study and as not found to correlate with particular occupations and exposure to such a degree as likely to explain the observations made with regard to the deviating risk patterns in relation to various jobs or exposures. The relationship of colon cancer, eating habits, previous diseases, and the use of various drugs will be discussed in forthcoming publications. 1842 CANCERMay I 1989 Vol. 63 REFERENCES I. Aries V, Crowther JS, Drasar BS, Hill MJ, Williams RE. Bacteria and the aetiology of cancer of the large bowel. Gut 1969; 10:334-335. 2. Hill MJ. Bacteria and the etiology of colonic cancer. Cancer 1974; 34:815-818. 3. Cummings JH, Hill MJ, Bone ES, Branch WJ, Jenkins DJ. The effect of meat protein and dietary fiber on colonic function and metabolism: 11. Bacterial metabolism in feces and urine. Am J CIinNutr 1979; 3212094-2101. 4. Weisburger JH, Reddy BS, Springarn NE, Wynder EL. Current views on the mechanisms in the etiologyof colorectal cancer. In: Winaver SJ, Schottenfeld D, Sherlock P, eds. Colorectal Cancer: Prevention, Epidemiology and Screening. New York Raven Press, 1980; 19-4 1. 5. Gupta I, Suzuki K, Bruce WR, Yates P, Krepinsky JJ. A model study of fecapentaens: Mutagens of bacterial origin with alkylating p r o p erties. Science 1984;225521423. 6. Bruce WR. What chemicals are responsible for colon cancer? J Cell Phys [Suppl] 1986;4:47-49. 7. Ziegler RG, Deves SS, Fraumeni J F Jr. Epidemiologic patterns of colorectal cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Important Advancesin Oncology. Philadelphia: JB Lippincot, 1986;209232. 8. Overall Evaluations of Carcinogenicity: An updating of IARC MonographsVolumes 1-42. In: IARC Monographs,supplement 7. Lyon: IARC, 1987; 108. 9. Schottenfeld D, Winawer SJ. Large intestine. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. Philadelphia: WB Saunders, 1982; 703-727. 10. Jones FA, Godding EW. Management of Constipation. London: Blackwell, 1972; 156-158. 1I. Hardell L. Relation of soft-tissue sarcoma, malignant lymphoma and colon cancer to phenoxy acids, chlorophenols and other agents. Scand J Work Environ Health 1981;7: 119- 130. 12. Garabrant DH, Peters JM, Mack TM, Bemstein L. Job activity and colon cancer risk. Am J Epidemiol 1984; 119:1005-1014. 13. Gerhardsson M, Norell SE, Kiviranta H, Pedersen NL, Ahlbom A. Sedentary jobs and colon cancer. Am J Epidemiol 1986; 125:775- 780. 14. Mantel N, Haenszel W. Statistical aspect of analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 23:719-748. 15. Miettinen 0s.Estimability and estimation in case-referent studies. Am J Epidemiol 1976; 103:226-235. 16. Burkitt DP, Walker ARP, Painter NS. Effect of dietary fiber on stools and transit times, and its roles in the causation of disease. Lancet 1972; 2: 1408-1412. 17. Greim H, Bonne G ,Radwan Z. Mutagenicityin vitroand potential carcinogenicityof chlorinatedethylenesasa function of metabolicoxirane formation. Biochem Pharmacol 1975; 24:2013. 18. Yamaguchi K, ShiraiT, Shimakura K efa/.Pneumatosis cystoides intestinalis and trichloroethylene exposure. Am J Gastroenterol 1985; 80:753-757.