Document YjroxRqpXK767mvzQgRJ0qbeV

441 Incidence of Colorectal Adenocarcinoma by Anatomic Subsite An Epidemiologic Study of Time Trends and Racial Differences in the Detroit, Michigan Area Raymond Y. Demers, m.d.1 Richard K. Severson, Ph.D.2 David Schottenfeld, m.d.3 Lisa Lazar, m.p.h.4 1 Cancer Center, Henry Ford Health System, De troit, Michigan. 2 Division of Epidemiology, Karmanos Cancer Institute; Department of Family Medicine, Wayne State University School of Medicine, De troit, Michigan. 3 Departments of Epidemiology and Internal Medicine, University of Michigan, School of Public Health and Medical School, Ann Arbor, Michigan. 4 University of Michigan, School of Public Health, Ann Arbor, Michigan. Supported by the National Cancer Institute Con tract N01-CN-05225. Address for reprints: Raymond Y. Demers, M.D., Director, Cancer Center, Henry Ford Health System, One Ford Place, Detroit, MI 48202. Received April 4, 1996; revisions received Au gust 1, 1996, and October 10, 1996; accepted October 10, 1996. 1997 American Cancer Society BACKGROUND. Colorectal adenocarcinoma may represent more than one disease process. Numerous epidemiologic studies suggest that rates of occurrence of colo rectal adenocarcinoma at particular anatomic subsites (e.g., right colon, left colon, and rectum) may be associated with distinctive geographic, demographic, and risk factor profiles. This study explored time trends over a 22-year period of the inci dence of adenocarcinoma of the colon and rectum at various subsites among patients of different race, gender, and stage of disease. methods. Data on the incidence of colorectal adenocarcinoma were obtained from a population-based cancer registry in the Detroit, Michigan area funded by the National Cancer Institute. Age-adjusted incidence rates were analyzed by year of diagnosis. Relative survival rates were also obtained for different race and gender categories, along with disease stage at diagnosis. results. A major rise was revealed in the incidence of adenocarcinoma in the right colon among African American men and women between the mid-1970s and the early 1980s. The rise was greatest among African American men and accounts for increases in late stage disease among them. Corresponding decreases in survival among African American men were noted. CONCLUSIONS. These findings indicated widely differing disease patterns based on anatomic subsite and patient demography and also indicated a need for targeted efforts at early detection of adenocarcinoma of the right colon among African Americans. Cancer 1997;79:441-7. 1997 American Cancer Society. KEYWORDS: colorectal, adenocarcinoma, race, subsite, trends. olorectal cancer among men and women in the U. S. is the third Cmost commonly diagnosed cancer and the second leading cause of cancer mortality. The American Cancer Society estimates that there will be 133,500 new cases of colorectal cancer in the U.S. in 1996, and 54,900 deaths due to this disease.1 Typically defined as a malignant neoplasm involving an organ extending from the cecum to the anus, cancer of the large intestine is receiving major research attention because of more coherent hypotheses regarding genetic and environ mental risk factors, and the recent publication of new screening guidelines.2 Research on the etiology of colorectal cancer has focused on dietary factors, specifically high fat and low fiber consumption, patterns of preferences for fruits and vegetables, vitamin supplemen tation, level of regular physical activity, and family history.3-8 Possible occupational risk factors include asbestos exposure and work in pat tern and model making.9,10 Risk factors vary in relation to race and ethnicity, country of birth and migration, and anatomic location in 442 CANCER February 1, 1997 / Volume 79 / Number 3 the large intestine.3 Increased knowledge of changing incidence trends may provide clues to further enhance understanding of the etiology of colorectal cancer. The potential for screening for colorectal cancer is largely dependent on the anatomic location of the premalignant or early phase of the tumor. The recent evolution of fiberoptic technology allows endoscopy to assist in direct viewing of the entire surface of this organ. Coupled with the knowledge that most colo rectal cancers originate as polypoid growths on the surface of the colon and rectum, endoscopy theoreti cally allows for visualization of adenomas and early invasive cancers. Once visualized, polyps can be re moved to prevent subsequent neoplastic disease.11-12 The revised U.S. Preventive Services Task Force guide lines suggest that the risk of colorectal cancer mortality in the general population can be reduced by annual fecal occult blood testing in conjunction with periodic flexible sigmoidoscopy in persons older than 50 years.13-17 The need to understand patterns of tumor location within the colon and rectum is underscored by the fact that screening endoscopy (flexible sigmoid oscopy) provides visualization of only the distal 60 cm of the colorectal surface, extending approximately to the splenic flexure. The evolving experimental and epi demiologic evidence regarding causal mechanisms and the potential efficacy of screening heighten the importance of developing additional strategies for pre vention of colorectal cancer. Descriptive population-based epidemiologic stud ies may provide insights into dynamic demographic patterns and their relationship to the natural history of colorectal cancer. Therefore, this study was under taken to analyze the interrelations of anatomic tumor location, race, gender, diagnostic stage, and time trend data for colon and rectal cancer in the Detroit Metro politan area from 1973-1994. METHODS Data were obtained from the Metropolitan Detroit Cancer Surveillance System, which represents one of the National Cancer Institute's Surveillance, Epidemi ology and End Results (SEER) Program sites. These data are population-based, encompassing the metro politan Detroit area of Wayne, Oakland, and Macomb counties. These counties have a combined 1990 popu lation of 3.9 million people, of whom 23.9% are African American. The study was limited to African Americans and whites with histologically confirmed first primary in vasive adenocarcinomas of the colon or rectum. Diag nosis of these adenocarcinomas was made inclusive of calendar years 1973-1994. Histologic types were limited to noncarcinoid adenocarcinomas, because these constitute the vast majority of colorectal malignancies.18 In addition to analyses by race, gender, and subsite of origin, the influence of stage of disease was also evaluated. Stage is defined according to SEER guidelines as local when the cancer is confined to the mucosa, submucosa, muscularis, or subserosa;regional when the cancer spreads to adjacent lymph nodes, adjacent organs, or the omentum; and distant when the cancer involves distant lymph nodes or other metastases. The subsite of anatomic origin constituted a key analytic variable and was grouped as the right colon, left colon, or rectum. The right colon subsite included the cecum, ascending colon, hepatic flexure, trans verse colon, and splenic flexure. The descending colon and sigmoid represented the left colon subsite. The rectum subsite included the rectum and rectosigmoid junction, with the latter location defined as the perito neal reflection, approximately 15 cm from the anal orifice. Incidence rates were age-adjusted to the 1970 U.S. standard population, and plotted for the colon, rec tum, and subsites of the colon and rectum, and by stage, gender, and race. Time trends were plotted us ing 3-year moving averages. Tests for linear trend in the age-adjusted rates were computed by simple linear regression.19 In addition, 5-year relative survival rates were determined by year of diagnosis (1973-1988) for adenocarcinomas in the right colon, left colon, or rec tum. RESULTS All cases among African Americans and whites from 1973-1994 are summarized in Table 1. The proportion of colorectal malignancies in the right colon was 30.8% amongwhite males, compared with41.1% amongAfrican American males. Among females, a greater pro portion of right colon adenocarcinomas was observed among African American females (44.9%) compared with white females (38.1%). The percentage of adeno carcinomas in the rectum and rectosigmoid junction was 37.4% in white males, compared with 27.5% among African American males. Similar racial differ ences were observed for rectal adenocarcinomas among white (31.0%) and African American (24.0%) females. Almost identical proportions were described for primary adenocarcinomas in the left colon in white and African American males and females. Annual age-adjusted incidence rates for adenocar cinomas of the colon and rectum combined between 1973 and 1994 are shown in Figure 1. Although the incidence rates appeared to increase for all four race/ gender groups from 1973-1988, increases were more evident among African Americans. Testing for trends Incidence of Colorectal Adenocarcinoma by Anatomic Subsite/Demers et al. 443 TABLE 1 Adenocarcinoma of the Colon and Rectum by Subsite, Race, and Gender Detroit Metropolitan Area, 1973-1994 Cases African American White males White females males Site n Rates" n Ratesa n Ratesa Colon Right Cecum Ascending Hepatic flexure Transverse Splenic flexure Left Descending Sigmoid Rectosigmoid junction and rectum Rectosigmoid junction Rectum Total colorectum 8704 4282 1825 1069 362 716 310 4422 802 3620 5159 1890 3269 13863 27.9 13.8 5.9 3.5 1.2 2.3 1.0 14.1 2.6 11.5 16.3 6.0 10.3 44.2 CO CO CO 8652 4793 2107 1112 901 285 3859 694 3165 3870 1481 2389 12,522 19.7 10.7 4.6 2.5 0.9 2.0 0.7 9.0 1.6 7.4 9.1 3.5 5.6 28.8 2016 28.9 1150 16.6 463 6.7 259 3.7 97 1.4 187 2.7 144 2.1 866 12.4 208 3.0 658 9.4 776 10.9 238 3.4 538 7.5 2792 39.8 a Rates are per 100,000 and are age-adjusted to the 1970 U. S. standard population. African American females n Ratesa 2290 24.3 1355 14.3 573 6.1 293 3.1 122 1.3 241 2.6 126 1.3 935 10.0 250 2.7 685 7.3 717 7.7 289 3.1 428 4.6 3007 32.0 FIGURE 1. Combined annual age-adjusted rates for adenocarcinomas of the colon and rectum between 1973 and 1994. FIGURE 2. Time trends for right colon adenocarcinoma incidence rates. showed statistically significant increases for white males (P< 0.001), African American males (P< 0.001), white females (P= 0.048), and African American females (P= 0.012). These trends showed a reversal (a decrease in incidence rates) from 1989-1994 for white males (P= 0.008), African American males (P= 0.009), and white females (P= 0.010), whereas rates for African American females have remained stable in recent years (P= 0.444). Figure 2 shows time trends for right colon adenocarcinoma incidence rates. Although right 444 CANCER February 1, 1997 / Volume 79 / Number 3 FIGURE 3. Time trends for left colon adenocarcinoma incidence rates. FIGURE 4. Incidence rates for rectal adenocarcinoma among white men compared with African American men and women of both races. colon adenocarcinoma rates remained relatively stable among both white males and females, these rates in creased notably in African American men from 1975 through 1983 and in African American women from 1975 through 1978. Rates for African American women surpassed those for white women in 1976 and re mained consistently higher through 1994. Although right colon adenocarcinoma rates were equal for both white and African American men in the early 1970s, the rates in African American men exceeded those in white men during the period 1976-1994. Figure 3 shows time trends for left colon adenocarcinoma inci dence rates. These rates remained essentially un changed among white men and women and African American women through the mid-1980s. Conversely, rates among African American men increased steadily from 1973 through 1980. Rates of left colon adenocar cinoma showed steady declines for African American and white men and women beginning in the mid to late l980s. Incidence rates of rectal adenocarcinoma reflected a persistent excess among white men com pared with African American men and women of both races with no obvious shift in the trends by race and gender (Fig. 4). The most striking temporal patterns appeared when racial trends for right colon adenocarcinoma were considered by stage at diagnosis (Fig. 5). Rates for distant stage disease for the right colon among African American men began to rise sharply in 1980, and were approximately double those of white men from 1981 through 1988. However, since 1988, distant stage right colon adenocarcinoma among African American men has been decreasing. A less dramatic, more gradual increase was observed among African American women, with this increase peaking in 1980 and reaching a plateau thereafter. Distant stage, right colon adenocarcinoma for white men and women ap peared to be stable from 1973-1994. With the consideration that a significant increase in the incidence of distant stage disease in the right colon might be associated with decreasing survival, 5year relative survival rates were computed for African American males and compared with that for white males diagnosed from 1973-1989 (Table 2). Males were selected for this analyses because of the observed increase in right colon, late stage disease noted among African American males (Fig. 5). Relative survival at 5 years for male patients with right colon adenocarci noma was approximately equal among both African Americans and whites diagnosed in 1973-1977 (49 50%) and rose to 61.3% for white males diagnosed in 1984-1989 (P= 0.005), although dropping to 38.5% for African American (P= 0.140). In contrast to that ob served for right colon adenocarcinoma among African Incidence of Colorectal Adenocarcinoma by Anatomic Subsite/Demers et al. 445 FIGURE 5. Racial trends for right colon adenocarcinoma considered by stage at diagnosis. American men, the trend for 5-year relative survival percentages improved from 39.8% in 1973-1977 to 58.4% in 1984-1989 for African American men with left colon and rectal adenocarcinoma, although the rates were generally improving, particularly after diag nosis in 1978-1983 (P= 0.002), as they were for white males from 1973-1977 (49.3%) to 1984-1989 (64.4%) (P< 0.001). DISCUSSION Results from this study revealed that there are differ ences in colorectal adenocarcinoma incidence rates among whites and African Americans. These results are not obvious in U.S. national data, which suggest generally equal and stable rates among African Ameri cans and whites.20 In the Detroit SEER data, an in crease in colorectal adenocarcinoma incidence was observed among African Americans and was primarily accounted for by increasing rates of right colon adeno carcinoma. Research findings indicate an increase in colorectal adenocarcinoma among African Americans, which is accounted for by increases in right colon ade nocarcinoma that was previously undocumented. Al though the rapid increase observed in rates of right colon adenocarcinoma among African Americans be ginning in the mid-1970s and extending to the mid- 1980s has subsequently shown a decrease among Afri can American men and level rates among African American women, incidence rates among African Americans continued to exceed those among white men and women. There are perplexing racial differences in the com parisons of tumors located in the proximal and distal segments of the colon and in the rectum that have been noted in other epidemiologic studies. In the U.S., Devesa et al. studied 120,000 colon cancer cases from 1976 through 1987 using data from the SEER program, and reported higher rates in African Americans com pared with whites for cancer of the transverse colon, splenic flexure, and descending colon.21 They advised further study using colorectal cancer subsite. Johnson et al. showed a greater proportion of proximal tumors among African Americans relative to whites.22 Similar epidemiologic findings were later published by Thomas et al.23 and reinforced by the findings of Offerhaus et al. in an ``epidemiologic necropsy'' study, which described the distribution of adenomatous pol yps concentrated in the proximal colon among African Americans.24 Elixhauser et al. reviewed the clinical re cords of 188,109 patients from the Hospital Discharge Survey and showed excessive rates per 1000 hospital discharges of colorectal cancer subsites among whites compared with African Americans. African Americans showed higher rates of ``unspecified'' subsite locations.25 What rationale is there for considering the hypoth esis that proximal and distal subsites of colorectal ade nocarcinoma may reflect distinctive pathogenic mech anisms? Subsites of the colon and rectum originate from the midgut and hind gut, respectively.26 Fecal transit time throughout the large intestine tends to be slower in the proximal than in the distal colon, allowing for differential contact time for the fecal stream with colonic and rectal mucosa.27 Animal fat consumption and body size are suggested risk factors for tumor development in the proximal colon.28,29 A literature review by Bufill summarized genetic and other biologic characteristics that may explain dif fering susceptibilities to neoplastic transformation be tween the proximal and distal colon.30 McMichael and Potter reviewed the biologic significance of bile acid metabolism and cholecystectomy in relation to the pathogenesis of carcinomas in the proximal colon.31 Offerhaus studied genetic alterations by tumor loca tion and found evidence to suggest that diploid can cers are found more frequently in the right colon, and 17p deletions are more common with left-sided malignancies.32 Overall, approximately 35.6% of colorectal adeno carcinomas among residents of the Detroit metropoli- 446 CANCER February 1, 1997 / Volume 79 / Number 3 TABLE 2 Five-Year Relative Survival for Right Colon and Left Colon/Rectum Subsites: Males by Race, 1973-1989 Year of diagnosis Right colon African Americans Whites Left colon and rectum African Americans Whites 1973-1977 1978-1983 1984-1989 48.6% 48.6% 38.5% 50.0% 57.0% 61.3% 39.8% 40.9% 58.4% 49.3% 56.9% 64.4% tan area are located in the right colon. Beginning in the early to mid-1970s, right-sided colon adenocarcinoma rates among African American men and women accel erated and exceeded the rates in whites, and were most evident among African American men with dis tant stage disease. These findings have major public health implications, especially because the rates of late stage disease for proximal colon adenocarcinomas greatly increased among African Americans beginning approximately in 1980 and were associated with de creased survival. Similarly, a decrease in distant stage right colon adenocarcinoma since 1988 among African American men warrants explanation. Advances in screening and early detection tech nologies may account for at least some of the changes in temporal trends for right-sided colon adenocarci noma incidence in whites and African Americans. However, this is unlikely to be the major explanation because flexible endoscopic screening usually focuses on the distal segment of the large intestine. Flexible sigmoidoscopy is currently recommended as a screen ing procedure for men and women at normal risk be ginning at age 50, but typically does not extend beyond 60 cm with maximal reach to the splenic flexure, po tentially encompassing 40-65% of colorectal adeno carcinomas. Although stool occult blood testing is cur rently recommended as a method of screening, it has limited sensitivity in detecting adenomas and early invasive cancers.33 Based on surveys conducted by the U.S. Centers for Disease Control and Prevention, utili zation of screening methods for colorectal adenocarci noma by whites exceeded the frequency reported dur ing an interval of 5 years by African Americans.34 Coding changes in the SEER program must also be considered when such marked changes in inci dence rates are observed. A careful review of historic procedures in the SEER program revealed no coding changes in registry data that would account for the findings identified in this study. In addition, any cod ing changes would likely affect incidence rates among all four race and gender subgroups. In the current study data, significant rate changes were observed only in African Americans. Race specific changes in environmental exposures must also be considered. However, the introduction of a putative causal factor would be difficult to identify historically and may be associated with an inductionlatency period of 10 to 30 years.35 Therefore, a signifi cant alteration in environmental exposure or lifestyle practices among African Americans that accounted for the increase in right colon adenocarcinoma incidence during the 1970s and 1980s would have been intro duced and sustained in the 1940s, 1950s, or 1960s. The search in population-based studies for such exposures or practices as occupational/environmental exposure, dietary fat, sedentary lifestyle, or use of tobacco and alcohol would provide the basis for future epidemio logic studies. As with many cancer registry-based studies, the current study has certain limitations. For example, data are limited to those that can be ascertained through existing hospital or pathology records. There fore, information on risk factors, comorbidity, or other personal data are not available. Also, it is not possible to identify the method of diagnosis or to capture changing patterns of diagnosis or early detection over time. This may be a significant limitation, because dif fering patterns of colorectal adenocarcinoma detec tion may, at least in part, explain racial and gender differences in the incidence of this malignancy. A principal strength is that this is a populationbased study, utilizing SEER data, with case ascertain ment approaching 100%. An additional strength is the fact that the cancer registry utilized for the current study has been in existence since 1973, with an accu mulation of 22 years of colorectal cancer cases docu mented through 1994. Also, the Detroit SEER program has within its geographic boundaries a high propor tion (approximately 24%) of African Americans. Future directions for research must be considered in light of the findings of the current study. Epidemio logic studies of colorectal adenocarcinoma should fo cus on anatomic subsites of colorectal disease. Spe cifically, studies emphasizing differences among races, by subsite of disease, pathology by stage at diagnosis, and other prognostic biomarkers may facilitate the identification of pathogenic mechanisms intrinsic to proximal compared with distal site neoplasia. These Incidence of Colorectal Adenocarcinoma by Anatomic Subsite/Demers et al. 447 studies should include biologic markers of exposure and susceptibility that are linked to environmental agents. The importance of needed epidemiologic, ge netic, and molecular studies that attempt to compare and contrast causal mechanisms for right-sided and left-sided colon adenocarcinoma is underscored by the observation that the incidence of right-sided colo rectal adenocarcinoma has risen significantly, particu larly among African American men and women. REFERENCES 1. Parker SL, Tong T, Bolden S, Wingo PA. Cancer Statistics 96. CA Cancer J Clin. 1996;46:5-27. 2. U.S. Preventive Services Task Force. Guide to clinical pre ventive services. Report of the U.S. Preventive Services Task Force. 2nd edition. Baltimore: Williams and Wilkins, 1996: 89-103. 3. West DW, Slattery ML, Robison LM, Schuman KL, Ford MH, Maloney AW, et al. Dietary intake and colon cancer: sex and anatomic site-specific associations. Am J Epidemiol 1989; 130:883-94. 4. Willett WC, Stampher MJ, Colditz GA, Rosner BA, Speizer FE. Relation of meat fat and fiber intake to the risk of colon cancer in a prospective study among women. N Engl J Med 1990;323:1664-72. 5. Peters RK, Pike MC, Garabrant D, Mack TM. Diet and colon cancer in Los Angeles County, California. Cancer Causes Control 1992;3:457-73. 6. Lee IM, Paffenbarger RS, Hsieh C. Physical activity and risk of developing colorectal cancer among college alumni. JNatl Cancer Inst 1991;83:1324-9. 7. Demers RY, Burns PB, Swanson GM. Construction occupa tions, asbestos exposure, and cancer of the colon and rec tum. J Occup Med 1994;36:1027-31. 8. Severson RK, Nomura A, Grove JS, Stemmermann GN. A prospective analysis of physical activity and cancer. Am J Epidemiol 1989; 130:522-9. 9. Gerhardsson S, le Verdier M, Plato N, Steineck G, Peters J. Occupational exposures and cancer of the colon and rectum. Am JIndMed 1992;22:291-303. 10. Swanson GM, Belle SH, Burrows RW Jr. Colon cancer inci dence among modelmakers and patternmakers in the auto mobile manufacturing industry. J Occup Med 1985;27:5679. 11. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 1992;326:658-62. 12. Winawer SJ, Zauber AG, Ho MN, O'Brien MJ, Gottlieb LS, Sternberg SS, et al. Prevention of colorectal cancer by colon oscopic polypectomy. N Engl J Med 1993;329:1977-81. 13. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, et al. Reducing mortality from colorectal can cer by screening for fecal occult blood. N Engl J Med 1993;328:1365-71. 14. Kewenter J, Bjork S, Haglind E, Smith L, Svanvik J, Ahren C. Screening and rescreening for colorectal cancer: a controlled trial of fecal occult blood testing in 27,700 subjects. Cancer 1988;62:645-51. 15. Selby JV, Friedman GD, Quesenberry JR, Weiss NS. A casecontrol study of screening, sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326:653-7. 16. Ahlquist DA, Wieand HS, Moertel CG, McGill DB, Loprinzi ZL, O'Connell MJ, et al. Accuracy of fecal occult blood screening for colorectal neoplasia: a prospect study using Hemoccult and HemoQuant tests. JAMA 1993;269:1262-7. 17. Demers RY, Parsons K. Colorectal cancer incidence in pat tern and modelmakers: evidence from a screening program. Am JInd Med 1994;26:33-45. 18. DiSario JA, Burt RW, Kendrick ML, McWhorter WP. Colo rectal cancers of rare histologic types compared with adeno carcinoma. Dis Colon Rectum 1994;37:1277-80. 19. Armitage P, Berry G. Statistical methods in medical research. Boston: Blackwell Scientific Publications, 1994:154-74. 20. Kosary CL, Ries LAG, Miller BA, Hankey BF, Harras A, Ed wards BK, editors. SEER Cancer Statistics Review, 1973 1992: Tables and Graphs. Bethesda, (MD): National Cancer Institute; 1995 NIH Pub. No. 96-2789. 21. Devesa SS, Chow W-H. Variation in colorectal cancer inci dence in the United State by subsite or origin. Cancer 1993;71:3819-26. 22. Johnson H Jr., Carstens R. Anatomical distribution of colonic carcinomas: interracial differences in a community hospital population. Cancer 1986;58:997-1000. 23. Thomas CR Jr., Jarosz R, Evans N. Racial differences in the anatomical distribution of colon cancer. Arch Surg 1992;127:1241-5. 24. Offerhaus GJ, Giardiello FM, Tersmette KWF, Mulder J-WR, Tersmette AC, Moore GW, et al. Ethnic differences in the anatomical location of colorectal adenomatous polyps. Int J Cancer 1991;49:641-4. 25. Elixhauser A, Ball JK. Black/white differences in colorectal tumor location in a national sample of hospitals. J Natl Med Assoc 1994;86:449-58. 26. Langman J. Digestive System. Medical embryology. 5th edi tion. Baltimore: Williams & Wilkins, 1985:235-43. 27. Ritchie JA. Colonic motor activity and bowel function. Part II. Distribution and incidence of motor activity at rest and after food and carbachol. Gut 1968;9:502-11. 28. Dietz AT, Newcomb PA, Marcus PM, Storer BE. The associa tion of body size and large bowel cancer risk in Wisconsin (United States) women. Cancer Causes Control 1995;6:306. 29. Steinmetz KA, Potter JD. Food-group consumption and co lon cancer in the Adelaide Case-Control Study. Int J Cancer 1993;53:720-7. 30. Bufill JA. Colorectal cancer: evidence for distinct genetic cat egories based on proximal or distal tumor location. Ann Intern Med 1990;113:779-88. 31. McMichael AJ, Potter JD. Host factors in carcinogenesis: cer tain bile-acid metabolic profiles that selectively increase the risk of proximal colon cancer. J Natl Cancer Inst 1985; 75:185-91. 32. Offerhaus GJ, De Feyter EP, Cornelisse DJ, Tersmette KWF, Floyd J, Kern SE, et al. The relationship of DNA aneuploidy to molecular genetic alterations in colorectal carcinoma. Gastroenterology 1992;102:1612-9. 33. Demers RY, Stawick LE, Demers P. Relative sensitivity of the fecal occult blood test and flexible sigmoidoscopy in detecting polyps. Prev Med 1985; 14:55-62. 34. Screening for colorectal cancer--United States. 1992-1993, and New Guidelines. MMWR Morb Mortal Wkly Rep 1996; 45:107-10. 35. Rothman KJ. Induction and latent periods. Am J Epidemiol 1981;114:253-9.