The relation of age, race, and gender to the subsite location of colorectal carcinoma

The shift of colon carcinomas to a more proximal location as people age may be even more pronounced than described by Nelson et al.1 We observed a similar shift in our study of the 77,163 cases of invasive colon adenocarcinoma from the California Cancer Registry (CCR) over the period 1988-1993.2 Like the Illinois Tumor Registry, the CCR is a total population-based registry with the advantages of absence of selection bias and large numbers. Registered cases consist of 39,337 cancers in men and 37,826 in women. This includes 61,142 in whites; 6201 in Hispanics; 4848 in blacks; 4506 in Asians, Pacific Islanders, and Native Americans; and 466 (0.6%) in patients of unknown ethnicity. Colorectal adenocarcinomas were divided into three subsites: 44,207 in the left colon (descending, sigmoid, and rectosigmoid colon and rectum); 10,316 in the transverse colon (hepatic flexure, transverse colon, and splenic flexure); and 20,369 in the right colon (cecum and ascending colon). If all 9 specific subsites were used, conclusions would be similar but presentation would be more difficult. In 2271 instances, the exact subsite of origin was not known. Starting at age 50 years, there is a significant "shift to the right" with increasing age (Fig. 1 (4K)). We feel that the percentage of tumors at each site more clearly demonstrates differences than does any attempt at visually comparing the two line graphs in the article by Nelson et al. Most of the decrease in the percentage of cases in the left colon is accounted for by an increase in the percentage of lesions in the right colon, but there also is an increase in transverse colon lesions. If plotted as semilogarithmic graphs, the changes with age after 50 years for all three subsites are straight lines, implying a constant rate of change in the percentages with age. Changes, with patient age, in the percentages of colorectal adenocarcinomas in the right, transverse, and left colon are shown; all patients are represented. Like Nelson et al., we feel that this shift has an impact on the choice of screening techniques, especially the use of colonoscopy versus flexible sigmoidoscopy. Based on experience, we assumed that lesions of the sigmoid colon, rectosigmoid colon, and rectum would be within reach of a flexible sigmoidoscope. This definition corresponds exactly to the definition of distal colon given by Nelson et al. Less than half of colon adenocarcinomas will be visualized by sigmoidoscopy, starting with the group age 75-79 years in all patients (ages 70-74 years in women and ages 80-84 years in men, Fig. 2 (4K)). Ethnic variations are clear (Fig. 3 (6K) and Table 1), with less than half of the lesions being within reach of the sigmoidoscope after the group age 40-44 years in blacks, after the group age 80-84 years in Hispanics, and never in the Asian/Pacific Islander/Native American population. Thus, with increasing age, total colonoscopy will be more likely to detect carcinomas. The exact age at which the shift to colonoscopy will occur will vary with both gender and ethnicity. Changes, with patient age, in the percentages of colorectal adenocarcinomas within reach of the flexible sigmoidoscope are shown; all ethnicities are represented. Changes, with patient age, in the percentages of colorectal adenocarcinomas within reach of the flexible sigmoidoscope are shown; both genders are represented. To some extent, the "shift to the right" contradicts Burkitt's classic ideas about colon carcinoma.3 He postulated that part of the reason these tumors were predominantly in the distal colon is that the bowel wall at that site is in prolonged contact with a concentrated stool containing carcinogens. If this were an overriding factor, one might expect an even higher percentage of distal colon adenocarcinomas as people age. Our data show the opposite. Sidney L Saltzstein M.D. M.P.H.*, Cynthia A. Behling M.D. Ph.D.*, Thomas J. Savides M.D.