- Email: [email protected]
Changes in colorectal cancer over a 15-year period in a single United States city
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2000 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 95, No. 12, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(00)02171-7
Changes in Colorectal Cancer Over a 15-Year Period in a Single United States City Virender K. Sharma, M.D., Rajeev Vasudeva, M.D., F.A.C.G., and Colin W. Howden, M.D., F.A.C.G. University of Arkansas for Medical Sciences, Little Rock, Akansas; University of South Carolina, Columbia, South Carolina; and Northwestern University Medical School, Chicago, Illinois
OBJECTIVE: The aim of this study was to identify colorectal cancer trends in different patient groups in one region. METHODS: We reviewed the records of all patients with colorectal cancer who were seen at two hospitals in Columbia, SC, between 1981 and 1995. Patients were divided into three 5-yr cohorts (cohort 1, 1981–1985; cohort 2, 1986 – 1990; and cohort 3, 1991–1995). We recorded demographic data, histological type, tumor stage, grade, and survival. RESULTS: A total of 1395 patients were diagnosed, of whom 1252 had adenocarcinoma. Mean age of diagnosis was 64.5 yr in black and 66 in white patients (p ⫽ 0.036). Adenocarcinoma proximal to the splenic flexure was found in 62% of women and 49% of men (p ⬍ 0.001), and in 51% of black and 44% of white patients (p ⫽ 0.08). Proximal cancers accounted for 61%, 48%, and 32% of cohorts 1, 2, and 3, respectively. The proportions of patients with advanced stage (III and IV) adenocarcinoma in the three cohorts were 54%, 47%, and 40%, respectively (p ⫽ 0.001). Median survival for adenocarcinoma was 23.7, 23.5, and 23.1 months in cohorts 1, 2, and 3 (p ⫽ 0.9). CONCLUSIONS: The prevalence and demographics of colorectal cancer have been fairly stable at this center. There has been a significant decrease in proximal cancers. Although adenocarcinoma is being diagnosed at an earlier stage, this has not affected survival. Colonoscopy may be the most appropriate screening method in women and in black patients, in whom proximal cancers are more often found. (Am J Gastroenterol 2000;95:3615–3619. © 2000 by Am. Coll. of Gastroenterology)
INTRODUCTION Colorectal cancer (CRC) is the second most common cause of cancer-related death in all racial and ethnic groups in the United States (1). The American Cancer Society predicted 130,200 new cases of CRC and 56,300 related deaths for the year 2000 (1). Various population-based studies have reported a rise (2– 4), no change (5, 6) or a decrease (7) in the
This work was presented at the 1998 annual meeting of the American Gastroenterological Association in New Orleans, LA, and published in abstract form (Gastroenterology 1998;114:A40).
incidence of CRC over the past few decades. Investigators from different regions have described changes in the anatomical distribution of CRC (2–14), with an increasing proportion of proximal cancers reported as a common trend (3– 6, 8 –10, 12, 14). Racial and gender differences in anatomical distribution have also been reported (2, 3, 7, 9, 11). The exact reasons for these trends are not clear. We have analyzed CRC trends over a 15-yr period in two hospitals in Columbia, SC, a metropolitan area of about 500,000 in the southeastern US.
MATERIALS AND METHODS Patients diagnosed with CRC at Richland Memorial Hospital (a university-affiliated teaching county hospital; hospital 1), or W.J.B. Dorn Veterans Affairs (VA) Medical Center (hospital 2) between 1981 and 1995 were identified using hospital tumor registry databases. Tumor registries at these two hospitals had prospectively collected data on all cancer patients diagnosed or treated there. The tumor registries track all living patients through clinic follow-ups, phone calls, and letters to patients. The most recent update of the tumor registry databases before our analysis was on November 30, 1997. Patients were divided into three 5-yr cohorts (cohort 1, 1981–1985; cohort 2, 1986 –1990; and cohort 3, 1991– 1995). Patient demographics, tumor histology, grade (i.e., well, moderately, or poorly differentiated), stage at presentation, location, and survival data were recorded. Surgical reports, pathology specimens, and colonoscopy and/or radiological evaluation were used to determine the location of the tumor. Cancers located proximal to the splenic flexure were defined as proximal CRC; those either at or distal to the splenic flexure were defined as distal CRC. Cancers of stages 0 –II were categorized as localized, and those of stage III or IV as advanced. Because adenocarcinoma (AC) is the predominant histological subtype of CRC, analysis was limited to these patients only. Categorical variables were compared using the 2 test. Continuous variables were analyzed using Student’s t test (two-tailed, unpaired) or analysis of variance on t tests. Survival analysis was performed using the life table method with statistical significance assessed by the log-rank test.
3616
Sharma et al.
AJG – Vol. 95, No. 12, 2000
Table 1. Patients With Colorectal Adenocarcinoma at the Two Hospitals During the 15-Yr Period of Study Cohort
Hospital 1
Hospital 2
Cohort 1 (1981–1985)
N ⫽ 227 White, 61.2%; black, 38.8% Men, 101 (44.5%) N ⫽ 266 White, 54.5%; black, 45.5% Men, 131 (49.2%) N ⫽ 244 White, 54.5%; black, 45.5% Men, 115 (47%)
N ⫽ 143 White, 63.1%; black, 36.9% Men, 140 (97.9%) N ⫽ 176 White, 65.1%; black, 36.9% Men, 140 (99.4%) N ⫽ 196 White, 59.7%; black, 40.3% Men, 193 (98.5%)
Cohort 2 (1986–1990) Cohort 3 (1991–1995)
Overall and pair-wise survival was compared using Wilcoxon (Gehan) statistics. For patients who could not be contacted or who were lost to follow-up, the last contact date was considered as the date of death in our survival analysis. Values of p ⱕ 0.05 were considered statistically significant. The Statistical Package for the Social Sciences (SPSS) for Windows, version 7.5 (SPSS, Chicago, IL) was used for all analyses.
RESULTS A total of 1395 patients (59% white and 41% black; 69% male; mean age 65.2 yr; range 9 –99) were diagnosed with CRC at the two hospitals between1981 and 1995. Histologicaly, 93.6% were diagnosed with AC, 1.2% squamous cancer, 0.8% leiomyosarcoma, 0.2% carcinoid, and 4.2% unspecified. AC accounted for 93.4%, 92.9%, and 94.4% of CRC in cohorts 1, 2, and 3, respectively (p ⫽ 0.3). Details are outlined in Table 1. There were 737 patients with AC at hospital 1 and 515 at hospital 2. All but seven patients at hospital 2 (the VA hospital) were men. Otherwise, there were no significant differences between the patients in the two hospitals with respect to age, race, histological type of tumor, and tumor grade or stage. Cohorts 1, 2, and 3 had 370 patients (mean age 64.8 yr, SD 12.4 yr), 442 patients (mean age 65.4 yr, SD 12.5 yr), and 440 patients (mean age 65.4 yr, SD 12.7 yr), respectively, with AC (p ⫽ 0.48). Details of these patients and principal results are summarized in Table 2. At the time of diagnosis, AC was present in stages 0, I, II, III, and IV in 8.4%, 16.9%, 27.7%, 25.8%, and 21.2% of patients, respectively. Among men, stage 0, I, II, III, and IV AC was diagnosed in 11.4%, 17.6%, 26.1%, 24.1%, and 20.8% of patients, respectively. The corresponding values in
women were 1.8%, 15.4%, 31.1%, 29.6%, and 22.1% (p ⬍ 0.0001). AC was diagnosed in stages 0, I, II, III, and IV in 9.5%, 17.7%, 25.1%, 23.4%, and 24.3% of black patients. The corresponding values in white patients were 7%, 15.7%, 30%, 27.8%, and 19.5%, respectively (p ⫽ 0.13). The mean age at diagnosis of AC in black and in white patients was 64.5 and 66 yr, respectively (p ⫽ 0.036). The mean age at diagnosis of AC was 65.4 yr in women and 65.3 yr in men (p ⫽ 0.9). The mean ages at diagnosis for localized and advanced cancers were 66.5 and 64.1 yr, respectively (p ⫽ 0.002). The mean age at diagnosis for well, moderately, and poorly differentiated AC was 67.1, 66.3, and 63.5 yr, respectively (p ⫽ 0.04). A total of 68% of well differentiated, 49.9% of moderately differentiated, and 17.2% of poorly differentiated AC presented at the localized stage (p ⬍ 0.001). Proximal AC was diagnosed in 46.4% of the patients. In all, 61% of women and 39% of men had proximal AC (p ⬍ 0.001), which was found in 51% of blacks and 44% of whites (p ⫽ 0.08). Data on median survival for the three cohorts are shown in Table 2 and Figure 1. There was no significant difference in survival among the three cohorts (p ⫽ 0.90 for overall comparison). Median survival for stage 0, I, II, III, and IV AC was 42.4, 53.9, 46.7, 24.5, and 8.04 months, respectively (p ⬍ 0.001; Fig. 2). There was no significant difference in survival among patients with stages 0, I, or II AC (p ⬎ 0.10 for all pairwise comparisons). Patients with stage III or IV AC had significantly worse survival compared to patients with localized stage 0, I, or II cancer (p ⬍ 0.001 for all pairwise comparisons). Patients with stage IV cancer had significantly worse survival than those with stage III (p ⬍ 0.0001). Median survival for patients with proximal and distal AC was 25.1 and 22.7 months, respectively (p ⫽ 0.79). Median survival was 24.3 months for women and
Table 2. Trends in Colorectal Adenocarcinoma Among the Three Cohorts
White:black (%) Male:female (%) Well:moderately:poorly differentiated (%) Proximal:distal (%) Localized:advanced stage (%) Median survival (months)
1981–1985
1986–1990
1991–1995
38:62 73:27 36:51:13 61:39 46:54 23.6
41:59 78:22 22:66:12 48:52 53:47 23.4
43:57 81:19 19:69:12 32:68 60:40 23.4
p Value for Trend 0.4 0.48 ⬍0.001 ⬍0.001 0.001 0.9
AJG – December, 2000
Changes in Colorectal Cancer Over 15 Years in One US City
3617
Figure 3. Trends in the site of colorectal adenocarcinoma in the three cohorts. (}) ⫽ transverse, (䊐) ⫽ ascending, (Œ) ⫽ transverse, (⫻) ⫽ descending, (*) ⫽ sigmoid, (●) ⫽ anorectum.
DISCUSSION Figure 1. Survival (months) in colorectal adenocarcinoma in the three cohorts.
22.8 months for men (p ⫽ 0.02). Median survival was 24.8 months for black and 21.4 months for white patients (p ⫽ 0.02). The trends in anatomical distribution of AC are shown in Figure 3. There was a statistically significant change in the pattern of distribution of AC over the 15-yr duration of the study (p ⬍ 0.001).
Figure 2. Survival (months) in colorectal adenocarcinoma according to stage.
Colorectal cancer is a common and preventable malignancy. It is the third leading cause of cancer death in both men and women, and the second most common overall (1). Some population-based studies have suggested a slow increase in the incidence of CRC over the past few decades, whereas others have shown either no change or a decrease in incidence (2–7). Analysis of our 15-yr data from two local hospitals revealed some interesting trends in CRC. AC accounted for 94% of all cases of CRC that we reviewed. The number of cases of CRC and the proportion of AC remained fairly stable over the period of review. We did not see any major change in the racial or gender distribution of the disease over the period studied. A proximal shift of CRC to the right colon was first observed by Ottenheimer et al. in 1955 (8). Since then, multiple studies have also documented this proximal shift in CRC (3– 6, 8 –10, 12, 14). However, this has not been confirmed in all studies (2, 7, 11, 13). Different reports on cancer in the left colon or rectosigmoid region have variously revealed increasing, stable, or decreasing incidence (3–14). We found a significant increase in the proportion of left-sided AC with a concomitant decrease in the proportion of right-sided AC. This was predominantly due to a significant increase in AC in the rectosigmoid region and a decrease in the ascending colon. A fall in the incidence of ascending colon cancer has not been reported before, and the reasons for this in our population remain to be elucidated. Adenocarcinoma was diagnosed at a statistically significantly younger age in black than in white patients. However, the absolute difference was small and may not be clinically relevant. AC was found at an earlier stage in men. This could reflect differences in health-seeking behavior between men and women. Alternatively, men may be more likely to be referred for colorectal cancer screening (14) because of
3618
Sharma et al.
the erroneous but widely held belief that CRC is predominantly a disease of men. There may also be differences in tumor biology between men and women (15). There was a significant decrease in the proportion of well differentiated AC over the study period. Cancers of advanced stage and those that were poorly differentiated presented at a significantly earlier age. As reported by other investigators (9), we found a stage/grade interaction with poorly differentiated tumors presenting at a younger age and at more advanced stages. Proximal cancers were seen more often in black than in white patients, and in women more often than in men. Other investigators have reported similar trends (11, 12, 16 –18). There was no significant difference in the age of presentation based on tumor location. However, others have reported an earlier age of diagnosis in patients with distal cancers (18). We found evidence that AC is being diagnosed at an earlier stage (Table 2). This may reflect the more widespread availability and use of appropriate endoscopic procedures and the effect of CRC screening programs. Unfortunately, and despite the earlier stage at diagnosis over the study period, we did not find any improvement in survival among the three cohorts (Fig. 1). We do not have any data on previous CRC screening in these patients or on the number of cancers that were detected as a result of screening for CRC. However, the lack of improvement in survival among the three cohorts may be partly explained by the inclusion in screening programs of patients who are poor candidates for screening and who have high risks for surgical intervention. Such patients with significant comorbidity may not have been offered surgery for CRC, may not have survived surgery, or may have succumbed to concomitant illnesses. The earlier diagnosis of CRC in such patients might not, therefore, be translated into improved survival. Our recent survey of practicing physicians found that many continue to offer screening for CRC to patients with substantial comorbidity or advanced age (19). Others have also reported a lack of survival advantage despite an earlier stage at diagnosis (20, 21). Advanced stage (Fig. 2) and poor histological grade adversely affect survival. In contrast to previous reports, we found that black patients had significantly better overall survival than white patients (17). Despite being diagnosed at a later stage, women had significantly better overall survival than men, consistent with possible differences in tumor biology between the sexes (15). We do not believe that our data are subject to major selection bias, as we identified all patients diagnosed at either hospital with CRC by using the hospitals’ cancer registry databases. However, because of local referral patterns, our study sample may be composed mainly of persons from lower socioeconomic groups. Individuals from higher socioeconomic strata with CRC may have been treated at other hospitals in the city. In conclusion, CRC continues to be a significant health care problem in this region of the US. Over the 15-yr study
AJG – Vol. 95, No. 12, 2000
period, we observed an increased incidence of distal cancers, as well as a more frequent occurrence of proximal cancers in black patients and in women. Although the explanation for these findings is unclear, this suggests that colonoscopy may be the most appropriate screening tool for these patient subgroups. Recent studies indicate that screening colonoscopy detects neoplastic lesions in the right colon that would have been missed by flexible sigmoidoscopy (22–24). Although we found that CRC is being diagnosed at an earlier stage, we did not observe any resulting improvement in survival. This may be partly explained by the inclusion in screening programs of patients who are inappropriate candidates for screening. Reprint requests and correspondence: Virender Kumar Sharma, M.D., University of Arkansas for Medical Sciences, 4301 West Markham, Slot # 567, Little Rock, AR 72205-7199. Received Feb. 29, 2000; accepted Aug. 16, 2000.
REFERENCES 1. Greenlee RT, Murray T, Bolden S, et al. Cancer statistics, 2000. CA Cancer J Clin 2000;50:7–33. 2. Chow EH, Devesa SS, Blot WJ. Colon cancer incidence: Recent trends in United State. Cancer Cause Control 1991;2: 419 –25. 3. Axtell LM, Chiazze L. Changing relative frequency of cancers of the colon and rectum in United States. Cancer 1966;19:750 – 4. 4. Devesa SS, Chow WH. Variation in colorectal cancer incidence in the United States by the subsite of origin. Cancer 1993;71:3819 –25. 5. Obrand DI, Gordon PH. Continued change in the distribution of colorectal carcinoma. Br J Surg 1998;85:246 – 8. 6. Beard CM, Spencer RJ, Weiland LH, et al. Trends in colorectal cancer over a half century in Rochester, Minnesota, 1940 –1989. Ann Epidemiol 1995;5:210 – 4. 7. Polednak AP. Trends in cancer incidence in Connecticut. Conn Med 1997;61:211– 8. 8. Ottenheimer EJ, Oughterson AW. Observations on cancer of the colon and rectum in Connecticut. N Engl J Med 1955;252: 561–7. 9. Jessup JM, McGinnis LS, Steele GD Jr, et al. The national cancer database. Report on colon cancer. Cancer 1996;78: 918 –26. 10. Steele GD Jr. The national cancer database. Report on colon cancer. Cancer 1994;74:1979 – 89. 11. Crerand S, Feely TM, Waldron RP, et al. Colorectal carcinoma over 30 year at one hospital: No evidence of a shift to the right. Int J Colorectal Dis 1991;6:184 –7. 12. Rosato FE, Marks G. Changing site distribution pattern of colorectal cancer at Thomas Jefferson Hospital. J Am Soc Colon Rectal Surg 1981;24:93–5. 13. Stewart RJ, Stewart AW, Turnbull PRG, et al. Sex differences in subsite incidence of large bowel cancer. J Am Soc Colon Rectal Surg 1983;26:658 – 60. 14. Vobecky J, Leduc C, Devroede G. Sex differences in the changing anatomic distribution of colorectal carcinoma. Cancer 1985;54:3065–9. 15. Alley PG, McNee RK. Age and sex differences in right colon cancer. Dis Colon Rectum 1986;29:227–9. 16. Johnson H. Untreated colorectal cancer in a community hospital. J Surg Oncol 185;29:198 –200.
AJG – December, 2000
17. Johnson H, Carstens R. Anatomic distribution of colonic carcinomas. Interracial differences in a community hospital population. Cancer 1986;58:997–1000. 18. Slattery ML, Friedman GD, Potter JD, et al. A description of age, sex, and site distributions of colonic carcinoma in three geographical areas. Cancer 1996;78:1666 –70. 19. Sharma VK, Vasudeva R, Howden CW. Colorectal cancer screening and surveillance practices by primary care physicians: Results of a national survey. Am J Gastroenterol 2000;95:1551– 6. 20. Ohman U. Colorectal carcinoma—Trends and results over a 30-year period. Dis Colon Rectum 1982;25:431– 40.
Changes in Colorectal Cancer Over 15 Years in One US City
3619
21. Becker N, Smith EM, Wahrendorf J. Time trends in cancer mortality in the Federal Republic of Germany: Progress against cancer. Int J Cancer 1989;43:245–9. 22. Lieberman DA, Weiss DG, Bond JH, et al. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. N Engl J Med 2000;343:162– 8. 23. Imperiale TF, Wagner DR, Lin CY, et al. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169 – 74. 24. Podolsky DK. Going the distance: The case for true colorectal cancer screening. N Engl J Med 2000;343:207– 8.
Vol. 95, No. 12, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(00)02171-7
Changes in Colorectal Cancer Over a 15-Year Period in a Single United States City Virender K. Sharma, M.D., Rajeev Vasudeva, M.D., F.A.C.G., and Colin W. Howden, M.D., F.A.C.G. University of Arkansas for Medical Sciences, Little Rock, Akansas; University of South Carolina, Columbia, South Carolina; and Northwestern University Medical School, Chicago, Illinois
OBJECTIVE: The aim of this study was to identify colorectal cancer trends in different patient groups in one region. METHODS: We reviewed the records of all patients with colorectal cancer who were seen at two hospitals in Columbia, SC, between 1981 and 1995. Patients were divided into three 5-yr cohorts (cohort 1, 1981–1985; cohort 2, 1986 – 1990; and cohort 3, 1991–1995). We recorded demographic data, histological type, tumor stage, grade, and survival. RESULTS: A total of 1395 patients were diagnosed, of whom 1252 had adenocarcinoma. Mean age of diagnosis was 64.5 yr in black and 66 in white patients (p ⫽ 0.036). Adenocarcinoma proximal to the splenic flexure was found in 62% of women and 49% of men (p ⬍ 0.001), and in 51% of black and 44% of white patients (p ⫽ 0.08). Proximal cancers accounted for 61%, 48%, and 32% of cohorts 1, 2, and 3, respectively. The proportions of patients with advanced stage (III and IV) adenocarcinoma in the three cohorts were 54%, 47%, and 40%, respectively (p ⫽ 0.001). Median survival for adenocarcinoma was 23.7, 23.5, and 23.1 months in cohorts 1, 2, and 3 (p ⫽ 0.9). CONCLUSIONS: The prevalence and demographics of colorectal cancer have been fairly stable at this center. There has been a significant decrease in proximal cancers. Although adenocarcinoma is being diagnosed at an earlier stage, this has not affected survival. Colonoscopy may be the most appropriate screening method in women and in black patients, in whom proximal cancers are more often found. (Am J Gastroenterol 2000;95:3615–3619. © 2000 by Am. Coll. of Gastroenterology)
INTRODUCTION Colorectal cancer (CRC) is the second most common cause of cancer-related death in all racial and ethnic groups in the United States (1). The American Cancer Society predicted 130,200 new cases of CRC and 56,300 related deaths for the year 2000 (1). Various population-based studies have reported a rise (2– 4), no change (5, 6) or a decrease (7) in the
This work was presented at the 1998 annual meeting of the American Gastroenterological Association in New Orleans, LA, and published in abstract form (Gastroenterology 1998;114:A40).
incidence of CRC over the past few decades. Investigators from different regions have described changes in the anatomical distribution of CRC (2–14), with an increasing proportion of proximal cancers reported as a common trend (3– 6, 8 –10, 12, 14). Racial and gender differences in anatomical distribution have also been reported (2, 3, 7, 9, 11). The exact reasons for these trends are not clear. We have analyzed CRC trends over a 15-yr period in two hospitals in Columbia, SC, a metropolitan area of about 500,000 in the southeastern US.
MATERIALS AND METHODS Patients diagnosed with CRC at Richland Memorial Hospital (a university-affiliated teaching county hospital; hospital 1), or W.J.B. Dorn Veterans Affairs (VA) Medical Center (hospital 2) between 1981 and 1995 were identified using hospital tumor registry databases. Tumor registries at these two hospitals had prospectively collected data on all cancer patients diagnosed or treated there. The tumor registries track all living patients through clinic follow-ups, phone calls, and letters to patients. The most recent update of the tumor registry databases before our analysis was on November 30, 1997. Patients were divided into three 5-yr cohorts (cohort 1, 1981–1985; cohort 2, 1986 –1990; and cohort 3, 1991– 1995). Patient demographics, tumor histology, grade (i.e., well, moderately, or poorly differentiated), stage at presentation, location, and survival data were recorded. Surgical reports, pathology specimens, and colonoscopy and/or radiological evaluation were used to determine the location of the tumor. Cancers located proximal to the splenic flexure were defined as proximal CRC; those either at or distal to the splenic flexure were defined as distal CRC. Cancers of stages 0 –II were categorized as localized, and those of stage III or IV as advanced. Because adenocarcinoma (AC) is the predominant histological subtype of CRC, analysis was limited to these patients only. Categorical variables were compared using the 2 test. Continuous variables were analyzed using Student’s t test (two-tailed, unpaired) or analysis of variance on t tests. Survival analysis was performed using the life table method with statistical significance assessed by the log-rank test.
3616
Sharma et al.
AJG – Vol. 95, No. 12, 2000
Table 1. Patients With Colorectal Adenocarcinoma at the Two Hospitals During the 15-Yr Period of Study Cohort
Hospital 1
Hospital 2
Cohort 1 (1981–1985)
N ⫽ 227 White, 61.2%; black, 38.8% Men, 101 (44.5%) N ⫽ 266 White, 54.5%; black, 45.5% Men, 131 (49.2%) N ⫽ 244 White, 54.5%; black, 45.5% Men, 115 (47%)
N ⫽ 143 White, 63.1%; black, 36.9% Men, 140 (97.9%) N ⫽ 176 White, 65.1%; black, 36.9% Men, 140 (99.4%) N ⫽ 196 White, 59.7%; black, 40.3% Men, 193 (98.5%)
Cohort 2 (1986–1990) Cohort 3 (1991–1995)
Overall and pair-wise survival was compared using Wilcoxon (Gehan) statistics. For patients who could not be contacted or who were lost to follow-up, the last contact date was considered as the date of death in our survival analysis. Values of p ⱕ 0.05 were considered statistically significant. The Statistical Package for the Social Sciences (SPSS) for Windows, version 7.5 (SPSS, Chicago, IL) was used for all analyses.
RESULTS A total of 1395 patients (59% white and 41% black; 69% male; mean age 65.2 yr; range 9 –99) were diagnosed with CRC at the two hospitals between1981 and 1995. Histologicaly, 93.6% were diagnosed with AC, 1.2% squamous cancer, 0.8% leiomyosarcoma, 0.2% carcinoid, and 4.2% unspecified. AC accounted for 93.4%, 92.9%, and 94.4% of CRC in cohorts 1, 2, and 3, respectively (p ⫽ 0.3). Details are outlined in Table 1. There were 737 patients with AC at hospital 1 and 515 at hospital 2. All but seven patients at hospital 2 (the VA hospital) were men. Otherwise, there were no significant differences between the patients in the two hospitals with respect to age, race, histological type of tumor, and tumor grade or stage. Cohorts 1, 2, and 3 had 370 patients (mean age 64.8 yr, SD 12.4 yr), 442 patients (mean age 65.4 yr, SD 12.5 yr), and 440 patients (mean age 65.4 yr, SD 12.7 yr), respectively, with AC (p ⫽ 0.48). Details of these patients and principal results are summarized in Table 2. At the time of diagnosis, AC was present in stages 0, I, II, III, and IV in 8.4%, 16.9%, 27.7%, 25.8%, and 21.2% of patients, respectively. Among men, stage 0, I, II, III, and IV AC was diagnosed in 11.4%, 17.6%, 26.1%, 24.1%, and 20.8% of patients, respectively. The corresponding values in
women were 1.8%, 15.4%, 31.1%, 29.6%, and 22.1% (p ⬍ 0.0001). AC was diagnosed in stages 0, I, II, III, and IV in 9.5%, 17.7%, 25.1%, 23.4%, and 24.3% of black patients. The corresponding values in white patients were 7%, 15.7%, 30%, 27.8%, and 19.5%, respectively (p ⫽ 0.13). The mean age at diagnosis of AC in black and in white patients was 64.5 and 66 yr, respectively (p ⫽ 0.036). The mean age at diagnosis of AC was 65.4 yr in women and 65.3 yr in men (p ⫽ 0.9). The mean ages at diagnosis for localized and advanced cancers were 66.5 and 64.1 yr, respectively (p ⫽ 0.002). The mean age at diagnosis for well, moderately, and poorly differentiated AC was 67.1, 66.3, and 63.5 yr, respectively (p ⫽ 0.04). A total of 68% of well differentiated, 49.9% of moderately differentiated, and 17.2% of poorly differentiated AC presented at the localized stage (p ⬍ 0.001). Proximal AC was diagnosed in 46.4% of the patients. In all, 61% of women and 39% of men had proximal AC (p ⬍ 0.001), which was found in 51% of blacks and 44% of whites (p ⫽ 0.08). Data on median survival for the three cohorts are shown in Table 2 and Figure 1. There was no significant difference in survival among the three cohorts (p ⫽ 0.90 for overall comparison). Median survival for stage 0, I, II, III, and IV AC was 42.4, 53.9, 46.7, 24.5, and 8.04 months, respectively (p ⬍ 0.001; Fig. 2). There was no significant difference in survival among patients with stages 0, I, or II AC (p ⬎ 0.10 for all pairwise comparisons). Patients with stage III or IV AC had significantly worse survival compared to patients with localized stage 0, I, or II cancer (p ⬍ 0.001 for all pairwise comparisons). Patients with stage IV cancer had significantly worse survival than those with stage III (p ⬍ 0.0001). Median survival for patients with proximal and distal AC was 25.1 and 22.7 months, respectively (p ⫽ 0.79). Median survival was 24.3 months for women and
Table 2. Trends in Colorectal Adenocarcinoma Among the Three Cohorts
White:black (%) Male:female (%) Well:moderately:poorly differentiated (%) Proximal:distal (%) Localized:advanced stage (%) Median survival (months)
1981–1985
1986–1990
1991–1995
38:62 73:27 36:51:13 61:39 46:54 23.6
41:59 78:22 22:66:12 48:52 53:47 23.4
43:57 81:19 19:69:12 32:68 60:40 23.4
p Value for Trend 0.4 0.48 ⬍0.001 ⬍0.001 0.001 0.9
AJG – December, 2000
Changes in Colorectal Cancer Over 15 Years in One US City
3617
Figure 3. Trends in the site of colorectal adenocarcinoma in the three cohorts. (}) ⫽ transverse, (䊐) ⫽ ascending, (Œ) ⫽ transverse, (⫻) ⫽ descending, (*) ⫽ sigmoid, (●) ⫽ anorectum.
DISCUSSION Figure 1. Survival (months) in colorectal adenocarcinoma in the three cohorts.
22.8 months for men (p ⫽ 0.02). Median survival was 24.8 months for black and 21.4 months for white patients (p ⫽ 0.02). The trends in anatomical distribution of AC are shown in Figure 3. There was a statistically significant change in the pattern of distribution of AC over the 15-yr duration of the study (p ⬍ 0.001).
Figure 2. Survival (months) in colorectal adenocarcinoma according to stage.
Colorectal cancer is a common and preventable malignancy. It is the third leading cause of cancer death in both men and women, and the second most common overall (1). Some population-based studies have suggested a slow increase in the incidence of CRC over the past few decades, whereas others have shown either no change or a decrease in incidence (2–7). Analysis of our 15-yr data from two local hospitals revealed some interesting trends in CRC. AC accounted for 94% of all cases of CRC that we reviewed. The number of cases of CRC and the proportion of AC remained fairly stable over the period of review. We did not see any major change in the racial or gender distribution of the disease over the period studied. A proximal shift of CRC to the right colon was first observed by Ottenheimer et al. in 1955 (8). Since then, multiple studies have also documented this proximal shift in CRC (3– 6, 8 –10, 12, 14). However, this has not been confirmed in all studies (2, 7, 11, 13). Different reports on cancer in the left colon or rectosigmoid region have variously revealed increasing, stable, or decreasing incidence (3–14). We found a significant increase in the proportion of left-sided AC with a concomitant decrease in the proportion of right-sided AC. This was predominantly due to a significant increase in AC in the rectosigmoid region and a decrease in the ascending colon. A fall in the incidence of ascending colon cancer has not been reported before, and the reasons for this in our population remain to be elucidated. Adenocarcinoma was diagnosed at a statistically significantly younger age in black than in white patients. However, the absolute difference was small and may not be clinically relevant. AC was found at an earlier stage in men. This could reflect differences in health-seeking behavior between men and women. Alternatively, men may be more likely to be referred for colorectal cancer screening (14) because of
3618
Sharma et al.
the erroneous but widely held belief that CRC is predominantly a disease of men. There may also be differences in tumor biology between men and women (15). There was a significant decrease in the proportion of well differentiated AC over the study period. Cancers of advanced stage and those that were poorly differentiated presented at a significantly earlier age. As reported by other investigators (9), we found a stage/grade interaction with poorly differentiated tumors presenting at a younger age and at more advanced stages. Proximal cancers were seen more often in black than in white patients, and in women more often than in men. Other investigators have reported similar trends (11, 12, 16 –18). There was no significant difference in the age of presentation based on tumor location. However, others have reported an earlier age of diagnosis in patients with distal cancers (18). We found evidence that AC is being diagnosed at an earlier stage (Table 2). This may reflect the more widespread availability and use of appropriate endoscopic procedures and the effect of CRC screening programs. Unfortunately, and despite the earlier stage at diagnosis over the study period, we did not find any improvement in survival among the three cohorts (Fig. 1). We do not have any data on previous CRC screening in these patients or on the number of cancers that were detected as a result of screening for CRC. However, the lack of improvement in survival among the three cohorts may be partly explained by the inclusion in screening programs of patients who are poor candidates for screening and who have high risks for surgical intervention. Such patients with significant comorbidity may not have been offered surgery for CRC, may not have survived surgery, or may have succumbed to concomitant illnesses. The earlier diagnosis of CRC in such patients might not, therefore, be translated into improved survival. Our recent survey of practicing physicians found that many continue to offer screening for CRC to patients with substantial comorbidity or advanced age (19). Others have also reported a lack of survival advantage despite an earlier stage at diagnosis (20, 21). Advanced stage (Fig. 2) and poor histological grade adversely affect survival. In contrast to previous reports, we found that black patients had significantly better overall survival than white patients (17). Despite being diagnosed at a later stage, women had significantly better overall survival than men, consistent with possible differences in tumor biology between the sexes (15). We do not believe that our data are subject to major selection bias, as we identified all patients diagnosed at either hospital with CRC by using the hospitals’ cancer registry databases. However, because of local referral patterns, our study sample may be composed mainly of persons from lower socioeconomic groups. Individuals from higher socioeconomic strata with CRC may have been treated at other hospitals in the city. In conclusion, CRC continues to be a significant health care problem in this region of the US. Over the 15-yr study
AJG – Vol. 95, No. 12, 2000
period, we observed an increased incidence of distal cancers, as well as a more frequent occurrence of proximal cancers in black patients and in women. Although the explanation for these findings is unclear, this suggests that colonoscopy may be the most appropriate screening tool for these patient subgroups. Recent studies indicate that screening colonoscopy detects neoplastic lesions in the right colon that would have been missed by flexible sigmoidoscopy (22–24). Although we found that CRC is being diagnosed at an earlier stage, we did not observe any resulting improvement in survival. This may be partly explained by the inclusion in screening programs of patients who are inappropriate candidates for screening. Reprint requests and correspondence: Virender Kumar Sharma, M.D., University of Arkansas for Medical Sciences, 4301 West Markham, Slot # 567, Little Rock, AR 72205-7199. Received Feb. 29, 2000; accepted Aug. 16, 2000.
REFERENCES 1. Greenlee RT, Murray T, Bolden S, et al. Cancer statistics, 2000. CA Cancer J Clin 2000;50:7–33. 2. Chow EH, Devesa SS, Blot WJ. Colon cancer incidence: Recent trends in United State. Cancer Cause Control 1991;2: 419 –25. 3. Axtell LM, Chiazze L. Changing relative frequency of cancers of the colon and rectum in United States. Cancer 1966;19:750 – 4. 4. Devesa SS, Chow WH. Variation in colorectal cancer incidence in the United States by the subsite of origin. Cancer 1993;71:3819 –25. 5. Obrand DI, Gordon PH. Continued change in the distribution of colorectal carcinoma. Br J Surg 1998;85:246 – 8. 6. Beard CM, Spencer RJ, Weiland LH, et al. Trends in colorectal cancer over a half century in Rochester, Minnesota, 1940 –1989. Ann Epidemiol 1995;5:210 – 4. 7. Polednak AP. Trends in cancer incidence in Connecticut. Conn Med 1997;61:211– 8. 8. Ottenheimer EJ, Oughterson AW. Observations on cancer of the colon and rectum in Connecticut. N Engl J Med 1955;252: 561–7. 9. Jessup JM, McGinnis LS, Steele GD Jr, et al. The national cancer database. Report on colon cancer. Cancer 1996;78: 918 –26. 10. Steele GD Jr. The national cancer database. Report on colon cancer. Cancer 1994;74:1979 – 89. 11. Crerand S, Feely TM, Waldron RP, et al. Colorectal carcinoma over 30 year at one hospital: No evidence of a shift to the right. Int J Colorectal Dis 1991;6:184 –7. 12. Rosato FE, Marks G. Changing site distribution pattern of colorectal cancer at Thomas Jefferson Hospital. J Am Soc Colon Rectal Surg 1981;24:93–5. 13. Stewart RJ, Stewart AW, Turnbull PRG, et al. Sex differences in subsite incidence of large bowel cancer. J Am Soc Colon Rectal Surg 1983;26:658 – 60. 14. Vobecky J, Leduc C, Devroede G. Sex differences in the changing anatomic distribution of colorectal carcinoma. Cancer 1985;54:3065–9. 15. Alley PG, McNee RK. Age and sex differences in right colon cancer. Dis Colon Rectum 1986;29:227–9. 16. Johnson H. Untreated colorectal cancer in a community hospital. J Surg Oncol 185;29:198 –200.
AJG – December, 2000
17. Johnson H, Carstens R. Anatomic distribution of colonic carcinomas. Interracial differences in a community hospital population. Cancer 1986;58:997–1000. 18. Slattery ML, Friedman GD, Potter JD, et al. A description of age, sex, and site distributions of colonic carcinoma in three geographical areas. Cancer 1996;78:1666 –70. 19. Sharma VK, Vasudeva R, Howden CW. Colorectal cancer screening and surveillance practices by primary care physicians: Results of a national survey. Am J Gastroenterol 2000;95:1551– 6. 20. Ohman U. Colorectal carcinoma—Trends and results over a 30-year period. Dis Colon Rectum 1982;25:431– 40.
Changes in Colorectal Cancer Over 15 Years in One US City
3619
21. Becker N, Smith EM, Wahrendorf J. Time trends in cancer mortality in the Federal Republic of Germany: Progress against cancer. Int J Cancer 1989;43:245–9. 22. Lieberman DA, Weiss DG, Bond JH, et al. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. N Engl J Med 2000;343:162– 8. 23. Imperiale TF, Wagner DR, Lin CY, et al. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169 – 74. 24. Podolsky DK. Going the distance: The case for true colorectal cancer screening. N Engl J Med 2000;343:207– 8.