- Email: [email protected]
Case-control study of screening colonoscopy in relatives of patients with colorectal cancer
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2003 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 98, No. 2, 2003 ISSN 0002-9270/03/$30.00 doi:10.1016/S0002-4975(02)05926-9
Case-Control Study of Screening Colonoscopy in Relatives of Patients With Colorectal Cancer Yaron Niv, M.D., Ram Dickman, M.D., Arie Figer, M.D., Galia Abuksis, R.N., M.Sc., and Gerald Fraser, M.D. Departments of Gastroenterology and Oncology, Rabin Medical Center, Beilinson Campus, Petah Tiqwa, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
OBJECTIVES: The efficacy of colonoscopic screening and polypectomy for the prevention of colorectal cancer (CRC) is well accepted but has never been documented in a prospective, controlled study. Screening by sigmoidoscopy has been found to reduce mortality from cancer of the rectum and distal colon. Case-control studies provide an alternative method for determining the efficacy of screening methods. METHODS: Between 1998 and 2000, a total of 40 subjects were found to have CRC (study group) and 160 had a normal colon (control group) among asymptomatic individuals participating in a screening colonoscopy program for a high-risk population of first-degree relatives of CRC patients. We compared these groups for screening by fecal occult blood testing, flexible sigmoidoscopy, barium enema, and colonoscopy in the 10-yr period before the index colonoscopy. RESULTS: Screening colonoscopy was performed in only 2.5% of the case subjects and 48.7% of controls (p ⬍ 0.0001), and all screening procedures in 12.5% and 73.7%, respectively (p ⬍ 0.0001). A statistically significant difference was also found for screening with fecal occult blood test, but not for flexible sigmoidoscopy or barium enema. Significant adenomatous polyps ⬎1 cm in diameter were detected and removed in 19% of the control group within 10 yr of the index colonoscopy. Six (15%) of the patients in the study group died of CRC. CONCLUSIONS: Screening by colonoscopy can prevent progression to CRC from adenomatous polyps and may reduce the mortality associated with this devastating disease. (Am J Gastroenterol 2003;98:486 – 489. © 2003 by Am. Coll. of Gastroenterology)
INTRODUCTION Several types of colorectal cancer (CRC) screening tests have been developed for the detection of early stage disease and disease precursors. The efficacy of colonoscopic screening and polypectomy is well accepted but has never been Arie Figer, M.D., is currently located at the Department of Oncology, Souraski Medical Center, Tel Aviv, Israel.
documented in a prospective, controlled study (1). Colonoscopy is relatively costly, uncomfortable, and not without complications (2, 3), and comprehensive screening of an average-risk population may not be feasible. However, it has been recommended as a possibility by a multidisciplinary expert panel in 1997 (4). Screening by fecal occult blood testing (FOBT) has been tested in large, prospective, case-controlled studies, and a significant reduction in mortality from CRC was reported (5–9). Recently, we have demonstrated a mortality reduction of 61% in a relatively small population (10). Screening by sigmoidoscopy was shown in two retrospective case-control studies to reduce mortality from cancer of the rectum and distal colon (11, 12). In the absence of a prospective, controlled study, a retrospective, case-control design may be a good alternative for estimating the efficacy of screening tests. In the present study, we investigated high-risk patients who underwent screening colonoscopy and other tests at our center. Those who were found to have CRC were compared with those who had a normal colonoscopy for use in CRC screening tests in the 5–10 yr before the index colonoscopy. Because all participants were asymptomatic, we overcame the potential bias of possible symptom-related procedures.
MATERIALS AND METHODS Study Subjects and Protocol A special clinic for early detection and secondary prevention of CRC was established in the Gastroenterology Department of Rabin Medical Center in 1995. The average-risk population was referred for FOBT, and high-risk individuals were referred for colonoscopy. Between January 1, 1998, and December 31, 1999, our department performed 4005 colonoscopies, including 512 in asymptomatic, healthy, high-risk individuals, of whom 484 had only one first degree relative with CRC but no clinical evidence of hereditary nonpolyposis colorectal cancer (HNPCC) or adenomatous polyposis coli. In all, 284 were found to have polyps on the index colonoscopy and were excluded from the study. Of the remaining 200 individuals, 40 had CRC (study group) and 160 had a normal colon (control group). None of the screenees were referred for colonoscopy because of positive
AJG – February, 2003
FOBT. Two screenees with a normal colon were later dropped from the study because technical difficulties made it impossible to reach the cecum during the testing. Colonoscopy was performed with an Olympus 140 series colonoscope. Colon preparation was done with Sophodex (Dexxon, Israel) or Golytly (Taro, Israel) according to the manufacturer’s instructions. Before the examination the patients participated in a patient education program that was conducted by a dedicated nurse, as previously described (13). The subjects’ outpatient and family practitioner medical records for the previous 10 yr were reviewed, and information was retrieved for FOBT, flexible sigmoidoscopy (FS), barium enema (BAE), and colonoscopy. The dates of testing and the findings were recorded. The hospital archives, files of the Oncology and Pathology Departments, and endoscopy reports were also searched and the relevant data recorded. All cases and controls completed a telephone survey conducted by two of the investigators (R.D. and G.A), which covered all colonic investigations performed during the last 10 yr, and the demographic data, clinical, endoscopic, and pathological findings. Because many procedures had been performed outside of the hospital at community health facilities, these interviews added considerable information to the patient files. Death certificates registered in the Ministry of Interior were examined December 31, 2000, for all names of the individuals screened. Statistical Analysis All of our analyses considered the 5- and 10-yr period just before the index colonoscopy in the case and control subjects. Because the official records were complemented by telephone interviews (which relied on patient memory) and family physician files, we considered the data for this period to be complete. An overall OR was calculated for having had any screening test within the 5- and 10-yr period before the index colonoscopy compared with no screening test. The same calculation was then performed for each procedure (FOBT, FS, BAE, and colonoscopy). Student’s t test and comparison of two proportions were used as needed. A p value ⬍ 0.05 was considered to be statistically significant.
Case-Control Study of Screening Colonoscopy
487
Table 1. Demographic Data of High-Risk, Asymptomatic Patients With CRC and Controls
No. of patients Sex Men Women Age (yr) Average ⫾ SD Range Origin Europe-America Asia-Africa Israel More than one 1st degree relative with CRC
Study Group n (%)
Control Group n (%)
40
160
32 (80) 8 (20)
91 (57) 69 (43)
0.013 0.013
61 ⫾ 11 33–76
58 ⫾ 11 24–79
0.124
25 (62) 15 (38) 0 (0) 11 (27)
64 (40) 37 (23) 59 (37) 41 (26)
0.020 0.108 0.000 0.842
p
CI for difference – 0.565 to – 0.235, p ⬍ 0.0001). The corresponding figures for the 10-yr period (Table 3) were 2.5% for the case subjects and 48.7% of the controls (difference of – 0.462 ⫾ 0.086, 95% CI – 0.631 to – 0.293, p ⬍ 0.0001). A statistically significant difference was also found for FOBT for the two periods but not for FS or BAE. Overall, in the 10-yr period before the index colonoscopy, 12.5% of the cases had undergone any screening test (FOBT, FS, BAE, or colonoscopy) compared with 73.7% of the controls (difference of – 0.612 ⫾ 0.086, 95% CI – 0.781 to – 0.448, p ⬍ 0.0001). In the control group, an adenomatous polyp ⬎1 cm in diameter had been removed in 29 subjects within the 10 yr before the index colonoscopy. The number of colonoscopies per 160 screenees per 10 yr was 78. By contrast, only one of the study subjects (3%) underwent colonoscopy in the 10 yr before the index colonoscopy. Six patients (15%) in this group later died of CRC, as confirmed by our review of the death certificates and the records of the tumor. Colonoscopy was performed during the 10-yr period at about the same frequency as FOBT (Tables 2 and 3). In all, 43% of the cancers in the study group and 40% of the significant adenomas in the controls were located proximal to the splenic flexure. No differences in use of aspirin or nonsteroidal anti-inflammatory drugs was found between the two groups.
RESULTS Demographic data for the study participants are shown in Table 1. There were 123 men and 77 women aged 24 –79 yr. The study group included more men than did the control group (p ⫽ 0.013), as well as more individuals of Israeli or European-American origin (p ⫽ 0.000, p ⫽ 0.020). There was no statistically significant difference between the groups in age or number of first degree relatives with CRC. The screening data for the 5-yr period before the index colonoscopy are shown in Table 2. Only 2.5% of the case subjects had undergone screening by colonoscopy compared with 42.5% of the controls (difference – 0.400 ⫾ 0.084, 95%
DISCUSSION The frequency with which colonoscopies were performed during the period preceding the diagnosis of cancer in patients with CRC and controls indicates that screening by colonoscopy can prevent the progression to CRC from adenomatous polyps in high-risk patients. Of the case subjects who were found to have CRC on the index colonoscopy, only 12.5% (five of 40) had undergone any previous screening compared with 73.7% (118/160) of the controls with a normal index colonoscopy. Furthermore, the control subjects who had undergone one or more screening examina-
488
Niv et al.
AJG – Vol. 98, No. 2, 2003
Table 2. Screening During the 5-Yr Period Before the Index Colonoscopy
No. of patients FOBT FS BAE Colonoscopy
Study Group n (%)
Control Group n (%)
⌬ ⫾ SE
40 3 (7.5) 2 (5.0) 3 (7.5) 1 (2.5)
160 80 (50.0) 2 (1.2) 18 (11.3) 68 (42.5)
⫺0.425 ⫾ 0.087 0.038 ⫾ 0.025 ⫺0.038 ⫾ 0.054 ⫺0.400 ⫾ 0.084
tions by FOBT, FS, or colonoscopy during the preceding 10 yr had a 25% (29/118) risk of an adenomatous polyp. The present study is noteworthy for the careful selection of the study sample. Only asymptomatic subjects with a first degree relative with CRC, but without FAP or HNPCC according to the Amsterdam criteria (4) were included. In this manner, we were able to rule out most of the potential confounding factors, namely history of adenoma or CRC before the 10-yr review period, diagnosis of inflammatory bowel disease, HNPCC, or FAP. These factors could lead to increased screening efforts or the risk of CRC, thereby obscuring the protective effect of screening. Although information about lifestyle and level of physical activity was not available from the medical records and was not addressed in the telephone interview, it is unlikely that these factors differed significantly between the case and the control groups, as all subjects were asymptomatic screenees. Our findings are in agreement with the study of Muller et al. (14) who found that patients with CRC had had significantly fewer endoscopic procedures during periods of up to 6 yr before the onset of the disease. However, they compared a heterogeneous population of all CRC patients in their files with controls that were matched for age, sex, and race. Their sample may have included symptomatic and asymptomatic individuals, as well as those at average and high risk. In all, 43% of the cancers in the study group and 40% of the polyps in the control group were found in the right colon, and would therefore have been missed by FS if not followed by total colonoscopy. Lieberman et al. (2) and Imperiale et al. (3) examined the role of colonoscopy in screening asymptomatic adults for CRC. Almost half of the screenees with advanced proximal neoplasm had no distal polyp. Combined, the two studies show that screening sigmoidoscopy will not detect—and thus will not prevent—CRC in
95% CI for Difference ⫺0.596 ⫺0.010 ⫺0.144 ⫺0.565
to to to to
⫺0.254 0.086 0.068 ⫺0.235
p 0.000 0.361 0.680 0.000
many asymptomatic, average-risk individuals. The studies also demonstrate the feasibility of colonoscopy for screening the average-risk population, as well as the importance of investigating the right colon. The approach to patients with one first degree relative with colon cancer is not firmly established (15, 16). Whether these patients should be monitored in the same way as average-risk patients or be screened more rigorously remains to be definitively determined. Selby et al. (11) compared the use of rigid sigmoidoscopy screening over a 10-yr period in a heterogeneous sample of patients who died of CRC, along with matched controls. By contrast, in the present study, we investigated a homogeneous, asymptomatic, high-risk population to determine whether the number of colonic investigations performed in the 10 yr before the index colonoscopy is different between CRC diagnosed cases and controls who had normal colonoscopic examination. Selby et al. (11) found that only 8.8% of the 261 case subjects had undergone previous screening sigmoidoscopy, compared with 24.2% of the 868 controls; the strength of the negative association was maintained even when the most recent sigmoidoscopy was done 9 to 10 yr before diagnosis. Thus, the risk of rectal cancer was markedly reduced for at least 10 yr after a single sigmoidoscopic examination. However, for the 268 subjects in the study of Selby et al. (11) with fatal lesion above the reach of the sigmoidoscope, and for the 268 controls, no significant difference in the number of sigmoidoscopic examinations was demonstrated. As in our study, the screening sigmoidoscopy led to the removal of polyps (12 adenomatous polyps of the 210 control subjects who had at least one screening sigmoidoscopy). In a related study, Newcomb et al. (12) noted a significantly lower number of sigmoidoscopies in patients with rectal cancer than in matched controls, but there was no such difference for digital rectal examination
Table 3. Screening During the 10-Yr Period Before the Index Colonoscopy
N FOBT FS BAE Colonoscopy Any screening test NSAID or aspirin
Study Group n (%)
Control Group n (%)
⌬ ⫾ SE
40 3 (7.5) 2 (5.0) 3 (7.5) 1 (2.5) 5 (12.5) 3 (7.5)
160 80 (50.5) 2 (1.2) 26 (16.3) 78 (48.7) 118 (73.7) 25 (15.6)
⫺0.425 ⫾ 0.087 0.038 ⫾ 0.025 ⫺0.088 ⫾ 0.062 ⫺0.462 ⫾ 0.086 ⫺0.612 ⫾ 0.086 ⫺0.081 ⫾ 0.061
NSAID ⫽ nonsteroidal anti-inflammatory drugs.
95% CI for Difference ⫺0.596 ⫺0.010 ⫺0.210 ⫺0.631 ⫺0.781 ⫺0.201
to to to to to to
⫺0.254 0.086 0.034 ⫺0.293 ⫺0.443 0.039
p 0.000 0.361 0.245 0.000 0.000 0.286
AJG – February, 2003
or FOBT. This finding differs from our data in which we show a significant difference in numbers of both colonoscopies and FOBTs (p ⬍ 0.001). A trend was also found for BAE (cases 6.8%, controls 14%), but it did not reach significance. Newcomb et al. (12) reported that individuals who had a single sigmoidoscopy had an 80% lower rate of rectal cancer than individuals who were never tested. Although these investigators tried to include only tests carried out for screening, it is possible that signs or symptoms prompting the examinations were not recorded. We overcame this problem by including only asymptomatic screenees. In conclusion, screening by FOBT or colonoscopy may prevent progression to CRC from adenomatous polyps and may thereby reduce the mortality rate of this devastating disease. Thus, we believe that FOBT or colonoscopy may be an option for screening the asymptomatic, relatively highrisk population. Colonoscopy is not widely used today for this purpose, because of lack of hard data to support this strategy, and because of a lack of financial, social, and organizational resources. However, recent cost analyses have noted that screening by colonoscopy is cost-effective relative to other medical interventions (17, 18).
Case-Control Study of Screening Colonoscopy
5. 6. 7. 8. 9.
10. 11. 12. 13.
Reprint requests and correspondence to: Yaron Niv, M.D., Department of Gastroenterology, Rabin Medical Center, Beilinson Campus, Petah Tiqwa, 49100, Israel. Received July 11, 2002; accepted Sep. 30, 2002.
14. 15.
REFERENCES 1. Winawer SJ, Wolff BG, Culp CE, et al. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009 –13. 2. 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. 3. 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. 4. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer
16. 17. 18.
489
screening: Clinical guidelines and rationale. Gastroenterology 1997;112:594 –642. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365–71. Kronberg O, Fenger C, Olsen J, et al. Randomized study of screening for colorectal cancer with fecal occult blood test. Lancet 1996;348:1467–71. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Randomized controlled trial of fecal occult blood screening for colorectal cancer. Lancet 1996;348:1472–7. Winawer SJ, Andrews M, Flehinger B. Progress report on controlled trial of fecal occult blood testing for the detection of colorectal neoplasia. Cancer 1980;45:2959. Kewenter J, Asztely M, Engaras B. A randomized trial of fecal occult blood testing for early detection of colorectal cancer: Results of screening and rescreening of 51,325 subjects. In: Miller AB, ed. Cancer screening. New York: Cambridge University Press, 1991:118. Niv Y, Lev-El M, Abuksis G, et al. Protective effect of fecal occult blood test screening for colorectal cancer: Worse prognosis for screening refusers. Gut 2002;50:33–7. Selby JV, Friedman GD, Quesenberry CP, Weiss NS. A casecontrol study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326:653–7. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992;84:1572–5. Abuksis G, Mor M, Segal N, et al. A patient education program is cost-effective for preventing failure of endoscopic procedures in a gastroenterology department. Am J Gastroenterol 2001;96:1786 –90. Muller AD, Sonnenberg A. Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case-control study of 32,702 veterans. Ann Intern Med 1995;123:904 –10. Boland CR. Malignant tumors of the colon. In: Yamada T, ed. Textbook of gastroenterology, 3rd ed. Philadelphia: Lippincott, 1999:2023–82. Bresalier RS. Malignant neoplasms of the large intestine. In: Sleisenger & Fordtran’s gastrointestinal and liver disease, 7th ed. Philadelphia: Saunders, 2215– 61. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 2000;284:1954 –61. Sonnenberg A, Delco F, Inadomi JM. Cost-effectiveness of colonoscopies in screening for colorectal cancer. Ann Intern Med 2000;113:573–84.
Vol. 98, No. 2, 2003 ISSN 0002-9270/03/$30.00 doi:10.1016/S0002-4975(02)05926-9
Case-Control Study of Screening Colonoscopy in Relatives of Patients With Colorectal Cancer Yaron Niv, M.D., Ram Dickman, M.D., Arie Figer, M.D., Galia Abuksis, R.N., M.Sc., and Gerald Fraser, M.D. Departments of Gastroenterology and Oncology, Rabin Medical Center, Beilinson Campus, Petah Tiqwa, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
OBJECTIVES: The efficacy of colonoscopic screening and polypectomy for the prevention of colorectal cancer (CRC) is well accepted but has never been documented in a prospective, controlled study. Screening by sigmoidoscopy has been found to reduce mortality from cancer of the rectum and distal colon. Case-control studies provide an alternative method for determining the efficacy of screening methods. METHODS: Between 1998 and 2000, a total of 40 subjects were found to have CRC (study group) and 160 had a normal colon (control group) among asymptomatic individuals participating in a screening colonoscopy program for a high-risk population of first-degree relatives of CRC patients. We compared these groups for screening by fecal occult blood testing, flexible sigmoidoscopy, barium enema, and colonoscopy in the 10-yr period before the index colonoscopy. RESULTS: Screening colonoscopy was performed in only 2.5% of the case subjects and 48.7% of controls (p ⬍ 0.0001), and all screening procedures in 12.5% and 73.7%, respectively (p ⬍ 0.0001). A statistically significant difference was also found for screening with fecal occult blood test, but not for flexible sigmoidoscopy or barium enema. Significant adenomatous polyps ⬎1 cm in diameter were detected and removed in 19% of the control group within 10 yr of the index colonoscopy. Six (15%) of the patients in the study group died of CRC. CONCLUSIONS: Screening by colonoscopy can prevent progression to CRC from adenomatous polyps and may reduce the mortality associated with this devastating disease. (Am J Gastroenterol 2003;98:486 – 489. © 2003 by Am. Coll. of Gastroenterology)
INTRODUCTION Several types of colorectal cancer (CRC) screening tests have been developed for the detection of early stage disease and disease precursors. The efficacy of colonoscopic screening and polypectomy is well accepted but has never been Arie Figer, M.D., is currently located at the Department of Oncology, Souraski Medical Center, Tel Aviv, Israel.
documented in a prospective, controlled study (1). Colonoscopy is relatively costly, uncomfortable, and not without complications (2, 3), and comprehensive screening of an average-risk population may not be feasible. However, it has been recommended as a possibility by a multidisciplinary expert panel in 1997 (4). Screening by fecal occult blood testing (FOBT) has been tested in large, prospective, case-controlled studies, and a significant reduction in mortality from CRC was reported (5–9). Recently, we have demonstrated a mortality reduction of 61% in a relatively small population (10). Screening by sigmoidoscopy was shown in two retrospective case-control studies to reduce mortality from cancer of the rectum and distal colon (11, 12). In the absence of a prospective, controlled study, a retrospective, case-control design may be a good alternative for estimating the efficacy of screening tests. In the present study, we investigated high-risk patients who underwent screening colonoscopy and other tests at our center. Those who were found to have CRC were compared with those who had a normal colonoscopy for use in CRC screening tests in the 5–10 yr before the index colonoscopy. Because all participants were asymptomatic, we overcame the potential bias of possible symptom-related procedures.
MATERIALS AND METHODS Study Subjects and Protocol A special clinic for early detection and secondary prevention of CRC was established in the Gastroenterology Department of Rabin Medical Center in 1995. The average-risk population was referred for FOBT, and high-risk individuals were referred for colonoscopy. Between January 1, 1998, and December 31, 1999, our department performed 4005 colonoscopies, including 512 in asymptomatic, healthy, high-risk individuals, of whom 484 had only one first degree relative with CRC but no clinical evidence of hereditary nonpolyposis colorectal cancer (HNPCC) or adenomatous polyposis coli. In all, 284 were found to have polyps on the index colonoscopy and were excluded from the study. Of the remaining 200 individuals, 40 had CRC (study group) and 160 had a normal colon (control group). None of the screenees were referred for colonoscopy because of positive
AJG – February, 2003
FOBT. Two screenees with a normal colon were later dropped from the study because technical difficulties made it impossible to reach the cecum during the testing. Colonoscopy was performed with an Olympus 140 series colonoscope. Colon preparation was done with Sophodex (Dexxon, Israel) or Golytly (Taro, Israel) according to the manufacturer’s instructions. Before the examination the patients participated in a patient education program that was conducted by a dedicated nurse, as previously described (13). The subjects’ outpatient and family practitioner medical records for the previous 10 yr were reviewed, and information was retrieved for FOBT, flexible sigmoidoscopy (FS), barium enema (BAE), and colonoscopy. The dates of testing and the findings were recorded. The hospital archives, files of the Oncology and Pathology Departments, and endoscopy reports were also searched and the relevant data recorded. All cases and controls completed a telephone survey conducted by two of the investigators (R.D. and G.A), which covered all colonic investigations performed during the last 10 yr, and the demographic data, clinical, endoscopic, and pathological findings. Because many procedures had been performed outside of the hospital at community health facilities, these interviews added considerable information to the patient files. Death certificates registered in the Ministry of Interior were examined December 31, 2000, for all names of the individuals screened. Statistical Analysis All of our analyses considered the 5- and 10-yr period just before the index colonoscopy in the case and control subjects. Because the official records were complemented by telephone interviews (which relied on patient memory) and family physician files, we considered the data for this period to be complete. An overall OR was calculated for having had any screening test within the 5- and 10-yr period before the index colonoscopy compared with no screening test. The same calculation was then performed for each procedure (FOBT, FS, BAE, and colonoscopy). Student’s t test and comparison of two proportions were used as needed. A p value ⬍ 0.05 was considered to be statistically significant.
Case-Control Study of Screening Colonoscopy
487
Table 1. Demographic Data of High-Risk, Asymptomatic Patients With CRC and Controls
No. of patients Sex Men Women Age (yr) Average ⫾ SD Range Origin Europe-America Asia-Africa Israel More than one 1st degree relative with CRC
Study Group n (%)
Control Group n (%)
40
160
32 (80) 8 (20)
91 (57) 69 (43)
0.013 0.013
61 ⫾ 11 33–76
58 ⫾ 11 24–79
0.124
25 (62) 15 (38) 0 (0) 11 (27)
64 (40) 37 (23) 59 (37) 41 (26)
0.020 0.108 0.000 0.842
p
CI for difference – 0.565 to – 0.235, p ⬍ 0.0001). The corresponding figures for the 10-yr period (Table 3) were 2.5% for the case subjects and 48.7% of the controls (difference of – 0.462 ⫾ 0.086, 95% CI – 0.631 to – 0.293, p ⬍ 0.0001). A statistically significant difference was also found for FOBT for the two periods but not for FS or BAE. Overall, in the 10-yr period before the index colonoscopy, 12.5% of the cases had undergone any screening test (FOBT, FS, BAE, or colonoscopy) compared with 73.7% of the controls (difference of – 0.612 ⫾ 0.086, 95% CI – 0.781 to – 0.448, p ⬍ 0.0001). In the control group, an adenomatous polyp ⬎1 cm in diameter had been removed in 29 subjects within the 10 yr before the index colonoscopy. The number of colonoscopies per 160 screenees per 10 yr was 78. By contrast, only one of the study subjects (3%) underwent colonoscopy in the 10 yr before the index colonoscopy. Six patients (15%) in this group later died of CRC, as confirmed by our review of the death certificates and the records of the tumor. Colonoscopy was performed during the 10-yr period at about the same frequency as FOBT (Tables 2 and 3). In all, 43% of the cancers in the study group and 40% of the significant adenomas in the controls were located proximal to the splenic flexure. No differences in use of aspirin or nonsteroidal anti-inflammatory drugs was found between the two groups.
RESULTS Demographic data for the study participants are shown in Table 1. There were 123 men and 77 women aged 24 –79 yr. The study group included more men than did the control group (p ⫽ 0.013), as well as more individuals of Israeli or European-American origin (p ⫽ 0.000, p ⫽ 0.020). There was no statistically significant difference between the groups in age or number of first degree relatives with CRC. The screening data for the 5-yr period before the index colonoscopy are shown in Table 2. Only 2.5% of the case subjects had undergone screening by colonoscopy compared with 42.5% of the controls (difference – 0.400 ⫾ 0.084, 95%
DISCUSSION The frequency with which colonoscopies were performed during the period preceding the diagnosis of cancer in patients with CRC and controls indicates that screening by colonoscopy can prevent the progression to CRC from adenomatous polyps in high-risk patients. Of the case subjects who were found to have CRC on the index colonoscopy, only 12.5% (five of 40) had undergone any previous screening compared with 73.7% (118/160) of the controls with a normal index colonoscopy. Furthermore, the control subjects who had undergone one or more screening examina-
488
Niv et al.
AJG – Vol. 98, No. 2, 2003
Table 2. Screening During the 5-Yr Period Before the Index Colonoscopy
No. of patients FOBT FS BAE Colonoscopy
Study Group n (%)
Control Group n (%)
⌬ ⫾ SE
40 3 (7.5) 2 (5.0) 3 (7.5) 1 (2.5)
160 80 (50.0) 2 (1.2) 18 (11.3) 68 (42.5)
⫺0.425 ⫾ 0.087 0.038 ⫾ 0.025 ⫺0.038 ⫾ 0.054 ⫺0.400 ⫾ 0.084
tions by FOBT, FS, or colonoscopy during the preceding 10 yr had a 25% (29/118) risk of an adenomatous polyp. The present study is noteworthy for the careful selection of the study sample. Only asymptomatic subjects with a first degree relative with CRC, but without FAP or HNPCC according to the Amsterdam criteria (4) were included. In this manner, we were able to rule out most of the potential confounding factors, namely history of adenoma or CRC before the 10-yr review period, diagnosis of inflammatory bowel disease, HNPCC, or FAP. These factors could lead to increased screening efforts or the risk of CRC, thereby obscuring the protective effect of screening. Although information about lifestyle and level of physical activity was not available from the medical records and was not addressed in the telephone interview, it is unlikely that these factors differed significantly between the case and the control groups, as all subjects were asymptomatic screenees. Our findings are in agreement with the study of Muller et al. (14) who found that patients with CRC had had significantly fewer endoscopic procedures during periods of up to 6 yr before the onset of the disease. However, they compared a heterogeneous population of all CRC patients in their files with controls that were matched for age, sex, and race. Their sample may have included symptomatic and asymptomatic individuals, as well as those at average and high risk. In all, 43% of the cancers in the study group and 40% of the polyps in the control group were found in the right colon, and would therefore have been missed by FS if not followed by total colonoscopy. Lieberman et al. (2) and Imperiale et al. (3) examined the role of colonoscopy in screening asymptomatic adults for CRC. Almost half of the screenees with advanced proximal neoplasm had no distal polyp. Combined, the two studies show that screening sigmoidoscopy will not detect—and thus will not prevent—CRC in
95% CI for Difference ⫺0.596 ⫺0.010 ⫺0.144 ⫺0.565
to to to to
⫺0.254 0.086 0.068 ⫺0.235
p 0.000 0.361 0.680 0.000
many asymptomatic, average-risk individuals. The studies also demonstrate the feasibility of colonoscopy for screening the average-risk population, as well as the importance of investigating the right colon. The approach to patients with one first degree relative with colon cancer is not firmly established (15, 16). Whether these patients should be monitored in the same way as average-risk patients or be screened more rigorously remains to be definitively determined. Selby et al. (11) compared the use of rigid sigmoidoscopy screening over a 10-yr period in a heterogeneous sample of patients who died of CRC, along with matched controls. By contrast, in the present study, we investigated a homogeneous, asymptomatic, high-risk population to determine whether the number of colonic investigations performed in the 10 yr before the index colonoscopy is different between CRC diagnosed cases and controls who had normal colonoscopic examination. Selby et al. (11) found that only 8.8% of the 261 case subjects had undergone previous screening sigmoidoscopy, compared with 24.2% of the 868 controls; the strength of the negative association was maintained even when the most recent sigmoidoscopy was done 9 to 10 yr before diagnosis. Thus, the risk of rectal cancer was markedly reduced for at least 10 yr after a single sigmoidoscopic examination. However, for the 268 subjects in the study of Selby et al. (11) with fatal lesion above the reach of the sigmoidoscope, and for the 268 controls, no significant difference in the number of sigmoidoscopic examinations was demonstrated. As in our study, the screening sigmoidoscopy led to the removal of polyps (12 adenomatous polyps of the 210 control subjects who had at least one screening sigmoidoscopy). In a related study, Newcomb et al. (12) noted a significantly lower number of sigmoidoscopies in patients with rectal cancer than in matched controls, but there was no such difference for digital rectal examination
Table 3. Screening During the 10-Yr Period Before the Index Colonoscopy
N FOBT FS BAE Colonoscopy Any screening test NSAID or aspirin
Study Group n (%)
Control Group n (%)
⌬ ⫾ SE
40 3 (7.5) 2 (5.0) 3 (7.5) 1 (2.5) 5 (12.5) 3 (7.5)
160 80 (50.5) 2 (1.2) 26 (16.3) 78 (48.7) 118 (73.7) 25 (15.6)
⫺0.425 ⫾ 0.087 0.038 ⫾ 0.025 ⫺0.088 ⫾ 0.062 ⫺0.462 ⫾ 0.086 ⫺0.612 ⫾ 0.086 ⫺0.081 ⫾ 0.061
NSAID ⫽ nonsteroidal anti-inflammatory drugs.
95% CI for Difference ⫺0.596 ⫺0.010 ⫺0.210 ⫺0.631 ⫺0.781 ⫺0.201
to to to to to to
⫺0.254 0.086 0.034 ⫺0.293 ⫺0.443 0.039
p 0.000 0.361 0.245 0.000 0.000 0.286
AJG – February, 2003
or FOBT. This finding differs from our data in which we show a significant difference in numbers of both colonoscopies and FOBTs (p ⬍ 0.001). A trend was also found for BAE (cases 6.8%, controls 14%), but it did not reach significance. Newcomb et al. (12) reported that individuals who had a single sigmoidoscopy had an 80% lower rate of rectal cancer than individuals who were never tested. Although these investigators tried to include only tests carried out for screening, it is possible that signs or symptoms prompting the examinations were not recorded. We overcame this problem by including only asymptomatic screenees. In conclusion, screening by FOBT or colonoscopy may prevent progression to CRC from adenomatous polyps and may thereby reduce the mortality rate of this devastating disease. Thus, we believe that FOBT or colonoscopy may be an option for screening the asymptomatic, relatively highrisk population. Colonoscopy is not widely used today for this purpose, because of lack of hard data to support this strategy, and because of a lack of financial, social, and organizational resources. However, recent cost analyses have noted that screening by colonoscopy is cost-effective relative to other medical interventions (17, 18).
Case-Control Study of Screening Colonoscopy
5. 6. 7. 8. 9.
10. 11. 12. 13.
Reprint requests and correspondence to: Yaron Niv, M.D., Department of Gastroenterology, Rabin Medical Center, Beilinson Campus, Petah Tiqwa, 49100, Israel. Received July 11, 2002; accepted Sep. 30, 2002.
14. 15.
REFERENCES 1. Winawer SJ, Wolff BG, Culp CE, et al. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009 –13. 2. 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. 3. 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. 4. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer
16. 17. 18.
489
screening: Clinical guidelines and rationale. Gastroenterology 1997;112:594 –642. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993;328:1365–71. Kronberg O, Fenger C, Olsen J, et al. Randomized study of screening for colorectal cancer with fecal occult blood test. Lancet 1996;348:1467–71. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Randomized controlled trial of fecal occult blood screening for colorectal cancer. Lancet 1996;348:1472–7. Winawer SJ, Andrews M, Flehinger B. Progress report on controlled trial of fecal occult blood testing for the detection of colorectal neoplasia. Cancer 1980;45:2959. Kewenter J, Asztely M, Engaras B. A randomized trial of fecal occult blood testing for early detection of colorectal cancer: Results of screening and rescreening of 51,325 subjects. In: Miller AB, ed. Cancer screening. New York: Cambridge University Press, 1991:118. Niv Y, Lev-El M, Abuksis G, et al. Protective effect of fecal occult blood test screening for colorectal cancer: Worse prognosis for screening refusers. Gut 2002;50:33–7. Selby JV, Friedman GD, Quesenberry CP, Weiss NS. A casecontrol study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326:653–7. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992;84:1572–5. Abuksis G, Mor M, Segal N, et al. A patient education program is cost-effective for preventing failure of endoscopic procedures in a gastroenterology department. Am J Gastroenterol 2001;96:1786 –90. Muller AD, Sonnenberg A. Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case-control study of 32,702 veterans. Ann Intern Med 1995;123:904 –10. Boland CR. Malignant tumors of the colon. In: Yamada T, ed. Textbook of gastroenterology, 3rd ed. Philadelphia: Lippincott, 1999:2023–82. Bresalier RS. Malignant neoplasms of the large intestine. In: Sleisenger & Fordtran’s gastrointestinal and liver disease, 7th ed. Philadelphia: Saunders, 2215– 61. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 2000;284:1954 –61. Sonnenberg A, Delco F, Inadomi JM. Cost-effectiveness of colonoscopies in screening for colorectal cancer. Ann Intern Med 2000;113:573–84.