The Prevalence Rate and Anatomic Location of Colorectal Adenoma and Cancer Detected by Colonoscopy in Average-Risk Individuals Aged Years

Transcription

1 American Journal of Gastroenterology ISSN C 2006 by Am. Coll. of Gastroenterology doi: /j x Published by Blackwell Publishing The Prevalence Rate and Anatomic Location of Colorectal Adenoma and Cancer Detected by Colonoscopy in Average-Risk Individuals Aged Years Hana Strul, M.D., 1,3 Revital Kariv, M.D., 1,3 Moshe Leshno, M.D., Ph.D., 3,4 Aharon Halak, M.D., 1,3 Markus Jakubowicz, M.D., 1,3 Moshe Santo, M.D., 1,3 Mark Umansky, M.D., 1 Haim Shirin, M.D., 5 Ya ara Degani, MRS, 1 Miri Revivo, MRS, 2 Zamir Halpern, M.D., 1,3 and Nadir Arber, M.D., M.Sc., M.H.A. 1,2,3,4 1 Departments of Gastroenterology and 2 Cancer Prevention, Tel-Aviv Sourasky Medical Center, 3 Sackler School of Medicine and 4 Faculty of Management, Tel-Aviv University; and 5 Department of Gastroenterology, Wolfson Medical Center, Holon, Israel BACKGROUND: OBJECTIVES: METHODS: RESULTS: CONCLUSIONS: The role of screening colonoscopy for colorectal (CR) neoplasia in average-risk population, remains to be determined. To evaluate the prevalence and anatomic location of CR adenoma and carcinoma and the morbidity of colonoscopy in individuals at average risk for CR cancer (CRC). A retrospective prevalence study of subjects aged yr, with no cancer-related symptoms, personal or family history of CR neoplasia, who underwent a colonoscopy. Enrolled were 1,177 persons; 183 aged yr (young), 917 aged yr, and 77 aged yr (elderly). The prevalence of overall CR neoplasia, advanced neoplasia, and cancer was 20.9%, 6.3%, and 1.1%, respectively. In the age group, the prevalence of overall adenoma, advanced neoplasia, and cancer was 21.3%, 6.7%, and 1.2%, respectively. Of the 206 neoplasia cases, 21 43% harbored proximal neoplasia beyond the reach of sigmoidoscopy, without distal lesions. Among the elderly, the prevalence of overall adenoma, advanced neoplasia, and cancer reached 26.0%, 14.3%, and 2.6%, respectively. In the young group, 9.8% had overall neoplasia, 1.1% had advanced adenoma, and none had CRC. Procedure-related morbidity rate was 0.1%, with no perforations, bleedings, or mortality. Screening colonoscopy in average-risk subjects demonstrated a considerable prevalence of CR neoplasia and proximal lesions beyond the reach of sigmoidoscopy. The morbidity rate was negligible. Primary screening colonoscopy should be considered in health programs for the average-risk population, beginning at the age of 50 yr. The significantly high rate of advanced and proximal neoplasia in the elderly, encourages the inclusion of healthy subjects aged yr in future prospective studies. (Am J Gastroenterol 2006;101: ) INTRODUCTION Colorectal cancer (CRC) is a leading cause of cancer mortality in the Western world, with more than 1,000,000 new cases per year, an estimated lifetime risk of 5 6% and nearly 50% mortality (1, 2). Incidence rates increase sharply after the age of 50 yr. Approximately 75% of new cases occur in individuals at average risk (3). In the third millennium, as preventive medicine becomes an integral element of our concept of health, CRC screening is an increasingly important goal for health providers, physicians, and the general population. CRC fits the criteria for a disease suitable for a populationbased screening program. It is a prevalent disease that is associated with considerable mortality and morbidity rates. Second, a pre-malignant precursor lesion (i.e., adenoma) usually precedes cancer. Third, CRC has a natural history of transition from precursor to malignant lesion that lasts an average of 5 10 yr, providing a window of opportunity for screening, effective intervention, and prevention in 76 90% of cases (4 8). Finally, polypectomy has been shown to reduce the incidence and related death from CRC as it was demonstrated in the National Polyp Study that patients that are maintained adenoma free are generally kept cancer free (6, 7). Although CRC screening in the average-risk population is recommended beginning at the age of 50 yr, there is no consensus on the optimal modality for a screening health program. Colonoscopy is undoubtedly the most sensitive and specific screening test (9, 10), combining both diagnostic 255

2 256 Strul et al. and therapeutic procedures. However, the role of primary colonoscopy as a screening strategy for the general population is still subject to debate. Currently, approaches for primary screening in the average-risk population include fecal occult blood test (FOBT), sigmoidoscopy, and colonoscopy. An annual FOBT can reduce CRC mortality by up to 33% (for review see Ref. 8). CRC mortality is reduced by up to 40 50% when sigmoidoscopy is performed every 5 yr (for review see Ref. 8). The National Polyp Study demonstrated a 76 90% decrease in the incidence of CRC at 6 yr after colonoscopic polypectomy, as compared to appropriately selected controls (6, 7). Two large-scale studies assessed the use of colonoscopy for asymptomatic subjects at the age of yr. In the study by Imperiale et al. (11), 17.7% of 1,994 screenees harbored adenoma. The point prevalence rate of advanced neoplasia (defined as villous adenoma, adenoma with high-grade dysplasia or carcinoma) was 3.1% in the distal colon and 2.5% in the proximal colon. Lieberman et al. (12) examined 3,121 asymptomatic subjects and reported point prevalence rates of 1%, 10.5%, and 36.5% for CRC, advanced neoplasia, and overall adenomas, respectively. In the same cohort, the FOBT obtained on three consecutive days before bowel preparation, was positive in only 23.9% of advanced neoplasia cases (13). Moreover, 52% of patients with advanced neoplasia in the proximal colon had no distal lesion. Hence, FOBT and sigmoidoscopy would have failed to detect 76.1% and 52% of advanced neoplasia, respectively (13). Following Lieberman s study, the procedural success and complications of colonoscopy were assessed in the same study cohort (14). Colonoscopy to the cecum was successful in 97.2% of the cases and no perforation or procedurerelated mortality was reported. Major morbidity (significant rectal bleeding, myocardial infarct, cerebrovascular accident, and thrombophlebitis) occurred overall in 0.3% of the cases and in 0.1% of solely diagnostic procedures. In a recent report (15), only 37 perforations and no mortalities occurred among 116,000 colonoscopies performed in a community setting that represents the setting for CRC screening more accurately. The current study was conducted to evaluate the prevalence rate and anatomic location of colorectal (CR) neoplasia detected by a primary colonoscopy and to assess the morbidity of the procedure in subjects with no cancer-related symptoms, personal or family history of CR neoplasia. PATIENTS AND METHODS Study Design A prevalence study was conducted on subjects who underwent a complete colonoscopy in the outpatient clinic of the Gastroenterology Department in Tel-Aviv Sourasky Medical Center or one of six outpatient gastroenterological clinics of Health Medical Organization in the Tel-Aviv area. All endoscopic reports and medical files have been computerized from January 1996 and onward. We assessed all the consecutive patients who underwent a complete colonoscopy in these clinics. Data were collected retrospectively from January 1996 through December 2001 and prospectively from January 2002 through July Different computerized information sources were used including all endoscopic reports, relevant pathological reports, and patients medical files of the family doctor and the gastroenterologist in the community and in the hospital. These databases were compared and cross-examined in order to determine the inclusion/exclusion criteria for each patient. A personal questioning of the gastroenterologist was employed to clarify various forms of expressions that may be used by different doctors. Patients The study included average-risk men and women aged yr. The participants were defined as asymptomatic regarding cancer-related symptoms or alarm signs. A primary colonoscopy was initiated by the patients or their family doctors. Average-risk study population was defined as subjects having no personal or familial history of CR adenoma or carcinoma. Stratification by age groups included a main group of screenees aged yr (the acceptable age range for screening) and two smaller groups of young screenees (aged yr) and elderly screenees (aged yr). Exclusion criteria included the following: a. Suggestive symptoms of neoplasia: Rectal bleeding, change in bowel habits, abdominal pain, or unexplained weight loss during the previous six months; b. Laboratory abnormalities: A positive FOBT, elevated CEA or iron deficiency anemia (in Israel, a complete blood count is practically done in all subjects prior to routine endoscopic procedures, including screening colonoscopy. Patients with hemoglobin value below the normal range are checked for iron and/or ferritin levels); c. Imaging abnormalities: Suspected CR lesion or evaluation of metastatic spread; d. Personal history of CR adenoma or carcinoma at any time; e. Family history of CR adenoma or carcinoma in firstdegree relatives aged 70 yr or in 2 family relatives at any age; f. Inflammatory bowel disease (ulcerative colitis or Crohn s disease); g. Severe co-morbidity (e.g., malignancy, cardiopulmonary, renal or hepatic failure); h. Colonoscopy, sigmoidoscopy, or barium enema within the previous 5 yr; or i. Unsatisfactory bowel cleansing or incomplete examination, when the cecum was not intubated. Mild complaints that are not regarded as neoplastic-related symptoms were not considered as exclusion criteria: chronic constipation without change in bowel habits, a blood spot on toilet paper only, minor anal or abdominal discomfort (such as flatulence or bloating), evaluation of diverticular disease, or prior to inguinal hernioplasty.

3 The Prevalence Rate and Anatomic Location of CR Adenoma and Cancer 257 COLOUR FIG. (95%) confidence intervals were calculated for all rates. The prevalence rates of neoplastic lesions were stratified according to age groups. In addition, standard logistic regression methods were used to calculate odds ratios for neoplasia with 95% confidence intervals. In the main age group of yr, odds ratios were adjusted for age in the analysis of the prevalence of neoplasia and the risk of proximal lesions according to the findings in the distal colon. RESULTS Figure 1. Anatomic definition of the distal colon 1 (Distal colon 1): The distal colon includes the rectum, sigmoid, and descending colon up to the splenic flexure. 2 (Distal colon 2): The distal colon includes only the rectum and sigmoid colon. Colonoscopy All participants consented to and underwent a full colonoscopic examination with an adequate bowel preparation. Subjects with unsatisfactory bowel cleansing or incomplete examination were excluded, unless a complete colonoscopy was performed within 6 months, in which case the combined results of the two procedures were included. Histopathological Evaluation All lesions were sampled and specimens were sent to the Pathology Institute at the Tel-Aviv Sourasky Medical Center. Neoplastic lesions included adenocarcinoma and adenoma of any size or type. All other types of polypoid findings (such as hyperplastic or inflammatory polyps) were considered as non-neoplastic lesions. Advanced neoplasia was defined as an adenoma with a diameter 10 mm, a villous or villotubular type, an adenoma with high-grade dysplasia, or carcinoma. The prevalence of histopathological findings was determined by classification of patients according to their most advanced lesion. In order to compare the findings of sigmoidoscopy and colonoscopy, the anatomic site was analyzed according to two definitions of the distal colon (see Fig. 1). The first definition included the desired extent of screening sigmoidoscopy that should include the rectum, sigmoid, and descending colon up to the splenic flexure (distal colon 1). The second definition is the actual visualization that is achieved in many screening sigmoidoscopies (16), e.g., the rectum and sigmoid colon (distal colon 2). Statistical Analysis Database management and all statistical analyses were performed with Microsoft Excel and SPSS Version 11. Rates and proportions were calculated for categorical data, and means and standard errors for continuous data. Ninety five percent A total of 15,737 individuals underwent colonoscopy, and 1,177 eligible average-risk Israeli Jews who met the inclusion criteria were enrolled. Among the study participants the point prevalence rates of overall CR neoplasia, advanced neoplasia, and CRC were 20.9%, 6.3%, and 1.1%, respectively. The study included a main group of 917 subjects aged yr (mean age 60.9 ± 7.3 yr), 183 subjects aged yr (mean age 45.1 ± 3.0 yr), and 77 subjects aged yr (mean age 77.7 ± 1.4 yr). Table 1 summarizes selected demographic features of the study population. A summary of the overall neoplastic findings is presented in Table 2. The main age group of yr consisted of 436 males (47.5%) and 481 females (52.5%). There were 409 persons (44.6%) aged yr, 348 persons (37.9%) aged yr, and 160 persons (17.4%) aged yr. In this group, 642 persons (70.0%) had no lesions and 69 persons (7.5%) had a non-neoplastic lesion (as defined above). CRC was detected in 11 (1.2%) screenees and adenoma was found in 195 (21.3%) screenees of the main age group (Table 3). Of the 206 (22.5%) neoplastic lesions, 61 (6.7%) were advanced (as defined above): 26 (2.8%) had a tubular adenoma with a diameter of 10 mm or more, 16 subjects (1.8%) had a villous adenoma, and 8 patients (0.9%) harbored an adenoma with high-grade dysplasia. Carcinoma in situ or Dukes A was diagnosed in five patients (0.5%) and Dukes B was diagnosed in six patients (0.7%). The main age group was further stratified into three age groups (50 59, 60 69, and yr) (Table 4). Nonadvanced adenoma was found in 51 (12.5%), 54 (15.5%), and 40 (25.0%) persons in the three age groups, respectively. Advanced adenoma was detected in 10 (2.4%), 23 (6.6%), and Table 1. Selected Demographic Features of 1,177 Screenees Variable Age yr Age yr Age yr Total number Male/female, 90/93 436/481 29/48 No. (%) (49.2/50.8) (47.5/52.5) (37.7/62.3) Mean age (years) 45.1 ± ± ± 1.4 Age range yr, NA 402 (43.8) NA No. (%) Age range yr, NA 347 (37.8) NA No. (%) Age range yr, NA 168 (18.3) NA No. (%)

4 258 Strul et al. Table 2. Summary of Overall Colorectal Neoplastic Lesions Age yr Age yr Age yr (No. = 183) (No. = 917) (No. = 77) 95% CI 95% CI 95% CI Findings No. (%) Lower Higher No. (%) Lower Higher No. (%) Lower Higher Non-advanced adenoma 16 (8.7) (15.8) (14.3) Advanced adenoma 2 (1.1) (5.5) (11.7) Invasive cancer 0 (0.0) (1.2) (2.6) Advanced adenoma was defined as size 10 mm, villous type, or high-grade dysplasia. 17 (10.6%) subjects, respectively. Invasive cancer was diagnosed in two (0.5%) patients in the subgroup, seven (2.0%) in the subgroup, and another two (1.3%) in the subgroup. The odds ratio for neoplastic findings with increasing severity, adjusted for gender was 1.78 per an increase of 10 yr of age ( p < , 95% confidence interval, ). Overall, in 21% of the 206 cases with proximal neoplasia, no distal neoplasia was detected according to the definition of distal colon 1 (as defined above). However, when the definition of distal colon 2 was used, 43% of the cases with proximal neoplasia had no distal lesions (Table 5). The odds Table 3. Histopathological Features of Colorectal Lesions in the Main Age Group of yr USA (12) (No. = 3,121) Israel (No. = 917) 95% CI Findings No. (%) No. (%) Lower Higher No lesions 1,441 (46.2) 642 (70.0) Non-neoplastic 509 (16.3) 69 (7.5) lesions Total neoplastic 1,171 (37.5) 206 (22.5) lesions Overall adenoma 1,141 (36.6) 195 (21.3) Tubular adenoma Size <10 mm 842 (27.0) 145 (15.8) adenomas 687 (22.0) 131 (14.3) adenomas 112 (3.6) 11 (1.2) adenomas 43 (1.4) 3 (0.3) Size 10 mm 155 (5.0) 26 (2.8) Villous adenoma Size <10 mm 24 (0.8) 7 (0.8) Size 10 mm 69 (2.2) 9 (1.0) Adenoma with high-grade dysplasia Size <10 mm 11 (0.4) 6 (0.7) Size 10 mm 40 (1.3) 2 (0.2) Advanced adenoma 299 (9.6) 50 (5.5) Advanced neoplasia 329 (10.5) 61 (6.7) (advanced adenoma or carcinoma) Invasive cancer 30 (1.0) 11 (1.2) Carcinoma in situ 3 (0.3) Dukes A 9 (0.3) 2 (0.2) Dukes B 13 (0.4) 6 (0.7) Dukes C or D 8 (0.3) 0 (0.0) Advanced adenoma defined as size 10 mm, villous type, or high-grade dysplasia. ratios for proximal neoplastic lesions in patients with distal neoplasia were (95% confidence interval, ) and (95% confidence interval, ) according to the definition of distal colon 1 and 2, respectively (Table 6). The prevalence rate of proximal neoplasia without distal lesions was further assessed when the main group of 50- to 75- yr-old screenees was divided into two subgroups of ages yr and yr. When the definition of distal colon 1 was used, a trend of higher prevalence that did not reach statistical significance was found in the subgroup, as compared to the subgroup [39% vs 28%, respectively ( p = 0.13)]. A significantly higher prevalence rate was demonstrated when the definition of distal colon 2 was employed [60% vs 43%, respectively ( p = 0.015)]. Among the elderly (aged yr), overall CR neoplasia, advanced neoplasia, and CRC were detected in 22 subjects (28.6%), 11 subjects (14.3%), and two (2.6%) subjects, respectively (Table 7). In the young age group (aged yr), overall CR neoplasia was found in 18 (9.8%) persons: 16 persons (8.7%) with non-advanced adenoma and only two persons (1.1%) with advanced adenoma. No CRC was diagnosed in this age group (Table 8). The overall rate of colonoscopy-related morbidity in the study was 0.1% (one patient with severe abdominal pain that was hospitalized for observation). No procedure-related perforations, bleedings, or mortality were recorded. DISCUSSION This study evaluates the yield of primary screening colonoscopy in average-risk men and women aged yr Table 4. Age Stratification of Colorectal Neoplastic Lesions in the Main Age Group of yr Age yr Age yr Age yr (No. = 409) (No. = 348) (No. = 160) Findings No. (%) No. (%) No. (%) No neoplasia 346 (84.6) 264 (75.9) 101 (63.1) Non-advanced 51 (12.5) 54 (15.5) 40 (25.0) adenoma Advanced adenoma 10 (2.4) 23 (6.6) 17 (10.6) Invasive cancer 2 (0.5) 7 (2.0) 2 (1.3) Advanced adenoma was defined as size 10 mm, villous type, or high-grade dysplasia.

5 The Prevalence Rate and Anatomic Location of CR Adenoma and Cancer 259 Table 5. Risk of Proximal Neoplasia According to Distal Neoplasia in the Age Group of yr Overall Proximal Total Number Neoplasia No. (%) Distal finding 1 No neoplastic lesion (8.0) Non-advanced adenoma (17.8) Advanced neoplasia (22.2) Distal finding 2 No neoplastic lesion (14.1) Non-advanced adenoma 51 9 (17.6) Advanced neoplasia (31.6) Distal colon 1 defined as rectum, sigmoid, and left colon until the splenic flexure; distal colon 2 defined as rectum and sigmoid colon; advanced neoplasia was defined as carcinoma or advanced adenoma (defined as size 10 mm, villous type, or high-grade dysplasia). with no cancer-related symptoms, personal or family history of CR adenoma or cancer. Overall CR neoplasia was detected in 28.6%, 22.5%, and 9.8% of the elderly age group (76 80 yr), the main age group (50 75 yr), and the younger group (40 49 yr), respectively. All the CRC cases were diagnosed at an early and curable stage (Dukes A or B). In the main age group (50 75 yr), analysis of the proximal findings without distal lesions demonstrated that sigmoidoscopy would have failed to detect 21 43% of overall CR neoplasia. The prevalence rate of proximal neoplasia without distal lesions was higher in the subgroup of age yr as compared to the subgroup of age yr. This rate was significantly higher when the definition of distal colon 2 was employed (60% vs 43%, respectively, p = 0.015). These findings suggest a possible proximal shift of neoplastic lesions in older age groups and support the use of colonoscopy rather than sigmoidoscopy as a screening modality in this age group. In the yr age group, overall neoplasia, advanced neoplasia, and CRC prevalence was 22.5%, 6.7%, and 1.2%, respectively. The prevalence of adenoma and advanced neoplasia was comparable to the results of Imperiale et al. (11). Although the prevalence of adenoma in our study was lower than the one reported by Lieberman et al. (12), a similar rate of CRC was detected. A possible explanation for this discrepancy is the study group definition. The current study population had a similar male/female ratio, while in Lieberman s study the vast majority (97%) of screenees were men Table 6. Odds Ratio for Proximal Neoplasia According to the Findings of Neoplasia in the Distal Colon 95% CI for OR B Sig. OR Lower Higher Distal colon Age Distal colon Age B = logistic regression coefficient. Distal colon 1 defined as rectum, sigmoid and left colon until the splenic flexure and distal colon 2 defined as rectum and sigmoid colon. Table 7. Histopathological Features of Colorectal Lesions in the Age Group of yr (Elderly) 95% CI Findings No. (%) Lower Higher No lesions or non-neoplastic lesions 55 (71.4) Total neoplastic lesions 22 (28.6) Overall adenoma 20 (26.0) Tubular adenoma Size <10 mm 11 (14.3) adenomas 10 (13.0) adenomas 1 (1.3) adenomas 0 (0.0) Size 10 mm 6 (7.8) Villous adenoma Size <10 mm 1 (1.3) Size 10 mm 0 (0.0) Adenoma with high-grade dysplasia Size <10 mm 1 (1.3) Size 10 mm 1 (1.3) Advanced adenoma 9 (11.7) Advanced neoplasia (advanced 11 (14.3) adenoma or carcinoma) Invasive cancer 2 (2.6) Advanced adenoma was defined as size 10 mm, villous type, or high-grade dysplasia. (12). Moreover, since persons with a family history of CR adenoma or cancer were excluded from our study, our participants represent average-risk subjects. Previous studies included asymptomatic but not necessarily average-risk subjects, since a family history of CR neoplasia did not exclude them from the study: 14% of participants in Lieberman s study had a family history of CRC in first-degree relatives and a family history of CR adenoma was not recorded (12). In Imperiale et al. s study, family history was not obtained at all (11). Other possible explanations for the lower rate of adenoma in our study may be related to differences between the study populations: ethnic variation, higher rates of malignant conversion from adenoma to carcinoma, or the higher rates of flat adenomas in Israel. Indeed, in previous earlydetection studies conducted in Israeli Jews, a prevalence rate Table 8. Histopathological Features of Colorectal Lesions in the Age Group of yr (Young) USA (20) (No. = 917) Israel (No. = 183) 95% CI Findings No. (%) No. (%) Lower Higher No lesions or 806 (88.9) 165 (90.2) non-neoplastic lesions Non-advanced 79 (8.7) 16 (8.7) adenoma Advanced adenoma 32 (3.5) 2 (1.1) Invasive cancer 0 (0) 0 (0.0) Advanced adenoma was defined as size 10 mm, villous type, or high-grade dysplasia.

6 260 Strul et al. of 10 11% for adenoma and 1% for CRC was detected. Nevertheless, the prevalence rate may be affected by the protocol of these studies, which comprises FOBT and sigmoidoscopy as the primary screening modalities (while colonoscopy was performed only in positive cases or in high-risk patients) (17, 18). Our main age group was further stratified into three age subgroups (50 59, 60 69, and yr). As expected, the prevalence and the severity of neoplastic changes from adenoma, through advanced adenoma to invasive cancer were significantly related to aging with an odds ratio of 1.78 per an increase of 10 yr of age ( p < , 95% confidence interval, ). The current study presents the first report of screening colonoscopy in a small group of elderly healthy individuals aged yr at average risk for CRC. Overall CR neoplasia was detected in 28.6%. The prevalence rates of advanced neoplasia and CRC (14.3% and 2.6%, respectively) were twice as high as the rates in the main age group of yr. Since the life expectancy of a 70-yr-old person with no functional limitations is 14.3 yr (19), these high prevalence rates should encourage the inclusion of healthy subjects of this age group in future studies of screening programs. In the young age group (40 49 yr), overall CR neoplasia was detected in 9.8% of the subjects, advanced adenoma in 1.1% of the subjects, and no cases of CRC were detected. These findings are similar to those of a recent study published by Imperiale et al. (20) and support the conclusion that screening before the age of 50 yr is not warranted in the general population. It is important to emphasize that screening colonoscopy in the average-risk population was safe and had a very low morbidity rate (0.1%). No procedure-related mortality was recorded. Comparable results were reported by other studies (14, 15). Possible limitations to our study are the prevalence nature of the study and the retrospective collection of data. Yet, meticulous efforts were made to minimize information/misclassification biases. The medical files, lab results, colonoscopic, and pathologic reports of all the patients are fully computerized and were completely assessed and double checked. Data from three different sources was crossexamined and verified: (a) the medical files of the family doctor, (b) the medical files of the gastroenterologist, and (c) the colonoscopic report. The patient s symptoms, signs, family history, and lab results are an integral part of the medical files of both the family physician and the gastroenterologist. Relevant data are also routinely documented as an essential part of the written indications in the colonoscopy report. We trust that the quality of the data is adequate. Regarding a potential information bias in the recollection of family history (and in particular the recollection of adenomatous polyps in relatives): The data on family history were cross-examined from several information sources. The largest database in our cancer prevention unit includes a detailed family history that is verified in many cases by receiving pathological reports of relevant family relatives. In cases of doubt, when the nature of polyps in family members was uncertain, we have chosen to exclude these patients. This approach may have excluded average-risk patients and decreased the total number of the study group. Potential differential bias in data collection or in patients inclusion was also decreased since the decision of inclusion or exclusion was based on the indications for colonoscopy. This decision was performed blindly and independently before the screener knew the results of the procedure. The possibility of selection biases include the following: a. Referral bias: The colonoscopies were performed in six outpatient gastroenterological clinics of Health Medical Organization in the greater Tel-Aviv area and the outpatient clinic of the Gastroenterology Department in Tel-Aviv Sourasky Medical Center (which serves as an ambulatory clinic for HMO patients). The participating gastroenterological and general medical clinics serve patients from a wide range of educational and socioeconomical states. The participants were offered colonoscopy for CRC screening by their family physician and some were self-referred. These are accepted approaches that are routinely employed for other screening modalities in Israel (such as FOBT or mammography). Subjects who were self-referred for colonoscopy and subjects who were regularly followed-up with family physicians, may be more health conscious as compared to the general population. These patients may also have healthier life-style habits. Therefore, our prevalence estimation of neoplasia may be underestimated and the true prevalence in the general population may be even higher. b. Berksonian bias: Hospitalized patients or patients with severe co-morbidities were excluded from the study. Indeed, one of the study limitations is that it was not conducted as a part of a prospective screening program. The current study may not adequatly represent the compliance rate of the general population. Potential confounders are as follows: a. Gender: Our participants included a similar ratio of men and women, although gender is not an established risk factor for CRC. b. Age: Patients were stratified into age groups and the results were analyzed accordingly. c. Socioeconomic status: The participating clinics represent a wide socioeconomic range, although it has not been reported as a risk factor for CRC. d. Life habits (e.g., cigarette smoking, alcohol consumption, or high BMI) were not assessed in the current study. As mentioned above, subjects who were self-referred may have healthier life-style habits. This possible confounder may in fact contribute to an underestimation of the prevalence rate of neoplasia.

7 The Prevalence Rate and Anatomic Location of CR Adenoma and Cancer 261 It is worthwhile to note that there is no proven association between these factors and the colonic location of neoplasia that might affect the detection rate of colonoscopy versus sigmoidoscopy. e. A personal or family history of CR adenoma or carcinoma was an exclusion criterion in our study but not in previous reports (11, 12). In a young age group (40 49 years), a low rate of adenoma was found without any cases of colon cancer similar to previous reports from the United States. This supports the conclusion that screening before the age of 50 is not warranted in the general population. Hence, our study participants may adequately represent the average-risk population with a reasonable generalizability of our prevalence estimates. The current study does not prove the effect of colonoscopy on cancer development or cancer-related mortality. Previous prospective studies [such as the National Polyp Study (6, 7)] have clearly demonstrated that adenoma removal significantly reduces the incidence of CRC. Although it is logic to conclude that prevention of cancer development should also affect cancer-related mortality, this effect has not been proved directly so far. Our study, being a prevalence study in its nature, cannot address this effect. In summary, the current study suggests that screening colonoscopy in average-risk, symptom-free population detects a considerable prevalence of CR neoplasia, including proximal neoplasia. These data support the current recommendations to consider primary screening colonoscopy as a part of the standard health package for the asymptomatic average-risk population. It should be offered beginning at the age of 50 yr. The significantly high rates of overall advanced neoplasia and proximal neoplasia in the elderly, encourage the inclusion of healthy subjects aged yr, in future prospective studies. The precise upper age range and the optimal interval between colonoscopies should be further assessed. STUDY HIGHLIGHTS What Is Current Knowledge Many countries have resisted instituting recommendations for screening colonoscopy for colon cancer. What Is New Here In average-risk Israeli Jews, there was a similar rate of cancer and a lower prevalence of adenoma, as compared to previous reports from the USA. Screening colonoscopy in a small group of elderly healthy individuals, ages 76-80, at average-risk for colon cancer showed the prevalence rates of advanced neoplasia and colon cancer were twice as high as the rates in the age group of years. There was a possible proximal shift neoplastic lesions in older ages, supporting the use of colonoscopy, rather than sigmoidoscopy, as a screening modality in this age group. Reprint requests and correspondence: Nadir Arber, Prof. of Medicine and Gastroenterology, Director Integrated Cancer Prevention Center, Tel-Aviv Sourasky Medical Center, 6 Weizmann Street, Tel-Aviv 64239, Israel. Received December 15, 2004; accepted August 19, REFERENCES 1. Jemal A, Tiwari RC, Murray T, et al. Cancer statistics CA Cancer J Clin 2004;54: Peeters M, Haller DG. Therapy for early stage colorectal cancer. Oncology 1999;13: Burt RW. Colon cancer screening. Gastroenterology 2000;119: Morson B. President s address. The polyp-cancer sequence in the large bowel. Proc R Soc Med 1974;67: Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61: Winawer SJ, Zauber AG, O Brien MJ. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. N Engl J Med 1993;328: Winawer SJ, Zauber AG, Ho MN, et al. The National Polyp Study Workgroup. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993;329: Lieberman DA. Cost-effectiveness model for colon cancer screening. Gastroenterology 1995;109: Hixson LJ, Fennerty MB, Sampliner RE, et al. Prospective blinded trial of the colonoscopic miss-rate of large colorectal polyps. Gastrointest Endosc 1991;37: Macafee DA, Scholefield JH. Antagonist: Population based endoscopic screening for colorectal cancer. Gut 2003;52: 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: Lieberman DA, Weiss DG, Bond JH, et al. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative group 380. N Engl J Med 2000;343: Lieberman DA, Weiss DG. One-time screening for colorectal cancer with combined fecal occult blood testing and examination of the distal colon. Veterans Affairs Cooperative group 380. N Engl J Med 2001;345: Nelson DB, McQuaid KR, Bond JH, et al. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55: Korman LY, Overholt BF, Box T, et al. Perforation during colonoscopy in endoscopic ambulatory surgical centers. Gastrointest Endosc 2003;58: Jensen J, Kewenter J, Swedenborg J. The anatomic range of examination by fiberoptic rectosigmoidoscopy. Scand J Gastroenterol 1992;27:842 4.

8 262 Strul et al. 17. Rozen P. Screening for colorectal neoplasia in the Tel-Aviv area: Cumulative data and initial conclusions. Isr J Med Sci 1992;28: Odes HS, Rozen P, Ron E, et al. Screening for colorectal neoplasia: A multicenter study in Israel. Isr J Med Sci 1992;28: Lubitz J, Cai L, Kramarow W, et al. Health, life expectancy and health care spending among the elderly. N Engl J Med 2003;349: Imperiale TF, Wagner DR, Lin CY, et al. Results of screening colonoscopy among persons 40 to 49 years of age. N Engl J Med 2002;346: