Screening for Colorectal Cancer in Chinese: Comparison of Fecal Occult Blood Test, Flexible Sigmoidoscopy, and Colonoscopy

Transcription

1 GASTROENTEROLOGY 2003;124: Screening for Colorectal Cancer in Chinese: Comparison of Fecal Occult Blood Test, Flexible Sigmoidoscopy, and Colonoscopy JOSEPH J. Y. SUNG,* FRANCIS K. L. CHAN,* WAI K. LEUNG,* JUSTIN C. Y. WU,* JAMES Y. W. LAU, JESSICA CHING,* KA F. TO, YUK T. LEE,* YIU W. LUK, NELSON N. S. KUNG, SAMUEL P. Y. KWOK, # MICHAEL K. W. LI,** and S. C. SYDNEY CHUNG Departments of *Medicine and Therapeutics, Surgery, and Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong; Departments of Medicine and **Surgery, Pamela Youde Nethersole Eastern Hospital, Hong Kong; and Departments Medicine and # Surgery, United Christian Hospital, Hong Kong Background & Aims: Fecal occult blood testing (FOBT), flexible sigmoidoscopy (FS), and colonoscopy are the most commonly recommended screening tests for colorectal cancer. The aim of this study was to compare the accuracy and safety of these 3 screening procedures in a general population of ethnic Chinese. Methods: Asymptomatic adults older than 50 years were recruited from the general public through health exhibitions. All enrolled subjects were offered FOBT and full colonoscopy under sedation. Advanced colonic lesions (defined as adenoma >10 mm, villous adenoma, adenoma with moderate or severe dysplasia, or invasive cancer) were recorded. Lesions at the distal 40 cm in the left colon and rectum were taken as findings of FS. Results: A total of 505 subjects (56% women; mean age SD, years) were enrolled, and 476 (94.3%) had a complete colonoscopy. Advanced colonic neoplasms were documented in 63 subjects (12.5%), of which 45 had lesions in the distal colon and 26 in the proximal colon. Among the 385 subjects with a normal distal colon, 14 (3.6%) had advanced lesions in the proximal colon that would be missed by FS alone. The sensitivity and specificity of FOBT for advanced colonic lesions were 14.3% and 79.2% and the sensitivity and specificity of FS were 77.8% and 83.9%, respectively. Combining FOBT with FS would not significantly improve the results of FS alone. Among these 505 subjects who underwent colonoscopy and 148 who underwent polypectomy, there was no perforation and only one occurrence of postpolypectomy bleeding recorded. Conclusions: Colonoscopy is a safe and accurate method for the screening of colorectal neoplasms in Chinese subjects. Colorectal cancer is the second leading cause of cancer death in the western world, and the incidence is increasing at an alarming rate among Chinese people. 1,2 The dramatic increase in cases of colorectal cancer is largely attributed to the increase in the number of people older than 50 years; incidence among the young has remained unchanged. Although screening with fecal occult blood testing (FOBT) 3,4 and flexible sigmoidoscopy (FS) 5,6 has been shown to reduce mortality from colorectal cancer in the West, no screening study in Asia has ever been reported. FOBT is limited by its low sensitivity of detecting polyps and frequent false-positive results. On the other hand, sigmoidoscopy can only reach the splenic flexure at best. It is not certain whether subjects found to have small adenomatous polyps at the distal colon would require more extensive investigation. A few reports have indicated that some advanced proximal neoplasms are missed if sigmoidoscopy is used as a primary screening procedure. 7,8 Colonoscopy has the advantage of allowing biopsy and/or removal of lesions during the same procedure. The marginal benefit of performing colonoscopy compared with other screening approaches is yet to be determined. In a follow-up of the VA study, Lieberman et al. showed that even combining FOBT with FS detects only about 75% of advanced lesions diagnosed by a full colonoscopy. 9 About one half of the proximal lesions would have been missed in the absence of a full colonoscopy. The screening population in the VA study was predominantly American men. Thus far, there are few data on the screening of colorectal cancer among Asian subjects. The aim of the present study was to assess the accuracy and safety of using FOBT, sigmoidoscopy, and combined FOBT with sigmoidoscopy compared with colonoscopy alone for screening of colorectal neoplasms in averagerisk ethnic Chinese subjects older than 50 years. Materials and Methods Study Subjects Asymptomatic subjects were recruited from the general public by conducting health exhibitions in community Abbreviations used in this paper: FOBT, fecal occult blood test; FS, flexible sigmoidoscopy by the American Gastroenterological Association /03/$30.00 doi: /gast

2 March 2003 SCREENING FOR CRC IN CHINESE SUBJECTS 609 centers and shopping malls. In the exhibitions, information about the increasing incidence of colorectal cancer, nature of the disease, various methods of screening, and procedure of colonoscopy in detail was introduced to the public. Interested individuals were invited to join the screening program on a voluntary basis, and their names and telephone numbers were collected during the exhibition. Each subject was given bottles for the collection of 3 consecutive stool samples. They were then referred to one of the participating hospitals for scheduling of a colonoscopy. Subjects who voluntarily entered the study as recommended by their relatives or friends were also accepted. The 3 participating hospitals were Prince of Wales Hospital, United Christian Hospital, and Pamela Youde Nethersole Eastern Hospital, each serving a community with a population of million. Only ethnic Chinese subjects older than 50 years were included. Subjects were excluded if they met the following criteria: (1) a personal or family history of colorectal cancer or polyps, (2) a history of screening tests (FOBT, barium studies, or endoscopy) performed within the past 5 years for colorectal cancer, (3) a history of altered bowel habits or rectal bleeding, (4) a known history of diverticular disease or inflammatory bowel diseases, (5) severe premorbid illnesses that might increase the risk of performing colonoscopy (e.g., cardiopulmonary insufficiency, bleeding disorders, or cirrhosis), (6) a history of prosthetic heart valve or vascular graft surgery, or (7) a history of using anticoagulants. This study was approved by the Clinical Trial Ethics Committee of the Chinese University of Hong Kong and the ethics committee of each participating hospital. Before undergoing colonoscopy, the subjects were interviewed by a research nurse using a structured questionnaire. The following information was recorded: (1) demographic data (age, sex, and income), (2) personal habits (smoking and drinking), (3) medical history (history of cancer, surgery, intake of aspirin or nonsteroidal anti-inflammatory drugs, or comorbid illness), (4) family history of colorectal cancer at age younger than 50 years or colorectal cancer among first-degree relatives, (5) alarming bowel symptoms (including change in bowel habits, rectal bleeding, tenesmus, or chronic diarrhea), and (6) number of bowel motions per week. Subjects were asked specifically about changes in bowel habits, rectal bleeding, tenesmus, weight loss, and anorexia, which might suggest underlying colorectal diseases. The possible risks of colonoscopy were explained to the subjects, and informed consent was obtained before enrollment in the study. FOBT Subjects were on an unrestricted diet because it would be impractical to eliminate peroxidase or pseudoperoxidasecontaining foods in the Chinese diet. Three consecutive stool samples were tested for the presence of occult blood before colonoscopy was performed. Stool samples were collected before bowel preparation for colonoscopy. The guaiac-based Hemoccult II (SmithKline Diagnostics Corp., San Jose, CA) was used to test unhydrated stool samples. Subjects were to return the Hemoccult II tests on the day of colonoscopy and were classified as FOBT positive if any of the 3 samples produced a positive result. Subjects underwent colonoscopy irrespective of the findings on FOBT. The tests were performed by trained nurses. The results of FOBT were not available to the endoscopists who performed the procedure. Colonoscopy Colonoscopy was performed as an outpatient procedure, with the subject under conscious sedation using intravenous diazepam and pethidine. The study investigators, who were selected because of their extensive experience in endoscopy, performed all colonoscopic procedures. Subjects were given phosphosoda (Fleet; CB Fleet, Lynchburg, VA) for bowel preparation together with instructions for use. A complete examination was defined as an endoscope reaching the cecum as documented by taking a picture of the ileocecal valve. If colonoscopy could not be completed because of unsatisfactory bowel preparation or bowel looping, subjects were given the option of repeating the examination at a later stage or undergoing examination by barium enema. All lesions found were photographed and documented for their site (distance from anal verge) and size (measured by biopsy forceps that opened up to 7 mm). Findings at the distal colon 40 cm from the anal verge on withdrawal of the colonoscope were taken as lesions found at FS. Subjects who underwent endoscopic polypectomy were admitted for observation for 24 hours and were discharged after confirmation of the absence of overt bleeding or a decrease in hemoglobin level after endoscopic polypectomy. Procedure-related complications and hospitalization were documented. Histology All colonoscopic biopsy samples were examined by a single pathologist (K.F.T.) at the Prince of Wales Hospital who is experienced in colonic pathology. Polyps were classified according to World Health Organization recommendations. 10 The most advanced lesions in the entire colon were determined. Subjects were classified according to the most advanced lesion in the proximal colon (at or above the splenic flexure) and distal colon (below the splenic flexure) for the documentation of pathology. Advanced colonic neoplasm was defined as an adenoma 10 mm in diameter, a villous adenoma (with at least 25% villous architecture), an adenoma with moderate to severe dysplasia, or invasive carcinoma. We included adenoma with moderate dysplasia as advanced colonic neoplasm based on the fact that their potential of developing into invasive cancer is similar that of an adenoma 1 cm in diameter. 11 Statistics Results were expressed as proportions, means, and SE where appropriate. 2 test was used to compare proportions. Mantel Haenszel method was used to adjust for the effect of age (SPSS 10.0; SPSS Inc., Chicago, IL). All P values were 2 tailed. Statistical significance was taken as P 0.05.

3 610 SUNG ET AL. GASTROENTEROLOGY Vol. 124, No. 3 Table 1. Characteristics of the 505 Subjects Age mean SD (yr) (72.5%) (26.3%) (1.2%) Sex Male 224 (44%) Female 281 (56%) Smoking Never packs/wk packs/wk 15 8 packs/wk 2 Stopped smoking 67 Drinking Never 363 Social drinker 131 Regular drinker 1 Daily drinker 17 Unknown 3 Recruitment method Health exhibition 427 Recommended by friends/family 78 Results Of the 510 subjects who were voluntarily recruited, 505 met the criteria and were enrolled. Subjects included 224 men and 281 women (mean age SD, years). Among these subjects, 12 were regular users of aspirin and 36 users of nonsteroidal antiinflammatory drugs. Characteristics of the subjects are summarized in Table 1. Five subjects were excluded from the study because of a history of per-rectal bleeding or a family history of colorectal cancer. All subjects underwent FOBT and colonoscopy. A total of 476 subjects (94.3%) underwent a colonoscopy reaching the cecum on the first attempt. The remaining 29 subjects had a subsequently completed examination by a second colonoscopy (n 21) or barium study (n 8). Among the 476 subjects with a successful colonoscopy on the first attempt, there were 148 cases of colorectal polyps, 25 cases of diverticular disease, 45 cases of hemorrhoids, 6 cases of colorectal ulcers, and 4 cases of invasive cancer diagnosed. A total of 120 subjects (23.7%) had neoplastic lesions (including 117 polyps and 3 invasive cancers) found at the distal colon below the level of the splenic flexure, and 59 (11.7%) had neoplastic lesions (including 58 polyps and 1 invasive cancer) found in the proximal colon at or above the splenic flexure (Table 2). A total of 148 subjects underwent endoscopic polypectomy during colonoscopy (29.3%). There was no perforation of bowel and only one case of postpolypectomy bleeding that required a blood transfusion. Only subjects who underwent a successful colonoscopy on the first attempt were analyzed for the results. Advanced colonic lesions were found in 63 subjects (12.5%); 37 were men, and 26 were women (mean age, 57.4 years). There were 4 invasive cancers (2 in the sigmoid colon, one in the descending colon, and one in the cecum), 35 tubular adenomas with moderate or severe dysplasia, 15 large ( 1 cm) tubular adenomas, and 9 villous adenomas. Forty-five (8.9%) were distal to the splenic flexure and 26 (5.1%) in the proximal colon (8 subjects had advanced lesions in both the proximal and distal colon) (Table 2). Among the 385 subjects with a normal distal colon, 14 (3.6%) had advanced lesions in the proximal colon. On the other hand, advanced proximal lesions were found in 12 of 120 subjects (10%) who had polyps at the distal colon. In total, 14 of 26 subjects (53.8%) with advanced lesions in the proximal colon had a normal distal colon that would not be detected by FS otherwise. A total of 101 subjects (20%) had at least one positive FOBT (Table 3). Colonic lesions were found in 29 cases Table 2. Findings According to the Most Advanced Lesions Rectosigmoid (A) Descending colon (B) Distal colon a (A B) Transverse colon (C) Ascending colon (D) Cecum (E) At or above splenic flexure a (C D E) Overall a (A B C D E) Nonadenomatous polyp Hyperplastic polyp Tubular adenoma (mm) (21) 17 (9) 59 (26) 12 (2) 19 (11) 6 (3) 36 (16) 78 (35) Villous adenoma (mm) Invasive cancer Total no. of polyps Total no. of advanced lesions NOTE. Numbers in parentheses represent tubular adenoma ( 10 mm) showing moderate or severe dysplasia. Advanced lesions included large tubular adenoma ( 10 mm), tubular adenoma with moderate or severe dysplasia, villous adenoma, and invasive cancers. a Lesions found in more than 1 site were counted once.

4 March 2003 SCREENING FOR CRC IN CHINESE SUBJECTS 611 Table 3. Results of FOBT in Relation to Endoscopic Findings FOBT positive (n 101) FOBT negative (n 404) Total (n 505) Normal colonoscopy Nonadenomatous polyp Proximal colon Distal colon Hyperplastic polyp Proximal colon Distal colon Tubular adenoma Proximal colon Distal colon Villous adenoma Proximal colon Distal colon Invasive cancer Proximal colon Distal colon (28.7%), and 9 (8.9%) were advanced lesions. Among the 404 subjects with negative FOBT, 123 (30.4%) had colonic lesions, including 54 (13.4%) with advanced lesions. The sensitivity, specificity, positive predictive value, and negative predictive value of FOBT for all kinds of colonic neoplasms were 19.1%, 79.6%, 28.7%, and 69.6%, respectively. The sensitivity, specificity, positive predictive value, and negative predictive value of FOBT for advanced colonic lesions were 14.3%, 79.2%, 8.9%, and 86.6%, respectively. Subjects with a positive test result did not have a significant increase in the risk of having advanced lesions compared with those with negative FOBT (relative risk, 0.63; 95% confidence interval, ). Overall, there were 101 subjects with positive FOBT, 120 with neoplastic lesions in the distal colon (that would have been detected by FS), and 198 with positive FOBT and/or neoplastic lesions found in the distal colon in this study. Assuming that these subjects were to undergo a full colonoscopic examination, the sensitivity, specificity, positive predictive value, and negative predictive value of the screening strategy, either alone or in combination, in detecting advanced colonic lesions are listed in Table 4. The number of procedures (including FS and colonoscopy) required to detect one advanced neoplastic lesion using FOBT alone as a screening test would be 11.2 (101 colonoscopies to detect 9 advanced lesions). We have also estimated the number of procedures required to detect one advanced lesion using FS as a screening test. Forty-five subjects had advanced lesions in the distal colon that could be detected by FS. In 4 subjects, distal lesions were found in association with proximal advanced lesions. These cases would have been picked up by colonoscopy triggered by positive FS. In total, 49 cases of advanced lesions would be diagnosed using FS as a screening test. Hence, the number of procedures required to detect one advanced lesion using FS alone as a screening test would be 12.8 (505 FS and 120 colonoscopies to detect 49 advanced lesions). If combining FS and FOBT in the screening strategy and offering colonoscopy to those with either test yielding a positive result, the number of procedures required would be 13.8 if sigmoidoscopy was performed first or 11.8 if FOBT was performed first. If all subjects underwent direct colonoscopy without a prior screening procedure, it would only take 8.0 colonoscopies to detect one advanced colonic lesion. On the other hand, using FOBT alone, FS alone, or combined FS with FOBT as screening tests, 85.7%, 22.2%, and 19.0% of advanced colonic lesions would be missed, respectively. The number of colonoscopies required to detect one of these missed diagnoses using FOBT alone, FS alone, or combined FS with FOBT was 7.5, 27.5, and 25.6, respectively. Thus, adding FOBT to FS did not improve the sensitivity of FS. The sensitivity of the screening program was further assessed in relation to the sex and age of the subjects. The prevalence of all colonic neoplasms and advanced lesions was significantly higher in men than in women (37.5% vs. 24.2% and 16.5% vs. 9.3%, respectively) (Table 5). Table 4. Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value of FOBT, FS, and Combined Screening for Advanced Colonic Neoplasia Screening program No. of patients with a positive test Sensitivity for advanced Specificity for advanced Positive predictive value for advanced Negative predictive value for advanced No. of procedures needed to find one advanced FOBT alone (9/63) 79.2 (350/442) 8.9 (9/101) 86.6 (350/404) 11.2 (101/9) FS alone (49/63) 83.9 (371/442) 40.8 (49/120) 96.4 (371/385) 12.8 (625/49) Combined FS and FOBT (51/63) 66.7 (295/442) 25.0 (51/198) 96.1 (295/307) 13.8 (703/51) a 11.8 (602/51) b a Screening with FS and, if negative, followed by FOBT. Colonoscopy if either of the tests yields a positive result. b Screening with FOBT and, if negative, followed by FS. Colonoscopy if either of the tests yields a positive result.

5 612 SUNG ET AL. GASTROENTEROLOGY Vol. 124, No. 3 Table 5. Sensitivity of Screening Program in Relation to Sex and Age Overall (n 505) Male (n 224) Female Age yr (n 281) P a (n 366) Age 60 yr (n 139) P Prevalence of all (%) 152/505 (30.1) 84/224 (37.5) 68/281 (24.2) /366 (26.5) 55/139 (39.6) Prevalence of advanced (%) 63/505 (12.5) 37/224 (16.5) 26/281 (9.3) /366 (10.9) 23/139 (16.5) 0.09 Sensitivity of FOBT (%) 9/63 (14.3) 6/37 (16.2) 3/26 (11.5) /40 (20) 1/23 (4.3) 0.14 Sensitivity of FS (%) 49/63 (77.8) 30/37 (81.1) 19/26 (73.1) /40 (77.5) 18/23 (78.3) 0.94 Sensitivity of FOBT in combination with FS (%) 51/63 (81) 32/37 (86.5) 19/26 (73.1) /40 (80) 19/23 (82.6) 1.00 a After adjustment for age using the Mantel Haenszel method. However, there was no difference between men and women in the sensitivity of FOBT alone, FS alone, or combined FOBT and sigmoidoscopy in detecting advanced colonic lesions. On the other hand, there were significantly more colonic neoplasms found in subjects older than 60 years and a trend toward more advanced colonic lesions in this group. However, no significant improvement in sensitivity or specificity of the screening programs (FOBT alone, FS alone, and combined FOBT and sigmoidoscopy) could be shown between the group older than and younger than 60 years. Discussion The 3 most commonly used screening tests for colorectal cancer reported in the literature are FOBT (hydrated and unhydrated), FS, and colonoscopy. FOBT is the only test shown in randomized studies to reduce the mortality of colorectal cancer. 3,4 While awaiting results from a prospective, randomized study, a goodquality analysis from a case-control study also suggested that FS reduces the mortality of cancer. 6 The use of colonoscopy as a screening procedure was only recently introduced. The procedure is reported to be safe and accurate in diagnosing colorectal cancer as well as premalignant lesions in the colon. 7,8 Despite the lack of data on survival benefit, colonoscopy is still an attractive option because it can be used to examine the entire colon and hence obviates the need for further screening tests for the next 5 10 years. Furthermore, colonoscopy seems to outperform FOBT and FS in the accuracy of diagnosing advanced colonic lesions in the VA study. 9 There has been a rapidly increasing trend in recent decades in the incidence of colorectal cancer in Asian countries, including China, Singapore, and Japan. 1,2,12,13 Age-period-cohort analysis suggests significant effects of birth cohort on the risk of developing this malignant disease, which has already become the second most common cancer in some of these countries. This phenomenon, which is more evident among subjects older than 50 years, has been attributed to environmental changes in the postwar period since the 1950s. With increasing affluence of the societies, westernization of lifestyle and changes in dietary habits have been identified as the most likely causes of the changing pattern of disease. Despite the alarming change in epidemiology, there are no published studies evaluating a screening program for colorectal cancer in Asia. Our study design differs from previous colonoscopy screening studies 7,8 in that we targeted a different ethnic group with exclusively Chinese subjects, included both men and women, and enrolled only asymptomatic, average-risk individuals. Subjects with a family history of colorectal cancers or those referred for the investigation of lower gastrointestinal symptoms were excluded. Overall, 12.5% of subjects in our study were found to have advanced colonic lesions compared with 10.5% reported by Lieberman et al. 8 and 5.6% reported by Imperiale et al. 7 It should be noted that the definition of advanced lesions used in these 3 studies is slightly different. Lieberman et al. included adenomas with high-grade (which is equivalent to severe) dysplasia and adenomas 10 mm as advanced lesions. Imperiale et al. included only adenomas with high-grade dysplasia irrespective of size of the lesion. We defined advanced lesions as adenomas with moderate to severe dysplasia or adenomas 10 mm. Applying the most stringent criteria of Imperiale et al., 5.5% of the VA study 7 and 4.4% of our Chinese study population had advanced lesions (Table 6). Given the differences in the definition of advanced lesions in these 3 studies, however, the distribution of colonic in our Chinese cohort is comparable to that reported in the West (Table 6). Among subjects with advanced colonic lesions, 40.8% were found in the proximal colon compared with 39% in the study by Lieberman et al. and 44.6% in the study by Imperiale et al. In our cohort, 3.6% of subjects had advanced proximal lesions in the absence of distal lesions. Similar figures of 2.7% and 1.5% were reported by Lieberman et al. and Imperiale et al., respectively. This comparable pattern implies that interpretation of screening strategies can be compared between the studies despite differences in ethnicity.

6 March 2003 SCREENING FOR CRC IN CHINESE SUBJECTS 613 Table 6. Comparison of Results From Chinese Versus 2 U.S. Studies Cohort Total screened population with advanced colonic lesions (%) Advanced colonic lesions in distal colon (%) Advanced colonic lesions in proximal colon (%) Advanced colonic lesions in proximal colon without distal lesions (%) Imperiale et al. 7a Lieberman et al. 8b 10.5 (5.5) Present study c 12.5 (4.4) NOTE. Numbers in parentheses represent percentages of the screened population with advanced colonic lesions according to Imperiale et al. 7 a The 3 studies had different definitions for advanced neoplastic lesions. In the study, by Imperiale et al., 7 advanced lesions were defined as a villous adenoma, an adenoma with high-grade dysplasia, or invasive cancer. Adenoma 10 mm was not included as an advanced lesion. b In the VA study by Lieberman et al., 8 advanced lesions were defined as an adenoma 10 mm in diameter, a villous adenoma, an adenoma with high-grade dysplasia, or invasive cancer. c In the present study, advanced lesions were defined as an adenoma 10 mm in diameter, a villous adenoma, an adenoma with moderate or severe dysplasia, or invasive cancer. Our study confirmed that colonoscopy is a safe and well-tolerated procedure for screening. Among 505 subjects undergoing colonoscopy and 148 undergoing colonoscopic polypectomy, there was no perforation and only one reported case of postpolypectomy bleeding requiring a blood transfusion. With slight intravenous sedation, the procedure was also tolerated by our subjects without much distress. Seventy-eight subjects (15.4%) in this series actually entered the study on the recommendation of their family or friends who had undergone the procedure themselves. Based on the results of this study, a significant number of colorectal neoplasms have been underdiagnosed with use of FOBT and FS. Because this study used unhydrated stool samples, our results using FOBT alone as a screening test are inferior to that reported by Lieberman et al. 9 (sensitivity, 14.3% vs. 23.9%). The trade-off of using a more sensitive FOBT method would be a higher rate of false-positive results and hence more-frequent unnecessary colonoscopies. Both our study as well as the study by Lieberman et al. showed that the use of FS alone (sensitivity, 77.8% vs. 70.3%) or combined FS with FOBT (sensitivity, 80.9% vs. 77.8%) yielded similar sensitivity. Even combining FOBT with FS, 19% of advanced proximal lesions will be missed. Without survival data, it is difficult to predict the significance of missing one fifth of advanced lesions. However, a combined screening procedure short of a complete examination of the colon may produce a false sense of security in those who have been reassured by negative test results. It is also noted, from both the present study and the data from Lieberman et al., that there is at best a modest gain by adding FOBT to FS. The results of FS alone are in fact very similar to that of combined FOBT and FS. In this cohort, male subjects have a higher prevalence of colonic neoplasms and advanced lesions. This translated into a higher detection rate of FOBT and FS for these lesions. However, there is no significant difference in sensitivity and specificity when comparing various screening strategies, including FOBT alone, FS alone, and combined FOBT with FS. The mean age of our cohort is comparable to that of Imperiale et al. 7 and slightly lower than that of Lieberman et al. 8 About one third of our subjects were older than 60 years. The prevalence of all neoplasms was significantly higher in those older than 60 years (39.6%) than in the younger subjects (26.5%). There was also a trend of higher prevalence of advanced lesions among the older subjects (16.5% vs. 10.9%), but the difference did not reach statistical significance. A recent study showed that, with advancing age, there is a tendency of right shift of colonic neoplasms. 14 However, with a relatively small number of subjects older than 60 years in this study, the interpretation of these data will be limited. Based on the results of a recent study by Imperiale et al., the low yield of screening colonoscopy for patients younger than 50 years cannot justify extending the program to younger age groups. 15 There are several limitations to the present study. First, this study is meant to be a population-based recruitment enrolling only voluntary subjects without symptoms of colorectal diseases. Like any studies that involve volunteers, there is always the possibility of self-selection bias. Subjects who volunteer for colonoscopy might have untold reasons to suspect that they have colorectal diseases. The use of direct questioning in our recruitment procedure was an attempt to detect subjects with concealed symptoms. However, the small percentage of subjects in this study with a habit of smoking and drinking suggests that the health-seeking behavior of the volunteers could modify the risk of targeted lesions. Second, this study had the same limitations as those previous studies in using the junction between the splenic flexure and descending colon as the definition of distal colon for the maximal reach of FS because it is not a randomized study. 7 9 Third, the definition of advanced colonic lesions used in these criteria is slightly different from previous studies. 7 9 Interpretation of findings

7 614 SUNG ET AL. GASTROENTEROLOGY Vol. 124, No. 3 should take this into consideration. Finally, the positive rate of FOBT in this study (20%) was higher than expected. Most studies have reported a positive rate of 3% 8% on FOBT used in the general population. 3,16,17 This discrepant result could be explained by the allowance of an unrestricted diet in this study. We did not impose a strict dietary control on the subjects based on practicality. Because a typical Chinese diet includes a fair amount of peroxidase-containing vegetables and nonhuman blood, false-positive results on FOBT could be a problem when a peroxidase-based occult blood test is used in such an ethnic group. Furthermore, although this study attempted to compare the 3 screening strategies, only one-time (3 consecutive) FOBT was used. The sensitivity of FOBT could have been better if annual testing was performed as recommended by the U.S. guidelines The choice between different screening strategies has never been easy because of various issues such as safety, accuracy, and ethnic or cultural differences. The complexity of the problem is further compounded by factors such as accessibility of diagnostic tools, cost-effectiveness, and public acceptance. This study, from a non- American population, confirmed the findings of previous reports that FS and FOBT would miss a number of advanced lesions. However, availability of equipment and expertise in colonoscopy vary among different Asian countries in relation to the economic conditions and health care systems in those countries. Colonoscopy has been evaluated and reported to be cost-effective for the screening of colorectal cancer in the United States. 21,22 However, the marginal benefit of discovering these lesions by colonoscopy and the cost-effectiveness of this screening strategy in Asia need to be addressed in future studies. References 1. Yuen ST, Chung LP, Leung SY, Luk IS, Chan SY, Ho JC, Ho JW, Wyllie AH. Colorectal carcinoma in Hong Kong: epidemiology and genetic mutations. Br J Cancer 1997;76: Hong Kong Cancer Registry. Cancer incidence and mortality in Hong Kong, Hospital Authority, Hong Kong. Available at: Accessed January Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomized study of screening for colorectal cancer with faecaloccult-blood test. Lancet 1996;348: Mandel JS, Church TR, Bond JH, Ederer F, Geisser MS, Morgin SJ, Snover DC, Schuman LM. The effect of fecal-occult-blood screening on the incidence of colorectal cancer. N Engl J Med 2000; 343: Newcomb PA, Norfleet RG, Storer BE, Surawicz TS, Marcus PM. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992;84: Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A casecontrol study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326: Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasm in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000; 343: Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G, for Veterans Affairs Cooperative Study Group 380. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. N Engl J Med 2000;343: Lieberman DA, Weiss DG, for Veterans Affairs Cooperative Study Group 380. One-time screening for colorectal cancer with combined fecal occult blood testing and examination of the distal colon. N Engl J Med 2001;345: Hamilton SR, Vogelstein B, Kudo S, Riboli E, Nakamura S, Hainaut P, Rubio CA, Sobin LH, Fogt F, Winawer SJ, Goldgar DE, Jass JR. Tumours of the colon and rectum. In: Hamilton SR, Aaltonen LA, eds. World Health Organization classifications of tumors: pathology and genetics of tumors of the digestive system. Lyon, France: IARC, Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975;36: Ji BT, Devesa SS, Chow WH, Jin F, Gao YT. Colorectal cancer incidence trend by subsite in urban Shanghai Cancer Epidemiol Biomarker Prev 1998;7: Muto T, Kotake K, Koyama Y. Colorectal cancer statistics in Japan: data from JSCCR registration, Int J Clin Oncol 2001;6: Okamoto M, Shiratori Y, Yamaji Y, Kato J, Ikenoue T, Togo G, Yoshida H, Kawabe T, Omata M. Relationship between age and site of colorectal cancer based on colonoscopy findings. Gastrointest Endosc 2002;55: Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Results of screening colonoscopy among persons 40 to 49 years of age. N Engl J Med 2002;346: Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, Ederer F. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 1993;328: Winawer SJ, Shottenfeld D, Flehinger BJ. Colorectal cancer screening. J Natl Cancer Inst 1991;83: Winawer SJ, Fletcher RH, Miller L, Godlee F, Stolar MH, Mulrow C, Woolf SH, Glick SN, Ganiats TG, Bond JH, Rosen L, Zapka JD, Olsen ST, Giardiello FM, Sisk JE, Van Antwerp PR, Brown-Davis C, Marciniak DA, Mayer RJ. Colorectal cancer screening clinical guidelines and rationale. Gastroenterology 1997;112: Byers T, Levin B, Rothenberger D, Dodd GD, Smith RA. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update CA Cancer J Clin 1997;47: Shike M, Kurtz RC, Hornsby-Lewis L, Gerdes H, Stewart ET, the US Preventive Services Task Force. Guide to clinical preventive services. 2nd ed. Baltimore: Williams & Wilkins, Sonnenberg A, Delco F, Inadomi JM. Cost-effectiveness of colonoscopy in screening for colorectal cancer. Ann Intern Med 2000;133: Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 2000;284: Received June 8, Accepted December 2, Address requests for reprints to: Joseph J. Y. Sung, M.D., Department of Medicine and Therapeutics, Prince of Wales Hospital, Shatin, NT, Hong Kong. joesung@cuhk.edu.hk; fax: (852) Sponsored by the Hong Kong Society of Digestive Endoscopy.