- Email: [email protected]
Repeat colonoscopy has a low yield even in symptomatic patients
ORIGINAL ARTICLE: Clinical Endoscopy
Repeat colonoscopy has a low yield even in symptomatic patients Cynthia H. Seow, MBBS (Hons), Hooi C. Ee, MBBS, PhD, FRACP, Alex B. Willson, MBBS, MPH, Ian F. Yusoff, MBBS, FRACP Perth, Western Australia
Background: In many regions, the demand for colonoscopy exceeds its availability. Patients undergoing repeat examinations comprise a significant proportion of those on waiting lists. Objective: To assess the yield of repeat colonoscopy in varied clinical settings. Design: Cohort study. Setting: Endoscopic database of an Australian tertiary referral hospital. Patients: Adults who had R2 colonoscopies between 1992 and 2004. Patients were excluded if the repeat procedure was for completion or for high-risk surveillance. Main Outcome Measurements: Yield for neoplasia by indication, interval to repeat examination, and appropriateness for surveillance (determined by National Australian guidelines). Results: A total of 4974 colonoscopies in 2075 patients were studied. The mean age was 63.1 years (range, 19.292.4 years). The mean number of examinations was 2.4 (range, 2-8), with a mean interval between examinations of 2.9 years. Colorectal cancer (CRC) was significantly more prevalent at initial colonoscopy compared with subsequent colonoscopies (7.9% vs 0.6%; prevalence ratio 14.2, 95% confidence interval [CI] 8.5-23.7, P !.001), as were advanced adenomas (15.3% vs 4.8%; prevalence ratio 3.2, 95% CI 2.6-3.9, P!.001). No CRCs were detected in symptomatic patients undergoing polyp surveillance examinations performed before the recommended interval. Limitations: Retrospective design. Conclusions: Yield of repeat colonoscopy is significantly lower than for initial colonoscopy, irrespective of indication. In symptomatic patients within a polyp surveillance program, the yield is negligible when a colonoscopy is performed before the recommended surveillance interval. The need for a repeat colonoscopy should be carefully considered, and patients who have never had a colonoscopy must take priority on waiting lists over those awaiting repeat examinations. (Gastrointest Endosc 2006;64:941-7.)
A colonoscopy is the criterion standard investigation to detect colorectal neoplasia; furthermore, adenoma removal reduces mortality from colorectal cancer (CRC).1 Because the risk of metachronous CRC is increased by a personal history of colorectal neoplasia, these patients are advised to undergo colonoscopic surveillance.1-4 Within the last decade in Australia, the average compound growth in the number of colonoscopies performed was 16% per annum (Medicare Australia, unpublished report, 2005). This is likely to be because of increasing public awareness of CRC and increased colonoscopic surveillance Copyright ª 2006 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2006.08.004
for patients at above-average risk of CRC. Population growth and an aging population contribute to a limited degree. As a result, the waiting times for public-sector elective colonoscopy in Western Australia exceed 6 months. Patients undergoing repeat colonoscopies constitute a significant proportion of the current workload, yet the yield of pathology outside highly selected populations has not been well described. Given the current constraints on endoscopic resources, appropriate patient prioritization for colonoscopy is important to ensure that patients most in need have the shortest wait for their procedures. This study aimed to compare the findings at repeat colonoscopy with those at initial colonoscopy, both in symptomatic and asymptomatic patients, within and outside of colorectal neoplasia surveillance programs.
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 941
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
PATIENTS AND METHODS Capsule Summary We performed an audit on individuals who had R2 colonoscopies between January 1, 1992, and January 1, 2004, at Sir Charles Gairdner Hospital, Western Australia. The study center is a tertiary referral hospital with an open-access endoscopy service. An integrated surveillance program for recalling above-average risk patients is provided. All endoscopic procedures were prospectively recorded on a comprehensive, secure, electronic database. Each procedure report was reviewed individually, and relevant data were extracted. When information in the report was ambiguous, patient case notes were reviewed. Adult patients (R18 years old) who had undergone more than 1 complete clearing colonoscopy were enrolled. This was defined as an examination to the cecum or the ileum, with the removal (or sampling, where appropriate) of suspected neoplasms. Patients were excluded if (1) the repeat procedure was to complete a previously incomplete or inadequate examination, or (2) the procedure was for high-risk surveillance (defined as actual or suspected familial adenomatous polyposis, hereditary nonpolyposis CRC, or inflammatory bowel disease surveillance). Data collected included patient demographics, date of procedure, interval from last procedure, indication(s) for initial and repeat procedure(s), endoscopic findings, and histopathology of sampled lesions (with specific documentation of advanced adenomas). An advanced adenoma was defined as an adenoma with a diameter of R10 mm, a villous adenoma (comprising at least 25% villous histology), and/or an adenoma with high-grade dysplasia. The concordance of each procedure with National Health and Medical Research Council of Australia (NHMRC) Guidelines was also analyzed (Table 1).5 For colonoscopic follow-up of colorectal neoplasia, these guidelines did not differ significantly from those of the American Gastroenterological Association at the time. Adherence to these guidelines was promoted by the gastroenterology department at the study center, and a copy of the guidelines accompanied colonoscopy reports to referring doctors. This study was approved by the institutional research and ethics committee.
Statistical analysis Rates and proportions were calculated for categorical data; means and standard errors were calculated for continuous data; and c2 tests were used to compare 2 independent proportions, on the assumption that all subjects with CRC or polyps were eradicated of the disease.
What is already known on this topic d
Patients with a history of colorectal neoplasia comprise a significant proportion of the population undergoing colonoscopy, yet the yield of pathology outside highly selected populations has not been well described.
What this study adds to our knowledge d
d
Retrospective results in 2075 adults who had R2 colonoscopies performed in an Australian tertiary hospital showed that CRC was significantly more prevalent at initial colonoscopy than at subsequent procedures (7.9% vs 0.6%; prevalence ratio, 14.2). No CRCs were detected in symptomatic patients undergoing polyp surveillance examinations performed before the recommended interval.
to 4. There was a slight female preponderance (54.2% vs 45.8%, P Z .48). The mean age of the patients was 63.1 years, and the mean age at entry into the study was 60.7 years. The median and mode number of colonoscopies was 2, with a mean of 2.4. Patients undergoing repeat colonoscopy represented 19.8% of all patients having colonoscopy during the study period. The mean interval duration between examinations was 2.9 years. The majority of repeat procedures were performed within surveillance programs (77.9%).
Initial versus repeat colonoscopy Initial colonoscopy (n Z 2075) had a significantly greater yield of colorectal neoplasia when compared with repeat colonoscopies (2899 colonoscopies performed in 2075 patients). CRC was more prevalent at initial colonoscopy (7.9% vs 0.6%; prevalence ratio 14.2, 95% confidence interval [CI] 8.5-23.7, P ! .001), as were advanced adenomas (15.3% vs 4.8%; prevalence ratio 3.2, 95% CI 2.6–3.9, P ! .001) (Fig. 1). Conversely, the absence of polyps (‘‘no polyps detected’’) was more prevalent at repeat colonoscopy compared with initial examinations (42.4% vs 58.5%; prevalence ratio 0.7, 95% CI 0.7–0.9, P ! .001) (Fig. 1).
Effect of surveillance programs on yield
There were 4974 procedures (on 2075 individuals) that met the study criteria and were included in the analysis. Patients and procedure details are outlined in Tables 2
Surveillance examinations comprised 77.9% (n Z 2259) of repeat examinations (Tables 3 and 4). Of these, adenomatous polyp surveillance constituted the majority (57.3%) of examinations. In a further 29.7% of surveillance examinations, there was a personal history of CRC, and, in 13.0%, there was a family history of CRC alone. The remaining repeat examinations (22.1%) did not fulfill NHMRC criteria for surveillance. The yield for CRC (0.6% vs 0.3%; prevalence ratio [PR] 2.0, 95% CI 0.4–8.7, P Z .36) and advanced adenomas (5.4% vs 2.7%; PR 2.0, 95% CI 1.2-3.3, P ! .05) at repeat colonoscopy was greater in
942 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
RESULTS
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
TABLE 1. National Health and Medical Research Council of Australia Guidelines for Management of Epithelial Polyps*
TABLE 3. Included colonoscopic examinations Initial colonoscopy, no.
2075
All patients with colorectal neoplasia completely removed at colonoscopy should be considered for colonoscopic surveillance according to the following protocols.
All repeat colonoscopies, no.
2899
Repeat within a surveillance program, no. (%)
2259 (77.9)
Within a year after incomplete or possible inadequate examination, e.g., in a subject with multiple adenomas.
CRC surveillance program Polyp surveillance program
At 3 years for subjects with large adenomas (O1 cm), adenomas with high-grade dysplasia, villous change in adenomas, 3 or more adenomas, or aged 60 years or more with a first-degree relative with colorectal neoplasia.
671 (29.7) 1294 (57.3)
Family history surveillance program
294 (13.0)
Repeat out of a surveillance program, no. (%)
640 (22.1)
At 4 to 6 years in subjects without the risk factors outlined above. *From Ref. 5.
TABLE 4. Excluded colonoscopic examinations Total no.
TABLE 2. Characteristics of the study participants No. patients
2075
Men, women, %
45.8, 54.2
Total no. colonoscopies
4974
No. repeat colonoscopies
2899
No. mean colonoscopies (range)
2.4 (2-8)
Mean interval duration between colonoscopies (IQR), y
2.9 (1.6-3.7)
Mean age of patients (IQR), y
63.1 (19.2–92.4)
Mean age at entry to study (IQR), y
60.7 (19.2–91.6)
IQR, Interquartile range.
those within any surveillance program with prevalence ratios of 2.0 compared with those who had a repeat procedure for symptoms alone (Fig. 2). There was a significantly disproportionate number of females in the nonsurveillance cohort (59.6% vs 51.3%; PR 1.2, 95% CI 1.1-1.3, P !.001). In symptomatic patients, the yield for significant neoplasia (CRC and advanced adenomas) presenting with bleeding was significantly higher at initial colonoscopy (31.7% vs 18.2%; PR 1.7, 95% CI 1.52.0, P ! .0001) but not at repeat colonoscopy (6.1% vs 5.1%; 95% CI 0.8-1.7, P Z .39).
CRCs detected at repeat colonoscopy Sixteen CRCs were detected at repeat colonoscopy. Fourteen of these were in individuals with a personal history of colorectal adenomas or CRCs. Five CRCs were detected in patients within a polyp surveillance program, and 9 CRCs were detected in patients with a personal history of CRC (8 CRCs were anastomotic recurrences, 1 CRC was a metachronous lesion). www.giejournal.org
2098
High-risk polyposis and nonpolyposis cancer syndromes, no. (%)
197 (9.4)
Inflammatory bowel disease, no. (%)
757 (36.1)
!2 Complete clearing colonoscopies, no. (%)
1144 (54.5)
No CRCs were identified in symptomatic patients within polyp surveillance programs who had colonoscopies performed before the recommended NHMRC interval. The yield of neoplasia was highest when colonoscopy was performed at or after the recommended NHMRC interval (Table 5). A description of the CRCs detected at a repeat colonoscopy within polyp surveillance programs is shown in Table 6. Two CRCs were detected at repeat examinations in symptomatic patients not meeting criteria to enter a surveillance program. The first individual was an 83-year-old woman, with a rectosigmoid CRC, detected approximately 4 years after an initial colonoscopy, which revealed diverticular disease alone. The indication for both initial and repeat colonoscopy was bleeding. The second individual, a 68-year-old female, had a benign-appearing cecal ulcer when she first presented with abdominal pain. Biopsy specimens suggested a benign etiology. She was rereferred with iron deficiency anemia 12 months later, and colonoscopy revealed a cecal CRC.
Number of colonoscopies needed to detect CRC The number of colonoscopies needed to detect CRC in different clinical settings is graphically represented in Figure 3. Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 943
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
Figure 1. Yield of pathology at initial and repeat colonoscopy (n Z 4974).
Figure 2. Yield of repeat colonoscopy in patients within and not in colorectal neoplasia surveillance programs (n Z 2899).
TABLE 5. Yield of neoplasia in a polyp surveillance program with respect to timing of colonoscopy Before NHMRC recommended interval with symptoms, %
Before NHMRC recommended interval without symptoms, %
At or after recommended NHMRC interval, %
CRC
0.0
0.0
1.0
Advanced adenoma
0.9
2.6
3.5
No polyps detected
8.0
20.2
16.8
Polyp surveillance
DISCUSSION Increasing public awareness of CRC and the introduction of screening programs have resulted in an unprecedented demand for colonoscopy. In many regions, demand exceeds supply6; thus, patients undergoing repeat examinations compete with symptomatic patients
who have never undergone a prior colonoscopy. Rex and Lieberman7 estimated that surveillance accounts for 25% of colonoscopies. At our institution, 20% of patients have repeat procedures for various indications. We showed, not surprisingly, that the yield of neoplasia at subsequent colonoscopy was significantly lower than at initial colonoscopy. Data extrapolated from the landmark study
944 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
TABLE 6. CRCs detected at repeat colonoscopy in patients within a polyp surveillance program
Age, y
Interval between procedures, y
Symptoms
Previous scope
Current scope
Comments
53
6.8
Nil
6-mm transverse colon tubular adenoma; 6-mm sigmoid tubular adenoma
Sigmoid CRC
Late examination
76
4.2
Abnormal radiology
15-mm ascending colon tubulovillous adenoma; 2 7-mm ascending colonic tubular adenomas
Sigmoid CRC
Late examination
83
7.8
Iron deficiency anemia
15-mm cecal tubular adenoma; 3 5-10–mm cecal polyps; 10-mm transverse colon polyp; 5-mm sigmoid polyp
Ascending colon CRC
Late examination
67
1.8
Nil
Small sigmoid polyp; suspected malignant ulcer in sigmoid; false-negative biopsy specimen from ulcer base
Sigmoid CRC
Late examination; should have had early colonoscopic follow-up for rebiopsy
67
1.0
Iron deficiency
10-mm ascending colon tubulovillous adenoma; multiple other left-sided polyps, 4-15 mm
Ascending colon CRC
At suggested NHMRC surveillance interval; multiple previous polyps
by Lieberman et al8 revealed that 100 colonoscopies were needed to detect 1 CRC in a screening program. Our data suggest that 362 colonoscopies are required to detect a CRC (excluding anastomotic recurrences) at a repeat examination for any indication. Although colorectal polyps are commonly found at repeat colonoscopy, our study focused on advanced neoplasia, given that the vast majority of small adenomas pose little long-term risk for CRC.2,3 We showed that the yield for significant neoplasia was very low at repeat examinations. Indeed, no CRCs were detected in symptomatic patients who had procedures performed before the recommended NHMRC interval in polyp surveillance programs. These data confirmed that little further benefit was derived from more frequent surveillance and that symptoms were not predictive of neoplasia.9 Patients presenting with a bleeding indication had an increased yield of CRC and advanced adenomas.10 We confirmed this finding at initial colonoscopy; however, we showed that bleeding was not predictive of neoplasia at repeat colonoscopy. Symptomatic patients undergoing repeat colonoscopy without a personal or family history or colorectal neoplasia had a significantly lower yield for advanced adenomas and CRC, despite similar mean intervals between examinations. An important ‘‘cost’’ of repeat examinations is the potential for cumulative complications to offset the benefit in
CRC reduction.11 Ransohoff12 calculated that the averagerisk person may benefit from a 1.2% CRC mortality reduction, but this is associated with a 1% risk of serious complications and a 0.1% chance of death. We previously reported that complications relate primarily to interventions rather than to indication.13 The perforation rate was 0.03% in the absence of intervention but rose to 0.25% when polypectomy was performed.13 It is conceivable that removal of trivial polyps at early repeat colonoscopy could increase ‘‘unnecessary’’ complications. In the current study, only 5 CRCs were detected within 1294 polyp-surveillance examinations, but only one of these was when the examination was not performed ‘‘late’’ (ie, after the recommended surveillance interval). This was detected at the recommended interval but occurred in the setting of multiple advanced adenomas, which confers a 3- to 4-fold increase in risk for subsequent CRC.14-16 Requests for early repeat examinations are a significant burden in an open-access practice. Alarm at new symptoms and the fear of litigation, compounded by ignorance regarding the slow rate of colorectal carcinogenesis in average-risk patients, are likely driving forces. Colonoscopy has become an important means of allaying patient and physician anxiety, and education programs for doctors and the general public are warranted.17 Two earlier studies from our institution18,19 found high rates of noncompliance with guidelines regarding surveillance for CRC because of
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 945
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
Figure 3. Number of colonoscopies needed to detect CRC in varying clinical settings.
too frequent examinations. Mysliwiec et al20 also reported on the disparity between clinical practice and physician knowledge, with a large number of inappropriate examinations for low-risk lesions. A strength of the current study is its generalizability, given the ‘‘real-world’’ setting. We studied a large nonselected patient population, with procedures performed by endoscopists of varied experience and expertise over 12 years (with 6018 person-years observed). In any year, up to 30% of procedures were performed by trainee endoscopists at varied stages of training, with varying degrees of supervision. Our interval CRC prevalence in the polyp surveillance cohort was 1.96 of 1000 person-years and is comparable with rates described in other major studies.16,21-24 These studies emphasize the importance of high-quality initial colonoscopies to minimize the frequency of missed lesions. This has important implications for training, maintenance of professional standards, and quality assurance, especially if substantial expansions in the colonoscopy workforce are required to deal with increasing demand. The majority of CRCs detected in this study occurred in patients with a personal history of CRC. This warrants clarification and comment. In 8 of 9 cases, these tumors were anastomotic recurrences. In itself, this may argue for more aggressive surveillance for early detection; however, our experience does not support this. We previously studied this group specifically and reported that anastomotic recurrences are uncommon, and resectable recurrences are even less common.25 Furthermore, current data suggest that intensive colonoscopic surveillance does not confer a survival benefit.26 The current study has convincingly shown that symptomatic patients who have never had a colonoscopy must take priority on waiting lists over patients awaiting repeat colonoscopy, even if the latter are symptomatic. These data, in concept, lend support to those who advocate ‘‘one-time screening’’ colonoscopy in 55- to 60-year-old
individuals27,28 as a way of streamlining resources. Although our study was not designed to address this specific issue, it provides indirect support for this argument, particularly in regions of limited resources. Similarly, these data support recent American Gastroenterological Association recommendations to prolong the interval to reexamination in patients with no polyps or with 1 to 2 small polyps.29 In conclusion, our major finding is that the yield of significant pathology at repeat colonoscopy is substantially lower than that at the initial examination, irrespective of the symptoms and indications. This has significant implications for clinical decision making and colonoscopic resource allocation.
946 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
ACKNOWLEDGMENTS We thank Dr Neville Hoffman and Dr Matthew Zimmerman for their assistance with the database.
DISCLOSURE None of the authors have any disclosures to make.
REFERENCES 1. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977-81. 2. Schoen RE. Surveillance after positive and negative colonoscopy examinations: issues, yields, and use. Am J Gastroenterol 2003;98: 1237-46. 3. Yamaji Y, Mitsushima T, Ikuma H, et al. Incidence and recurrence rates of colorectal adenomas estimated by annually repeated colonoscopies on asymptomatic Japanese. Gut 2004;53:568-72. 4. Neugut AI, Jacobson JS, Ahsan H, et al. Incidence and recurrence rates of colorectal adenomas: a prospective study. Gastroenterology 1995;108:402-8.
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
5. National Health and Medical Research Council. Clinical practice guidelines for the prevention, early detection and management of colorectal cancer: the Cancer Council Australia/Australia Cancer Network. Canberra, Australia: Biotext; 2005. 6. Baron TH, Kimery BD, Sorbi D, et al. Strategies to address increased demand for colonoscopy: guidelines in an open endoscopy practice. Clin Gastroenterol Hepatol 2004;2:178-82. 7. Rex DK, Lieberman DA. Feasibility of colonoscopy screening: discussion of issues and recommendations regarding implementation. Gastrointest Endosc 2001;54:662-7. 8. Lieberman DA, Weiss DG, Bond JH, et al. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000;343:162-8. 9. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. The National Polyp Study Workgroup. N Engl J Med 1993;328:901-6. 10. Rex DK. Colonoscopy: a review of its yield for cancers and adenomas by indication. Am J Gastroenterol 1995;90:353-65. 11. Nelson DB, McQuaid KR, Bond JH, et al. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55:307-14. 12. Ransohoff DF. Screening colonoscopy in balance. Issues of implementation. Gastroenterol Clin North Am 2002;31:1031-44, vii. 13. Viiala CH, Zimmerman M, Cullen DJ, et al. Complication rates of colonoscopy in an Australian teaching hospital environment. Intern Med J 2003;33:355-9. 14. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 1992;326: 658-62. 15. Avidan B, Sonnenberg A, Schnell TG, et al. New occurrence and recurrence of neoplasms within 5 years of a screening colonoscopy. Am J Gastroenterol 2002;97:1524-9. 16. Robertson DJ, Greenberg ER, Beach M, et al. Colorectal cancer in patients under close colonoscopic surveillance. Gastroenterology 2005;129:34-41. 17. Petersen GM. Barriers to preventive intervention. Gastroenterol Clin North Am 2002;31:1061-8, viii. 18. Yusoff IF, Hoffman NE, Ee HC. Colonoscopic surveillance for family history of colorectal cancer: are NHMRC guidelines being followed? Med J Aust 2002;176:151-4.
19. Leber JM, Yusoff IF, Ee HC. Inappropriate use of colonoscopy in a polyp follow-up surveillance program at a tertiary referral centre [abstract]. J Gastroenterol Hepatol 2004;19:A193. 20. Mysliwiec PA, Brown ML, Klabunde CN, et al. Are physicians doing too much colonoscopy? A national survey of colorectal surveillance after polypectomy. Ann Intern Med 2004;141:264-71. 21. Jorgensen OD, Kronborg O, Fenger C. The Funen Adenoma Follow-up Study. Incidence and death from colorectal carcinoma in an adenoma surveillance program. Scand J Gastroenterol 1993;28:869-74. 22. Pabby A, Schoen RE, Weissfeld JL, et al. Analysis of colorectal cancer occurrence during surveillance in the dietary Polyp Prevention Trial. Gastrointest Endosc 2005;61:385-91. 23. Alberts DS, Martinez ME, Roe DJ, et al. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians’ Network. N Engl J Med 2000;342:1156-62. 24. Ee HC, Semmens JB, Hoffman NE. Complete colonoscopy rarely misses cancer. Gastrointest Endosc 2002;55:167-71. 25. Yusoff IF, Hoffman NE, Ee HC. Colonoscopic surveillance after surgery for colorectal cancer. ANZ J Surg 2003;73:3-7. 26. Ee HC, Yusoff IF. Colonoscopic surveillance after curative surgery for colorectal cancer. ANZ J Surg 2005;75:257-8. 27. Lieberman D. Endoscopic colon screening: is less more? Gastroenterology 1996;111:1385-7. 28. Rex DK, Cummings OW, Helper DJ, et al. 5-year incidence of adenomas after negative colonoscopy in asymptomatic average-risk persons [see comment]. Gastroenterology 1996;111:1178-81. 29. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US MultiSociety Task Force on Colorectal Cancer and the American Cancer Society. Gastroenterology 2006;130:1872-85.
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 947
Received April 5, 2006. Accepted August 7, 2006. Current affiliations: Department of Gastroenterology (C.H.S., H.C.E., and A.B.W.), School of Medicine and Pharmacology (I.F.Y.), Sir Charles Gairdner Hospital Unit, The University of Western Australia, Nedlands, Perth, Western Australia. Reprint requests: Ian Yusoff, MBBS, Department of Gastroenterology, Sir Charles Gairdner Hospital, Hospital Ave, Nedlands, Western Australia 6009.
Repeat colonoscopy has a low yield even in symptomatic patients Cynthia H. Seow, MBBS (Hons), Hooi C. Ee, MBBS, PhD, FRACP, Alex B. Willson, MBBS, MPH, Ian F. Yusoff, MBBS, FRACP Perth, Western Australia
Background: In many regions, the demand for colonoscopy exceeds its availability. Patients undergoing repeat examinations comprise a significant proportion of those on waiting lists. Objective: To assess the yield of repeat colonoscopy in varied clinical settings. Design: Cohort study. Setting: Endoscopic database of an Australian tertiary referral hospital. Patients: Adults who had R2 colonoscopies between 1992 and 2004. Patients were excluded if the repeat procedure was for completion or for high-risk surveillance. Main Outcome Measurements: Yield for neoplasia by indication, interval to repeat examination, and appropriateness for surveillance (determined by National Australian guidelines). Results: A total of 4974 colonoscopies in 2075 patients were studied. The mean age was 63.1 years (range, 19.292.4 years). The mean number of examinations was 2.4 (range, 2-8), with a mean interval between examinations of 2.9 years. Colorectal cancer (CRC) was significantly more prevalent at initial colonoscopy compared with subsequent colonoscopies (7.9% vs 0.6%; prevalence ratio 14.2, 95% confidence interval [CI] 8.5-23.7, P !.001), as were advanced adenomas (15.3% vs 4.8%; prevalence ratio 3.2, 95% CI 2.6-3.9, P!.001). No CRCs were detected in symptomatic patients undergoing polyp surveillance examinations performed before the recommended interval. Limitations: Retrospective design. Conclusions: Yield of repeat colonoscopy is significantly lower than for initial colonoscopy, irrespective of indication. In symptomatic patients within a polyp surveillance program, the yield is negligible when a colonoscopy is performed before the recommended surveillance interval. The need for a repeat colonoscopy should be carefully considered, and patients who have never had a colonoscopy must take priority on waiting lists over those awaiting repeat examinations. (Gastrointest Endosc 2006;64:941-7.)
A colonoscopy is the criterion standard investigation to detect colorectal neoplasia; furthermore, adenoma removal reduces mortality from colorectal cancer (CRC).1 Because the risk of metachronous CRC is increased by a personal history of colorectal neoplasia, these patients are advised to undergo colonoscopic surveillance.1-4 Within the last decade in Australia, the average compound growth in the number of colonoscopies performed was 16% per annum (Medicare Australia, unpublished report, 2005). This is likely to be because of increasing public awareness of CRC and increased colonoscopic surveillance Copyright ª 2006 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2006.08.004
for patients at above-average risk of CRC. Population growth and an aging population contribute to a limited degree. As a result, the waiting times for public-sector elective colonoscopy in Western Australia exceed 6 months. Patients undergoing repeat colonoscopies constitute a significant proportion of the current workload, yet the yield of pathology outside highly selected populations has not been well described. Given the current constraints on endoscopic resources, appropriate patient prioritization for colonoscopy is important to ensure that patients most in need have the shortest wait for their procedures. This study aimed to compare the findings at repeat colonoscopy with those at initial colonoscopy, both in symptomatic and asymptomatic patients, within and outside of colorectal neoplasia surveillance programs.
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 941
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
PATIENTS AND METHODS Capsule Summary We performed an audit on individuals who had R2 colonoscopies between January 1, 1992, and January 1, 2004, at Sir Charles Gairdner Hospital, Western Australia. The study center is a tertiary referral hospital with an open-access endoscopy service. An integrated surveillance program for recalling above-average risk patients is provided. All endoscopic procedures were prospectively recorded on a comprehensive, secure, electronic database. Each procedure report was reviewed individually, and relevant data were extracted. When information in the report was ambiguous, patient case notes were reviewed. Adult patients (R18 years old) who had undergone more than 1 complete clearing colonoscopy were enrolled. This was defined as an examination to the cecum or the ileum, with the removal (or sampling, where appropriate) of suspected neoplasms. Patients were excluded if (1) the repeat procedure was to complete a previously incomplete or inadequate examination, or (2) the procedure was for high-risk surveillance (defined as actual or suspected familial adenomatous polyposis, hereditary nonpolyposis CRC, or inflammatory bowel disease surveillance). Data collected included patient demographics, date of procedure, interval from last procedure, indication(s) for initial and repeat procedure(s), endoscopic findings, and histopathology of sampled lesions (with specific documentation of advanced adenomas). An advanced adenoma was defined as an adenoma with a diameter of R10 mm, a villous adenoma (comprising at least 25% villous histology), and/or an adenoma with high-grade dysplasia. The concordance of each procedure with National Health and Medical Research Council of Australia (NHMRC) Guidelines was also analyzed (Table 1).5 For colonoscopic follow-up of colorectal neoplasia, these guidelines did not differ significantly from those of the American Gastroenterological Association at the time. Adherence to these guidelines was promoted by the gastroenterology department at the study center, and a copy of the guidelines accompanied colonoscopy reports to referring doctors. This study was approved by the institutional research and ethics committee.
Statistical analysis Rates and proportions were calculated for categorical data; means and standard errors were calculated for continuous data; and c2 tests were used to compare 2 independent proportions, on the assumption that all subjects with CRC or polyps were eradicated of the disease.
What is already known on this topic d
Patients with a history of colorectal neoplasia comprise a significant proportion of the population undergoing colonoscopy, yet the yield of pathology outside highly selected populations has not been well described.
What this study adds to our knowledge d
d
Retrospective results in 2075 adults who had R2 colonoscopies performed in an Australian tertiary hospital showed that CRC was significantly more prevalent at initial colonoscopy than at subsequent procedures (7.9% vs 0.6%; prevalence ratio, 14.2). No CRCs were detected in symptomatic patients undergoing polyp surveillance examinations performed before the recommended interval.
to 4. There was a slight female preponderance (54.2% vs 45.8%, P Z .48). The mean age of the patients was 63.1 years, and the mean age at entry into the study was 60.7 years. The median and mode number of colonoscopies was 2, with a mean of 2.4. Patients undergoing repeat colonoscopy represented 19.8% of all patients having colonoscopy during the study period. The mean interval duration between examinations was 2.9 years. The majority of repeat procedures were performed within surveillance programs (77.9%).
Initial versus repeat colonoscopy Initial colonoscopy (n Z 2075) had a significantly greater yield of colorectal neoplasia when compared with repeat colonoscopies (2899 colonoscopies performed in 2075 patients). CRC was more prevalent at initial colonoscopy (7.9% vs 0.6%; prevalence ratio 14.2, 95% confidence interval [CI] 8.5-23.7, P ! .001), as were advanced adenomas (15.3% vs 4.8%; prevalence ratio 3.2, 95% CI 2.6–3.9, P ! .001) (Fig. 1). Conversely, the absence of polyps (‘‘no polyps detected’’) was more prevalent at repeat colonoscopy compared with initial examinations (42.4% vs 58.5%; prevalence ratio 0.7, 95% CI 0.7–0.9, P ! .001) (Fig. 1).
Effect of surveillance programs on yield
There were 4974 procedures (on 2075 individuals) that met the study criteria and were included in the analysis. Patients and procedure details are outlined in Tables 2
Surveillance examinations comprised 77.9% (n Z 2259) of repeat examinations (Tables 3 and 4). Of these, adenomatous polyp surveillance constituted the majority (57.3%) of examinations. In a further 29.7% of surveillance examinations, there was a personal history of CRC, and, in 13.0%, there was a family history of CRC alone. The remaining repeat examinations (22.1%) did not fulfill NHMRC criteria for surveillance. The yield for CRC (0.6% vs 0.3%; prevalence ratio [PR] 2.0, 95% CI 0.4–8.7, P Z .36) and advanced adenomas (5.4% vs 2.7%; PR 2.0, 95% CI 1.2-3.3, P ! .05) at repeat colonoscopy was greater in
942 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
RESULTS
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
TABLE 1. National Health and Medical Research Council of Australia Guidelines for Management of Epithelial Polyps*
TABLE 3. Included colonoscopic examinations Initial colonoscopy, no.
2075
All patients with colorectal neoplasia completely removed at colonoscopy should be considered for colonoscopic surveillance according to the following protocols.
All repeat colonoscopies, no.
2899
Repeat within a surveillance program, no. (%)
2259 (77.9)
Within a year after incomplete or possible inadequate examination, e.g., in a subject with multiple adenomas.
CRC surveillance program Polyp surveillance program
At 3 years for subjects with large adenomas (O1 cm), adenomas with high-grade dysplasia, villous change in adenomas, 3 or more adenomas, or aged 60 years or more with a first-degree relative with colorectal neoplasia.
671 (29.7) 1294 (57.3)
Family history surveillance program
294 (13.0)
Repeat out of a surveillance program, no. (%)
640 (22.1)
At 4 to 6 years in subjects without the risk factors outlined above. *From Ref. 5.
TABLE 4. Excluded colonoscopic examinations Total no.
TABLE 2. Characteristics of the study participants No. patients
2075
Men, women, %
45.8, 54.2
Total no. colonoscopies
4974
No. repeat colonoscopies
2899
No. mean colonoscopies (range)
2.4 (2-8)
Mean interval duration between colonoscopies (IQR), y
2.9 (1.6-3.7)
Mean age of patients (IQR), y
63.1 (19.2–92.4)
Mean age at entry to study (IQR), y
60.7 (19.2–91.6)
IQR, Interquartile range.
those within any surveillance program with prevalence ratios of 2.0 compared with those who had a repeat procedure for symptoms alone (Fig. 2). There was a significantly disproportionate number of females in the nonsurveillance cohort (59.6% vs 51.3%; PR 1.2, 95% CI 1.1-1.3, P !.001). In symptomatic patients, the yield for significant neoplasia (CRC and advanced adenomas) presenting with bleeding was significantly higher at initial colonoscopy (31.7% vs 18.2%; PR 1.7, 95% CI 1.52.0, P ! .0001) but not at repeat colonoscopy (6.1% vs 5.1%; 95% CI 0.8-1.7, P Z .39).
CRCs detected at repeat colonoscopy Sixteen CRCs were detected at repeat colonoscopy. Fourteen of these were in individuals with a personal history of colorectal adenomas or CRCs. Five CRCs were detected in patients within a polyp surveillance program, and 9 CRCs were detected in patients with a personal history of CRC (8 CRCs were anastomotic recurrences, 1 CRC was a metachronous lesion). www.giejournal.org
2098
High-risk polyposis and nonpolyposis cancer syndromes, no. (%)
197 (9.4)
Inflammatory bowel disease, no. (%)
757 (36.1)
!2 Complete clearing colonoscopies, no. (%)
1144 (54.5)
No CRCs were identified in symptomatic patients within polyp surveillance programs who had colonoscopies performed before the recommended NHMRC interval. The yield of neoplasia was highest when colonoscopy was performed at or after the recommended NHMRC interval (Table 5). A description of the CRCs detected at a repeat colonoscopy within polyp surveillance programs is shown in Table 6. Two CRCs were detected at repeat examinations in symptomatic patients not meeting criteria to enter a surveillance program. The first individual was an 83-year-old woman, with a rectosigmoid CRC, detected approximately 4 years after an initial colonoscopy, which revealed diverticular disease alone. The indication for both initial and repeat colonoscopy was bleeding. The second individual, a 68-year-old female, had a benign-appearing cecal ulcer when she first presented with abdominal pain. Biopsy specimens suggested a benign etiology. She was rereferred with iron deficiency anemia 12 months later, and colonoscopy revealed a cecal CRC.
Number of colonoscopies needed to detect CRC The number of colonoscopies needed to detect CRC in different clinical settings is graphically represented in Figure 3. Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 943
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
Figure 1. Yield of pathology at initial and repeat colonoscopy (n Z 4974).
Figure 2. Yield of repeat colonoscopy in patients within and not in colorectal neoplasia surveillance programs (n Z 2899).
TABLE 5. Yield of neoplasia in a polyp surveillance program with respect to timing of colonoscopy Before NHMRC recommended interval with symptoms, %
Before NHMRC recommended interval without symptoms, %
At or after recommended NHMRC interval, %
CRC
0.0
0.0
1.0
Advanced adenoma
0.9
2.6
3.5
No polyps detected
8.0
20.2
16.8
Polyp surveillance
DISCUSSION Increasing public awareness of CRC and the introduction of screening programs have resulted in an unprecedented demand for colonoscopy. In many regions, demand exceeds supply6; thus, patients undergoing repeat examinations compete with symptomatic patients
who have never undergone a prior colonoscopy. Rex and Lieberman7 estimated that surveillance accounts for 25% of colonoscopies. At our institution, 20% of patients have repeat procedures for various indications. We showed, not surprisingly, that the yield of neoplasia at subsequent colonoscopy was significantly lower than at initial colonoscopy. Data extrapolated from the landmark study
944 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
TABLE 6. CRCs detected at repeat colonoscopy in patients within a polyp surveillance program
Age, y
Interval between procedures, y
Symptoms
Previous scope
Current scope
Comments
53
6.8
Nil
6-mm transverse colon tubular adenoma; 6-mm sigmoid tubular adenoma
Sigmoid CRC
Late examination
76
4.2
Abnormal radiology
15-mm ascending colon tubulovillous adenoma; 2 7-mm ascending colonic tubular adenomas
Sigmoid CRC
Late examination
83
7.8
Iron deficiency anemia
15-mm cecal tubular adenoma; 3 5-10–mm cecal polyps; 10-mm transverse colon polyp; 5-mm sigmoid polyp
Ascending colon CRC
Late examination
67
1.8
Nil
Small sigmoid polyp; suspected malignant ulcer in sigmoid; false-negative biopsy specimen from ulcer base
Sigmoid CRC
Late examination; should have had early colonoscopic follow-up for rebiopsy
67
1.0
Iron deficiency
10-mm ascending colon tubulovillous adenoma; multiple other left-sided polyps, 4-15 mm
Ascending colon CRC
At suggested NHMRC surveillance interval; multiple previous polyps
by Lieberman et al8 revealed that 100 colonoscopies were needed to detect 1 CRC in a screening program. Our data suggest that 362 colonoscopies are required to detect a CRC (excluding anastomotic recurrences) at a repeat examination for any indication. Although colorectal polyps are commonly found at repeat colonoscopy, our study focused on advanced neoplasia, given that the vast majority of small adenomas pose little long-term risk for CRC.2,3 We showed that the yield for significant neoplasia was very low at repeat examinations. Indeed, no CRCs were detected in symptomatic patients who had procedures performed before the recommended NHMRC interval in polyp surveillance programs. These data confirmed that little further benefit was derived from more frequent surveillance and that symptoms were not predictive of neoplasia.9 Patients presenting with a bleeding indication had an increased yield of CRC and advanced adenomas.10 We confirmed this finding at initial colonoscopy; however, we showed that bleeding was not predictive of neoplasia at repeat colonoscopy. Symptomatic patients undergoing repeat colonoscopy without a personal or family history or colorectal neoplasia had a significantly lower yield for advanced adenomas and CRC, despite similar mean intervals between examinations. An important ‘‘cost’’ of repeat examinations is the potential for cumulative complications to offset the benefit in
CRC reduction.11 Ransohoff12 calculated that the averagerisk person may benefit from a 1.2% CRC mortality reduction, but this is associated with a 1% risk of serious complications and a 0.1% chance of death. We previously reported that complications relate primarily to interventions rather than to indication.13 The perforation rate was 0.03% in the absence of intervention but rose to 0.25% when polypectomy was performed.13 It is conceivable that removal of trivial polyps at early repeat colonoscopy could increase ‘‘unnecessary’’ complications. In the current study, only 5 CRCs were detected within 1294 polyp-surveillance examinations, but only one of these was when the examination was not performed ‘‘late’’ (ie, after the recommended surveillance interval). This was detected at the recommended interval but occurred in the setting of multiple advanced adenomas, which confers a 3- to 4-fold increase in risk for subsequent CRC.14-16 Requests for early repeat examinations are a significant burden in an open-access practice. Alarm at new symptoms and the fear of litigation, compounded by ignorance regarding the slow rate of colorectal carcinogenesis in average-risk patients, are likely driving forces. Colonoscopy has become an important means of allaying patient and physician anxiety, and education programs for doctors and the general public are warranted.17 Two earlier studies from our institution18,19 found high rates of noncompliance with guidelines regarding surveillance for CRC because of
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 945
Repeat colonoscopy has a low yield even in symptomatic patients
Seow et al
Figure 3. Number of colonoscopies needed to detect CRC in varying clinical settings.
too frequent examinations. Mysliwiec et al20 also reported on the disparity between clinical practice and physician knowledge, with a large number of inappropriate examinations for low-risk lesions. A strength of the current study is its generalizability, given the ‘‘real-world’’ setting. We studied a large nonselected patient population, with procedures performed by endoscopists of varied experience and expertise over 12 years (with 6018 person-years observed). In any year, up to 30% of procedures were performed by trainee endoscopists at varied stages of training, with varying degrees of supervision. Our interval CRC prevalence in the polyp surveillance cohort was 1.96 of 1000 person-years and is comparable with rates described in other major studies.16,21-24 These studies emphasize the importance of high-quality initial colonoscopies to minimize the frequency of missed lesions. This has important implications for training, maintenance of professional standards, and quality assurance, especially if substantial expansions in the colonoscopy workforce are required to deal with increasing demand. The majority of CRCs detected in this study occurred in patients with a personal history of CRC. This warrants clarification and comment. In 8 of 9 cases, these tumors were anastomotic recurrences. In itself, this may argue for more aggressive surveillance for early detection; however, our experience does not support this. We previously studied this group specifically and reported that anastomotic recurrences are uncommon, and resectable recurrences are even less common.25 Furthermore, current data suggest that intensive colonoscopic surveillance does not confer a survival benefit.26 The current study has convincingly shown that symptomatic patients who have never had a colonoscopy must take priority on waiting lists over patients awaiting repeat colonoscopy, even if the latter are symptomatic. These data, in concept, lend support to those who advocate ‘‘one-time screening’’ colonoscopy in 55- to 60-year-old
individuals27,28 as a way of streamlining resources. Although our study was not designed to address this specific issue, it provides indirect support for this argument, particularly in regions of limited resources. Similarly, these data support recent American Gastroenterological Association recommendations to prolong the interval to reexamination in patients with no polyps or with 1 to 2 small polyps.29 In conclusion, our major finding is that the yield of significant pathology at repeat colonoscopy is substantially lower than that at the initial examination, irrespective of the symptoms and indications. This has significant implications for clinical decision making and colonoscopic resource allocation.
946 GASTROINTESTINAL ENDOSCOPY Volume 64, No. 6 : 2006
www.giejournal.org
ACKNOWLEDGMENTS We thank Dr Neville Hoffman and Dr Matthew Zimmerman for their assistance with the database.
DISCLOSURE None of the authors have any disclosures to make.
REFERENCES 1. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977-81. 2. Schoen RE. Surveillance after positive and negative colonoscopy examinations: issues, yields, and use. Am J Gastroenterol 2003;98: 1237-46. 3. Yamaji Y, Mitsushima T, Ikuma H, et al. Incidence and recurrence rates of colorectal adenomas estimated by annually repeated colonoscopies on asymptomatic Japanese. Gut 2004;53:568-72. 4. Neugut AI, Jacobson JS, Ahsan H, et al. Incidence and recurrence rates of colorectal adenomas: a prospective study. Gastroenterology 1995;108:402-8.
Seow et al
Repeat colonoscopy has a low yield even in symptomatic patients
5. National Health and Medical Research Council. Clinical practice guidelines for the prevention, early detection and management of colorectal cancer: the Cancer Council Australia/Australia Cancer Network. Canberra, Australia: Biotext; 2005. 6. Baron TH, Kimery BD, Sorbi D, et al. Strategies to address increased demand for colonoscopy: guidelines in an open endoscopy practice. Clin Gastroenterol Hepatol 2004;2:178-82. 7. Rex DK, Lieberman DA. Feasibility of colonoscopy screening: discussion of issues and recommendations regarding implementation. Gastrointest Endosc 2001;54:662-7. 8. Lieberman DA, Weiss DG, Bond JH, et al. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000;343:162-8. 9. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. The National Polyp Study Workgroup. N Engl J Med 1993;328:901-6. 10. Rex DK. Colonoscopy: a review of its yield for cancers and adenomas by indication. Am J Gastroenterol 1995;90:353-65. 11. Nelson DB, McQuaid KR, Bond JH, et al. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55:307-14. 12. Ransohoff DF. Screening colonoscopy in balance. Issues of implementation. Gastroenterol Clin North Am 2002;31:1031-44, vii. 13. Viiala CH, Zimmerman M, Cullen DJ, et al. Complication rates of colonoscopy in an Australian teaching hospital environment. Intern Med J 2003;33:355-9. 14. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 1992;326: 658-62. 15. Avidan B, Sonnenberg A, Schnell TG, et al. New occurrence and recurrence of neoplasms within 5 years of a screening colonoscopy. Am J Gastroenterol 2002;97:1524-9. 16. Robertson DJ, Greenberg ER, Beach M, et al. Colorectal cancer in patients under close colonoscopic surveillance. Gastroenterology 2005;129:34-41. 17. Petersen GM. Barriers to preventive intervention. Gastroenterol Clin North Am 2002;31:1061-8, viii. 18. Yusoff IF, Hoffman NE, Ee HC. Colonoscopic surveillance for family history of colorectal cancer: are NHMRC guidelines being followed? Med J Aust 2002;176:151-4.
19. Leber JM, Yusoff IF, Ee HC. Inappropriate use of colonoscopy in a polyp follow-up surveillance program at a tertiary referral centre [abstract]. J Gastroenterol Hepatol 2004;19:A193. 20. Mysliwiec PA, Brown ML, Klabunde CN, et al. Are physicians doing too much colonoscopy? A national survey of colorectal surveillance after polypectomy. Ann Intern Med 2004;141:264-71. 21. Jorgensen OD, Kronborg O, Fenger C. The Funen Adenoma Follow-up Study. Incidence and death from colorectal carcinoma in an adenoma surveillance program. Scand J Gastroenterol 1993;28:869-74. 22. Pabby A, Schoen RE, Weissfeld JL, et al. Analysis of colorectal cancer occurrence during surveillance in the dietary Polyp Prevention Trial. Gastrointest Endosc 2005;61:385-91. 23. Alberts DS, Martinez ME, Roe DJ, et al. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians’ Network. N Engl J Med 2000;342:1156-62. 24. Ee HC, Semmens JB, Hoffman NE. Complete colonoscopy rarely misses cancer. Gastrointest Endosc 2002;55:167-71. 25. Yusoff IF, Hoffman NE, Ee HC. Colonoscopic surveillance after surgery for colorectal cancer. ANZ J Surg 2003;73:3-7. 26. Ee HC, Yusoff IF. Colonoscopic surveillance after curative surgery for colorectal cancer. ANZ J Surg 2005;75:257-8. 27. Lieberman D. Endoscopic colon screening: is less more? Gastroenterology 1996;111:1385-7. 28. Rex DK, Cummings OW, Helper DJ, et al. 5-year incidence of adenomas after negative colonoscopy in asymptomatic average-risk persons [see comment]. Gastroenterology 1996;111:1178-81. 29. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US MultiSociety Task Force on Colorectal Cancer and the American Cancer Society. Gastroenterology 2006;130:1872-85.
www.giejournal.org
Volume 64, No. 6 : 2006 GASTROINTESTINAL ENDOSCOPY 947
Received April 5, 2006. Accepted August 7, 2006. Current affiliations: Department of Gastroenterology (C.H.S., H.C.E., and A.B.W.), School of Medicine and Pharmacology (I.F.Y.), Sir Charles Gairdner Hospital Unit, The University of Western Australia, Nedlands, Perth, Western Australia. Reprint requests: Ian Yusoff, MBBS, Department of Gastroenterology, Sir Charles Gairdner Hospital, Hospital Ave, Nedlands, Western Australia 6009.