- Email: [email protected]
Impact of colonoscopy preparation quality on detection of suspected colonic neoplasia
Impact of colonoscopy preparation quality on detection of suspected colonic neoplasia Gavin C. Harewood, MD, MSc, Virender K. Sharma, MD, Pat de Garmo, ANP Rochester, Minnesota, Scottsdale, Arizona, Portland, Oregon
Background: Suboptimal bowel preparation for colonoscopy can lead to missed colonic lesions. The aim of this study was to describe the impact of preparation quality on detection of suspected colonic neoplasia. Methods: Data from the Clinical Outcomes Research Initiative national endoscopic database for the period January 1, 2000 to December 31, 2001, were analyzed. Patient demographics, quality of preparation, and colonoscopy findings were abstracted from the database. Results: Overall, 93,004 colonoscopies with adequate documentation were reviewed. Preparation was adequate for 71,501 (76.9%) of these procedures. On multivariate analysis, preparation adequacy was associated with colonic lesion detection, odds ratio (OR) 1.21: 95% CI [1.16, 1.25]. Adequate preparation demonstrated a closer association with identification of “nonsignificant” lesions (polyps ≤9 mm), OR 1.23: 95% CI [1.19, 1.28], compared with “significant” lesion detection (mass lesion, polyps >9 mm), OR 1.05: 95% CI [0.98, 1.11]. Conclusions: Bowel preparation is inadequate for almost a quarter of patients undergoing colonoscopy. These results suggest that inadequate preparation quality only hinders detection of smaller lesions, while having negligible impact on detection of larger lesions. These results should be confirmed in prospective studies. (Gastrointest Endosc 2003;58:76-9.)
Colon cancer is the second leading cause of cancer death in North America, but mortality from this disease can be reduced by routine screening starting at age 50 years.1-5 Colonoscopy has been shown to be more effective than double-contrast radiography of the colon and flexible sigmoidoscopy for detection of colonic neoplasia.6,7 With the increasing use of screening colonoscopy, the impact of bowel preparation adequacy on the yield of the examination has become an important but incompletely investigated clinical question. This study used a national endoscopy database, the Clinical Outcomes Research Initiative (CORI), to characterize the impact of bowel preparation adequacy on detection of colonic lesions. PATIENTS AND METHODS Clinical Outcomes Research Initiative database The CORI database represents a consortium of 580 specialists in GI diseases at 88 sites in 24 states, selected to obtain a cross section of endoscopic practice in the United States. Participants use a computerized endoReceived December 21, 2002. Accepted March 19, 2003. Current affiliations: Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota, Department of Gastroenterology, Mayo Clinic, Scottsdale, Arizona, Oregon Health and Science University, Portland, Oregon. Reprint requests: Gavin C. Harewood, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, Charlton 8, 200 First St. SW, Rochester MN 55905. Copyright © 2003 by the American Society for Gastrointestinal Endoscopy 0016-5107/2003/$30.00 + 0 doi:10.1067/mge.2003.294 76
GASTROINTESTINAL ENDOSCOPY
scopic report generator to produce all endoscopic reports. Reports contain standard elements determined by the Standards and Practice Committee of the American Society for Gastrointestinal Endoscopy (ASGE).8 The report generator produces a paper copy of the report and simultaneously creates a data file composed of elements from the report. Each week, data from each site are uploaded to a central data repository in Portland, Oregon. These merged data can be used to examine the indications for procedures, what is found, and interventions performed at endoscopy. By linking indications and findings, it is possible to determine and compare the diagnostic yield of endoscopy based on the procedure indication. Data files are stripped of all patient and specific physician identifiers to protect both patient and physician confidentiality. After reception at the central data bank, the data files are screened for key missing fields and then merged with data from other sites for analysis. All colonoscopies performed from January 1, 2000 to December 31, 2001, were queried. Demographic information (age, gender), quality of bowel preparation, and colonic lesion detection (polyp ≤9 mm, polyp >9 mm, suspected cancer) were recorded. Polyp size was determined by the endoscopist. Suspected tumors were defined as mass lesions with characteristics suggestive of adenocarcinoma. Unfortunately, histopathologic data were not consistently retrievable for the database. The bowel preparation quality was classified as “excellent,” “good,” “fair,” “fair with adequate examination,” “fair with compromised examination,” or “poor” as judged by the endoscopist. The preparation quality classifications were dichotomized into “adequate” (excellent and good results pooled) and “inadequate” (fair and poor results pooled) for the purposes of analysis. Of VOLUME 58, NO. 1, 2003
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
G Harewood, V Sharma, P de Garmo
Table 1. Predictive value of adequate preparation (vs. inadequate preparation) for detection of colonic lesions of different sizes Lesion size
Adjusted OR*
95% CI
1.21 1.23 1.05
1.16, 1.25 1.18, 1.28 0.98, 1.11
All lesions Polyp ≤9 mm Polyp >9 mm, suspected cancer
*Obtained from multivariate logistic regression.
note, definitions for excellent, good, fair, or poor preparation quality were not provided to the endoscopists. Statistical analysis Proportions were compared by using the chi-square test. A p value of 0.05 was considered statistically significant. To assess the association between preparation adequacy and colonic lesion detection, a multivariate logistic regression analysis was performed to account for the possible confounding effects of age and gender. Separate logistic regression models were developed to assess the predictive value of adequate bowel preparation on both “significant” lesions (suspected cancer and polyps >9 mm) and “nonsignificant” lesions (polyps ≤9 mm). Odds ratios and their 95% confidence intervals served to describe the influence of preparation adequacy on lesion detection.
lesions, OR 1.23: 95% CI [1.19, 1.28] compared with significant lesions, OR 1.05: 95% CI [0.98, 1.11] (Table 1).
RESULTS
DISCUSSION
Data from a total of 113,272 colonoscopy procedures were recorded in the CORI database between January 1, 2000 and December 31, 2001, of which 93,004 were complete procedures (i.e., completed to cecum) with complete documentation of age, gender, preparation quality, and endoscopic findings. This latter group comprised the patient cohort for analysis. Of the 93,004 colonoscopes with complete data, the bowel preparation was rated as adequate in 71,501 (76.9%). Overall, suspected neoplasia was detected at 26,490 (28.5%) colonoscopies, 20,822 (29.1%) with adequate preparation and 5668 (26.4%) with inadequate preparation (p < 0.0001; Fig. 1). Significant lesions (polyp >9 mm, mass lesion) were identified at 6783 colonoscopies (7.3%), 5207 (7.3%) with adequate preparation, and 1576 (7.3%) with inadequate preparation (p = 0.82). Nonsignificant lesions (polyp ≤9 mm) were identified at 19,707 colonoscopies (21.2%), 15,615 (21.8%) with adequate preparation, and 4092 (19.0%) with inadequate preparation (p < 0.0001). Adjusting for age and gender, adequate preparation quality was predictive of detection of suspected neoplasia, OR 1.21: 95% CI [1.16, 1.25]. However, when lesions were classified into significant (suspected cancer, polyp >9 mm) and nonsignificant (polyp ≤9 mm), adequate preparation quality only appeared to influence detection of nonsignificant
This study addresses an important clinical question, namely, the impact of preparation adequacy on identification of suspected colonic neoplasia. The findings demonstrate that preparation adequacy primarily impacts the ability to discern smaller lesions while having negligible effect on the detection of larger lesions. Some data exist with respect to usual rates of adequate bowel preparation for colonoscopy. The finding in the present study of adequate quality preparation in 77% of all patients is similar to quoted rates from prior studies of 79%,9 78%,10 and 70%.11,12 However, these data were gathered in tertiary-care settings, thereby limiting generalizability. The strength of the CORI database lies in its ability to provide data from a diverse array of endoscopy centers representing all levels of medical care: community, academic, and Veterans Affairs institutions settings.13,14 The classification of significant and nonsignificant lesions used in the current study, which was based on lesion size and endoscopic appearance, requires comment. This arbitrary classification has come into use in part because of the low incidence and prevalence of colorectal cancer, especially in screening studies.6,15,16 Although the large polyp is a convenient proxy for colorectal cancer, its use as an outcome measure may be misleading because the
VOLUME 58, NO. 1, 2003
Figure 1. Colonic lesion detection rates according to preparation quality.
GASTROINTESTINAL ENDOSCOPY
77
G Harewood, V Sharma, P de Garmo
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
natural history of this lesion is unknown. Stryker et al.17 found that large polyps (>1 cm in diameter), when left intact, progressed to colorectal cancer at a rate of about 1% per year; although polyps 9 mm or less in diameter were not assessed, they are considered to grow at a slower rate. In the present study, inadequate preparation primarily impacted detection of polyps 9 mm or less in diameter. Therefore, if screening colonoscopy is repeated within 5 years, it is likely that a polyp 9 mm or less in diameter at the index colonoscopy will be removed before progression to cancer. However, if the screening interval is widened to 10 years, the likelihood of adenomatous progression is higher. Recognizing that Medicare currently provides coverage for screening colonoscopy every 10 years, this compromised ability to detect polyps 9 mm or less in diameter may represent an increased risk for adenomatous progression over a 10-year time interval. Database research has many important limitations with respect to clinical practice. First, the CORI database lacks consistent recording of histopathologic data and of clinical follow-up, thereby depriving the researcher of a true reference standard that can be used to determine the nature of colonic lesions (incorporation of pathology reports into the CORI database is currently being studied). In the present study, a surrogate end point for serious pathology was used (polyp >9 mm, mass lesion), based on the assumption that most of these lesions would be neoplastic. Second, the scale used to measure bowel preparation quality (adequate vs. inadequate) has not been standardized or validated. Formal definitions of excellent, good, fair, or poor quality were not provided to endoscopists, thereby creating a potential for interobserver variation. Nevertheless, use of the definitions “adequate” and “inadequate” has been shown to be subject to less variability than a more complex scoring system.18 Third, this study is limited by incomplete documentation of procedures. A goal of the CORI project is to minimize inconvenience to the physician by making data entry part of routine practice. To achieve this goal, the computer program makes entry of information mandatory for only a few fields. Consequently, data provided were sufficient for analysis for only 82% of the colonoscopy procedures. This incomplete documentation increases the potential for bias in the data. Fourth, the physicians in the CORI cohort may not be representative of the “average GI specialist.” Although CORI attempts to reflect all levels of GI practice across the United States, it may be argued that physicians willing to use a computerized endoscopic report generator are highly motivated clinicians who want to use the information about 78
GASTROINTESTINAL ENDOSCOPY
their practice to enhance quality of care. Thus, physician practices that use CORI may differ from those that do not incorporate such quality improvement measures. However, the CORI participating physicians in the data set used in the present study were not aware that the analysis would be performed, thus reducing the opportunity for selection bias by the participating GI physicians. Nevertheless, the limitations outlined above underscore the importance of prospective studies with longitudinal followup to characterize the impact of bowel preparation adequacy on colonic neoplasia detection. In summary, although the CORI database facilitates statistically powerful descriptive studies, it is primarily a hypothesis-generating instrument that can be used to identify clinical questions to be pursued in prospective studies. The present study is such an example. It represents a “snapshot in time” of colonoscopy practice in a variety of patient care settings. The findings suggest that inadequate colonoscopy preparation quality only hinders detection of smaller lesions. However, this retrospective analysis should serve primarily to prompt prospective studies of the impact of bowel preparation quality on lesion detection. ACKNOWLEDGMENT The data in this manuscript were obtained from the Clinical Research Initiative National Endoscopic Database (CORI-NED); CORI is supported by grants of the ASGE, Bard Interventional Products, AstraZeneca Pharmaceuticals, and support from National Institutes of Health (NIDDK) U01-DK57132-01. REFERENCES 1. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 1993;328:1365-71. 2. Hardcastle JD, Chamberlain JO, Robinson MH, Moss SM, Amar SS, Balfour TW, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472-7. 3. Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467-71. 4. Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326:653-7. 5. Newcomb PA, Norfleet RG, Storer BE, Surawicz TS, Marcus PM. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992;84:1572-5. 6. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000;343:162-8. 7. Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169-74. VOLUME 58, NO. 1, 2003
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
8. Computer Committee. Standard format and content of the endoscopic procedure report. Manchester (MA): American Society for Gastrointestinal Endoscopy; 1992. 9. Cohen SM, Wexner SD, Binderow SR, Nogueras JJ, Daniel N, Ehrenpreis ED, et al. Prospective, randomized, endoscopicblinded trial comparing precolonoscopy bowel cleansing methods. Dis Colon Rectum 1994;37:689-96. 10. Ness RM, Manam R, Hoen H, Chalasani N. Predictors of inadequate bowel preparation for colonoscopy. Am J Gastroenterol 2001;96:1797-802. 11. Kolts BE, Lyles WE, Achem SR, Burton L, Geller AJ, MacMath T. A comparison of the effectiveness and patient tolerance of oral sodium phosphate, castor oil, and standard electrolyte lavage for colonoscopy or sigmoidoscopy preparation. Am J Gastroenterol 1993;88:1218-23. 12. Marshall JB, Pineda JJ, Barthel JS, King PD. Prospective, randomized trial comparing sodium phosphate solution with polyethylene glycol-electrolyte lavage for colonoscopy preparation. Gastrointest Endosc 1993;39:631-4.
G Harewood, V Sharma, P de Garmo
13. Lieberman DA, De Garmo PL, Fleischer DE, Eisen GM, Helfand M. Patterns of endoscopy use in the United States. Gastroenterology 2000;118:619-24. 14. McCashland TM, Brand R, Lyden E, de Garmo P. Gender differences in colorectal polyps and tumors. Am J Gastroenterol 2001;96:882-6. 15. Levin TR, Palitz A, Grossman S, Conell C, Finkler L, Ackerson L, et al. Predicting advanced proximal colonic neoplasia with screening sigmoidoscopy. JAMA 1999;281:1611-7. 16. Read TE, Read JD, Butterly LF. Importance of adenomas 5 mm or less in diameter that are detected by sigmoidoscopy. N Engl J Med 1997;336:8-12. 17. Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009-13. 18. Aronchick CA, Lipshutz WH, Wright SH, Dufrayne F, Bergman G. A novel tableted purgative for colonoscopic preparation: efficacy and safety comparisons with Colyte and Fleet Phospho-Soda. Gastrointest Endosc 2000;52:346-52.
Submission of ACCEPTED manuscript on diskette Gastrointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1). Please refer to complete Instructions to Authors in the most recent January or July issue of the Journal.
VOLUME 58, NO. 1, 2003
GASTROINTESTINAL ENDOSCOPY
79
Background: Suboptimal bowel preparation for colonoscopy can lead to missed colonic lesions. The aim of this study was to describe the impact of preparation quality on detection of suspected colonic neoplasia. Methods: Data from the Clinical Outcomes Research Initiative national endoscopic database for the period January 1, 2000 to December 31, 2001, were analyzed. Patient demographics, quality of preparation, and colonoscopy findings were abstracted from the database. Results: Overall, 93,004 colonoscopies with adequate documentation were reviewed. Preparation was adequate for 71,501 (76.9%) of these procedures. On multivariate analysis, preparation adequacy was associated with colonic lesion detection, odds ratio (OR) 1.21: 95% CI [1.16, 1.25]. Adequate preparation demonstrated a closer association with identification of “nonsignificant” lesions (polyps ≤9 mm), OR 1.23: 95% CI [1.19, 1.28], compared with “significant” lesion detection (mass lesion, polyps >9 mm), OR 1.05: 95% CI [0.98, 1.11]. Conclusions: Bowel preparation is inadequate for almost a quarter of patients undergoing colonoscopy. These results suggest that inadequate preparation quality only hinders detection of smaller lesions, while having negligible impact on detection of larger lesions. These results should be confirmed in prospective studies. (Gastrointest Endosc 2003;58:76-9.)
Colon cancer is the second leading cause of cancer death in North America, but mortality from this disease can be reduced by routine screening starting at age 50 years.1-5 Colonoscopy has been shown to be more effective than double-contrast radiography of the colon and flexible sigmoidoscopy for detection of colonic neoplasia.6,7 With the increasing use of screening colonoscopy, the impact of bowel preparation adequacy on the yield of the examination has become an important but incompletely investigated clinical question. This study used a national endoscopy database, the Clinical Outcomes Research Initiative (CORI), to characterize the impact of bowel preparation adequacy on detection of colonic lesions. PATIENTS AND METHODS Clinical Outcomes Research Initiative database The CORI database represents a consortium of 580 specialists in GI diseases at 88 sites in 24 states, selected to obtain a cross section of endoscopic practice in the United States. Participants use a computerized endoReceived December 21, 2002. Accepted March 19, 2003. Current affiliations: Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota, Department of Gastroenterology, Mayo Clinic, Scottsdale, Arizona, Oregon Health and Science University, Portland, Oregon. Reprint requests: Gavin C. Harewood, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, Charlton 8, 200 First St. SW, Rochester MN 55905. Copyright © 2003 by the American Society for Gastrointestinal Endoscopy 0016-5107/2003/$30.00 + 0 doi:10.1067/mge.2003.294 76
GASTROINTESTINAL ENDOSCOPY
scopic report generator to produce all endoscopic reports. Reports contain standard elements determined by the Standards and Practice Committee of the American Society for Gastrointestinal Endoscopy (ASGE).8 The report generator produces a paper copy of the report and simultaneously creates a data file composed of elements from the report. Each week, data from each site are uploaded to a central data repository in Portland, Oregon. These merged data can be used to examine the indications for procedures, what is found, and interventions performed at endoscopy. By linking indications and findings, it is possible to determine and compare the diagnostic yield of endoscopy based on the procedure indication. Data files are stripped of all patient and specific physician identifiers to protect both patient and physician confidentiality. After reception at the central data bank, the data files are screened for key missing fields and then merged with data from other sites for analysis. All colonoscopies performed from January 1, 2000 to December 31, 2001, were queried. Demographic information (age, gender), quality of bowel preparation, and colonic lesion detection (polyp ≤9 mm, polyp >9 mm, suspected cancer) were recorded. Polyp size was determined by the endoscopist. Suspected tumors were defined as mass lesions with characteristics suggestive of adenocarcinoma. Unfortunately, histopathologic data were not consistently retrievable for the database. The bowel preparation quality was classified as “excellent,” “good,” “fair,” “fair with adequate examination,” “fair with compromised examination,” or “poor” as judged by the endoscopist. The preparation quality classifications were dichotomized into “adequate” (excellent and good results pooled) and “inadequate” (fair and poor results pooled) for the purposes of analysis. Of VOLUME 58, NO. 1, 2003
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
G Harewood, V Sharma, P de Garmo
Table 1. Predictive value of adequate preparation (vs. inadequate preparation) for detection of colonic lesions of different sizes Lesion size
Adjusted OR*
95% CI
1.21 1.23 1.05
1.16, 1.25 1.18, 1.28 0.98, 1.11
All lesions Polyp ≤9 mm Polyp >9 mm, suspected cancer
*Obtained from multivariate logistic regression.
note, definitions for excellent, good, fair, or poor preparation quality were not provided to the endoscopists. Statistical analysis Proportions were compared by using the chi-square test. A p value of 0.05 was considered statistically significant. To assess the association between preparation adequacy and colonic lesion detection, a multivariate logistic regression analysis was performed to account for the possible confounding effects of age and gender. Separate logistic regression models were developed to assess the predictive value of adequate bowel preparation on both “significant” lesions (suspected cancer and polyps >9 mm) and “nonsignificant” lesions (polyps ≤9 mm). Odds ratios and their 95% confidence intervals served to describe the influence of preparation adequacy on lesion detection.
lesions, OR 1.23: 95% CI [1.19, 1.28] compared with significant lesions, OR 1.05: 95% CI [0.98, 1.11] (Table 1).
RESULTS
DISCUSSION
Data from a total of 113,272 colonoscopy procedures were recorded in the CORI database between January 1, 2000 and December 31, 2001, of which 93,004 were complete procedures (i.e., completed to cecum) with complete documentation of age, gender, preparation quality, and endoscopic findings. This latter group comprised the patient cohort for analysis. Of the 93,004 colonoscopes with complete data, the bowel preparation was rated as adequate in 71,501 (76.9%). Overall, suspected neoplasia was detected at 26,490 (28.5%) colonoscopies, 20,822 (29.1%) with adequate preparation and 5668 (26.4%) with inadequate preparation (p < 0.0001; Fig. 1). Significant lesions (polyp >9 mm, mass lesion) were identified at 6783 colonoscopies (7.3%), 5207 (7.3%) with adequate preparation, and 1576 (7.3%) with inadequate preparation (p = 0.82). Nonsignificant lesions (polyp ≤9 mm) were identified at 19,707 colonoscopies (21.2%), 15,615 (21.8%) with adequate preparation, and 4092 (19.0%) with inadequate preparation (p < 0.0001). Adjusting for age and gender, adequate preparation quality was predictive of detection of suspected neoplasia, OR 1.21: 95% CI [1.16, 1.25]. However, when lesions were classified into significant (suspected cancer, polyp >9 mm) and nonsignificant (polyp ≤9 mm), adequate preparation quality only appeared to influence detection of nonsignificant
This study addresses an important clinical question, namely, the impact of preparation adequacy on identification of suspected colonic neoplasia. The findings demonstrate that preparation adequacy primarily impacts the ability to discern smaller lesions while having negligible effect on the detection of larger lesions. Some data exist with respect to usual rates of adequate bowel preparation for colonoscopy. The finding in the present study of adequate quality preparation in 77% of all patients is similar to quoted rates from prior studies of 79%,9 78%,10 and 70%.11,12 However, these data were gathered in tertiary-care settings, thereby limiting generalizability. The strength of the CORI database lies in its ability to provide data from a diverse array of endoscopy centers representing all levels of medical care: community, academic, and Veterans Affairs institutions settings.13,14 The classification of significant and nonsignificant lesions used in the current study, which was based on lesion size and endoscopic appearance, requires comment. This arbitrary classification has come into use in part because of the low incidence and prevalence of colorectal cancer, especially in screening studies.6,15,16 Although the large polyp is a convenient proxy for colorectal cancer, its use as an outcome measure may be misleading because the
VOLUME 58, NO. 1, 2003
Figure 1. Colonic lesion detection rates according to preparation quality.
GASTROINTESTINAL ENDOSCOPY
77
G Harewood, V Sharma, P de Garmo
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
natural history of this lesion is unknown. Stryker et al.17 found that large polyps (>1 cm in diameter), when left intact, progressed to colorectal cancer at a rate of about 1% per year; although polyps 9 mm or less in diameter were not assessed, they are considered to grow at a slower rate. In the present study, inadequate preparation primarily impacted detection of polyps 9 mm or less in diameter. Therefore, if screening colonoscopy is repeated within 5 years, it is likely that a polyp 9 mm or less in diameter at the index colonoscopy will be removed before progression to cancer. However, if the screening interval is widened to 10 years, the likelihood of adenomatous progression is higher. Recognizing that Medicare currently provides coverage for screening colonoscopy every 10 years, this compromised ability to detect polyps 9 mm or less in diameter may represent an increased risk for adenomatous progression over a 10-year time interval. Database research has many important limitations with respect to clinical practice. First, the CORI database lacks consistent recording of histopathologic data and of clinical follow-up, thereby depriving the researcher of a true reference standard that can be used to determine the nature of colonic lesions (incorporation of pathology reports into the CORI database is currently being studied). In the present study, a surrogate end point for serious pathology was used (polyp >9 mm, mass lesion), based on the assumption that most of these lesions would be neoplastic. Second, the scale used to measure bowel preparation quality (adequate vs. inadequate) has not been standardized or validated. Formal definitions of excellent, good, fair, or poor quality were not provided to endoscopists, thereby creating a potential for interobserver variation. Nevertheless, use of the definitions “adequate” and “inadequate” has been shown to be subject to less variability than a more complex scoring system.18 Third, this study is limited by incomplete documentation of procedures. A goal of the CORI project is to minimize inconvenience to the physician by making data entry part of routine practice. To achieve this goal, the computer program makes entry of information mandatory for only a few fields. Consequently, data provided were sufficient for analysis for only 82% of the colonoscopy procedures. This incomplete documentation increases the potential for bias in the data. Fourth, the physicians in the CORI cohort may not be representative of the “average GI specialist.” Although CORI attempts to reflect all levels of GI practice across the United States, it may be argued that physicians willing to use a computerized endoscopic report generator are highly motivated clinicians who want to use the information about 78
GASTROINTESTINAL ENDOSCOPY
their practice to enhance quality of care. Thus, physician practices that use CORI may differ from those that do not incorporate such quality improvement measures. However, the CORI participating physicians in the data set used in the present study were not aware that the analysis would be performed, thus reducing the opportunity for selection bias by the participating GI physicians. Nevertheless, the limitations outlined above underscore the importance of prospective studies with longitudinal followup to characterize the impact of bowel preparation adequacy on colonic neoplasia detection. In summary, although the CORI database facilitates statistically powerful descriptive studies, it is primarily a hypothesis-generating instrument that can be used to identify clinical questions to be pursued in prospective studies. The present study is such an example. It represents a “snapshot in time” of colonoscopy practice in a variety of patient care settings. The findings suggest that inadequate colonoscopy preparation quality only hinders detection of smaller lesions. However, this retrospective analysis should serve primarily to prompt prospective studies of the impact of bowel preparation quality on lesion detection. ACKNOWLEDGMENT The data in this manuscript were obtained from the Clinical Research Initiative National Endoscopic Database (CORI-NED); CORI is supported by grants of the ASGE, Bard Interventional Products, AstraZeneca Pharmaceuticals, and support from National Institutes of Health (NIDDK) U01-DK57132-01. REFERENCES 1. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schuman LM, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 1993;328:1365-71. 2. Hardcastle JD, Chamberlain JO, Robinson MH, Moss SM, Amar SS, Balfour TW, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472-7. 3. Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467-71. 4. Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992;326:653-7. 5. Newcomb PA, Norfleet RG, Storer BE, Surawicz TS, Marcus PM. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992;84:1572-5. 6. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000;343:162-8. 7. Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169-74. VOLUME 58, NO. 1, 2003
Colonic neoplasia: detection and quality of colonoscopy bowel preparation
8. Computer Committee. Standard format and content of the endoscopic procedure report. Manchester (MA): American Society for Gastrointestinal Endoscopy; 1992. 9. Cohen SM, Wexner SD, Binderow SR, Nogueras JJ, Daniel N, Ehrenpreis ED, et al. Prospective, randomized, endoscopicblinded trial comparing precolonoscopy bowel cleansing methods. Dis Colon Rectum 1994;37:689-96. 10. Ness RM, Manam R, Hoen H, Chalasani N. Predictors of inadequate bowel preparation for colonoscopy. Am J Gastroenterol 2001;96:1797-802. 11. Kolts BE, Lyles WE, Achem SR, Burton L, Geller AJ, MacMath T. A comparison of the effectiveness and patient tolerance of oral sodium phosphate, castor oil, and standard electrolyte lavage for colonoscopy or sigmoidoscopy preparation. Am J Gastroenterol 1993;88:1218-23. 12. Marshall JB, Pineda JJ, Barthel JS, King PD. Prospective, randomized trial comparing sodium phosphate solution with polyethylene glycol-electrolyte lavage for colonoscopy preparation. Gastrointest Endosc 1993;39:631-4.
G Harewood, V Sharma, P de Garmo
13. Lieberman DA, De Garmo PL, Fleischer DE, Eisen GM, Helfand M. Patterns of endoscopy use in the United States. Gastroenterology 2000;118:619-24. 14. McCashland TM, Brand R, Lyden E, de Garmo P. Gender differences in colorectal polyps and tumors. Am J Gastroenterol 2001;96:882-6. 15. Levin TR, Palitz A, Grossman S, Conell C, Finkler L, Ackerson L, et al. Predicting advanced proximal colonic neoplasia with screening sigmoidoscopy. JAMA 1999;281:1611-7. 16. Read TE, Read JD, Butterly LF. Importance of adenomas 5 mm or less in diameter that are detected by sigmoidoscopy. N Engl J Med 1997;336:8-12. 17. Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL. Natural history of untreated colonic polyps. Gastroenterology 1987;93:1009-13. 18. Aronchick CA, Lipshutz WH, Wright SH, Dufrayne F, Bergman G. A novel tableted purgative for colonoscopic preparation: efficacy and safety comparisons with Colyte and Fleet Phospho-Soda. Gastrointest Endosc 2000;52:346-52.
Submission of ACCEPTED manuscript on diskette Gastrointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1). Please refer to complete Instructions to Authors in the most recent January or July issue of the Journal.
VOLUME 58, NO. 1, 2003
GASTROINTESTINAL ENDOSCOPY
79