- Email: [email protected]
Tu1355 The Boston Classification of Excrement and Residue (B-Clear): a Valid and Reliable Descriptive Scheme for Bowel Preparation Research
Abstracts
preparation has a deleterious effect on adenoma detection, physicians may have varying standards for declaring bowel preparation to be suboptimal. We hypothesized that providers who have higher standards for optimal preparation, and hence higher suboptimal preparation rates (SPRs) may be more fastidious in their examination of the colon and therefore have higher ADRs. Methods: Patients ⱖ50 years old who underwent screening colonoscopy at our medical center during 2011 were identified via the electronic endoscopy database. Examinations were excluded if the provider performed fewer than 50 screening colonoscopies during the study period, or if the bowel preparation rating was not recorded. Preparations labeled “fair,” “poor,” or “unsatisfactory” were considered suboptimal. Adenomas were confirmed via cross-reference with the histology database. The ADR was calculated for each provider, and was correlated with the provider’s suboptimal preparation rate (SPR). Multivariate logistic regression was performed to identify independent patient and provider factors associated with ADR, including provider quartile for SPR. Results: Of 2,134 screening colonoscopies performed during the study period, 1,649 examinations from 11 separate gastroenterologists were eligible for inclusion. The majority (59%) of subjects were female, and the mean age was 61 years. Preparation was considered suboptimal in 22% of examinations. Procedure volume ranged from 52 to 353 examinations per provider, with a mean of 150. SPR varied widely among providers, ranging from 3% to 40%. Overall ADR was 22.6%, with a range of 13.4% to 31.6%. SPR was not significantly correlated with ADR (r⫽-0.22, p⫽0.51). Adenoma detection was more common in male patients (26.9% vs 19.6%, p ⫽ 0.0005) and in older patients (37.2% in ⱖ80y vs 19.0% in 50-59y, p⫽0.003), Adenoma detection was lower among patients with suboptimal preparation (19.5% vs 23.5%, p⫽0.10). On multivariate analysis, adenoma detection was associated with male gender (OR 1.55, 95% CI 1.231.97), and older age (OR 2.58 for ⱖ80 vs 50-59, 95% CI 1.35-4.95). After adjusting for age and gender, adenoma detection was associated with the highest quartile of provider procedure volume (OR 1.59, 95% CI 1.05-2.42), but remained independent of provider SPR (p⫽0.28). Sensitivity analysis with exclusion of “fair” exams did not change these associations. Conclusions: SPRs vary widely between providers, but SPR was not correlated with ADR. This suggests that a high SPR is not a marker of higher quality standards and expectations by the provider. Procedure volume was correlated with increased adenoma detection. The impact of physician personality traits on colonoscopy performance requires further study.
or PEG 3350 solution mixed with ascorbic acid (MoviPrep) but does provide patients with an overall better prep experience. Descriptive Summary Table 1. Adjusted Analysis - Based on Logistic Regression (1) Variable Gender Male Female Age Median (IQR) Mean (SD) Range % Prep Taken ⬎80% 60-80% 40-60% Time of Procedure A.M. P.M. Prep Experience Easy Acceptable Difficult Very Difficult Unable to complete
All (nⴝ317)
2L (nⴝ161)
4L (nⴝ156)
164 (51.7) 153 (48.3)
79 (49.1) 82 (50.9)
85 (54.5) 71 (45.5)
57 (50.1, 62.9) 56 (11.3) 20-87
56 (50.2, 62.2) 56 (10.6) 25-80
58 (49.9, 63.8) 57 (12.0) 20-87
308 (97.2) 7 (2.2) 2 (0.6)
158 (98.1) 3 (1.9) 0 (0)
150 (96.2) 4 (2.6) 2 (1.3)
187 (59) 130 (41)
96 (59.6) 65 (40.4)
91 (58.3) 65 (41.7)
110 (34.7) 151 (47.6) 43 (13.6) 12 (3.8) 1 (0.3)
69 (42.9) 76 (47.2) 15 (9.3) 1 (0.6) 0 (0.0)
41 (26.3) 75 (48.1) 28 (17.9) 11 (7.1) 1 (0.6)
Table 2. (1) Adjusted for age and gender. Also adjusted for endoscopist and timing of procedure for the analysis of BBPS (2) Effect of 2L relative to 4L Peglyte Outcome BBPS⬎8 BBPS⬎7 Had Adverse Event Would be Willing to Repeat Found Prep Easy or Acceptable Not Able to Complete Prep (⬍80%)
Tu1354 Efficacy and Patient Tolerability of Split Dose 2L vs. Split Dose 4L Peglyte: a Prospective Randomized Controlled Trial Mayur Brahmania*, George Ou, Andrew M. Round, Jessica Tong, Ricky W. Kwok, Cherry Galorport, Hin Hin Ko, Eric C. LAM, Brian Bressler, Jennifer J. Telford, Robert A. Enns Gastroenterology, St Paul’s Hospital, Vancouver, BC, Canada Introduction: Despite the evolution of bowel preparation for endoscopic investigation it remains one of the most frequently cited deterrents for patient acceptance and cooperation prior to colonoscopy. In an effort to advance the quality of preparations used and patient compliance we compared split dosing of 2L Peglyte with 15mg of bisacodyl vs split dosing of 4L Peglyte. Our primary endpoint was to assess the efficacy of prep and secondarily patient tolerability. Patients and Methods: This study was a randomized, controlled, single centre, endoscopist-blinded, non-inferiority trial involving five endoscopists. All adults (ⱖ19 years of age) who were referred for outpatient colonoscopy were considered for our study. Exclusion criteria included those with constipation (⬍3 BM/week with or without laxative use), suspected small bowel obstruction, severe inflammatory bowel disease with known stricturing or penetrating disease, and history of colonic resection. Eligible patients were randomized to one of two bowel preparations through the use of concealed allocation by a scheduling assistant (blinded) in a one-to-one ratio. Bowel preparations included: (1) 4L of PegLyte (4L PEG 3350 electrolyte solution) or (2) 2L of Peglyte (2L PEG 3350 electrolyte solution and 15mg bisacodyl) as currently manufactured and marketed. Both preparations were taken as a split dose. The primary endpoint was bowel cleansing efficacy as rated by a Boston Bowel Prep Score (BBPS) score of greater than 8 (scores of 7 or above are generally felt to provide a good visualization of the colon). Sample size study was based on an assumption of alpha⫽0.05 and beta⫽0.20. The estimated sample size was 158 patients in each arm. Our secondary outcome was patient tolerability to 2L PegLyte based on a six-question survey. Results: Our study finds using 2L PegLyte in a split dosing regimen is as efficacious as using a 4L split dosing regimen using the Boston Bowel Prep Score. Additionally, 2L PegLyte in a split dose showed statistically improved patient tolerability and willingness to repeat colonoscopic procedures. See Tables 1 & 2 Conclusion: The current standard of gut lavage (4L PegLyte in a split dose) is often a large volume for patients to consume, often leading to non compliance and poor bowel preparation. Our non inferiority study demonstrates endoscopists can safely use 2L PegLyte in a split dosing fashion to achieve similar results using 4L of PegLyte in a split dose. In addition, the low volume PEG was statistically better tolerated than the conventional 4L volume. It is unclear how 2L PEG split dose compares to other low volume agents available such as split dose sodium picosulfate (Pico- Salax)
www.giejournal.org
Odds Ratio (2)
95% CI
p-Value
0.81 0.72 1.30 34.72 3.58
(0.51, 1.29) (0.42, 1.24) (0.82, 2.04) (8.12, 148.43) (1.87, 6.85)
0.369 0.231 0.265 ⬍0.001 ⬍0.001
0.40
(0.10, 1.68)
0.212
Tu1355 The Boston Classification of Excrement and Residue (B-Clear): a Valid and Reliable Descriptive Scheme for Bowel Preparation Research Pushpak Taunk*, Jonah N. Rubin, Loc Ton, Helen Lee, Audrey H. Calderwood, Brian C. Jacobson Boston University Medical Center, Boston, MA Purpose: Existing bowel preparation rating scales are useful for clinical purposes but may not be sufficiently descriptive of the residual fecal material and types of liquids encountered during colonoscopy for all types of bowel preparation research. We sought to validate a novel bowel preparation classification scheme that describes both solids and liquids encountered during colonoscopy. Methods: The B-CLEAR assigns a categorical variable based on the type of solid material (NS - no solid, MS - movable solid, IS - immovable solid) and liquid material (NL - no liquid, L - liquid, F- film, G- gelatinous) encountered during colonoscopy. These categories are then combined to describe what was found in each segment of the colon (right, transverse, left). For example, R:NS/NL would be a perfectly clean right colon while T:IS/G would be a very poor preparation in the transverse colon. When more than one type of solid or liquid is found in the same segment, the “worse” of the two types is listed. Washing/suctioning is permitted, but this won’t improve the rating; it is used to help assign a classification (e.g. can a solid be moved to permit mucosal visualization?). After reviewing a set of training images to learn the new classification scheme, 14 gastroenterologists classified a testing set of 16 different images representing the full spectrum of categories. Inter-rater reliability was measured using intraclass correlation coefficients (ICCs). B-CLEAR categories were then assigned prospectively during outpatient colonoscopies and validated by correlation with volume of water flushes used, volume of effluent suctioned and colonoscope withdrawal times. We also compared B-CLEAR assignments with segmental Boston Bowel Preparation Scale (BBPS) scores. Results: The ICCs for solid, liquid and combined solid⫹liquid B-CLEAR classifications were 0.78, 0.89 and 0.79,
Volume 77, No. 5S : 2013
GASTROINTESTINAL ENDOSCOPY
AB511
Abstracts
respectively, indicating good inter-observer reliability. In the validation portion of the study, the worst B-CLEAR score in each case correlated with washing and suctioning volumes (Pearson’s correlation coefficient 0.56, p⫽0.003 for both) and was non-significantly correlated with colonoscope withdrawal time (0.37, p⫽0.06). B-CLEAR assignments and BPPS scores were not correlated, suggesting each scale provides unique information about the preparation. Conclusion: The B-CLEAR classification scheme is a valid and reliable tool for describing residual fecal material encountered during colonoscopy. With this new scheme, we can determine how various types of residual material affect an endoscopist’s ability to visualize the colon. Because the B-CLEAR captures specific information about what is encountered in each colon segment, it may prove a useful tool for studying the impact of novel washing devices or comparing dietary instructions prior to colonoscopy.
Tu1356 Adenoma Detection in Excellent Versus Good Bowel Preparation for Colonoscopy Danielle M. Tholey*1, Archana Anantharaman1, Corbett E. Shelton1, Robert Frankel1, Paurush Shah1, Amy Coan2, Sarah E. Hegarty3, Benjamin E. Leiby3, David M. Kastenberg4 1 Medicine, TJUH, Philadelphia, PA; 2Medicine, Thomas Jefferson Medical College, Philadelphia, PA; 3Biostatistics, TJUH, Philadelphia, PA; 4Gastroenterology, TJUH, Philadelphia, PA Adequate bowel preparation (prep) permits detection of ⬎5 mm polyps anywhere in the colon. Our definition of adequate also requires endoscopist adherence to published surveillance intervals not be shortened in response to prep quality. Within the category of adequate, a good prep requires excessive flushing and/or suctioning whereas excellent does not. Aim: To evaluate whether the adenoma detection rate (ADR) is superior with an excellent vs. good prep. Methods: Retrospective review of consecutive colonoscopies with good or excellent prep from 9/1/11-9/1/12 at a single academic center. Endoworks™ (Olympus), EMR, and pathology data bases utilized for prep quality, indication, demographics, medical history, adenoma, and cancer data. Primary outcome was ADR (# pt’s with ⬎ 1 adenoma); secondary were adenomas/colonoscopy (# adenomas/# colonoscopies), adenoma distribution, advanced and sessile serrated adenomas (SSA), and cancer (CRC). Exclusions ⫽ age ⬍18, IBD, or Familial Polyposis. ADR was compared for excellent and good preparations using logistic regression. Secondary endpoints were compared using Poisson regression. Outcome comparisons adjusted for confounding variables in demography and history. Assuming an ADR⫽30%, with 70% good and 30% excellent preps, ⬃ 4,400 colonoscopies were necessary to find a 1.15 RR between groups with 82% power using a 2-sided alpha⫽.05. Results: 5,113 colonoscopies with adequate prep were reviewed (good ⫽ 3112, excellent ⫽ 2001). The excellent group was younger and more often had a FHx of CRC and polyps, and indication of screening; and, less frequently, a personal history of CRC and adenoma, and surveillance indication (all, P⬍0.0001). Neither ADR nor adenoma/colonoscopy was superior for the excellent group (respectively, OR 0.97, CI 0.85-1.11, p⫽0.618; IRR 0.98, CI 0.90-1.07, p⫽0.705). These findings were consistent for the proximal and distal colon (respectively, p⫽0.426, p⫽0.764). ADR and adenomas/ colonoscopy for good and excellent were, respectively, 29% vs 26% and 0.499 vs 0.437. The excellent group had superior detection of SSA’s (IRR 1.66, CI 1.142.40, p⫽0.008) and advanced adenomas (IRR 1.37, CI 1.09-1.72, p⫽0.007), but was not superior in detecting CRC (IRR .268, CI .08-.91, p⫽0.035). ADR and adenoma/colonoscopy increased consistently with age (p⬍0.001), and was greater in men (p⬍0.001) and those with FHx of polyps (respectively, p⫽0.002; p⫽0.001), personal history of adenoma (respectively, p⬍0.001; p⬍0.001), and a surveillance indication (respectively, p⬍0.003; p⫽0.007). Adenoma/colonoscopy increased with a personal history of CRC (p⫽0.003). Conclusion: In pts with adequate colon preparation, excellent cleansing as compared to good does not improve ADR or adenoma/colonoscopy. However, excellent cleansing is associated with superior detection of high risk lesions including advanced adenomas and SSA’s.
Tu1357 Colonoscopy Preparation Quality - Is Fair Good Enough? ADAM B. Gluskin*1, Nikhil Martis2, Weihua Gao3, Vijay S. Khiani2 1 College of Medicine, University of Illinois at Chicago, Chicago, IL; 2 Medicine, University of Illinois at Chicago, Chicago, IL; 3Center for Clinical and Translational Science, University of Illinois at Chicago, Chicago, IL Introduction: Many factors influence colonoscopy quality, one of which is bowel preparation. Time to next colonoscopy is frequently shortened when prep is rated as poor. However, there is little data regarding the proper screening interval for individuals with fair prep quality. In addition to prep quality, two other important indicators, Adenoma Detection Rate (ADR) and Adenoma Miss Rate (AMR), have emerged as tools to assess colonoscopy quality. ADR is the percentage of colonoscopies that have a detected adenoma and AMR is the
percent of adenomas that are detected in a repeat colonoscopy divided by the total number of adenomas found in both the original and repeat procedures. This study examined how ADR and AMR vary across preparation quality. Methods: A retrospective analysis was performed that examined 4909 colonoscopies completed at the University of Illinois Hospital between 2007 and 2011. The exclusion criteria were a history of colorectal cancer, a history of Inflammatory Bowel Disease, patients with Familial Adenomatous Polyposis, a history of colon resection, procedures failing to reach the cecum, procedures not identifying a preparation quality or with multiple preparation qualities cited, procedures in which polyps were removed but not retrieved, procedures in which no polyp size was cited, and patients under 50 for any indication. Eligible procedures were scrutinized for the prep quality, ADR, and AMR. Bowel preparation was rated as adequate (combined excellent or good), fair, or poor. A repeat procedure within 3 years of the original was used to calculate AMR for a total of 359 procedures. Results: The patient population was as follows: 56.6% Female; 54.8% Black, 18.9% Caucasian, 18.6% Hispanic, and 7.7% other. 72.8% of procedures were for screening or surveillance purposes. The ADR for adequate prep was 24.5%, for fair prep was 24.2%, and for poor prep was 23.3% (p-value .89). The ADR for screening procedures was 25.4% and for all other indications was 21.7% (p-value .0071). The AMR for adequate prep was 32.9%, for fair prep was 43.2%, and for poor prep was 40.5% (p-value .18). The AMR for screening procedures was 38.7% and for all other indications was 36.0% (p-value .60). Conclusion: Fair prep is associated with an equal ADR compared to adequate prep. However, the AMR for fair prep is as high as it is for poor prep. These findings may suggest that fair prep, while finding a similar number of adenomas as adequate prep, misses more as well. Thus, fair prep may not be sufficient for detecting all adenomas and time to follow-up should be shortened in these patients. The lack of AMR statistical significance may be due to a limited AMR sample size in this study. A larger sample size study looking specifically at AMR by prep quality may thus reveal a potential benefit to a shorter interval for follow-up in patients with fair prep.
Tu1358 Conventional Bowel Preparation May Not Be Enough for Diabetic Patients: a Cohort Analysis in an Urban Minority Setting Manhal Olaywi*, Shashideep Singhal, Kinesh Changela, Deepanshu Jain, Devin Lane, Mojdeh Momeni, Mahesh Krishnaiah, Sushil Duddempudi, Sury Anand Gastorenterology, The Brooklyn Hospital Center, Brooklyn, NY Introduction: Adequacy of bowel preparation is critical for the quality of colonoscopy. Suboptimal preparation may lead to missed pathology, increase in healthcare cost and procedure related risks. Effect of diabetes on gastrointestinal motility is considered to be multidimensional. The study was designed to evaluate adequacy of bowel preparation and outcome of colonoscopy in diabetic subjects in comparison to controls. Methods: In this cohort study, the subjects were selected from the Colonoscopy Outcomes in Minority Population (COMP) registry at our institution enrolled through colonoscopy patient navigator program. Out of the patients presenting for average risk screening colonoscopy, patients with Diabetes Mellitus (DM) were identified using ICD 9 code and were compared to the matched controls from the study population. In addition to clear liquid diet on the day prior, all patients had received one gallon of polyethylene glycol and 10 mg of dulcolax as bowel preparation the evening prior to procedure. The demographic, clinical, and pathological data of both groups were collected and analyzed using SPSS statistical software. Results: Out of 4015 colonoscopic encounters, 1168 encounters satisfied the inclusion criteria for the study. There were 58.5% African Americans, 31.9% Hispanics, 8% Caucasians and 1.6% others. The Mean age for DM group was 63.5 years (SD 9.3) and 60.6 years (SD 9.15) for the controls. There were no significant differences in the demographics of the study and control group. The bowel preparation was rated as good in 52.6% diabetics vs 64.4% in controls (p⫽0.001). Suboptimal or poor preparation was found in 47.4 % of diabetics compared to only 35.6% of the control (P⫽0.001). Also the completion rates in diabetics with cecal intubation rate of 92.2% were lower in comparison to controls 97.1% (P 0.001). Adenoma Detection Rate (ADR) was 22% in diabetic vs 26.1% in controls, and advanced adenoma detection rates (AaDR) 12.1% in DM2 and 15.8% in controls. There was no statistically significant difference in ADR and AADR between the two groups on accounting for bowel preparation and race. Conclusions: Diabetic patients tend to more often have suboptimal or poor bowel preparation during screening colonoscopy. Our study did not show a statistically significant difference in ADR and AaDR between diabetics and non diabetics. Further studies are needed to devise interventions to improve quality of colonoscopy in diabetic population.
AB512 GASTROINTESTINAL ENDOSCOPY Volume 77, No. 5S : 2013
www.giejournal.org
preparation has a deleterious effect on adenoma detection, physicians may have varying standards for declaring bowel preparation to be suboptimal. We hypothesized that providers who have higher standards for optimal preparation, and hence higher suboptimal preparation rates (SPRs) may be more fastidious in their examination of the colon and therefore have higher ADRs. Methods: Patients ⱖ50 years old who underwent screening colonoscopy at our medical center during 2011 were identified via the electronic endoscopy database. Examinations were excluded if the provider performed fewer than 50 screening colonoscopies during the study period, or if the bowel preparation rating was not recorded. Preparations labeled “fair,” “poor,” or “unsatisfactory” were considered suboptimal. Adenomas were confirmed via cross-reference with the histology database. The ADR was calculated for each provider, and was correlated with the provider’s suboptimal preparation rate (SPR). Multivariate logistic regression was performed to identify independent patient and provider factors associated with ADR, including provider quartile for SPR. Results: Of 2,134 screening colonoscopies performed during the study period, 1,649 examinations from 11 separate gastroenterologists were eligible for inclusion. The majority (59%) of subjects were female, and the mean age was 61 years. Preparation was considered suboptimal in 22% of examinations. Procedure volume ranged from 52 to 353 examinations per provider, with a mean of 150. SPR varied widely among providers, ranging from 3% to 40%. Overall ADR was 22.6%, with a range of 13.4% to 31.6%. SPR was not significantly correlated with ADR (r⫽-0.22, p⫽0.51). Adenoma detection was more common in male patients (26.9% vs 19.6%, p ⫽ 0.0005) and in older patients (37.2% in ⱖ80y vs 19.0% in 50-59y, p⫽0.003), Adenoma detection was lower among patients with suboptimal preparation (19.5% vs 23.5%, p⫽0.10). On multivariate analysis, adenoma detection was associated with male gender (OR 1.55, 95% CI 1.231.97), and older age (OR 2.58 for ⱖ80 vs 50-59, 95% CI 1.35-4.95). After adjusting for age and gender, adenoma detection was associated with the highest quartile of provider procedure volume (OR 1.59, 95% CI 1.05-2.42), but remained independent of provider SPR (p⫽0.28). Sensitivity analysis with exclusion of “fair” exams did not change these associations. Conclusions: SPRs vary widely between providers, but SPR was not correlated with ADR. This suggests that a high SPR is not a marker of higher quality standards and expectations by the provider. Procedure volume was correlated with increased adenoma detection. The impact of physician personality traits on colonoscopy performance requires further study.
or PEG 3350 solution mixed with ascorbic acid (MoviPrep) but does provide patients with an overall better prep experience. Descriptive Summary Table 1. Adjusted Analysis - Based on Logistic Regression (1) Variable Gender Male Female Age Median (IQR) Mean (SD) Range % Prep Taken ⬎80% 60-80% 40-60% Time of Procedure A.M. P.M. Prep Experience Easy Acceptable Difficult Very Difficult Unable to complete
All (nⴝ317)
2L (nⴝ161)
4L (nⴝ156)
164 (51.7) 153 (48.3)
79 (49.1) 82 (50.9)
85 (54.5) 71 (45.5)
57 (50.1, 62.9) 56 (11.3) 20-87
56 (50.2, 62.2) 56 (10.6) 25-80
58 (49.9, 63.8) 57 (12.0) 20-87
308 (97.2) 7 (2.2) 2 (0.6)
158 (98.1) 3 (1.9) 0 (0)
150 (96.2) 4 (2.6) 2 (1.3)
187 (59) 130 (41)
96 (59.6) 65 (40.4)
91 (58.3) 65 (41.7)
110 (34.7) 151 (47.6) 43 (13.6) 12 (3.8) 1 (0.3)
69 (42.9) 76 (47.2) 15 (9.3) 1 (0.6) 0 (0.0)
41 (26.3) 75 (48.1) 28 (17.9) 11 (7.1) 1 (0.6)
Table 2. (1) Adjusted for age and gender. Also adjusted for endoscopist and timing of procedure for the analysis of BBPS (2) Effect of 2L relative to 4L Peglyte Outcome BBPS⬎8 BBPS⬎7 Had Adverse Event Would be Willing to Repeat Found Prep Easy or Acceptable Not Able to Complete Prep (⬍80%)
Tu1354 Efficacy and Patient Tolerability of Split Dose 2L vs. Split Dose 4L Peglyte: a Prospective Randomized Controlled Trial Mayur Brahmania*, George Ou, Andrew M. Round, Jessica Tong, Ricky W. Kwok, Cherry Galorport, Hin Hin Ko, Eric C. LAM, Brian Bressler, Jennifer J. Telford, Robert A. Enns Gastroenterology, St Paul’s Hospital, Vancouver, BC, Canada Introduction: Despite the evolution of bowel preparation for endoscopic investigation it remains one of the most frequently cited deterrents for patient acceptance and cooperation prior to colonoscopy. In an effort to advance the quality of preparations used and patient compliance we compared split dosing of 2L Peglyte with 15mg of bisacodyl vs split dosing of 4L Peglyte. Our primary endpoint was to assess the efficacy of prep and secondarily patient tolerability. Patients and Methods: This study was a randomized, controlled, single centre, endoscopist-blinded, non-inferiority trial involving five endoscopists. All adults (ⱖ19 years of age) who were referred for outpatient colonoscopy were considered for our study. Exclusion criteria included those with constipation (⬍3 BM/week with or without laxative use), suspected small bowel obstruction, severe inflammatory bowel disease with known stricturing or penetrating disease, and history of colonic resection. Eligible patients were randomized to one of two bowel preparations through the use of concealed allocation by a scheduling assistant (blinded) in a one-to-one ratio. Bowel preparations included: (1) 4L of PegLyte (4L PEG 3350 electrolyte solution) or (2) 2L of Peglyte (2L PEG 3350 electrolyte solution and 15mg bisacodyl) as currently manufactured and marketed. Both preparations were taken as a split dose. The primary endpoint was bowel cleansing efficacy as rated by a Boston Bowel Prep Score (BBPS) score of greater than 8 (scores of 7 or above are generally felt to provide a good visualization of the colon). Sample size study was based on an assumption of alpha⫽0.05 and beta⫽0.20. The estimated sample size was 158 patients in each arm. Our secondary outcome was patient tolerability to 2L PegLyte based on a six-question survey. Results: Our study finds using 2L PegLyte in a split dosing regimen is as efficacious as using a 4L split dosing regimen using the Boston Bowel Prep Score. Additionally, 2L PegLyte in a split dose showed statistically improved patient tolerability and willingness to repeat colonoscopic procedures. See Tables 1 & 2 Conclusion: The current standard of gut lavage (4L PegLyte in a split dose) is often a large volume for patients to consume, often leading to non compliance and poor bowel preparation. Our non inferiority study demonstrates endoscopists can safely use 2L PegLyte in a split dosing fashion to achieve similar results using 4L of PegLyte in a split dose. In addition, the low volume PEG was statistically better tolerated than the conventional 4L volume. It is unclear how 2L PEG split dose compares to other low volume agents available such as split dose sodium picosulfate (Pico- Salax)
www.giejournal.org
Odds Ratio (2)
95% CI
p-Value
0.81 0.72 1.30 34.72 3.58
(0.51, 1.29) (0.42, 1.24) (0.82, 2.04) (8.12, 148.43) (1.87, 6.85)
0.369 0.231 0.265 ⬍0.001 ⬍0.001
0.40
(0.10, 1.68)
0.212
Tu1355 The Boston Classification of Excrement and Residue (B-Clear): a Valid and Reliable Descriptive Scheme for Bowel Preparation Research Pushpak Taunk*, Jonah N. Rubin, Loc Ton, Helen Lee, Audrey H. Calderwood, Brian C. Jacobson Boston University Medical Center, Boston, MA Purpose: Existing bowel preparation rating scales are useful for clinical purposes but may not be sufficiently descriptive of the residual fecal material and types of liquids encountered during colonoscopy for all types of bowel preparation research. We sought to validate a novel bowel preparation classification scheme that describes both solids and liquids encountered during colonoscopy. Methods: The B-CLEAR assigns a categorical variable based on the type of solid material (NS - no solid, MS - movable solid, IS - immovable solid) and liquid material (NL - no liquid, L - liquid, F- film, G- gelatinous) encountered during colonoscopy. These categories are then combined to describe what was found in each segment of the colon (right, transverse, left). For example, R:NS/NL would be a perfectly clean right colon while T:IS/G would be a very poor preparation in the transverse colon. When more than one type of solid or liquid is found in the same segment, the “worse” of the two types is listed. Washing/suctioning is permitted, but this won’t improve the rating; it is used to help assign a classification (e.g. can a solid be moved to permit mucosal visualization?). After reviewing a set of training images to learn the new classification scheme, 14 gastroenterologists classified a testing set of 16 different images representing the full spectrum of categories. Inter-rater reliability was measured using intraclass correlation coefficients (ICCs). B-CLEAR categories were then assigned prospectively during outpatient colonoscopies and validated by correlation with volume of water flushes used, volume of effluent suctioned and colonoscope withdrawal times. We also compared B-CLEAR assignments with segmental Boston Bowel Preparation Scale (BBPS) scores. Results: The ICCs for solid, liquid and combined solid⫹liquid B-CLEAR classifications were 0.78, 0.89 and 0.79,
Volume 77, No. 5S : 2013
GASTROINTESTINAL ENDOSCOPY
AB511
Abstracts
respectively, indicating good inter-observer reliability. In the validation portion of the study, the worst B-CLEAR score in each case correlated with washing and suctioning volumes (Pearson’s correlation coefficient 0.56, p⫽0.003 for both) and was non-significantly correlated with colonoscope withdrawal time (0.37, p⫽0.06). B-CLEAR assignments and BPPS scores were not correlated, suggesting each scale provides unique information about the preparation. Conclusion: The B-CLEAR classification scheme is a valid and reliable tool for describing residual fecal material encountered during colonoscopy. With this new scheme, we can determine how various types of residual material affect an endoscopist’s ability to visualize the colon. Because the B-CLEAR captures specific information about what is encountered in each colon segment, it may prove a useful tool for studying the impact of novel washing devices or comparing dietary instructions prior to colonoscopy.
Tu1356 Adenoma Detection in Excellent Versus Good Bowel Preparation for Colonoscopy Danielle M. Tholey*1, Archana Anantharaman1, Corbett E. Shelton1, Robert Frankel1, Paurush Shah1, Amy Coan2, Sarah E. Hegarty3, Benjamin E. Leiby3, David M. Kastenberg4 1 Medicine, TJUH, Philadelphia, PA; 2Medicine, Thomas Jefferson Medical College, Philadelphia, PA; 3Biostatistics, TJUH, Philadelphia, PA; 4Gastroenterology, TJUH, Philadelphia, PA Adequate bowel preparation (prep) permits detection of ⬎5 mm polyps anywhere in the colon. Our definition of adequate also requires endoscopist adherence to published surveillance intervals not be shortened in response to prep quality. Within the category of adequate, a good prep requires excessive flushing and/or suctioning whereas excellent does not. Aim: To evaluate whether the adenoma detection rate (ADR) is superior with an excellent vs. good prep. Methods: Retrospective review of consecutive colonoscopies with good or excellent prep from 9/1/11-9/1/12 at a single academic center. Endoworks™ (Olympus), EMR, and pathology data bases utilized for prep quality, indication, demographics, medical history, adenoma, and cancer data. Primary outcome was ADR (# pt’s with ⬎ 1 adenoma); secondary were adenomas/colonoscopy (# adenomas/# colonoscopies), adenoma distribution, advanced and sessile serrated adenomas (SSA), and cancer (CRC). Exclusions ⫽ age ⬍18, IBD, or Familial Polyposis. ADR was compared for excellent and good preparations using logistic regression. Secondary endpoints were compared using Poisson regression. Outcome comparisons adjusted for confounding variables in demography and history. Assuming an ADR⫽30%, with 70% good and 30% excellent preps, ⬃ 4,400 colonoscopies were necessary to find a 1.15 RR between groups with 82% power using a 2-sided alpha⫽.05. Results: 5,113 colonoscopies with adequate prep were reviewed (good ⫽ 3112, excellent ⫽ 2001). The excellent group was younger and more often had a FHx of CRC and polyps, and indication of screening; and, less frequently, a personal history of CRC and adenoma, and surveillance indication (all, P⬍0.0001). Neither ADR nor adenoma/colonoscopy was superior for the excellent group (respectively, OR 0.97, CI 0.85-1.11, p⫽0.618; IRR 0.98, CI 0.90-1.07, p⫽0.705). These findings were consistent for the proximal and distal colon (respectively, p⫽0.426, p⫽0.764). ADR and adenomas/ colonoscopy for good and excellent were, respectively, 29% vs 26% and 0.499 vs 0.437. The excellent group had superior detection of SSA’s (IRR 1.66, CI 1.142.40, p⫽0.008) and advanced adenomas (IRR 1.37, CI 1.09-1.72, p⫽0.007), but was not superior in detecting CRC (IRR .268, CI .08-.91, p⫽0.035). ADR and adenoma/colonoscopy increased consistently with age (p⬍0.001), and was greater in men (p⬍0.001) and those with FHx of polyps (respectively, p⫽0.002; p⫽0.001), personal history of adenoma (respectively, p⬍0.001; p⬍0.001), and a surveillance indication (respectively, p⬍0.003; p⫽0.007). Adenoma/colonoscopy increased with a personal history of CRC (p⫽0.003). Conclusion: In pts with adequate colon preparation, excellent cleansing as compared to good does not improve ADR or adenoma/colonoscopy. However, excellent cleansing is associated with superior detection of high risk lesions including advanced adenomas and SSA’s.
Tu1357 Colonoscopy Preparation Quality - Is Fair Good Enough? ADAM B. Gluskin*1, Nikhil Martis2, Weihua Gao3, Vijay S. Khiani2 1 College of Medicine, University of Illinois at Chicago, Chicago, IL; 2 Medicine, University of Illinois at Chicago, Chicago, IL; 3Center for Clinical and Translational Science, University of Illinois at Chicago, Chicago, IL Introduction: Many factors influence colonoscopy quality, one of which is bowel preparation. Time to next colonoscopy is frequently shortened when prep is rated as poor. However, there is little data regarding the proper screening interval for individuals with fair prep quality. In addition to prep quality, two other important indicators, Adenoma Detection Rate (ADR) and Adenoma Miss Rate (AMR), have emerged as tools to assess colonoscopy quality. ADR is the percentage of colonoscopies that have a detected adenoma and AMR is the
percent of adenomas that are detected in a repeat colonoscopy divided by the total number of adenomas found in both the original and repeat procedures. This study examined how ADR and AMR vary across preparation quality. Methods: A retrospective analysis was performed that examined 4909 colonoscopies completed at the University of Illinois Hospital between 2007 and 2011. The exclusion criteria were a history of colorectal cancer, a history of Inflammatory Bowel Disease, patients with Familial Adenomatous Polyposis, a history of colon resection, procedures failing to reach the cecum, procedures not identifying a preparation quality or with multiple preparation qualities cited, procedures in which polyps were removed but not retrieved, procedures in which no polyp size was cited, and patients under 50 for any indication. Eligible procedures were scrutinized for the prep quality, ADR, and AMR. Bowel preparation was rated as adequate (combined excellent or good), fair, or poor. A repeat procedure within 3 years of the original was used to calculate AMR for a total of 359 procedures. Results: The patient population was as follows: 56.6% Female; 54.8% Black, 18.9% Caucasian, 18.6% Hispanic, and 7.7% other. 72.8% of procedures were for screening or surveillance purposes. The ADR for adequate prep was 24.5%, for fair prep was 24.2%, and for poor prep was 23.3% (p-value .89). The ADR for screening procedures was 25.4% and for all other indications was 21.7% (p-value .0071). The AMR for adequate prep was 32.9%, for fair prep was 43.2%, and for poor prep was 40.5% (p-value .18). The AMR for screening procedures was 38.7% and for all other indications was 36.0% (p-value .60). Conclusion: Fair prep is associated with an equal ADR compared to adequate prep. However, the AMR for fair prep is as high as it is for poor prep. These findings may suggest that fair prep, while finding a similar number of adenomas as adequate prep, misses more as well. Thus, fair prep may not be sufficient for detecting all adenomas and time to follow-up should be shortened in these patients. The lack of AMR statistical significance may be due to a limited AMR sample size in this study. A larger sample size study looking specifically at AMR by prep quality may thus reveal a potential benefit to a shorter interval for follow-up in patients with fair prep.
Tu1358 Conventional Bowel Preparation May Not Be Enough for Diabetic Patients: a Cohort Analysis in an Urban Minority Setting Manhal Olaywi*, Shashideep Singhal, Kinesh Changela, Deepanshu Jain, Devin Lane, Mojdeh Momeni, Mahesh Krishnaiah, Sushil Duddempudi, Sury Anand Gastorenterology, The Brooklyn Hospital Center, Brooklyn, NY Introduction: Adequacy of bowel preparation is critical for the quality of colonoscopy. Suboptimal preparation may lead to missed pathology, increase in healthcare cost and procedure related risks. Effect of diabetes on gastrointestinal motility is considered to be multidimensional. The study was designed to evaluate adequacy of bowel preparation and outcome of colonoscopy in diabetic subjects in comparison to controls. Methods: In this cohort study, the subjects were selected from the Colonoscopy Outcomes in Minority Population (COMP) registry at our institution enrolled through colonoscopy patient navigator program. Out of the patients presenting for average risk screening colonoscopy, patients with Diabetes Mellitus (DM) were identified using ICD 9 code and were compared to the matched controls from the study population. In addition to clear liquid diet on the day prior, all patients had received one gallon of polyethylene glycol and 10 mg of dulcolax as bowel preparation the evening prior to procedure. The demographic, clinical, and pathological data of both groups were collected and analyzed using SPSS statistical software. Results: Out of 4015 colonoscopic encounters, 1168 encounters satisfied the inclusion criteria for the study. There were 58.5% African Americans, 31.9% Hispanics, 8% Caucasians and 1.6% others. The Mean age for DM group was 63.5 years (SD 9.3) and 60.6 years (SD 9.15) for the controls. There were no significant differences in the demographics of the study and control group. The bowel preparation was rated as good in 52.6% diabetics vs 64.4% in controls (p⫽0.001). Suboptimal or poor preparation was found in 47.4 % of diabetics compared to only 35.6% of the control (P⫽0.001). Also the completion rates in diabetics with cecal intubation rate of 92.2% were lower in comparison to controls 97.1% (P 0.001). Adenoma Detection Rate (ADR) was 22% in diabetic vs 26.1% in controls, and advanced adenoma detection rates (AaDR) 12.1% in DM2 and 15.8% in controls. There was no statistically significant difference in ADR and AADR between the two groups on accounting for bowel preparation and race. Conclusions: Diabetic patients tend to more often have suboptimal or poor bowel preparation during screening colonoscopy. Our study did not show a statistically significant difference in ADR and AaDR between diabetics and non diabetics. Further studies are needed to devise interventions to improve quality of colonoscopy in diabetic population.
AB512 GASTROINTESTINAL ENDOSCOPY Volume 77, No. 5S : 2013
www.giejournal.org