Abstracts

39 experienced (O100) operators of variable experience were asked to perform264 episodes of colonoscopy on the HT Immersion simulator at random. All theprocedure recorded the total amounts of air a) insufflated and left behind b) Painclassified by the simulator as absent, mild, moderate, severe and extreme. The painwas weighed into three categories as Grades 1, 11 and III according to increasingseverity. A Pearsons Correlation coefficient was calculated for demonstrating anyassociative relationship between; a) amounts of air insufflated and pain, b) amountof air left behind and pain. Data was also evaluated according to the level ofexperience of the operator. Results: Two tailed correlation coefficients of - 0.078and -0.084 were demonstrated between the amounts of air insufflated, / air leftbehind respectively and Pain. When data was filtered according to experience of theoperator, the correlation coefficients were novice (-0.119, -0.113), intermediate(0.004, -0.059) and expert (-0.076, -0.103) respectively. This indicates that pain is notassociated with amount of air insufflated or left behind during simulatedcolonoscopy and that this holds true even when corrected for the experience of theoperators. Discussion: During a simulated colonoscopy, if pain is not affected by airin the colon in any major way, then what represents pain? Possible contenders areexcessive looping, exertion of surplus force on the colonoscopy shaft, bowelhypersensitivity and even machine error. What exactly is meant by pain insimulation and the nature of the concept needs to be fully elicited. In context ofsimulated Colonoscopy, we need to clarify what proportions of each of the abovelisted contenders contribute in marking pain. Further studies must validate thisaspect of the colonoscopy simulator because it is now a recognised teaching andassessment tool.

W1508

Nurse Endoscopists May Take OverPaul G. Van Putten, Monique E. Van Leerdam, E.J. KuipersBackground: Colorectal cancer screening and surveillance is still not feasible in themajority of Europe. Nurse endoscopists (NE) may provide a solution for the highendoscopic demand and shortage of endoscopists. The aim of the present studywas to determine expectations, opinions and future perspectives of nurseendoscopy in a Western European community. Methods: A postal questionnaire wassend to all registered gastroenterologists (GE) (n Z 301) and gastroenterologyresidents (n Z 79) in the Netherlands in 2007. Topics that were evaluated included:presumed endoscopic quality, patient experiences and costs. In addition theattitude towards and potential endoscopic procedures to be performed by NE wereevaluated. Results: 165 of 380 (43%) GE and residents completed the questionnaire.The distribution of academic versus general hospital employees was 30 vs 70% witha 79 to 21% M/F distribution. Overall, 52% had a positive attitude towardsintroduction of NE, whereas 18% were neutral and 30% negative. No difference inendoscopic quality between physician and NE was expected by 42% of respondents,35% expected physicians to perform better and 23% expected nurses to performbetter. With regard to patient experiences: 66% of respondents expected physiciansto perform better than NE, 18% expected nurses and physicians to perform equallyand 16% expected NE to perform better. The majority expected a waiting list andcost reduction of NE. Screening sigmoidoscopy and colonoscopy were consideredappropriate procedures to be performed by nurses according to respectively 92%and 73% of respondents. In contrast, only half of the respondents judged thatdiagnostic endoscopies would be appropriate for NE. Only 45% judgedpolypectomies of polyps smaller than 10 mm appropriate for NE. Respondentsagreed that polypectomies of larger polyps and other therapeutic proceduresshould not be performed by nurses (respectively 75% and O90%). GE andresidents who expected better or comparable endoscopic quality of NE incomparison with physicians, were significantly more positive towards theintroduction of NE (70 and 55% vs 38%; p Z 0.004). Those who expected NE toperform better than physicians with regard to patient experiences, were alsosignificantly more positive (77% vs 48%; p Z 0.007). Respondents age, gender ortype of hospital was not related with the attitude towards NE. Conclusion: Themajority of GE is positive towards the introduction of NE. Screening sigmoidoscopyand colonoscopy may well be performed by NE. To determine the exact place ofnurse endoscopy, precise assessment of endoscopic quality and patient experiencesof well trained NE is highly needed.

W1509

Prospective Results of Endoscopic Ultrasound Drainage of

Pancreatic Pseudocysts in a Centre Starting with This ProcedureRiadh Sadik, Claes JonssonOnly a few prospective studies exist on the results of endoscopic pseudocystdrainage. The largest prospective study from a large centre showed a complicationrate of 19% (Kahaleh et al, Endoscopy. 2006 Apr;38:355-9.). However, reports fromcentres starting with the procedure are lacking. The aim of this study was toprospectively analyze the results and complication rate of pseudocyst drainage ata centre starting with the procedure. Methods: All patients referred for endoscopicpseudocyst drainage were prospectively included. The procedures were performedby one endoscopist with long endoscopy experience. The results and complicationrate were recorded. Results: In a period of 20 months 10 patients were included.One patient had high Carcino Emberionic Anigen level and was found to havepancreatic cancer. The results are shown in the table. Inspection of the cyst witha gastroscope and flushing with saline and/or necrosectomy was performed in four

AB302 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 5 : 2008

patients. Conclusion: Complications of pseudocyst drainage are associated with cystinfection. The results in centres starting with endoscopic pseudocyst drainage arecomparable to results from larger centres.

Patient(Age) Cyst size

Infectionin the Number

Hospitalstay Final

Gender E

thiology in cm cyst o f stents Complications

ww

(days)

w.giejo

result

1 (52) M I

diopathic 13 � 10 No 1 (10F) No 1 Total regress 2 (66) M M alignancy 6 � 7 No 1 (10F) No 1 Total regress 3 (61) M A lcohol 15 � 13 No 3 (7F) No 2 Total regress 4 (76) F G allstone 7 � 7 Yes 3 (7F) No 1 Total regress 5 (76) M A lcohol 5 � 5 No 1 (10F) No 1 Total regress 6 (47) M A lcohol 7 � 5 No 3 (7F) No 1 Total regress 7 (61) M A lcohol 10 � 10 No 2 (10F) No 1 Total regress 8 (57) M A lcohol 5 � 5 Yes 0 Bleeding 6 Operation 9 (68) F G allstone 18 � 13 Yes 5 (7F)þone

naso-cystic

No 4 Total regress

10 (51)M

A

lcohol 8 � 3 Yes 2 (7F) Perforation 5 Initial regress,perforationduringa secondprocedure

W1510

Construct Validity of the Simbionix GI Mentor II Endoscopy

Simulator for GI Fellow Trainees in ColonoscopyJenifer R. Lightdale, Meghan E. Fredette, Paul A. Rufo, Victor L. Fox,John R. Saltzman, John M. PonerosThe Simbionix GI Mentor II Endoscopy Simulator is designed to provide detailedfeedback, including time to anatomic landmarks, intraprocedural patientdiscomfort, and visualization of the mucosa. While construct validation studies haveshown that the GI Mentor II can distinguish beginners from experts, its capacity todiscriminate between fellows in a 3-year training program is unknown. Aim: Todetermine whether feedback measures provided by the GI Mentor II differ between1st and 3rd year fellows. Methods: Fellows in their 1st and 3rd years of training atChildren’s Hospital Boston and the Brigham and Women’s Hospital were invited torecord colonoscopies (6/06 - 11/07) on a GI Mentor II. Fellows were instructed toreach the cecum as quickly as possible, while visualizing the entire mucosa andavoiding colonoscopic looping. Patient discomfort was defined as the # of and %procedure time excessive loops were formed causing local pressure. Othersimulator parameters included % time spent with clear view of the lumen, and # oftimes view of the lumen was lost. Results: 16 GI Trainees (10 1st Yrs, 6 3rd Yrs; 9male; median age 32 yrs (IQR 29, 35)) recorded a total of 63 simulatedcolonoscopies. No differences were found between groups in # of pediatric fellows(1st Yr: 80%; 3rd Yr: 50%, p Z .225), nor median # of simulations (1st Yr: 3(2, 7);3rd Yr: 3 (3, 5). p Z 0.865). 1st year fellows did not spend more total time persimulated procedure than 3rd year fellows (13 min (9, 19) vs. 9 (8, 15), p Z 0.084),but did take longer to reach the cecum (7 min (4, 11) vs. 4 (2, 6), p Z 0.002). Fewfellows formed excessive loops (1st Yrs: 0 (0, 1); 3rd Yrs: 0 (0, 1), p Z 0.383), andthere was no difference in ‘‘virtual’’ patient discomfort (1st Yr: 1% procedural timein pain (0, 5); 3rd Yr: 0% (0, 2), p Z .108). While 1st and 3rd year fellows did notdiffer in terms of their propensity to lose view of the lumen (1 time (0, 2) vs. 0 (0,1),p Z 0.102), 1st Years spent less time visualizing the entire mucosa (84% (77, 88) vs.3rd Yr: 89% (81, 91), p Z .011) and were less efficient in screening as calculated bythe simulator (64% (39, 81) vs. 84% (64, 89), p Z .006). Conclusions: Testparameters concerning time to anatomic landmarks and visualization of the mucosagenerated by the GI Mentor II Endoscopy Simulator can discriminate between 1stand 3rd year GI fellows. Although the simulator did not detect differences in patientcomfort, our findings suggest progression in endoscopy skills over a 3 yearfellowship training period. Further study is needed to understand howperformance enhancement on simulator measures translates into improvedendoscopy in live patients.

W1511

Comparative Study of Conventional Colonoscopy and

Pan-Colonic Narrow-Band Imaging System in the Detection of

Neoplastic Colonic Polyps: A Randomized, Controlled TrialTakuya Inoue, Mitsuyuki Murano, Naoko Murano, Takanori Kuramoto,Ken Kawakami, Yosuke Abe, Eijiro Morita, Ken Toshina, Yutaro Egashira,Eiji Umegaki, Kazuhide HiguchiBackground: Detection and removal of adenomas by colonoscopy is an importantmeans for preventing cancer; however, small adenomas may be missed duringcolonoscopy. The narrow-band imaging (NBI) system clearly enhances themicrovasculature in neoplastic lesions, making it appear as a dark complex.Therefore, the NBI system may improve the detection of colonic neoplasias. Theaim of this randomized prospective study was to determine whether the NBIsystem could increase the adenoma detection rate in routine use. Aim: Weconducted a randomized, controlled trial to determine the efficacy of the

urnal.org

Abstracts

pan-colonic NBI system in adenoma detection. Methods: Two hundred forty-threepatients were randomized, 121 to conventional colonoscopy and 122 to pan-colonicNBI system. Demographics, indication for colonoscopy, and quality of preparationwere similar between groups. Results: Extubation time was not significantlydifferent between the conventional colonoscopy and pan-colonic NBI system. Theproportions of patients with at least one adenoma and those with multipleadenomas were not significantly different between groups. However, the pan-colonic NBI system significantly increased the total number of adenomas detected(p ! 0.05) and the number of diminutive (!5 mm) adenomas detected (p !0.05). The pan-colonic NBI system allowed detection of more diminutive adenomasin the distal colon than did conventional colonoscopy (p ! 0.01), and morepatients in the NBI group had at least one diminutive adenoma than in the controlgroup (p ! 0.05). Conclusions: The pan-colonic NBI system improves the totalnumber of adenomas detected, including significantly more diminutive adenomas,without prolongation of extubation time. These results indicate that routine use ofthe NBI system for surveillance of diminutive adenomas may be recommended.

Polyp findings for the whole colon and distal/proximal to the sigmoid-descending junction

Distal colon Proximal colon Entire colon

www.giejo

urnal.org

Cn

ontrolZ 121

Nn

BIZ 122 p

Controln Z 121

NBIn Z 122

p

Cn

ontrolZ 121

Nn

BIZ 122 p

Patients with atleast one polyp,n(%)

2

1 (17%) 3 2 (26%) 0 .094) 34 (28%) 34 (28%) 0.97) 4 7 (39%) 6 2 (51%) 0 .061)

Total No. ofpolyps

2

6 4 8 0 .047y 52 79 0.078y 7 8 1 27 0 .014y

Patients withat least oneadenoma, n(%)

1

7 (14%) 2 7 (22%) 0 .10) 30 (25%) 42 (34%) 0.10) 4 1 (34%) 5 1 (42%) 0 .20)

Total No. ofadenomas

1

9 3 7 0 .069y 46 65 0.15y 6 5 1 02 0 .046y

Patients withat least oneadenoma!5 mm, n(%)

2

(2%) 1 3 (11%) 0 .0035) 18 (15%) 29 (24%) 0.079) 2 0 (17%) 3 7 (30%) 0 .011)

Total No. ofadenomas!5 mm

2

1 4 0 .0034y 24 39 0.12y 2 6 5 3 0 .020y

)Chi-squared testyStudent’s t-test

W1512

A Randomized Controlled Trial of the Usefulness of Cap-Fitted

Colonoscopy On the Polyp DetectionNoriyuki Takano, Yutaka Yamaji, Haruka Kajiwara, Takafumi Sugimoto,Atsuo Yamada, Masao Akanuma, Goichi Togo, Keiji Ogura,Makoto Okamoto, Haruhiko Yoshida, Takao Kawabe, Masao OmataBackground: In our previous trial, a transparent hood attached to the tip of thecolonoscope was shown to be helpful in shortening the cecal intubation time (Am JGastroenterol 2007;102:75). That trial also suggested the hood improved thedetectability of colorectal polyps including large polyps over 10 mm. Objectives: Toevaluate the effect of using a transparent hood on the polyp detectability. Methods:Patients for colonoscopy from August 2005 to July 2007 in Tokyo University Hospitalwere enrolled to this prospective, randomized controlled trial, and assigned tocolonoscopy with or without transparent hood. Colonoscopies were performed bytrainees without sedation. Exclusion criteria were prior colon resection, knownpolyps or fulminant colitis. Main outcomes were detection rates of all polyps andlarge polyps of R10 mm. Number of detected polyps, size and location stratifieddetection rates were secondary outcomes. Results: A total of 2502 patients wererandomly assigned to transparent hood (n Z 1287), and no hood (n Z 1215)groups. The cecal intubation rate was slightly higher in the transparent hood groupthan in the no hood group (94.7% vs 92.6%, P Z 0.03). The polyp detectability wasevaluated for the cases of complete total colonosopies. The detection rate of allpolyps was higher in the hood group than in the no hood group (47.0% vs 42.6%, PZ 0.03), but the detection rates of large polyps of R10 mm showed no significantdifference (8.2% vs 7.2%, P Z 0.329). The mean number of all polyps was higher inthe hood group than in the no hood group (0.951 vs 0.819 polyps per patient, P Z0.01), while the number of large polyps was not (0.095 vs 0.092 polyps per patient,P Z 0.378). Detection rates of polyps of 6-9 mm were 12.6% and 10.7% (p Z0.155), and those for small polyps of %5 mm were 39.8% and 35.2% (p ! 0.05), inthe hood and no hood groups, respectively. While the polyp detection rates weregenerally higher in the hood group in all locations, the difference was significantonly in the sigmoid colon (21.4% and 17.1%, p Z 0.01). Conclusion: Thetransparent hood in colonoscopy improves polyp detection. The effect wasprominent for small polyps in the sigmoid colon.

W1513

Comparison of the Performance Characteristics of Endoscopic

Submucosal Dissection with Laparoscopic Assisted Colectomy

for En-Bloc Resection of Colorectal NeoplasmShinsuke Kiriyama, Yutaka Saito, Takahisa Matsuda, Takeshi Nakajima,Tsuyoshi Kikuchi, Seiichiro Yamamoto

Vo

Background: Laparoscopic assisted colectomy (LAC) has become an acceptedminimally-invasive alternative to open surgery for the resection of colorectalneoplasm. Endoscopic submucosal dissection (ESD) has recently become anadditional alternative, but is reportedly associated with a steep learning curve andsignificant risk of complications. Aim: The indication for ESD in our hospital isintramucosal or sm1 lesion, and that for LAC is sm2 or deeper cancer. However, thestandard treatment for large colorectal adenoma or intramucosal lesions in manyhospitals worldwide is LAC. We therefore compared ESD in the treatment ofcolorectal neoplasm with LAC used to resect colorectal carcinoma. Methods: Thesubjects had early stage colorectal tumors of preoperative T1 or less (LAC), orcolorectal adenoma/ Tis (intramucosal) and sm1 (less than 1000mm) carcinoma(ESD) from February 1998 to October 2007. We assessed clinical results andcomplications in both groups. Result: There were 300 cases in the ESD group, and272 cases in LAC. There was no difference between the 2 groups in terms of age andgender. The cases, stratified by depth of invasion (if carcinoma), were as follows: m/sm1/sm2-/mp/ss were 152/35/36/2/0 in ESD, and 43/20/197/10/1 in LAC groups. Inthe ESD cases, the mean tumor size was 36 mm (7-140). The en-bloc resection rateand curative resection rates were 86% and 83% respectively. The average time ofprocedure was 109 mins (15-480). There were 12 cases complicated by perforation(4%) and 4 with post-procedure bleeding (1.3%) among the ESD patients. Allcomplications were managed endoscopically except for one patient withperforation. In the LAC group, the mean tumor size was 19 mm (3-80). The averagetime of operation was 210 mins (95-500). There were 31 wound infections, 2 pelvicabscesses, 3 anastomotic leaks and 1 patient with anastomotic bleeding. Stomaswere necessary in 75% cases in LAC of rectal lesions below the peritoneal reflection(11 temporary ileostomies and 1 permanent colostomy). No blood transfusionswere required for either ESD or LAC cases. Conclusion: In this series, ESD wasassociated with few complications and a high rate of en-bloc resection when usedto treat superficial colorectal neoplastic lesions including a significant number ofearly carcinomas. Its safety profile and en-bloc resection rate were favourable whencompared with LAC. ESD represents a promising minimally-invasive therapy forsuperficial colorectal carcinoma; randomized studies and outcome studies areneeded.

W1514

Examination of Current Colorectal Cancer Screening Practices:

Are We Adherent to American Gastroenterological Association

(AGA) Guidelines?Paul Crainich, Eric K. Ganguly, Richard ZubarikBackground: There is limited data on how adherent primary care providers (PCP)are to AGA cancer screening guidelines. Aims: 1) To determine if patients in thecommunity are being referred for colorectal cancer screening using colonoscopyaccording to AGA guidelines. 2) To assess if any barriers to screening exist. 3) Toidentify differences in referral practices between and within primary care providergroups. Methods: This prospective cohort study queried consecutive patientspresenting for colonoscopic colorectal cancer screening using a questionnaire andchart review when necessary. Patients were eligible if they were at least 18 years ofage and if this was their first colorectal cancer screening referral for colonoscopy.Overall, 1006 patients were mailed questionnaires along with their screeningcolonoscopy information packets. Of these, 641 (64%) consenting patientscompleted questionnaires, and 621 of these met the aforementioned eligibilityrequirements and were used for data analysis. Referrals for screening wereconsidered according to AGA guidelines if they occurred within three years of therecommended age. Results: Of the 621 eligible patients, 48% were male and theaverage age was 56 years, range 30-83 years. Data for adherence to AGA guidelinesare illustrated in the table below. The most common reasons cited by patientspresenting outside the recommended screening period (i.e age R53) were (1) The‘‘PCP did not recommend it before now’’ (34%), (2) The patient ‘‘did not feel testwas necessary before now’’ (21%) and (3) The patient was ‘‘concerned that the testmight be painful or uncomfortable’’ (16%). The University PCP’s were significantlymore adherent to guidelines than non-University PCP’s (p Z 0.002). This remainedsignificant (p Z 0.007) even if patients that had any other prior screening testswere excluded. There was no significant difference between individual referringPCP groups (p Z 0.576) in adherence to guidelines. Conclusions: Approximatelyhalf of patients are referred outside of AGA guidelines for colonoscopic colorectalcancer screening. The most commonly cited reasons for this are that the PCP doesnot recommend it, the patient does not feel it is necessary and the patient isconcerned the procedure may be painful.

Number (%) of patients referred for screening colonoscopy as stratified by age.

# of patients R53 R55 R60

lume 67, No. 5 : 2008 GA

STROINTEST INAL END OSCOPY

All eligible patients

621 359 (58%) 288 (46%) 169 (27%) Patients without prior screening) 391 178 (46%) 126 (32%) 61 (16%)

)Defined as any combination of prior sigmoidoscopy, fecal occult blood test orbarium enema performed at any time.

AB303