Abstracts

T1521

Outcomes Following Gastrostomy: Radiologically Inserted Vs.

Percutaneous Endoscopic GastrostomyJohn S. Leeds, Mark E. Mcalindon, Julia Grant, Helen E. Robson,Stephen Morley, Fred K. Lee, David S. SandersIntroduction: Gastrostomy insertion has been demonstrated to be of benefit inselected patients. Percutaneous endoscopic gastrostomy (PEG) using the pull throughtechnique is the most widely used insertion method, but it is recognised to havesignificant complications particularly in patients with respiratory risk factors. Analternative is a radiologically inserted gastrostomy (RIG). It has been suggested that RIGmay be advantageous in patients who are potentially ‘high risk’ from respiratorycomplications. However there are no large studies comparing PEG against RIG.Methods All patients referred for a gastrostomy are prospectively included in a databasealong with demographic, biochemical and outcome data. Analysis of gastrostomyinsertions over the period February 2004 to February 2007 was performed withreference to method of insertion and outcome at 30 days. Selection for method ofinsertion is left to the discretion of the referring clinician. Patients were sub groupedinto cognitive impairment (n Z 5), dysphagic stroke (n Z 36), nasopharygeal cancer (nZ 175), neurological (n Z 116) and other (n Z 71). Results Over the study period 170RIG’s and 233 PEG’s were inserted (mean age 62, 268 males). There were no differencesin age between the RIG and PEG group and case mix was comparable except in thenasopharyngeal cancer group (proportionally more RIG’s). The RIG 30 day mortalitywas 26/170 (15.3%) and the PEG 30-day mortality was 25/233 (10.7%) (p Z 0.17). Oneyear mortality was 92/170 (54.1%) for RIG and 131/233 (56.7%) for PEG (p Z 0.60).Within sub groups the only significant difference in 30 day mortality was in thosepatients with nasopharyngeal cancer; 14/106 (13.2%) for RIG and 1/69 (1.4%) for PEG(p Z 0.005). However, patients referred for RIG were significantly older than thosereferred for PEG (mean age 59.7 vs. 64, p Z 0.019) and had a higher prevalence ofsignificant comorbidities (21.1% in the PEG group and 37.7% in the RIG group).Conclusions Overall RIG and PEG appear to have similar 30 day mortality rates. Inpatients with nasopharyngeal cancer there was a higher mortality in those referred forRIG however pre-selection by the referring clinician due to perceived risk ofendoscopic insertion may have biased the outcome. A randomized trial comparingboth methods in this sub group is needed.

T1522

Predicting Outcomes Following Gastrostomy Insertion Using

the Sheffield Gastrostomy Score. A Prospectively Devised

Scoring System with a Validation CohortJohn S. Leeds, Stephen Morley, Roger Marr, Helen E. Robson,Julia Grant, Fred K. Lee, Mark T. Donnelly, Mark E. Mcalindon,David S. SandersIntroduction: Several scoring systems are used in the field of Gastroenterology.Although, previous studies have demonstrated the substantial risk of death followinggastrostomy insertion and some risk factors have been identified - no previousinvestigators have described a scoring system for gastrostomy insertion. Weundertook a prospective, unselected, dual centre study in order to establish therelative importance of risk factors for mortality after gastrostomy insertion. We thenformulated a simple numerical scoring system to categorize patients’ risk of death. Wethen sought to validate our scoring system on an independent second cohort ofpatients. Methods All patients referred for gastrostomy are prospectively included ina database along with demographic, biochemical and outcome data. We analysedgastrostomy insertions from 2 teaching hospitals from February 2004 to February2007. There were 887 referrals resulting in 837 gastrostomy insertions. The largestcohort was at site A and was used to construct a risk stratification scoring system. SiteB was used to validate the scoring system. Results Site A received 552 referrals and 403new gastrostomies were inserted (median age 64, 268 males). Overall 30 day mortalityrate was 51/403 (12.7%) with the highest risk in those with dementia (40%) followedby stroke (22.2%). Univariate analysis identified that 30 day mortality was associatedwith age (OR 3.4), albumin (OR 5.6), cardiac comorbidity (OR 2.0) and neurologicalcomorbidity (OR 1.7). On multivariate analysis only age and albumin remainedsignificant (both p ! 0.001) and were then modelled and attributed scores with agescoring 0 or 1 and albumin scoring 0, 1 or 2 giving composite scores from 0 to 3. Scoresof 0, 1, 2 and 3 gave 30 day mortalities (95% confidence interval) of 0% (0 - 2.1), 7% (2.9- 13.9), 21.3% (13.5 - 30.9) and 37.3% (24.1 - 51.9) respectively. Kaplan-Meier curvesstratified by total score showed a significantly increased mortality at 7 (p Z 0.0003), 30(p ! 0.0001) and 90 (p ! 0.0001) days. Site B (validation cohort) received 335referrals and inserted 153 new gastrostomies (median age 77, 64 males) with a 30 daymortality of 24/153 (15.7%). Application of the scoring system in this validation cohortgave comparable 30 day mortality figures of 0%, 7.7%, 15.6% and 29.3% for scores 0, 1,2 and 3 respectively. Conclusions: The Sheffield gastrostomy score can be used tocategorise patients being considered for gastrostomy insertion and to calculate risk ofdeath at 30 days. Further external validation is required.

T1523

ERCP-Related Perforations: Analysis of Patient-Outcome

After ERCP in Over 2700 PatientsSumit Kapoor, Devi Mukkai Krishnamurty, Frederick Eckhauser,Anthony N. Kalloo, Patrick Okolo, Sanjay B. Jagannath

AB238 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 5 : 2008

ERCP-related perforation is a serious complication. Identifying clinical factors thatinfluence hospital course may lead to better targeted clinical intervention andimprove patient outcome. AIM: To identify the clinical factors that affect outcome inpatients who suffered an ERCP-related perforation. Methods: IRB-approved,retrospective analysis of ERCP-related perforations at our institution from January2003 to November 2007. Patient demographics, procedural factors, laboratory andradiographic imaging results, and post-ERCP complications were recorded andanalyzed using length of stay (LOS), surgical intervention, and readmission rates asprimary endpoints. All procedures were performed by ERCP experts. Results: 2759ERCP procedures were performed during this time period, and 30 (1.1%) patients(21 Females, mean age 47.4 years) suffered an ERCP-related perforation. Indicationsfor ERCP were: SOD (40.0%), pancreatitis (16.7%), stone/stricture (13.3%),ampullary mass (10.0%), pancreas divisum (6.7%), PSC, abdominal pain, biloma,dilated pancreatic duct (3.3% each). The odds of perforation were higher (OR 40,CI 18-87) in SOD patients, as compared to non-SOD patients. 25 patientsunderwent endoscopic sphincterotomy and 16 patients had an endoprosthesisplaced. 96.7% (29/30) of the perforations were detected within 72 hours; 85.7%were detected by CT. A delay in perforation diagnosis significantly increased themean LOS (7.0 days if detected at ERCP vs. 15.7 days if detected within 24 hours vs.23.0 days if detected between 24-72 hours; p Z 0.048), as did the presence ordevelopment of one or more of the following clinical features: ascites, abscess, fluidcollection, pneumothorax, pneumomediastinum, pleural effusion,thrombophlebitis. In 13 patients who had concomitant post-ERCP pancreatitis, theodds of developing a retroperitoneal fluid collection (OR 13.2; p Z 0.006) washigher than in patients with perforation alone; however, there was no increase inLOS or need for surgical intervention. Patients who underwent surgical intervention(9/30) had a significantly longer LOS (37.7 days vs. 11.1 days; p Z 0.006) ascompared to patients managed non-operatively, and surgical patients were morelikely to be re-admitted (p ! 0.03). There was no mortality in this cohort.Conclusions: Early diagnosis of the perforation and non-operative management wasassociated with shorter hospitalization. LOS is increased with the development ofseveral clinical features. Surgical patients had higher readmission rates compared tonon-operative patients. This is the first study to identify clinical factors associatedwith patient outcome after suffering an ERCP-related perforation.

T1524

Comprehensive Colonoscopy Quality in a Community

Gastroenterology PracticeScott W. Rathgaber, Laura E. Black, Jacob Gundrum, Wendy Berth,Michelle A. MathiasonBackground: The efficacy of colonscopy as a diagnostic and therapeutic tool dependsupon the quality of the exam. The quality of colonoscopy within communitygastroenterology practices is largely unknown. Gastroenterology professionalsocieties advocate the collection and reporting of quality data. Purpose: To documenta range of colonoscopy quality measures within a community gastroenterologypractice. Methods: All colonoscopies attempted at a single endoscopy center between11/1/06 and 4/30/07 were reviewed for quality measures. Additional information wasobtained from screening exams only. All seven board-certified gastroenterologistswithin a single community gastroenterology group participated. Results: 2840consecutive colonoscopy exams were included, (48.9% male). 1374 exams were forscreening (48.4%). 509 exams were for surveillance (17.9%). Consent was documentedin the endoscopy note in 98.8%. Cecal intubation rate was 98.6%. Photodocumentationof the cecum was completed in 88.3%. The indications for colonoscopy followed ASGEguidelines in 95.6%. The most common indications for colonoscopy not included inASGE recommendations were abdominal pain (2.1%) and constipation (2.0%).Adenocarcinomas were detected in 40 patients, more commonly in patients referredfor symptoms (3.2%) versus screening (0.5%) or surveillance (0.4%) (p ! 0.0001).Three synchronous cancers were identified (7.5%), all in symptomatic patients. Of1374 screening exams (47.2% male), adenomas were detected in 38.7% of males and22.9% of females (p ! 0.0001). Polyps 1 cm and larger were detected in 5.9%. 50adenomas per 100 colonoscopies were detected within this screening group. Noscreening colonoscopy had a withdrawal time of less than 6 minutes. No differences incecal intubation rate or large adenoma detection were noted between individualgastroenterologists. Total adenoma detection (range 25�71 adenomas per 100colonoscopies) and withdrawal times (9.7�15.2 minutes) varied widely. However,higher detection did not correlate with slower withdrawal times (R Z 0.06).Conclusions: Community gastroenterology practices can demonstrate colonoscopyquality data that compare favorably to academic centers. Colorectal cancer is morelikely to be found in patients presenting with symptoms rather than for screening orsurveillance. Differences in polyp detection rates exist between gastroenterologists.Factors other than withdrawal times contribute to polyp detection rates.

T1525

Endoscopic Balloon Dilatation to Treat Stricture Caused By

Circumferential Resection of the Gastric Antrum By Endoscopic

Submucosal DissectionKotaro Mannen, Seiji Tsunada, Shinichi Ogata, Seiichiro Arima,Yasuhisa Sakata, Ryosuke Shiraishi, Ryo Shimoda, Hibiki Ootani,

www.giejournal.org