- Email: [email protected]
Idiopathic Recurrent Pancreatitis (IRP): Long Term Success of Pancreatic Endotherapy
Abstracts
were performed within 6 weeks (median: 4 weeks) (group I) and ten stent extractions were performed after 6 to 84 weeks (median: 7 weeks) (group II). Stent extractions were successful and without complications in all patients in group I (100%) vs 5 patients in group II (50%) (pZ0.001). Six complications occurred in group II, including 2 self-limiting bleedings, 3 stent fractures, and 1 stent removal by gastrotomy. In 2 patients complete stent removal was performed after 3 attempts and in 1 patient the stent was incompletely removed after 9 attempts. This patient subsequently developed a stenosis and a thoracic empyema and required esophageal resection 4 years after initial therapy. Conclusion: In patients with a benign esophageal perforation, stenting is effective and surgery can be avoided in 72% of the patients. Complications due to SEMS removal can be prevented by removal of the SEMS within 6 weeks after insertion.
influenced by years of experience or specialty of reviewer (ERCPist, radiologist or pancreatologist). Conclusion: Ability to diagnose AIP based on ERP features alone varied widely in different countries. ERP had the highest sensitivity for AIP when interpreted by experts from Asia, where it is most often used to make the diagnosis of AIP. Highlighting features of AIP on ERP may improve diagnostic yield and reduce interobserver variability in centers where ERP is not utilized as a diagnostic modality. Results
N. America (USA) NZ8
792 Endoscopic Pancreatic Sphincterotomy and Stenting for Preoperative Prophylaxis of Pancreatic Fistula After Distal Pancreatectomy Bernhard Rieder, Daniel Krampulz, Jost Adolf, Albrecht Pfeiffer Introduction: Distal pancreatectomy (DP) is the procedure of choice for pancreatic resections left of the superior mesenteric vein. Pancreatic fistula (PF) is the most common (0-61 %) postoperative complication. While the prevention of PF remains a major problem in DP, endoscopic pancreatic sphincterotomy and drainage has been shown to be an effective treatment of PF. Recently preoperative endoscopic pancreatic stenting was proposed to prevent PF. Aim: We investigated, if preoperative pancreatic sphincterotomy and stenting could prevent the incidence of pancreatic fistula in patients with DP. Method: Preoperative endoscopic pancreatic sphincterotomy and stenting was intended for the prophylaxis of PF after DP in 25 patients between 7/2004 and 10/2008. The incidence of PF was compared with a control group of 23 patients who underwent DP between 1/2001 and 3/2004 without preoperative endoscopic intervention. DP was performed by the same surgeon in both periods with identical management of the pancreatic remnant. PF was defined as an amylase level in surgical drainage fluid more than 5-fold the serum level. Results: Overall a cohort of 48 Patients underwent DP for adenocarcinoma (26), neuroendocrine tumor (7), intraductal papillary mucinous tumor (3), cystadenoma (4) and focal pancreatitis or pancreatic pseudocyst (8) between 1/2001 and 10/2008. In all 25 patients who underwent preoperative endoscopic pancreatic intervention sphincterotomy was successfully performed. A 5 French pancreatic stent was placed in all but two cases. Access through the minor papilla due to a pancreas divisum was required in two patients. One patient with an intraductal papillary mucinous tumor showed mild pancreatitis after the stent placement.PF developed in none of the 25 patients in the endoscopic intervention group. In the 23 patients without preoperative endoscopic intervention, 5 patients (22 %) developed a PF. The observed difference in the incidence of PF was statistically significant (P Z 0.02). Conclusion: Preoperative pancreatic sphincterotomy and stenting is a feasible and safe procedure. Prophylactic preoperative endoscopic intervention prevents pancreatic fistula development after distal pancreatectomy. These results encourage to perform a prospective trial.
793 Utility of Endoscopic Retrograde Pancreatogram (ERP) to Diagnose Autoimmune Pancreatitis (AIP): An International, Double Blind, Randomized, Multicenter Study Aravind Sugumar, Michael J. Levy, Terumi Kamisawa, Myung-Hwan Kim, Felicity Enders, Zahir Amin, Todd H. Baron, Mike H. Chapman, Nicholas I. Church, Jonathan E. Clain, Naoto Egawa, Gavin J. Johnson, Kazuichi Okazaki, Randall K. Pearson, Stephen P. Pereira, Bret T. Petersen, Samantha Read, Raghuwansh P. Sah, Neomal S. Sandanayake, Naoki Takahashi, Mark D. Topazian, Kazushige Uchida, Santhi Swaroop Vege, George Webster, Suresh T. Chari Background: The various diagnostic criteria for AIP differ in their emphasis on the need for ERP: the Japanese criteria mandate performance of ERP, the Korean criteria mandate either magnetic resonance pancreatogram or ERP and the Mayo Clinic HISORt criteria do not require either. The performance characteristics of ERP to diagnose AIP are unknown. Methods: ERPs (nZ164) of AIP, chronic pancreatitis (CP) and pancreatic cancer (PaC) obtained from centers in USA, Japan and UK were screened for quality by an experienced endoscopist (MJL) unaware of clinical diagnoses. A final set of 48 pancreatograms (20 AIP, 10 CP, 10 PaC and 8 internal duplicates) arranged in random order were presented to 21 reviewers from North America (USA), Europe (UK) and Asia (Japan and S. Korea) who were unaware of clinical data or diagnoses. Reviewers noted presence or absence of key ERP features and provided their 3 most probable diagnoses given as % confidence (O95%, 75%, 50% and 25%), the total not exceeding 100%. We used a O75% confidence in the diagnosis of AIP to determine sensitivity and specificity for that condition. The Kappa statistic for inter-observer agreement between reviews was also calculated. Results: The specificity of ERP for AIP was high and similar at different centers (Table). However, its sensitivity for AIP varied widely. As a group, Asian readers had a significantly higher sensitivity for diagnosis of AIP (see table) and had a higher inter-observer agreement. The performance characteristics of ERP were not
AB124 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 5 : 2009
Sensitivity for AIP (CI) 33% (26-41%) Specificity for AIP (CI) 92% (86-96%) Inter-observer agreement 0.16 for AIP (kappa)
Asia Japan (NZ4)/ S. p value Europe (UK) Korea (NZ1) Asia vs NZ5 Rest NZ8 40% (32-48%) 71% (61-79%) 0.00009 95% (90-98%) 89% (81-94%) NS 0.16 0.52
794 The Therapeutic Strategy for Autoimmune Pancreatitis Is Subject to Endoscopic Features of the Duodenal Papilla Kensuke Kubota, Shingo Kato, Noritoshi Kobayashi, Atsushi Nakajima Background: The factors related to spontaneous remission and relapse in cases of Autoimmune pancreatitis (AIP)have not yet been fully elucidated. Aim: To review the predictive factors for spontaneous remission and relapse included in the endoscopic features of AIP. Patients and Methods: 53 consecutive AIP patients were enrolled from April 2001 to Nov 2008. We conducted a review to determine if the following factors might be potentially related to spontaneous remission and relapse in patients with AIP: serological data (IgG, IgG4); the presence/absence of obstructive jaundice; the presence/absence of diabetes mellitus (DM); the presence/absence of sclerosing cholangitis affecting the hilar bile duct; duodenal papillary findings (the presence/ absence of swelling); distribution of the pancreatic swelling (focal/diffuse); the presence/absence of complicating extra-pancreatic lesions; the presence/absence of spontaneous remission; and the presence/absence of relapse. Results: The results of a univariate analysis revealed a significant association of spontaneous remission with serum negativity for IgG (!1800 mg/dl), the absence of obstructive jaundice, the absence of underlying DM, the absence of swelling of the duodenal papilla and localized as opposed to diffuse pancreatic swelling (p!0.05). A multivariate analysis revealed the absence of swollen duodenal papilla as the only significant independent factor predictive of spontaneous remission in these cases (ORZ1.96, pZ0.022). On the other hand, according to the results of the univariate analysis, relapse was associated with diffuse pancreatic swelling, the presence of duodenal papilla, the presence of extra-pancreatic lesions and a history of corticosteroid therapy (p!0.05), whereas the multivariate analysis revealed the presence of swollen duodenal papilla (ORZ14.8, pZ0.0182) and the presence of extra-pancreatic lesions (ORZ12.7, pZ0.0248) as the significant independent factors predictive of relapse in these cases. Conclusions: Our results suggested that the absence of a swollen duodenal papilla is a predictive factor for spontaneous remission, and that the administration of corticosteroids is unnecessary in such AIP patients. On the contrary, the presence of swollen duodenal papillae and the presence of extra-pancreatic lesions are risk factors for relapse, and those AIP patients are candidates for maintenance corticosteroid therapy to reduce relapse. Therefore, selection of the optimum therapeutic strategy, such as the indication for corticosteroid administration, depends on the endoscopic features of the duodenal papilla.
795 Idiopathic Recurrent Pancreatitis (IRP): Long Term Success of Pancreatic Endotherapy Marc F. Catalano, Michael H. Lee, Vivaik Tyagi, Nalini M. Guda, Roberto M. Gamarra, Joseph E. Geenen IRP presents a significant diagnostic and therapeutic challenge. An etiologic association can be identified in 60-80% of pts. In a minority of pts no abnormality can be detected despite exhaustive laboratory studies, radiographic (US, CT, MRI, ERCP) studies and history data. Several authors believe that many pts presenting with IRP may have microlithiasis, dysfunction of the biliary and/or pancreatic segment of the sphincter of Oddi, or anatomical variations. Several reports have described pancreatic endotherapy; pancreatic duct stenting and/or sphincterotomy as a primary treatment option for IRP: Aim: To evaluate the long term endoscopic treatment options in IRP; pancreatic duct stenting and/or sphincterotomy. Methods: Over a 5 year period 106 pts with acute recurrent pancreatitis (ARP) presented for evaluation and treatment. All underwent exhaustive laboratory and radiographic studies followed by ERCP with sphincter of Oddi manometry. Of these pts, 30/106 met the criteria for IRP (11 men/19 women, age 24-81, mean 50.4). Endoscopic treatment was individualized based on physician and patient preference; PD stent alone, sphincterotomy (ES) alone or dual therapy (simultaneous or sequential). Outcome measurement included, 1) episodes of
www.giejournal.org
Abstracts
acute pancreatitis (AP) following therapy, 2) # of ER visits or hospitalizations, 3) procedure related complications (AP, bleeding, stent related changes). Results: Thirty pts received a total of 93 stents (3.01 per pt). Size of stents included 7Fr (nZ62), 5Fr (nZ30), 10Fr (nZ1). Twelve of 30 pts underwent endoscopic sphincterotomy; this included Biliary ES (nZ4), pancreatic ES (nZ4) & dual ES (nZ4). Sphincterotomy was performed after stent therapy in 7pts & before stent therapy in 5 pts. Follow-up was a mean of 8.86 yrs (range 3-19yrs) & was available in only 22 of 30 pts. 11/22 (50%) pts had complete resolution of AP episodes. 8/22 (36%) had improved outcome (R50% reduction AP) while 3/22 (13%) had no improvement or worsening symptoms. Complications included post procedure pancreatitis 8/93 (8.5%). No stent induced changes were seen in any of our pts. 20 of 93 stents had distally migrated stents while 21 other stents were occluded at the time of removal. Conclusions: Pancreatic stent therapy in pts with IRP results in complete resolution or improvement in episodes of AP in the vast majority of pts. The addition of ES may further benefit outcomes. Randomized, controlled studies comparing stent alone vs. ES alone vs. combined therapy will reveal the most successful endoscopic treatment option.
Outcome
PT (n)
% Total
No AP following Stents Rx Improved (R50%) Worse/No Change AP
11 8 3 22
50% 36% 14% 100%
796 Pancreatic Stent-Induced Ductal Injury: Clinical Presentation and Outcomes of Endoscopic Therapy Yan Bakman, Martin L. Freeman Background: Pancreatic duct (PD) stent placement is increasingly performed for prevention of post-ERCP pancreatitis in patients at high risk. PD stents can result in injury especially in normal ducts, but the clinical significance and outcomes of subsequent endoscopic therapy are not known. Methods: Patients included all those referred from outside facilities to a single pancreaticobiliary center for management of stent-induced PD injury over 3 years. All pts had initially normal PD diameter without evidence of chronic pancreatitis, and all PD stents were placed for prevention of post-ERCP pancreatitis. Results: 8 pts were referred for symptomatic PD stent-induced duct injury. At initial outside ERCP, all had small caliber (! or Z 3 mm) PD with normal PD anatomy (nZ5) or pancreas divisum (nZ3). Indication for initial ERCP with prophylactic PD stent placement was suspected SOD with empirical biliary þ/- pancreatic sphincterotomy without SO manometry (nZ5), or minor papillotomy for pancreas divisum with acute recurrent pancreatitis (nZ1) or marginally elevated lipase (nZ2). All had conventional polyethylene 5F (nZ6) or 7F (nZ2) stents !4 cm in length in place for !2 weeks, except one pt who was lost to follow up with stent in place for one year. At presentation at a mean interval of 18 months after 1st ERCP, all pts had recurrent abdominal pain requiring hospitalization; of these 5/7 (71%) had acute recurrent pancreatitis, and 3/8 (38%) were on daily narcotics. By secretin MRCP/EUS/ERCP, all had a PD stricture within 2cm of major or minor papilla and 5/8 (63 %) upstream PD dilation (mean 4.3 mm, [2.2 mm-8.2 mm]). 3/6 (50%) pts undergoing EUS had new parenchymal changes suggestive of chronic pancreatitis. All patients were treated with pancreatic sphincterotomy (if not done already), balloon dilation of stricture, and placement of multiple 3-5 F soft polymer pancreatic stents depending on duct diameter; Six required multiple (2-11) ERCPs for therapy. All had improvement or resolution of pancreatic strictures and recurrent pancreatitis. 50% had sustained clinical response with resolution of pain, 13% fair response with repeated ERCPs, and 25% with poor response underwent total pancreatectomy with islet autotransplantat. Conclusions: PD stent-induced ductal injury with significant clinical consequences can occur after a relatively brief interval of stenting using conventional polyethylene 5F stents. Endoscopic therapy is moderately effective but some patients develop irreversible damage. Further investigation is required to determine prevalence and risk factors for stent-induced injury and to improve configuration and material of stents.
Randomization was by variable block size. CO2 insufflation was regulated by the Olympus Endoscopic CO2 Regulation Unit (Olympus Medical). Patients, endoscopists, nurses, and data analysts were blinded as to allocation. A nurse not involved with the procedure or data collection set-up and masked the equipment. Conscious sedation with midazolam and fentanyl was used. Primary outcomes were abdominal pain assessed on a 100mm visual analogue scale pre- and 1, 3, 6, and 24 hrs post-ERCP, and abdominal girth (cm) pre- and post-ERCP. Key secondary outcomes were pCO2 levels (mm Hg) measured with transcutaneous capnography and procedural complications. Outcomes were compared using t-tests and chi-square. This is a planned interim analysis. Results: Baseline demographic characteristics, indications for ERCP, and comorbidities were similar in the air (nZ26) and CO2 (nZ24) groups, though cannulation time, procedure time, and procedure failure rate were all higher in the CO2 group. Mean pain scores in the air and CO2 groups were 11 and 12 at baseline (pZ0.89), 8 and 18 at 1 hr (pZ0.08), 12 and 23 at 3 hr (pZ0.12), 13 and 20 at 6 hr (pZ0.41), and 14 and 17 at 24 hr (pZ0.66). Abdominal girth increased a mean of 0.8 cm in the air group and 0.3 cm in the CO2 group (pZ0.73). Baseline pCO2 levels were 41 in the air group and 40 in the CO2 group (pZ0.68), and increased to a mean of 50 in the air group and 48 in the CO2 group (pZ0.57). One patient in the air group and none in the CO2 group had respiratory depression. There were no differences in other complications. Conclusions: The use of CO2 for insufflation during ERCP was safe in a tertiary care referral population. While there was a mild rise in pCO2 in both groups, this was not associated with respiratory complications. Despite procedures being more difficult in the CO2 group, there was no difference between the air and CO2 groups in pain scores or abdominal distention. Patient accrual is ongoing for this study.
798 Enhanced Harmonic Endoscopic Ultrasonography for Assessment of Pancreaticobiliary Malignancies Kazuki Sumiyama, Hiroo Imazu, Hisao Tajiri Background: Endoscopic ultrasonography (EUS) is well established as the most sensitive imaging modality in the detection, differentiation and T-staging of pancreaticobiliary (PB) malignancies. Recently, Sonazoid, a second generation microbubble contrast agent for acoustic imaging, was developed to improve visualization of hepatic lesions for abdominal ultrasonography. This novel contrast agent is visualized with the resonance of gas bubbles at lower acoustic power relative to first generation contrast agents. It is therefore hypothesized that Sonazoid may be well suited for EUS. Aim: To evaluate if the use of Sonazoid can improve EUS assessment of PB cancers in a blinded design. Method: The study protocol was approved by the institutional review board of Jikei University School of Medicine. Patients referred for EUS to evaluate suspected pancreatic and extrahepatic biliary malignancies from October 2007 to October 2008 at Jikei University hospital were enrolled in this study, and all patients provided written informed consent. First, standard screening of the PB system with harmonic imaging (H-EUS) was performed. Enhanced harmonic (EH) EUS was subsequently performed after intravenous infusion of 0.015 mL/kg of Sonazoid. Ultrasound data from H-EUS and EH-EUS were digitally recorded, and the unedited videos were reviewed by a blinded interpreter for assessment of tumor detection and T-staging. A definite diagnosis was made of either EUS-guided fine-needle aspiration or surgical pathology in tumor detection, and surgical pathology in T-staging. Results: Sixty patients underwent EUS evaluation for suspected PB malignancies. Sensitivity and specificity of tumor detection were 92.5% and 90%, respectively, with H-EUS, and 100% and 95%, respectively, with EH-EUS (not significant). In 25 of 60 patients, surgical excision of the primary tumor was successful and eligible for T-stage assessment. Overall accuracy of H-EUS and EH-EUS for T-staging was 70% and 91.7%, respectively (p!0.05). There was discrepancy between H-EUS and EH-EUS for T-staging in six cases. In each of these cases, use of the contrast indicated misdiagnosis with H-EUS (2/10 pancreatic cancers, 1/7 gallbladder cancers, 2/4 bile duct cancers and 1/4 ampullary cancers). Imaging contrast between the anatomic tissue layers such as the mucosa, submucosa and vessel walls was more distinct with EH-EUS. Conclusion: This is the first report to use the second generation ultrasound contrast with EUS staging for PB malignancies in a blinded design. Use of the contrast may improve the accuracy of EUS staging further.
797 A Randomized, Controlled, Double-Blind Trial of Air vs Carbon Dioxide Insufflation During ERCP Evan S. Dellon, Arumugam Velayudham, Bridger W. Clarke, Kim L. Isaacs, Lisa M. Gangarosa, Joseph Galanko, Ian S. Grimm
799 EUS-Guided Ethanol Lavage with Paclitaxel Injection (EUS-EP) for Cystic Tumors of the Pancreas (CTP): Long-Term (More Than One Year) Follow-Up Hyoung-Chul OH, Dong Wan Seo, Tae Jun Song, Sung-Hoon Moon, Do Hyun Park, Sang Soo Lee, Sung Koo Lee, Myung-Hwan Kim
Background: Insufflation of carbon dioxide (CO2) has been shown to decrease patient pain and bloating after colonoscopy, but has not been as extensively studied for use during ERCP. Aim: To assess the safety (transcutaneously measured pCO2; respiratory complications) and efficacy (post-procedural abdominal pain and distention) of CO2 insufflation during ERCP. Methods: We performed a randomized, controlled, doubleblind trial of air vs CO2 insufflation for patients undergoing ERCP at University of North Carolina. Exclusion criteria were age !18, COPD requiring oxygen, known CO2 retention, same-day second endoscopy, use of chronic opiates for pain, and pregnancy.
Background: EUS-guided intervention has been recently tried for the treatment of CTP. Preliminary reports showed the safety and feasibility of EUS-guided intervention. Treatment responses were also somewhat promising, but differed among studies as previous studies involved small number of patients and shortterm follow-up. Its effectiveness, therefore, should be evaluated by the study with larger population and longer follow-up. The present study analyzed the treatment response of EUS-EP over longer period among larger study population, and factors which may influence the treatment response. Methods: Thirty-three patients were
www.giejournal.org
Volume 69, No. 5 : 2009 GASTROINTESTINAL ENDOSCOPY AB125
were performed within 6 weeks (median: 4 weeks) (group I) and ten stent extractions were performed after 6 to 84 weeks (median: 7 weeks) (group II). Stent extractions were successful and without complications in all patients in group I (100%) vs 5 patients in group II (50%) (pZ0.001). Six complications occurred in group II, including 2 self-limiting bleedings, 3 stent fractures, and 1 stent removal by gastrotomy. In 2 patients complete stent removal was performed after 3 attempts and in 1 patient the stent was incompletely removed after 9 attempts. This patient subsequently developed a stenosis and a thoracic empyema and required esophageal resection 4 years after initial therapy. Conclusion: In patients with a benign esophageal perforation, stenting is effective and surgery can be avoided in 72% of the patients. Complications due to SEMS removal can be prevented by removal of the SEMS within 6 weeks after insertion.
influenced by years of experience or specialty of reviewer (ERCPist, radiologist or pancreatologist). Conclusion: Ability to diagnose AIP based on ERP features alone varied widely in different countries. ERP had the highest sensitivity for AIP when interpreted by experts from Asia, where it is most often used to make the diagnosis of AIP. Highlighting features of AIP on ERP may improve diagnostic yield and reduce interobserver variability in centers where ERP is not utilized as a diagnostic modality. Results
N. America (USA) NZ8
792 Endoscopic Pancreatic Sphincterotomy and Stenting for Preoperative Prophylaxis of Pancreatic Fistula After Distal Pancreatectomy Bernhard Rieder, Daniel Krampulz, Jost Adolf, Albrecht Pfeiffer Introduction: Distal pancreatectomy (DP) is the procedure of choice for pancreatic resections left of the superior mesenteric vein. Pancreatic fistula (PF) is the most common (0-61 %) postoperative complication. While the prevention of PF remains a major problem in DP, endoscopic pancreatic sphincterotomy and drainage has been shown to be an effective treatment of PF. Recently preoperative endoscopic pancreatic stenting was proposed to prevent PF. Aim: We investigated, if preoperative pancreatic sphincterotomy and stenting could prevent the incidence of pancreatic fistula in patients with DP. Method: Preoperative endoscopic pancreatic sphincterotomy and stenting was intended for the prophylaxis of PF after DP in 25 patients between 7/2004 and 10/2008. The incidence of PF was compared with a control group of 23 patients who underwent DP between 1/2001 and 3/2004 without preoperative endoscopic intervention. DP was performed by the same surgeon in both periods with identical management of the pancreatic remnant. PF was defined as an amylase level in surgical drainage fluid more than 5-fold the serum level. Results: Overall a cohort of 48 Patients underwent DP for adenocarcinoma (26), neuroendocrine tumor (7), intraductal papillary mucinous tumor (3), cystadenoma (4) and focal pancreatitis or pancreatic pseudocyst (8) between 1/2001 and 10/2008. In all 25 patients who underwent preoperative endoscopic pancreatic intervention sphincterotomy was successfully performed. A 5 French pancreatic stent was placed in all but two cases. Access through the minor papilla due to a pancreas divisum was required in two patients. One patient with an intraductal papillary mucinous tumor showed mild pancreatitis after the stent placement.PF developed in none of the 25 patients in the endoscopic intervention group. In the 23 patients without preoperative endoscopic intervention, 5 patients (22 %) developed a PF. The observed difference in the incidence of PF was statistically significant (P Z 0.02). Conclusion: Preoperative pancreatic sphincterotomy and stenting is a feasible and safe procedure. Prophylactic preoperative endoscopic intervention prevents pancreatic fistula development after distal pancreatectomy. These results encourage to perform a prospective trial.
793 Utility of Endoscopic Retrograde Pancreatogram (ERP) to Diagnose Autoimmune Pancreatitis (AIP): An International, Double Blind, Randomized, Multicenter Study Aravind Sugumar, Michael J. Levy, Terumi Kamisawa, Myung-Hwan Kim, Felicity Enders, Zahir Amin, Todd H. Baron, Mike H. Chapman, Nicholas I. Church, Jonathan E. Clain, Naoto Egawa, Gavin J. Johnson, Kazuichi Okazaki, Randall K. Pearson, Stephen P. Pereira, Bret T. Petersen, Samantha Read, Raghuwansh P. Sah, Neomal S. Sandanayake, Naoki Takahashi, Mark D. Topazian, Kazushige Uchida, Santhi Swaroop Vege, George Webster, Suresh T. Chari Background: The various diagnostic criteria for AIP differ in their emphasis on the need for ERP: the Japanese criteria mandate performance of ERP, the Korean criteria mandate either magnetic resonance pancreatogram or ERP and the Mayo Clinic HISORt criteria do not require either. The performance characteristics of ERP to diagnose AIP are unknown. Methods: ERPs (nZ164) of AIP, chronic pancreatitis (CP) and pancreatic cancer (PaC) obtained from centers in USA, Japan and UK were screened for quality by an experienced endoscopist (MJL) unaware of clinical diagnoses. A final set of 48 pancreatograms (20 AIP, 10 CP, 10 PaC and 8 internal duplicates) arranged in random order were presented to 21 reviewers from North America (USA), Europe (UK) and Asia (Japan and S. Korea) who were unaware of clinical data or diagnoses. Reviewers noted presence or absence of key ERP features and provided their 3 most probable diagnoses given as % confidence (O95%, 75%, 50% and 25%), the total not exceeding 100%. We used a O75% confidence in the diagnosis of AIP to determine sensitivity and specificity for that condition. The Kappa statistic for inter-observer agreement between reviews was also calculated. Results: The specificity of ERP for AIP was high and similar at different centers (Table). However, its sensitivity for AIP varied widely. As a group, Asian readers had a significantly higher sensitivity for diagnosis of AIP (see table) and had a higher inter-observer agreement. The performance characteristics of ERP were not
AB124 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 5 : 2009
Sensitivity for AIP (CI) 33% (26-41%) Specificity for AIP (CI) 92% (86-96%) Inter-observer agreement 0.16 for AIP (kappa)
Asia Japan (NZ4)/ S. p value Europe (UK) Korea (NZ1) Asia vs NZ5 Rest NZ8 40% (32-48%) 71% (61-79%) 0.00009 95% (90-98%) 89% (81-94%) NS 0.16 0.52
794 The Therapeutic Strategy for Autoimmune Pancreatitis Is Subject to Endoscopic Features of the Duodenal Papilla Kensuke Kubota, Shingo Kato, Noritoshi Kobayashi, Atsushi Nakajima Background: The factors related to spontaneous remission and relapse in cases of Autoimmune pancreatitis (AIP)have not yet been fully elucidated. Aim: To review the predictive factors for spontaneous remission and relapse included in the endoscopic features of AIP. Patients and Methods: 53 consecutive AIP patients were enrolled from April 2001 to Nov 2008. We conducted a review to determine if the following factors might be potentially related to spontaneous remission and relapse in patients with AIP: serological data (IgG, IgG4); the presence/absence of obstructive jaundice; the presence/absence of diabetes mellitus (DM); the presence/absence of sclerosing cholangitis affecting the hilar bile duct; duodenal papillary findings (the presence/ absence of swelling); distribution of the pancreatic swelling (focal/diffuse); the presence/absence of complicating extra-pancreatic lesions; the presence/absence of spontaneous remission; and the presence/absence of relapse. Results: The results of a univariate analysis revealed a significant association of spontaneous remission with serum negativity for IgG (!1800 mg/dl), the absence of obstructive jaundice, the absence of underlying DM, the absence of swelling of the duodenal papilla and localized as opposed to diffuse pancreatic swelling (p!0.05). A multivariate analysis revealed the absence of swollen duodenal papilla as the only significant independent factor predictive of spontaneous remission in these cases (ORZ1.96, pZ0.022). On the other hand, according to the results of the univariate analysis, relapse was associated with diffuse pancreatic swelling, the presence of duodenal papilla, the presence of extra-pancreatic lesions and a history of corticosteroid therapy (p!0.05), whereas the multivariate analysis revealed the presence of swollen duodenal papilla (ORZ14.8, pZ0.0182) and the presence of extra-pancreatic lesions (ORZ12.7, pZ0.0248) as the significant independent factors predictive of relapse in these cases. Conclusions: Our results suggested that the absence of a swollen duodenal papilla is a predictive factor for spontaneous remission, and that the administration of corticosteroids is unnecessary in such AIP patients. On the contrary, the presence of swollen duodenal papillae and the presence of extra-pancreatic lesions are risk factors for relapse, and those AIP patients are candidates for maintenance corticosteroid therapy to reduce relapse. Therefore, selection of the optimum therapeutic strategy, such as the indication for corticosteroid administration, depends on the endoscopic features of the duodenal papilla.
795 Idiopathic Recurrent Pancreatitis (IRP): Long Term Success of Pancreatic Endotherapy Marc F. Catalano, Michael H. Lee, Vivaik Tyagi, Nalini M. Guda, Roberto M. Gamarra, Joseph E. Geenen IRP presents a significant diagnostic and therapeutic challenge. An etiologic association can be identified in 60-80% of pts. In a minority of pts no abnormality can be detected despite exhaustive laboratory studies, radiographic (US, CT, MRI, ERCP) studies and history data. Several authors believe that many pts presenting with IRP may have microlithiasis, dysfunction of the biliary and/or pancreatic segment of the sphincter of Oddi, or anatomical variations. Several reports have described pancreatic endotherapy; pancreatic duct stenting and/or sphincterotomy as a primary treatment option for IRP: Aim: To evaluate the long term endoscopic treatment options in IRP; pancreatic duct stenting and/or sphincterotomy. Methods: Over a 5 year period 106 pts with acute recurrent pancreatitis (ARP) presented for evaluation and treatment. All underwent exhaustive laboratory and radiographic studies followed by ERCP with sphincter of Oddi manometry. Of these pts, 30/106 met the criteria for IRP (11 men/19 women, age 24-81, mean 50.4). Endoscopic treatment was individualized based on physician and patient preference; PD stent alone, sphincterotomy (ES) alone or dual therapy (simultaneous or sequential). Outcome measurement included, 1) episodes of
www.giejournal.org
Abstracts
acute pancreatitis (AP) following therapy, 2) # of ER visits or hospitalizations, 3) procedure related complications (AP, bleeding, stent related changes). Results: Thirty pts received a total of 93 stents (3.01 per pt). Size of stents included 7Fr (nZ62), 5Fr (nZ30), 10Fr (nZ1). Twelve of 30 pts underwent endoscopic sphincterotomy; this included Biliary ES (nZ4), pancreatic ES (nZ4) & dual ES (nZ4). Sphincterotomy was performed after stent therapy in 7pts & before stent therapy in 5 pts. Follow-up was a mean of 8.86 yrs (range 3-19yrs) & was available in only 22 of 30 pts. 11/22 (50%) pts had complete resolution of AP episodes. 8/22 (36%) had improved outcome (R50% reduction AP) while 3/22 (13%) had no improvement or worsening symptoms. Complications included post procedure pancreatitis 8/93 (8.5%). No stent induced changes were seen in any of our pts. 20 of 93 stents had distally migrated stents while 21 other stents were occluded at the time of removal. Conclusions: Pancreatic stent therapy in pts with IRP results in complete resolution or improvement in episodes of AP in the vast majority of pts. The addition of ES may further benefit outcomes. Randomized, controlled studies comparing stent alone vs. ES alone vs. combined therapy will reveal the most successful endoscopic treatment option.
Outcome
PT (n)
% Total
No AP following Stents Rx Improved (R50%) Worse/No Change AP
11 8 3 22
50% 36% 14% 100%
796 Pancreatic Stent-Induced Ductal Injury: Clinical Presentation and Outcomes of Endoscopic Therapy Yan Bakman, Martin L. Freeman Background: Pancreatic duct (PD) stent placement is increasingly performed for prevention of post-ERCP pancreatitis in patients at high risk. PD stents can result in injury especially in normal ducts, but the clinical significance and outcomes of subsequent endoscopic therapy are not known. Methods: Patients included all those referred from outside facilities to a single pancreaticobiliary center for management of stent-induced PD injury over 3 years. All pts had initially normal PD diameter without evidence of chronic pancreatitis, and all PD stents were placed for prevention of post-ERCP pancreatitis. Results: 8 pts were referred for symptomatic PD stent-induced duct injury. At initial outside ERCP, all had small caliber (! or Z 3 mm) PD with normal PD anatomy (nZ5) or pancreas divisum (nZ3). Indication for initial ERCP with prophylactic PD stent placement was suspected SOD with empirical biliary þ/- pancreatic sphincterotomy without SO manometry (nZ5), or minor papillotomy for pancreas divisum with acute recurrent pancreatitis (nZ1) or marginally elevated lipase (nZ2). All had conventional polyethylene 5F (nZ6) or 7F (nZ2) stents !4 cm in length in place for !2 weeks, except one pt who was lost to follow up with stent in place for one year. At presentation at a mean interval of 18 months after 1st ERCP, all pts had recurrent abdominal pain requiring hospitalization; of these 5/7 (71%) had acute recurrent pancreatitis, and 3/8 (38%) were on daily narcotics. By secretin MRCP/EUS/ERCP, all had a PD stricture within 2cm of major or minor papilla and 5/8 (63 %) upstream PD dilation (mean 4.3 mm, [2.2 mm-8.2 mm]). 3/6 (50%) pts undergoing EUS had new parenchymal changes suggestive of chronic pancreatitis. All patients were treated with pancreatic sphincterotomy (if not done already), balloon dilation of stricture, and placement of multiple 3-5 F soft polymer pancreatic stents depending on duct diameter; Six required multiple (2-11) ERCPs for therapy. All had improvement or resolution of pancreatic strictures and recurrent pancreatitis. 50% had sustained clinical response with resolution of pain, 13% fair response with repeated ERCPs, and 25% with poor response underwent total pancreatectomy with islet autotransplantat. Conclusions: PD stent-induced ductal injury with significant clinical consequences can occur after a relatively brief interval of stenting using conventional polyethylene 5F stents. Endoscopic therapy is moderately effective but some patients develop irreversible damage. Further investigation is required to determine prevalence and risk factors for stent-induced injury and to improve configuration and material of stents.
Randomization was by variable block size. CO2 insufflation was regulated by the Olympus Endoscopic CO2 Regulation Unit (Olympus Medical). Patients, endoscopists, nurses, and data analysts were blinded as to allocation. A nurse not involved with the procedure or data collection set-up and masked the equipment. Conscious sedation with midazolam and fentanyl was used. Primary outcomes were abdominal pain assessed on a 100mm visual analogue scale pre- and 1, 3, 6, and 24 hrs post-ERCP, and abdominal girth (cm) pre- and post-ERCP. Key secondary outcomes were pCO2 levels (mm Hg) measured with transcutaneous capnography and procedural complications. Outcomes were compared using t-tests and chi-square. This is a planned interim analysis. Results: Baseline demographic characteristics, indications for ERCP, and comorbidities were similar in the air (nZ26) and CO2 (nZ24) groups, though cannulation time, procedure time, and procedure failure rate were all higher in the CO2 group. Mean pain scores in the air and CO2 groups were 11 and 12 at baseline (pZ0.89), 8 and 18 at 1 hr (pZ0.08), 12 and 23 at 3 hr (pZ0.12), 13 and 20 at 6 hr (pZ0.41), and 14 and 17 at 24 hr (pZ0.66). Abdominal girth increased a mean of 0.8 cm in the air group and 0.3 cm in the CO2 group (pZ0.73). Baseline pCO2 levels were 41 in the air group and 40 in the CO2 group (pZ0.68), and increased to a mean of 50 in the air group and 48 in the CO2 group (pZ0.57). One patient in the air group and none in the CO2 group had respiratory depression. There were no differences in other complications. Conclusions: The use of CO2 for insufflation during ERCP was safe in a tertiary care referral population. While there was a mild rise in pCO2 in both groups, this was not associated with respiratory complications. Despite procedures being more difficult in the CO2 group, there was no difference between the air and CO2 groups in pain scores or abdominal distention. Patient accrual is ongoing for this study.
798 Enhanced Harmonic Endoscopic Ultrasonography for Assessment of Pancreaticobiliary Malignancies Kazuki Sumiyama, Hiroo Imazu, Hisao Tajiri Background: Endoscopic ultrasonography (EUS) is well established as the most sensitive imaging modality in the detection, differentiation and T-staging of pancreaticobiliary (PB) malignancies. Recently, Sonazoid, a second generation microbubble contrast agent for acoustic imaging, was developed to improve visualization of hepatic lesions for abdominal ultrasonography. This novel contrast agent is visualized with the resonance of gas bubbles at lower acoustic power relative to first generation contrast agents. It is therefore hypothesized that Sonazoid may be well suited for EUS. Aim: To evaluate if the use of Sonazoid can improve EUS assessment of PB cancers in a blinded design. Method: The study protocol was approved by the institutional review board of Jikei University School of Medicine. Patients referred for EUS to evaluate suspected pancreatic and extrahepatic biliary malignancies from October 2007 to October 2008 at Jikei University hospital were enrolled in this study, and all patients provided written informed consent. First, standard screening of the PB system with harmonic imaging (H-EUS) was performed. Enhanced harmonic (EH) EUS was subsequently performed after intravenous infusion of 0.015 mL/kg of Sonazoid. Ultrasound data from H-EUS and EH-EUS were digitally recorded, and the unedited videos were reviewed by a blinded interpreter for assessment of tumor detection and T-staging. A definite diagnosis was made of either EUS-guided fine-needle aspiration or surgical pathology in tumor detection, and surgical pathology in T-staging. Results: Sixty patients underwent EUS evaluation for suspected PB malignancies. Sensitivity and specificity of tumor detection were 92.5% and 90%, respectively, with H-EUS, and 100% and 95%, respectively, with EH-EUS (not significant). In 25 of 60 patients, surgical excision of the primary tumor was successful and eligible for T-stage assessment. Overall accuracy of H-EUS and EH-EUS for T-staging was 70% and 91.7%, respectively (p!0.05). There was discrepancy between H-EUS and EH-EUS for T-staging in six cases. In each of these cases, use of the contrast indicated misdiagnosis with H-EUS (2/10 pancreatic cancers, 1/7 gallbladder cancers, 2/4 bile duct cancers and 1/4 ampullary cancers). Imaging contrast between the anatomic tissue layers such as the mucosa, submucosa and vessel walls was more distinct with EH-EUS. Conclusion: This is the first report to use the second generation ultrasound contrast with EUS staging for PB malignancies in a blinded design. Use of the contrast may improve the accuracy of EUS staging further.
797 A Randomized, Controlled, Double-Blind Trial of Air vs Carbon Dioxide Insufflation During ERCP Evan S. Dellon, Arumugam Velayudham, Bridger W. Clarke, Kim L. Isaacs, Lisa M. Gangarosa, Joseph Galanko, Ian S. Grimm
799 EUS-Guided Ethanol Lavage with Paclitaxel Injection (EUS-EP) for Cystic Tumors of the Pancreas (CTP): Long-Term (More Than One Year) Follow-Up Hyoung-Chul OH, Dong Wan Seo, Tae Jun Song, Sung-Hoon Moon, Do Hyun Park, Sang Soo Lee, Sung Koo Lee, Myung-Hwan Kim
Background: Insufflation of carbon dioxide (CO2) has been shown to decrease patient pain and bloating after colonoscopy, but has not been as extensively studied for use during ERCP. Aim: To assess the safety (transcutaneously measured pCO2; respiratory complications) and efficacy (post-procedural abdominal pain and distention) of CO2 insufflation during ERCP. Methods: We performed a randomized, controlled, doubleblind trial of air vs CO2 insufflation for patients undergoing ERCP at University of North Carolina. Exclusion criteria were age !18, COPD requiring oxygen, known CO2 retention, same-day second endoscopy, use of chronic opiates for pain, and pregnancy.
Background: EUS-guided intervention has been recently tried for the treatment of CTP. Preliminary reports showed the safety and feasibility of EUS-guided intervention. Treatment responses were also somewhat promising, but differed among studies as previous studies involved small number of patients and shortterm follow-up. Its effectiveness, therefore, should be evaluated by the study with larger population and longer follow-up. The present study analyzed the treatment response of EUS-EP over longer period among larger study population, and factors which may influence the treatment response. Methods: Thirty-three patients were
www.giejournal.org
Volume 69, No. 5 : 2009 GASTROINTESTINAL ENDOSCOPY AB125