- Email: [email protected]
Successful Endoscopic Placement of Feeding Tubes Into the Small Intestine Using the Inscope Endorail
Abstracts
T1334 Successful Endoscopic Placement of Feeding Tubes Into the Small Intestine Using the Inscope Endorail Douglas O. Faigel, Suzanne Thompson, David Stefanchik Background: Endoscopic placement of feeding tubes through the pylorus into the small intestine is technically difficult. When pulled down along side of the endoscope, retrograde migration out of the small intestine proximally into the stomach or esophagus is frequently observed. We used a new device, the EndoRail, to achieve transpyloric placement of feeding tubes. Methods: The EndoRail consists of a thin plastic sheath and a plastic rail. The scope is inserted through the sheath allowing the rail to serve as an exterior channel along which accessories may be passed. We fashioned 140 cm length 10 Fr Dobb-Hoff-type feeding tubes (Viasys Healthcare Inc.) with a compatible rail to allow insertion along the EndoRail affixed to a pediatric colonoscope (Olympus PCF100). The colonoscope was inserted perorally, deeply intubated into the small intestine and the feeding tube then slid along the EndoRail to the tip of the scope. The tube is left in place by backing the scope out while advancing the tube along the rail. We compared the EndoRail technique to a standard technique of using a polypectomy snare to drag a similar tube through the pylorus and deeply into the small intestine in 4 mongrel dogs [mean weight 21.6 kg (Range 18.6 to 24.1 kg)]; 3 attempts with each technique/ dog. Two metallic clips were affixed to the prepyloric area to allow fluoroscopic monitoring. Successful placement was defined as the feeding ports remaining in the small intestine as seen on fluoroscopy. The study was powered to detect a success rate difference of 80% vs. 30%. Two tailed p-values were calculated with Fisher’s exact test. Results: In all cases deep intubation of the small intestine was achieved (130-140 cm from the incisors). Successful placement was achieved in 9/12 (75%) attempts with the EndoRail but only 1/12 (8%) attempts with the snare (p Z 0.001). Failures with the snare were all due to proximal tube migration during scope withdrawal. The 3 EndoRail failures occurred in 3 different dogs and appeared to be due to difficulty in passing the feeding tube over the rail and through the pylorus. The time to place feeding tubes with EndoRail was 5.24 G 1.17 mins (mean G SD). Conclusion: The EndoRail allows for quick and successful transpyloric placement of feeding tubes into the small intestine. Optimization of the design should further improve the success rate and ease of use.
T1335 Inter-Observer Variability with Enhanced Magnification Endoscopy of the Squamocolumnar Junction Dawn D. Ferguson, Kenneth R. DeVault, David S. Loeb, Michael B. Wallace Background: Enhanced Magnification Endoscopy (EME) is an endoscopic technique that combines the application of acetic acid to the surface of the esophagus and endoscopic imaging with a magnifying endoscope. This technique has been identified as highly accurate in predicting specialized intestinal metaplasia (SIM) in Barrett’s esophagus based on surface pattern types. Previous investigators have noted that there is high inter-observer variability in identifying pattern types that may limit the utility of this technique. We sought to reexamine inter-observer variability in a group of images obtained as part of an ongoing, prospective study. Methods: 44 patients with GERD symptoms were prospectively evaluated with EME. Pictures were taken of representative pattern types. 50 endoscopic pictures were shown in a blinded fashion to four experienced endoscopists who are participating in a prospective trial of EME. Each endoscopist was asked to identify the pattern type based on previously published criteria and a provided atlas of previously published images. The patterns were Type I) round pits, a characteristic pit pattern of orderly arranged circular dots; II) reticular, pits circular or oval and regular in arrangement; III) villous, no pits present but a fine villiform appearance, with the villi having a regular arrangement; and IV) ridged, no pits present, but a thick convoluted appearance (cerebriform) with a regular arrangement of the villi. Results: The value of Cohen’s Kappa corresponding to the agreement in pattern identification of the 4 observers on 50 endoscopic photographs is 0.41. All four observers agreed on the pattern type for 14 pictures (28%); three observers agreed 19 (38%) times; and only two observers agreed on the remaining 17 (34%). In the 33 patients where there was
www.mosby.com/gie
agreement between either 3 or 4 of the observers, the best agreement was seen in patients with predominantly type IV (figure). Conclusions: The previously published criteria for identifying SIM using EME are associated with a high level of inter-observer variability. Of the 4 described pattern types, it appears the best accuracy between observers is seen with Type IV. This inter-observer variability may limit the applicability of this type of technique in the diagnosis of SIM in endoscopically apparent Barrett’s esophagus.
T1336 Enhanced Magnification Directed Biopsies for Detection of Intestinal Metaplasia in Patients with GERD Dawn D. Ferguson, Kenneth R. DeVault, David S. Loeb, Murli Krishna, Michael B. Wallace Background: The diagnosis of Barrett’s esophagus (BE) requires histologic confirmation of specialized intestinal metaplasia (SIM) through biopsy, a technique prone to sampling error. A promising method to improve the yield of biopsy uses acetic acid with magnification endoscopy: Enhanced Magnification Endoscopy (EME). This technique identifies mucosal surface patterns and of these, pattern types III and IV have been associated with SIM. Methods: We conducted a prospective, randomized trial to evaluate EME directed biopsies and standard endoscopy with random biopsies in patients with symptoms of GERD. Patients in the standard endoscopy group with evidence of BE had 4 quadrant random biopsies taken every 2 cms. If there was no BE, 4 quadrant biopsies were taken at the SCJ. Patients in the EME group had the mucosa at the SCJ classified as Type I) round pits, Type II) reticular, III) villous or IV) ridged based on published criteria. In some cases, the visualized pattern did not fall into one of these categories and was labeled Type V. Biopsies were taken from each pattern type. Results: 93 patients enrolled (44 with EME, 49 with standard endoscopy). 36 (39%) had endoscopic evidence of BE (12 standard endoscopy, 24 EME). The yield of SIM using EME for each pattern type is shown (table). Of the 12 patients with apparent BE on standard endoscopy, random biopsies confirmed SIM on a per patient basis in 6 (50%) and in 6 (25%) for the 24 with EME directed biopsies. Overall, the prevalence of SIM was higher in the standard endoscopy group (50%) vs. EME (25%), but if pattern type III or IV were identified, EME confirmed SIM in 75% (6/8 patients). Patients without apparent BE (57) had SIM of the SCJ confirmed in 16% (6/37) with standard endoscopy and random biopsies compared with 5% (1/20) using magnification endoscopy. Conclusions: Random biopsies of endoscopically apparent BE yield SIM at slightly lower rates than targeted biopsies with EME in patients with pattern types III or IV. However, overall it appears that the yield of biopsy confirming SIM using these two techniques is similar. This calls into question the utility of this technique in reducing sampling error to identify SIM.
Volume 61, No. 5 : 2005 GASTROINTESTINAL ENDOSCOPY AB225
T1334 Successful Endoscopic Placement of Feeding Tubes Into the Small Intestine Using the Inscope Endorail Douglas O. Faigel, Suzanne Thompson, David Stefanchik Background: Endoscopic placement of feeding tubes through the pylorus into the small intestine is technically difficult. When pulled down along side of the endoscope, retrograde migration out of the small intestine proximally into the stomach or esophagus is frequently observed. We used a new device, the EndoRail, to achieve transpyloric placement of feeding tubes. Methods: The EndoRail consists of a thin plastic sheath and a plastic rail. The scope is inserted through the sheath allowing the rail to serve as an exterior channel along which accessories may be passed. We fashioned 140 cm length 10 Fr Dobb-Hoff-type feeding tubes (Viasys Healthcare Inc.) with a compatible rail to allow insertion along the EndoRail affixed to a pediatric colonoscope (Olympus PCF100). The colonoscope was inserted perorally, deeply intubated into the small intestine and the feeding tube then slid along the EndoRail to the tip of the scope. The tube is left in place by backing the scope out while advancing the tube along the rail. We compared the EndoRail technique to a standard technique of using a polypectomy snare to drag a similar tube through the pylorus and deeply into the small intestine in 4 mongrel dogs [mean weight 21.6 kg (Range 18.6 to 24.1 kg)]; 3 attempts with each technique/ dog. Two metallic clips were affixed to the prepyloric area to allow fluoroscopic monitoring. Successful placement was defined as the feeding ports remaining in the small intestine as seen on fluoroscopy. The study was powered to detect a success rate difference of 80% vs. 30%. Two tailed p-values were calculated with Fisher’s exact test. Results: In all cases deep intubation of the small intestine was achieved (130-140 cm from the incisors). Successful placement was achieved in 9/12 (75%) attempts with the EndoRail but only 1/12 (8%) attempts with the snare (p Z 0.001). Failures with the snare were all due to proximal tube migration during scope withdrawal. The 3 EndoRail failures occurred in 3 different dogs and appeared to be due to difficulty in passing the feeding tube over the rail and through the pylorus. The time to place feeding tubes with EndoRail was 5.24 G 1.17 mins (mean G SD). Conclusion: The EndoRail allows for quick and successful transpyloric placement of feeding tubes into the small intestine. Optimization of the design should further improve the success rate and ease of use.
T1335 Inter-Observer Variability with Enhanced Magnification Endoscopy of the Squamocolumnar Junction Dawn D. Ferguson, Kenneth R. DeVault, David S. Loeb, Michael B. Wallace Background: Enhanced Magnification Endoscopy (EME) is an endoscopic technique that combines the application of acetic acid to the surface of the esophagus and endoscopic imaging with a magnifying endoscope. This technique has been identified as highly accurate in predicting specialized intestinal metaplasia (SIM) in Barrett’s esophagus based on surface pattern types. Previous investigators have noted that there is high inter-observer variability in identifying pattern types that may limit the utility of this technique. We sought to reexamine inter-observer variability in a group of images obtained as part of an ongoing, prospective study. Methods: 44 patients with GERD symptoms were prospectively evaluated with EME. Pictures were taken of representative pattern types. 50 endoscopic pictures were shown in a blinded fashion to four experienced endoscopists who are participating in a prospective trial of EME. Each endoscopist was asked to identify the pattern type based on previously published criteria and a provided atlas of previously published images. The patterns were Type I) round pits, a characteristic pit pattern of orderly arranged circular dots; II) reticular, pits circular or oval and regular in arrangement; III) villous, no pits present but a fine villiform appearance, with the villi having a regular arrangement; and IV) ridged, no pits present, but a thick convoluted appearance (cerebriform) with a regular arrangement of the villi. Results: The value of Cohen’s Kappa corresponding to the agreement in pattern identification of the 4 observers on 50 endoscopic photographs is 0.41. All four observers agreed on the pattern type for 14 pictures (28%); three observers agreed 19 (38%) times; and only two observers agreed on the remaining 17 (34%). In the 33 patients where there was
www.mosby.com/gie
agreement between either 3 or 4 of the observers, the best agreement was seen in patients with predominantly type IV (figure). Conclusions: The previously published criteria for identifying SIM using EME are associated with a high level of inter-observer variability. Of the 4 described pattern types, it appears the best accuracy between observers is seen with Type IV. This inter-observer variability may limit the applicability of this type of technique in the diagnosis of SIM in endoscopically apparent Barrett’s esophagus.
T1336 Enhanced Magnification Directed Biopsies for Detection of Intestinal Metaplasia in Patients with GERD Dawn D. Ferguson, Kenneth R. DeVault, David S. Loeb, Murli Krishna, Michael B. Wallace Background: The diagnosis of Barrett’s esophagus (BE) requires histologic confirmation of specialized intestinal metaplasia (SIM) through biopsy, a technique prone to sampling error. A promising method to improve the yield of biopsy uses acetic acid with magnification endoscopy: Enhanced Magnification Endoscopy (EME). This technique identifies mucosal surface patterns and of these, pattern types III and IV have been associated with SIM. Methods: We conducted a prospective, randomized trial to evaluate EME directed biopsies and standard endoscopy with random biopsies in patients with symptoms of GERD. Patients in the standard endoscopy group with evidence of BE had 4 quadrant random biopsies taken every 2 cms. If there was no BE, 4 quadrant biopsies were taken at the SCJ. Patients in the EME group had the mucosa at the SCJ classified as Type I) round pits, Type II) reticular, III) villous or IV) ridged based on published criteria. In some cases, the visualized pattern did not fall into one of these categories and was labeled Type V. Biopsies were taken from each pattern type. Results: 93 patients enrolled (44 with EME, 49 with standard endoscopy). 36 (39%) had endoscopic evidence of BE (12 standard endoscopy, 24 EME). The yield of SIM using EME for each pattern type is shown (table). Of the 12 patients with apparent BE on standard endoscopy, random biopsies confirmed SIM on a per patient basis in 6 (50%) and in 6 (25%) for the 24 with EME directed biopsies. Overall, the prevalence of SIM was higher in the standard endoscopy group (50%) vs. EME (25%), but if pattern type III or IV were identified, EME confirmed SIM in 75% (6/8 patients). Patients without apparent BE (57) had SIM of the SCJ confirmed in 16% (6/37) with standard endoscopy and random biopsies compared with 5% (1/20) using magnification endoscopy. Conclusions: Random biopsies of endoscopically apparent BE yield SIM at slightly lower rates than targeted biopsies with EME in patients with pattern types III or IV. However, overall it appears that the yield of biopsy confirming SIM using these two techniques is similar. This calls into question the utility of this technique in reducing sampling error to identify SIM.
Volume 61, No. 5 : 2005 GASTROINTESTINAL ENDOSCOPY AB225