Translumenal Esophageal Anastomosisfor Natural Orifice Translumenal Endoscopic Surgery:

An Ex Vivo Feasibility Study

Tetsuya Ishimaru, MD, Tadashi Iwanaka, MD, PhD, Akira Hatanaka, MD,Hiroshi Kawashima, MD, and Kan Terawaki, MD, PhD

Abstract

Aim: This study aimed to develop a novel procedure for esophagoesophageal anastomosis for natural orificetranslumenal endoscopic surgery (NOTES).Materials and Methods: An ex vivo feasibility study was performed in eight porcine models. The procedure wasas follows: (1) A BraceBar� (Olympus Medical Systems Corp., Tokyo, Japan), a double T-bar suturing device,was placed endoscopically at the blind end of the upper esophagus (UE). (2) The blind end was incised, and thescope was advanced out of the esophagus. (3) A balloon catheter was inserted into the lower esophagus (LE). (4)The catheter and a thread on the BraceBar were withdrawn so that the end of the UE was inverted, and the LEwas pulled into the UE. (5) After the catheter was removed, a short tube was placed inside the duplicated part ofthe esophagus via the transgastric route. (6) A double ligature was performed using a ligating device over thetube. A liquid leak test was performed after the procedure.Results: All steps in this procedure were technically successful under the endoscopic visualization without anyassistance from outside of the esophagus. The median time of this procedure was 31 (23–66) minutes. Themedian internal pressure of the UE was 122 (82–142) mm Hg when the anastomosed esophagus was separatedinto two specimens during the leak test.Conclusions: Translumenal esophagoesophageal anastomosis was feasible. The duration of the procedure wasshort, and the anastomoses appear to have sufficient strength for use in clinical practice. An in vivo survivalstudy is needed to confirm the safety and reliability of this NOTES procedure.

Introduction

Natural orifice translumenal endoscopic surgery(NOTES), the first case of which was reported by Kalloo

et al.,1 shows promise as the ultimate version of minimallyinvasive surgery because it enables the performance of no-scar surgery. However, several challenges must be overcomebefore NOTES can be widely applied in clinical practice,2–4

and translumenal bowel anastomosis is one of the mostchallenging and imperative assignments.

We previously reported a laparoscopic gastric pull-upusing NOTES techniques and succeeded in translumenalesophagoesophageal anastomosis in acute nonsurvival ex-periments.5 However, that procedure was complicated andassociated with risk; therefore, a simpler and safer method isneeded. The aim of this study is to develop a novel procedurefor esophagoesophageal anastomosis for NOTES.

Materials and Methods

The study protocol was approved by the Animal Care andUse Committee of The University of Tokyo and Japan NOTES.An ex vivo feasibility study was performed in eight porcinemodels. Eight esophagi and whole stomachs were harvestedfrom 100-kg pigs, and each esophagus was transected at themiddle and closed at both ends with interrupted sutures. ATOP Overtube (model 16632; Top Corp., Tokyo, Japan) wasintroduced into the upper esophagus (UE) and mounted in aphantom of the human upper body. The lower esophagus(LE) and whole stomach was set in the phantom (Fig. 1).

Surgical procedure

A two-channel endoscope (model GIF-2TQ260M; OlympusMedical Systems Corp., Tokyo) was introduced into the upper

Department of Pediatric Surgery, The University of Tokyo Hospital, Tokyo, Japan,

JOURNAL OF LAPAROENDOSCOPIC & ADVANCED SURGICAL TECHNIQUESVolume 22, Number 7, 2012ª Mary Ann Liebert, Inc.DOI: 10.1089/lap.2012.0017

724

esophagus, and a BraceBar� (Olympus Medical SystemsCorp.), a prototype of the double T-bar suturing deviceoriginally developed to establish endoscopic full-thicknessclosure of large gastric perforations, was placed at the blindend. This device consisted of an applicator, which was aneedle catheter with a slidable outer sheath (Fig. 2a) and asuture unit composed of a bifurcated nylon suture with threeT-tags, two of which were fixed at the distal ends of theseparated nylon threads, with the remaining tag placed at aposition proximal to both (Fig. 2b). The suture unit wasloaded inside the applicator in advance (Fig. 3a). The appli-cator was then advanced through the scope, and all layers ofthe blind end were penetrated by the needle, after which thefirst T-tag and thread were released out of the applicatorblindly behind the intestinal wall (Fig. 3b). The applicator waswithdrawn into the lumen, leaving the T-tag and thread inplace. Next, another puncture was made on the other side ofthe blind end at a position equidistant from the center as thefirst puncture, and the second T-tag was released in the samemanner as the first. After the needle was pulled back intothe lumen, the outer sheath was advanced, and the proximaltag, which could slide over the thread, was pushed to cinchthe pair of T-tags loosely (Fig. 3c).

The esophageal wall between the T-tags was incised, andthe scope was advanced out of the esophagus, after which theLE was identified. The blind end of the LE was grasped, anda small hole was made at its center. A balloon catheter(Multi3V�; Olympus Medical Systems Corp.) was insertedinto the LE through the hole, and the balloon was expandedwith water (Fig. 3d). The scope was withdrawn back into theUE, leaving the balloon catheter in place, and the thread onthe BraceBar was grasped by a forceps. After the tip of theovertube was advanced up to 2–3 cm proximal from the blindend, the scope, together with the catheter and the forceps, waspulled so that the edge of the UE was inverted, and the LE wasdragged into the UE (Fig. 3e). The catheter was removed, anda short (approximately 3 cm) 16 French silicon tube wasplaced inside the duplicated part of the LE via the transgastricroute. Finally, a double ligature over the transanastomotictube was performed using a ligating device (Olympus Med-ical Systems Corp.) with the expectation that a wide opening

would be created after the ligated tissue became necrotic anddropped off together with the tube (Fig. 3f and g).

Liquid leak test

The following experimental system was set up to evaluatethe strength of the anastomosis. An extension tube was in-serted into the UE through the oral opening and ligatedtightly. The other end of the tube was connected to a Tru-Wave� disposable pressure transducer (Edwards LifesciencesLLC, Irvine, CA), and the internal pressure of the UE wasdisplayed in real time on a monitor. A small hole was made onthe side wall of the UE, and another extension tube, whichwas connected to a syringe, was inserted and ligated firmlywith double purse-string sutures. The LE was clampedwith two intestinal forceps just proximal to the cardia. In-digocarmin aqueous solution in a syringe was manually in-fused into the UE, and the state of the specimens and thepressure values were recorded. Leakage pressure was definedas the value at the time when the esophageal duplication wasreleased and the anastomosed esophagus was separated into

FIG. 1. Experimental setting.

FIG. 2. The BraceBar, a prototype of a double tissue an-choring device, consisting of two parts: (a) an applicator and(b) a suture unit. (a) The applicator is a needle catheter with amovable outer sheath. The needle is hidden inside the cath-eter. (b) The suture unit has a pair of T-tags fixed at the endof bifurcated threads and a slidable tag.

TRANSLUMENAL ESOPHAGEAL ANASTOMOSIS FOR NOTES 725

two specimens, or when the solution was seen leaking fromthe anastomotic part.

Results

The procedure was technically successful under endo-scopic visualization without any manipulations from out-side of the esophagus in all attempts except one, in whichthe ligating device broke down. Endoscopic views and theappearance of the exterior of the esophagus during theprocedure are shown in Figure 4. The overtube played animportant role in inverting the end of the UE, and advancingthe overtube close to the blind end was an essential process.The median time of this procedure was 31 (23–66) minutes(Table 1).

The leak test was performed in all seven specimens forwhich the procedure was completed. Table 1 shows results ofthe liquid leak test. The internal pressure of the UE was ad-equately measured in four of the seven specimens (Cases 3,4, 6, and 8), and the median value was 122 (82–142) mm Hg.In these cases, as the solution was infused, both the UE andLE were expanded, which indicated patency of the anasto-

mosis (Fig. 5). Finally, the LE was pushed out of the UE, as ifan intussusception was reduced by a high-pressure enema,and the anastomosed esophagus was separated into twospecimens. In contrast, there was no expansion of the LEduring infusion of the solution in two of the three specimensin which the leakage pressure was not measured correctly.The reason was that the transanastomotic tube slipped fromthe duplicated part when we transferred the anastomosedesophagus from the phantom of the human upper body tothe experimental setting of the leak test. As a result, thepatency of the anastomosis was lost (Cases 1 and 2). Theinternal pressure of one of those specimens reached themaximum limit (300 mm Hg) without separating theesophagus into two segments (Case 1). In the other case, theinternal pressure was 248 mm Hg when the LE was pushedout of the UE (Case 2). Leakage from the needle holes wasobserved in one (Case 7).

Discussion

Many studies to develop methods for gastrotomy or co-lotomy closure for NOTES have been performed in both ex vivo

FIG. 3. Schema of translumenal esophagoesophageal anastomosis. (a) The suture unit is loaded inside the applicator inadvance. (b) After the needle penetrates all layers of the esophageal wall, the first T-tag is released behind the wall. (c) Thesecond T-tag is placed in the same manner, and the pair of T-tags is cinched loosely. The suture unit is placed at the blind end.(d) The scope is advanced out of the upper esophagus, and a balloon catheter is inserted into the lower esophagus. (e) Anovertube is advanced close to the blind end, and the balloon catheter and the thread of the BraceBar are pulled. (f) The end ofthe upper esophagus is inverted, and the lower esophagus is pulled into the upper esophagus. A short transanastomotic tubeis inserted inside the duplicated part of the lower esophagus, and a double ligature is performed using the ligating device. (g)A wide opening is created after the ligated part drops off.

726 ISHIMARU ET AL.

and in vivo settings. However, only a few studies have aimedto establish a method for bowel anastomosis for NOTES.Rolanda et al.6 reported peroral esophageal segmentectomyand anastomosis, but a single transthoracic trocar was usedto assist the procedure. Kantsevoy et al.7 and Bergstormet al.8 described endoscopic gastrojejunostomy, but theseprocedures were performed in the stomach. The stomachprovides a large working space, which enables easy ma-nipulation of the scope and various instruments. The targetof our research is esophagoesophageal anastomosis, which

presents difficulty in handling the scope or instruments be-cause of the limited space. Our previous method achievedtranslumenal esophagoesophageal anastomosis but wascomplicated, and the blind puncture of the upper esophaguspresented risks of injury.5 Subsequently we devised a newprocedure and herein report its feasibility. The minimumduration of the procedure was 23 minutes, a time sufficientlyshort for clinical practice. Moreover, with more experienceand with more cases, the time could be shortened to less than20 minutes.

FIG. 4. Endoscopic views and outer appearances during the procedure. (a) The BraceBar is placed at the blind end of theupper esophagus. (b and c) A small hole is made at the center of the lower esophagus: (b) endoscopic view and (c) outerappearance. (d and e) The edge of the upper esophagus is inverted, and the lower esophagus is pulled into the upperesophagus: (d) endoscopic view and (e) outer appearance. (f) A short transanastomotic tube (arrow) is inserted via thetransgastric route, and a loop of the ligating device (arrowhead) is fastened over it.

TRANSLUMENAL ESOPHAGEAL ANASTOMOSIS FOR NOTES 727

We performed the liquid leak test to evaluate the strengthof the anastomosis. Hookey et al.9 reported that a liquid leakwas detected at the site of hand-sewn gastrotomy closure at apressure of 80.8 mm Hg. Sodergren et al.10 noted leakage ofwater dyed blue from a colotomy site that was closed withinterrupted sutures at a pressure of 23 mm Hg. Of course, in

these studies leak tests were not performed for the esophagus,and the procedures examined were not anastomoses. How-ever, our result showing a pressure of 122 mm Hg was muchhigher than in those studies described above and indicatesthat the anastomosis formed by our procedure was suffi-ciently strong to be used in clinical practice.

This study has several limitations. To begin with, in thisstudy the experiments were ex vivo. Considering future clin-ical application, we admit concern about how to dissect theUE and LE. Our previous report described that dissection ofthe postmediastinum could be easily performed,5 and webelieve that dissection of the LE is possible, even thoughchallenging.

We performed an in vivo experiment on a 40-kg pig toconfirm the feasibility of dissection of the UE (authors’ un-published data). First, the LE was transected through a rightthoracotomy, taking care not to injure the upper and themiddle mediastinum. Next, a BraceBar was placed at the endof the UE. Then, a thread of the BraceBar was grasped endo-scopically, and the endoscope was withdrawn. The outerappearance of the esophagus was observed under direct vi-sion through the thoracotomy. As a result, the edge of theesophagus was inverted, and the fold was moved fromthe lower mediastinum through the upper mediastinum to theoutside of the thorax at the end of the procedure. This meantthat dissection of the UE, including the area between thetrachea and the esophagus, was feasible just by traction of theBraceBar in the porcine model.

In addition, we have yet to confirm that the tissue ligated bythe ligating device becomes necrotic and drops off together withthe short transanastomotic tube. If these phenomena do not oc-cur and only the tube drops off, leaving behind living ligatedtissue, our procedure may result in a kind of papillary formation.That this could be a concern was indicated in Cases 1 and 2,where displacement of the tube caused loss of anastomotic pa-tency. Animal survival experiments are needed in future.

Our long-term aim is to use this technique in neonates orinfants with esophageal atresia. However, the endoscope andinstruments used in the current study were those developedfor adults and thus were too large to be used in infants.Therefore, in this study we had to use specimens larger thanthe esophagus of neonates or infants.

Despite these limitations, our results showed the feasibilityof translumenal esophagoesophageal anastomosis. Moreover,the current procedure was simpler and safer than our previ-ous one.5 This method might after further study becomeavailable in general surgery such as colorectal resection andcould contribute to widening the indication for NOTES in thefuture. Although a transgastric maneuver to insert the tubewas used in this study, our final goal is to accomplish theanastomosis only by the oral approach. An improved proce-dure is under development.

Acknowledgments

This study was supported by grant 23592626 from theMinistry of Education, Culture, Sports, Science and Technol-ogy of Japan.

Disclosure Statement

No competing financial interests exist.

Table 1. Procedure Time and Leakage Pressure

CaseProcedure

time (minutes)Leakage

pressure (mm Hg)

1 65 (300)a

2 25 (248)a

3 66 1424 31 1085 Not completed6 47 1357 23 —b

8 23 82Median (range) 31 (23–66) 122 (82–142)

aDislocation of the transanastomotic tube occurred, and leakagepressure could not be measured correctly. Pressure values are shownin parentheses as reference.

bLeakage from needle holes was observed.

FIG. 5. State of the specimen during the liquid leak test. (a)State before infusion; the upper esophagus is placed at theleft side. (b) State during infusion; both the upper and thelower esophagus are expanded.

728 ISHIMARU ET AL.

References

1. Kalloo AN, Singh VK, Jagannath SB, et al. Flexible trans-gastric peritoneoscopy: A novel approach to diagnostic andtherapeutic interventions in the peritoneal cavity. Gastro-intest Endosc 2004;60:114–117.

2. Rattner DW, Hawes R, Schwaitzberg S, et al. The SecondSAGES/ASGE White Paper on natural orifice transluminalendoscopic surgery: 5 years of progress. Surg Endosc 2011;25:2441–2448.

3. Rattner D, Kalloo A. ASGE/SAGES Working Group onNatural Orifice Translumenal Endoscopic Surgery. October2005. Surg Endosc 2006;20:329–333.

4. ASGE, SAGES. ASGE/SAGES Working Group on NaturalOrifice Translumenal Endoscopic Surgery White PaperOctober 2005. Gastrointest Endosc 2006;63:199–203.

5. Ishimaru T, Iwanaka T, Kawashima H, et al. A pilot study oflaparoscopic gastric pull-up by using the natural orificetranslumenal endoscopic surgery technique: A novel pro-cedure for treating long-gap esophageal atresia (type a).J Laparoendosc Adv Surg Tech A 2011;21:851–857.

6. Rolanda C, Silva D, Branco C, et al. Peroral esophageal seg-mentectomy and anastomosis with single transthoracic trocar:A step forward in thoracic NOTES. Endoscopy 2011;43:14–20.

7. Kantsevoy SV, Jagannath SB, Niiyama H, et al. Endoscopicgastrojejunostomy with survival in a porcine model. Gas-trointest Endosc 2005;62:287–292.

8. Bergstrom M, Ikeda K, Swain P, et al. Transgastric anasto-mosis by using flexible endoscopy in a porcine model (withvideo). Gastrointest Endosc 2006;63:307–312.

9. Hookey LC, Khokhotva V, Bielawska B, et al. The Queen’sclosure: A novel technique for closure of endoscopic gas-trotomy for natural-orifice transluminal endoscopic surgery.Endoscopy 2009;41:149–153.

10. Sodergren M, Clark J, Beardsley J, et al. A novel flexibleendoluminal stapling device for use in NOTES colotomyclosure: A feasibility study using an ex vivo porcine model.Surg Endosc 2011;25:3266–3272.

Address correspondence to:Tetsuya Ishimaru, MD

Department of Pediatric SurgeryThe University of Tokyo Hospital

7-3-1 Hongo, Bunkyo, Tokyo 113-8655Japan

E-mail: i-tetsuya@umin.ac.jp

TRANSLUMENAL ESOPHAGEAL ANASTOMOSIS FOR NOTES 729