- Email: [email protected]
Self-approximating transluminal access technique for natural orifice transluminal endoscopic surgery: a porcine survival study (with video)
ORIGINAL ARTICLE: Experimental Endoscopy
Self-approximating transluminal access technique for natural orifice transluminal endoscopic surgery: a porcine survival study (with video) Eric M. Pauli, MD, Matthew T. Moyer, MD, Randy S. Haluck, MD, FACS, Abraham Mathew, MD Hershey, Pennsylvania, USA
Background: The ability to access the abdominal cavity though a direct (modified-PEG type) gastric incision to perform natural orifice transluminal endoscopic surgery (NOTES) has been demonstrated in the literature. However, the optimal technique to access the abdomen remains unknown. Objective: The aim of this study was to evaluate the safety and feasibility of a transgastric approach to the abdominal cavity through an extended submucosal tunnel. Design: Animal feasibility study. Interventions: Transgastric endoscopic peritoneoscopy was performed in 7 anesthetized swine, including 2 acute and 5 survival animals. After the creation of a 10-cm to 12-cm tunnel in the gastric submucosal plane, the peritoneal cavity was accessed by needle-knife puncture through the gastric wall near the greater curvature. The peritoneal cavity was examined before the gastric mucosal incision was closed with endoclips. Survival animals were euthanized two weeks after the procedure, and a necropsy was performed. Results: The abdominal cavity was successfully entered without complication in all 7 animals. The mucosal incisions were able to be closed by endoscopy. In the survival experiments, all animals recovered and gained weight. Two animals experienced clinically unapparent infectious complications. Limitations: Small sample size. Conclusions: A peroral transgastric approach to the abdominal cavity through an extended submucosal tunnel is technically feasible and allows safe abdominal access and reliable closure with currently available technology. It has potential benefits as an alternative to direct transgastric access for NOTES procedures. (Gastrointest Endosc 2008;67:690-7.)
Published reports of animal models demonstrate the feasibility of a transgastric approach to the abdominal cavity for natural orifice transluminal endoscopic surgery (NOTES) procedures. The preliminary work by Kalloo et al1 validated the concept of endoscopic peritoneoscopy. Subsequently, other successful transgastric procedures have been reported, all using a porcine model. These include liver biopsy, tubal ligation, gastrojejunostomy,
Copyright ª 2008 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2007.09.023
cholecystectomy, splenectomy, oophorectomy, and partial hysterectomy.2-13 Despite early success, it is recognized that concerns of safety and reliability remain before NOTES should be introduced into clinical practice.14-18 For transgastric interventions, many of the significant difficulties relate to the formation of the gastrotomy. The optimal method of achieving transgastric access to both avoid iatrogenic injury and allow reliable closure has yet to be delineated.14,17-19 To date, the majority of investigators accessed the abdomen through a direct puncture gastrotomy, typically on the anterior portion of the stomach.1-13,19 This technique requires full-thickness closure of the gastrotomy site. Recently, we, and other investigators, reported on the technical ability to endoscopically dissect within the submucosal plane and to use these dissection techniques to achieve transluminal
690 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
Abbreviations: IM, intramuscular; NOTES, natural orifice transluminal endoscopic surgery; NSS, normal saline solution; O2, oxygen; STAT, self-approximating transluminal access technique.
Pauli et al
access.20-23 This approach spatially separates the mucosal and seromuscular portions of the enterotomy, potentially allowing for a more simplified gastrotomy closure. In this study, we investigated the technical feasibility of obtaining transgastric abdominal access by using our previously described self-approximating transluminal access technique (STAT)20 in an in vivo porcine model with longterm survival. We hypothesized that animals would survive for two weeks without complications if the STAT would be used for abdominal access and mucosal closure. Our results demonstrated this technique to be a safe means of abdominal access and a reliable means of gastrotomy closure.
Self-approximating transluminal access technique
Capsule Summary What is already known on this topic d
Transgastric procedures, including liver biopsy, gastrojejunostomy, cholecystectomy, and splenectomy, have been successful in a porcine model, but formation of the gastrotomy can cause significant difficulties.
What this study adds to our knowledge d
A peroral transgastric approach to the abdominal cavity through an extended submucosal tunnel was technically feasible in 7 pigs and allowed safe abdominal access and reliable closure.
MATERIALS AND METHODS Animals
Operative procedure
Seven female domestic farm swine (Sus scrofus domesticus) that weighed 22 to 30 kg were used in this study (Reinecker Farms, York Springs, Pa). All animals were housed at the animal resource facility and were cared for by the veterinary staff of the Penn State College of Medicine Department of Comparative Medicine, Hershey, Pennsylvania. The study was conducted in accordance with the U.S. Department of Agriculture Animal Welfare Act. The protocol was approved by the animal care and use committee at the Penn State College of Medicine. Survival experiments were performed after a series of acute experiments.
Animals were fed standard chow during their required quarantine period. Beginning 24 hours before the procedure, the animals were kept without food (access to water ad libitum). In the final 3 animals of the series, a liquid diet (Ensure; Abbott Laboratories, Abbott Park, Ill) was begun 72 hours before the procedure. For survival experiments, one gram of IV cefazolin and 600,000 units of intramuscular (IM) penicillin G benzathine plus penicillin G procainebased antibiotic were given immediately before the start of the procedure.
General anesthesia was induced by an IM injection of Telazol (500 mg/kg; Fort Dodge Animal Health, Fort Dodge, Iowa), medetomidine (70-80 mg/kg), and butorphanol (300 mg/kg). After endotracheal intubation, anesthesia was maintained with 1% to 2% isoflurane delivered in 100% oxygen (O2), with mechanical ventilation through a semi-closed circuit. The end-tidal carbon dioxide level, O2 saturation, respiratory rate, and pulse were monitored and recorded throughout the procedure. With the animal in a supine position, a gastric-length overtube (Gardus; US Endoscopy, Mentor, Ohio) was advanced into the esophagus with an adult single-channel gastroscope inside. The endoscope was used to lavage the stomach with sterile normal saline solution (NSS) until the stomach was free of gross food material. After evacuation of the NSS irrigant, 900 mL of an antibiotic solution (40 mg neomycin in one liter NSS) was instilled for 10 minutes before being aspirated endoscopically. The remaining 100 mL of antibiotic solution was held in reserve. By using an injection needle (InjectaFlow; Cook Endoscopy, Winston-Salem, NC), 5 to 10 mL NSS was introduced submucosally to elevate the mucosa from the underlying muscle (Fig. 1A). This mucosal pillow was created at a technically convenient position on the posterior wall in the cardiac portion of the stomach, approximately 2 cm distal to the pars esophagea tunicae mucosae. Next, a 1-cm to 1.5-cm incision was created on the pillow with a 4-mm needle-knife (Huibregtse; Cook) (Fig. 1B). A rat-tooth grasping forceps (FG-47L; Olympus Ltd, Tokyo, Japan) was used to begin the dissection of the loose areolar tissue of the submucosal plane, separating the mucosa from the adjacent circular muscle layer. Once a sufficient area had been dissected, the endoscope was advanced into the submucosal plane. By using a combination of dissection with the forceps, as well as blunt dissection with the leading edge of the endoscope, a 10-cm to 12-cm submucosal tunnel was created, directed toward the greater curvature of the stomach (Fig. 1C). The tunnel length was determined by observing the distance
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 691
Endoscopes and instruments An adult single-channel gastroduodenoscope (EG 2940; Pentax America Inc, Mountvale, NJ) was used for the procedure. For the survival experiments, the endoscope and all reusable, nonautoclavable endoscopic equipment (needle-knives, esophageal overtubes, and injection needles) were cleansed and chemically sterilized with 2.4% glutaraldehyde (Cidex; Johnson and Johnson, New Brunswick, NJ) and were air-dried before the start of the procedure. All reusable equipment capable of undergoing steam sterilization (endoscopic forceps and clip appliers) was autoclaved. Endoclips were supplied by Microvasive Endoscopy, Boston Scientific Corp, Natick, Mass, and Olympus America Inc, Center Valley, Pa.
Preoperative care
Self-approximating transluminal access technique
Pauli et al
Figure 2. Arterial-venous pair identified within the submucosa.
to 1.5-cm linear incision in the seromuscular layer of the stomach at the distal end of the submucosal tract. This completed the gastrotomy and allowed the endoscope to advance into the abdomen (Fig. 1D). On one occasion (with animal no. 5), a 15-mm controlled radial expansion balloon (CRE; Microvasive) was used to dilate the distal tract to allow easier passage of the endoscope into the abdomen. Pneumoperitoneum was established, with the judicious use of the endoscope air pump. A brief peritoneoscopy was performed by using standard endoscopic techniques to assess for evidence of iatrogenic injury. When necessary, the omentum was incised with the needle-knife. At the conclusion of our abdominal exploration, the pneumoperitoneum was evacuated by using the suction channel of the endoscope. The endoscope was withdrawn from the submucosal tract. The seromuscular incision was not closed. The gastric mucosal incision was closed with an endoscopic clip fixing device (HX-5QR-1 and Quick Clip [Olympus] or Resolution [Microvasive]) (Fig. 1E). The endoscope and overtube were removed, and the animal was awakened from anesthesia. The accompanying video demonstrates this technique (Video 1, available online at www.giejournal.org). Figure 1. The transgastric STAT. A, Mucosal pillow creation with saline solution injection. B, Needle-knife incision to enter the submucosa. C, Controlled submucosal tunneling with a forceps. D, Seromuscular incision to complete the transgastric access. E, Endoclip closure of the mucosal incision.
Postoperative care
markers as the endoscope was withdrawn from the tract. As perforating blood vessels to the mucosa were identified, dissection was carried out around the vasculature (Fig. 2). Once the desired length and direction of the tunnel was achieved, the remaining 100 mL of reserved antibiotic solution was used to irrigate the tract before entering the abdomen. The needle-knife was then used to make a 1-cm
Nonsurvival animals were euthanized after completing the operative procedure. For the survival animals, a regular diet was restarted the morning after the procedure. The animals had immediate free access to water and were monitored twice daily by the veterinary staff for evidence of altered behavior, clinical distress, or appetite changes. Routine postoperative analgesics were not administered. Survival animals were weighed weekly during their twoweek survival period to assess for appropriate weight gain. All animals were euthanized with an overdose of pentobarbital sodium (O100 mg/kg IV), after the induction of anesthesia.
692 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
Pauli et al
Self-approximating transluminal access technique
TABLE 1. Results of survival experiments Animal
1
2
3
4
5
Time to complete gastrotomy (min)
22
26
26
20
42
Good
Good
Good
Good
Moderate
5.9
4.5
5.4
2.7
3.6
7
6
3.5
6.5
5
Well healed
Well healed
Well healed
Well healed
Well healed
Normal
Normal
Left upper quadrant adhesion
Normal
Fibrinous adhesions
2/2
3/3
3/2
4/4
3/3
None
None
Microabscess
None
Submucosal abscess
Food intake Weight gain (kg) Tunnel length at necropsy (cm) Gastrotomy Abdominal cavity Clips placed/retained Complications
Necropsy Immediately after euthanasia, a necropsy was performed on all animals in the presence of a veterinary physician. The peritoneal cavity was evaluated for evidence of abscess formation, adhesive disease, and unrecognized iatrogenic injuries. The entire length of the submucosal tract (from the mucosal clips to the seromuscular incision site) was assessed by gross examination and submitted, in 10% formalin solution, for microscopic histopathologic examination. The specimens were later imbedded in paraffin by using a tissue processor (Tissue-Tek VIP5; Sakura Finetek, Torrance, Calif), sectioned on a manual microtome, and stained with Harris-modified hematoxylin solution and eosin. All slides were reviewed by a veterinary pathologist.
of iatrogenic injury to the abdominal wall or to nearby organs. There was no evidence of bleeding within the submucosal tunnel or from the seromuscular incision itself. When immersed in water, neither of the two esophagogastric organ specimens demonstrated evidence of air leak.
Survival experiments
The submucosal dissection and gastrotomy were performed without complication in two animals. We were able to control the submucosal dissection to guide the tunnel direction and to avoid disruption of perforating mucosal blood vessels. The average time to complete the procedure and to enter the abdominal cavity was 13.5 minutes. By using standard endoscopic techniques (eg, retroflexing and torquing), we were able to explore the abdomen. The gastric mucosal incisions were closed without difficulty. At the necropsy, there was no evidence
The STAT was performed without intraoperative complication in 5 animals. The results of these experiments are summarized in Table 1. In all animals, the tunnel began at a convenient location on the posterior gastric wall. All tunnels were deliberately directed inferiorly toward the greater curvature, and all were measured to be 10 to 12 cm at the time of the operation. We were again able to control the direction of the submucosal dissection and to navigate around perforating blood vessels. The seromuscular incision site was consistently made on the posterior gastric wall near the leaves of the greater omentum. The mean standard deviation (SD) procedure time (from submucosal injection to entry of the endoscope into the abdominal cavity) was 27.2 8.7 minutes. Once inside the abdomen, we undertook peritoneoscopy. Pneumoperitoneum was successfully created by using the endoscope air pump and suction. No untoward physiologic effects from the unmonitored pneumoperitoneum were noted. During endoscopic exploration, we were able to identify the stomach, small bowel, colon, spleen, liver, gall bladder, pancreas, urinary bladder, and uterine horns. The movement of the endoscope within the abdomen was not perceived to be limited by the tunnel. No iatrogenic injuries were suspected at the time of the gastrotomy. Clip closure of the mucosal incision was accomplished with 2 to 4 endoclips (mean 3). There was minimal change in the mucosal incision site (edema or tearing) at the conclusion of the procedure, despite manipulation of the endoscope within the abdomen. During the survival period, all the animals ate freely. The mean (SD) weight gain was 4.42 1.30 kg (18.8% 6.82% increase from baseline weight). None of the animals required analgesics in the postoperative period.
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 693
Data collection The procedures were digitally recorded by using video software (Pinnacle Studio Plus 700 USB, Version 10; Pinnacle Systems, Mountain View, Calif) on a laptop computer (Dell Latitude D820; Dell Inc, Round Rock, Tex). Significant portions of the procedures were also recorded by using a Sony video recorder (Sony Digital Handycam, DCR-VX2000; Sony Corp, Tokyo, Japan). Video editing was performed with Premiere Pro, Version 1.5 (Adobe Systems Inc, San Jose, Calif).
RESULTS Acute experiments
Self-approximating transluminal access technique
Pauli et al
Figure 3. Necropsy images. A, Mucosal incision site with endoclips in place. B, Seromuscular incision, demonstrating healing tissue, without evidence of mucosal necrosis.
At the necropsy, we noted a high rate of endoclip retention (93.3%) (Fig. 3A). All of the mucosal and seromuscular incisions were well healed (Fig. 3B). The mean tunnel length (measured from the endoclips to the seromuscular incision) was 5.6 cm. Two of the animals had mild intra-abdominal adhesions. None of the animals had evidence of iatrogenic injury or of intra-abdominal abscess formation. The final animal in the series (animal no. 5) had a 2 2-cm submucosal abscess noted in the distal portion of the STAT tunnel, as identified at the necropsy (Fig. 4). This abscess did not communicate with either the abdominal cavity or the mucosal incision. This process was subclinical in nature, because the animal thrived and gained weight after the operation. Microscopic tissue examination of the submucosal tracts revealed the following: (1) a gastrotomy site and a submucosal tract with inflammatory changes (Fig. 5), (2) fibrovascular tissue formation within the submucosa, and (3) submucosal microabscess (animal no. 3 only).
Figure 4. Luminal view of the intragastric abscess noted at the necropsy in animal no. 5. The endoclips mark the mucosal incision of the STAT tunnel. The abscess was at the distal portion of the tunnel.
The optimal location and means for obtaining transgastric access for NOTES has yet to be determined. Here we report a submucosal dissection technique performed under direct endoscopic visualization. The method combines standard endoscopic and surgical techniques to produce a transgastric access method that is easily reproducible and that has a small technical learning curve. The application of the STAT to an in vivo model allowed us to reliably and safely access the peritoneal cavity, which suggests that this methodology of achieving transgastric abdominal access may prove to be an enabling technique for further NOTES research. Observations made during this study led us to believe that the STAT may have several advantages over other proposed transgastric access techniques: 1. We were able to withdraw from and reenter the STAT tunnel, without difficulty in identifying or entering
the tract. Once the dissection began, the initial mucosal incision tended to stay tented up. Also, from within the gastric lumen, the full length and direction of the tract was visible (Video 1). This made finding and reentering the mucosal incision technically easy, because the tunnel could be followed back to the initial mucosal incision. A reassessment of the tunnel did not require the placement of a transgastric or transabdominal guidewire, although the technique does not exclude the use of one. 2. By performing the submucosal dissection under direct visualization with a grasping forceps, we were able to finely control the direction and width of the tunnel. This transgastric dissection technique differs from that recently reported by Sumiyama et al.22,23 Unlike high-pressure gas or an endoscopic balloon dissection, the STAT allowed us to visually identify and to intentionally dissect around perforating blood vessels without causing bleeding. We believe that this also reduced the risk of unintended muscular or mucosal perforation seen with a needle-knife, carbon dioxide, or monofilament snare dissection.23-25 By leaving the vasculature intact, we not only had an unobscured view of the tissue plane, but we also preserved the blood supply to the mucosa. We did not observe the partial
694 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
DISCUSSION
Pauli et al
Self-approximating transluminal access technique
Figure 5. Photomicrographs. A, Mucosal incision site (top), with pyogranulomatous submucosal inflammation and intact musculature (bottom) (H&E, orig. mag. 100). B, Intact mucosa and muscularis mucosa (top), focal inflammation, and fibrosis within the submucosa (including giant-cell foreignbody-type reaction) and the seromuscular defect (bottom) (H&E, orig. mag. 100).
necrosis of the mucosa that other investigators reported after submucosal dissection.22 The STAT tunnel also may provide a more ideal endoscope/instrument guide than the giant (8 cm) bleb created via submucosal gas insufflation.22,23 3. Despite the use of aggressive maneuvers during the peritoneoscopy (including torque and angulation of the endoscope), the mucosal incision remained at its original 1-cm to 1.5-cm size and did not develop significant edema or tearing. This allowed for endoclip closure with a minimal number of clips. Although the brevity of our peritoneoscopy may have provided a limited opportunity for the mucosal incision to lengthen, there are several other reasons why our technique may not have resulted in mucosal tearing. These include the following: appropriate selection of the mucosal entry and exit sites, snugness of the tunnel around the endoscope, and the overall length of the tunnel. 4. After withdrawal of the endoscope, the long submucosal tunnel naturally collapsed or self approximated. This was evidenced by the fact that immediately
explanted specimens remained airtight under pressure. We believe it is possible for this method to further seal the gastrotomy (by collapse of the tunnel) under conditions of high intra-abdominal or intragastric pressure, such as coughing or vomiting (Fig. 6). 5. With the endoscope in the submucosal plane, it was possible to visualize the intra-abdominal organs through the yet uncut seromuscular layer (see Video 1). This allowed us to select the exit site from the tunnel in a location where potential iatrogenic injuries could be minimized. Despite these advantages, our technique requires further refinement. We created 10-cm to 12-cm tunnels at the time of the operation. These distances were measured under partial gastric distention with the tissue stretched. At the necropsy, the length of the tract in the decompressed stomach averaged only 6 cm. We do not yet know the optimal length for the STAT, but the tunnel length can be varied without difficulty. Our abdominalaccess technique was also more time consuming than other investigators reported in the literature.19,22
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 695
Self-approximating transluminal access technique
Pauli et al
In conclusion, in this porcine survival model, we were able to successfully enter the abdominal cavity by using the STAT method. Transgastric access was safe and reproducible. The separation of the mucosal and seromuscular incisions proved a reliable means to close the gastrotomy. The finding of a submucosal abscess in one animal suggests that further investigation is warranted. We remain optimistic that the STAT may be an enabling technique to allow transgastric abdominal access for NOTES procedures. Figure 6. Schematic representation of how tunneled gastrotomy seal improves under conditions of high intraluminal or intra-abdominal pressure. Pressure (arrows) leads to increased apposition of the seromuscular and mucosal layers and prevents gastrotomy leak.
However, the potential for a simplified, reliable closure method may ultimately yield a time savings. Although we used high-level instrument disinfection, systemic preoperative antibiotics, and antibacterial gastric lavage, 2 of our 5 survival animals (40%) developed infection complications. One of our animals developed an abscess in the distal portion of the STAT tunnel (animal no. 5), and another animal had a microabscess on pathologic examination (animal no. 3). It should be noted, however, that none of our animals developed intra-abdominal sepsis from a gastrotomy leak or demonstrated signs of infection. Clinically, these animals behaved normally and gained weight appropriately. Microabscess formation was noted on histopathology, even with nontunneled transluminal access.1,26 At the necropsy, we found no gross evidence of intra-abdominal abscess formation. These collections were well circumscribed and did not communicate with either the seromuscular or the mucosal incisions. The remainder of these tracts were well healed. Although we did not perform repeat endoscopy on our animals, the abscess in animal no. 5 would have been identifiable and treatable with currently available endoscopic techniques.27,28 Nevertheless, infection issues are an obvious concern with transluminal-access techniques. Tunneling beneath the mucosal turns creates a void. Subsequent closure of bacteria into this space provides an excellent environment for abscess formation. In recognizing this fact, we flushed the tract with antibiotic irrigation before entering the abdomen. This method may be improved upon in a number of ways, including altering the antibiotics used in the irrigant, irrigating the tunnel at the end of the procedure rather than at the beginning or irrigating with dilute iodine solution rather than antibiotic solution. Providing routine postoperative antibiotics may also be of benefit in procedures that use a STAT gastrotomy. It is clear that further research is required to define the optimum protocol for tunneled transgastric access that provides a reliable closure method and minimizes the risk of infectious complications during the postoperative period. 696 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
ACKNOWLEDGMENT We thank the dedicated staff of the Department of Comparative Medicine (Drs Wilson, Carney, and Griffith, and Ms Joy Ellwanger) and Ms Pat Grosh for their help with this project.
DISCLOSURE The authors report that there are no disclosures relevant to this publication. Microvasive Endoscopy, Boston Scientific, Inc, and Olympus Medical, Inc provided endoscopic clipping devices. Pentax Medical, Inc provided endoscopy equipment.
REFERENCES 1. Kalloo AN, Singh VK, Jagannath SB, et al. Flexible transgastric peritoneoscopy: a novel approach to diagnostic and therapeutic interventions. Gastrointest Endosc 2004;60:114-7. 2. Jagannath SB, Kantsevoy SV, Vaughn CA, et al. Peroral transgastric endoscopic ligation of fallopian tubes with long-term survival in a porcine model. Gastrointest Endosc 2005;61:49-53. 3. Kantsevoy SV, Jagannath SB, Niiyama H, et al. Endoscopic gastrojejunostomy with survival in a porcine model. Gastrointest Endosc 2005;62:278-92. 4. Kantsevoy SV, Niiyama H, Jagannath SB, et al. The endoscopic transilluminator: an endoscopic device for identification of the proximal jejunum for transgastric endoscopic gastrojejunostomy. Gastrointest Endosc 2006;63:1055-8. 5. Bergstrom M, Ikeda K, Swain P, et al. Transgastric anastomosis using flexible endoscopy in a porcine model (with video). Gastrointest Endosc 2006;63:307-12. 6. Park PO, Bergstrom M, Ikeda K, et al. Experimental studies of transgastric gallbladder surgery: cholecystectomy and cholecystogastric anastomosis. Gastrointest Endosc 2005;61:601-6. 7. Rolanda C, Lima E, Pego JM, et al. Third-generation cholecystectomy by natural orifices: transgastric and transvesicular combined approach (with video). Gastrointest Endosc 2007;65:111-7. 8. Swanstrom LL, Kozarek R, Pasricha PJ, et al. Development of a new access device for transgastric surgery. J Gastrointest Surg 2005;9: 1129-36. 9. Kantsevoy SV, Hu B, Jagannath SB, et al. Transgastric endoscopic splenectomy: is it possible? Surg Endosc 2006;20:522-5. 10. Wagh MS, Merrifield BF, Thompson CC. Survival studies after endoscopic transgastric oophorectomy and tubectomy in a porcine model. Gastrointest Endosc 2006;63:473-8.
www.giejournal.org
Pauli et al
Self-approximating transluminal access technique
11. Wagh MS, Merrifield BF, Thompson CC. Endoscopic transgastric abdominal exploration and organ resection: initial experience in a porcine model. Clin Gastroenterol Hepatol 2005;3:892-6. 12. Merrifield BF, Wagh MS, Thompson CC. Peroral transgastric organ resection: a feasibility study in animals. Gastrointest Endosc 2006;63: 693-7. 13. Sumiyama K, Gostout CJ, Rajan E, et al. Pilot study of the porcine uterine horn as an in vivo appendicitis model for development of endoscopic transgastric appendectomy. Gastrointest Endosc 2006;64:808-12. 14. American Society for Gastrointestinal Endoscopy/SAGES Working Group on Natural Orifice Translumenal Endoscopic Surgery white paper October 2005. Gastrointest Endosc 2006;63:199-204. 15. Hochberger J, Lamade W. Transgastric surgery in the abdomen: the dawn of a new era? Gastrointest Endosc 2005;62:293-6. 16. Lamade W, Hochberger J. Transgastric surgery: avoiding pitfalls in the development of a new technique. Gastrointest Endosc 2006;63: 698-700. 17. Ponsky JL. Gastroenterologists as surgeons: what they need to know. Gastrointest Endosc 2005;61:454. 18. Sclabas GM, Swain P, Swanstrom LL. Endoluminal methods for gastrotomy closure in natural orifice transenteric surgery (NOTES). Surg Innov 2006;13:23-30. 19. Kantsevoy SV, Jagannath SB, Niiyama H, et al. A novel safe approach to the peritoneal cavity for per-oral transgastric endoscopic procedures. Gastrointest Endosc 2007;65:497-500. 20. Moyer MT, Pauli EM, Haluck RS, et al. Self-approximating translumenal access technique (STAT) for potential use in NOTES: an ex vivo porcine model (with video). Gastrointest Endosc 2007;66:974-8. 21. Sumiyama K, Gostout CJ, Rajan E, et al. Transesophageal mediastinoscopy by submucosal endoscopy with mucosal flap safety valve technique. Gastrointest Endosc 2007;65:679-83.
22. Sumiyama K, Gostout CJ, Rajan E, et al. Submucosal endoscopy with mucosal flap safety valve. Gastrointest Endosc 2007;65:688-94. 23. Sumiyama K, Gostout CJ, Rajan E, et al. Transgastric cholecystectomy: transgastric accessibility to the gallbladder improved with the SEMF method and a novel multibending therapeutic endoscope. Gastrointest Endosc 2007;65:1028-34. 24. Ro¨sch T, Sarbia M, Schumacher B, et al. Attempted endoscopic en bloc resection of mucosal and submucosal tumors using insulatedtip knives; a pilot series (including videos). Endoscopy 2004;36: 788-801. 25. Kamler JP, Borsatto R, Binmoeller KF. Circumferential endoscopic mucosal resection in the swine esophagus assisted by a cap attachment. Gastrointest Endosc 2002;55:923-8. 26. Pai RD, Fong DG, Bundga ME, et al. Transcolonic endoscopic cholecystectomy: a NOTES survival study in a porcine model (with video). Gastrointest Endosc 2006;64:428-34. 27. Ruiz-Rebollo ML, Atienza-Sa´nchez R, Go´mez-Corral J. Gastric wall abscess caused by an ingested toothpick. Gastrointest Endosc 2007;65:518-9. 28. Choong NWW, Levy MJ, Rajan E, et al. Intramural gastric abscess: case history and review. Gastrointest Endosc 2003;58:627-9.
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 697
Received May 24, 2007. Accepted September 4, 2007. Current affiliations: Division of Minimally Invasive and Bariatric Surgery (E.M.P., R.S.H.), Division of Gastroenterology and Hepatology (M.T.M., A.M.), The Milton S. Hershey Medical Center, Penn State College of Medicine, Hershey, Pennsylvania, USA. Reprint requests: Eric M. Pauli, MD, Department of Surgery, PO Box 850 MC H149, Penn State College of Medicine, Hershey, PA 17033.
Self-approximating transluminal access technique for natural orifice transluminal endoscopic surgery: a porcine survival study (with video) Eric M. Pauli, MD, Matthew T. Moyer, MD, Randy S. Haluck, MD, FACS, Abraham Mathew, MD Hershey, Pennsylvania, USA
Background: The ability to access the abdominal cavity though a direct (modified-PEG type) gastric incision to perform natural orifice transluminal endoscopic surgery (NOTES) has been demonstrated in the literature. However, the optimal technique to access the abdomen remains unknown. Objective: The aim of this study was to evaluate the safety and feasibility of a transgastric approach to the abdominal cavity through an extended submucosal tunnel. Design: Animal feasibility study. Interventions: Transgastric endoscopic peritoneoscopy was performed in 7 anesthetized swine, including 2 acute and 5 survival animals. After the creation of a 10-cm to 12-cm tunnel in the gastric submucosal plane, the peritoneal cavity was accessed by needle-knife puncture through the gastric wall near the greater curvature. The peritoneal cavity was examined before the gastric mucosal incision was closed with endoclips. Survival animals were euthanized two weeks after the procedure, and a necropsy was performed. Results: The abdominal cavity was successfully entered without complication in all 7 animals. The mucosal incisions were able to be closed by endoscopy. In the survival experiments, all animals recovered and gained weight. Two animals experienced clinically unapparent infectious complications. Limitations: Small sample size. Conclusions: A peroral transgastric approach to the abdominal cavity through an extended submucosal tunnel is technically feasible and allows safe abdominal access and reliable closure with currently available technology. It has potential benefits as an alternative to direct transgastric access for NOTES procedures. (Gastrointest Endosc 2008;67:690-7.)
Published reports of animal models demonstrate the feasibility of a transgastric approach to the abdominal cavity for natural orifice transluminal endoscopic surgery (NOTES) procedures. The preliminary work by Kalloo et al1 validated the concept of endoscopic peritoneoscopy. Subsequently, other successful transgastric procedures have been reported, all using a porcine model. These include liver biopsy, tubal ligation, gastrojejunostomy,
Copyright ª 2008 by the American Society for Gastrointestinal Endoscopy 0016-5107/$32.00 doi:10.1016/j.gie.2007.09.023
cholecystectomy, splenectomy, oophorectomy, and partial hysterectomy.2-13 Despite early success, it is recognized that concerns of safety and reliability remain before NOTES should be introduced into clinical practice.14-18 For transgastric interventions, many of the significant difficulties relate to the formation of the gastrotomy. The optimal method of achieving transgastric access to both avoid iatrogenic injury and allow reliable closure has yet to be delineated.14,17-19 To date, the majority of investigators accessed the abdomen through a direct puncture gastrotomy, typically on the anterior portion of the stomach.1-13,19 This technique requires full-thickness closure of the gastrotomy site. Recently, we, and other investigators, reported on the technical ability to endoscopically dissect within the submucosal plane and to use these dissection techniques to achieve transluminal
690 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
Abbreviations: IM, intramuscular; NOTES, natural orifice transluminal endoscopic surgery; NSS, normal saline solution; O2, oxygen; STAT, self-approximating transluminal access technique.
Pauli et al
access.20-23 This approach spatially separates the mucosal and seromuscular portions of the enterotomy, potentially allowing for a more simplified gastrotomy closure. In this study, we investigated the technical feasibility of obtaining transgastric abdominal access by using our previously described self-approximating transluminal access technique (STAT)20 in an in vivo porcine model with longterm survival. We hypothesized that animals would survive for two weeks without complications if the STAT would be used for abdominal access and mucosal closure. Our results demonstrated this technique to be a safe means of abdominal access and a reliable means of gastrotomy closure.
Self-approximating transluminal access technique
Capsule Summary What is already known on this topic d
Transgastric procedures, including liver biopsy, gastrojejunostomy, cholecystectomy, and splenectomy, have been successful in a porcine model, but formation of the gastrotomy can cause significant difficulties.
What this study adds to our knowledge d
A peroral transgastric approach to the abdominal cavity through an extended submucosal tunnel was technically feasible in 7 pigs and allowed safe abdominal access and reliable closure.
MATERIALS AND METHODS Animals
Operative procedure
Seven female domestic farm swine (Sus scrofus domesticus) that weighed 22 to 30 kg were used in this study (Reinecker Farms, York Springs, Pa). All animals were housed at the animal resource facility and were cared for by the veterinary staff of the Penn State College of Medicine Department of Comparative Medicine, Hershey, Pennsylvania. The study was conducted in accordance with the U.S. Department of Agriculture Animal Welfare Act. The protocol was approved by the animal care and use committee at the Penn State College of Medicine. Survival experiments were performed after a series of acute experiments.
Animals were fed standard chow during their required quarantine period. Beginning 24 hours before the procedure, the animals were kept without food (access to water ad libitum). In the final 3 animals of the series, a liquid diet (Ensure; Abbott Laboratories, Abbott Park, Ill) was begun 72 hours before the procedure. For survival experiments, one gram of IV cefazolin and 600,000 units of intramuscular (IM) penicillin G benzathine plus penicillin G procainebased antibiotic were given immediately before the start of the procedure.
General anesthesia was induced by an IM injection of Telazol (500 mg/kg; Fort Dodge Animal Health, Fort Dodge, Iowa), medetomidine (70-80 mg/kg), and butorphanol (300 mg/kg). After endotracheal intubation, anesthesia was maintained with 1% to 2% isoflurane delivered in 100% oxygen (O2), with mechanical ventilation through a semi-closed circuit. The end-tidal carbon dioxide level, O2 saturation, respiratory rate, and pulse were monitored and recorded throughout the procedure. With the animal in a supine position, a gastric-length overtube (Gardus; US Endoscopy, Mentor, Ohio) was advanced into the esophagus with an adult single-channel gastroscope inside. The endoscope was used to lavage the stomach with sterile normal saline solution (NSS) until the stomach was free of gross food material. After evacuation of the NSS irrigant, 900 mL of an antibiotic solution (40 mg neomycin in one liter NSS) was instilled for 10 minutes before being aspirated endoscopically. The remaining 100 mL of antibiotic solution was held in reserve. By using an injection needle (InjectaFlow; Cook Endoscopy, Winston-Salem, NC), 5 to 10 mL NSS was introduced submucosally to elevate the mucosa from the underlying muscle (Fig. 1A). This mucosal pillow was created at a technically convenient position on the posterior wall in the cardiac portion of the stomach, approximately 2 cm distal to the pars esophagea tunicae mucosae. Next, a 1-cm to 1.5-cm incision was created on the pillow with a 4-mm needle-knife (Huibregtse; Cook) (Fig. 1B). A rat-tooth grasping forceps (FG-47L; Olympus Ltd, Tokyo, Japan) was used to begin the dissection of the loose areolar tissue of the submucosal plane, separating the mucosa from the adjacent circular muscle layer. Once a sufficient area had been dissected, the endoscope was advanced into the submucosal plane. By using a combination of dissection with the forceps, as well as blunt dissection with the leading edge of the endoscope, a 10-cm to 12-cm submucosal tunnel was created, directed toward the greater curvature of the stomach (Fig. 1C). The tunnel length was determined by observing the distance
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 691
Endoscopes and instruments An adult single-channel gastroduodenoscope (EG 2940; Pentax America Inc, Mountvale, NJ) was used for the procedure. For the survival experiments, the endoscope and all reusable, nonautoclavable endoscopic equipment (needle-knives, esophageal overtubes, and injection needles) were cleansed and chemically sterilized with 2.4% glutaraldehyde (Cidex; Johnson and Johnson, New Brunswick, NJ) and were air-dried before the start of the procedure. All reusable equipment capable of undergoing steam sterilization (endoscopic forceps and clip appliers) was autoclaved. Endoclips were supplied by Microvasive Endoscopy, Boston Scientific Corp, Natick, Mass, and Olympus America Inc, Center Valley, Pa.
Preoperative care
Self-approximating transluminal access technique
Pauli et al
Figure 2. Arterial-venous pair identified within the submucosa.
to 1.5-cm linear incision in the seromuscular layer of the stomach at the distal end of the submucosal tract. This completed the gastrotomy and allowed the endoscope to advance into the abdomen (Fig. 1D). On one occasion (with animal no. 5), a 15-mm controlled radial expansion balloon (CRE; Microvasive) was used to dilate the distal tract to allow easier passage of the endoscope into the abdomen. Pneumoperitoneum was established, with the judicious use of the endoscope air pump. A brief peritoneoscopy was performed by using standard endoscopic techniques to assess for evidence of iatrogenic injury. When necessary, the omentum was incised with the needle-knife. At the conclusion of our abdominal exploration, the pneumoperitoneum was evacuated by using the suction channel of the endoscope. The endoscope was withdrawn from the submucosal tract. The seromuscular incision was not closed. The gastric mucosal incision was closed with an endoscopic clip fixing device (HX-5QR-1 and Quick Clip [Olympus] or Resolution [Microvasive]) (Fig. 1E). The endoscope and overtube were removed, and the animal was awakened from anesthesia. The accompanying video demonstrates this technique (Video 1, available online at www.giejournal.org). Figure 1. The transgastric STAT. A, Mucosal pillow creation with saline solution injection. B, Needle-knife incision to enter the submucosa. C, Controlled submucosal tunneling with a forceps. D, Seromuscular incision to complete the transgastric access. E, Endoclip closure of the mucosal incision.
Postoperative care
markers as the endoscope was withdrawn from the tract. As perforating blood vessels to the mucosa were identified, dissection was carried out around the vasculature (Fig. 2). Once the desired length and direction of the tunnel was achieved, the remaining 100 mL of reserved antibiotic solution was used to irrigate the tract before entering the abdomen. The needle-knife was then used to make a 1-cm
Nonsurvival animals were euthanized after completing the operative procedure. For the survival animals, a regular diet was restarted the morning after the procedure. The animals had immediate free access to water and were monitored twice daily by the veterinary staff for evidence of altered behavior, clinical distress, or appetite changes. Routine postoperative analgesics were not administered. Survival animals were weighed weekly during their twoweek survival period to assess for appropriate weight gain. All animals were euthanized with an overdose of pentobarbital sodium (O100 mg/kg IV), after the induction of anesthesia.
692 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
Pauli et al
Self-approximating transluminal access technique
TABLE 1. Results of survival experiments Animal
1
2
3
4
5
Time to complete gastrotomy (min)
22
26
26
20
42
Good
Good
Good
Good
Moderate
5.9
4.5
5.4
2.7
3.6
7
6
3.5
6.5
5
Well healed
Well healed
Well healed
Well healed
Well healed
Normal
Normal
Left upper quadrant adhesion
Normal
Fibrinous adhesions
2/2
3/3
3/2
4/4
3/3
None
None
Microabscess
None
Submucosal abscess
Food intake Weight gain (kg) Tunnel length at necropsy (cm) Gastrotomy Abdominal cavity Clips placed/retained Complications
Necropsy Immediately after euthanasia, a necropsy was performed on all animals in the presence of a veterinary physician. The peritoneal cavity was evaluated for evidence of abscess formation, adhesive disease, and unrecognized iatrogenic injuries. The entire length of the submucosal tract (from the mucosal clips to the seromuscular incision site) was assessed by gross examination and submitted, in 10% formalin solution, for microscopic histopathologic examination. The specimens were later imbedded in paraffin by using a tissue processor (Tissue-Tek VIP5; Sakura Finetek, Torrance, Calif), sectioned on a manual microtome, and stained with Harris-modified hematoxylin solution and eosin. All slides were reviewed by a veterinary pathologist.
of iatrogenic injury to the abdominal wall or to nearby organs. There was no evidence of bleeding within the submucosal tunnel or from the seromuscular incision itself. When immersed in water, neither of the two esophagogastric organ specimens demonstrated evidence of air leak.
Survival experiments
The submucosal dissection and gastrotomy were performed without complication in two animals. We were able to control the submucosal dissection to guide the tunnel direction and to avoid disruption of perforating mucosal blood vessels. The average time to complete the procedure and to enter the abdominal cavity was 13.5 minutes. By using standard endoscopic techniques (eg, retroflexing and torquing), we were able to explore the abdomen. The gastric mucosal incisions were closed without difficulty. At the necropsy, there was no evidence
The STAT was performed without intraoperative complication in 5 animals. The results of these experiments are summarized in Table 1. In all animals, the tunnel began at a convenient location on the posterior gastric wall. All tunnels were deliberately directed inferiorly toward the greater curvature, and all were measured to be 10 to 12 cm at the time of the operation. We were again able to control the direction of the submucosal dissection and to navigate around perforating blood vessels. The seromuscular incision site was consistently made on the posterior gastric wall near the leaves of the greater omentum. The mean standard deviation (SD) procedure time (from submucosal injection to entry of the endoscope into the abdominal cavity) was 27.2 8.7 minutes. Once inside the abdomen, we undertook peritoneoscopy. Pneumoperitoneum was successfully created by using the endoscope air pump and suction. No untoward physiologic effects from the unmonitored pneumoperitoneum were noted. During endoscopic exploration, we were able to identify the stomach, small bowel, colon, spleen, liver, gall bladder, pancreas, urinary bladder, and uterine horns. The movement of the endoscope within the abdomen was not perceived to be limited by the tunnel. No iatrogenic injuries were suspected at the time of the gastrotomy. Clip closure of the mucosal incision was accomplished with 2 to 4 endoclips (mean 3). There was minimal change in the mucosal incision site (edema or tearing) at the conclusion of the procedure, despite manipulation of the endoscope within the abdomen. During the survival period, all the animals ate freely. The mean (SD) weight gain was 4.42 1.30 kg (18.8% 6.82% increase from baseline weight). None of the animals required analgesics in the postoperative period.
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 693
Data collection The procedures were digitally recorded by using video software (Pinnacle Studio Plus 700 USB, Version 10; Pinnacle Systems, Mountain View, Calif) on a laptop computer (Dell Latitude D820; Dell Inc, Round Rock, Tex). Significant portions of the procedures were also recorded by using a Sony video recorder (Sony Digital Handycam, DCR-VX2000; Sony Corp, Tokyo, Japan). Video editing was performed with Premiere Pro, Version 1.5 (Adobe Systems Inc, San Jose, Calif).
RESULTS Acute experiments
Self-approximating transluminal access technique
Pauli et al
Figure 3. Necropsy images. A, Mucosal incision site with endoclips in place. B, Seromuscular incision, demonstrating healing tissue, without evidence of mucosal necrosis.
At the necropsy, we noted a high rate of endoclip retention (93.3%) (Fig. 3A). All of the mucosal and seromuscular incisions were well healed (Fig. 3B). The mean tunnel length (measured from the endoclips to the seromuscular incision) was 5.6 cm. Two of the animals had mild intra-abdominal adhesions. None of the animals had evidence of iatrogenic injury or of intra-abdominal abscess formation. The final animal in the series (animal no. 5) had a 2 2-cm submucosal abscess noted in the distal portion of the STAT tunnel, as identified at the necropsy (Fig. 4). This abscess did not communicate with either the abdominal cavity or the mucosal incision. This process was subclinical in nature, because the animal thrived and gained weight after the operation. Microscopic tissue examination of the submucosal tracts revealed the following: (1) a gastrotomy site and a submucosal tract with inflammatory changes (Fig. 5), (2) fibrovascular tissue formation within the submucosa, and (3) submucosal microabscess (animal no. 3 only).
Figure 4. Luminal view of the intragastric abscess noted at the necropsy in animal no. 5. The endoclips mark the mucosal incision of the STAT tunnel. The abscess was at the distal portion of the tunnel.
The optimal location and means for obtaining transgastric access for NOTES has yet to be determined. Here we report a submucosal dissection technique performed under direct endoscopic visualization. The method combines standard endoscopic and surgical techniques to produce a transgastric access method that is easily reproducible and that has a small technical learning curve. The application of the STAT to an in vivo model allowed us to reliably and safely access the peritoneal cavity, which suggests that this methodology of achieving transgastric abdominal access may prove to be an enabling technique for further NOTES research. Observations made during this study led us to believe that the STAT may have several advantages over other proposed transgastric access techniques: 1. We were able to withdraw from and reenter the STAT tunnel, without difficulty in identifying or entering
the tract. Once the dissection began, the initial mucosal incision tended to stay tented up. Also, from within the gastric lumen, the full length and direction of the tract was visible (Video 1). This made finding and reentering the mucosal incision technically easy, because the tunnel could be followed back to the initial mucosal incision. A reassessment of the tunnel did not require the placement of a transgastric or transabdominal guidewire, although the technique does not exclude the use of one. 2. By performing the submucosal dissection under direct visualization with a grasping forceps, we were able to finely control the direction and width of the tunnel. This transgastric dissection technique differs from that recently reported by Sumiyama et al.22,23 Unlike high-pressure gas or an endoscopic balloon dissection, the STAT allowed us to visually identify and to intentionally dissect around perforating blood vessels without causing bleeding. We believe that this also reduced the risk of unintended muscular or mucosal perforation seen with a needle-knife, carbon dioxide, or monofilament snare dissection.23-25 By leaving the vasculature intact, we not only had an unobscured view of the tissue plane, but we also preserved the blood supply to the mucosa. We did not observe the partial
694 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
www.giejournal.org
DISCUSSION
Pauli et al
Self-approximating transluminal access technique
Figure 5. Photomicrographs. A, Mucosal incision site (top), with pyogranulomatous submucosal inflammation and intact musculature (bottom) (H&E, orig. mag. 100). B, Intact mucosa and muscularis mucosa (top), focal inflammation, and fibrosis within the submucosa (including giant-cell foreignbody-type reaction) and the seromuscular defect (bottom) (H&E, orig. mag. 100).
necrosis of the mucosa that other investigators reported after submucosal dissection.22 The STAT tunnel also may provide a more ideal endoscope/instrument guide than the giant (8 cm) bleb created via submucosal gas insufflation.22,23 3. Despite the use of aggressive maneuvers during the peritoneoscopy (including torque and angulation of the endoscope), the mucosal incision remained at its original 1-cm to 1.5-cm size and did not develop significant edema or tearing. This allowed for endoclip closure with a minimal number of clips. Although the brevity of our peritoneoscopy may have provided a limited opportunity for the mucosal incision to lengthen, there are several other reasons why our technique may not have resulted in mucosal tearing. These include the following: appropriate selection of the mucosal entry and exit sites, snugness of the tunnel around the endoscope, and the overall length of the tunnel. 4. After withdrawal of the endoscope, the long submucosal tunnel naturally collapsed or self approximated. This was evidenced by the fact that immediately
explanted specimens remained airtight under pressure. We believe it is possible for this method to further seal the gastrotomy (by collapse of the tunnel) under conditions of high intra-abdominal or intragastric pressure, such as coughing or vomiting (Fig. 6). 5. With the endoscope in the submucosal plane, it was possible to visualize the intra-abdominal organs through the yet uncut seromuscular layer (see Video 1). This allowed us to select the exit site from the tunnel in a location where potential iatrogenic injuries could be minimized. Despite these advantages, our technique requires further refinement. We created 10-cm to 12-cm tunnels at the time of the operation. These distances were measured under partial gastric distention with the tissue stretched. At the necropsy, the length of the tract in the decompressed stomach averaged only 6 cm. We do not yet know the optimal length for the STAT, but the tunnel length can be varied without difficulty. Our abdominalaccess technique was also more time consuming than other investigators reported in the literature.19,22
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 695
Self-approximating transluminal access technique
Pauli et al
In conclusion, in this porcine survival model, we were able to successfully enter the abdominal cavity by using the STAT method. Transgastric access was safe and reproducible. The separation of the mucosal and seromuscular incisions proved a reliable means to close the gastrotomy. The finding of a submucosal abscess in one animal suggests that further investigation is warranted. We remain optimistic that the STAT may be an enabling technique to allow transgastric abdominal access for NOTES procedures. Figure 6. Schematic representation of how tunneled gastrotomy seal improves under conditions of high intraluminal or intra-abdominal pressure. Pressure (arrows) leads to increased apposition of the seromuscular and mucosal layers and prevents gastrotomy leak.
However, the potential for a simplified, reliable closure method may ultimately yield a time savings. Although we used high-level instrument disinfection, systemic preoperative antibiotics, and antibacterial gastric lavage, 2 of our 5 survival animals (40%) developed infection complications. One of our animals developed an abscess in the distal portion of the STAT tunnel (animal no. 5), and another animal had a microabscess on pathologic examination (animal no. 3). It should be noted, however, that none of our animals developed intra-abdominal sepsis from a gastrotomy leak or demonstrated signs of infection. Clinically, these animals behaved normally and gained weight appropriately. Microabscess formation was noted on histopathology, even with nontunneled transluminal access.1,26 At the necropsy, we found no gross evidence of intra-abdominal abscess formation. These collections were well circumscribed and did not communicate with either the seromuscular or the mucosal incisions. The remainder of these tracts were well healed. Although we did not perform repeat endoscopy on our animals, the abscess in animal no. 5 would have been identifiable and treatable with currently available endoscopic techniques.27,28 Nevertheless, infection issues are an obvious concern with transluminal-access techniques. Tunneling beneath the mucosal turns creates a void. Subsequent closure of bacteria into this space provides an excellent environment for abscess formation. In recognizing this fact, we flushed the tract with antibiotic irrigation before entering the abdomen. This method may be improved upon in a number of ways, including altering the antibiotics used in the irrigant, irrigating the tunnel at the end of the procedure rather than at the beginning or irrigating with dilute iodine solution rather than antibiotic solution. Providing routine postoperative antibiotics may also be of benefit in procedures that use a STAT gastrotomy. It is clear that further research is required to define the optimum protocol for tunneled transgastric access that provides a reliable closure method and minimizes the risk of infectious complications during the postoperative period. 696 GASTROINTESTINAL ENDOSCOPY Volume 67, No. 4 : 2008
ACKNOWLEDGMENT We thank the dedicated staff of the Department of Comparative Medicine (Drs Wilson, Carney, and Griffith, and Ms Joy Ellwanger) and Ms Pat Grosh for their help with this project.
DISCLOSURE The authors report that there are no disclosures relevant to this publication. Microvasive Endoscopy, Boston Scientific, Inc, and Olympus Medical, Inc provided endoscopic clipping devices. Pentax Medical, Inc provided endoscopy equipment.
REFERENCES 1. Kalloo AN, Singh VK, Jagannath SB, et al. Flexible transgastric peritoneoscopy: a novel approach to diagnostic and therapeutic interventions. Gastrointest Endosc 2004;60:114-7. 2. Jagannath SB, Kantsevoy SV, Vaughn CA, et al. Peroral transgastric endoscopic ligation of fallopian tubes with long-term survival in a porcine model. Gastrointest Endosc 2005;61:49-53. 3. Kantsevoy SV, Jagannath SB, Niiyama H, et al. Endoscopic gastrojejunostomy with survival in a porcine model. Gastrointest Endosc 2005;62:278-92. 4. Kantsevoy SV, Niiyama H, Jagannath SB, et al. The endoscopic transilluminator: an endoscopic device for identification of the proximal jejunum for transgastric endoscopic gastrojejunostomy. Gastrointest Endosc 2006;63:1055-8. 5. Bergstrom M, Ikeda K, Swain P, et al. Transgastric anastomosis using flexible endoscopy in a porcine model (with video). Gastrointest Endosc 2006;63:307-12. 6. Park PO, Bergstrom M, Ikeda K, et al. Experimental studies of transgastric gallbladder surgery: cholecystectomy and cholecystogastric anastomosis. Gastrointest Endosc 2005;61:601-6. 7. Rolanda C, Lima E, Pego JM, et al. Third-generation cholecystectomy by natural orifices: transgastric and transvesicular combined approach (with video). Gastrointest Endosc 2007;65:111-7. 8. Swanstrom LL, Kozarek R, Pasricha PJ, et al. Development of a new access device for transgastric surgery. J Gastrointest Surg 2005;9: 1129-36. 9. Kantsevoy SV, Hu B, Jagannath SB, et al. Transgastric endoscopic splenectomy: is it possible? Surg Endosc 2006;20:522-5. 10. Wagh MS, Merrifield BF, Thompson CC. Survival studies after endoscopic transgastric oophorectomy and tubectomy in a porcine model. Gastrointest Endosc 2006;63:473-8.
www.giejournal.org
Pauli et al
Self-approximating transluminal access technique
11. Wagh MS, Merrifield BF, Thompson CC. Endoscopic transgastric abdominal exploration and organ resection: initial experience in a porcine model. Clin Gastroenterol Hepatol 2005;3:892-6. 12. Merrifield BF, Wagh MS, Thompson CC. Peroral transgastric organ resection: a feasibility study in animals. Gastrointest Endosc 2006;63: 693-7. 13. Sumiyama K, Gostout CJ, Rajan E, et al. Pilot study of the porcine uterine horn as an in vivo appendicitis model for development of endoscopic transgastric appendectomy. Gastrointest Endosc 2006;64:808-12. 14. American Society for Gastrointestinal Endoscopy/SAGES Working Group on Natural Orifice Translumenal Endoscopic Surgery white paper October 2005. Gastrointest Endosc 2006;63:199-204. 15. Hochberger J, Lamade W. Transgastric surgery in the abdomen: the dawn of a new era? Gastrointest Endosc 2005;62:293-6. 16. Lamade W, Hochberger J. Transgastric surgery: avoiding pitfalls in the development of a new technique. Gastrointest Endosc 2006;63: 698-700. 17. Ponsky JL. Gastroenterologists as surgeons: what they need to know. Gastrointest Endosc 2005;61:454. 18. Sclabas GM, Swain P, Swanstrom LL. Endoluminal methods for gastrotomy closure in natural orifice transenteric surgery (NOTES). Surg Innov 2006;13:23-30. 19. Kantsevoy SV, Jagannath SB, Niiyama H, et al. A novel safe approach to the peritoneal cavity for per-oral transgastric endoscopic procedures. Gastrointest Endosc 2007;65:497-500. 20. Moyer MT, Pauli EM, Haluck RS, et al. Self-approximating translumenal access technique (STAT) for potential use in NOTES: an ex vivo porcine model (with video). Gastrointest Endosc 2007;66:974-8. 21. Sumiyama K, Gostout CJ, Rajan E, et al. Transesophageal mediastinoscopy by submucosal endoscopy with mucosal flap safety valve technique. Gastrointest Endosc 2007;65:679-83.
22. Sumiyama K, Gostout CJ, Rajan E, et al. Submucosal endoscopy with mucosal flap safety valve. Gastrointest Endosc 2007;65:688-94. 23. Sumiyama K, Gostout CJ, Rajan E, et al. Transgastric cholecystectomy: transgastric accessibility to the gallbladder improved with the SEMF method and a novel multibending therapeutic endoscope. Gastrointest Endosc 2007;65:1028-34. 24. Ro¨sch T, Sarbia M, Schumacher B, et al. Attempted endoscopic en bloc resection of mucosal and submucosal tumors using insulatedtip knives; a pilot series (including videos). Endoscopy 2004;36: 788-801. 25. Kamler JP, Borsatto R, Binmoeller KF. Circumferential endoscopic mucosal resection in the swine esophagus assisted by a cap attachment. Gastrointest Endosc 2002;55:923-8. 26. Pai RD, Fong DG, Bundga ME, et al. Transcolonic endoscopic cholecystectomy: a NOTES survival study in a porcine model (with video). Gastrointest Endosc 2006;64:428-34. 27. Ruiz-Rebollo ML, Atienza-Sa´nchez R, Go´mez-Corral J. Gastric wall abscess caused by an ingested toothpick. Gastrointest Endosc 2007;65:518-9. 28. Choong NWW, Levy MJ, Rajan E, et al. Intramural gastric abscess: case history and review. Gastrointest Endosc 2003;58:627-9.
www.giejournal.org
Volume 67, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 697
Received May 24, 2007. Accepted September 4, 2007. Current affiliations: Division of Minimally Invasive and Bariatric Surgery (E.M.P., R.S.H.), Division of Gastroenterology and Hepatology (M.T.M., A.M.), The Milton S. Hershey Medical Center, Penn State College of Medicine, Hershey, Pennsylvania, USA. Reprint requests: Eric M. Pauli, MD, Department of Surgery, PO Box 850 MC H149, Penn State College of Medicine, Hershey, PA 17033.