Transluminal closure for NOTES: an ex vivo study comparing leak pressures of various gastrotomy and colotomy closure modalities

• Introduction
• Methods
• Results
• Discussion
• Acknowledgments
• References

Background and study aims: Transluminal closure is fundamental to the safe introduction of natural orifice transluminal endoscopic surgery (NOTES) into humans. Suture, staples, and clips have been used. We aimed to evaluate the acute strength of various gastrotomy and colotomy closure techniques in an ex vivo porcine model by assessing air leak pressures.

Patients and methods: The following closure modalities were assessed with at least five samples per arm: conventional open/laparoscopic suturing techniques including full-thickness interrupted sutures, double-layer sutures, and purse-string sutures, as well as endoscopic clips and endoscopic staples. Historical values for transgastric closures with hand-sewn interrupted sutures, endoscopic clips, and a prototype endoscopic suture device were used from our laboratory’s prior study.

Results: Using Kruskal-Wallis analysis, the overall comparisons were significant (P = 0.0038 for gastrotomy closure; P = 0.0018 for colotomy closure). Post hoc paired comparisons revealed that the difference between all closure arms versus negative control were significant. Significance could not be established among the various closure arms. However, trends suggested hand-sewn double-layer sutures, endoscopic staples, and both hand-sewn and endoscopically-placed purse-string sutures produced the strongest closures. Furthermore, endoscopic clips appeared sufficient for colotomy closure when ideally placed.

Conclusions: Suture (both hand-sewn and endoscopically deployed) appears to produce the strongest closures in both stomach and colon, with the important caveats that (1) a continuous through-thickness suture track be avoided, such as in the full-thickness closure, or (2) suture holes be buried, such as in the purse-string configuration. When suture tracks are full-thickness, they can serve as leak sites. Staples and clips can produce comparable closures, but only under ideal conditions.

Introduction

The 2006 natural orifice transluminal endoscopic surgery (NOTES) White Paper by the ASGE/SAGES Working Group highlighted secure transluminal closure (specifically gastric closure) as one of the most fundamental challenges to the safe introduction of NOTES into humans [1]. Following the original focus on transgastric access, other routes of transluminal entry have been explored, including transcolonic [2] [3], transvesical [4], and transvaginal [5]. Regardless of route, the importance of secure closure remains.

In addition to the various routes of access, several modalities of transluminal closure have been utilized, including various endoscopic suturing devices [6] [7] endoscopic clips [8] [9], and endoscopic staplers [10]. The purpose of this study was to evaluate the acute strength of gastrotomy and colotomy closure via various modalities by assessing air leak pressures in an ex vivo porcine model.

The following modalities were utilized in this study: conventional open and laparoscopic suturing techniques such as full-thickness interrupted sutures, double-layer sutures, and purse-string sutures. Additionally, endoscopic clips and staples were included. Historical values for transgastric closures with hand-sewn interrupted sutures, endoscopic clips, and a prototype endoscopic purse-string suture device were used from a prior study [11].

Methods

Whole stomachs and distal colons were harvested from adult white pigs weighing 30 - 50 kg (Adams Farm, Athol, Massachusetts, USA). The volume of the stomachs ranged from 1000 - 1500 ml, comparable to a small adult human stomach. The distal 20 cm of the porcine colon was harvested, as previous porcine transcolonic experiments have targeted the site of peritoneal access at 10 - 15 cm from the anal verge. The specimens were only refrigerated and not frozen, and were used within 3 days of procurement.

Subcentimeter gastrotomy and colotomy incisions were created by needle knife (Microvasive, Boston Scientific, Natick, Massachusetts, USA). The harvested organs were everted to replicate endoluminal application of electrocautery ([Fig. 1]). In order to standardize the size of the incisions, they were dilated with an 18-mm controlled radial expansion (CRE) balloon (Boston Scientific), inflated to a pressure of 20 cm H₂O for 1 minute ([Fig. 2]).

Stomachs were then assigned to one of four closure arms, with at least five stomachs used per arm: (1) purse-string suture with 2 - 0 polypropylene; (2) double-layer closure with 3 - 0 polypropylene; (3) prototype endoscopic linear stapler (Power Medical Interventions, Langhorne, Pennsylvania, USA); (4) no closure (negative control). These arms were compared with historical values obtained from a previous experiment, which featured: (5) full-thickness interrupted sutures using 3 - 0 polypropylene; (6) QuikClips (Olympus America, Inc., Center Valley, Pennsylvania, USA); and (7) prototype purse-string closure device (LSI Solutions, Victor, New York, USA). Therefore, seven total modalities were compared. One stomach was used for each individual closure event ([Fig. 3]).

Colons were assigned to one of six closure arms: (1) full-thickness interrupted sutures with 3 - 0 polypropylene; (2) purse-string suture with 2 - 0 polypropylene; (3) double-layer sutures using full-thickness running chromic and interrupted serosal sutures with 3 - 0 polypropylene; (4) QuikClips (Olympus America, Inc.); (5) prototype endoscopic linear stapler (Power Medical Interventions); and (6) no closure (negative control). One colon was used for each individual closure event ([Fig. 3]).

Following otomy closure, pressure recordings were obtained using a digital pressure gauge connected by noncollapsible tubing to a 16-gauge Veris needle that resided within the organ lumen. The organs were slowly insufflated using compressed air introduced through another 16-gauge Veris needle. The esophageal stump, the pylorus, and the ends of the colonic segment were sealed using surgical clamps. The organs were then submerged, and the pressure (mm Hg) to achieve air leak (i. e. bubble leak) was recorded ([Fig. 4]).

Again, two prototype devices were featured in this study. The prototype device from LSI Solutions ([Fig. 5]) uses vacuum assist to secure tissue within its functional distal tip. The creation of transluminal access and deployment of the circumscribing purse-string suture are combined into one process. The purse-string is closed and secured by a separate knot-tying device. The endoscopic linear stapler from Power Medical Interventions ([Fig. 6]) uses computer guidance to sense proper tissue compression prior to automated stapler firing and transection of tissue. Of note, only blue loads (staples designed for thinner tissue such as colon) were used in this experiment due to product availability. [Tables 1] and [2] provide details for every closure modality.

Results

Statistical analysis was carried out by the Kruskal-Wallis test. Post hoc paired comparisons were performed using the MULTTEST procedures with Bonferroni and bootstrap adjustments. All analyses were performed with SAS version 9.1 (Cary, North Carolina, USA).

[Fig. 7] reports the mean and median values obtained for each method of gastrotomy closure, and [Fig. 8] reports those values for colotomy closure. For the gastrotomy closures, the Kruskal-Wallis test was significant (P = 0.0038). Post hoc paired comparisons showed all arms to be significant compared with negative control. However, significance could not be established for comparisons among the various closure arms. Likewise, for colotomy closures, the Krusal-Wallis test was significant (P = 0.0018). Post hoc paired comparisons showed all colotomy closure arms to be significant compared with negative control, but again, the comparisons among the closure arms were not statistically significant.

Nevertheless, certain trends are suggested by the data. For gastrotomy closures, double-layer sutures, purse-string sutures, and staples appeared strongest. Full-thickness interrupted sutures and clips were less robust. For colotomy closures, double-layer sutures, purse-string sutures, staples, and clips appeared strong. Again, full-thickness interrupted sutures were not.

The following important observations were made. (1) Leaks consistently occurred at sites where suture, staples or clips breached tissue integrity. (2) Clips, even when deployed under ”ideal“ conditions (i. e. by hand, with back-pressure), achieved mucosal closure but never serosal closure for gastrotomies. Even in thinner colonic tissue, clips rarely achieved serosal closure. (3) Blue load staples achieved full-thickness closures in colon but only mucosal/seromuscular closure in stomach. This was not surprising given that blue loads are designed for thinner tissue. (4) Secure mucosal and/or seromuscular closures without serosal closures (i. e. clips and blue load staples) were sufficient to yield strong closures that were more robust than expected.

Discussion

The following limitations regarding this study should be highlighted. First, this study was sufficiently powered to determine statistical significance against the control arm, but was not adequately powered to detect difference among test arms. Therefore, the ensuing discussion is based on trends that are suggested by the data. Second, this study deals with the acute strength of closure in ex vivo organs. Future studies may focus on survival animals to assess the interaction between closure modalities and living tissue (e. g. blood flow, ischemia, tissue healing) over a period of time. Third, the endoscopic modalities tested were executed by hand, representing ideal closure conditions. Reproduction of true endoscopic conditions could potentially yield lower leak pressures. Finally, although this study aims to identify a hierarchy of closure modalities, it may be the case that any method that yields strength of closure above a certain critical in vivo leak-pressure threshold will be sufficient for maintaining gut wall integrity.

Despite these caveats, this study presents some interesting observations that may impact the development and future clinical adoption of NOTES closure modalities. Probably the most significant observation is the fact that, in the acute setting, full-thickness closures are not necessarily superior to mucosal or seromuscular closures. The reason is that the site of tissue breach by suture or staple ultimately becomes the leak site. This phenomenon played out in every modality tested. For example, this phenomenon is well illustrated in the difference between full-thickness sutures and double-layer or purse-string sutures. Full-thickness sutures create through-thickness suture tracts by which air can easily leak. Consequently, full-thickness sutures yielded some of the weakest closures tested. In contrast, double-layer sutures close each layer separately, such that a through-thickness tract is avoided. Purse-string sutures mechanically bolster their suture sites in another way: purse-strings produce tissue involution and thereby bury the suture holes. This was true for both hand-sewn and prototype endoscopically delivered suture. Consequently, double-layer and purse-string sutures yielded some of the strongest gastrotomy and colotomy closures.

The staple arm also illustrated the same phenomenon. Due to product availability, only blue loads (closed staple heights of 1.3 mm, ideal for colon) could be used in this study. Ideally, green loads (closed staple heights of 1.7 mm) would have been utilized for gastric tissue [12]. To our surprise, blue loads fared well for gastrotomy closure, producing almost the same leak pressures as colotomy closures despite being able to secure only mucosal and seromuscular closure in the stomach. We predict that future gastrotomy closure experiments with the appropriate green loads will yield higher leak pressures. Even though staples are designed for full-thickness closure, the presence of an adjacent row of appropriately sized staples will likely augment mechanical support [13]. (So, too, will the use of staple-line reinforcement, such as nonabsorbable expanded polytetrafluoroethylene [ePTFE] or biodegradable buttressing, which are commonly used but were not featured in this study.) Of note, the stapler has already been successfully used for endoscopic full-thickness gastric resections in humans [12]. Furthermore, it should be pointed out that that the current version of the endoscopic stapler does present some difficulty with in vivo access to the transgastric incision given (1) the relatively short length of the working shaft (60cm), and (2) the need for a parallel endoscope to provide visualization [14].

Finally, clips provided surprisingly strong closures, particularly in colon, despite producing almost entirely nonserosal closures. It is surmised that clip colotomy closures yielded such high leak pressures because the colonic mucosa is thinner and permits the clips to bite into more seromuscular tissue. Again, it should be emphasized that these clips were manually placed with appropriate back-pressure, and replicating this task endoscopically would likely be more difficult. Merrifield et al. has reported inadequate closure of the transgastric incision with endoclips, with septic complications occurring by 2 weeks in two of five animals [15]. In their work on transesophageal access, Fritscher-Ravens et al. has also reported that endo-clip closures yield superficial closures with some discontinuation of the muscular layer, which could potentially lead to the formation of diverticula [16]. However, the Raju group has shown in their extensive work with clip closures of colonic perforations that clips are capable of leak-proof closures of even large (5 cm) transmural wounds [17], and that the introduction of the Multi-Clip Applier (InScope, Ethicon Endo-Surgery, Cincinnati, Ohio, USA) may further facilitate the task [18]. Given the results of our ex vivo work and the growing body of literature examining in vivo endoclip closure for various NOTES access sites, perhaps it would be prudent to restrict endoclip closures to transcolonic access.

The optimal in vivo leak-pressure is not only unknown but likely different for both stomach and colon. Key physiologic variables include native intraluminal pressures, vascular redundancy, and microbial density. From the perspective of physiologic pressures, intragastric pressures in humans as high as 200 mm Hg have been reported during emesis. While no closure modality tested thus far has generated leak pressures this high, and porcine transgastric survival studies have demonstrated clinical effectiveness with ”weaker” closure modalities, this may represent a target for gastric closure devices. In general, it is reasonable to assume that transgastric closures will see higher physiologic pressures compared with transcolonic closures, and different devices may be ideal for each setting. A second critical variable is vascular supply. For example, whereas the stomach and the rectum have redundant vascular supplies, the sigmoid colon in comparison occupies a relative watershed area and could be more susceptible to ischemia. Thus, an excessively strong trans-sigmoid closure may cause more significant tissue necrosis. Finally, local microbial density represents an important variable. From the perspective of peritonitis, a small colonic leak will likely pose a bigger hazard than a small gastric leak. Ultimately, it is unclear how and to what extent these variables contribute to an optimal range of leak pressures for stomach and colon, but survival studies will hopefully shed more light on this matter.

In conclusion, suture, staples, and clips all appear to be capable of yielding strong closures under different circumstances; however, none have been acceptably adapted for NOTES in the clinical setting. Several strategies for maximizing strength of acute closure were identified during this study: (1) full-thickness closure while avoiding a through-thickness tract (e. g. double-layer closure or purse-string suture), (2) seromuscular bites without serosal breach (e. g. clips in colon and blue load staples in stomach), and (3) use of full-thickness staples with an adjacent row for mechanical support. Although it is not clear which of these modalities is best, or if a device yet to be studied will prevail, the task of transluminal closure using an endoscopic platform will require an approach that maximizes efficiency and yields consistent results.

Acknowledgments

The authors would like to thank Rie Maurer for her assistance with the biostatistics aspect of this work.

Competing interests: None

References

• 1 ASGE/SAGES Working Group on Natural Orifice Translumenal Endoscopic Surgery . White Paper, October 2005.  Gastrointest Endosc. 2006;  63 199-203
  CrossrefPubMedSearch in Google Scholar
• 2 Fong D G, Pai R D, Thompson C C. Transcolonic endoscopic abdominal exploration: a NOTES survival study in the porcine model.  Gastrointest Endosc. 2007;  65 312-318
  CrossrefPubMedSearch in Google Scholar
• 3 Pai R D, Fong D G, Bundga M E. et al . Transcolonic endoscopic cholecystectomy: a NOTES survival study in a porcine model.  Gastrointest Endosc. 2006;  64 428-434
  CrossrefPubMedSearch in Google Scholar
• 4 Rolanda C, Lima E, Pego J M. et al . Third-generation cholecystectomy by natural orifices: transgastric and transvesical combined approach (with video).  Gastrointest Endosc. 2007;  65 111-117
  CrossrefPubMedSearch in Google Scholar
• 5 Tsin D A, Sequeria R J, Giannikas G. Culdolaparoscopic cholecysetectomy during vaginal hysterectomy.  JSLS. 2003;  7 171-172
  PubMedSearch in Google Scholar
• 6 Hu B, Chung S C, Sun L C. et al . Endoscopic suturing without extracorporeal knots: a laboratory study.  Gastrointest Endosc. 2005;  62 266-270
  CrossrefPubMedSearch in Google Scholar
• 7 Swain P. Endoscopic suturing: now and incoming.  Gastrointest Endosc Clin N Am. 2007;  17 505-520
  CrossrefPubMedSearch in Google Scholar
• 8 Kalloo A N, Singh V K, Jagannath S B. et al . Flexible transgastric peritoneoscopy: a novel approach to diagnostic and therapeutic interventions in the peritoneal cavity.  Gastrointest Endosc. 2004;  60 114-117
  CrossrefPubMedSearch in Google Scholar
• 9 Wagh M S, Merrifield B F, Thompson C C. Endoscopic transgastric abdominal exploration and organ resection: initial experience in the porcine model.  Clin Gastroenterol Hepatol. 2005;  3 892-896
  CrossrefPubMedSearch in Google Scholar
• 10 Kaehler G F, Langner C, Suchan K L. et al . Endoscopic full-thickness resection of the stomach: an experimental approach.  Surg Endosc. 2006;  20 519-521
  CrossrefPubMedSearch in Google Scholar
• 11 Ryou M, Pai R, Sauer J. et al . Evaluating an optimal gastric closure method for transgastric surgery.  Surg Endosc. 2007;  21 677-680
  CrossrefPubMedSearch in Google Scholar
• 12 Kaehler G, Grobholz R, Langner C. et al . A new technique of endoscopic full-thickness resection using a flexible stapler.  Endoscopy. 2006;  38 86-89
  Thieme ConnectPubMedSearch in Google Scholar
• 13 Baker R S, Foote J, Kemmeter P. et al . The science of stapling and leaks.  Obes Surg. 2004;  14 1290-1298
  CrossrefPubMedSearch in Google Scholar
• 14 Magno P, Giday S A, Chung S S. et al . A new stapler-based full-thickness transgastric access closure: results from an animal pilot trial.  Endoscopy. 2007;  39 876-880
  Thieme ConnectPubMedSearch in Google Scholar
• 15 Merrifield B F, Wagh M S, Thompson C C. Peroral transgastric organ resection: a feasibility study in pigs.  Gastrointest Endosc. 2006;  63 693-697
  CrossrefPubMedSearch in Google Scholar
• 16 Fritscher-Ravens A, Patel K, Ghanbari A. et al . Natural orifice transluminal endoscopic surgery (NOTES) in the mediastinum: long-term survival animal experiments in transesophageal access, including minor surgical procedures.  Endoscopy. 2007;  39 870-875
  Thieme ConnectPubMedSearch in Google Scholar
• 17 Raju G S, Ahmed I, Brining D. et al . Endoluminal closure of large perforations of colon with clips in a porcine model (with video).  Gastrointest Endosc. 2006;  64 640-646
  CrossrefPubMedSearch in Google Scholar
• 18 Raju G S, Ahmed I, Xiao S Y. et al . Controlled trial of immediate endoluminal closure of colon perforations in a porcine model by use of a novel clip device (with videos).  Gastrointest Endosc. 2006;  64 989-997
  CrossrefPubMedSearch in Google Scholar

C. C. Thompson, MD

Brigham and Women’s Hospital

Division of Gastroenterology and Hepatology

75 Francis St.

Boston, MA 02115

USA

Fax: +1-617-525-8266

Email: ccthompson@partners.org

References

• 1 ASGE/SAGES Working Group on Natural Orifice Translumenal Endoscopic Surgery . White Paper, October 2005.  Gastrointest Endosc. 2006;  63 199-203
  CrossrefPubMedSearch in Google Scholar
• 2 Fong D G, Pai R D, Thompson C C. Transcolonic endoscopic abdominal exploration: a NOTES survival study in the porcine model.  Gastrointest Endosc. 2007;  65 312-318
  CrossrefPubMedSearch in Google Scholar
• 3 Pai R D, Fong D G, Bundga M E. et al . Transcolonic endoscopic cholecystectomy: a NOTES survival study in a porcine model.  Gastrointest Endosc. 2006;  64 428-434
  CrossrefPubMedSearch in Google Scholar
• 4 Rolanda C, Lima E, Pego J M. et al . Third-generation cholecystectomy by natural orifices: transgastric and transvesical combined approach (with video).  Gastrointest Endosc. 2007;  65 111-117
  CrossrefPubMedSearch in Google Scholar
• 5 Tsin D A, Sequeria R J, Giannikas G. Culdolaparoscopic cholecysetectomy during vaginal hysterectomy.  JSLS. 2003;  7 171-172
  PubMedSearch in Google Scholar
• 6 Hu B, Chung S C, Sun L C. et al . Endoscopic suturing without extracorporeal knots: a laboratory study.  Gastrointest Endosc. 2005;  62 266-270
  CrossrefPubMedSearch in Google Scholar
• 7 Swain P. Endoscopic suturing: now and incoming.  Gastrointest Endosc Clin N Am. 2007;  17 505-520
  CrossrefPubMedSearch in Google Scholar
• 8 Kalloo A N, Singh V K, Jagannath S B. et al . Flexible transgastric peritoneoscopy: a novel approach to diagnostic and therapeutic interventions in the peritoneal cavity.  Gastrointest Endosc. 2004;  60 114-117
  CrossrefPubMedSearch in Google Scholar
• 9 Wagh M S, Merrifield B F, Thompson C C. Endoscopic transgastric abdominal exploration and organ resection: initial experience in the porcine model.  Clin Gastroenterol Hepatol. 2005;  3 892-896
  CrossrefPubMedSearch in Google Scholar
• 10 Kaehler G F, Langner C, Suchan K L. et al . Endoscopic full-thickness resection of the stomach: an experimental approach.  Surg Endosc. 2006;  20 519-521
  CrossrefPubMedSearch in Google Scholar
• 11 Ryou M, Pai R, Sauer J. et al . Evaluating an optimal gastric closure method for transgastric surgery.  Surg Endosc. 2007;  21 677-680
  CrossrefPubMedSearch in Google Scholar
• 12 Kaehler G, Grobholz R, Langner C. et al . A new technique of endoscopic full-thickness resection using a flexible stapler.  Endoscopy. 2006;  38 86-89
  Thieme ConnectPubMedSearch in Google Scholar
• 13 Baker R S, Foote J, Kemmeter P. et al . The science of stapling and leaks.  Obes Surg. 2004;  14 1290-1298
  CrossrefPubMedSearch in Google Scholar
• 14 Magno P, Giday S A, Chung S S. et al . A new stapler-based full-thickness transgastric access closure: results from an animal pilot trial.  Endoscopy. 2007;  39 876-880
  Thieme ConnectPubMedSearch in Google Scholar
• 15 Merrifield B F, Wagh M S, Thompson C C. Peroral transgastric organ resection: a feasibility study in pigs.  Gastrointest Endosc. 2006;  63 693-697
  CrossrefPubMedSearch in Google Scholar
• 16 Fritscher-Ravens A, Patel K, Ghanbari A. et al . Natural orifice transluminal endoscopic surgery (NOTES) in the mediastinum: long-term survival animal experiments in transesophageal access, including minor surgical procedures.  Endoscopy. 2007;  39 870-875
  Thieme ConnectPubMedSearch in Google Scholar
• 17 Raju G S, Ahmed I, Brining D. et al . Endoluminal closure of large perforations of colon with clips in a porcine model (with video).  Gastrointest Endosc. 2006;  64 640-646
  CrossrefPubMedSearch in Google Scholar
• 18 Raju G S, Ahmed I, Xiao S Y. et al . Controlled trial of immediate endoluminal closure of colon perforations in a porcine model by use of a novel clip device (with videos).  Gastrointest Endosc. 2006;  64 989-997
  CrossrefPubMedSearch in Google Scholar

C. C. Thompson, MD

Brigham and Women’s Hospital

Division of Gastroenterology and Hepatology

75 Francis St.

Boston, MA 02115

USA

Fax: +1-617-525-8266

Email: ccthompson@partners.org