6
Full Reports Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model David J. Desilets, MD, PhD, 1 John R. Romanelli, MD, 2 David B. Earle, MD, 2 and Christopher N. Chapman, MD 3 Abstract Background and Study Aims: The success of natural orifice surgery depends on secure closure of the transmural gut opening, so a rapid, secure, and easy-to-place closure method is desirable. Our aim was to determine whether a gastrotomy can be closed safely and effectively from within the stomach in a survival model by using a novel, endoscopically placed device: the Padlock-G system. Patients and Methods: This was a pilot study of 4 survival animals in an animal laboratory setting. Gastrotomies were made in the stomachs of laboratory swine, and the abdomen was explored by using a standard gastro- scope. Gastrotomies were then closed by using the Padlock-G system. Survival for 2 or 6 weeks was the primary outcome measurement. Secondary outcomes included ease of use, visual assessment of closure integrity im- mediately and at necropsy, presence of adhesions, evidence of infection, and histologic appearance at the closure sites. Results: All animals thrived, ate normally, and gained weight. None developed fever, tachycardia, or signs of peritoneal irritation. Closure-site inspection at necropsy revealed excellent healing, with epithelial growth over the Padlock-G. There were no ulcers, serosal surfaces were tightly closed, and no defects could be seen. There were no signs of peritoneal inflammation, intra-abdominal adhesions, or gastric spillage. Histologic evaluation showed organizing granulation tissue with fibrosis, vascular proliferation, and mild chronic inflammatory in- filtrate (i.e., scar). Conclusions: The Padlock-G is easy to place, provides a durable closure, and allows survival animals to thrive without adverse sequellae. This device provides a suitable closure system for transgastric NOTES. Introduction N atural orifice transluminal endoscopic surgery (NOTES) is an emerging technique where surgery is performed without a skin incision, using a natural bodily orifice to provide access. An enterotomy is made across the lumen of a hollow organ, and a body cavity or potential space is entered to perform the surgery. For intra-abdominal NOTES, the stomach is one useful portal of entry. Gastrotomy closure is therefore a key consideration. At present, there are no rapid, reproducible, reliable gastrotomy closure methods for transgastric NOTES. The aim of this study was to evaluate gastrotomy closure in a survival model by using a hexagonal Nitinol clip: the Padlock-G. Our previously published explant and nonsurvival animal data describe the first successful application of this device in a live animal. 1 We now report the results of our survival study using this closure system in 4 laboratory swine that were survived for 2–6 weeks. Materials and Methods The Lock-It System Ò (Aponos Medical, Kingston, NH) comprises the Padlock-G, a Nitinol clip that is 18 mm in di- ameter and 0.76 mm thick (Fig. 1), as well as a scope-mounted delivery pod (Fig. 2). The Padlock-G is folded to fit into the pod but springs back into its original shape when deployed. This locking device is six-sided and has six inner prongs that embed into the gastric wall around the gastrotomy site. When the device is deployed and snaps back into its original flat or disc-like shape, these six prongs gather the edges of the 1 Division of Gastroenterology, Department of Medicine, Baystate Medical Center, Tufts University School of Medicine, Springfield, Massachusetts. 2 Department of Surgery, Baystate Medical Center, Tufts University School of Medicine, Springfield, Massachusetts. 3 Department of Pathology, Hartford Hospital, Hartford, Connecticut. JOURNAL OF LAPAROENDOSCOPIC & ADVANCED SURGICAL TECHNIQUES Volume 20, Number 8, 2010 ª Mary Ann Liebert, Inc. DOI: 10.1089/lap.2010.0076 671

Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

Embed Size (px)

Citation preview

Page 1: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

Full Reports

Gastrotomy Closure with the Lock-It System and thePadlock-G clip: A Survival Study in a Porcine Model

David J. Desilets, MD, PhD,1 John R. Romanelli, MD,2

David B. Earle, MD,2 and Christopher N. Chapman, MD3

Abstract

Background and Study Aims: The success of natural orifice surgery depends on secure closure of the transmuralgut opening, so a rapid, secure, and easy-to-place closure method is desirable. Our aim was to determinewhether a gastrotomy can be closed safely and effectively from within the stomach in a survival model by usinga novel, endoscopically placed device: the Padlock-G system.Patients and Methods: This was a pilot study of 4 survival animals in an animal laboratory setting. Gastrotomieswere made in the stomachs of laboratory swine, and the abdomen was explored by using a standard gastro-scope. Gastrotomies were then closed by using the Padlock-G system. Survival for 2 or 6 weeks was the primaryoutcome measurement. Secondary outcomes included ease of use, visual assessment of closure integrity im-mediately and at necropsy, presence of adhesions, evidence of infection, and histologic appearance at the closuresites.Results: All animals thrived, ate normally, and gained weight. None developed fever, tachycardia, or signs ofperitoneal irritation. Closure-site inspection at necropsy revealed excellent healing, with epithelial growth overthe Padlock-G. There were no ulcers, serosal surfaces were tightly closed, and no defects could be seen. Therewere no signs of peritoneal inflammation, intra-abdominal adhesions, or gastric spillage. Histologic evaluationshowed organizing granulation tissue with fibrosis, vascular proliferation, and mild chronic inflammatory in-filtrate (i.e., scar).Conclusions: The Padlock-G is easy to place, provides a durable closure, and allows survival animals to thrivewithout adverse sequellae. This device provides a suitable closure system for transgastric NOTES.

Introduction

Natural orifice transluminal endoscopic surgery

(NOTES) is an emerging technique where surgery isperformed without a skin incision, using a natural bodilyorifice to provide access. An enterotomy is made across thelumen of a hollow organ, and a body cavity or potential spaceis entered to perform the surgery. For intra-abdominalNOTES, the stomach is one useful portal of entry. Gastrotomyclosure is therefore a key consideration. At present, there areno rapid, reproducible, reliable gastrotomy closure methodsfor transgastric NOTES. The aim of this study was to evaluategastrotomy closure in a survival model by using a hexagonalNitinol clip: the Padlock-G. Our previously published explantand nonsurvival animal data describe the first successful

application of this device in a live animal.1 We now report theresults of our survival study using this closure system in 4laboratory swine that were survived for 2–6 weeks.

Materials and Methods

The Lock-It System� (Aponos Medical, Kingston, NH)comprises the Padlock-G, a Nitinol clip that is 18 mm in di-ameter and 0.76 mm thick (Fig. 1), as well as a scope-mounteddelivery pod (Fig. 2). The Padlock-G is folded to fit into thepod but springs back into its original shape when deployed.This locking device is six-sided and has six inner prongs thatembed into the gastric wall around the gastrotomy site. Whenthe device is deployed and snaps back into its original flator disc-like shape, these six prongs gather the edges of the

1Division of Gastroenterology, Department of Medicine, Baystate Medical Center, Tufts University School of Medicine, Springfield,Massachusetts.

2Department of Surgery, Baystate Medical Center, Tufts University School of Medicine, Springfield, Massachusetts.3Department of Pathology, Hartford Hospital, Hartford, Connecticut.

JOURNAL OF LAPAROENDOSCOPIC & ADVANCED SURGICAL TECHNIQUESVolume 20, Number 8, 2010ª Mary Ann Liebert, Inc.DOI: 10.1089/lap.2010.0076

671

Page 2: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

gastrotomy into a secure, watertight bundle, thus sealing thegastrotomy (Fig. 3).

All animals were treated humanely according to UnitedStates Department of Agriculture (USDA) guidelines. Thework was done according to a protocol approved by theBaystate Medical Center (Springfield, MA) Institutional Ani-mal Care and Use Committee. Procedures were performed ina clean operating theater intended for sterile procedures withsurvival of the experimental subject, according to USDAguidelines and regulations. Four 30- to 35-kg laboratory swinewere fasted overnight and then anesthetized with an in-tramuscular cocktail of ketamine, xylazine, and Telazol,followed by isofluorane anesthesia with endotracheal intu-bation. No paralytic agents were given. Gastroscopes weredisinfected by using a standard reprocessing regimen withglutaraldehyde. Further disinfection was performed bysoaking the endoscopes for 5–10 minutes in Betadine solutionat 1:10 dilution, and this solution was injected into, and as-pirated from, the working channel. Sterile water was used inthe reservoir used for lens washing.

All 4 animals were treated in the same way. Each animalwas treated preoperatively with penicillin G (1 cc per 20 kgbody weight). They were then placed in the supine position onthe operating table and secured with rope ties. The mouth andoropharynx were sprayed with a solution of Betadine at 1:10dilution. A previously disinfected, double-channel, thera-

peutic gastroscope (TGF-130; Olympus America, CenterValley, PA) was inserted into the oral cavity, and additionalapplication of Betadine was done to visibly untreated areas ofthe oropharynx by injecting forcefully down the endoscopicchannels under endoscopic vision. The esophagus and stom-ach were similarly treated. For the first 2 animals, a gastricovertube was used (U.S. Endoscopy, Mentor, OH). We sub-sequently dispensed with the overtube, having learned thatthe scope-mounted pod could be inserted directly into theesophagus endoscopically, analogous to placement of anendoscope-mounted band ligator. Five hundred milliliters ofdiluted Betadine solution were injected into the stomach andallowed to remain for 1 minute, while the animal was rockedfrom side to side to ensure even coating and complete treat-ment of the mucosa. As much residual debris and sawdustbedding as possible was then aspirated from the stomach,along with the excess Betadine solution.

Gastrotomy was performed by the so-called percutaneousendoscopic gastrostomy (PEG) technique, similar to theplacement of a PEG.2 After sterile prep of the abdomen withBetadine, an 18-G spinal needle was inserted aseptically intothe abdomen under endoscopic vision. A 0.035-in Teflon-coated guidewire (Tracer Metro; Cook Medical, Winston-Salem, NC) was inserted into the stomach through the needleand brought out through the right scope channel (Fig. 4). Anadditional area on the abdominal wall was sterilely prepped,

FIG. 1. The Padlock-G. The outer circle in the computer drawing is 18 mm in diameter.

FIG. 2. Two versions of the Lock-It system. (A) Early version mounted on a standard gastroscope. (B) Later version,designed for use with large- or small-diameter endoscopes.

672 DESILETS ET AL.

Page 3: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

and a Veress needle was then inserted into the abdomen. Thepneumoperitoneum was then initiated with carbon dioxide(CO2) to a pressure of 12 mm Hg.

The left scope channel was then used to place two endo-scopic T-tags (Cook Medical) transmurally in the stomach bystandard techniques. That is, a 19-G endoscopic needle loadedwith a T-tag on a 2-0 nylon suture was used to puncture thegastric wall near the guidewire, and the T-tag was deployedthrough the stomach wall into the abdomen. The fact that thewire was inserted on the anterior portion of the stomachduring PEG-type placement, and the fact that there was nowthe pneumoperitoneum lifting the abdominal wall away fromthe stomach, ensured that there was a lower risk of punctur-ing other viscera, a solid organ, or the abdominal wall duringT-tag placement. A second T-tag was placed in an identicalfashion on the opposite side of the guidewire (Fig. 4). Excessguidewire was then coiled deeply in the abdomen by pushingit through the skin puncture site.

A wire-guided, 12-mm endoscopic dilation balloon (CookMedical) was then passed over the guidewire, still in the rightscope channel, and passed through the needle tract in thegastric wall. The balloon was positioned across the gastricwall and inflated for 1 minute to 12 mm in diameter, creating a12-mm gastrotomy (Fig. 4). The balloon was deflated andremoved. Leaving the wire in place, the therapeutic gastro-scope was removed. A diagnostic gastroscope with a diame-

ter of 9.5 mm (GIF-140; Olympus America) was then passedalongside the guidewire or, in some cases, over the guidewire,and we explored the abdomen endoscopically. The guidewirewas then removed.

The Padlock-G was then loaded into its deployment pod.The loaded pod was mounted on the tip of the diagnosticgastroscope with a friction-fit adaptor, and the sutures werebrought out through the working channel with a snare. Theendoscope with attached deployment pod was then re-introduced over the sutures until the pod was abutting thegastrotomy. The sutures attached to the extramural T-tagswere then used to pull the edges of the gastrotomy physicallyinto the deployment pod under endoscopic vision. Suctionwas applied to assist with prolapsing as much gastric wallinto the pod as possible. The Padlock-G was then deployedfrom the pod by retracting the inner sheath of the pod, usingthe attached tripwire. The Padlock-G, once deployed imme-diately resumed its original shape, entrapping the gastric walllayers on its six prongs, and effectively sealing the gastrotomy(Fig. 3A). The T-tag sutures were then cut short with endo-scopic scissors, the endoscope was withdrawn, and the pro-cedure was terminated.

The first 2 animals were survived for 2 weeks prior tonecropsy. The second 2 animals were survived for 6 weeks. Atnecropsy gross examination of the abdominal cavity and ofthe closure sites, both on the mucosal and serosal sides, was

FIG. 3. Appearance immediately after deployment of the Padlock-G. (A) Endoscopic view of 12-mm gastrotomy closurewith the Padlock-G. (B) Serosal side of stomach after closure of gastrotomy with the Padlock-G.

FIG. 4. Sequential placement of guidewire, T-bags, and then balloon dilation of gastrotomy. T-bags are placed on oppositesides of the gastrotomy site to facilitate prolapse of gastric wall into the deployment pod.

GASTRIC CLOSURE WITH THE PADLOCK-G 673

Page 4: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

performed. The closure sites were then excised for micro-scopic evaluation.

Results

All animals thrived, ate normally, and gained weight. The2-week survivors gained 2.8 kg each, while the 6-week sur-vivors more than doubled in weight (>35 kg). None of theanimals developed fever, tachycardia, or signs of peritonealirritation. The stomachs were harvested, and inspection of theclosure sites revealed excellent healing with epithelial growthover the Padlock-G (Fig. 5). There were no ulcers, serosalsurfaces were tightly closed, and no defects could be seen.There were no signs of peritoneal inflammation, intra-abdominal adhesions, or gastric spillage. There were localizedserosal adhesions and adherent omentum at the closure sitesinvolving the T-tags, but not the Padlock-Gs, which were onthe mucosal side. Histologic evaluation showed organizinggranulation tissue with fibrosis, vascular proliferation, and

mild chronic inflammatory infiltrate (i.e., scar). Acute in-flammatory infiltrate was not seen (Fig. 6).

Discussion

Although one group has reported a successful porcinesurvival study where the gastrotomy was left open,3 fewwould be willing to perform this experiment in humans. It iscommon practice in experimental animals to close the en-terotomy by using endoscopic clips,4–8 but this is technicallydifficult, expensive, and does not guarantee a secure closure.Other closure methods include tissue anchors and loopedT-anchors,9–11 suturing techniques,12 self-approximating tissueflaps,13,14 and even deployment of a cardiac septal occluder,15

among others. However, these can be time-consuming ordifficult to place (mucosal flaps, endoscopic sutures), canpotentially lead to unintended damage to other organs on theopposite side of the gastric wall (T-tags), or are expensive(cardiac septal occluder). A cheaper, faster, more reliable, and

FIG. 5. Appearance of the Padlock-G at necropsy and after tissue fixation. This is from one of the 2-week survival animals.

FIG. 6. On left: low-power image of closure at 6 weeks. (A) Regeneration of normal gastric mucosa over the gastrotomy site.Running diagonally is the closure site (between lines), consisting of granulation tissue and early organizing fibrosis (B). Thedefect (C) within the muscularis propria is where a T-bag suture was removed prior to fixation. There are associated serosaladhesions (D) from adherence to the T-bag. There is a chronic lymphohistiocytic infiltrate, with focal, mild acute infiltratewithin the wound tract. Severe acute or suppurative inflammation is not seen. On right: higher power image showing fibrosisand chronic lymphohistiocytic infiltrate (wound) between muscle bundles (double arrow). There are scattered neutrophils(focal, mild acute inflammation) and some multinucleated giant cells immediately adjacent to the path of the suture (C).

674 DESILETS ET AL.

Page 5: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

secure closure method is needed before transgastric NOTESprocedures will be ready for widespread adoption.

The specific aim of this study was to determine whether agastrotomy can be closed effectively from within the stomachby using a novel, endoscopically placed device: the Padlock-Gsystem. This device relies on the shape-memory alloy prop-erties of Nitinol to gather up the tissue around the gastrotomyand pinch it firmly together into a watertight bundle. Explantdata reveal that after gastrotomy closure with the Padlock-G,burst pressures in explanted stomachs average 68 mm Hg.1 Atthese pressures the inflated stomachs are tightly stretchedand are twice their normal capacity or larger. Internal gastricpressures of this magnitude are unlikely to be generatedin vivo, and the Padlock-G should maintain integrity of thegastric wall. We have confirmed this by means of this survivalstudy.

It would have been very informative to burst test thestomachs from this experiment after a 2- or 6-week survival.However, we wanted to be certain to obtain histologic data ineach case, so burst testing could not be done. Also, in order toobtain meaningful data, several such burst tests would needto have been done. This is perhaps an experiment that can beperformed in the future if sufficient funding is available. Onecriticism of the experiment is the technique utilized to createthe gastrotomy. We chose balloon dilatation, rather thanneedle knife cautery, to open the stomach wall, as we haveobserved a higher risk of bleeding with the needle knife. Also,we have found that needle knife cautery makes the gastrot-omy irregularly shaped and thus harder to close.

In this study, we used the PEG technique to prevent inad-vertent T-tag insertion into nearby solid organs or viscera.Exiting the stomach on its anterior surface and then generat-ing the pneumoperitoneum ensures a dome of CO2 over theanterior stomach, such that needle puncture and placement ofT-tags has a lower likelihood of misplacement of T-tags. De-spite this, in 1 animal, a T-tag was still found to be embeddedin the anterior abdominal wall upon endoscopic explorationof the abdomen. This tag was removed with endoscopic for-ceps, and a new tag was placed in the proper positionalongside the gastrotomy. We had no inadvertent organpunctures in this study, as determined during abdominalexploration and also at necropsy.

Examination of the stomachs at necropsy showed that theT-tags could be a nidus for adhesion formation, althoughthere were no abscesses or other evidence of infection. There isstill a theoretic risk of inadvertent organ puncture, and, in-deed, this occurred on some of our nonsurvival studies. Wehave subsequently determined that a satisfactory closure canbe achieved by pulling the edge of the gastrotomy into thepod by using rat-toothed forceps with either a single-channelgastroscope or by using two such forceps and a double-channel scope. Closure in this fashion would be expected toeliminate the risk of errant T-tag puncture, adhesion forma-tion, and so on.

Finally, it appeared, on the 6-week survivors, that thePadlock-G tends to migrate along the surface of the mucosatoward the antrum/pylorus as it slowly sloughs or rises to thesurface. Indeed, in these animals, the Padlock-G was found3–4 cm away from the closure site, which was covered withnormal-appearing mucosa and which would be unrecogniz-able were it not for the two nylon sutures marking the spot. Itis unclear if it will slough before reaching the pylorus, and, of

course, this will depend on where it is placed. Consideringthat the Padlock-G is about the same size and thickness as aU.S. dime (i.e., 10-cent coin), and since these pass readilythrough the gastrointestinal tract in children and adults, wetheorize that the Padlock-G should pass spontaneously once itsloughs from the inner surface of the stomach. This is also tobe a subject of future research.

Conclusions

In summary, transgastric NOTES will depend on being ableto enter and leave the abdominal cavity through the stomach,and therefore closure of the gastric opening is of paramountimportance. Much of the published data involve gastric clo-sure from using endoscopic clips or T-tags. Disadvantages ofthese approaches include the cost of numerous endoscopicclips, the technical difficulties of trying to close a hole frominside the stomach, and the potentially poor mechanicalstrength of these closures. By using the the Lock-It System andthe Padlock-G, we have demonstrated that effective, durablegastrotomy closure can be achieved. The method is relativelyfast, simple, and reproducible.

Acknowledgments

Aponos Medical funded the animal labs for this study andprovided the Padlock-G System for our use. Cook Medicalprovided engineer support and some laboratory supplies insupport of this project. Olympus America donated the en-doscopes and endoscopic towers used in this study. The au-thors received no salary support from Cook, Olympus, orAponos Medical, have no stock or other interests in thesecompanies, and have no other disclosures related to this article.

Disclosure Statement

No competing financial interests exist.

References

1. Romanelli JR, Desilets DJ, Earle DB. NOTES gastrotomyclosure in porcine gastric explants with the Padlock-G clipusing the Lock-It system. Submitted for publication to En-doscopy.

2. Kantsevoy SV, Jagannath SB, Hideaki N, Isakovich NV,Chung SSC, et al. A novel safe approach to the peritonealcavity for per-oral transgastric endoscopic procedures.Gastrointest Endosc 2007;65:497–500.

3. Jagannath SB, Kantsevoy SV, Vaughn CA, Chung SS, CottonPB, et al. Per-oral transgastric endoscopic ligation of fallo-pian tubes with long-term survival in a porcine model.Gastrointest Endosc 2005;61:449–453.

4. Rao GV, Reddy N. Transgastric appendectomy. Oral pre-sentation, SAGES Annual Meeting, Ft. Lauderdale, FL, April13–16, 2005.

5. Wagh MS, Merrifield BF, Thompson CC. Endoscopic trans-gastric abdominal exploration and organ resection: Initialexperience in a porcine model. Clin Gastro Hep 2005;3:892–896.

6. Wagh MS, Merrifield BF, Thompson CC. Survival studiesafter endoscopic transgastric oophorectomy and tubectomyin a porcine model. Gastrointest Endosc 2006;63:473–478.

7. Merrifield BF, Wagh MS, Thompson CC. Peroral transgastricorgan resection: A feasibility study in pigs. GastrointestEndosc 2006;63:693–697.

GASTRIC CLOSURE WITH THE PADLOCK-G 675

Page 6: Gastrotomy Closure with the Lock-It System and the Padlock-G clip: A Survival Study in a Porcine Model

8. Hookey LC, Khokhotva V, Bielawska B, Samis A, Jalink D,Hurlbut D, Mercer D. The Queen’s closure: A novel tech-nique for closure of endoscopic gastrotomy for natural-orifice transluminal endoscopic surgery. Endoscopy 2009;41:149–153.

9. Trunzo JA, Cavazzola LT, Elmunzer BJ, Poulose BK, McGeeMF, Schomish S, Ponsky JL, Marks JM. Facilitating gastrot-omy closure during natural-orifice transluminal endoscopicsurgery using tissue anchors. Endoscopy 2009;41:487–492.

10. Sumiyama K, Gostout CJ, Rajan E, Bakken TA, Deters JL,Knipschield MA. Endoscopic full-thickness closure of largegastric perforations by use of tissue anchors. GastrointestEndosc 2007;65:134–139.

11. Desilets DJ, Romanelli JR, Earle DB, Surti VC, WillinghamFF, Brugge WR. Loop-anchor purse-string vs. endoscopicclips for gastric closure—a NOTES comparison study usingburst pressures. Gastrointest Endosc (accepted for publication).

12. Philip W, Chiu PW, Lau JY, Ng EK, Lam CC, Hui M, To KF,Sung JS, Chung SS. Closure of a gastrotomy after transgas-tric tubal ligation by using the Eagle Claw VII: A survivalexperiment in a porcine model (with video). GastrointestEndosc 2008;68:554–559.

13. Sumiyama K, Gostout CJ, Rajan E, Bakken TA, KnipschieldMA, Chung S, Cotton PB, Hawes RH, Kalloo AN, Kantsevoy

SV, Pasricha PJ. Pilot study of transesophageal endoscopicepicardial coagulation by submucosal endoscopy with themucosal flap safety valve technique (with videos). Gastro-intest Endosc 2008;67:497–501.

14. Pauli EM, Moyer MT, Haluck RS, Mathew A. Self-approximating transluminal access technique for natural or-ifice transluminal endoscopic surgery: A porcine survivalstudy (with video). Gastrointest Endosc 2008;67:690–697.

15. Perretta S, Sereno S, Forgione A, Dallemagne B, CoumarosD, Boosfeld C, Moll C, Marescaux J. A new method to closethe gastrotomy by using a cardiac septal occluder: Long-term survival study in a porcine model. Gastrointest Endosc2007;66:809–813.

Address correspondence to:David J. Desilets, MD, PhDDivision of Gastroenterology

Department of MedicineBaystate Medical Center

759 Chestnut StreetSpringfield, MA 01199

E-mail: [email protected]

676 DESILETS ET AL.