Repair of Full-Thickness Defects in Alimentary Tract Wall With Patches of Expanded Polytetrafluoroethylene

Journal List > Ann Surg > v.235(5); May 2002

Ann Surg. 2002 May; 235(5): 708–712.

PMCID: PMC1422497

Repair of Full-Thickness Defects in Alimentary Tract Wall With Patches of Expanded Polytetrafluoroethylene

Daniel S. Oh, MD,^* Melanie M. Manning, DVM,^† Janson Emmanuel, MS,^† Stuart E. Broyles, PhD,^† and H. Harlan Stone, MD^*

From the *Department of Surgery (Phoenix Campus), University of Arizona College of Medicine, Phoenix, Arizona, and †W. L. Gore & Associates (Science Center), Flagstaff, Arizona

This article has been cited by other articles in PMC.

Abstract

Objectives

To test the efficacy of patches of expanded polytetrafluoroethylene (ePTFE) for the repair of full-thickness defects in alimentary tract wall.

Summary Background Data

A recent report of successful replacement of duodenal wall with patches of ePTFE was met with skepticism and clearly warranted confirmation as well as evaluation in repair of other segments of the abdominal intestinal tract.

Methods

Defects of 4 cm² were created in various segments of canine abdominal alimentary tract (stomach, duodenum, small bowel, and colon) as well as in bladder dome. For the duodenum in 13 dogs, three different ePTFE fabrications were used: CVX (cardiovascular), PDX (preclude dura membrane), and DLM (dual mesh plus). In repair of the other areas in six dogs, the PDX patch was used. When the animals were killed, both gross inspection of the parietes and tissue for histologic study became the basis for evaluation. Peritoneal and intraluminal cultures of the specific study viscera were also taken.

Results

There were no patch failures. Only six significant adhesions were noted in 3 of the 19 dogs. Serosal surface healing was complete without exception by 1 week in all animals. Patches of CVX and PDX had heaping mucosa at the margin of well-sealed patch edges in the study involving duodenum. However, the DLM patch had an undergrowth of mucosa with partial patch separation by 1 week, beginning patch extrusion into gut lumen at 3 weeks, and total separation of patch with complete mucosal repair at 6 weeks. The fate of the PDX patches at 6 weeks in stomach, small bowel, colon, and bladder was identical to what had been observed for the PDX patch in the duodenum. All peritoneal and bladder cultures had no growth, whereas the contents of the alimentary tract grew expected flora.

Conclusions

These observations suggest that ePTFE may well be an acceptable membrane for at least temporary replacement of full-thickness hollow viscus defects, even in the face of heavy bacterial contamination, and that certain structural configurations of ePTFE may provide a base for increasing absorptive mucosal surface area.

Editorial review of a manuscript touting the successful expanded polytetrafluoroethylene (ePTFE) patch repair of experimentally produced full-thickness duodenal wall defects in the rat was initially met with disbelief. ¹ However, a supporting reference in that paper gave some credence, which thereby seemed to demand that a confirmatory study be conducted. ² Collaborative efforts were then begun between various authors of this present work.

METHODS

The initial plan included a preliminary test in the duodenum of a single pig, subsequently more detailed evaluation of three different ePTFE fabrications in the duodenal wall of dogs, and finally selection of a given type of ePTFE patch that would remain in place for at least 6 weeks for other intraabdominal hollow viscera. The pilot study was first conducted on a 45-kg female laboratory pig to test three different patches of ePTFE in the lateral duodenal wall. This was followed by phase 1, a study involving 13 mixed-breed male dogs weighing 20 to 30 kg. The final phase 2 study involved six 20- to 30-kg mixed-breed female dogs.

All animal selection, experiments, subsequent care, and sacrifice adhered to established facility guidelines and were always under the supervision of the Animal Care Committee of W. L. Gore & Associates Science Center. There was strict compliance to all regulations as set forth by the U.S. Department of Agriculture, the NIH, and the National Research Council, thereby ensuring appropriate and humane treatment of study animals.

All procedures were carried out under aseptic conditions in the research laboratory of W. L. Gore & Associates Science Center (Flagstaff, AZ). The protocols for anesthesia, postoperative care, and sacrifice were identical for all animals. Anesthesia was induced by intravenous diazepam (10 mg) and propofol (6 mg/kg). After endotracheal intubation, the animals were maintained on controlled ventilation with isoflurane and 100% oxygen. Ringer’s lactate was given intravenously throughout the operative procedure at a rate of 10 mL/kg/h.

With the animal supine, the abdomen was shaved and prepared with an iodophor. The area was draped as a sterile field, and a midline incision was made. Visceral defects were excised with scissors from the antimesenteric border of bowel and dome of bladder. Patch fixation and duodenal incision closures were performed with Gore-Tex (ePTFE) CV-5 sutures on either a PT-13 or TH-13 needle. Abdominal closure was in layers with a running suture of 2-0 polypropylene for fascia, a running 4-0 polyglycolic suture for the subcutaneous plane, and staples for skin.

Established Science Center protocol for antibiotic prophylaxis was followed. Cefazolin (30 mg/kg) was given intravenously on induction of anesthesia. A postoperative antibiotic was administered for a total of 12 days (enrofloxacin 68 mg intramuscularly every 12 hours until an oral diet could be tolerated). Thereafter, a single 136-mg capsule of enrofloxacin was given daily until the full course had been completed. Diet was standard dog chow. Other care was routine for the research facility. Animals were housed in individual cages.

The animals were killed by an intravenous overdose of barbiturate. After death the abdomen was entered using the same antiseptic precautions as at the initial operation. Both aerobic and anaerobic cultures were taken on peritoneal cavity entry and of visceral lumens at the time of excising the individual patches. The patch and its surrounding soft tissues were placed in fixative and processed for histologic study with stains of standard hematoxylin and eosin as well as Alcian blue for mucosal cells.

Patches used in the pilot study and in phase 1 were CVX (cardiovascular), PDX (preclude dura membrane), and DLM (dual mesh plus). For phase 2, only the PDX fabric was inserted. In the pilot study patches were 1 cm²; in phases 1 and 2 the fabric area was increased to 4 cm².

For the pilot study, all three of the different patches were used to replace 1-cm² full-thickness duodenal defects, beginning 5 cm distal to the pylorus and then spaced every 5 cm beyond. The animals were killed at 21 days.

In phase 1, the three different patches were used to replace 4-cm² antimesenteric defects in the duodenum at 10-cm intervals, beginning 10 cm distal to the pylorus and randomly alternating with a fourth site of a suture-repaired 2-cm linear incision. Four dogs were killed at 1 week, another four at 3 weeks, and the final five at 6 weeks.

In phase 2, involving six dogs, a single PDX patch was used to replace 4-cm² defects in anterior stomach wall, an equal antimesenteric gap in small bowel as well as transverse colon, and bladder dome. All animals were killed at 6 weeks.

RESULTS

In the pilot study, no leaks were noted, nor was there any evidence of peritonitis. Serosa had healed over the peritoneal surface of each patch, whereas on the lumen side the fabric appeared as the base of an ulcer for the CDX and PDX fabrics. The DLM patch was partially separating. All three areas were firm in repair. Until the animal was killed, the animal had eaten well and had displayed no unusual behavior.

For phase 1, there was no evidence of leak or peritonitis, the patch was well fixed, and serosa had overgrown the peritoneal surface of all patches without exception, even after only 1 week. To the 26 patches of either CVX or PDX there were only two significant adhesions. On the lumen side, mucosa was heaped at the patch edge for a distance of 1 to 2 mm, leaving the remainder of the fabric exposed as an ulcer base. Histologic study confirmed these gross observations; there was no mucosal ingrowth beyond the heaped-up edge (Fig. 1). Gram-positive cocci and rods as well as gram-negative rods were seen on Gram stain within the mucosa at the defect edge.

Figure 1. Section of duodenal repair with a patch of CVX (cardiovascular); heaping mucosal edges without mucosa healing are obvious (42-day retrieval).

Similar findings were noted with the DLM patch on the peritoneal side. However, mucosa had proliferated beneath the patch within the lumen by partial ingrowth at 1 week. There was almost complete mucosal healing beneath the patch by 3 weeks (Fig. 2), with one patch having been extruded, and total healing on both sides was noted with absent patch at 6 weeks. However, on histologic examination, mucosa was seen merely to be a cell or two thick and failed to have developed any configuration suggestive of normal villi (Fig. 3).

Figure 2. Section of duodenal repair with a patch of DLM (dual mesh plus); mucosa is proliferating beneath the patch (21-day retrieval).

Figure 3. Section of duodenal wall after DLM (dual mesh plus) patch extrusion; stain is Alcian blue of mucosa only a single cell thick (42-day retrieval).

Each repaired duodenal incision had healed without adhesion formation. All peritoneal cultures were negative, and isolates grown from lumen samples were routine expected flora.

In phase 2, all patches of PDX had firmly healed without evidence of peritonitis when the animals were killed at 6 weeks. There were adhesions to the small bowel patch in two animals and a single adhesion to the colon in one of these. Serosal healing was uniformly complete. On the lumen side, margins of heaped-up mucosa for a distance of 1 to 2 mm were noted around a central ulcerlike base of patch in stomach, small bowel, and colon; this was identical to what had been noted with the CVX and PDX patches in phase 1. A similar histologic appearance was noted with respect to the bladder. Samples of peritoneal fluid and urine gave no growth.

DISCUSSION

Many axioms have existed for ages in medicine and many are certainly still valid, yet even the most sacred cow may at times warrant a challenge. Often progress can take place only when an irrefutable dogma is questioned. That obviously is the case for replacement of a portion of gastrointestinal wall by a foreign body.

The present study and others have clearly shown that a water-impenetrable fabric can function, at least for short periods of time, as a reliable barrier against the escape of intestinal contents. ^1–4 An open mesh would without doubt fail. In addition, placement of sutures to fix the material must also prevent leak of chyme and digestive juices through the resultant perforating holes. Synthetic prostheses and needle-swedged sutures of ePTFE appear to satisfy both requirements. When using this particular material, no significant leaks have been reported to date. ^1–4 Complete resurfacing by serosa has been observed to occur within a few days from the time of ePTFE patch insertion as replacement for full-thickness defects in gastrointestinal wall. The same has been true for intraperitoneal urinary bladder. ⁵ Even when a patch has partial separation, the mesothelial layer remains intact and is reliably confining to luminal contents.

Thoroughness of mucosal regrowth, however, cannot be substantiated. Careful review of prior reports failed to disclose accurate details as to exactly which ePTFE fabrication was tested. ^1–5 Indeed, there did not appear to have been actual mucosal resurfacing over the ePTFE prostheses in what was displayed in the microscopic illustrations already published or in our own work. ^1–4 Heaping mucosa at the margins of the insert seemed to fold over toward the middle for a distance of 4 mm or so, and thus a mucosal clump from the opposite side was met toward the center of the patch, thereby giving an illusion of complete luminal healing by mucosal overgrowth. However, a wider patch of 2 cm across prevented this abutment and thus obviated such a false impression. Mucosa never extended more than 5 mm beyond its supporting vascularized bed in our own preparations. As a result, with a wider prosthesis the central area became the base of what appeared to be a punched-out ulcer.

Hopes for inserted patches of ePTFE to provide a matrix for mucosal overgrowth and thereby to increase absorptive surface seem to be dashed if the prosthesis is retained in bowel wall. Nevertheless, changes in fabrication of the material can evoke an entirely different mucosal response. CVX and PDX patches have remained in position as placed, yet the DLM product uniformly allowed mucosal proliferation beneath and eventually led to patch extrusion into bowel lumen. The result left an intact serosa, a narrow area of scar between opposing muscularis, and a one- or two-cell layer of mucosa. There was absolutely no villus formation, and accordingly any increase in mucosal absorptive surface was quite small at best. It therefore appears that ePTFE, as currently fabricated, is not the matrix for mucosal area expansion, yet the fact that a synthetic prosthesis can be accepted into bowel wall and has some durability is at least a major step forward.

Why the several fabrications of ePTFE patches evoke different responses in mucosal repair is not immediately apparent. Each fosters serosal healing within a matter of only a few days, but mucosal repair is another matter. Porosity and material thickness do not seem to influence these discrepancies (Table 1). However, when an inert middle layer of elastomeric fluoropolymer is sandwiched between thin sheets of ePTFE, there has been neither mucosal overgrowth nor mucosal proliferation beneath the patch. Perhaps some future fabrication may permit the desired repair of mucosa to occur.

Table 1. PATCHES OF ePTFE

Possibly more immediate as well as more promising applications of ePTFE for a bowel patch include closing of the difficult duodenal stump, replacement of sizeable areas of traumatized duodenal wall, and perhaps in small bowel stricturoplasty. Animal experiments to evaluate its practicality for esophageal lengthening also appear to be viable options. However, advances in ePTFE fabrication might even bring into consideration its use for repair in other areas.

Discussion

Dr. Timothy C. Fabian (Memphis, TN): This study is obviously quite provocative. Dr. Stone and his colleagues have once again challenged conventional wisdom. How is it possible that a nonabsorbable foreign body could be used to establish integrity of the GI or GU lumens? Ridiculous, anathema, heresy! This sounds as crazy as abbreviating a laparotomy in the face of shock or sepsis by placing a prosthesis or a zipper in the abdominal wall for closure and coming back to fight on a more physiologically even footing; or not exteriorizing a colon wound in an acutely injured patient. The latter two concepts were widely believed to be reckless approaches which violated established surgical principles when they were initially proposed. Well, they are now standards of care which were initiated by Dr. Stone.

Will alimentary tract substitution by PTFE ever become clinically adopted? I don’t know. But certainly the possibilities are intriguing based on the results that he presented here today. I would like to ask Dr. Stone a few methodologic questions related to the study.

Were the fascia secured with running or interrupted suture techniques?

Second, the dual mesh was proven to heal the mucosal surfaces completely, whereas the cardiovascular and dura membrane prostheses had incomplete healing with small central ulcerations over the PTFE. Why didn’t you use the dual mesh in the phase 2 studies? And along that line, if you were going to use this patch clinically to cover a duodenal defect with patch material, what would you use today? Regarding species selection, after wrestling with the 100-pound pilot pig, did you decide to change to the canine model?

Could you please speculate briefly on three areas? First, what about using this for common duct replacement, as that is a common clinical problem? Second, what about somewhat larger defects? And third, what about adjunctive use of things such as growth factors in such models?

I would like to thank the Association for allowing me to discuss the paper. I have long felt that Dr. Stone was often a decade ahead of his time, ground which can be tricky to tread on. And as Samuel Clemens once said, “The man with a new idea is a crank until the idea succeeds.”

Dr. Maurice J. Jurkiewicz (Atlanta, GA): With this presentation, Harlan Stone has reinforced his standing as a world-class surgical iconoclast.

In my interaction with Harlan Stone well over a decade in the 1970s, I had a very rich and rewarding experience. Harlan Stone is a consummate generalist. He has a probing mind and has a well-developed sense of challenging established surgical dogma.

As one who has dealt extensively with holes in the upper aerodigestive tract, my initial response was exactly like his, that this was in fact a ridiculous idea. As I read more in the manuscript and reflected on it, there may well be room for such approach in fixed areas in the oral pharynx and the upper aerodigestive tract.

The palate is such a fixed area. Cleft palate, oral-antral fistula after resection of cancer—difficult areas to repair, particularly the latter. The mucosa is not as specialized as the mucosa of the lower gastrointestinal tract. Squamous epithelium might be able to proliferate over such a patch. However, in the remainder of the upper aerodigestive tract there is considerable mobility in the act of swallowing, for example, and second, juxtaposition to major vessels in the neck.

An approach was attempted in the past with a single-stage reconstruction of the cervical esophagus in 1952. However, the net result was erosion of the tantalum mesh covered with a skin graft into the vascular system, which resulted in exsanguinating hemorrhage. This leads me to certain questions.

In mobile areas of the gastrointestinal tract, lower down in the abdomen, you showed heaped-up mucosa at the edges of the patch. I wonder if this heaped-up mucosa could not act as a focus for intussusception because of its polypoid configuration.

A few years ago I was interested in short-gut syndrome. We developed a model in dogs where we created a blood supply to the antimesenteric border of the small intestine. And then the idea was to use skin expanders, modified, of course, put them in, and try to expand the gut to double its width. What happened? This was a 100% model for the creation of intussusception. Will a patch of Gore-Tex in a mobile portion of the intestine act as a focus for intussusception?

Again, a very, very intriguing presentation, Harlan. Thank you very much for bringing it to our attention. One last question. What is your next step? What are you going to do next?

Dr. Hiram C. Polk, jr. (Louisville, KY): For 35 years Dr. Stone has been my best friend. And he has spent almost all of that time a decade ahead of the rest of thought in American surgery. That includes all of you. He really has been innovative and creative in the extreme.

He also, from a Journal editor’s point of view, is almost the only person you can send an off-the-wall manuscript to and know that it will get a straight evaluation. Back to the Journal editor issues, it has been mentioned many times, and Harlan is kind enough to give credit to AJS for the creation of this.

There are two things worthwhile. We have spent so much time dealing with petty salami-slicing, but how to deal with really, really creative off-the-wall kind of proposals doesn’t get much attention. It is interesting that this is the direct result of the decision of the lady who was president of Turkey nearly a decade ago, to insist that the only way in Turkish medical schools you can now achieve something like tenure is through English language, preferably American medical publications. So we see lots of things from people who are now not the traditional staged sort of American approach to given things. This is where this came from. And Harlan, of course, is the ideal person to try to take that on.

I also want to say how important he is as the second, or third, if you will, concurring report as to the utility of this material. It is always the concurring report that is more important than the initial one. The report by Bill Ledger 30 years ago about confirming our observations on the value of prophylactic antibiotics is the most important paper I ever read. It is the confirmation that is much more important than the original report. And the fact that someone else can critically do this in a respected environment is really special.

The hard part for Harlan and all of us is the fact that this is bought and paid for by Gore-Tex. You lack any degree of independence and any degree of being able to tackle this straight up. The Journal editors are beginning to grapple with the pervasive effect of industry in these kinds of endeavors, and it is extraordinarily hard for even the most moral and straight-minded of us to separate ourselves from those obligations. Harlan made it straight up front what he has done, and sometimes you simply can’t deal with this other than to be extra-careful to admit a potential conflict exists.

The final negative point about this is the ungodly cost of this material, which is totally unjustified. And if it is going to be used, it needs to be dealt with in something like a sensible way. This is first-class work, but it has been typical of what Dr. Stone has done for three decades. My congratulations.

Dr. David V. Feliciano (Atlanta, GA): You put these patches in heavily vascularized structures. When PTFE patches were used to replace bile duct holes in the dog lab, they all failed with stricture and dehiscence. And I am wondering if your success may be unique to GI tract replacement because of the vascularity.

Second, there are tremendous species differences in the results that you get when you put in prosthetics such as vascular tube grafts. Do you think that is going to be a factor here, and maybe you ought to move up to higher mammalian models as the next step?

Dr. Max R. Langham, JR. (Gainesville, FL): I enjoyed this a great deal. I wanted to ask Dr. Stone whether he thinks that Gore-Tex as a material is as important as the concept of using a patch. Mike Chen in our group has used SurgiSys, which is a collagen matrix, to do very similar work in a dog esophagus. This has been published in the Journal of Surgical Research, and I think that the concept is a great one. I was just curious as to whether the material is the important thing or if an absorbable material might accomplish the same goal.

Dr. H. Harlan Stone (Phoenix, AZ): First, I want to thank the discussants for staying so late in the morning for my show and tell.

Dr. Fabian, you had a number of questions and suggested application to different animals. That must be a dig at me, because when you were a fellow in Atlanta, I conned you into being immunized against rabies. The monkeys we were to use had been in the test for efficacy of a new vaccine against rabies. Thus, we both were immunized against rabies. As you will recall, word soon went around that our barks were far worse than our bites.

We did use running sutures. Patch selection was based on our not wanting to have a patch extrude and be loose in the bladder. There was one animal we thought might have had dysuria. We gave it a single dose of Pyridium and never did it show such signs again.

I think to close a duodenal defect or stump I probably would use a DLM patch. A Colombian surgeon told me he had used one to close a duodenal stump and another to correct a stricture in the duodenum due to ulcer. Both were reputed successes. So I think it could have some human application in this area.

Regarding its application to the bile duct, I don’t know. I think that may depend on the patch chosen. I would worry if the patch were to extrude into the lumen and thereby create a problem with obstruction, that is unless you had done a papilloplasty before or the patient were later to have an endoscopic papillotomy.

I don’t know how large a defect can be safely spanned. We went to 2-cm squares, which is 4 cm². We would like to test larger areas as well as insert one circumferentially for a distance. These extensions are on the planning board. I don’t know about growth factor; it may indeed play a role.

At Emory, Dr. Jurkiewicz was always one to ask a question at Grand Rounds for which you never had an answer. I thought he would be the perfect one to put this to other uses. This perhaps might offer a readily available modality for reconstruction of the oral pharynx. Its use for a palate reconstruction certainly should be pursued. I don’t know about tantalum mesh. It certainly did provide many around-the-world trips for a famous Baltimore surgeon.

With respect to intussusception, that indeed could be a problem. We see such occur with polyps. None of our animals had it complicate their course, but they weren’t followed that long. If one used the DLM patch, then they would not have the heaped-up mucosa.

I don’t know what the next step will be, other than we are planning to replace esophageal segments and repair divots taken out of the trachea.

Hiram, thank you for your kind comments. After all, it was your sending me this article to review that started it all. The Gore people have been very supportive of our effort. I had control over everything except one item, that being the administration of the antibiotics. Their protocol was such that the animal had to receive 12 days of enrofloxacin. For those of you who are not up on enrofloxacin, it is veterinary Cipro. Otherwise, they gave me total freedom.

David, I mentioned before about the common duct. You may have used a differently fabricated product, because DLM is a new mesh. That may well have made the difference. I think the patch would be great for certain procedures, such as stricturoplasty, thereby obviating the need to do a Roux-en-Y in the difficult patient.

As far as species differences, there probably are. We noticed a different response between what was reported for rabbits, dogs, and the pig, but these also were different patch sizes. As far as other fabrics, I do not know that much about them. I am sure there are other materials that could be used, but I am just not familiar with any of them.

Again, I want to thank you for staying so late this morning—and particularly for my son, who must soon catch a plane to return to Phoenix.

Footnotes

Presented at the 113th Annual Session of the Southern Surgical Association, December 3–5, 2001, Hot Springs, Virginia.

Supported by an in-house grant from W. L. Gore & Associates, Flagstaff, Arizona.

Correspondence: H. Harlan Stone, MD, Phoenix Integrated Surgical Residency, Suite #619, 1300 North 12th Street, Phoenix, AZ 85006.

Accepted for publication December 2001.

References

Astarcioglu H, Kocdor MA, Sokmen S, et al. Comparison of different surgical repairs in the treatment of experimental duodenal injuries. Am J Surg 2001; 181: 309–312. [PubMed].

Ozlem N, Erdogan B, Gultekin S, et al. Repairing great duodenal defects in rabbits by ePTFE patch. Acta Chir Belg 1998; 98: 17–21.

Caga T, Gurer F. Polytetrafluoroethylene patch grafting for closure of stomach defects in the rat. Br J Surg 1993; 80: 1013–1014. [PubMed].

Watson LC, Friedman HI, Griffin DG, et al. Small bowel neomucosa. J Surg Res 1980; 28: 280–291. [PubMed].

Chamorra MV, Meli BG, Casado JS, et al. Estudio experimental de la ampliacion vesical con lamina de Gore-tex: aspectos biomecanicos, bioquimicos y biostructurales. Arch Esp de Urol 1994; 47: 958–966.