Small Intestinal Submucosa for Vascular Reconstruction in the Presence of Gastrointestinal Contamination

Abstract

Introduction:

Surgical options for vascular reconstruction in a contaminated field are limited and include prosthetic reconstruction or ligation with extra-anatomic bypass. With prosthetic insertion, rates of graft infection and failures (pseudoaneurysms and thrombosis) are high. In the emergent situations, extra-anatomic bypass is time-consuming and complex, and it produces marginal long-term results. Small intestinal submucosa (SIS) is a cell-free collagen matrix derived from porcine small intestine. Preliminary studies have demonstrated its ability to be remodeled into host tissue. In this study, we compared SIS to polytetrafluoroethylene (PTFE) as a vascular patch for arterial repair in the presence of massive gastrointestinal contamination to evaluate graft patency, incorporation, infection, and aneurysm formation.

Methods:

Adult mongrel pigs underwent general anesthesia with Isoflurane and were then randomized to 1 of 3 groups: control, contamination (colon puncture with stool contamination of the pelvis), or shock + contamination (40% blood volume for 1 hour, then resuscitation with shed blood and crystalloid, plus contamination). All groups then underwent a left common iliac arteriotomy and further randomized to a 1 × 3–cm patch angioplasty with either SIS or PTFE. All received cefotetan for 24 hours. All animals were sacrificed between 2 and 4 weeks, and necropsy was performed. Grafts were cultured, and microscopic analysis with hematoxylin and eosin and trichrome was performed. Outcomes included pulse quality (normal or diminished) compared with opposite side, graft infection, and pseudoaneurysm; all were determined by a blinded investigator.

Results:

Forty animals were randomized, and 1 died of abdominal sepsis. All control animals had normal distal pulses, no pseudoaneurysms, and no patch infections. The pseudoaneurysm rate for the contaminated PTFE patches was 25% compared with 0% in the SIS group (P = 0.09). Patch infection occurred in 73% of all PTFE patches compared with 8% of SIS patches (P < 0.03). Organisms present in the infected grafts included Escherichia coli, Bacteroides species, and other Gram-negative enterics. Histopathology demonstrated the presence of neointima in both SIS and PTFE. Only SIS was completely incorporated, with infiltration of collagen fibrils and lymphocytes.

Conclusions:

SIS was associated with improved graft patency, less infection, complete incorporation, and no false aneurysm formation when compared with PTFE. This may be due to its ability to provide a durable scaffold for cellularization and tissue remodeling. This material may offer a superior alternative to more complex vascular reconstruction techniques in contaminated fields.

Prosthetic grafts are prone to infection, thrombosis, and pseudoaneurysm formation when placed in contaminated fields. The authors compared porcine small intestinal submucosa (SIS), a cell-free collagen matrix, with polytetrafluoroethylene (PTFE) as a patch angioplasty in the face of gastrointestinal contamination. SIS was associated with improved graft patency, less infection, no pseudoaneurysm formation, and complete incorporation compared with PTFE.

Surgical options for vascular reconstruction in a contaminated field are limited. If primary repair is not possible, the choices include autogenous or prosthetic reconstruction or vessel ligation with extra-anatomic bypass. Autogenous vein is not always available or may be of poor quality. In addition, it is not completely resistant to infection and is associated with pseudoaneurysm formation and rupture when used in contaminated fields.^1–3 Prosthetic insertion in the face of bacterial contamination is associated with high rates of graft infection, leading to pseudoaneurysm formation, graft thrombosis, or disruption with catastrophic hemorrhage.^3–5

Clinical scenarios that may require emergent vascular reconstruction in the presence of bacterial contamination include graft infections with false aneurysms, arterioenteric fistulae, and traumatic vascular wounds with associated hollow organ injuries. The trauma patient with penetrating colon injury, massive fecal contamination, hemorrhagic shock, and major vascular injury is perhaps the most difficult to manage with a good functional outcome. A prosthetic material that is resistant to infection, readily available, and reliable would be ideal in this situation.

Biologic products for vascular repair may provide a viable alternative to prosthetics or extra-anatomic bypass. Small intestinal submucosa (SIS) is a cell-free collagen matrix that is harvested from porcine small intestine. It provides a scaffolding for cellular migration and early capillary ingrowth. Preliminary studies have demonstrated that SIS is remodeled into host tissue.^6,7 It has been used in humans for various indications, including hernia repair and urologic reconstruction procedures. In addition, initial studies using SIS as a vascular conduit have shown early endothelialization and high patency rates.^7,8 However, there are scant data regarding its utility for vascular reconstruction in contaminated fields.

The purpose of this study is to compare SIS to a standard vascular prosthetic—expanded polytetrafluoroethylene (PTFE)—as a vascular patch for arterial repair in the presence of gastrointestinal contamination. To accomplish this, we conducted an animal study designed to mimic the most severe conditions of massive gastrointestinal contamination and hemorrhagic shock.

METHODS

Mongrel pigs weighing 20 to 40 kg were used and acclimated for 7 days in the Animal Care Facility at The University of Tennessee Health Science Center before the surgical procedure. The experiment was approved by the Animal Care and Utilization Committee of the University of Tennessee Health Science Center. All care complied with the Principles of Laboratory Animal Care and the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Bethesda, MD).

Animals underwent induction of anesthesia with an intramuscular injection of ketamine and xylazine, and general anesthesia was maintained with isoflurane. Right carotid artery and right external jugular vein catheters were inserted. The animals were monitored continuously during surgery with ECG, pulse oximetry, and continuous arterial blood pressure measurements. All animals received perioperative antibiotics (cefotetan), intramuscular injections of 1 g preoperatively and 1 g at 12 and 24 hours postoperatively.

After instrumentation, the animals were randomized into 1 of 3 groups. The experiments were designed to mimic the clinical scenario of penetrating colon and major vascular injury. In the control group, animals underwent celiotomy and left common iliac arteriotomy and 1 × 3–cm patch angioplasty with either SIS (Surgisis; Cook Surgical, Bloomington, IN) or standard PTFE (Gore-Tex; W. L. Gore and Associates, Flagstaff, AZ). There were 2 experimental contamination groups. In 1 contamination group, animals had the same operative procedures with patch angioplasty plus sigmoid colotomy with spillage of stool into the pelvis and on the patch angioplasty. The stool was left in place for 15 minutes and then irrigated away with 1 L of normal saline. The colotomy was closed in a standard 2-layered fashion. The other contamination group, shock plus contamination, underwent celiotomy, patch angioplasty, fecal contamination, and 40% arterial hemorrhage shock for 1 hour. The animals were resuscitated with shed blood and crystalloid to normalization of heart rate and blood pressure. After abdominal closure, all animals were awakened, returned to their cages, and allowed food and water ad libitum during the recovery period. During this recovery period, each animal was examined twice daily by a blinded investigator for evaluation of limb viability and characterization of distal pulses. Pulses in the left hind leg were compared with the right hind leg and defined as 2+ (equal in both hind legs), 1+ (diminished in the left compared with right), or 0 (absent in the left).

Animals recovered for 2 or 4 weeks to determine degrees of patch incorporation at 2 distinct time points. After this recovery period, animals were anesthetized and underwent celiotomy. The left common iliac artery patch angioplasty was inspected for the presence of pseudoaneurysm (defined as partial anastomotic disruption with a defined blood-filled cavity), cultured, and excised en bloc and placed in 10% buffered formalin solution. Patch infection was defined as lack of incorporation with surrounding tissue with either positive cultures or a collection of pus in direct contact with the patch. Animals were then sacrificed with overdoses of pentobarbital and potassium chloride. The specimens were allowed to fix for 7 days. Sections were made, embedded in paraffin, and cut into 6-micrometer sections. Each section was stained with hematoxylin and eosin (H & E) and trichrome. Slides were inspected and interpreted by a pathologist who was blinded to the treatment group.

RESULTS

A total of 40 animals were randomized: 16 to the control group, 16 to the contamination group, and 8 to the shock plus contamination group (Fig. 1). One animal in the contamination group died on the fourth postoperative day due to abdominal sepsis. All remaining animals survived until necropsy at either 2 or 4 weeks.

FIGURE 1. Study design. *One animal died of abdominal sepsis on the fourth postoperative day.

A blinded investigator assessed each animal twice daily for general condition, limb viability, and distal hind leg pulses. There were no instances of ischemic complications in any animal. There were no pulse discrepancies in either the SIS or PTFE control animals at any time. In the contamination groups, all 8 SIS animals had normal pulses, but only 3 of 7 with PTFE had normal pulses. The shock plus contamination groups had similar findings: no pulse discrepancy in SIS and 2 of 4 in PTFE animals. Since the addition of shock to the model did not seem to affect the pulse measurements, all animals with contamination were combined. In these groups, all SIS (12 of 12) animals had normal pulses, and 5 of 11 with PTFE had normal pulses (100% vs. 45%; P < 0.05).

Similar analyses were performed regarding the presence of a pseudoaneurysm and presence of graft infection. No pseudoaneurysms were seen in any animals in the control group. For the contamination group, no SIS animal had a pseudoaneurysm and 2 of 7 with PTFE had pseudoaneurysms (Fig. 2). One PTFE animal in the shock plus contamination group had a pseudoaneurysm. Combining the contamination groups, no SIS and 3 of 11 (27%) PTFE animals had pseudoaneurysms (P = 0.09).

FIGURE 2. Gross specimen demonstrating severe narrowing in (a) the contaminated PTFE group compared with the (b) patent SIS explant. The arrow denotes a pseudoaneurysm.

No patch infections were seen in either group of control animals, and 1 infection (Acinetobacter species) was present in the SIS contamination group. However, 5 of the 7 PTFE contamination group had infection (13% SIS versus 71% PTFE; P < 0.04). Cultured organisms included Escherichia coli, Bacteroides species, Streptococcus viridans, Acinetobacter species, Lactobacillus species, enterococcus, and diphtheroids. There was also a statistical difference between SIS and PTFE in the combined contamination groups (8% versus 73%; P < 0.03). In the presence of fecal contamination, patch angioplasty with PTFE was associated with diminished distal pulses, development of pseudoaneurysm, and perigraft infection when compared with patch angioplasty with SIS.

Histopathology

All specimens were examined by a pathologist who was blinded to study group. Specimens were stained with H & E and trichrome. H & E demonstrated that all SIS and PTFE patches were endothelialized after 2 weeks regardless of study group or type of patch. All SIS patches were completely incorporated by host tissue at 2 and 4 weeks regardless of study group. However, no PTFE patch from any group was incorporated by native tissue at either time point (Fig. 3). There was also infiltration of lymphocytes and fibroblasts with proliferation within the SIS patches; none of which was seen in the PTFE patches. In the contaminated groups, there was not an intense inflammatory reaction with neutrophil infiltration in the SIS groups, but neutrophil migration was seen in the PTFE groups. Trichrome staining demonstrated the deposition of new collagen and elastin within the SIS patches but none in the PTFE patches. After 4 weeks, the SIS patch was nearly unrecognizable from native vessel due to this infiltration of collagen and elastin (Fig. 4).

FIGURE 3. H & E slide from contamination group demonstrating lack of native tissue incorporation of the (a) PTFE patch compared with the remodeling of (b) the SIS patch into native vessel. The lumen is inferior.

FIGURE 4. Trichrome staining of both patches from the contamination group demonstrating lack of ingrowth in (a) the PTFE patch compared with fibroblast and elastin infiltration in (b) the SIS patch. The lumen is inferior.

DISCUSSION

The ideal vascular graft should be strong, pliable, readily available, and resistant to infection and possess long-term patency. Unfortunately, such a graft remains elusive. Available grafts, including Dacron, PTFE, autogenous vein, homologous vessels, and heterografts, all have advantages and disadvantages. None meet the above criteria for the ideal graft.

Autogenous vein has been considered the first choice for arterial reconstruction in the presence of contamination. A 1972 report by Rich and Hughes⁴ documented significant complication and amputation rates in patients with prosthetic arterial repair. Bricker et al⁹ questioned the superiority of vein in an experimental model. While Dacron grafts had lower patency rates, 10% of animals with infected vein grafts died of exsanguination. In the presence of infection, poor results have also been seen with arterial homografts and bovine heterografts.^10,11 Thus, while there were problems using autogenous vein for arterial reconstruction in contaminated fields, the overall results were better than those seen with Dacron, homografts, or heterografts.

This tendency for vein graft dissolution with subsequent hemorrhage led investigators to advocate the use of PTFE in contaminated wounds.2,3,12–14 This prosthetic material was thought to be an ideal conduit because it is inert, nonthrombogenic, and impervious. Initial enthusiasm was followed by later reports that identified problems with PTFE, especially in contaminated wounds. Feliciano reported a 5 year experience with PTFE grafts in vascular wounds and documented lower patency rates when compared with vein grafts.¹⁴ The PTFE grafts were not resistant to infection when exposed to significant contamination, although these grafts seemed to tolerate contamination better with aggressive wound care and antibiotic therapy. Similar relatively poor responses to contamination were reported by others.^5,15 When infected, a lack of tissue incorporation and spotty neointima formation present in the PTFE grafts were noted.⁵ While vein tended to necrose with subsequent hemorrhage in the presence of infection, PTFE was prone to thrombose–especially when smaller (4 mm) grafts were used. In the present study, there was obvious lack of incorporation of the PTFE patch in all contamination groups. In addition, thrombosis was significantly more frequent in the contaminated PTFE groups. In nontraumatic instances of graft infection, PTFE appears to be more resistant than Dacron, and has been used as an in situ replacement of infected arterial prostheses.^16,17

While arterial substitutes in open extremity traumatic wounds are problematic and may result in limb loss, major abdominal vascular wounds are usually life threatening. Iliac arterial injuries are particularly difficult to manage due to location and frequency of associated hollow organ injury and contamination.^18–20 Principles of management include vascular control and control of intestinal spillage. Carrillo reported prosthetic iliac arterial repair in 17 patients, and 8 had associated colon wounds.¹⁹ There were no graft infections. There were 3 patients with massive colonic contamination who had extra-anatomic bypass grafts with 1 death in a patient with delayed reconstruction. Feliciano et al¹⁴ reported 7 patients with PTFE reconstruction after iliac artery injury, with 1 asymptomatic graft thrombosis, but there was no mention of enteric contamination. Similar results of higher rates of thrombosis and diminished distal pulses with PTFE were seen in the present study. Others have suggested ligation with ipsilateral fasciotomy in the presence of massive contamination and extra-anatomic bypass if signs of ischemia develop.^21,22

Since infection remains the most significant problem following arterial prosthetic insertion, other tissue substitutes have been developed, including urinary bladder submucosa, cadaveric fascia, amniotic membrane tissue, and acellular dermis.^23–26 Another promising material is small intestinal submucosa, which is derived from porcine jejunum. It is primarily an acellular collagen matrix that provides a scaffold for tissue regeneration. After implantation, it elicits a host-tissue response for angiogenesis, tissue deposition, and ultimately restoration of structure and function that is specific to the implantation site.⁶ The mechanism for this host response is likely mediated in part by the presence of fibroblast growth factor and transforming growth factor-β, which have been identified in SIS extracts.²⁷ The material has been used clinically in various patient populations with promising initial success. Favorable results have been seen when used for repair of large ventral hernias,²⁸ paraesophageal hernias,²⁹ and as a buttress for the gastrojejunostomy in laparoscopic bariatric procedures.³⁰

In the present study, the animal model was designed to mimic the scenario of penetrating colon and iliac arterial injury with massive gastrointestinal contamination. Since PTFE is the most common prosthetic used in the face of contamination, it was compared with the newer biomaterial SIS. In animals with the SIS patches, there was no vessel narrowing, no false aneurysm formation, and substantial resistance to infection. The acellular matrix proved to be an excellent scaffold for arterial remodeling. Histopathology demonstrated complete incorporation in all SIS implants regardless of the degree of shock or amount of contamination. This resistance to infection displayed by SIS has previously been demonstrated.³¹ Extracts from SIS exhibit antimicrobial activity and can inhibit bacterial growth for at least 13 hours. This property has also been shown in a canine model of orthopedic soft tissue repair.³²

The results of the present study must be interpreted with caution. The superiority of SIS over PTFE patch angioplasty in the presence of massive gastrointestinal contamination does not necessarily translate into similar results when each material is used as a conduit. Preliminary work regarding SIS as a vascular graft is encouraging. When used as an autograft, allograft, or xenograft, there was compatibility and high patency rates in the aorta, carotid and femoral arteries, and superior vena cava after 28 days.⁷ Another study using 5-cm segments of a 4-mm graft demonstrated 89% patency up to 9 weeks.⁸ It is clear that more work is necessary before this biomaterial is ready for widespread use. Since the material acts as a scaffold, there may be a limit to its effective length as a graft. In addition, the effect of contamination by organisms that produce collagenase, such as Pseudomonas species, must be tested. Nonetheless, these preliminary results are quite promising.

Discussions

Dr. David V. Feliciano (Atlanta, Georgia): Despite the extraordinary progress in vascular surgery over the past 50 years, there are unsolved problems, including the choice of patch in contaminated and already infected fields. This is a particular problem with abdominal gunshot or stab wounds involving the common or iliac arteries as most have associated enteric or colonic wounds with contamination.

Dr. Croce has nicely demonstrated in this study that the outcome of 1-by-3 centimeter patch angioplasties with a porcine SIS versus standard PTFE is incredibly different if these grafts are exposed to stool for 15 minutes. In all categories of postoperative results including palpation of left hind limb pulses, presence of pseudoaneurysms at the patch site at autopsy, presence of patch infection, and patch incorporation by host tissue, the SIS patch was far superior to the PTFE patch. These striking differences prompt some comments and questions.

Dr. Croce, first, I understand that your model was directed at creating infections in prostheses as they are so susceptible, and it was most successful in the PTFE group. In the clinical setting an iliac PTFE patch or tube graft inserted in a contaminated field would be thin-walled to encourage incorporation. Secondly, the area would be irrigated with 4 to 5 liters of saline containing antibiotics. And thirdly, the patch or graft would be wrapped in a viable omental pedicle containing milky spots.

My first question is, therefore, did you cover these patches in any way or were they continuously exposed to pelvic cellulitis or a separate pelvic abscess? If so, this does not nearly mimic the clinical situation.

Secondly, where did the new collagen and elastin come from when you did the trichrome staining of the SIS patches? Are there actually stem cells in porcine intestinal submucosa? Are there viable collagen or elastin progenitors in the processed SIS product? Or do you really believe that these are secreted by the new presence of fibroblasts invading the patch?

Third, in the manuscript you touched on something that you didn't mention in the presentation, and that is there is an implication that there is an antimicrobial substance in processed SIS patches. Is this a new class of antibiotics? Can you clarify this for the audience?

And fourth, with these encouraging results, what will your next project be? The one I would really like to see would be insertion of tube grafts with long-term follow-up to see if dilatation occurs. It is hard to believe these are going to turn out to be the same as native artery.

This was a very directed and simple study, with quite promising results. I am most grateful to Drs. Jernigan, Croce, and Fabian and their colleagues for the opportunity to discuss it.

Dr. Edward E. Cornwell, III (Baltimore, Maryland): Drs. McGinnis and Townsend, members and guests. I congratulate Dr. Croce and colleagues on their outstanding work and thank them for the opportunity to discuss it and to review the manuscript in advance of the meeting.

Several months ago I had the opportunity to spend much of a warm night in East Baltimore operating on an 18-year-old with a transpelvic GSW with injuries to the rectum, small bowel, the external iliac artery and vein with large tissue defects. The patient ultimately did well following bowel repair, vascular ligation, pelvic packing, and extra anatomic femoral-femoral bypass with PTFE; but he represents my latest opportunity to experience a sense of uncertainty regarding among other things, the long-term patency of his graft supply to his leg.

While early experimental work with porcine (SIS) suggests to so readily remodeled into host tissue, there has heretofore been little data regarding its appropriateness for vascular reconstruction in a contaminated field. The authors had taken a major step in attempting to address the gaps in our knowledge and to fill this gap in knowledge, and this area of uncertainty in one of the most challenging clinical problems facing trauma surgeons and vascular surgeons; and have aided the pursuit of the ideal vascular graft; one that is, to borrow their words, strong, pliable, readily available, resistant to infection, and have long-term patency.

The dramatic differences on blinded histologic analysis between complete tissue incorporation with infiltration of lymphocytes and fibroblasts in all pigs with SIS patches, and the total absence of those findings in the pigs with PTFE grafts are the major findings in this study, along with early differences in pulse intensity between the 2 groups.

I have 2 questions for the authors. First, your methodology called for presence of infection to be determined by inspection of the grafts in recovered animals at 2 and 4 weeks. Given this, was it really possible to blind the investigators as to which graft was used?

Second, since the intestinal mucosa acts as a cell free matrix for tissue remodeling, can you speculate on the maximum length of graft, if there is one, that is appropriate for reconstruction?

If it is determined that there is a length limitation, we may need to keep vascular ligation and extra anatomic bypass in our armamentarium for the more devastating injuries.

I would like to thank the Association for the opportunity to discuss this paper and for the privilege of membership.

Dr. Basil A. Pruitt, Jr. (San Antonio, Texas): I understood you to say that you did not give the animals any antibiotics. That certainly doesn't mimic the clinical situation. Consequently I wonder, whether you have a group to which you gave antibiotics?

Secondly, do you have any experience with impregnating the material with an antibiotic? In a totally different vein, do you think impregnation with a growth factor would accelerate the incorporation of the SIS membrane? Basic fibroblast growth factor or VEGF would be 2 likely candidates for such incorporation.

Dr. Martin A. Croce (Memphis, Tennessee): First I will address Dr. Feliciano and in essence also address Dr. Pruitt's questions. We tried to create the harshest environment possible for this particular experimental model. We did not use any omental coverage of the vascular patches. We did not reperitonealize when closing the abdomen, nor did we use any antibiotic irrigation. And that is where they got no antibiotics, we used no antibiotic irrigation. Yes, that is not quite the way it is in the clinical scenario however, it was our hypothesis if that these materials worked under these conditions then certainly we would be perhaps a leg up if in fact this makes the next step into the clinical arena.

Regarding the second question, there has been some work with these SIS extracts where they have been analyzed, and they have identified a fibroblast growth factor and transforming growth factor beta from these particular materials. Now, whether that comes from the material itself, which again these were fresh specimens so perhaps it does come from the material itself, or whether this collagen acellular matrix promotes the release from native tissue of these various growth factors, I am unsure. However, it is very intriguing.

Regarding the antimicrobial properties of this particular material, it is unlikely that we have identified a new antibiotic class. In fact, others have looked at this particular issue. And there is a peptide that is present in fresh specimens of this SIS that has antimicrobial properties. In fact, it is much more potent against staphylococcus aureus that it is against others, but it is primarily against staph aureus and also against the E coli. This peptide is not unique to SIS. It has also been extracted from urinary bladder submucosa and it is similar to the ones that are seen in various species, including insects. Whether this remains present in the packaged process material or whether it is unique to fresh specimens, I am unsure about that.

As far as the future goes, to address both Dr. Feliciano and Dr. Cornwell, it is clear that longer grafts and further work is needed. In the literature, 5 centimeter grafts have been tolerated quite well in animal models. Whether that is the limit or not, I don't really know. I hope not, because this is a fairly promising endeavor. Our plan is to use gradually longer lengths of graft to see if in fact there is a limit, and also infect these things with various other organisms, primarily mucin-producing staphylococcus epidermidis, which would be the most common organism seen in routine graft infections, and also in the presence of organisms that produce collagenase such as pseudomonas. We don't really know how this particular material will withstand that.

As far as the blinding goes, Dr. Cornwell, the pathologist obviously can't be blinded relative to SIS or PTFE, but he was blinded to the study group. So he didn't know which group the animals were.

As far as the pulse check, those were all done by Dr. Jernigan, who would enter the pig suite when all these animals were post-op and perhaps sing a song to them or read them a story so that they were all calmed down and then they would allow him to check their pulses. He knew which main group, whether they were controls or shock or shock plus contamination, but he didn't know which graft they had. So that is how we determined that.

Again, Dr. Pruitt, as far as the antibiotic use, they got perioperative intramuscular injections of cefotetan because that is what we do in the clinical scenario. But they did not have antibiotic irrigation.

As far as impregnating these things, you read our mind. It may very well be beneficial to impregnate the grafts with antibiotics or perhaps growth factors or perhaps other endothelial cells. And those are all things that we were working on in the lab right now.

I'd like to thank the Association for the honor of the floor.