- Email: [email protected]
Differences in Neointima Formation between Impervious and Porous Polytetrafluoroethylene Vascular Patch Material
1
Differences in Neointima Formation between Impervious and Porous Polytetra¯uoroethylene Vascular Patch Material Peter Wong, MD, Steven Hopkins, MD, David Vincente, MD, Katrina Williams, RN, Nicholas Macri, DVM, MS, PhD, and Ramon Berguer, MD, PhD, Detroit, Michigan
The aim of this study was to compare the neointima formation and blood loss of an impervious ePTFE with those of the porous ePTFE patch. Ten mongrel dogs were selected for the study. Both the impervious and porous ePTFE patches were implanted into the common iliac arteries in each dog. The blood loss of each patch was recorded and the patches were explanted 30-60 days later for microscopic analysis. The arteries with either the impervious or the porous ePTFE patch were all patent at the end of the study. The impervious ePTFE patch had a signi®cant reduction in blood loss when compared with the porous ePTFE patch (p = 0.04). The neointima covering both patches showed no statistically signi®cant difference in its thickness or in its cellular composition. It has been speculated that wall porosity is essential for tissue ingrowth, endothelial cell proliferation, and neointima formation. In this study, the impervious ePTFE patch did not inhibit neointima formation. Our study shows that graft porosity does not improve neointima formation or patency. Neointimization of the impervious ePTFE patch is the result of pannus ingrowth and deposition of circulating blood elements. There is a statistically signi®cant reduction in blood loss (p = 0.04) with impervious ePTFE patch, compared with that of the porous ePTFE patch.
INTRODUCTION Closure of the carotid after endarterectomy is usually achieved by employing a synthetic patch. When using a standard extended polytetra¯uoroethylene (ePTFE) patch (Gore-texÒ Cardiovascular Patch, W.L. Gore & Associates, Flagstaff, AZ), the multiple needle holes created may prolong bleedDivision of Vascular Surgery, Harper University Hospital, Wayne State University, Detroit, MI. Correspondence to: R. Berguer, Division of Vascular Surgery, Harper University Hospital, Wayne State University, 3990 John R, Detroit, MI 48201, USA. N. Macri is from W.L. Gore and Associates. Ann Vasc Surg 2002; 16: 407-412 DOI: 10.1007/s10016-001-0163-z Ó Annals of Vascular Surgery Inc. Published online: 29 July 2002
ing and require that pressure be held for an additional period at the end of the operation. An impervious ePTFE patch (Gore-tex Acuseal Cardiovascular patch, W.L. Gore & Associates) has been engineered to decrease the blood loss from needle holes and, hence, the pressure holding time. However, there has not been any study to compare results with impervious and porous ePTFE patches. Since the new patch (Acuseal) is impervious in the middle layer, tissue ingrowth will not occur and there is a question as to whether this will cause the neointima growth to be compromised. The Acuseal patch, referred to here as the impervious patch, was introduced as a modi®cation to the standard ePTFE patch (referred here as the porous patch) to decrease bleeding through needle holes. This is accomplished by introducing a layer of elastic ¯uoropolymer (the ``sealing'' element) 407
408 Wong et al.
Annals of Vascular Surgery
Table I. Sponge weight in grams of blood loss
3 Total weight Mean SD
Impervious patch
Porous patch
81 g 8.1 g 5.82
135 g 13.5 g 12.23
Paired t-test, p = 0.04.
between two sheets of ePTFE. Given the traditional belief that some porosity is required for appropriate healing of arterial prosthetics, we set out to compare the healing of an impervious ePTFE patch with that of a porous ePTFE patch. This investigation was intended to answer the following questions: 1. Will neointima develop with an impervious elastic ¯uoropolymer layer sandwiched between two ePTFE layers? 2. If neointima develops in the impervious patch, will it cover the luminal surface entirely? 3. Is the neointima thinner or thicker with an impervious ePTFE patch or a porous ePTFE patch? 4. Will there be a signi®cant reduction in blood loss with the impervious ePTFE patch?
MATERIALS AND METHODS PTFE patches were provided by Gore (Flagstaff, AZ) in the standard porous form (ePTFE) and in the new impervious layered form (Acuseal) in precut rectangular shapes of 10 ´ 15 mm size. Patches were designed wide (10 mm) to study endothelialization along a centerline that would be at least 4 mm from each suture line. Each patch was received in a sterile, individually marked package. The patched segment was considerably larger in diameter than the nonpatched segment. The reason for increasing the proportions of patch to arterial wall in the cross section was to decrease the chance of pannus ingrowth across the width of the patch (width = 8 mm). All experimental protocols and animal care complied with the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, Commission on Life Sciences, NaÂtional Research Council (Washington: National Academy Press, 1996) and were approved by the Institutional Animal and Use Committee at our institution. Ten large mongrel dogs (20-40 kg) were chosen for the evaluation. After an acclimation period of 2 weeks, each dog received cefazolin at 30 mg/kg (SmithKline Beecham, Pittsburgh, PA) 1 hr
preoperatively and was taken to the operating theater for the procedure. Dogs were anesthetized by intravenous (IV) administration of pentothal (10-20 mg/kg) and anesthesia maintained by intubation and inhalation of iso¯urane (2-3%). Once a surgical plane of anesthesia was achieved, a urinary catheter was inserted and an arterial line was placed in the carotid artery to monitor the blood pressure. The hair over the midline and along the medial aspect of each groin was clipped and a sterile ®eld developed. A midline incision was made extending from the xyphoid to the pubic area. Blunt and sharp dissection was used to expose the distal aorta and both common iliac arteries over a length of approximately 4 cm. Systemic anticoagulation was achieved with heparin (1 mg/Kg) prior to clamping the vessels. The distal aorta was then clamped, as were both distal common iliac arteries and the central caudal artery. A section of the common iliac artery along the anterior surface was selected and a rectangular shape of the vessel wall was excised corresponding to the exact shape and size of the patch. The patch was sutured with 6.0 prolene (Ethicon, NJ) suture. The order in which the two types of patches were sutured was alternated for each dog. After completion of the second anastomosis, the aortic clamp was released, restoring ¯ow to both lower extremities, and the heparin was reversed with protamine (1 mg/100 units of heparin IV). The volume of blood escaping through the needle holes was collected and quanti®ed individually for each patch. Surgical sponges were applied with gentle pressure to the sutured patch and removed and exchanged after 1-min intervals until dry. Sponges were weighed for each side separately. With hemostasis achieved (indicated by absence of bleeding from the sutured patch site after 4 mins), the laparotomy incision was closed in two layers with vicryl suture. At the completion of surgery, the arterial catheter was removed. Postoperatively the animals were evaluated for pain, activity, healing, and ¯ow to the lower extremities. Pain was controlled with buprenorphine (0.01-0.02 mg/kg) intramuscularly (IM) as needed. Following the 30- to 60-day period of healing, the dogs were again anesthetized by IV administration of pentothal (10-20 mg/kg) and maintained by intubation and inhalation of iso¯urane (2-3%). After isolation of the iliac vessels, each segment containing the patch was marked for direction of ¯ow and excised. The specimens were ®xed in glutaraldehyde for future histologic evaluation. The animals were then euthanized with saturated potassium chloride (5 mEq/kg) intravenously.
Vol. 16, No. 4, 2002
Neointima formation with impervious and porous ePTFE 409
Table II. Impervious patch
Porous patch
Neointimal coverage of luminal surface
Completely covered in 10/10 dogs
Neointimal thickness at the proximal end patch (lm)a Neointimal thickness at the middle patch (lm)a Neointimal thickness at the distal end patch (lm)a Vascularization of the neointima Vascularization of the adventitial layer Thrombosis of host vessel or luminal side of patch surface Fibrous connective tissue in®ltrating the interstices of the luminal and adventitial sides Endothelial cells in neointima
77.5)285 109.52 mean = 179.25 67.3)283 119.67 mean = 175.13 50.4)254.0 53.26 mean = 152.2 6/10 dogs 10/10 dogs None
Completely covered in 7/10 dogs and 90% covered in 3/10 of the remainder dogs 101.3)349 156.46 mean = 225.13 107.8)304.5 170.59 mean = 206.13 164.9)345.5 288.15 mean = 265.18 5/10 dogs 10/10 dogs None
Present in 10/10 dogs
Present in 10/10 dogs
Present multifocally in 7/10 dogs Present multifocally in 2/10 dogs (dogs 1 and 3) Present in 10/10 dogs on both sides Present in 9/10 dogs on both sides
Present multifocally in 6/10 dogs Present multifocally in 1/10 dogs (dog 1) Present in 10/10 dogs on both sides Present in 5/10 dogs on the luminal side and 10/10 dogs on the adventitial side
Villous hyperplasia in host vessel and neointima Lymphocyte in®ltration of the luminal and adventitial sides Macrophage in®ltration of the luminal and adventitial sides
4 aStatistical analysis of neointimal thickness at the proximal end, middle, and distal end patch, using paired t-test with subsampling: p > 0.05.
Table III.
Sacri®ce day
Dog no.
Impervious patch (mean neointimal thickness in lm)a
Day Day Day Day Day Day Day Day Day Day
1 2 3 4 5 6 7 8 9 10
62.92 80.41 107.08 180 97.67 287.5 230.83 210.33 213.5 218.33
a
30 30 30 30 30 60 60 60 60 60
Porous patch (mean neointimal thickness in lm)b 142.08 558.33 76.25 203.33 270.83 283.75 195.42 266 163.33 162.08
Paired t-test for day 30 and day 60, p = 0.007. Paired t-test for day 30 and day 60, p = 0.367.
b
RESULTS Ten dogs completed the study with no clinical signs of ischemia, infection, or vascular complications. The impervious and the porous ePTFE patches were all patent at the end of the study.
Mean and standard deviation (SD) of the blood loss for the impervious and porous patches were 8.1 5.82 g and 13.5 2.23 g, respectively. The impervious patch leaked less blood than the porous 2 patch. A paired t-test showed that this difference is statistically signi®cant (p < 0.04), (Table I).
410 Wong et al.
Fig. 1. A Gore-texÒ Acuseal cardiovascular patch, 30 days postimplantation. The luminal surface of the implant (L) is covered by a thin layer of ®brous connective tissue (arrows). There is no evidence of thrombosis in the implant lumen. Bands of ®brous connective tissue (arrowheads) are deeply in®ltrating the luminal interstices (LI) of the implant. The barrier material (B) is devoid of cellular material. The adventitial surface of the implant (A) is covered by a layer of vascularized ®brous connective tissue. Milligan's trichrome; ´10 magni®cation. B Higher magni®cation of A showing a thin layer of ®brous connective tissue (arrows) covering the luminal surface (L) of the implant. Bands of ®brous connective tissue (arrowheads) are deeply in®ltrating the luminal interstices (LI) of the implant. The barrier layer (B) is devoid of cellular material. Small amounts of ®brous connective tissue (arrowheads) are in®ltrating the adventitial interstices (AI) of the implant. Milligan's trichrome; ´20 magni®cation.
Annals of Vascular Surgery
Fig. 2. A Gore-texÒ cardiovascular patch, 30 days postimplantation. The luminal surface (L) of the implant is covered by a layer of ®brous connective tissue (arrows), which is super®cially in®ltrating the luminal interstices of the implant. The lumen of the implant is free of thrombosis. The adventitial surface (A) is covered by a layer of vascularized ®brous connective tissue. Milligan's trichrome; ´4 magni®cation. B Higher magni®cation of A showing a layer of ®brous connective tissue (arrows) covering the luminal surface (L) of the implant. Bands of ®brous connective tissue (arrowheads) are super®cially in®ltrating the luminal interstices of the implant. The adventitia of the implant (A) is covered by a layer of vascularized ®brous connective tissue. Small amounts of ®brous connective tissue (closed arrows) are in®ltrating the adventitial interstices of the implant. Milligan's trichrome; ´10 magni®cation.
Vol. 16, No. 4, 2002
Histopathology Findings At 30 to 60 days postimplantation, the healing process of the impervious patch was compared to that of the porous patch (see Table II). There was no substantial impact on the histological ®ndings, based on the timing of analysis except in the thickness of the impervious patch in the ®rst 30 days (see Table III). Three cross sections and two longitudinal sections, representing the proximal, middle, and distal aspects of each patch, were evaluated (see Figs. 1 and 2). There were no signi®cant differences in thickness or microscopic cell composition.
DISCUSSION Investigators have traditionally correlated the porosity of a vascular graft to its tissue ingrowth, neointima formation, endothelium cell proliferation, and ultimately, enhancement of prosthetic graft healing and improved graft patency. In 1961, Wesolowski et al.1 reported that the more porous the graft, the faster the neointima formation. Meijne,2 Mackenzie and Loewenthal,3 and Florey et al.4 reported that the endothelium lining of the synthetic vascular graft originates from the endothelium of the anastomotic sites and the scattered detached endothelium islands. Matsumoto et al.5 concluded in 1973 that the cellular neointima developed in the synthetic grafts is antithrombogenic, and the high porosity of the graft was essential for the formation of neointima, which derived its nutrients directly from the blood stream and the ingrowth blood vessels. The perigraft capillaries are responsible for the ingrowth of blood vessels through the graft interstices. In 1981, Kusaba et al.6 reported that the in®ltration of ®broblasts and the ingrowth of blood vessels through the graft interstices correlate directly with the porosity of the graft. Investigators have tried to determine the optimal internodal distance (porosity) for the PTFE grafts. In 1986, in a primate model, Clowes et al.7 found that endothelium originated from the adjacent artery in low-porosity grafts, growing as a continuous monolayer towards the middle of the graft for about 1 cm in length. However, the middle zone of the low-porosity grafts was covered with pseudointima (small clumps of thrombus and leukocytes) and remained unhealed. In contrast, high-porosity grafts were entirely covered by endothelium in 2 weeks.
Neointima formation with impervious and porous ePTFE 411
Other studies reported opposite conclusions regarding ePTFE porosity. Campbell et al.8 concluded that low-porosity grafts have thinner neointima and less tissue ingrowth and hence a more compatible blood tissue interface and a higher patency rate. Conversely, high-porosity grafts have a thicker neointima that sometimes thickens progressively and may lead to thrombosis. These highly porous grafts have lower patency rates than low-porosity grafts. Obviously, the in¯uence of porosity on ePTFE graft healing is not yet fully elucidated. Several studies done in animals have shown that neointima formation in a synthetic graft is based on the porosity of the graft's material. It has also been suggested that, in order for tissue ingrowth, endothelial cell proliferation, and neointimization to occur, adequate wall porosity is needed.1-7 This study compares a new impervious patch (made of an impervious layer, an elastic ¯uoropolymer sandwiched between two layers of ePTFE) with the traditional ePTFE porous patch. The midlayer of elastic ¯uoropolymer is impervious but both the adventitial and luminal layers are porous ePTFE. The middle barrier layer effectively blocks migration of the cells from both the adventitial and luminal sides (Fig. 1). Neointima covered the entire graft, despite the fact that the cellular migration was blocked from both sides. Therefore, we concluded that the neointimization of the impervious patch is the result of pannus ingrowth (host vessel tissue growth across anastomotic site) and direct deposition by circulating blood elements. In addition to the pannus ingrowth and direct deposition, the neointima of the porous patch also derived from the contribution of transinterstitial tissue migration through its pores (Fig. 2). This is in agreement with Campbell et al.'s8 ®ndings. In our study, wall porosity was not a prerequisite for the development of the neointima. Neointima can also form from pannus ingrowth and direct deposition of circulating blood elements. Wall porosity allows transinterstitial tissue migration, which adds to the neointima. The study by Contreras et al.9 of 2-mm grafts implanted in rabbits, showed that, in their model, the porous graft has the same (100%) patency rate as the impervious graft. Our results did not show any advantage of porous over impervious grafts for neointimization, patency, and general healing. However, they did show a signi®cant decrease in blood loss with the impervious patch. Our data are limited to experimental animals, especially dogs. Even though we cannot extrapolate these observations to human clinical practice, Kohler et al. have shown that in-
412 Wong et al.
creased graft porosity does not improve the con¯uence of the endothelial lining and it does not improve graft patency in humans.10
Our thanks go to Mr. William Montgomery and Tom Shook of W.L. Gore & Associates, Inc. (Flagstaff, AZ) for their support of this study. Dr. Nicholas Macri processed the histological specimens. Gore-texÒ is a registered trademark of W.L. Gore & Associates.
REFERENCES 1. Wesolowski EA, Fries CC, De Bakey M, et al. Porosity: primary determinant of ultimate fate of synthetic vascular grafts. Surgery 1961;50:91-96. 2. Meijne NC. Endothelial growth in nylon vascular prosthesis. Arc Chir Neesland 1959;11:41-56. 3. Mackenzie DC, Loewenthal J. Endothelial growth in nylon vascular grafts. Br J Surg 1960;48:212-217.
Annals of Vascular Surgery
4. Florey HW, Greer SJ, Kiser J, et al. The development of the pseudointima lining fabric grafts of the aorta. Br J Exp Pathol 1962;93:655-660. 5. Matsumoto H, Fuse K, Yamamoto M, et al. Studies on the porous polytetra¯uoroethylene as the vascular prosthesis. Jinko Zoki 1972;1:44-50. 6. Kusaba A, Fischer CR, Matuelweski TJ, et al. Experimental study of the in¯uence of porosity on development of neointima in Gore-Tex grafts: a method to increase long-term patency rates. Am Surg 1981;47:347-354. 7. Clowes AW, Kirkman TR, Reidy MA. Mechanisms of graft healing. Rapid transmural capillary ingrowth provides a source of intimal endothelium and smooth muscle in porous PTFE prothesis. Am J Pathol 1986;123:221-230. 8. Campbell CD, Goldfarb D, Roe R. A small arterial substitute: expanded microporous polytetra¯uoroethylene: patency versus porosity. Ann Surg 1975;182:138-143. 9. Contreras MA, Quist WC, Logerfo FW. Effect of porosity on small diameter vascular graft healing. Microsurgery 2000;20:15-21. 10. Kohler TR, Stratton JR, Kirkman TR, et al. Conventional versus high porosity polytetra¯uoroethylene grafts: clinical evaluation. Surgery 1992;112:901-909.
Differences in Neointima Formation between Impervious and Porous Polytetra¯uoroethylene Vascular Patch Material Peter Wong, MD, Steven Hopkins, MD, David Vincente, MD, Katrina Williams, RN, Nicholas Macri, DVM, MS, PhD, and Ramon Berguer, MD, PhD, Detroit, Michigan
The aim of this study was to compare the neointima formation and blood loss of an impervious ePTFE with those of the porous ePTFE patch. Ten mongrel dogs were selected for the study. Both the impervious and porous ePTFE patches were implanted into the common iliac arteries in each dog. The blood loss of each patch was recorded and the patches were explanted 30-60 days later for microscopic analysis. The arteries with either the impervious or the porous ePTFE patch were all patent at the end of the study. The impervious ePTFE patch had a signi®cant reduction in blood loss when compared with the porous ePTFE patch (p = 0.04). The neointima covering both patches showed no statistically signi®cant difference in its thickness or in its cellular composition. It has been speculated that wall porosity is essential for tissue ingrowth, endothelial cell proliferation, and neointima formation. In this study, the impervious ePTFE patch did not inhibit neointima formation. Our study shows that graft porosity does not improve neointima formation or patency. Neointimization of the impervious ePTFE patch is the result of pannus ingrowth and deposition of circulating blood elements. There is a statistically signi®cant reduction in blood loss (p = 0.04) with impervious ePTFE patch, compared with that of the porous ePTFE patch.
INTRODUCTION Closure of the carotid after endarterectomy is usually achieved by employing a synthetic patch. When using a standard extended polytetra¯uoroethylene (ePTFE) patch (Gore-texÒ Cardiovascular Patch, W.L. Gore & Associates, Flagstaff, AZ), the multiple needle holes created may prolong bleedDivision of Vascular Surgery, Harper University Hospital, Wayne State University, Detroit, MI. Correspondence to: R. Berguer, Division of Vascular Surgery, Harper University Hospital, Wayne State University, 3990 John R, Detroit, MI 48201, USA. N. Macri is from W.L. Gore and Associates. Ann Vasc Surg 2002; 16: 407-412 DOI: 10.1007/s10016-001-0163-z Ó Annals of Vascular Surgery Inc. Published online: 29 July 2002
ing and require that pressure be held for an additional period at the end of the operation. An impervious ePTFE patch (Gore-tex Acuseal Cardiovascular patch, W.L. Gore & Associates) has been engineered to decrease the blood loss from needle holes and, hence, the pressure holding time. However, there has not been any study to compare results with impervious and porous ePTFE patches. Since the new patch (Acuseal) is impervious in the middle layer, tissue ingrowth will not occur and there is a question as to whether this will cause the neointima growth to be compromised. The Acuseal patch, referred to here as the impervious patch, was introduced as a modi®cation to the standard ePTFE patch (referred here as the porous patch) to decrease bleeding through needle holes. This is accomplished by introducing a layer of elastic ¯uoropolymer (the ``sealing'' element) 407
408 Wong et al.
Annals of Vascular Surgery
Table I. Sponge weight in grams of blood loss
3 Total weight Mean SD
Impervious patch
Porous patch
81 g 8.1 g 5.82
135 g 13.5 g 12.23
Paired t-test, p = 0.04.
between two sheets of ePTFE. Given the traditional belief that some porosity is required for appropriate healing of arterial prosthetics, we set out to compare the healing of an impervious ePTFE patch with that of a porous ePTFE patch. This investigation was intended to answer the following questions: 1. Will neointima develop with an impervious elastic ¯uoropolymer layer sandwiched between two ePTFE layers? 2. If neointima develops in the impervious patch, will it cover the luminal surface entirely? 3. Is the neointima thinner or thicker with an impervious ePTFE patch or a porous ePTFE patch? 4. Will there be a signi®cant reduction in blood loss with the impervious ePTFE patch?
MATERIALS AND METHODS PTFE patches were provided by Gore (Flagstaff, AZ) in the standard porous form (ePTFE) and in the new impervious layered form (Acuseal) in precut rectangular shapes of 10 ´ 15 mm size. Patches were designed wide (10 mm) to study endothelialization along a centerline that would be at least 4 mm from each suture line. Each patch was received in a sterile, individually marked package. The patched segment was considerably larger in diameter than the nonpatched segment. The reason for increasing the proportions of patch to arterial wall in the cross section was to decrease the chance of pannus ingrowth across the width of the patch (width = 8 mm). All experimental protocols and animal care complied with the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, Commission on Life Sciences, NaÂtional Research Council (Washington: National Academy Press, 1996) and were approved by the Institutional Animal and Use Committee at our institution. Ten large mongrel dogs (20-40 kg) were chosen for the evaluation. After an acclimation period of 2 weeks, each dog received cefazolin at 30 mg/kg (SmithKline Beecham, Pittsburgh, PA) 1 hr
preoperatively and was taken to the operating theater for the procedure. Dogs were anesthetized by intravenous (IV) administration of pentothal (10-20 mg/kg) and anesthesia maintained by intubation and inhalation of iso¯urane (2-3%). Once a surgical plane of anesthesia was achieved, a urinary catheter was inserted and an arterial line was placed in the carotid artery to monitor the blood pressure. The hair over the midline and along the medial aspect of each groin was clipped and a sterile ®eld developed. A midline incision was made extending from the xyphoid to the pubic area. Blunt and sharp dissection was used to expose the distal aorta and both common iliac arteries over a length of approximately 4 cm. Systemic anticoagulation was achieved with heparin (1 mg/Kg) prior to clamping the vessels. The distal aorta was then clamped, as were both distal common iliac arteries and the central caudal artery. A section of the common iliac artery along the anterior surface was selected and a rectangular shape of the vessel wall was excised corresponding to the exact shape and size of the patch. The patch was sutured with 6.0 prolene (Ethicon, NJ) suture. The order in which the two types of patches were sutured was alternated for each dog. After completion of the second anastomosis, the aortic clamp was released, restoring ¯ow to both lower extremities, and the heparin was reversed with protamine (1 mg/100 units of heparin IV). The volume of blood escaping through the needle holes was collected and quanti®ed individually for each patch. Surgical sponges were applied with gentle pressure to the sutured patch and removed and exchanged after 1-min intervals until dry. Sponges were weighed for each side separately. With hemostasis achieved (indicated by absence of bleeding from the sutured patch site after 4 mins), the laparotomy incision was closed in two layers with vicryl suture. At the completion of surgery, the arterial catheter was removed. Postoperatively the animals were evaluated for pain, activity, healing, and ¯ow to the lower extremities. Pain was controlled with buprenorphine (0.01-0.02 mg/kg) intramuscularly (IM) as needed. Following the 30- to 60-day period of healing, the dogs were again anesthetized by IV administration of pentothal (10-20 mg/kg) and maintained by intubation and inhalation of iso¯urane (2-3%). After isolation of the iliac vessels, each segment containing the patch was marked for direction of ¯ow and excised. The specimens were ®xed in glutaraldehyde for future histologic evaluation. The animals were then euthanized with saturated potassium chloride (5 mEq/kg) intravenously.
Vol. 16, No. 4, 2002
Neointima formation with impervious and porous ePTFE 409
Table II. Impervious patch
Porous patch
Neointimal coverage of luminal surface
Completely covered in 10/10 dogs
Neointimal thickness at the proximal end patch (lm)a Neointimal thickness at the middle patch (lm)a Neointimal thickness at the distal end patch (lm)a Vascularization of the neointima Vascularization of the adventitial layer Thrombosis of host vessel or luminal side of patch surface Fibrous connective tissue in®ltrating the interstices of the luminal and adventitial sides Endothelial cells in neointima
77.5)285 109.52 mean = 179.25 67.3)283 119.67 mean = 175.13 50.4)254.0 53.26 mean = 152.2 6/10 dogs 10/10 dogs None
Completely covered in 7/10 dogs and 90% covered in 3/10 of the remainder dogs 101.3)349 156.46 mean = 225.13 107.8)304.5 170.59 mean = 206.13 164.9)345.5 288.15 mean = 265.18 5/10 dogs 10/10 dogs None
Present in 10/10 dogs
Present in 10/10 dogs
Present multifocally in 7/10 dogs Present multifocally in 2/10 dogs (dogs 1 and 3) Present in 10/10 dogs on both sides Present in 9/10 dogs on both sides
Present multifocally in 6/10 dogs Present multifocally in 1/10 dogs (dog 1) Present in 10/10 dogs on both sides Present in 5/10 dogs on the luminal side and 10/10 dogs on the adventitial side
Villous hyperplasia in host vessel and neointima Lymphocyte in®ltration of the luminal and adventitial sides Macrophage in®ltration of the luminal and adventitial sides
4 aStatistical analysis of neointimal thickness at the proximal end, middle, and distal end patch, using paired t-test with subsampling: p > 0.05.
Table III.
Sacri®ce day
Dog no.
Impervious patch (mean neointimal thickness in lm)a
Day Day Day Day Day Day Day Day Day Day
1 2 3 4 5 6 7 8 9 10
62.92 80.41 107.08 180 97.67 287.5 230.83 210.33 213.5 218.33
a
30 30 30 30 30 60 60 60 60 60
Porous patch (mean neointimal thickness in lm)b 142.08 558.33 76.25 203.33 270.83 283.75 195.42 266 163.33 162.08
Paired t-test for day 30 and day 60, p = 0.007. Paired t-test for day 30 and day 60, p = 0.367.
b
RESULTS Ten dogs completed the study with no clinical signs of ischemia, infection, or vascular complications. The impervious and the porous ePTFE patches were all patent at the end of the study.
Mean and standard deviation (SD) of the blood loss for the impervious and porous patches were 8.1 5.82 g and 13.5 2.23 g, respectively. The impervious patch leaked less blood than the porous 2 patch. A paired t-test showed that this difference is statistically signi®cant (p < 0.04), (Table I).
410 Wong et al.
Fig. 1. A Gore-texÒ Acuseal cardiovascular patch, 30 days postimplantation. The luminal surface of the implant (L) is covered by a thin layer of ®brous connective tissue (arrows). There is no evidence of thrombosis in the implant lumen. Bands of ®brous connective tissue (arrowheads) are deeply in®ltrating the luminal interstices (LI) of the implant. The barrier material (B) is devoid of cellular material. The adventitial surface of the implant (A) is covered by a layer of vascularized ®brous connective tissue. Milligan's trichrome; ´10 magni®cation. B Higher magni®cation of A showing a thin layer of ®brous connective tissue (arrows) covering the luminal surface (L) of the implant. Bands of ®brous connective tissue (arrowheads) are deeply in®ltrating the luminal interstices (LI) of the implant. The barrier layer (B) is devoid of cellular material. Small amounts of ®brous connective tissue (arrowheads) are in®ltrating the adventitial interstices (AI) of the implant. Milligan's trichrome; ´20 magni®cation.
Annals of Vascular Surgery
Fig. 2. A Gore-texÒ cardiovascular patch, 30 days postimplantation. The luminal surface (L) of the implant is covered by a layer of ®brous connective tissue (arrows), which is super®cially in®ltrating the luminal interstices of the implant. The lumen of the implant is free of thrombosis. The adventitial surface (A) is covered by a layer of vascularized ®brous connective tissue. Milligan's trichrome; ´4 magni®cation. B Higher magni®cation of A showing a layer of ®brous connective tissue (arrows) covering the luminal surface (L) of the implant. Bands of ®brous connective tissue (arrowheads) are super®cially in®ltrating the luminal interstices of the implant. The adventitia of the implant (A) is covered by a layer of vascularized ®brous connective tissue. Small amounts of ®brous connective tissue (closed arrows) are in®ltrating the adventitial interstices of the implant. Milligan's trichrome; ´10 magni®cation.
Vol. 16, No. 4, 2002
Histopathology Findings At 30 to 60 days postimplantation, the healing process of the impervious patch was compared to that of the porous patch (see Table II). There was no substantial impact on the histological ®ndings, based on the timing of analysis except in the thickness of the impervious patch in the ®rst 30 days (see Table III). Three cross sections and two longitudinal sections, representing the proximal, middle, and distal aspects of each patch, were evaluated (see Figs. 1 and 2). There were no signi®cant differences in thickness or microscopic cell composition.
DISCUSSION Investigators have traditionally correlated the porosity of a vascular graft to its tissue ingrowth, neointima formation, endothelium cell proliferation, and ultimately, enhancement of prosthetic graft healing and improved graft patency. In 1961, Wesolowski et al.1 reported that the more porous the graft, the faster the neointima formation. Meijne,2 Mackenzie and Loewenthal,3 and Florey et al.4 reported that the endothelium lining of the synthetic vascular graft originates from the endothelium of the anastomotic sites and the scattered detached endothelium islands. Matsumoto et al.5 concluded in 1973 that the cellular neointima developed in the synthetic grafts is antithrombogenic, and the high porosity of the graft was essential for the formation of neointima, which derived its nutrients directly from the blood stream and the ingrowth blood vessels. The perigraft capillaries are responsible for the ingrowth of blood vessels through the graft interstices. In 1981, Kusaba et al.6 reported that the in®ltration of ®broblasts and the ingrowth of blood vessels through the graft interstices correlate directly with the porosity of the graft. Investigators have tried to determine the optimal internodal distance (porosity) for the PTFE grafts. In 1986, in a primate model, Clowes et al.7 found that endothelium originated from the adjacent artery in low-porosity grafts, growing as a continuous monolayer towards the middle of the graft for about 1 cm in length. However, the middle zone of the low-porosity grafts was covered with pseudointima (small clumps of thrombus and leukocytes) and remained unhealed. In contrast, high-porosity grafts were entirely covered by endothelium in 2 weeks.
Neointima formation with impervious and porous ePTFE 411
Other studies reported opposite conclusions regarding ePTFE porosity. Campbell et al.8 concluded that low-porosity grafts have thinner neointima and less tissue ingrowth and hence a more compatible blood tissue interface and a higher patency rate. Conversely, high-porosity grafts have a thicker neointima that sometimes thickens progressively and may lead to thrombosis. These highly porous grafts have lower patency rates than low-porosity grafts. Obviously, the in¯uence of porosity on ePTFE graft healing is not yet fully elucidated. Several studies done in animals have shown that neointima formation in a synthetic graft is based on the porosity of the graft's material. It has also been suggested that, in order for tissue ingrowth, endothelial cell proliferation, and neointimization to occur, adequate wall porosity is needed.1-7 This study compares a new impervious patch (made of an impervious layer, an elastic ¯uoropolymer sandwiched between two layers of ePTFE) with the traditional ePTFE porous patch. The midlayer of elastic ¯uoropolymer is impervious but both the adventitial and luminal layers are porous ePTFE. The middle barrier layer effectively blocks migration of the cells from both the adventitial and luminal sides (Fig. 1). Neointima covered the entire graft, despite the fact that the cellular migration was blocked from both sides. Therefore, we concluded that the neointimization of the impervious patch is the result of pannus ingrowth (host vessel tissue growth across anastomotic site) and direct deposition by circulating blood elements. In addition to the pannus ingrowth and direct deposition, the neointima of the porous patch also derived from the contribution of transinterstitial tissue migration through its pores (Fig. 2). This is in agreement with Campbell et al.'s8 ®ndings. In our study, wall porosity was not a prerequisite for the development of the neointima. Neointima can also form from pannus ingrowth and direct deposition of circulating blood elements. Wall porosity allows transinterstitial tissue migration, which adds to the neointima. The study by Contreras et al.9 of 2-mm grafts implanted in rabbits, showed that, in their model, the porous graft has the same (100%) patency rate as the impervious graft. Our results did not show any advantage of porous over impervious grafts for neointimization, patency, and general healing. However, they did show a signi®cant decrease in blood loss with the impervious patch. Our data are limited to experimental animals, especially dogs. Even though we cannot extrapolate these observations to human clinical practice, Kohler et al. have shown that in-
412 Wong et al.
creased graft porosity does not improve the con¯uence of the endothelial lining and it does not improve graft patency in humans.10
Our thanks go to Mr. William Montgomery and Tom Shook of W.L. Gore & Associates, Inc. (Flagstaff, AZ) for their support of this study. Dr. Nicholas Macri processed the histological specimens. Gore-texÒ is a registered trademark of W.L. Gore & Associates.
REFERENCES 1. Wesolowski EA, Fries CC, De Bakey M, et al. Porosity: primary determinant of ultimate fate of synthetic vascular grafts. Surgery 1961;50:91-96. 2. Meijne NC. Endothelial growth in nylon vascular prosthesis. Arc Chir Neesland 1959;11:41-56. 3. Mackenzie DC, Loewenthal J. Endothelial growth in nylon vascular grafts. Br J Surg 1960;48:212-217.
Annals of Vascular Surgery
4. Florey HW, Greer SJ, Kiser J, et al. The development of the pseudointima lining fabric grafts of the aorta. Br J Exp Pathol 1962;93:655-660. 5. Matsumoto H, Fuse K, Yamamoto M, et al. Studies on the porous polytetra¯uoroethylene as the vascular prosthesis. Jinko Zoki 1972;1:44-50. 6. Kusaba A, Fischer CR, Matuelweski TJ, et al. Experimental study of the in¯uence of porosity on development of neointima in Gore-Tex grafts: a method to increase long-term patency rates. Am Surg 1981;47:347-354. 7. Clowes AW, Kirkman TR, Reidy MA. Mechanisms of graft healing. Rapid transmural capillary ingrowth provides a source of intimal endothelium and smooth muscle in porous PTFE prothesis. Am J Pathol 1986;123:221-230. 8. Campbell CD, Goldfarb D, Roe R. A small arterial substitute: expanded microporous polytetra¯uoroethylene: patency versus porosity. Ann Surg 1975;182:138-143. 9. Contreras MA, Quist WC, Logerfo FW. Effect of porosity on small diameter vascular graft healing. Microsurgery 2000;20:15-21. 10. Kohler TR, Stratton JR, Kirkman TR, et al. Conventional versus high porosity polytetra¯uoroethylene grafts: clinical evaluation. Surgery 1992;112:901-909.