- Email: [email protected]
Dr. Gary J. Becker Young Investigator Award: Comparison of Small-Diameter Type 1 Collagen Stent-Grafts and PTFE Stent-Grafts in a Canine Model—Work in Progress
Special Communication
Dr. Gary J. Becker Young Investigator Award: Comparison of Small-Diameter Type 1 Collagen Stent-Grafts and PTFE Stent-Grafts in a Canine Model—Work in Progress David F. Kallmes, MD, Horng-Ban Lin, PhD,1 Naomi H. Fujiwara, MD, John G. Short, MD,2 Klaus D. Hagspiel, MD, Shu-Tung Li, PhD,1 and Alan H. Matsumoto, MD
PURPOSE: To report an in-progress experiment in a canine model in which two types of small-diameter stent-grafts— one constructed of polytetrafluoroethylene (PTFE) and the other of a new, type 1 collagen material—were compared regarding vessel patency, intimal hyperplasia formation, and tissue reaction. MATERIALS AND METHODS: Six mongrel dogs weighing 30 –35 kg were used. Stent-grafts of 4-mm diameter and 20-mm length were constructed with use of balloon-expandable stainless-steel stents wrapped with either PTFE or a new type 1 collagen graft. Stent-grafts were placed in deep femoral arteries bilaterally (PTFE on one side, collagen on the other). Animals were followed for 2 weeks (n ⴝ 2), 6 weeks (n ⴝ 2), or 12 weeks (n ⴝ 2). Percent stenosis based on angiographic findings as well as thickness and area of neointimal hyperplasia were compared at each time point and compared with use of the Student t test. RESULTS: All devices were patent in the immediate postimplantation period. Five of six collagen stent-grafts and five of six PTFE implants were patent at follow-up. In-stent stenosis was undetectable angiographically in all five patent collagen stent-grafts. All five patent PTFE stent-grafts showed demonstrable in-stent stenosis (10%– 60%), indicating a trend toward improved patency in collagen stent-grafts versus PTFE stent-grafts (P ⴝ .07). Neointimal hyperplasia was absent at 2 weeks in the collagen stent-grafts. Neointimal thickness increased to a maximum of 360 m at 12 weeks in the collagen stent-grafts. For PTFE stent-grafts, neointimal hyperplasia was present in all samples and reached a maximum of 770 m at 12 weeks (P ⴝ .03). CONCLUSIONS: Even in small-diameter vessels, type 1 collagen stent-grafts demonstrate excellent patency rates and favorable histologic findings. The type 1 collagen stent-graft technology merits further developmental efforts in preclinical models. Index terms:
Endovascular stent-grafts • Stents and prostheses • SCVIR Annual Meeting, 2001
J Vasc Interv Radiol 2001; 12:1127–1133 Abbreviations:
ePTFE ⫽ expanded polytetrafluoroethylene, SMA ⫽ smooth muscle actin
IN THE past several years, there has been intense interest within the inter From the Department of Radiology (D.F.K., N.H.F., J.G.S., K.D.H., A.H.M.), University of Virginia, Charlottesville, Virginia; and Collagen Matrix, Inc. (H.B.L., S.T.L.), Franklin Lakes, New Jersey. Received January 25, 2001; revision requested May 11; final revision received and accepted May 21. This project was supported by NIH grant number 1R43HL61049. From the 2001 SCVIR Annual Meeting. Address correspondence to D.F.K., Box 800170, Department of Radiology, University of Virginia Health Services, Charlottesville, VA 22908; E-mail: [email protected] 1 These authors have disclosed the existence of a potential conflict of interest. 2 Current address: Asheville Radiology Associates, Asheville, North Carolina. © SCVIR, 2001
ventional radiology community regarding the development of stentgrafts. Large-diameter stent-grafts for treatment of abdominal aortic aneurysms rapidly progressed from preclinical work to clinical trials to approval by the US Food and Drug Administration (1–3). These large-diameter stent-grafts demonstrate outstanding long-term patency in animals and humans and represent one of the most exciting recent developments in minimally invasive therapies. In contradistinction to the success enjoyed by large-diameter stent-grafts, the development of small-diameter stent-grafts has been frustratingly slow
(4 – 6). Small-diameter stent-grafts offer the theoretic advantage over uncovered stents of reducing intimal hyperplasia by inhibiting cellular migration from the vessel wall into the stent lumen. Unfortunately, these devices, typically constructed of synthetic polymer such as polyester or polytetrafluoroethylene (PTFE), have shown suboptimal patency rates in normal and atherosclerotic vessels. Although the pathophysiology leading to poor long-term patency has not been clearly defined, thrombogenicity and tissue reaction related to the synthetic polymer likely contribute to progressive vessel compromise (7).
1127
1128
•
October 2001
Collagen versus PTFE Stent-Grafts in a Canine Model
JVIR
Angiographic Findings (Percent Stenosis) Stenosis (%) Time point (wk) 2 6 12
Collagen-based Endoprosthesis
ePTFE-based Endoprosthesis
0 0 5 Occluded 0 4
10 Occluded 15 25 60 15
Figure 1. Collagen stent-graft after balloon inflation.
We and others have proposed the use of natural biologic polymers, specifically type 1 collagen, for the construction of small-diameter stent-grafts (8,9). Type 1 collagen is a natural biopolymer found in high concentration in the vessel wall, which demonstrates outstanding tensile strength and is nonimmunogenic based on highly preserved tertiary protein structures among various species. Collagen also represents an excellent vector for local delivery of bioactive substances, including growth factors (10). One potentially limiting factor in the endovascular delivery of type 1 collagen regards its stimulation of the clotting cascade (11). However, its thrombogenicity might be modulated through incorporation of anticoagulants into the material (8,12). In the current study, we report our work in progress comparing patency rates of small-diameter stent-grafts constructed of type 1 collagen to those constructed of PTFE. Although most previous preclinical work on stent-grafts has focused on iliac artery implantation (13–16), in which setting devices are typically approximately 6–8 mm in diameter, we used stent-grafts of 4 mm diameter implanted into deep femoral arteries in dogs. Because vessel diameter represents one of the most important predictors of stent-graft performance (7)—smaller diameters usually result in lower patency rates—we considered the use of 4-mm vessels as a significant challenge to the type 1 collagen stentgraft technology.
MATERIALS AND METHODS Collagen Stent-Graft The collagen stent-graft (Fig 1) was engineered and supplied by Collagen Matrix, Inc. (Franklin Lakes, NJ). Purified type I collagen material was iso-
lated and purified from bovine Achilles tendon. The purified type I collagen was tested for any residual noncollagenous materials to ensure that the material was purified to the extent suitable for human implantation via an Food-and-DrugAdministration–recommended biocompatibility testing program. Heparin sodium (180 USP U/mg; Diosynth, Chicago, IL) was added to the collagen material during processing. The amount of incorporated heparin in a collagen membrane was determined by sulfur analysis, which demonstrated that approximately 20% of the dry weight was heparin. The heparin stability was evaluated in a closed-loop, in vitro system with use of circulating saline solution at ambient temperature. Within 4 days, approximately 40% of the initial heparin content had dissociated, but there was no further decrease in heparin content within 14 days. These kinetics suggest two pools of heparin within the collagen matrix, one pool loosely bound and one pool tightly bound. The processed collagen/heparin material was coated onto a Teflon mandrel, then freeze-dried and chemically cross-linked for in vivo stability. Thickness of the collagen graft was 0.15 mm ⫾ 0.03. The length of the tubular collagen membrane was trimmed to 20 mm to fabricate the prototypes. A balloon-expandable stent was fabricated by a contract stent manufacturer (Laser Technology Center, San Jose, CA). Medical quality 316-L ultrathin stainless-steel wall tubing (0.0625inch outer diameter, 0.004-inch wall thickness) was cut into a stent with use of a pulsed yttrium aluminum garnet laser system (model 126; Lasag, Arlington Heights, IL). The stent was visually inspected under a microscope for accuracy, electro-polished, and cleaned in isopropanol before use. A Marshall Balloon dilation catheter
(Boston Scientific/Medi-tech, Watertown, MA) with a 4-mm ⫻ 20-mm noncompliant balloon was used to fabricate the delivery system. The collagen membrane was affixed to the stent with use of 6 – 0 prolene suture (Ethicon, New Brunswick, NJ), with the collagen material affixed to the outer surface of the stent. Because the collagen material was radiolucent, a 2-mm segment of platinum coil was sutured to the outer surface of the collagen to lend radiopacity to the device. The collagen stent-graft was crimped onto the balloon catheter. PTFE Stent-Graft Assembly For fabrication of PTFE stent-grafts, a commercial expandable PTFE (ePTFE) graft (3-mm internal diameter; Impra/ C.R. Bard, Tempe, AZ) that did not have an external wrap was used. The ePTFE graft was cut to 20 mm long and placed over a stainless-steel stent. The ends of the ePTFE graft were sutured onto the stent with 6 – 0 prolene sutures. Because the graft material was radiolucent, a 2-mm segment of platinum coil was sutured to the outer surface of the ePTFE to lend radiopacity to the device. The stainless-steel stent/ePTFE construct was firmly crimped onto the balloon of a Marshall balloon dilation catheter (fully expanded balloon size ⫽ 4 mm outer diameter ⫻ 20 mm). Animal Protocol The study protocol was approved by the animal care and use committee at our institution and complied with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. A total of seven adult
Volume 12
Number 10
Kallmes et al
•
1129
Figure 3. ePTFE stent-graft 12 weeks after implantation. (a) Angiogram obtained immediately after stent implantation. Mid-portion of stent is denoted by arrow. (b) Angiogram 12 weeks after implantation demonstrates severe stenosis at the distal aspect of the stent (dashed arrow). The solid arrow denotes mid-portion of stent.
Figure 2. Collagen stent-graft 12 weeks after implantation. Angiogram shows no evidence for luminal compromise (solid arrow). Two side branches have recanalized (dashed arrows).
Mongrel dogs (25–30 kg) were used in the investigation. One subject died from a surgical complication at the time of stent implantation. Each of the six surviving dogs was implanted with one collagen stent-graft and one ePTFE stent-graft in its right and left deep femoral arteries, respectively. To avoid any compromise of downstream vessels that might arise from vascular access via the femoral artery, vascular access via the carotid artery was used. The implantation site (ie, right versus left) for each type of stent-graft was randomized to minimize any possible bias. Atropine 0.04 mg/kg was administered intramuscularly as a preanesthetic. Anesthesia was induced with use of 10 mg/kg pentobarbital intravenously and was maintained with use of halothane 0.5%. The skin over the right common carotid artery was prepared with use of standard sterile techniques. Surgical cutdown was performed on the carotid artery. Vascular access was obtained with use of the standard Seldinger technique with serial dilations to 12 F. A 10-F vascular sheath (60 cm long; Cook, Bloomington, IN) was advanced retrogradely into the common carotid artery and advanced into the mid-portion of the descending aorta. After intravenous administration of 5,000 U of heparin, a 5-F catheter was advanced to the distal aorta. The vascular tree was visualized by injection of 10 mm3. Omnipaque 300 (Nycomed, Prince-
ton, NJ). External sizing devices consisting of stainless-steel spheres 2–7 mm in diameter were placed over the femoral region to determine the vessel size. A road-map injection was performed to facilitate advancement of the guide wire and stent-graft delivery system (Angioskop D33/Polytron 1000; Siemens, Erlangen, Germany). An angled 0.035-inch guide wire (Terumo/Scimed, Tokyo, Japan) was advanced to the distal aspect of the deep femoral artery. The stent-graft delivery system was advanced over the guide wire until it reached the proximal aspect of the deep femoral artery. The stent-graft prototype was deployed by inflation of the balloon to 6 atm. When necessary, a second inflation was performed at the proximal or distal end of the stent to ensure that both ends of the stent were adequately expanded. Immediate postplacement angiography was performed to confirm the patency of the stent-graft within the implanted artery. Subsequently, the second type of stent-graft was deployed in the contralateral deep femoral artery with use of the same techniques. Immediate postplacement angiography was performed to confirm the patency of the implanted artery with a 5-F diagnostic catheter placed in the iliac artery. Eight milliliters of iodinated contrast material (Omnipaque 300; Nycomed) was handinjected during digital subtraction angiography (two frames per second). Heparinization was reversed with the appropriate amount of protamine (Fujisawa, Deerfield, IL) to avoid prolonged oozing at the surgical site. The vascular sheath was removed, the carotid artery was ligated, and the tissues and skin were closed with ab-
sorbable sutures (Ethicon/Johnson & Johnson, Somerville, NJ). The animals were allowed to recover. There was no postprocedural anticoagulation or antiplatelet therapy administered. Two dogs each were killed at 2, 6, and 12 weeks, respectively, with an overdose of pentobarbital (120 mg/kg intravenous), for angiographic and histologic analyses. Angiograms were obtained immediately before the animals were killed. Immediate pre-sacrifice angiography was performed from a femoral approach. A 5-F diagnostic catheter was placed in the iliac artery and 8 mL of iodinated contrast material (Omnipaque 300; Nycomed) was hand injected during digital subtraction angiography (2 frames per second). For histologic analysis, each deep femoral artery was isolated, carefully excised, and fixed with a buffered 10% formalin solution for approximately 60 minutes. The excised vessels were embedded in paraffin. With use of a rotating saw with a diamond blade, the vessel and stent-graft were sectioned axially at 2–3-mm intervals. Under a dissecting microscope, the metal stent was carefully removed. The samples were reembedded in paraffin, sectioned with use of a microtome, and stained with hematoxylin & eosin and trichome stains. Selected samples were studied with use of anti–smooth muscle actin (SMA) and anti–Factor VIII immunohistochemistry or propidium iodide staining. Angiographic Analysis Digital subtraction angiographic images were obtained before and after implantation and immediately before sacrifice. The images were transferred to film
1130
•
October 2001
Collagen versus PTFE Stent-Grafts in a Canine Model
JVIR
Figure 4. ePTFE stent-graft 12 weeks after implantation. (a) Hematoxylin & eosin stain, original magnification ⫻20. Massive neointimal hyperplasia (arrows) is seen along the luminal surface of the PTFE. “L” indicates vessel lumen. (b) Histologic specimen with anti-SMC actin stain and original magnification of ⫻100. Cells lining the graft, which were SMA-negative at 2 weeks, are now SMA-positive. (c) Histologic specimen with propidium iodide stain and an original magnification of ⫻100. As in the 6-week ePTFE sample, the ePTFE graft is relatively acellular, suggesting that transgraft cell migration may not explain the origin of neointimal hyperplasia (NIH).
with use of a laser printer. An experienced observer blinded to the prosthesis type measured the residual lumen (defined as the lumen immediately deep to the internal aspect of the metallic stent strut) immediately after implanting the endoprosthesis and immediately before sacrifice. The ePTFE and collagen devices appeared absolutely identical on angiography; both radiolucent membranes were marked with similar fragments of platinum coil. Therefore, the reader was blinded to the type of endoprosthesis. The largest and smallest residual luminal diameter, corresponding to the areas of minimum and maximum stenosis, respectively, were used for this measurement. The percent lumen stenosis was calculated. Histologic Analysis Histologic specimens were evaluated by an experienced observer. Because the ePTFE vascular graft material is clearly
obvious in the section, there was no feasible method to blind the observer to the treatment type; ie, collagen versus ePTFE. Axial sections were taken from the midportion of the segment with the stent and viewed under an Olympus BH2 microscope (Lake Success, NY) connected to an MTI digital camera (Dage-MTI, Michigan City, IN). A calibrated slide with 10-m minor units and 100-m major units was scanned into ImagePro Plus 3.0 software (Media Cybernetics, Silver Spring, MD) and used to calibrate each histologic specimen. Thickness of neointima was measured, taking into account the minimum and maximum thickness points. After measurement of neointimal thickness, neointimal areas were calculated with either automatic (for the lumen/neointima interface) or manual segmentation (for neointima/stent interface). The values of neointimal area were calculated with use of ImagePro Plus 3.0.
Statistical Analysis Neointimal thickness and neointimal areas were compared for the collagen-based endoprosthesis versus the control specimens with use of oneway analysis of variance followed by an unpaired Student t-test to evaluate two-tailed levels of significance. Minimum and maximum neointimal thickness were pooled for this comparison. SPSS Base 10.0 (SPSS, Chicago, IL) was used for these calculations.
RESULTS The collagen and ePTFE stent-grafts were deployed in the deep femoral arteries for 2, 6, and 12 weeks. Two dogs were used for each time point. Implantation of the devices was uneventful, except for ePTFE stent-grafts designated for 2- and 6-week implantation. Placement of two
Volume 12
Number 10
Kallmes et al
•
1131
Figure 5. Collagen stent-graft 12 weeks after implantation. (a) Histologic specimen with hematoxylin & eosin stain and an original magnification of 20⫻. Only minimal hyperplasia is seen along most of the lumen (solid arrows), and most of the lumen demonstrates extremely thin neointima (dashed arrow). Arrowheads indicate lucencies from the stent struts that were removed. “L” indicates vessel lumen. (b) Histologic specimen with anti–Factor VIII immunohistochemical stain and an original magnification of 400⫻. Dark-staining granules along the luminal surface of the intima indicate endothelial cells (long arrows). Smooth muscle cells of the neointima stain blue are indicated by short arrows.
Figure 6. Mean minimum (a) and maximum (b) neointimal thickness measurements.
ePTFE vascular-graft–based endoprosthesis prototypes were problematic given the slightly increased stiffness of the device compared to uncovered stents. One of these two stent-grafts was placed in the iliac artery; the other was placed in the deep femoral artery with a small segment remaining in the adjacent iliac artery. Angiographic Analysis Statistical analysis of angiography is shown in the Table. Representative angiograms are shown in Figures 2–3.
Although the difference did not reach significance, there was a strong trend toward higher percent luminal stenosis in the ePTFE stent-grafts versus the collagen stent-grafts (P ⫽ .07; twotailed t-test). Histologic Analysis Representative histologic sections are shown in Figures 4 and 5 and summary data are presented in Figure 6. For ePTFE stent-grafts, measurement of neointimal hyperplasia was straightfor-
ward because the characteristic appearance of ePTFE graft material allowed clear demarcation from adjacent neointimal hyperplasia (Fig 4). Conversely, it was very difficult, if not impossible, to exactly define the thickness of neointimal hyperplasia on the collagen-based stent-grafts at 6 and 12 weeks, because the implanted collagen membrane was extremely well incorporated into host tissues (Fig 5). In these cases, the location of the stent struts were used as a surrogate marker for the measurement of neointimal thickness. This technique
1132
•
Collagen versus PTFE Stent-Grafts in a Canine Model
Figure 7. Mean neointimal thickness measurements.
resulted in systematic overestimation of neointimal thickness for the collagen samples because the implanted collagen membrane was included in the measurement. Therefore, it should be kept in mind that the histologic analysis at 6 and 12 weeks represents the worst case scenario for neointimal hyperplasia with the collagen stent-graft. SMA stain demonstrates that the cells lining the graft, which were SMAnegative at 2 weeks, are SMA-positive at 12 weeks (Fig 4b). As in the 6-week implantation sample, propidium iodide stain (Fig 4c) demonstrated that the ePTFE was relatively acellular, calling into question the importance of transgraft cell migration and suggesting that blood elements may be responsible for cells lining the grafts. Statistical Analysis Neointimal thickness for ePTFE stent-grafts was significantly greater than that for the collagen stent-grafts (P ⫽ .03, paired t-test with pooled data for minimum and maximum thickness measurements) (Figs 6,7). Although there was a trend toward greater neointimal area formation in the ePTFE stent-grafts versus the collagen stentgrafts, this difference did not reach significance (P ⫽ .19, paired t-test). There was excellent correlation between the angiographic percent stenosis and the mean thickness of neointimal hyperplasia (P ⫽ .01, two-tailed Pearson correlation).
DISCUSSION This pilot study was conducted to compare, in small vessels, the in vivo
performance regarding vessel patency and tissue reaction of a commerciallyavailable synthetic endovascular graft material, ePTFE, and a new, type 1 collagen graft material. Our data suggest some superiority of the type 1 collagen graft over ePTFE. There was no measurable neointimal hyperplasia seen at 2 weeks on the collagen stentgraft surface. At 6 weeks, there was a thin layer of neointima of approximately 160 m covering the collagen stent-graft. Twelve weeks after implantation, the degree of neointimal hyperplasia seen with the collagen stent-graft was similar to that seen at 6 weeks after implantation. There was no evidence of inflammatory response at any time. At all time points, the degree of neointimal hyperplasia seen with the ePTFE stent-graft was greater than that seen with the collagen stent-graft, and a significant difference in neointimal thickness was seen in our data, even with a small number of subjects. These data support the need for ongoing development and testing of the type 1 collagen stent-graft for use in treating small vessel pathologies, including but not limited to de novo atherosclerosis and neointimal hyperplasia after angioplasty. Numerous previous investigators have studied, in both preclinical and clinical series, various stent-graft materials for use in small vessels. Most previous studies have evaluated synthetic materials used routinely for surgical bypass grafts, including ePTFE and Dacron. Although these materials demonstrate excellent patency in large vessels, patency rates in vessels
October 2001
JVIR
smaller than 6 mm in diameter in clinical series has been poor. The pathophysiology leading to poor patency rates remains unclear, but at least in the case of Dacron, may result from inflammation in and around the graft (5,17,18). Patency in animal models with ePTFE in vessels 6 – 8 mm diameter is relatively good, but no previous data are available regarding ePTFE stent-grafts in vessels 4 mm in diameter or smaller (13). Indeed, the ePTFE implants in our study demonstrated marked neointimal hyperplasia at 12 weeks; because the mean neointimal thickness increased between 6 and 12 weeks in our series, it remains possible that longer duration of implantation may have resulted in late-onset occlusions in the ePTFE stent-grafts, which appears concordant with previous authors’ observations. Collagen has been proposed by our group and others as a material of construction for stent-grafts (8,9). We demonstrated excellent patency rates as long as 6 months after implantation of heparin-impregnated type 1 collagen grafts in rabbit aortas; in that study, the collagen graft was not affixed along its length to a stent (8). Results of the current study are in agreement with that previous study, with excellent long-term patency and favorable tissue reaction. Goodwin et al (9) recently reported a study evaluating heparin-impregnated collagen stent-grafts in swine with dismal results; patency rates at 4 weeks were approximately 22%. The disparity between our and their work may relate to the animal model, in which we used canine and they used swine arteries, or the technique of processing of the collagen. We have implanted our own type 1 collagen stent-grafts in swine with favorable results (unpublished data, January–December 2000), so we consider that differences in collagen processing account for observed differences between the present study and previous work. Other authors have previously suggested the incorporation of heparin into graft materials to improve patency rates (9,17). However, patency rates of heparin-incorporated collagen and heparin-incorporated Dacron were poor in experimental models. Although we have not tested type 1 collagen stent-grafts without heparin incorporation, we consider the use of
Volume 12
Number 10
heparin binding to be beneficial. Heparin has been shown to modulate thrombogenicity in coronary stents, probably through inhibition of thrombin generation (12). In addition, heparin has been shown, in vitro and in vivo, to inhibit smooth muscle cell proliferation (19,20). Collagen is widely used clinically and elicits mild, if any, inflammatory response. The tertiary structure of collagen is highly conserved among species, which may account for excellent tolerance of xenografted collagen. One particularly unexpected finding in our study was the observation that, in delayed fashion, small arterial branches initially covered by the collagen membrane underwent recanalization. This observation suggests that the collagen membrane in its current formulation has a microporous structure that allows gradual reopening of side branches. When used for an atherosclerotic application in the coronary or iliofemoral system, this characteristic of the collagenbased endoprosthesis is considered highly favorable because side branches are numerous in these systems. This study, although it demonstrates favorable attributes for the collagen stentgraft, has several limitations. First, the numbers of implants at each time point was small; available funds supported only six animal subjects. However, even within this limitation, we were able to show strong trends toward improved performance of type 1 collagen over ePTFE. Second, one vessel in each group had thrombosis. In the collagen group, the occluded stent-graft had required a second inflation to achieve full expansion; vessel injury might explain this stent-graft occlusion. The current study did not use antiplatelet therapy at any point and used anticoagulants in only the periprocedural period. Antiplatelet therapy may have prevented these episodes of stent-graft thrombosis, but we were attempting to challenge the collagen stent-graft performance in this study. Last, we chose to use dogs rather than pigs in this study; patency rates of implanted devices are typically lower in the latter species than in the former. After
Kallmes et al
demonstrating excellent patency in dogs in this work, we have since moved on to the use of swine for further development of the stent-graft. Acknowledgments: The authors thank Cynthia Dodson for assistance in animal care, Thomas Young for assistance in histologic processing, and Joyce Henderson for administrative assistance. References 1. Buth J. Endovascular repair of abdominal aortic aneurysms. Results from the EUROSTAR registry. EUROpean collaborators on Stent-graft Techniques for abdominal aortic Aneurysm Repair. Semin Interv Cardiol 2000; 5:29 –33. 2. Chuter TA, Gordon RL, Reilly LM, et al. Abdominal aortic aneurysm in high-risk patients: short- to intermediate-term results of endovascular repair. Radiology 1999; 210:361–365. 3. Lipsitz EC, Ohki T, Veith FJ. Overview of techniques and devices for endovascular abdominal aortic aneurysm repair. Semin Interv Cardiol 2000; 5:21– 8. 4. Castaneda F, Ball SM, Wyffels PL, Young K, Li R. Assessment of a polyester-covered nitinol stent in an atherosclerotic swine model. J Vasc Interv Radiol 2000; 11:483– 491. 5. Henry M, Amor M, Cragg A, et al. Occlusive and aneurysmal peripheral arterial disease: assessment of a stent-graft system [see comments]. Radiology 1996; 201:717–724. 6. Maynar M, Reyes R, Ferral H, et al. Cragg Endopro System I: early experience. I. Femoral arteries. J Vasc Interv Radiol 1997; 8:203–207. 7. Palmaz JC. Review of polymeric graft materials for endovascular applications. J Vasc Interv Radiol 1998; 9:7–13. 8. Cloft HJ, Kallmes DF, Lin HB, et al. Bovine type I collagen as an endovascular stent-graft material: biocompatibility study in rabbits. Radiology 2000; 214:557–262. 9. Goodwin SC, Yoon HC, Wong GC, Bonilla SM, Vedantham S, Arora LC. Percutaneous delivery of a heparin-impregnated collagen stent-graft in a porcine model of atherosclerotic disease. Invest Radiol 2000; 35:420 – 425. 10. van Wachem PB, Plantinga JA, Wissink MJ, et al. In vivo biocompatibility of
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
•
1133
carbodiimide-crosslinked collagen matrices: Effects of crosslink density, heparin immobilization, and bFGF loading. J Biomed Mater Res 2001; 55:368 – 378. Baugh RF, Hougie C. The chemistry of blood coagulation. Clin Haematol 1979; 8:3–30. Blezer R, Cahalan L, Cahalan PT, Lindhout T. Heparin coating of tantalum coronary stents reduces surface thrombin generation but not factor IXa generation. Blood Coagul Fibrinolysis 1998; 9:435– 440. Dolmatch BL, Tio FO, Li XD, Dong YH. Patency and tissue response related to two types of polytetrafluoroethylenecovered stents in the dog. J Vasc Interv Radiol 1996; 7:641– 649. Schurmann K, Haage P, Meyer J, et al. Comparison of two stent-grafts with different porosity: in vivo studies in a sheep model. J Vasc Interv Radiol 2000; 11:493–502. Yee DC, Williams SK, Salzmann DL, et al. Stent versus endovascular graft healing characteristics in the porcine iliac artery. J Vasc Interv Radiol 1998; 9:609 – 617. Virmani R, Kolodgie FD, Dake MD, et al. Histopathologic evaluation of an expanded polytetrafluoroethylene-nitinol stent endoprosthesis in canine iliofemoral arteries. J Vasc Interv Radiol 1999; 10:445– 456. Schurmann K, Vorwerk D, Uppenkamp R, Klosterhalfen B, Bucker A, Gunther RW. Iliac arteries: plain and heparin-coated Dacron-covered stentgrafts compared with noncovered metal stents—an experimental study. Radiology 1997; 203:55– 63. Schurmann K, Vorwerk D, Bucker A, et al. Perigraft inflammation due to Dacron-covered stent-grafts in sheep iliac arteries: correlation of MR imaging and histopathologic findings. Radiology 1997; 204:757–763. Snow AD, Bolender RP, Wight TN, Clowes AW. Heparin modulates the composition of the extracellular matrix domain surrounding arterial smooth muscle cells. Am J Pathol 1990; 137: 313–330. Clowes AW, Clowes MM. Kinetics of cellular proliferation after arterial injury. II. Inhibition of smooth muscle growth by heparin. Lab Invest 1985; 52:611– 616.
Dr. Gary J. Becker Young Investigator Award: Comparison of Small-Diameter Type 1 Collagen Stent-Grafts and PTFE Stent-Grafts in a Canine Model—Work in Progress David F. Kallmes, MD, Horng-Ban Lin, PhD,1 Naomi H. Fujiwara, MD, John G. Short, MD,2 Klaus D. Hagspiel, MD, Shu-Tung Li, PhD,1 and Alan H. Matsumoto, MD
PURPOSE: To report an in-progress experiment in a canine model in which two types of small-diameter stent-grafts— one constructed of polytetrafluoroethylene (PTFE) and the other of a new, type 1 collagen material—were compared regarding vessel patency, intimal hyperplasia formation, and tissue reaction. MATERIALS AND METHODS: Six mongrel dogs weighing 30 –35 kg were used. Stent-grafts of 4-mm diameter and 20-mm length were constructed with use of balloon-expandable stainless-steel stents wrapped with either PTFE or a new type 1 collagen graft. Stent-grafts were placed in deep femoral arteries bilaterally (PTFE on one side, collagen on the other). Animals were followed for 2 weeks (n ⴝ 2), 6 weeks (n ⴝ 2), or 12 weeks (n ⴝ 2). Percent stenosis based on angiographic findings as well as thickness and area of neointimal hyperplasia were compared at each time point and compared with use of the Student t test. RESULTS: All devices were patent in the immediate postimplantation period. Five of six collagen stent-grafts and five of six PTFE implants were patent at follow-up. In-stent stenosis was undetectable angiographically in all five patent collagen stent-grafts. All five patent PTFE stent-grafts showed demonstrable in-stent stenosis (10%– 60%), indicating a trend toward improved patency in collagen stent-grafts versus PTFE stent-grafts (P ⴝ .07). Neointimal hyperplasia was absent at 2 weeks in the collagen stent-grafts. Neointimal thickness increased to a maximum of 360 m at 12 weeks in the collagen stent-grafts. For PTFE stent-grafts, neointimal hyperplasia was present in all samples and reached a maximum of 770 m at 12 weeks (P ⴝ .03). CONCLUSIONS: Even in small-diameter vessels, type 1 collagen stent-grafts demonstrate excellent patency rates and favorable histologic findings. The type 1 collagen stent-graft technology merits further developmental efforts in preclinical models. Index terms:
Endovascular stent-grafts • Stents and prostheses • SCVIR Annual Meeting, 2001
J Vasc Interv Radiol 2001; 12:1127–1133 Abbreviations:
ePTFE ⫽ expanded polytetrafluoroethylene, SMA ⫽ smooth muscle actin
IN THE past several years, there has been intense interest within the inter From the Department of Radiology (D.F.K., N.H.F., J.G.S., K.D.H., A.H.M.), University of Virginia, Charlottesville, Virginia; and Collagen Matrix, Inc. (H.B.L., S.T.L.), Franklin Lakes, New Jersey. Received January 25, 2001; revision requested May 11; final revision received and accepted May 21. This project was supported by NIH grant number 1R43HL61049. From the 2001 SCVIR Annual Meeting. Address correspondence to D.F.K., Box 800170, Department of Radiology, University of Virginia Health Services, Charlottesville, VA 22908; E-mail: [email protected] 1 These authors have disclosed the existence of a potential conflict of interest. 2 Current address: Asheville Radiology Associates, Asheville, North Carolina. © SCVIR, 2001
ventional radiology community regarding the development of stentgrafts. Large-diameter stent-grafts for treatment of abdominal aortic aneurysms rapidly progressed from preclinical work to clinical trials to approval by the US Food and Drug Administration (1–3). These large-diameter stent-grafts demonstrate outstanding long-term patency in animals and humans and represent one of the most exciting recent developments in minimally invasive therapies. In contradistinction to the success enjoyed by large-diameter stent-grafts, the development of small-diameter stent-grafts has been frustratingly slow
(4 – 6). Small-diameter stent-grafts offer the theoretic advantage over uncovered stents of reducing intimal hyperplasia by inhibiting cellular migration from the vessel wall into the stent lumen. Unfortunately, these devices, typically constructed of synthetic polymer such as polyester or polytetrafluoroethylene (PTFE), have shown suboptimal patency rates in normal and atherosclerotic vessels. Although the pathophysiology leading to poor long-term patency has not been clearly defined, thrombogenicity and tissue reaction related to the synthetic polymer likely contribute to progressive vessel compromise (7).
1127
1128
•
October 2001
Collagen versus PTFE Stent-Grafts in a Canine Model
JVIR
Angiographic Findings (Percent Stenosis) Stenosis (%) Time point (wk) 2 6 12
Collagen-based Endoprosthesis
ePTFE-based Endoprosthesis
0 0 5 Occluded 0 4
10 Occluded 15 25 60 15
Figure 1. Collagen stent-graft after balloon inflation.
We and others have proposed the use of natural biologic polymers, specifically type 1 collagen, for the construction of small-diameter stent-grafts (8,9). Type 1 collagen is a natural biopolymer found in high concentration in the vessel wall, which demonstrates outstanding tensile strength and is nonimmunogenic based on highly preserved tertiary protein structures among various species. Collagen also represents an excellent vector for local delivery of bioactive substances, including growth factors (10). One potentially limiting factor in the endovascular delivery of type 1 collagen regards its stimulation of the clotting cascade (11). However, its thrombogenicity might be modulated through incorporation of anticoagulants into the material (8,12). In the current study, we report our work in progress comparing patency rates of small-diameter stent-grafts constructed of type 1 collagen to those constructed of PTFE. Although most previous preclinical work on stent-grafts has focused on iliac artery implantation (13–16), in which setting devices are typically approximately 6–8 mm in diameter, we used stent-grafts of 4 mm diameter implanted into deep femoral arteries in dogs. Because vessel diameter represents one of the most important predictors of stent-graft performance (7)—smaller diameters usually result in lower patency rates—we considered the use of 4-mm vessels as a significant challenge to the type 1 collagen stentgraft technology.
MATERIALS AND METHODS Collagen Stent-Graft The collagen stent-graft (Fig 1) was engineered and supplied by Collagen Matrix, Inc. (Franklin Lakes, NJ). Purified type I collagen material was iso-
lated and purified from bovine Achilles tendon. The purified type I collagen was tested for any residual noncollagenous materials to ensure that the material was purified to the extent suitable for human implantation via an Food-and-DrugAdministration–recommended biocompatibility testing program. Heparin sodium (180 USP U/mg; Diosynth, Chicago, IL) was added to the collagen material during processing. The amount of incorporated heparin in a collagen membrane was determined by sulfur analysis, which demonstrated that approximately 20% of the dry weight was heparin. The heparin stability was evaluated in a closed-loop, in vitro system with use of circulating saline solution at ambient temperature. Within 4 days, approximately 40% of the initial heparin content had dissociated, but there was no further decrease in heparin content within 14 days. These kinetics suggest two pools of heparin within the collagen matrix, one pool loosely bound and one pool tightly bound. The processed collagen/heparin material was coated onto a Teflon mandrel, then freeze-dried and chemically cross-linked for in vivo stability. Thickness of the collagen graft was 0.15 mm ⫾ 0.03. The length of the tubular collagen membrane was trimmed to 20 mm to fabricate the prototypes. A balloon-expandable stent was fabricated by a contract stent manufacturer (Laser Technology Center, San Jose, CA). Medical quality 316-L ultrathin stainless-steel wall tubing (0.0625inch outer diameter, 0.004-inch wall thickness) was cut into a stent with use of a pulsed yttrium aluminum garnet laser system (model 126; Lasag, Arlington Heights, IL). The stent was visually inspected under a microscope for accuracy, electro-polished, and cleaned in isopropanol before use. A Marshall Balloon dilation catheter
(Boston Scientific/Medi-tech, Watertown, MA) with a 4-mm ⫻ 20-mm noncompliant balloon was used to fabricate the delivery system. The collagen membrane was affixed to the stent with use of 6 – 0 prolene suture (Ethicon, New Brunswick, NJ), with the collagen material affixed to the outer surface of the stent. Because the collagen material was radiolucent, a 2-mm segment of platinum coil was sutured to the outer surface of the collagen to lend radiopacity to the device. The collagen stent-graft was crimped onto the balloon catheter. PTFE Stent-Graft Assembly For fabrication of PTFE stent-grafts, a commercial expandable PTFE (ePTFE) graft (3-mm internal diameter; Impra/ C.R. Bard, Tempe, AZ) that did not have an external wrap was used. The ePTFE graft was cut to 20 mm long and placed over a stainless-steel stent. The ends of the ePTFE graft were sutured onto the stent with 6 – 0 prolene sutures. Because the graft material was radiolucent, a 2-mm segment of platinum coil was sutured to the outer surface of the ePTFE to lend radiopacity to the device. The stainless-steel stent/ePTFE construct was firmly crimped onto the balloon of a Marshall balloon dilation catheter (fully expanded balloon size ⫽ 4 mm outer diameter ⫻ 20 mm). Animal Protocol The study protocol was approved by the animal care and use committee at our institution and complied with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. A total of seven adult
Volume 12
Number 10
Kallmes et al
•
1129
Figure 3. ePTFE stent-graft 12 weeks after implantation. (a) Angiogram obtained immediately after stent implantation. Mid-portion of stent is denoted by arrow. (b) Angiogram 12 weeks after implantation demonstrates severe stenosis at the distal aspect of the stent (dashed arrow). The solid arrow denotes mid-portion of stent.
Figure 2. Collagen stent-graft 12 weeks after implantation. Angiogram shows no evidence for luminal compromise (solid arrow). Two side branches have recanalized (dashed arrows).
Mongrel dogs (25–30 kg) were used in the investigation. One subject died from a surgical complication at the time of stent implantation. Each of the six surviving dogs was implanted with one collagen stent-graft and one ePTFE stent-graft in its right and left deep femoral arteries, respectively. To avoid any compromise of downstream vessels that might arise from vascular access via the femoral artery, vascular access via the carotid artery was used. The implantation site (ie, right versus left) for each type of stent-graft was randomized to minimize any possible bias. Atropine 0.04 mg/kg was administered intramuscularly as a preanesthetic. Anesthesia was induced with use of 10 mg/kg pentobarbital intravenously and was maintained with use of halothane 0.5%. The skin over the right common carotid artery was prepared with use of standard sterile techniques. Surgical cutdown was performed on the carotid artery. Vascular access was obtained with use of the standard Seldinger technique with serial dilations to 12 F. A 10-F vascular sheath (60 cm long; Cook, Bloomington, IN) was advanced retrogradely into the common carotid artery and advanced into the mid-portion of the descending aorta. After intravenous administration of 5,000 U of heparin, a 5-F catheter was advanced to the distal aorta. The vascular tree was visualized by injection of 10 mm3. Omnipaque 300 (Nycomed, Prince-
ton, NJ). External sizing devices consisting of stainless-steel spheres 2–7 mm in diameter were placed over the femoral region to determine the vessel size. A road-map injection was performed to facilitate advancement of the guide wire and stent-graft delivery system (Angioskop D33/Polytron 1000; Siemens, Erlangen, Germany). An angled 0.035-inch guide wire (Terumo/Scimed, Tokyo, Japan) was advanced to the distal aspect of the deep femoral artery. The stent-graft delivery system was advanced over the guide wire until it reached the proximal aspect of the deep femoral artery. The stent-graft prototype was deployed by inflation of the balloon to 6 atm. When necessary, a second inflation was performed at the proximal or distal end of the stent to ensure that both ends of the stent were adequately expanded. Immediate postplacement angiography was performed to confirm the patency of the stent-graft within the implanted artery. Subsequently, the second type of stent-graft was deployed in the contralateral deep femoral artery with use of the same techniques. Immediate postplacement angiography was performed to confirm the patency of the implanted artery with a 5-F diagnostic catheter placed in the iliac artery. Eight milliliters of iodinated contrast material (Omnipaque 300; Nycomed) was handinjected during digital subtraction angiography (two frames per second). Heparinization was reversed with the appropriate amount of protamine (Fujisawa, Deerfield, IL) to avoid prolonged oozing at the surgical site. The vascular sheath was removed, the carotid artery was ligated, and the tissues and skin were closed with ab-
sorbable sutures (Ethicon/Johnson & Johnson, Somerville, NJ). The animals were allowed to recover. There was no postprocedural anticoagulation or antiplatelet therapy administered. Two dogs each were killed at 2, 6, and 12 weeks, respectively, with an overdose of pentobarbital (120 mg/kg intravenous), for angiographic and histologic analyses. Angiograms were obtained immediately before the animals were killed. Immediate pre-sacrifice angiography was performed from a femoral approach. A 5-F diagnostic catheter was placed in the iliac artery and 8 mL of iodinated contrast material (Omnipaque 300; Nycomed) was hand injected during digital subtraction angiography (2 frames per second). For histologic analysis, each deep femoral artery was isolated, carefully excised, and fixed with a buffered 10% formalin solution for approximately 60 minutes. The excised vessels were embedded in paraffin. With use of a rotating saw with a diamond blade, the vessel and stent-graft were sectioned axially at 2–3-mm intervals. Under a dissecting microscope, the metal stent was carefully removed. The samples were reembedded in paraffin, sectioned with use of a microtome, and stained with hematoxylin & eosin and trichome stains. Selected samples were studied with use of anti–smooth muscle actin (SMA) and anti–Factor VIII immunohistochemistry or propidium iodide staining. Angiographic Analysis Digital subtraction angiographic images were obtained before and after implantation and immediately before sacrifice. The images were transferred to film
1130
•
October 2001
Collagen versus PTFE Stent-Grafts in a Canine Model
JVIR
Figure 4. ePTFE stent-graft 12 weeks after implantation. (a) Hematoxylin & eosin stain, original magnification ⫻20. Massive neointimal hyperplasia (arrows) is seen along the luminal surface of the PTFE. “L” indicates vessel lumen. (b) Histologic specimen with anti-SMC actin stain and original magnification of ⫻100. Cells lining the graft, which were SMA-negative at 2 weeks, are now SMA-positive. (c) Histologic specimen with propidium iodide stain and an original magnification of ⫻100. As in the 6-week ePTFE sample, the ePTFE graft is relatively acellular, suggesting that transgraft cell migration may not explain the origin of neointimal hyperplasia (NIH).
with use of a laser printer. An experienced observer blinded to the prosthesis type measured the residual lumen (defined as the lumen immediately deep to the internal aspect of the metallic stent strut) immediately after implanting the endoprosthesis and immediately before sacrifice. The ePTFE and collagen devices appeared absolutely identical on angiography; both radiolucent membranes were marked with similar fragments of platinum coil. Therefore, the reader was blinded to the type of endoprosthesis. The largest and smallest residual luminal diameter, corresponding to the areas of minimum and maximum stenosis, respectively, were used for this measurement. The percent lumen stenosis was calculated. Histologic Analysis Histologic specimens were evaluated by an experienced observer. Because the ePTFE vascular graft material is clearly
obvious in the section, there was no feasible method to blind the observer to the treatment type; ie, collagen versus ePTFE. Axial sections were taken from the midportion of the segment with the stent and viewed under an Olympus BH2 microscope (Lake Success, NY) connected to an MTI digital camera (Dage-MTI, Michigan City, IN). A calibrated slide with 10-m minor units and 100-m major units was scanned into ImagePro Plus 3.0 software (Media Cybernetics, Silver Spring, MD) and used to calibrate each histologic specimen. Thickness of neointima was measured, taking into account the minimum and maximum thickness points. After measurement of neointimal thickness, neointimal areas were calculated with either automatic (for the lumen/neointima interface) or manual segmentation (for neointima/stent interface). The values of neointimal area were calculated with use of ImagePro Plus 3.0.
Statistical Analysis Neointimal thickness and neointimal areas were compared for the collagen-based endoprosthesis versus the control specimens with use of oneway analysis of variance followed by an unpaired Student t-test to evaluate two-tailed levels of significance. Minimum and maximum neointimal thickness were pooled for this comparison. SPSS Base 10.0 (SPSS, Chicago, IL) was used for these calculations.
RESULTS The collagen and ePTFE stent-grafts were deployed in the deep femoral arteries for 2, 6, and 12 weeks. Two dogs were used for each time point. Implantation of the devices was uneventful, except for ePTFE stent-grafts designated for 2- and 6-week implantation. Placement of two
Volume 12
Number 10
Kallmes et al
•
1131
Figure 5. Collagen stent-graft 12 weeks after implantation. (a) Histologic specimen with hematoxylin & eosin stain and an original magnification of 20⫻. Only minimal hyperplasia is seen along most of the lumen (solid arrows), and most of the lumen demonstrates extremely thin neointima (dashed arrow). Arrowheads indicate lucencies from the stent struts that were removed. “L” indicates vessel lumen. (b) Histologic specimen with anti–Factor VIII immunohistochemical stain and an original magnification of 400⫻. Dark-staining granules along the luminal surface of the intima indicate endothelial cells (long arrows). Smooth muscle cells of the neointima stain blue are indicated by short arrows.
Figure 6. Mean minimum (a) and maximum (b) neointimal thickness measurements.
ePTFE vascular-graft–based endoprosthesis prototypes were problematic given the slightly increased stiffness of the device compared to uncovered stents. One of these two stent-grafts was placed in the iliac artery; the other was placed in the deep femoral artery with a small segment remaining in the adjacent iliac artery. Angiographic Analysis Statistical analysis of angiography is shown in the Table. Representative angiograms are shown in Figures 2–3.
Although the difference did not reach significance, there was a strong trend toward higher percent luminal stenosis in the ePTFE stent-grafts versus the collagen stent-grafts (P ⫽ .07; twotailed t-test). Histologic Analysis Representative histologic sections are shown in Figures 4 and 5 and summary data are presented in Figure 6. For ePTFE stent-grafts, measurement of neointimal hyperplasia was straightfor-
ward because the characteristic appearance of ePTFE graft material allowed clear demarcation from adjacent neointimal hyperplasia (Fig 4). Conversely, it was very difficult, if not impossible, to exactly define the thickness of neointimal hyperplasia on the collagen-based stent-grafts at 6 and 12 weeks, because the implanted collagen membrane was extremely well incorporated into host tissues (Fig 5). In these cases, the location of the stent struts were used as a surrogate marker for the measurement of neointimal thickness. This technique
1132
•
Collagen versus PTFE Stent-Grafts in a Canine Model
Figure 7. Mean neointimal thickness measurements.
resulted in systematic overestimation of neointimal thickness for the collagen samples because the implanted collagen membrane was included in the measurement. Therefore, it should be kept in mind that the histologic analysis at 6 and 12 weeks represents the worst case scenario for neointimal hyperplasia with the collagen stent-graft. SMA stain demonstrates that the cells lining the graft, which were SMAnegative at 2 weeks, are SMA-positive at 12 weeks (Fig 4b). As in the 6-week implantation sample, propidium iodide stain (Fig 4c) demonstrated that the ePTFE was relatively acellular, calling into question the importance of transgraft cell migration and suggesting that blood elements may be responsible for cells lining the grafts. Statistical Analysis Neointimal thickness for ePTFE stent-grafts was significantly greater than that for the collagen stent-grafts (P ⫽ .03, paired t-test with pooled data for minimum and maximum thickness measurements) (Figs 6,7). Although there was a trend toward greater neointimal area formation in the ePTFE stent-grafts versus the collagen stentgrafts, this difference did not reach significance (P ⫽ .19, paired t-test). There was excellent correlation between the angiographic percent stenosis and the mean thickness of neointimal hyperplasia (P ⫽ .01, two-tailed Pearson correlation).
DISCUSSION This pilot study was conducted to compare, in small vessels, the in vivo
performance regarding vessel patency and tissue reaction of a commerciallyavailable synthetic endovascular graft material, ePTFE, and a new, type 1 collagen graft material. Our data suggest some superiority of the type 1 collagen graft over ePTFE. There was no measurable neointimal hyperplasia seen at 2 weeks on the collagen stentgraft surface. At 6 weeks, there was a thin layer of neointima of approximately 160 m covering the collagen stent-graft. Twelve weeks after implantation, the degree of neointimal hyperplasia seen with the collagen stent-graft was similar to that seen at 6 weeks after implantation. There was no evidence of inflammatory response at any time. At all time points, the degree of neointimal hyperplasia seen with the ePTFE stent-graft was greater than that seen with the collagen stent-graft, and a significant difference in neointimal thickness was seen in our data, even with a small number of subjects. These data support the need for ongoing development and testing of the type 1 collagen stent-graft for use in treating small vessel pathologies, including but not limited to de novo atherosclerosis and neointimal hyperplasia after angioplasty. Numerous previous investigators have studied, in both preclinical and clinical series, various stent-graft materials for use in small vessels. Most previous studies have evaluated synthetic materials used routinely for surgical bypass grafts, including ePTFE and Dacron. Although these materials demonstrate excellent patency in large vessels, patency rates in vessels
October 2001
JVIR
smaller than 6 mm in diameter in clinical series has been poor. The pathophysiology leading to poor patency rates remains unclear, but at least in the case of Dacron, may result from inflammation in and around the graft (5,17,18). Patency in animal models with ePTFE in vessels 6 – 8 mm diameter is relatively good, but no previous data are available regarding ePTFE stent-grafts in vessels 4 mm in diameter or smaller (13). Indeed, the ePTFE implants in our study demonstrated marked neointimal hyperplasia at 12 weeks; because the mean neointimal thickness increased between 6 and 12 weeks in our series, it remains possible that longer duration of implantation may have resulted in late-onset occlusions in the ePTFE stent-grafts, which appears concordant with previous authors’ observations. Collagen has been proposed by our group and others as a material of construction for stent-grafts (8,9). We demonstrated excellent patency rates as long as 6 months after implantation of heparin-impregnated type 1 collagen grafts in rabbit aortas; in that study, the collagen graft was not affixed along its length to a stent (8). Results of the current study are in agreement with that previous study, with excellent long-term patency and favorable tissue reaction. Goodwin et al (9) recently reported a study evaluating heparin-impregnated collagen stent-grafts in swine with dismal results; patency rates at 4 weeks were approximately 22%. The disparity between our and their work may relate to the animal model, in which we used canine and they used swine arteries, or the technique of processing of the collagen. We have implanted our own type 1 collagen stent-grafts in swine with favorable results (unpublished data, January–December 2000), so we consider that differences in collagen processing account for observed differences between the present study and previous work. Other authors have previously suggested the incorporation of heparin into graft materials to improve patency rates (9,17). However, patency rates of heparin-incorporated collagen and heparin-incorporated Dacron were poor in experimental models. Although we have not tested type 1 collagen stent-grafts without heparin incorporation, we consider the use of
Volume 12
Number 10
heparin binding to be beneficial. Heparin has been shown to modulate thrombogenicity in coronary stents, probably through inhibition of thrombin generation (12). In addition, heparin has been shown, in vitro and in vivo, to inhibit smooth muscle cell proliferation (19,20). Collagen is widely used clinically and elicits mild, if any, inflammatory response. The tertiary structure of collagen is highly conserved among species, which may account for excellent tolerance of xenografted collagen. One particularly unexpected finding in our study was the observation that, in delayed fashion, small arterial branches initially covered by the collagen membrane underwent recanalization. This observation suggests that the collagen membrane in its current formulation has a microporous structure that allows gradual reopening of side branches. When used for an atherosclerotic application in the coronary or iliofemoral system, this characteristic of the collagenbased endoprosthesis is considered highly favorable because side branches are numerous in these systems. This study, although it demonstrates favorable attributes for the collagen stentgraft, has several limitations. First, the numbers of implants at each time point was small; available funds supported only six animal subjects. However, even within this limitation, we were able to show strong trends toward improved performance of type 1 collagen over ePTFE. Second, one vessel in each group had thrombosis. In the collagen group, the occluded stent-graft had required a second inflation to achieve full expansion; vessel injury might explain this stent-graft occlusion. The current study did not use antiplatelet therapy at any point and used anticoagulants in only the periprocedural period. Antiplatelet therapy may have prevented these episodes of stent-graft thrombosis, but we were attempting to challenge the collagen stent-graft performance in this study. Last, we chose to use dogs rather than pigs in this study; patency rates of implanted devices are typically lower in the latter species than in the former. After
Kallmes et al
demonstrating excellent patency in dogs in this work, we have since moved on to the use of swine for further development of the stent-graft. Acknowledgments: The authors thank Cynthia Dodson for assistance in animal care, Thomas Young for assistance in histologic processing, and Joyce Henderson for administrative assistance. References 1. Buth J. Endovascular repair of abdominal aortic aneurysms. Results from the EUROSTAR registry. EUROpean collaborators on Stent-graft Techniques for abdominal aortic Aneurysm Repair. Semin Interv Cardiol 2000; 5:29 –33. 2. Chuter TA, Gordon RL, Reilly LM, et al. Abdominal aortic aneurysm in high-risk patients: short- to intermediate-term results of endovascular repair. Radiology 1999; 210:361–365. 3. Lipsitz EC, Ohki T, Veith FJ. Overview of techniques and devices for endovascular abdominal aortic aneurysm repair. Semin Interv Cardiol 2000; 5:21– 8. 4. Castaneda F, Ball SM, Wyffels PL, Young K, Li R. Assessment of a polyester-covered nitinol stent in an atherosclerotic swine model. J Vasc Interv Radiol 2000; 11:483– 491. 5. Henry M, Amor M, Cragg A, et al. Occlusive and aneurysmal peripheral arterial disease: assessment of a stent-graft system [see comments]. Radiology 1996; 201:717–724. 6. Maynar M, Reyes R, Ferral H, et al. Cragg Endopro System I: early experience. I. Femoral arteries. J Vasc Interv Radiol 1997; 8:203–207. 7. Palmaz JC. Review of polymeric graft materials for endovascular applications. J Vasc Interv Radiol 1998; 9:7–13. 8. Cloft HJ, Kallmes DF, Lin HB, et al. Bovine type I collagen as an endovascular stent-graft material: biocompatibility study in rabbits. Radiology 2000; 214:557–262. 9. Goodwin SC, Yoon HC, Wong GC, Bonilla SM, Vedantham S, Arora LC. Percutaneous delivery of a heparin-impregnated collagen stent-graft in a porcine model of atherosclerotic disease. Invest Radiol 2000; 35:420 – 425. 10. van Wachem PB, Plantinga JA, Wissink MJ, et al. In vivo biocompatibility of
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
•
1133
carbodiimide-crosslinked collagen matrices: Effects of crosslink density, heparin immobilization, and bFGF loading. J Biomed Mater Res 2001; 55:368 – 378. Baugh RF, Hougie C. The chemistry of blood coagulation. Clin Haematol 1979; 8:3–30. Blezer R, Cahalan L, Cahalan PT, Lindhout T. Heparin coating of tantalum coronary stents reduces surface thrombin generation but not factor IXa generation. Blood Coagul Fibrinolysis 1998; 9:435– 440. Dolmatch BL, Tio FO, Li XD, Dong YH. Patency and tissue response related to two types of polytetrafluoroethylenecovered stents in the dog. J Vasc Interv Radiol 1996; 7:641– 649. Schurmann K, Haage P, Meyer J, et al. Comparison of two stent-grafts with different porosity: in vivo studies in a sheep model. J Vasc Interv Radiol 2000; 11:493–502. Yee DC, Williams SK, Salzmann DL, et al. Stent versus endovascular graft healing characteristics in the porcine iliac artery. J Vasc Interv Radiol 1998; 9:609 – 617. Virmani R, Kolodgie FD, Dake MD, et al. Histopathologic evaluation of an expanded polytetrafluoroethylene-nitinol stent endoprosthesis in canine iliofemoral arteries. J Vasc Interv Radiol 1999; 10:445– 456. Schurmann K, Vorwerk D, Uppenkamp R, Klosterhalfen B, Bucker A, Gunther RW. Iliac arteries: plain and heparin-coated Dacron-covered stentgrafts compared with noncovered metal stents—an experimental study. Radiology 1997; 203:55– 63. Schurmann K, Vorwerk D, Bucker A, et al. Perigraft inflammation due to Dacron-covered stent-grafts in sheep iliac arteries: correlation of MR imaging and histopathologic findings. Radiology 1997; 204:757–763. Snow AD, Bolender RP, Wight TN, Clowes AW. Heparin modulates the composition of the extracellular matrix domain surrounding arterial smooth muscle cells. Am J Pathol 1990; 137: 313–330. Clowes AW, Clowes MM. Kinetics of cellular proliferation after arterial injury. II. Inhibition of smooth muscle growth by heparin. Lab Invest 1985; 52:611– 616.