- Email: [email protected]
Vein patching reduces neointimal thickening associated with prosthetic graft implantation
Vein Patching Reduces Neointimal Thickening Associated with Prosthetic Graft Implantation Andrew T. Gentile, MD, Joseph L. Mills, MD, Michael A. Gooden, MD, R. Daniel Hagerty, MS, Scott S. Berman, MD, John D. Hughes, MD, Leigh B. Kleinert, Stuart K. Williams, PhD, Tucson, Arizona
PURPOSE: Modified anastomotic techniques utilizing autogenous vein-cuffs or patches have been devised with the hope of improving prosthetic graft patency. The mechanisms of the presumed improvement in patched anastomoses have never been elucidated and remain speculative. We characterized the healing response of the Taylor vein patch in prosthetic arteriovenous fistulae in a canine model of intimal hyperplasia. METHODS: Six adult dogs underwent placement of bilateral (6 patched, 6 control) 4-mm diameter expanded polytetrafluoroethylene loop femoral artery-vein fistulae. Serial duplex ultrasound examinations confirmed graft patency until explant at 6 weeks. Differential light microscopy with computerized image analysis was performed on serial 5-mm sections. Intimal thickness through the venous anastomosis and outflow veins of Taylor patch and control (nonpatched) grafts were compared. Cell type-specific immunocytochemical antibody stains for smooth muscle cells (a SMC actin) and endothelial cells (von Willebrand factor) were performed. RESULTS: Eleven of 12 grafts remained patent for 6 weeks, 1 control graft thrombosed. Mean duplex-derived peak systolic velocities of patched (96 cm/sec) and control (108 cm/sec) grafts were similar. Microscopy revealed more intimal pannus anastomotic suture line ingrowth in controls (mean thickness 5 178 mm) than Taylor patched grafts (mean 147 mm, p 5 0.0002). Significantly less intimal thickening was present in the outflow vein of patched (mean thickness 5 90 mm) versus control grafts (mean 195 mm, P <0.0001). The intima maintained a single cell layer of vWF 1 endothelial cells, while the majority of the cells comprising the lesion expressed a SMC actin.
From the Division of Vascular Surgery (ATG, JLM, MAG, RDH, SSB, JDH) and the Department of Surgical Research (LBK, SKW), University of Arizona Health Sciences Center, Tucson, Arizona. This work was supported in part by a Research Development Grant from the Impra Pharmaceuticals Company, Tempe, Arizona. Requests for reprints should be addressed to Joseph L. Mills, MD, Professor of Surgery, Chief, Division of Vascular Surgery, AHSC PO Box 245072, 1501 N. Campbell Avenue, Tucson, Arizona 85724-5072. Presented at the 50th Annual Meeting of the South western Surgical Congress, San Antonio, Texas, April 19 –22, 1998. © 1998 by Excerpta Medica, Inc. All rights reserved.
CONCLUSION: Perianastomotic pannus is primarily composed of intimal smooth muscle cells. Neointimal thickening is significantly reduced in prosthetic arteriovenous fistulae created with the Taylor vein patch in a canine model. Reduction in perianastomotic intimal thickening may explain the reported clinical improvement in prosthetic bypass graft patency when modified with vein patch techniques. Am J Surg. 1998;176: 601– 607. © 1998 by Excerpta Medica, Inc.
I
n an effort to increase the patency of vascular reconstructions requiring the use of prosthetic bypass conduits, various graft configurations and anastomotic modifications have been proposed and applied clinically in several small human trials. The most common reasons infrainguinal prosthetic grafts are placed are lack or insufficient length of usable greater saphenous vein. Although results of lower extremity revascularization with autogenous vein are clearly superior to prosthetics with at least a 30% difference in primary 5-year patency,1 approximately 10% to 30% of patients do not have suitable saphenous vein for lower extremity arterial bypass.2,3 The failure of prosthetic grafts in distal reconstructions is thought to result largely from fibrointimal hyperplasia at the distal anastomosis.4 The concept of a prosthetic/autogenous composite was developed in response to the generally poor patency achieved with prosthetic grafts for distal arterial bypass in the setting of lack of suitable vein. Prostheticvein composite graft patencies are generally intermediate between those of all-autogenous and all prosthetic bypass grafts,5 with reported mean 2-year cumulative patencies ranging from 30% to 64%.6,7 Theories abound concerning the pathogenesis of the intimal lesions responsible for prosthetic, as well as autogenous graft-anastomotic failures. Various operative adjuncts, including vein cuffs, collars, and patches, have been placed in humans to modify these perianastomotic factors and improve graft patency. The mechanisms of the presumed improvement in patched anastomoses have never been elucidated and remain speculative. The common denominator in prosthetic graft failure appears to be related to the perianastomotic healing response of the recipient vessel. This is true not only for arterial reconstructions but also for prosthetic arteriovenous fistuae with venous outflow stenosis responsible for the majority of hemodialysis fistula failure. In order to characterize the healing response of a modified prosthetic anastomosis, Taylor vein patched arteriovenous fistulae were implanted in a canine model of intimal hyperplasia. 0002-9610/98/$19.00 PII S0002-9610(98)00286-4
601
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
MATERIALS AND METHODS Six adult mongrel dogs of either gender each weighing 24 to 28 kg were used in this study. Study design and research methods were approved by the Institutional Animal Care and Use Committee at the University of Arizona Health Sciences Center. All animal experiments were performed in accordance with the “Principles of Laboratory Animal Care and the Guide for the Care and Use of Laboratory Animals” (NIH Publication no. 80-23, 1985). Graft Implantation Dogs underwent preoperative general medical examination and were found fit to undergo femoral artery to vein prosthetic graft implantation. Prior to surgery, each animal received 50 mg dipyridamole orally twice per day for 4 days and 325 mg aspirin orally once per day for 1 day. These medications were continued during the entire postoperative period. All animals received perioperative antibiotics. On the day of surgery, animals were sedated with a mixture of atropine, ketamine, and acepromazine injected intramuscularly. General anesthesia was induced using intravenous sodium pentothal, and the animal was intubated and maintained on mechanical ventilation with a mixture of inhaled halothane and oxygen. Grafts were prepared using standard, 4-mm internal diameter, expanded polytetrafluoroethylene (e-PTFE; IMPRA Inc. Tempe, Arizona). Graft implantation was performed through diagonal, femoral incisions with standard vascular surgical techniques. Endgraft to side-femoral arterial and venous anastomoses were performed (6 study and 6 controls). A single segment of native internal jugular vein was harvested, flushed, opened, and cut to a diamond shape for autogenous Taylor patch as previously described.8 Six Taylor vein-patch anastomoses were constructed for the modified venous-outflow anastomosis. All anastomoses were performed with 6-0 Prolene suture (Ethicon Inc., Somerville, New Jersey). Graft Explant Serial duplex ultrasonography was performed to confirm graft patency and measure graft blood flow velocities at implant, 3 weeks, and immediately preceding graft explant at 6 weeks. At explant, animals received identical preoperative sedation and general anesthesia. Dogs were systemically heparinized, and the grafts were explanted in-toto along with approximately 3 cm of inflow artery and outflow vein segments. Grafts were gently flushed with heparinized saline and pressure fixed at 80- to 100 mm Hg with 10% neutral buffered formalin. After fixation, grafts were sequentially sectioned in the direction of blood flow through both anastomoses. Graft segments were dehydrated and paraffin embedded. Five-micron cross sections were floated on glass slides, deparaffinized, and stained with hematoxylin & eosin, and Masson’s trichrome. Using the Nikon Optiphot microscope with image transfer through a DAGE SIT-66 low light level camera, serial micrograph cross sections of the venous anastomoses were compared at 20 to 403 magnification. Analysis of the histologic characteristics and neointimal layer thickness was performed using digitized computerized image analysis (MetaMorph Image Analysis Systems) program linked through Microsoft Windows Applications. Serial cross sec602
Figure 1. Cross section of a well-incorporated (light, outer adventitial covering) explanted ePTFE graft (dark middle layer) with computerized image analysis for determination of mean thickness of the neointimal lining (light, hatched inner luminal layer; 320).
Figure 2. Schematic drawing of Taylor-patched arteriovenous anastomosis denoting multiple sample sites used to measure neointimal thickness at cross sections (narrow lines) through the anastomosis and separate sample area at outflow vein segments (wide line).
tions were compared through the venous anastomosis of both control and Taylor patched grafts. The image analysis software graphics allows calculation of neointimal area and average neointimal thickness was obtained for each graft segment (Figure 1). The average intimal thickness throughout the entire venous anastomosis as well as the intimal thickness in a clearly defined segment of outflow vein just distal to the anastomosis were compared between groups (Figure 2). All measurements were electronically transferred from MetaMorph to Microsoft Excel spreadsheets. Statistical comparison between control and study venous anastomosis and outflow vein segments were performed using the Student’s t test for continuous variables on standard PC compatible software (StatView; Abacus Concepts, Berkeley, California). A P value of #0.05 was considered statistically significant.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
Figure 3. Representative gross photograph at graft explant revealing a well-healed, tightly looped 4-mm ePTFE arteriovenous fistula. Vascular loops encircle inflow arterial (dark loops) and outflow vein (light loops) anatomy.
Immunohistologic Analysis Photomicrographs of cross sections were analyzed for morphologic features and neointimal pannus ingrowth. Single-label immunocytochemistry was performed after deparaffinization with xylene and rehydration in graded ethanol solutions. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide, and primary antibodies to either von Willebrand factor (vWF; kindly provided by Dr. James Catalfano, New York State Health Department, Albany, New York) or a smooth muscle cell actin (Sigma Co., Bushfield, Tennessee) were incubated at room temperature for 60 minutes. The sections were rinsed with phosphate buffered saline, a biotinylated secondary antibody was applied for 30 minutes, followed by a streptavidin peroxidase conjugate (Dako Corp. Carpinteria, California). Standard peroxidase enzyme substrate, 3939-diaminobenzidine or alkaline phosphatase substrate were added to yield a brown or red reaction product. Nuclei were lightly counterstained with dilute methyl green.
RESULTS This canine study compared the patency and healing characteristics of standard and a modified anastomotic configuration of small caliber prosthetic arteriovenous fistulae. All dogs tolerated the operative procedures without adverse events. Eleven of 12 grafts remained patent for 6 weeks, 1 control graft thrombosed. Mean duplex-derived perianastomotic peak systolic velocities of patched (96 6 22 cm/sec) and control (108 6 24 cm/sec) grafts were similar. At explant, all grafts were well incorporated without signs of infection. The grafts were implanted as smooth U-shaped loops; however, during healing the grafts often formed tightly associated proximal and distal segments (Figure 3). Gross examination of the luminal surface of the grafts at the time of explant revealed clean body and midgraft segments without an obvious inner lining or associated blood components. The perianastomotic areas revealed only very small amounts of a glistening neointimal layer without adherence of thrombus or platelets. Microscopic Results Morphologic analysis of grafts from both groups revealed very clean luminal surfaces throughout most of the grafts.
The Taylor patch segments were well incorporated and without microscopic signs of infection. The patch remained viable as evidenced by normal appearing vein histology consisting of endothelium, organized intimal layer bounded by an inner elastic lamina and muscular medial layer surrounded by loose connective tissue and adventitia. There was minimal change in the intimal layer morphology of the Taylor patched segment when compared with normal internal jugular vein specimens (Figure 4A). The perianastomotic areas had obvious neointimal pannus ingrowth most often at the cut edge of the vein and anastomotic suture line. This perianastomotic neointimal pannus was qualitatively more prominent in control grafts than in the Taylor-patched grafts (Figure 4B). The anastomotic neointima contained a layer of well-organized smooth muscle cells covered by a single layer of polygonalshaped endothelial cells. Since grafts were sectioned transversely, representative cross sections were analyzed to quantify the neointima within the inner surface of the grafts through the anastomosis and at the outflow vein segments. Microscopy revealed distinct areas of neointimal ingrowth starting at the heel of the PTFE with neointimal ingrowth at the margins of the graft into the recipient vessel. Quantitatively there was more intimal pannus anastomotic suture line ingrowth in controls (mean thickness 5 178 6 129 mm) than Taylor patched grafts (mean 147 6 103 mm, P 5 0.0002; Table). In addition, significantly less intimal thickness was present in the outflow vein of patched (mean thickness 5 90 6 62 mm) versus control grafts (mean 195 6 107 mm, P ,0.0001). Immunohistologic Results Single-label immunocytochemical staining produced positive staining of anti-vWf antibody in areas of neointima covered by endothelial cells and staining for anti-a SMC actin in subendothelial regions. The neointima maintained an organized morphology with a confluent single cell layer of vWF 1 endothelial cells along the neointima lining the luminal surface of the grafts, (Figure 4C) while the majority of the cells comprising the neointimal lesion expressed a SMC actin consistent with smooth muscle cell ingrowth. (Figure 4D)
COMMENTS Polytetrafluoroethylene grafts have been used extensively for infrainguinal vascular reconstructions either as the conduit of choice or as an alternative conduit when saphenous vein is unavailable. Although studies have reported acceptable early patency rates, the long-term efficacy of these grafts has been less satisfactory, especially in the infrapopliteal position. Results of nearly all studies concur that infrainguinal PTFE prostheses provide significantly inferior results when compared with autogenous vein.1,9 –11 Rutherford et al12 analyzed the factors affecting the patency of infrainguinal bypass grafts and confirmed the often held dictum that the quality of graft conduit and the anastomotic level (below-knee versus above-knee) are two important variables affecting bypass performance. Because results for infrapopliteal revascularization with prosthetic grafts have been so disappointing, some authors have considered primary amputation preferable to infrapopliteal ar-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
603
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
Figure 4. Photomicrographs demonstrating perianastomotic intimal pannus. A. Cross section through a Taylor-patched anastomosis at 103 reveals a normal appearing vein patch segment (arrow) and outflow vein segment (*) with minimal intimal thickening. B. At higher magnification (Trichrome stain, 320) neointimal pannus ingrowth into the PTFE is observed at anastomotic heel regions and along prolene suture lines. C. Immunocytochemical staining for anti-vWF antibody demonstrates dark staining endothelial monolayer. D. The majority of the neointimal lesion is composed of cells staining positive for anti-a SMC actin antibody.
TABLE Canine Taylor Patch Mean Intimal Thickness Measurements Graft Taylor patch (n 5 6) Control (n 5 5) Significance t test
Venous Anastomosis
Venous Outflow
147 6 103 mm (318 counts)* 178 6 129 mm (588 counts) P 5 0.0002
90 6 62 mm (31 counts) 195 6 107 mm (64 counts) P ,0.0001
* Counts 5 individual thickness measurements.
tery bypass with PTFE in patients without usable autogenous vein.13 Recently, however, when faced with the more frequent clinical dilemma of elderly patients requiring secondary extremity revascularizations and lacking adequate autogenous conduit, investigators have applied various technical adjuncts in hopes of improving prosthetic bypass patency rates and limb salvage. These techniques include various vein patches, cuffs, or collars, and distal arteriovenous fistulae.14 Early uncontrolled studies of patency rates of PTFE with vein cuffs in the femoral-distal position reported encouraging initial results. Miller et al (1984)15 and Tyrell and Wolfe (1991)16 reported 72% and 78% 8 to 12 month patencies, respectively. Subsequent work on the influence of vein cuffs on PTFE femoral-popliteal bypass by Raptis 604
and Miller17 retrospectively reported no difference in above-knee PTFE patency with or without vein cuffs at 36 months (68% versus 69%), but a significant improvement in below-knee PTFE patency was demonstrated with the use of adjuvant vein cuffs (57% versus 29%). In a recent randomized, prospective trial reported by Stonebridge et al,18 vein cuffed below-knee prosthetic bypass produced a 65% 2-year patency rate compared with 29% patency in uncuffed grafts, with a sixfold greater early graft failure rate in the uncuffed grafts (13%) versus cuffed grafts (2%). The most promising results of a vein cuff clinical series, by Pappas et al,19 report 62% cumulative 2-year patency for 23 infrapopliteal cuffed prosthetic grafts. Reported results of uncontrolled series of infrainguinal bypasses using the Taylor vein patch approach those of revascularizations
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
with autogenous conduit, with an overall primary patency rate of 65% at 5 years.8 We have had anecdotal experience with distal anastomotic graft modifications in several patients requiring infrapopliteal bypass without usable vein, and have been impressed by the proliferation of clinical reports, trials and animal studies on the potential for vein collars and other enhancements to improve infrainguinal (and possibly arteriovenous dialysis fistula) prosthetic bypass results. The most widely held opinions on the etiology of prosthetic bypass graft failure have focused on graft narrowing from distal perianastomotic intimal hyperplasia.4 Theories abound concerning the pathogenesis of the intimal lesions responsible for prosthetic, as well as autogenous graft-anastomotic failures. The rationale for the addition of a vein cuff or patch to the distal anastomosis is to limit neointimal hyperplastic ingrowth. Plausible, though unproven, explanations have described various hemodynamic causes such as abnormal velocity patterns, turbulence, and shear stress;20 mechanical causes such as compliance mismatch, geometry, and technical factors,21 all of which may contribute to the (mal)adaptive vascular remodeling of the perianastomotic area. The mechanisms of the reported clinical patency improvement in vein-patched or cuffed anastomoses have never been elucidated and remain speculative. The common denominator in prosthetic graft failure appears to be the perianastomotic healing response of the recipient vessel. This current report was designed to investigate the healing properties of the Taylor vein patched anastomosis in a well-characterized canine model of intimal hyperplasia.22 Although we noted no difference in initial patency of control and study grafts, the addition of the Taylor patch had a significant affect on limiting neointimal pannus ingrowth at the perianastomotic suture line, and also importantly decreases intimal thickening at the juxta-anastomotic venous outflow segments. This decrease in neointimal thickening associated with a modified vascular anastomosis in this experimental model of intimal hyperplasia, the canine arteriovenous fistula, has not been previously demonstrated. Suggs et al23 demonstrated decreased patency of non-cuffed PTFE but similar histologic findings in a canine model using small caliber PTFE carotid bypass grafts. Norberto et al24 demonstrated no histologic difference in jacketed-vein cuffed and non-cuffed carotid PTFE anastomoses, concluded the protective effect of vein cuffs are not mechanical, and suggested this effect may be humoral in nature. We were able to clearly demonstrate a significant reduction in neointimal thickening in prosthetic arteriovenous fistulae created with the Taylor vein patch in a canine model. Microscopic and immunocytochemical analysis demonstrated that perianastomotic pannus is primarily composed of intimal smooth muscle cells. The luminal surface of the neointimal pannus maintains a confluent monolayer of endothelial cells. The juxta-anastomotic neointimal pannus is presumed to be derived from smooth muscle cell migration from the adjacent recipient vessel. Although this project demonstrated a reduction in neointimal thickening associated with vein-patched prosthetic anastomoses, the precise mechanisms responsible for this
difference were not elucidated, and further investigations into hemodynamic and/or humoral mechanisms are needed to explain these findings. One proposed weakness for the Taylor patched type of anastomosis is that approximately one third (heel portion) of the prosthetic graft is in contact with the recipient vessel and may be affected by suture line pannus ingrowth. Other investigators have concluded that results of all composite vein-PTFE anastomoses will be affected by myointimal hyperplasia, thus shifting this graft-threatening process more proximally in the graft. By bringing the autogenous patch across the heel of the anastomosis, the Taylor patch provides a somewhat patulous anastomotic hood, which may make any neointimal thickening at the suture line insignificant in affecting graft patency. We demonstrated an altered healing response of vein-patched anastomoses with reduction in perianastomotic intimal thickening when compared with controls. Operative adjuncts to prosthetic graft anastomoses may limit neointimal thickening by providing an “anastomotic buffer zone” and may explain the reported clinical improvement in prosthetic bypass graft patency when modified with vein patch techniques. Further investigation of the precise mechanisms for these findings is ongoing.
REFERENCES 1. Veith FJ, Gupta SK, Ascer EA, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J Vasc Surg. 1986;3:104 –114. 2. Wolfe JHN, Tyrell MR. Justifying arterial reconstruction to crural vessels-even with prosthetic graft. Br J Surg. 1991;78:897– 899. 3. Brewster DC, Lasalle AJ, Robinson JG, et al. Femoropopliteal graft failures: Clinical consequences and success of secondary reconstructions. Arch Surg. 1983;118:1043–1047. 4. Clowes AW, Gown AM, Hanson SR, Reidy MA. Mechanisms of arterial graft failure. I. Role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol. 1985;118:43–54. 5. Demasi RJ, Snyder SO. The current status of prosthetic-vein composite grafts for lower extremity revascularization. Surg Clin North Am. 1995;75:741–752. 6. Feinberg RI, Winter RP, Wheeler JR, et al. The use of composite grafts in femorcrural bypasses performed for limb salvage: a review of 108 consecutive cases and comparison with 57 in situ saphenous vein bypasses. J Vasc Surg. 1990;12:257–263. 7. Gregory RT, Raithel D, Snyder SO, et al. Composite grafts: an alternative to saphenous vein for lower extremity arterial reconstruction. J Cardiovasc Surg. 1983;24:53–57. 8. Taylor RS, Loh A, McFarland RJ, et al. Improved technique for polytetrafluoroethylene bypass grafting: long-term results using anastomotic vein patches. Br J Surg. 1992;79:348 –354. 9. Whittemore AD, Kent KC, Donaldson MC, et al. What is the proper role of polytetrafluoroethylene grafts in infrainguinal reconstruction. J Vasc Surg. 1989;10:299 –305. 10. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J Vasc Surg. 1986;3:104 –114. 11. Hobson RW II, Lynch TG, Zafar J, et al. Results of revascularization and amputation in severe lower extremity ischemia: a five-year clinical experience. J Vasc Surg. 1985;2:174 –185. 12. Rutherford RD, Jones DN, Bergentz SE, et al. Factors affecting the patency of infrainguinal bypass. J Vasc Surg. 1988;8:236 –246. 13. Yeager RA, Hobson RW II, Jamil Z, et al. Differential patency
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
605
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
and limb salvage for polytetrafluoroethylene and autogenous saphenous vein in severe lower extremity ischemia. Surgery. 1982;91:99 – 103. 14. Webb T. Operative adjuncts for distal revascularization. Surg Clin North Am. 1995;75:753–758. 15. Miller JH, Foreman RK, Ferguson L, Faris I. Interposition vein cuff for anastomosis of prosthesis to small arteries. Aust NZ J Surg. 1984;54:283–285. 16. Tyrell MR, Wolfe JHN. New prosthetic venous collar anastomotic technique: combining the best of the other procedures. Br J Surg. 1991;78:1016 –1017. 17. Raptis S, Miller JH. Influence of a vein cuff on polytetrafluoroethylene graft for primary femoropopliteal bypass. Br J Surg. 1995;82:487– 491. 18. Stonebridge PA, Prescott RJ, Ruckley CV. Randomized trial comparing infrainguinal polytetrafluoroethylene bypass grafting with and without vein interposition cuff at the distal anastomosis. J Vasc Surg. 1997;26:543–550.
19. Pappas PJ, Hobson RW II, Meyers MG, et al. Patency of infrainguinal polytetrafluoroethylene bypass grafts with distal interposition vein cuffs. Cardio Vasc Surg. 1998;6:19 –26. 20. Binns RL, Ku DN, Stewart MT, et al. Optimal graft diameter: effect of wall shear stress on vascular healing. J Vasc Surg. 1989; 10:326 –337. 21. Abbott WM, Megerman J, Hasson JE, et al. Effect of compliance mismatch on vascular graft patency. J Vasc Surg. 1987;5:376 – 381. 22. Williams SK, Jarrell BE, Kleinert LB. Endothelial cell transplantation onto polymeric arteriovenous grafts evaluated using a canine model. J Invest Surg. 1994;7:503–517. 23. Suggs WD, Henriques HF, DePalma RG. Vein cuff interposition prevents juxta-anastomotic neointimal hyperplasia. Ann Surg. 1988;207:717–723. 24. Norberto JJ, Sidawy AN, Trad KS, et al. The protective effect of vein cuffed anastomoses in not mechanical in origin. J Vasc Surg. 1995;21:558 –566.
DISCUSSION
luminal surface, with a relatively acellular zone between the graft and the smooth muscle cells, what appeared to be like matrix. Where do you think the smooth muscle cells originate? Richard C. Pennell, MD (St. Louis, Missouri): When I was doing my vascular research fellowship, one of the things that we found was that arterial and venous endothelium do have different characteristics. Certainly, your experiment is designed to analyze the way things interact in the venous endothelium. Do you have any plans to perform ligated femoral artery bypasses in which you would try and test this same process compared with arterial endothelium rather than venous?
B. Timothy Baxter, MD (Omaha, Nebraska): In some senses, this is a preliminary study, because the end points involve really only looking at neointimal thickness and looking at one time point. However, I think the study is very important, and it’s also very timely for all of us, because we’re quite interested in the fact that using a vein patch at the end of a distal prosthetic graft may help to improve patency. The initial reports of this were uncontrolled trials from Europe, where they have a lot more experience with it. But we know now from control trials in below-knee prosthetic femoropopliteal bypasses, that the vein collars do improve patency, and there’s a strong suggestion that that’s also true in the tibial vessels. We’re also all encountering more patients in which we’re having a more difficult time finding enough vein to do the redo-redo procedures. This study also included immunohistochemical analysis of the cell types in the pannus. The investigators report a decrease in thickness in the Taylor patch group compared with the control group. In both areas where the measurements were made, the differences were significant, although the actual thickness was quite small. I have several questions for the investigators. First of all, a methodologic question: Why were more counts done in the control group than in the Taylor patch group, and with the measurements made in the venous outflow, were they done in a blinded fashion? The second question is, did you look at proliferation in some way? The third question, how did you choose the time frame used in your study? Fourth question; the patch would appear that it might change the angle of the anastomosis. Did you correct for this in the control group? Larry W. Kraiss, MD (Salt Lake City, Utah): The vein patch is already being used clinically in prosthetic lower extremity bypasses. This was an a-v fistula model. Has your group applied the vein patch to the venous outflow anastomoses of hemodialysis access grafts, and if so, do you have any impression on the effect on patency that the patch might have? Glenn C. Hunter, MD (San Antonio, Texas): I was intrigued by the last immunohistochemistry slide. All the smooth muscle cells seemed to be aligned closer to the 606
CLOSING Andrew T. Gentile, MD: Dr. Baxter, you had several good questions regarding methodology of the analysis. There were just quantitatively many more sites of intimal thickening within the control grafts than in the patched grafts themselves. So statistically there was much more ability to look at the intimal thickening from the central grafts. The actual measurements were at every single spot that I was able to find intimal thickening. That’s why there was a difference in the number of data points for comparison. The power of the statistics is provided by multiple observations. There were nearly a thousand data points through the venous anastomoses and almost a hundred data points through the venous outflow segment. You asked about proliferative activity in these grafts. Proliferation assays were not specifically performed in this model for several reasons. We have done some studies with PCNA, proliferating cell nuclear antigen, in human models of explanted vein graft stenotic segments, and that has shown that there is a very low level of proliferative activity. In other experimental projects, using rat vein graft stenosis models, there is a moderate increase in proliferative activity in these particular lesions. In the dog model, it’s unknown. There happens to be fewer monoclonal antibodies available for canine studies, and we haven’t been able to successfully cross-react the antibodies of the human with the canine model. You asked about the time frame of the explant. This was
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
chosen based on other previous studies, both at our institution and others, with a similar flow model. In the dog A-V fistula model, the intimal thickness increases and probably peaks somewhere between 5 and 6 weeks. There is believed to be a leveling off or a plateauing of the intimal thickening, and then actually there may be some reduction in intimal thickening with time. I think it’s a good suggestion to do another series and follow them out longer and see if this process continues. I think the origin of the cells that are present most likely is from the pannus. It has been shown years ago that the recipient vessel does promote healing at the area of the cut edge, and this pannus was demonstrated. It appears to be composed mostly of smooth muscle cells or cells that express smooth muscle cell alpha actin antibody. And there were areas of some myxoid type of matrix within the grafts, but other immunochemical stains for matrix components were not performed. The angle of the anastomosis is probably a critical factor, not only in the velocity turbulence measurements, but also in how the graph will lie. Some investigators from Joseph Archie’s lab have shown in an engineering computer-based flow model that an angle of the heel coming down on the recipient vessel of about 10 to 15 degrees may be the optimal angle.
This angle is almost reproduced with the Taylor patch. By patching a piece of vein over the anastomosis, you are able to really bring a much smoother tapered anastomosis and get approximately that same angle. I think this does bring up several more questions, which we’re looking at. The angles for the control grafts were not modified in any way. They were just standard anastomoses. Dr. Kraiss asked about our experience with Taylor patch in human patients. We’ve had only anecdotal experience with several patients undergoing repeat leg bypass without useable lengths of saphenous or other alternative sources of vein. We don’t really have any data to support that, but we have proposed that A-V fistula Taylor patched grafts be placed and followed up in selected patients. Dr. Hunter questioned about the smooth muscle cell origin. I have to presume it’s from the recipient vessel. Dr. Pennell also asked about if this has been done in an arterial model. In our experience, the A-V fistula model is a pretty good representation of the intimal thickening from a flow-induced model. We haven’t done them in arterial lesions. It can be done. Several other animal projects have been reported in the last couple of years looking at carotid interposition grafts with patched grafts, and they show very similar histology and morphology to the canine A-V fistula model.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
607
PURPOSE: Modified anastomotic techniques utilizing autogenous vein-cuffs or patches have been devised with the hope of improving prosthetic graft patency. The mechanisms of the presumed improvement in patched anastomoses have never been elucidated and remain speculative. We characterized the healing response of the Taylor vein patch in prosthetic arteriovenous fistulae in a canine model of intimal hyperplasia. METHODS: Six adult dogs underwent placement of bilateral (6 patched, 6 control) 4-mm diameter expanded polytetrafluoroethylene loop femoral artery-vein fistulae. Serial duplex ultrasound examinations confirmed graft patency until explant at 6 weeks. Differential light microscopy with computerized image analysis was performed on serial 5-mm sections. Intimal thickness through the venous anastomosis and outflow veins of Taylor patch and control (nonpatched) grafts were compared. Cell type-specific immunocytochemical antibody stains for smooth muscle cells (a SMC actin) and endothelial cells (von Willebrand factor) were performed. RESULTS: Eleven of 12 grafts remained patent for 6 weeks, 1 control graft thrombosed. Mean duplex-derived peak systolic velocities of patched (96 cm/sec) and control (108 cm/sec) grafts were similar. Microscopy revealed more intimal pannus anastomotic suture line ingrowth in controls (mean thickness 5 178 mm) than Taylor patched grafts (mean 147 mm, p 5 0.0002). Significantly less intimal thickening was present in the outflow vein of patched (mean thickness 5 90 mm) versus control grafts (mean 195 mm, P <0.0001). The intima maintained a single cell layer of vWF 1 endothelial cells, while the majority of the cells comprising the lesion expressed a SMC actin.
From the Division of Vascular Surgery (ATG, JLM, MAG, RDH, SSB, JDH) and the Department of Surgical Research (LBK, SKW), University of Arizona Health Sciences Center, Tucson, Arizona. This work was supported in part by a Research Development Grant from the Impra Pharmaceuticals Company, Tempe, Arizona. Requests for reprints should be addressed to Joseph L. Mills, MD, Professor of Surgery, Chief, Division of Vascular Surgery, AHSC PO Box 245072, 1501 N. Campbell Avenue, Tucson, Arizona 85724-5072. Presented at the 50th Annual Meeting of the South western Surgical Congress, San Antonio, Texas, April 19 –22, 1998. © 1998 by Excerpta Medica, Inc. All rights reserved.
CONCLUSION: Perianastomotic pannus is primarily composed of intimal smooth muscle cells. Neointimal thickening is significantly reduced in prosthetic arteriovenous fistulae created with the Taylor vein patch in a canine model. Reduction in perianastomotic intimal thickening may explain the reported clinical improvement in prosthetic bypass graft patency when modified with vein patch techniques. Am J Surg. 1998;176: 601– 607. © 1998 by Excerpta Medica, Inc.
I
n an effort to increase the patency of vascular reconstructions requiring the use of prosthetic bypass conduits, various graft configurations and anastomotic modifications have been proposed and applied clinically in several small human trials. The most common reasons infrainguinal prosthetic grafts are placed are lack or insufficient length of usable greater saphenous vein. Although results of lower extremity revascularization with autogenous vein are clearly superior to prosthetics with at least a 30% difference in primary 5-year patency,1 approximately 10% to 30% of patients do not have suitable saphenous vein for lower extremity arterial bypass.2,3 The failure of prosthetic grafts in distal reconstructions is thought to result largely from fibrointimal hyperplasia at the distal anastomosis.4 The concept of a prosthetic/autogenous composite was developed in response to the generally poor patency achieved with prosthetic grafts for distal arterial bypass in the setting of lack of suitable vein. Prostheticvein composite graft patencies are generally intermediate between those of all-autogenous and all prosthetic bypass grafts,5 with reported mean 2-year cumulative patencies ranging from 30% to 64%.6,7 Theories abound concerning the pathogenesis of the intimal lesions responsible for prosthetic, as well as autogenous graft-anastomotic failures. Various operative adjuncts, including vein cuffs, collars, and patches, have been placed in humans to modify these perianastomotic factors and improve graft patency. The mechanisms of the presumed improvement in patched anastomoses have never been elucidated and remain speculative. The common denominator in prosthetic graft failure appears to be related to the perianastomotic healing response of the recipient vessel. This is true not only for arterial reconstructions but also for prosthetic arteriovenous fistuae with venous outflow stenosis responsible for the majority of hemodialysis fistula failure. In order to characterize the healing response of a modified prosthetic anastomosis, Taylor vein patched arteriovenous fistulae were implanted in a canine model of intimal hyperplasia. 0002-9610/98/$19.00 PII S0002-9610(98)00286-4
601
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
MATERIALS AND METHODS Six adult mongrel dogs of either gender each weighing 24 to 28 kg were used in this study. Study design and research methods were approved by the Institutional Animal Care and Use Committee at the University of Arizona Health Sciences Center. All animal experiments were performed in accordance with the “Principles of Laboratory Animal Care and the Guide for the Care and Use of Laboratory Animals” (NIH Publication no. 80-23, 1985). Graft Implantation Dogs underwent preoperative general medical examination and were found fit to undergo femoral artery to vein prosthetic graft implantation. Prior to surgery, each animal received 50 mg dipyridamole orally twice per day for 4 days and 325 mg aspirin orally once per day for 1 day. These medications were continued during the entire postoperative period. All animals received perioperative antibiotics. On the day of surgery, animals were sedated with a mixture of atropine, ketamine, and acepromazine injected intramuscularly. General anesthesia was induced using intravenous sodium pentothal, and the animal was intubated and maintained on mechanical ventilation with a mixture of inhaled halothane and oxygen. Grafts were prepared using standard, 4-mm internal diameter, expanded polytetrafluoroethylene (e-PTFE; IMPRA Inc. Tempe, Arizona). Graft implantation was performed through diagonal, femoral incisions with standard vascular surgical techniques. Endgraft to side-femoral arterial and venous anastomoses were performed (6 study and 6 controls). A single segment of native internal jugular vein was harvested, flushed, opened, and cut to a diamond shape for autogenous Taylor patch as previously described.8 Six Taylor vein-patch anastomoses were constructed for the modified venous-outflow anastomosis. All anastomoses were performed with 6-0 Prolene suture (Ethicon Inc., Somerville, New Jersey). Graft Explant Serial duplex ultrasonography was performed to confirm graft patency and measure graft blood flow velocities at implant, 3 weeks, and immediately preceding graft explant at 6 weeks. At explant, animals received identical preoperative sedation and general anesthesia. Dogs were systemically heparinized, and the grafts were explanted in-toto along with approximately 3 cm of inflow artery and outflow vein segments. Grafts were gently flushed with heparinized saline and pressure fixed at 80- to 100 mm Hg with 10% neutral buffered formalin. After fixation, grafts were sequentially sectioned in the direction of blood flow through both anastomoses. Graft segments were dehydrated and paraffin embedded. Five-micron cross sections were floated on glass slides, deparaffinized, and stained with hematoxylin & eosin, and Masson’s trichrome. Using the Nikon Optiphot microscope with image transfer through a DAGE SIT-66 low light level camera, serial micrograph cross sections of the venous anastomoses were compared at 20 to 403 magnification. Analysis of the histologic characteristics and neointimal layer thickness was performed using digitized computerized image analysis (MetaMorph Image Analysis Systems) program linked through Microsoft Windows Applications. Serial cross sec602
Figure 1. Cross section of a well-incorporated (light, outer adventitial covering) explanted ePTFE graft (dark middle layer) with computerized image analysis for determination of mean thickness of the neointimal lining (light, hatched inner luminal layer; 320).
Figure 2. Schematic drawing of Taylor-patched arteriovenous anastomosis denoting multiple sample sites used to measure neointimal thickness at cross sections (narrow lines) through the anastomosis and separate sample area at outflow vein segments (wide line).
tions were compared through the venous anastomosis of both control and Taylor patched grafts. The image analysis software graphics allows calculation of neointimal area and average neointimal thickness was obtained for each graft segment (Figure 1). The average intimal thickness throughout the entire venous anastomosis as well as the intimal thickness in a clearly defined segment of outflow vein just distal to the anastomosis were compared between groups (Figure 2). All measurements were electronically transferred from MetaMorph to Microsoft Excel spreadsheets. Statistical comparison between control and study venous anastomosis and outflow vein segments were performed using the Student’s t test for continuous variables on standard PC compatible software (StatView; Abacus Concepts, Berkeley, California). A P value of #0.05 was considered statistically significant.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
Figure 3. Representative gross photograph at graft explant revealing a well-healed, tightly looped 4-mm ePTFE arteriovenous fistula. Vascular loops encircle inflow arterial (dark loops) and outflow vein (light loops) anatomy.
Immunohistologic Analysis Photomicrographs of cross sections were analyzed for morphologic features and neointimal pannus ingrowth. Single-label immunocytochemistry was performed after deparaffinization with xylene and rehydration in graded ethanol solutions. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide, and primary antibodies to either von Willebrand factor (vWF; kindly provided by Dr. James Catalfano, New York State Health Department, Albany, New York) or a smooth muscle cell actin (Sigma Co., Bushfield, Tennessee) were incubated at room temperature for 60 minutes. The sections were rinsed with phosphate buffered saline, a biotinylated secondary antibody was applied for 30 minutes, followed by a streptavidin peroxidase conjugate (Dako Corp. Carpinteria, California). Standard peroxidase enzyme substrate, 3939-diaminobenzidine or alkaline phosphatase substrate were added to yield a brown or red reaction product. Nuclei were lightly counterstained with dilute methyl green.
RESULTS This canine study compared the patency and healing characteristics of standard and a modified anastomotic configuration of small caliber prosthetic arteriovenous fistulae. All dogs tolerated the operative procedures without adverse events. Eleven of 12 grafts remained patent for 6 weeks, 1 control graft thrombosed. Mean duplex-derived perianastomotic peak systolic velocities of patched (96 6 22 cm/sec) and control (108 6 24 cm/sec) grafts were similar. At explant, all grafts were well incorporated without signs of infection. The grafts were implanted as smooth U-shaped loops; however, during healing the grafts often formed tightly associated proximal and distal segments (Figure 3). Gross examination of the luminal surface of the grafts at the time of explant revealed clean body and midgraft segments without an obvious inner lining or associated blood components. The perianastomotic areas revealed only very small amounts of a glistening neointimal layer without adherence of thrombus or platelets. Microscopic Results Morphologic analysis of grafts from both groups revealed very clean luminal surfaces throughout most of the grafts.
The Taylor patch segments were well incorporated and without microscopic signs of infection. The patch remained viable as evidenced by normal appearing vein histology consisting of endothelium, organized intimal layer bounded by an inner elastic lamina and muscular medial layer surrounded by loose connective tissue and adventitia. There was minimal change in the intimal layer morphology of the Taylor patched segment when compared with normal internal jugular vein specimens (Figure 4A). The perianastomotic areas had obvious neointimal pannus ingrowth most often at the cut edge of the vein and anastomotic suture line. This perianastomotic neointimal pannus was qualitatively more prominent in control grafts than in the Taylor-patched grafts (Figure 4B). The anastomotic neointima contained a layer of well-organized smooth muscle cells covered by a single layer of polygonalshaped endothelial cells. Since grafts were sectioned transversely, representative cross sections were analyzed to quantify the neointima within the inner surface of the grafts through the anastomosis and at the outflow vein segments. Microscopy revealed distinct areas of neointimal ingrowth starting at the heel of the PTFE with neointimal ingrowth at the margins of the graft into the recipient vessel. Quantitatively there was more intimal pannus anastomotic suture line ingrowth in controls (mean thickness 5 178 6 129 mm) than Taylor patched grafts (mean 147 6 103 mm, P 5 0.0002; Table). In addition, significantly less intimal thickness was present in the outflow vein of patched (mean thickness 5 90 6 62 mm) versus control grafts (mean 195 6 107 mm, P ,0.0001). Immunohistologic Results Single-label immunocytochemical staining produced positive staining of anti-vWf antibody in areas of neointima covered by endothelial cells and staining for anti-a SMC actin in subendothelial regions. The neointima maintained an organized morphology with a confluent single cell layer of vWF 1 endothelial cells along the neointima lining the luminal surface of the grafts, (Figure 4C) while the majority of the cells comprising the neointimal lesion expressed a SMC actin consistent with smooth muscle cell ingrowth. (Figure 4D)
COMMENTS Polytetrafluoroethylene grafts have been used extensively for infrainguinal vascular reconstructions either as the conduit of choice or as an alternative conduit when saphenous vein is unavailable. Although studies have reported acceptable early patency rates, the long-term efficacy of these grafts has been less satisfactory, especially in the infrapopliteal position. Results of nearly all studies concur that infrainguinal PTFE prostheses provide significantly inferior results when compared with autogenous vein.1,9 –11 Rutherford et al12 analyzed the factors affecting the patency of infrainguinal bypass grafts and confirmed the often held dictum that the quality of graft conduit and the anastomotic level (below-knee versus above-knee) are two important variables affecting bypass performance. Because results for infrapopliteal revascularization with prosthetic grafts have been so disappointing, some authors have considered primary amputation preferable to infrapopliteal ar-
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
603
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
Figure 4. Photomicrographs demonstrating perianastomotic intimal pannus. A. Cross section through a Taylor-patched anastomosis at 103 reveals a normal appearing vein patch segment (arrow) and outflow vein segment (*) with minimal intimal thickening. B. At higher magnification (Trichrome stain, 320) neointimal pannus ingrowth into the PTFE is observed at anastomotic heel regions and along prolene suture lines. C. Immunocytochemical staining for anti-vWF antibody demonstrates dark staining endothelial monolayer. D. The majority of the neointimal lesion is composed of cells staining positive for anti-a SMC actin antibody.
TABLE Canine Taylor Patch Mean Intimal Thickness Measurements Graft Taylor patch (n 5 6) Control (n 5 5) Significance t test
Venous Anastomosis
Venous Outflow
147 6 103 mm (318 counts)* 178 6 129 mm (588 counts) P 5 0.0002
90 6 62 mm (31 counts) 195 6 107 mm (64 counts) P ,0.0001
* Counts 5 individual thickness measurements.
tery bypass with PTFE in patients without usable autogenous vein.13 Recently, however, when faced with the more frequent clinical dilemma of elderly patients requiring secondary extremity revascularizations and lacking adequate autogenous conduit, investigators have applied various technical adjuncts in hopes of improving prosthetic bypass patency rates and limb salvage. These techniques include various vein patches, cuffs, or collars, and distal arteriovenous fistulae.14 Early uncontrolled studies of patency rates of PTFE with vein cuffs in the femoral-distal position reported encouraging initial results. Miller et al (1984)15 and Tyrell and Wolfe (1991)16 reported 72% and 78% 8 to 12 month patencies, respectively. Subsequent work on the influence of vein cuffs on PTFE femoral-popliteal bypass by Raptis 604
and Miller17 retrospectively reported no difference in above-knee PTFE patency with or without vein cuffs at 36 months (68% versus 69%), but a significant improvement in below-knee PTFE patency was demonstrated with the use of adjuvant vein cuffs (57% versus 29%). In a recent randomized, prospective trial reported by Stonebridge et al,18 vein cuffed below-knee prosthetic bypass produced a 65% 2-year patency rate compared with 29% patency in uncuffed grafts, with a sixfold greater early graft failure rate in the uncuffed grafts (13%) versus cuffed grafts (2%). The most promising results of a vein cuff clinical series, by Pappas et al,19 report 62% cumulative 2-year patency for 23 infrapopliteal cuffed prosthetic grafts. Reported results of uncontrolled series of infrainguinal bypasses using the Taylor vein patch approach those of revascularizations
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
with autogenous conduit, with an overall primary patency rate of 65% at 5 years.8 We have had anecdotal experience with distal anastomotic graft modifications in several patients requiring infrapopliteal bypass without usable vein, and have been impressed by the proliferation of clinical reports, trials and animal studies on the potential for vein collars and other enhancements to improve infrainguinal (and possibly arteriovenous dialysis fistula) prosthetic bypass results. The most widely held opinions on the etiology of prosthetic bypass graft failure have focused on graft narrowing from distal perianastomotic intimal hyperplasia.4 Theories abound concerning the pathogenesis of the intimal lesions responsible for prosthetic, as well as autogenous graft-anastomotic failures. The rationale for the addition of a vein cuff or patch to the distal anastomosis is to limit neointimal hyperplastic ingrowth. Plausible, though unproven, explanations have described various hemodynamic causes such as abnormal velocity patterns, turbulence, and shear stress;20 mechanical causes such as compliance mismatch, geometry, and technical factors,21 all of which may contribute to the (mal)adaptive vascular remodeling of the perianastomotic area. The mechanisms of the reported clinical patency improvement in vein-patched or cuffed anastomoses have never been elucidated and remain speculative. The common denominator in prosthetic graft failure appears to be the perianastomotic healing response of the recipient vessel. This current report was designed to investigate the healing properties of the Taylor vein patched anastomosis in a well-characterized canine model of intimal hyperplasia.22 Although we noted no difference in initial patency of control and study grafts, the addition of the Taylor patch had a significant affect on limiting neointimal pannus ingrowth at the perianastomotic suture line, and also importantly decreases intimal thickening at the juxta-anastomotic venous outflow segments. This decrease in neointimal thickening associated with a modified vascular anastomosis in this experimental model of intimal hyperplasia, the canine arteriovenous fistula, has not been previously demonstrated. Suggs et al23 demonstrated decreased patency of non-cuffed PTFE but similar histologic findings in a canine model using small caliber PTFE carotid bypass grafts. Norberto et al24 demonstrated no histologic difference in jacketed-vein cuffed and non-cuffed carotid PTFE anastomoses, concluded the protective effect of vein cuffs are not mechanical, and suggested this effect may be humoral in nature. We were able to clearly demonstrate a significant reduction in neointimal thickening in prosthetic arteriovenous fistulae created with the Taylor vein patch in a canine model. Microscopic and immunocytochemical analysis demonstrated that perianastomotic pannus is primarily composed of intimal smooth muscle cells. The luminal surface of the neointimal pannus maintains a confluent monolayer of endothelial cells. The juxta-anastomotic neointimal pannus is presumed to be derived from smooth muscle cell migration from the adjacent recipient vessel. Although this project demonstrated a reduction in neointimal thickening associated with vein-patched prosthetic anastomoses, the precise mechanisms responsible for this
difference were not elucidated, and further investigations into hemodynamic and/or humoral mechanisms are needed to explain these findings. One proposed weakness for the Taylor patched type of anastomosis is that approximately one third (heel portion) of the prosthetic graft is in contact with the recipient vessel and may be affected by suture line pannus ingrowth. Other investigators have concluded that results of all composite vein-PTFE anastomoses will be affected by myointimal hyperplasia, thus shifting this graft-threatening process more proximally in the graft. By bringing the autogenous patch across the heel of the anastomosis, the Taylor patch provides a somewhat patulous anastomotic hood, which may make any neointimal thickening at the suture line insignificant in affecting graft patency. We demonstrated an altered healing response of vein-patched anastomoses with reduction in perianastomotic intimal thickening when compared with controls. Operative adjuncts to prosthetic graft anastomoses may limit neointimal thickening by providing an “anastomotic buffer zone” and may explain the reported clinical improvement in prosthetic bypass graft patency when modified with vein patch techniques. Further investigation of the precise mechanisms for these findings is ongoing.
REFERENCES 1. Veith FJ, Gupta SK, Ascer EA, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J Vasc Surg. 1986;3:104 –114. 2. Wolfe JHN, Tyrell MR. Justifying arterial reconstruction to crural vessels-even with prosthetic graft. Br J Surg. 1991;78:897– 899. 3. Brewster DC, Lasalle AJ, Robinson JG, et al. Femoropopliteal graft failures: Clinical consequences and success of secondary reconstructions. Arch Surg. 1983;118:1043–1047. 4. Clowes AW, Gown AM, Hanson SR, Reidy MA. Mechanisms of arterial graft failure. I. Role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol. 1985;118:43–54. 5. Demasi RJ, Snyder SO. The current status of prosthetic-vein composite grafts for lower extremity revascularization. Surg Clin North Am. 1995;75:741–752. 6. Feinberg RI, Winter RP, Wheeler JR, et al. The use of composite grafts in femorcrural bypasses performed for limb salvage: a review of 108 consecutive cases and comparison with 57 in situ saphenous vein bypasses. J Vasc Surg. 1990;12:257–263. 7. Gregory RT, Raithel D, Snyder SO, et al. Composite grafts: an alternative to saphenous vein for lower extremity arterial reconstruction. J Cardiovasc Surg. 1983;24:53–57. 8. Taylor RS, Loh A, McFarland RJ, et al. Improved technique for polytetrafluoroethylene bypass grafting: long-term results using anastomotic vein patches. Br J Surg. 1992;79:348 –354. 9. Whittemore AD, Kent KC, Donaldson MC, et al. What is the proper role of polytetrafluoroethylene grafts in infrainguinal reconstruction. J Vasc Surg. 1989;10:299 –305. 10. Veith FJ, Gupta SK, Ascer E, et al. Six-year prospective multicenter randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial reconstructions. J Vasc Surg. 1986;3:104 –114. 11. Hobson RW II, Lynch TG, Zafar J, et al. Results of revascularization and amputation in severe lower extremity ischemia: a five-year clinical experience. J Vasc Surg. 1985;2:174 –185. 12. Rutherford RD, Jones DN, Bergentz SE, et al. Factors affecting the patency of infrainguinal bypass. J Vasc Surg. 1988;8:236 –246. 13. Yeager RA, Hobson RW II, Jamil Z, et al. Differential patency
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
605
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
and limb salvage for polytetrafluoroethylene and autogenous saphenous vein in severe lower extremity ischemia. Surgery. 1982;91:99 – 103. 14. Webb T. Operative adjuncts for distal revascularization. Surg Clin North Am. 1995;75:753–758. 15. Miller JH, Foreman RK, Ferguson L, Faris I. Interposition vein cuff for anastomosis of prosthesis to small arteries. Aust NZ J Surg. 1984;54:283–285. 16. Tyrell MR, Wolfe JHN. New prosthetic venous collar anastomotic technique: combining the best of the other procedures. Br J Surg. 1991;78:1016 –1017. 17. Raptis S, Miller JH. Influence of a vein cuff on polytetrafluoroethylene graft for primary femoropopliteal bypass. Br J Surg. 1995;82:487– 491. 18. Stonebridge PA, Prescott RJ, Ruckley CV. Randomized trial comparing infrainguinal polytetrafluoroethylene bypass grafting with and without vein interposition cuff at the distal anastomosis. J Vasc Surg. 1997;26:543–550.
19. Pappas PJ, Hobson RW II, Meyers MG, et al. Patency of infrainguinal polytetrafluoroethylene bypass grafts with distal interposition vein cuffs. Cardio Vasc Surg. 1998;6:19 –26. 20. Binns RL, Ku DN, Stewart MT, et al. Optimal graft diameter: effect of wall shear stress on vascular healing. J Vasc Surg. 1989; 10:326 –337. 21. Abbott WM, Megerman J, Hasson JE, et al. Effect of compliance mismatch on vascular graft patency. J Vasc Surg. 1987;5:376 – 381. 22. Williams SK, Jarrell BE, Kleinert LB. Endothelial cell transplantation onto polymeric arteriovenous grafts evaluated using a canine model. J Invest Surg. 1994;7:503–517. 23. Suggs WD, Henriques HF, DePalma RG. Vein cuff interposition prevents juxta-anastomotic neointimal hyperplasia. Ann Surg. 1988;207:717–723. 24. Norberto JJ, Sidawy AN, Trad KS, et al. The protective effect of vein cuffed anastomoses in not mechanical in origin. J Vasc Surg. 1995;21:558 –566.
DISCUSSION
luminal surface, with a relatively acellular zone between the graft and the smooth muscle cells, what appeared to be like matrix. Where do you think the smooth muscle cells originate? Richard C. Pennell, MD (St. Louis, Missouri): When I was doing my vascular research fellowship, one of the things that we found was that arterial and venous endothelium do have different characteristics. Certainly, your experiment is designed to analyze the way things interact in the venous endothelium. Do you have any plans to perform ligated femoral artery bypasses in which you would try and test this same process compared with arterial endothelium rather than venous?
B. Timothy Baxter, MD (Omaha, Nebraska): In some senses, this is a preliminary study, because the end points involve really only looking at neointimal thickness and looking at one time point. However, I think the study is very important, and it’s also very timely for all of us, because we’re quite interested in the fact that using a vein patch at the end of a distal prosthetic graft may help to improve patency. The initial reports of this were uncontrolled trials from Europe, where they have a lot more experience with it. But we know now from control trials in below-knee prosthetic femoropopliteal bypasses, that the vein collars do improve patency, and there’s a strong suggestion that that’s also true in the tibial vessels. We’re also all encountering more patients in which we’re having a more difficult time finding enough vein to do the redo-redo procedures. This study also included immunohistochemical analysis of the cell types in the pannus. The investigators report a decrease in thickness in the Taylor patch group compared with the control group. In both areas where the measurements were made, the differences were significant, although the actual thickness was quite small. I have several questions for the investigators. First of all, a methodologic question: Why were more counts done in the control group than in the Taylor patch group, and with the measurements made in the venous outflow, were they done in a blinded fashion? The second question is, did you look at proliferation in some way? The third question, how did you choose the time frame used in your study? Fourth question; the patch would appear that it might change the angle of the anastomosis. Did you correct for this in the control group? Larry W. Kraiss, MD (Salt Lake City, Utah): The vein patch is already being used clinically in prosthetic lower extremity bypasses. This was an a-v fistula model. Has your group applied the vein patch to the venous outflow anastomoses of hemodialysis access grafts, and if so, do you have any impression on the effect on patency that the patch might have? Glenn C. Hunter, MD (San Antonio, Texas): I was intrigued by the last immunohistochemistry slide. All the smooth muscle cells seemed to be aligned closer to the 606
CLOSING Andrew T. Gentile, MD: Dr. Baxter, you had several good questions regarding methodology of the analysis. There were just quantitatively many more sites of intimal thickening within the control grafts than in the patched grafts themselves. So statistically there was much more ability to look at the intimal thickening from the central grafts. The actual measurements were at every single spot that I was able to find intimal thickening. That’s why there was a difference in the number of data points for comparison. The power of the statistics is provided by multiple observations. There were nearly a thousand data points through the venous anastomoses and almost a hundred data points through the venous outflow segment. You asked about proliferative activity in these grafts. Proliferation assays were not specifically performed in this model for several reasons. We have done some studies with PCNA, proliferating cell nuclear antigen, in human models of explanted vein graft stenotic segments, and that has shown that there is a very low level of proliferative activity. In other experimental projects, using rat vein graft stenosis models, there is a moderate increase in proliferative activity in these particular lesions. In the dog model, it’s unknown. There happens to be fewer monoclonal antibodies available for canine studies, and we haven’t been able to successfully cross-react the antibodies of the human with the canine model. You asked about the time frame of the explant. This was
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
VEIN PATCHING REDUCES NEOINTIMAL THICKENING/GENTILE ET AL.
chosen based on other previous studies, both at our institution and others, with a similar flow model. In the dog A-V fistula model, the intimal thickness increases and probably peaks somewhere between 5 and 6 weeks. There is believed to be a leveling off or a plateauing of the intimal thickening, and then actually there may be some reduction in intimal thickening with time. I think it’s a good suggestion to do another series and follow them out longer and see if this process continues. I think the origin of the cells that are present most likely is from the pannus. It has been shown years ago that the recipient vessel does promote healing at the area of the cut edge, and this pannus was demonstrated. It appears to be composed mostly of smooth muscle cells or cells that express smooth muscle cell alpha actin antibody. And there were areas of some myxoid type of matrix within the grafts, but other immunochemical stains for matrix components were not performed. The angle of the anastomosis is probably a critical factor, not only in the velocity turbulence measurements, but also in how the graph will lie. Some investigators from Joseph Archie’s lab have shown in an engineering computer-based flow model that an angle of the heel coming down on the recipient vessel of about 10 to 15 degrees may be the optimal angle.
This angle is almost reproduced with the Taylor patch. By patching a piece of vein over the anastomosis, you are able to really bring a much smoother tapered anastomosis and get approximately that same angle. I think this does bring up several more questions, which we’re looking at. The angles for the control grafts were not modified in any way. They were just standard anastomoses. Dr. Kraiss asked about our experience with Taylor patch in human patients. We’ve had only anecdotal experience with several patients undergoing repeat leg bypass without useable lengths of saphenous or other alternative sources of vein. We don’t really have any data to support that, but we have proposed that A-V fistula Taylor patched grafts be placed and followed up in selected patients. Dr. Hunter questioned about the smooth muscle cell origin. I have to presume it’s from the recipient vessel. Dr. Pennell also asked about if this has been done in an arterial model. In our experience, the A-V fistula model is a pretty good representation of the intimal thickening from a flow-induced model. We haven’t done them in arterial lesions. It can be done. Several other animal projects have been reported in the last couple of years looking at carotid interposition grafts with patched grafts, and they show very similar histology and morphology to the canine A-V fistula model.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
607