In the absence of suitable autogenous tissue, syn-thetic vascular grafts are the primary alternative foruse in the construction of access grafts for hemodial-

ysis. Unfortunately, the poor patency of prostheticarteriovenous (AV) grafts and the subsequent treat-ment of graft thrombosis contribute to significantmorbidity and mortality in the growing number ofpatients with end-stage renal disease worldwide.1-4

Stenosis at the venous anastomosis accounts formost AV access thrombosis.3-5 Presently, polypropy-lene suture is used to approximate prosthetic grafts tonative vessel. In addition to being time consuming,sutured anastomoses can subject the vessel to factorsassociated with increased risk of intimal thickening.Perforations of the native vessel wall expose thrombo-genic extracellular matrix to the turbulent flowobserved in AV access grafts.7-9 Moreover, the sutures

Anastomotic tissue response associatedwith expanded polytetrafluoroethyleneaccess grafts constructed by usingnonpenetrating clipsDonny B. Dal Ponte, MS, Scott S. Berman, MD, Vangie B. Patula, Leigh B.Kleinert, and Stuart K. Williams, PhD, Tucson, Ariz

Purpose: The gross, light microscopic, and scanning microscopic appearance of arterialand venous anastomoses in expanded polytetrafluoroethylene (ePTFE) access grafts con-structed with nonpenetrating clips were compared with that of those constructed withpolypropylene suture. We hypothesized that clip-constructed anastomoses would pro-vide controlled approximation of native vessel intimal and medial components with theePTFE grafts. We further hypothesized that anastomotic healing with clips wouldinvolve primarily an intimal cellular response, as compared with suture-constructedanastomoses in which cells within the media and adventitia walls participate. Methods: Femoral artery to femoral vein arteriovenous (AV) grafts were constructed infive dogs using 4-mm internal diameter ePTFE graft material. Each animal received oneAV graft with anastomoses constructed by using polypropylene sutures in one leg andone AV graft with anastomoses constructed with Vascular Closure System clips in thecontralateral leg. Animals were given aspirin for the duration of the study, and graftswere explanted at 5 weeks. At the time of explantation, graft segments were grossly eval-uated and then underwent light and scanning electron microscopic analysis.Results: At the time of explantation, all access grafts were patent. Joining the ePTFEgrafts to the native vessels with clips resulted in minimal vessel wall damage. The lume-nal contours of the discontinuous approximation were smooth and without grossendothelial disruption. These observations are in contrast to the lumenal compromiseand endothelial disturbance associated with the sutured anastomoses. Furthermore,hemostasis was achieved immediately in the clipped grafts, decreasing the incidence ofperianastomic hematoma. Finally, cellular reconstitution occurred at the anastomoticcleft in both the sutured and the clipped junctions. The neointima exhibited an endothe-lial cell lining on the lumenal surface and the presence of α-smooth muscle cell actin pos-itive cells within the subendothelial layer. Conclusion: Vascular Closure System clips are a viable alternative to suture for the approx-imation of ePTFE AV access grafts to native blood vessels. The use of the clips resultedin a more streamlined anastomosis, with decreased vessel wall damage, immediate hemo-stasis, and a trend toward shorter procedure times. (J Vasc Surg 1999;30:325-33.)

325

From the Department of Biomedical Engineering, University ofArizona.

Supported by a grant from the United States SurgicalCorporation, Norwalk, Conn.

Reprint requests: Dr Stuart K. Williams, Biomedical Engineering,University of Arizona, 1501 N Campbell Ave, Tucson, AZ85724.

Copyright © 1999 by The Society for Vascular Surgery andInternational Society for Cardiovascular Surgery, NorthAmerican Chapter.

0741-524/99/$8.00 + 0 24/1/95423

themselves project into the anastomosis, potentiallyirritating the inside of the vessel, triggering thrombo-sis. Closely spaced sutures may cause mural ischemiawith resultant necrosis of the vessel margins.10,11

Needle-hole bleeding and blood seepage at suturedanastomoses also allow for localized platelet deposi-tion, hematoma formation, and secondary edema.Such irregularities may contribute to an environmentconducive to abnormal healing and intimal thickening.

Research has demonstrated that the healing atanastomoses performed with Vascular Closure System(VCS) clips is equivalent, if not superior, when com-pared with conventional, sutured anastomosis.11-13

The rapid hemostasis and antithrombogenic anasto-motic construction attained with VCS clips mayaccount, at least in part, for these findings. The cur-rent study compared the healing associated with vas-cular anastomoses made with VCS clips with that of vascular anastomoses made with conventionalpolypropylene suture in a canine model of AV access.

MATERIALS AND METHODSAnimal selection and welfare. All animal stud-

ies were performed in accordance with the protocolsapproved by the University of Arizona Animal ReviewCommittee and the National Institutes of Health’s“Guide for the Care and Use of Laboratory Animals”(NIH publication #85-23, revised 1985). All sur-geries were performed in accordance with and all ani-mals were housed in facilities approved by theAmerican Association for Accreditation of LaboratoryAnimal Care.

Performance of trial. Five mongrel dogs (20 to32 kg) were orally medicated with 325 mg of aspirin1 day before surgery and maintained on this dosagefor the duration of the study. On the day of surgery,animals were sedated with a mixture of atropine,ketamine, and acepromazine injected intramuscular-ly. General anesthesia was induced with intravenoussodium Pentothal, and the animals were intubated.Anesthesia was maintained with a mixture of inhaledhalothane and oxygen. Intravenous cephalexin (500mg) was given before the initial incision, and anoth-er 500 mg cephalexin was delivered intravenouslyapproximately 1 hour later. No analysis was per-formed to determine interanimal variability withrespect to platelet or coagulant reactivity.

Graft implantation. Femoral AV grafts wereconstructed using 25-cm segments of 4-mm internaldiameter, standard-walled expanded polytetrafluo-roethylene (ePTFE; W.L. Gore and Associates,Flagstaff, Ariz). Bilateral transverse incisions parallelto the inguinal ligament were made to expose the

femoral arteries and veins. Heparin (3000 units) wasadministered intravenously 5 minutes before vesselclamping. The femoral artery and femoral vein werethen occluded with vascular loops, and an AV shuntwas constructed with end-to-side anastomoses madewith 6-0 polypropylene (US Surgical, Norwalk,Conn) in one leg and VCS clips (US Surgical) in thecontralateral leg.

In all animals, the venous/prosthetic anasto-moses were performed first, followed by the arteri-al/prosthetic construction. The sutured anasto-moses were handsewn in a continuous stitch. 6-0polypropylene stay sutures were used to fix the“heel” and “toe” of the clipped anastomoses to facil-itate clip application. Vessel and graft edges werethen everted at 90 degrees by means of speciallydesigned approximation forceps and placed againsteach other. Medium-size clips were placed perpen-dicular to the suture line, as close to each other aspossible. If a clip was mis-set, a clip remover wasavailable for expedient replacement. The VCS clipapplier delivered the clips mechanically through adistributing cartridge of 25 titanium clips. Aftercompletion of the anastomoses, flow was restored,the skin lateral to the incision was separated from themuscle by means of a hemostat, the grafts weretacked to the muscle layer at the most distal point ofthe fistula with a single suture, and the skin wasclosed in two layers. In this layout, the outline of theAV fistula can be readily seen and palpated throughthe skin. Wounds were closed in two layers, and theanimals recovered while under direct observation for6 hours.

Two additional AV fistulas were implanted,immediately perfused-fixed, and explanted for theexpress purpose of gross anastomotic examination atthe time of implantation. These constructs areshown in Fig 2.

Patency evaluation. Graft patency was assessedby means of a flow probe (Transonic Systems,Ithaca, NY) immediately after graft implantationswere complete and flow was restored. Patency wasfurther evaluated at 2, 3, and 4 weeks by means of ahand-held Doppler ultrasound scanner (MedaSonics, Mountain View, Calif) and at explant bymeans of the flow probe.

Graft explantation. After 5 weeks, the animalswere anesthetized and prepared for surgery asdescribed earlier. The grafts were exposed andassessed for patency by means of direct needle punc-ture. Heparin (100 U/kg) was given, vessels wereclamped, and graft explantations were performed.Removed samples were immediately flushed with

JOURNAL OF VASCULAR SURGERY326 Dal Ponte et al August 1999

warm (37°C) Dulbecco’s cation free–phosphatebuffer saline containing 0.1% bovine serum albumin.Subsequently, the grafts were sectioned into seg-ments and prepared for light histology, scanning elec-tron microscopy (SEM), and immunocytochemistry.

Light microscopy. Representative samples ofgraft, surrounding tissue, and native vessels werefixed in Histochoice (AMRESCO, Solon, Ohio),dehydrated, and brought to paraffin. Clips wereremoved (in the clip group only) from the anasto-moses, and all samples were embedded into blocksand sectioned perpendicular to the suture line. Thesections were then deparaffinized and stained withboth hematoxylin/eosin and a modified Masson

trichrome stain. The slides were examined, and pho-tomicrographs were obtained with a Nikon Opti-phot microscope.

Immunochemistry. Paraffin-embedded sectionswere obtained and fixed on slides. Sections werethen reacted with primary antibodies to vonWillebrand factor (vWF; provided by Dr JamesCatalfamo, New York State Health Department,Albany, NY) or α-smooth muscle cell actin (αSMC;Sigma) and visualized by means of peroxidase conju-gated secondary antibody. Nuclei were lightly coun-terstained with methyl green. Sections were thenevaluated by means of a Nikon Optiphot micro-scope.

JOURNAL OF VASCULAR SURGERYVolume 30, Number 2 Dal Ponte et al 327

Fig 1. Light micrographs (original magnification, ×1) of AV fistulas joined to the native ves-sel with suture (A) or vascular closure system clips (B) shown at implantation (time, T = 0),before the release of the vascular loops. The suture holes associated with the conventional anas-tomoses allowed blood to flow freely for the first several minutes after the release of loops (C),whereas hemostasis was achieved almost instantaneously on release of vessel loops in the anas-tomoses constructed with the clips (D).

SutureT=0

SutureT=0

VCST=0

VCST=0

Scanning electron microscopy. Samples forelectron microscopic evaluation were fixed in 3%glutaraldehyde in PIPES (Piperazine-N,N′-bis-[2-ethanesulfonic acid]) buffer (pH, 7.4), dehydratedin a graded series of acetone, and critical point dried.Dry samples were sputter coated with a gold targetand evaluated by means of a JOEL 820 scanningelectron microscope.

Data analysis. The flow data were analyzed bymeans of an analysis of variance, followed by aBonferroni post hoc test. All data are expressed asmeans ± standard deviation.

RESULTSAfter graft implantation, all animals recovered

rapidly, with no indication of congestive heart failureor other postoperative complications related to thepresence of two high-flow AV grafts. The averageprocedure duration from application of the venousvessel loops to the release of all (venous and arterial)vessel loops subsequent to the completion of the AVfistulas was approximately 20 minutes for both thesutured and the clipped groups. The range of proce-dure duration for the sutured groups was 18 to 22minutes, whereas the range for the clipped groupwas 19 to 30 minutes. There was a trend towardshorter procedure times when the clips were used,and the authors believe that if more fistulas had beenconstructed, significantly reduced procedure dura-tions, such as those observed in other studies, couldhave been achieved.11,12 Nevertheless, no significanteffect of anastomosis approximation was observedon graft patency and/or thrombogenesis. All grafts(five sutured, five clipped) remained patent and weresubjected to further histologic and immunocyto-chemical evaluation after explantation.

Preflow findings. Representative AV fistulasjoined to the native vessel with suture or VCS clipsare shown in Fig 1 before the release of the vascularloops. The continuous stitch used in the suturedanastomoses resulted in vessel wall overlap and“purse stringing” in certain areas along the sutureline. This observation was noted more often at thevenous unions. Furthermore, vessel wall distentionwas noted in other areas. Such imperfections lead toan irregular anastomotic line (Fig 2A and C).Penetration of the vessel wall and protrusion of thesuture material into the vessel lumen was alsoobserved.

A representative anastomosis constructed withthe VCS clips, on the other hand, shows neither per-foration of the vessel wall nor obstruction of thelumen caused by a foreign body (Fig 2B and D).

Application of the clips resulted in focal compressionof the media between the clip tips. Furthermore, thelumenal contours of the clipped anastomoses weresmooth without endothelial disruption, in contrastto the lumenal compromise and endothelial distur-bances associated with the sutured anastomoses.

Postflow findings. Hemostasis was achievedalmost instantaneously on release of vessel loops inthe anastomoses constructed with the clips (Fig 1D),whereas the suture holes associated with the con-ventional anastomoses allowed blood to flow freelyfor the first several minutes after the release of theloops (Fig 1C). No statistically significant differ-ences were observed between the arterial flow ratesbefore the placement of the access grafts. Flow ratesdid significantly rise (P = .003) relative to preimplantvalues after AV graft construction in both thesutured and clipped groups. Nevertheless, no differ-ences were seen between the stapled and suturedgroups. At explantation, flow rates remained elevat-ed when compared with preimplant values (P =.001), but they were not statistically different fromthe flow rates recorded immediately after implanta-tion (Table I).

Similar to the findings of past studies, gross evalu-ation at explant showed no differences between thesutured and the clipped groups in lumen diameter.6,12

Within 5 weeks, a glistening white cellular surface wasevident at the anastomotic cleft of both the suturedand clipped anastomoses. A cellular lining that wasmorphologically consistent with endothelial cells wasrevealed by means of SEM (Fig 3F). No fibrin orplatelet deposits were noted on this lining, suggestinga nonthrombogenic tissue/blood interface.

This neointima was further characterized bymeans of antibodies against vWF (Fig 3E) andαSMC actin. The cells lining the lumenal surfacestained positive for vWF, suggesting an endothelialcell origin. αSMC-actin staining predominated inthe cells residing in the subendothelial cell layers. Wecannot definitively establish that these cells were ofsmooth muscle cell origin, because this actin subtypecan be expressed by endothelial cells under specificconditions.14

Fig 4 compares the variations in tissue approxi-mation seen with sutured anastomoses to a clippedanastomosis. Ideally, the vessel media abut the edgeof the prosthetic material, thereby excluding throm-bogenic matrix components of the media from theflow surface (Fig 4A).7,15 Because of the drastic dif-ference in compliance between the native vessel,whether artery or vein, and the prosthetic grafts, thisconfiguration is almost impossible to achieve, even

JOURNAL OF VASCULAR SURGERY328 Dal Ponte et al August 1999

with end-to-end anastomoses. Moreover, most anas-tomoses performed for AV access grafts involve anend-to-side construct. In this circumstance, thecommonly attained result is the vessel overlappingthe graft or vice versa (Fig 4C and D). However, amore desirable result, that of an everted vessel/graftjunction, is easily attained with clips (Fig 4B).

DISCUSSION

The experimental use of metal clips to performblood vessel anastomoses was first introduced morethan 40 years ago.6,11,12 At that time, the proceduredid not gain wide acceptance and fell into disfavor.More recently, however, the dissatisfaction withsuturing as the method for anastomosing the super-

JOURNAL OF VASCULAR SURGERYVolume 30, Number 2 Dal Ponte et al 329

Table I. Arterial flow-rate comparisons between the sutured and clipped groups before arteriovenousaccess graft implantation, immediately after graft placement, and at explantation

Pre Post Explant

Group Sutured Clipped Sutured Clipped Sutured Clipped

Flow rates 109.8 ± 50.1 109.0 ± 39.1 800.8 ± 130.3* 602.0 ± 253.9* 763.8 ± 251.7* 739.2 ± 223.9*(mL/min)

*Significantly different from prearteriovenous access flow rates.

Fig 2. Light (A and B; original magnification, ×2.5) and scanning electron micrographs (C and D; bar, 100 mm) looking down on the graft/vessel interface of sutured (A and C) andclipped (B and D) anastomoses immediately after implantation (time, T = 0). In all cases, thearterial anastomoses are displayed. Venous approximations appear identical and are thereforenot shown. g, vascular graft; v, native vessel.

SutureT=0

SutureT=0

VCST=0

VCST=0

ficial temporal artery to the middle cerebral arteryhas spurred the development of a nonpenetrating,arcuate clip design.11,12 To date, vessel clips havebeen successfully used in both laboratory and oper-

ating room settings.7 Endoscopic tissue reconstruc-tion, free-flap transfer, and skin grafting are just afew of the procedures in which clips have beenused.16 This particular study was performed to eval-

JOURNAL OF VASCULAR SURGERY330 Dal Ponte et al August 1999

Fig 3. Side-view scanning electron micrographs of sutured (A and C) and clipped (B and D)approximations after explantation (time, T = 5 weeks). Pannus ingrowth (C and D) was observedat all anastomoses. Immunohistochemical evaluation of this neointima with von Willebrand Factor(light micrographs, E; original magnification, ×25) revealed a lumenal layer of positively stainedcells. This monolayer was morphologically consistent with endothelial cells (scanning electronmicrograph F). In all cases, the arterial anastomoses are displayed. Bars, 100 µm; g, vascular graft;v, native vessel; s, staple.

SutureT=5wk

SutureT=5wk

VCST=5wk

VCST=5wk

SutureT=5wk

VCST=5wk

uate the efficacy of using vascular clips to approxi-mate synthetic vascular grafts to native blood vesselsin the construction of an AV fistula for hemodialysis.

Suture. Approximation of the vessel wall to thevascular graft material with suture resulted in imper-fections in the flow surface (Fig 1A and C; Fig 4A,C, and D). Furthermore, the sutures themselvesproject into the lumen of the anastomosis, causingvessel wall damage. Such damage can result in aninherently thrombogenic surface.7-11 Althoughimmediate AV graft thrombosis is an unusual clinicalevent, suture projections may serve as a nidus for

chronic buildup of fibrin and may establish a scaffoldfor matrix deposition and the subsequent stenosisoften observed at the venous anastomosis. Such aprocess may be accelerated in diseased vessels. Theturbulent blood flow characteristic of AV grafts canonly further exacerbate the activation of the throm-bus formation.17 What role these factors play in inti-mal thickening is largely undefined, but certainlymust be considered when mechanisms of anasto-motic stenosis are studied.

Finally, sutured anastomoses have a baseline levelof leakage through the needle holes, which can

JOURNAL OF VASCULAR SURGERYVolume 30, Number 2 Dal Ponte et al 331

Fig 4. Light micrographs of the tissue approximations routinely observed with sutured (A, C,and D) and clipped (B) anastomoses (original magnification, ×25) at 5 weeks. Ideally, the suturedvessel media abuts the edge of the prosthetic material (A), thereby excluding thrombogenicmatrix components of the media from the flow surface 157. More commonly sutured vesselsoverlap the graft or vice versa (C and D). A desirable result, that of an everted vessel/graft junc-tion, is easily attained with clips (B). In all cases, the arterial anastomoses are displayed. g, vascu-lar graft; v, native vessel.

SutureT=5wk

SutureT=5wk

VCST=5wk

SutureT=5wk

result in the formation of hematomas (Fig 2C).Although most of the hematomas are of no clinicalconsequence in blood loss or surgical-wound com-plications, they are accompanied by significantedema formation and are proinflammatory in thisrespect. Recruitment of inflammatory cells inducedby perianastomotic hematomas may also contributeto anastomotic stenosis by providing additionalmilieu components that promote prolific rather thancontrolled healing of the flow surface.18-20 Althoughthe current study did not specifically address thisissue, the striking difference between the anasto-motic blood loss in the sutured anastomoses and thevirtual absence of anastomotic bleeding seen withthe VCS anastomoses will need to be assessed in thecontext of graft failure.

Clips. The use of VCS clips bypasses many of theshortcomings of the sutured anastomoses. Becauseno vessel wall perforation occurs, the only area inwhich medial matrix components could be exposedto blood flow is the vessel edge. By necessity, thisedge is excluded from contact with flowing bloodbecause it is everted to accomplish a clipped anasto-mosis (Fig 3B). The resultant anastomotic ring iscomposed solely of an intima/graft interface withonly minimal exception—that being at the proximaland distal stay-suture sites. This positioning alonemay reduce the risk of intimal proliferation andthereby promote wound healing.21

Aside from the morphologic differences in flowsurface offered by using the VCS clip for vascularanastomoses, other advantages may be realized. Clipapplication is discontinuous; therefore, the opportu-nity does exist for vessel growth both in the longitu-dinal and transverse planes (Fig 2B). This may notbe critical in the AV graft application; however, thisadvantage may prove useful when anastomosing ves-sels that have not yet completed their growth, suchas those in children. This ability for anastomoticgrowth may also prove a useful characteristic in lightof the popularity of catheter-based balloon interven-tions to treat AV access failure. When coupled witha compliant graft material, the clipped anastomosismay prove to be a more intervention-friendly con-figuration for subsequent balloon angioplasty ofanastomotic stenosis.

In summary, this report has demonstrated theefficacy of constructing vascular anastomosis with theVCS clip in a canine model of AV access grafts.Compared with conventional continuous polypropy-lene-sutured vascular anastomoses, the VCS-clippedapproximations were more hemostatic, resulted in asmoother lumenal-flow profile, and presented a flush

endothelial cell/graft interface. Moreover, perianas-tomotic hematoma and subsequent edema werenonexistent with the VCS anastomoses. We believethese differences suggest that the VCS-clipped anas-tomoses will be less likely to develop anastomoticthickening; more long-term studies will be necessaryto confirm these early results.

REFERENCES

1. Windus D. Permanent vascular access: a nephrologist’s view.Am J Kidney Dis 1993;21:457-71.

2. Feldman H, Held P, Hutchinson J, Stabber E, Hartigan M,Berlin J. Hemodialysis vascular access morbidity in theUnited States. Kidney Int 1993;43:1091-6.

3. Painter T. Myointimal hyperplasia: pathogenesis and implica-tions. 1. In vitro characteristics. Artif Organs 1991;15:42-55.

4. Sottiurai V, Yao J, Batson R, Sue S, Jones R, Nakamura Y.Distal anastomotic intimal hyperplasia: histopathologic char-acter and biogenesis. Ann Vasc Surg 1989;3:26-33.

5. Clowes A, Gown A, Hanson S, Reidy M. Mechanisms ofarterial graft failure. 1. Role of cellular proliferation in earlyhealing of PTFE prostheses. Am J Pathol 1985;118:43-54.

6. Leppaniemi A, Wherry D, Soltero R, Pikoalis E, Hufnagel H,Fishback N, et al. A quick and simple method to close vascu-lar, biliary, and urinary tract incisions using the new VascularClosure Staples: a preliminary report. Surg Endosc 1996;10:771-4.

7. Zhu YH, Kirsch WM. A new surgical technique for ven-ous reconstruction: the nonpenetrating clip. In: Modernvascular surgery. Vol 5. New York: Springer-Verlag; 1992.p. 425-63.

8. Pagnanelli D, Pait T, Rizzoli H, Kobrine A. Scanning electronmicrographic study of vascular lesions caused by microvascu-lar needles and suture. J Neurosurg 1980;53:32-6.

9. Longa E, Weinstein P, Chater G. Scanning electron micro-scopy studies of needle and suture damage in rat carotid andfemoral arteries. Microsurgery 1984;5:169-74.

10. Padubidri A, Browne E, Kononov A. Fibrin glue-assistedend-to-side anastomosis of rat femoral vessels: comparisonwith conventional suture methods. Ann Plast Surg 1996;37:41-7.

11. Leppaniemi A, Wherry D, Pikoulis E, Hufnagel H, WaasdorpC, Fishback N, et al. Arterial and venous repair with vascularclips: comparison with suture closure. J Vasc Surg 1997;26:24-8.

12. Pikoulis E, Burris D, Rhee P, Nishibe T, Leppaniemi AK,Wherry DC, et al. Rapid arterial anastomosis with titaniumclips. Am J Surg 1998;175:494-6.

13. Kirsch W, Zhu Y, Hardesty R, Chapolini R. A new methodfor microvascular anastomoses: report of experimental andclinical research. Am Surg 1992;58:722-7.

14. Kocher O, Madri J. Modulation of actin mRNAs in culturedvascular cells by matrix components and TGF-beta 1. InVitro Cell Dev Biol 1989;25:424-34.

15. Kirsch WM, Zhu YH, Boukouvalas Z, Cushman R, HardestyRA, Chrisler J. The anastomoses of small arteries and veins byclips: a description of instruments. In: Vascular Surgery.Beijing: International Academic Publishers; 1993. p. 41-6.

16. Mital D, Foster P, Jensik S, del Rio J, Sankary H, McChesneyL, et al. Renal transplantation without sutures using the vas-cular clipping system for renal artery and vein anastomosis—a new technique. Transplantation 1996;62:1171-3.

JOURNAL OF VASCULAR SURGERY332 Dal Ponte et al August 1999

17. Merino A, Cohen M, Badimon J, Fuster V, Badimon L.Synergistic action of severe wall injury and shear forces onthrombus formation in arterial stenosis: definition of athrombotic shear rate threshold. J Am Coll Cardiol 1994;24:1091-7.

18. Anderson J. Inflammatory response to implants. ASAIOTrans 1988;34:101-7.

19. Jansen J, Paquay G, van der Waerden J. Tissue reaction to soft-tissue anchored percutaneous implants in rabbits. J BiomedMaterial Res 1994;28:1047-54.

20. Diegelmann R, Cohen I, Kaplan A. The role of macrophagesin wound repair: a review. Plast Reconstr Surg 1981;68:107-13.

21. Heijmen R, Grundeman P, Borst C. Intima-adventitia appo-sition in end-to-side arterial anastomosis: an experimentalstudy in the pig. Ann Thorac Surg 1998;65:705-11.

Submitted Jul 27, 1998; accepted Oct 21, 1998.

JOURNAL OF VASCULAR SURGERYVolume 30, Number 2 Dal Ponte et al 333

Don’t miss a single issue of the journal! To ensure prompt service when you change your address,please photocopy and complete the form below.

Please send your change of address notification at least six weeks before your move to ensure continued ser-vice. We regret we cannot guarantee replacement of issues missed due to late notification.

JOURNAL TITLE:Fill in the title of the journal here.

OLD ADDRESS:Affix the address label from a recent issue of the journal here.

COPY AND MAIL THIS FORM TO: OR FAX TO: OR PHONE:Journal Subscription Services 314-432-1158 1-800-453-4351Mosby, Inc Outside the US, call11830 Westline Industrial Dr 314-453-4351St Louis, MO 63146-3318

NEW ADDRESS:Clearly print your new address here.

Name

Address

City/State/ZIP

Send us your new address at least six weeks aheadO N THE MOVE?