Vascular Access for Hemodialysis: Pathologic Features of Surgically Excised ePTFE Grafts Jean-Marc Delorme, MD, Robert Guidoin, PhD, Sheny Canizales, MD, Jamal Charara, PhD, Thien How, PhD, Yves Marois, MSc, Michel Batt, MD*, Pierre Hallade, MD*, Michael Ricci, MD ~, Carlo Picetti, MD**, Serge Contard, MD t, Quebec City, Canada
We analyzed 52 surgically excised ePTFE grafts used as secondary vascular access in chronic hemodialysis patients, structurally and histopathologically. Pseudoaneurysm formation at the site of repeated venipuncture was the main cause of surgical removal later than two years after implantation. Repeated needle punctures, twice per treatment, two or three times a week may result in a perigraft fibrous tissue capsule directly above areas where the graft was punctured. The delicate microporous structure of the graft wall was shown to be disrupted by needle punctures. The needle puncture sites were filled by surrounding connective tissue, and in one case, capillary formation was observed within the puncture sites. Examination by both light and scanning electron microscopy demonstrated identical patterns of pseudointir:la on the luminal surface. A thin pannus of endothelium-like cells, confineu to the vicinity of the anastomoses, was noted in only four cases. On other areas of the luminal surface without endothelium, a red coagulum incorporating blood cells and fibrin was observed. Histological evidence of acute infection was absent in 61% of the cases and only 27% were considered to be clinically infected. Careful needle puncture technique, systematic rotation of puncture sites, and the use of rigorous aseptic technique are essential in preserving the long-term structural integrity of the vascular access, despite the good mechanical properties and reasonable good resistance to infection of ePTFE grafts. (Ann Vasc Surg 1992;6: 517-524). KEY WORDS: disease.
Expanded PTFE prostheses; vascular access; hemodialysis; renal
For patients with end-stage renal disease who are not candidates for renal transplantation, hemodialysis remains the primary therapeutic modality. The arterioFrom the Department of Surgery, Laval University, and Biomaterials Institute, St-Franqois d'Assise Hospital, Quebec City, Quebec, Canada. *CHU, Nice, France, ~HoteI-Dieu, Mont St-Martin, France, §University Health Center, Burlington, Vermont, **Ospedale Santa Chiara, Trento, Italy. Reprint requests: Dr. R. Guidoin, Laboratoire de Chirurgie expOrimentale, Local 1701, Pavilion de Services, Universitd Laval, QuObec, QC G1K 7P4, Canada.
venous fistula constructed near the wrist, as described by Brescia and Cimino in 1969 , is the vascular access of choice and is associated with the best longterm patency of all available methods of angioaccess [2,3]. Winsett and Wolma  noted in their report of 390 autogenous arteriovenous (AV) fistulae that the two-year fistula patency rate, after exclusion of 27% of the cases of early failure, was 86% and the complication rate 9%. However, when anatomic prerequisites are lacking or the autogenous AV fistula has failed because of progressive complications, alternative access sites are required. For a number of patients, the secondary procedure involves the use of a biological or
ANNALS OF VASCULAR SURGERY
V A S C U L A R ACCESS FOR HEMODIAL YSIS
a synthetic graft implanted as a bridge between an artery and a suitable vein. The multitude of reports on the many types of available arterial substitutes attests to the universal failure of acceptance of any particular graft [5-7]. Over the past 20 years, different types of prostheses have been proposed for vascular access: autologous vein in 1969 , bovine carotid artery in 1972 , Dacron graft in 1972  and umbilical vein graft in 1976 . The expanded polytetrafluoroethylene (ePTFE) vascular prosthesis, first introduced as an AV fistula by Volker  in 1973, has been advocated as the preferred conduit for secondary vascular access procedures. This is due to the relative inertness of the material , minimal inflammation and little tissue or blood reactivity. However, it is generally agreed that the ePTFE is not the ideal conduit, and the complication rate related to the puncture technique during hemodialysis procedure remains high . We report here the morphologic and histopathological changes of 52 surgically-excised ePTFE grafts subjected to repeated external injuries due to needle puncture for hemodialysis two or three times a week.
METHODS During the last decade, 54 ePTFE specimens were provided by surgeons from 13 centers both in North America (USA and Canada) and Europe (France and Italy). This work forms part of our cooperative retrieval program for collecting and evaluating prosthetic arterial implants in humans. Graft harvesting
After excision at surgery, each graft was opened longitudinally, carefully rinsed with heparinized serum, fixed in a buffered solution of 1.5% glutaraldehyde and shipped with the relevant clinical data to the Biomaterials Institute at St-Franqois d'Assise Hospital, Quebec City, Quebec, Canada. Graft pathologic feature study
Graft processing." The grafts were photographed and representative areas of internal and external capsules were selected for pathologic investigation. Each area was divided into two subspecimens. The first one was post-fixed in a Perfix ® solution, dehydrated with ethanol and clarified with toluene. The second was immersed in glutaraldehyde (2.5% in a 0.2 M sodium cacodylate buffer) for 24 hours before the processing for scanning electron microscopy. The remaining graft specimens were cleaned, preferably by boiling in a 5% NaHCO3 solution to remove any adherent tissue, then rinsed with distilled water at room temperature for two hours.
Light microscopy: The first specimens were mounted in paraffin and sections, each 4 microns thick and were stained with Weigert and Masson trichrome for elastin and collagen, Dahl stain for mineralization and Gram stain for bacteria. Scanning electron microscopy (SEM): The specimens were post-fixed in thiocarbohydrazide (1% TCH) and osmium tetroxide (1%). Dehydration was achieved by immersion in graded solutions of ethanol. After critical point drying with liquid carbon dioxide used as the transfer medium, the specimens were coated with gold palladium and then observed in a JSM 35CF scanning electron microscope at 15 to 20 kV accelerating voltage. RESULTS Clinical data
Fifty-four ePTFE graft specimens were available, of which 52 met our minimum criteria for inclusion in the study, that is, documentation of cause of explantation and duration of implantation. Of these 52 ePTFE grafts, 42 (81%) were reinforced Gore-tex ® ePTFE vascular grafts, 8 (15%) were Impra ®vascular prostheses, and 2 (4%) were Vitagraft® vascular grafts. These 52 specimens were obtained from 48 patients (21 men, 25 women, and 2 sex not specified). Two patients were reoperated twice, while a third was reoperated three times. At the time of implantation, the patients' average age was 47.5 years (range 22 to 75 years). In all cases, the indication for the initial operation was secondary vascular access for maintenance chronic hemodialysis. The reasons for excision of the ePTFE grafts are given in Table I. In several cases there was more than one cause, but only the main cause is indicated in Table I. The leading causes for excision were pseudoaneurysm formation at the site of puncture (29% of the grafts), thrombosis (29%), and clinically suspected infection (27%). The cause of surgical removal versus the type of ePTFE is shown in Table II. The period of implantation was less than one month for two specimens (one each removed for infection and thrombosis); one to six months for nine specimens (five infections, two thromboses, TABLE I.mSurgical causes for explantation* Patients Percentage (n = 52) (%) Pseudoaneurysm formation Thrombosis Clinically suspected infection Stenosis Hemorrhage Arterial steal syndrome
15 15 14 5 1 2
29 29 27 9 2 4
*Some prostheses were explanted for more than one complication but only principal cause is indicated.
VOLUME6 No 6 - 1992
VASCULAR ACCESS FOR HEMODIAL YS1S
TABLE II.--Causes of surgical removal versus type of ePTFE graft Gore-tex ~ Impra ® Vitagraft ~¢ (n = 42) (n = 8) (n = 2) (n/%) (n/%) (n/%) Pseudoaneurysm formation Thrombosis Clinically suspected infection Stenosis Hemorrhage Arterial steal syndrome
1/12.5% 1/12.5% --
and two arterial steal syndrome); six to 12 months for nine specimens (four infections, two thromboses, one dilatation, one hemorrhage and one stenosis); one to two years for 12 specimens (six thromboses, three infections, two stenosis and one pseudoaneurysm); and greater than two years for 20 specimens (13 pseudoaneurysms, four thromboses, one infection and two stenosis). The mean duration of implantation was 25.3 months (range 0.4 to 95.2 months). The relationship between surgical cause of explantation and the period of implantation is shown in Table III. Macroscopic observations
The external capsules of the explanted prostheses were examined and classified in four different stages of encapsulation, depending upon the degree of fibrous coating: 0 = absence of encapsulation; 1 = thin fibrous coating; 2 = intermediate fibrous coating; and 3 = thick fibrous coating involving fatty tissue (Fig. 1). We noted that higher degrees of encapsulation were associated with increasing duration of implantation (Table IV). However, development of fibrous tissue was absent within the repeated needle puncture area (Fig. 2).
The microporous structure of the ePTFE grafts was often seen to be shredded by needle punctures. The degree to which connective tissue had penetrated the perforations in the microporous structure was observed and classified according to three stages: 0 -- absence of connective tissue penetration with no alteration of the microporous structure; 1 = penetration and disruption with deformation of the wall structure; and 2 = obliteration of the structure (Table V). We noted that in 54% of the cases, the connective tissue had infiltrated the microporous structure of the graft while in 27%, the wall structure of the graft was obliterated. Connective tissue filled the needle puncture holes and it appeared to blend with the surrounding fibrous tissue of the external capsule (Fig. 3). In addition, it is clear that both the degree of connective tissue penetration and damage to the wall structure, particularly near the site of repeated venipuncture, increased with longer periods of implantation and cannulation. In addition, we observed that the outer PTFE reinforcing film of the Gore-tex ® grafts was detached from the graft wall in 64% of the cases. This was probably due to the repeated needle punctures. A chronic inflammatory reaction consisting of mononucleate cells and foreign body giant cells was observed within the external capsule. In one case, an intense inflammatory reaction was seen to be located against the detached outer reinforcing film of the graft (Fig. 4). The luminal surface of the prostheses demonstrated little healing. Collagen was observed over 26% of the luminal surface. Neocapillary formation was noted in only five of the 52 grafts studied. In four cases this was observed on the outer surface and in one case it was confined to the needle puncture hole (Fig. 5). Scanning electron microscopy (SEM) demonstrated an identical pattern of healing in all three graft types. Endothelium-like cells were present on the luminal surface of four
TABLE III.mSurgical cause of explantation versus period of implantation 1 month N u m b e r of e x p l a n t e d grafts Surgical c a u s e Pseudoaneurysm formation Thrombosis Infection Stenosis Hemorrhage Arterial steal syndrome
Period of implantation 1-6 6-12 1-2 months months years
Mean duration of implantation (months)
1 1 o ---
2 5 _ -2
2 4 1 1 --
6 3 2
4 t 2 ---
22.8 10.7 23.4 9.2 2.7
VASCULAR ACCESS FOR HEMODIAL YSIS
ANNALS OF VASCULAR SURGERY
Fi~. 1. Explanted ePTFE grafts showing various stages of encapsulation. (A) Absence of encapsulation in reinforced Gore-tex ~ graft surgically removed because of clinical signs of infection eight months after implantation. (B) High degree of encapsulation with fatty deposits in Vitagraft* removed 31 months after implantation because of graft thrombosis.
grafts (Fig. 6). In almost all the other cases, the flow surface consisted of a red cell fibrin coagulum, Histological evidence of infection seen by light microscopic examination (i.e. the presence of polynuclear leukocytes, macrophages and lymphoplas-
mocytes) was observed in 12% of the grafts. Heavy infection with microabcesses was not visible in any of the grafts studied. Bacteria (Gram+) were observed in five of the 52 grafts. Only three of these five grafts had been surgically excised because of
TABLE IV.--Effect of duration of implantation on degree of encapsulation Number of prostheses (n =52) 0 1 2 3
= = = =
absence of encapsulation thin fibrous coating intermediate fibrous coating thick fibrous coating involving fatty tissue
16 14 9
Percentage (%) 13 31 27 17
Mean duration of implantation (months) 25 15.7 21.2 32.5 49.4
Fig. 2. Photographs showing segments of reinforced Gore-tex ~ graft implanted as secondary access procedure in 65-year-old patient. Graft was surgically removed after 45,3 months because of pseudoaneurysm formation at site of repeated puncture. (A) View of excised specimen. (B) View of cleaned specimen showing destruction of graft wall, Integrity of conduit is maintained only by fibrous tissue.
VOLUME6 No 6 - 1992
V A S C U L A R A C C E S S FOR HEMOD1AL YSIS
TABLE V.--Connective tissue penetration into the graft wall Degree of connective tissue penetration
Number of grafts
Mean duration of implantation (months)
0 = absence of penetration 1 = penetration of wall structure by connective tissue 2 = obliteration of the structure
clinically suspected infection. The other two grafts which were excised for pseudoaneurysm formation and arterial steal syndrome were found to harbor bacteria, although no infection was noted clinically. Moreover, the infection was found to be located preferentially on the external part of the graft. SEM examination of the luminal surface revealed the presence of leukocytes on three of the 14 grafts with clinical signs of infection
(21.4%). Bacteremic colonization was seen in two of the cases (14.3%) (Fig. 7). In addition, leukocytes and bacteremic colonization were found in 8.3% and 16.6% of the 38 noninfected grafts, respectively, while three grafts had both (8.3%). Both light and scanning electron microscopic evidence of infection were present in one graft, which was surgically removed because of clinical signs of infection (Table VI).
Fig. 3. Section of thrombosed Gore-tex ~ graft explanted after 39.3 months in 35-year-old patient. Connective tissue fills needle puncture hole, disrupting graft wall (FT). Note needle damage to luminal surface (arrow) (Masson X100).
Fig. 4. Light microscopy photograph of reinforced Gore-tex ~ graft implanted for 40 months and surgically removed because of thrombosis: neocapiliary formation in needle puncture hole (Masson X400).
VASCULAR ACCESS FOR HEMODIAL YSIS
ANNALS OF VASCULAR SURGERY
Fig. 5. Section of reinforced Gore-rex ~":graft removed because of infection after 44.7 months of imptantation in 44-year-old patient. Area of chronic inflammatory reaction is located under ePTFE reinforcing film, denoted by arrow (Masson X100).
DISCUSSION The use of subcutaneous arteriovenous fistula for primary angioaccess in patients with chronic renal failure is widely accepted. Continued hemodialysis has been made possible through the development of second stage procedures employing many types of arterial substitutes, including ePTFE grafts. Since ePTFE grafts were first used experimentally as a venous substitute by Soyer in 1972  and as arterial bypass by Matsumoto in 1973 , they have become the most popular of the synthetic alternatives to autologous veins because of desirable characteristics such as ease of handling, availability in a wide range of sizes and configurations and low thrombogenicity [16-18]. Although excellent results have been obtained with this graft,
Fig. 6. Scanning electron microphotograph of luminal surface of reinforced Gore-tex ~ graft implanted for 36 months. Endothelial-like cells are seen on luminal surface (original magnification X1000).
puncture site complications remain as a result of the repetitive cannulation of the graft. Pseudoaneurysm formation at the site of venipuncture is of particular importance, as it limits the area of the graft available for cannulation. The reported incidence of this complication with ePTFE grafts is variable [4,17,19]. Complication rates related to the techniques of graft puncture have been reported previously . We identified the importance of the diameter of the needle, and the needle puncture technique in the degree of damage to the PTFE graft wall structure. Repeated puncture of the graft may cause fragmentation of the microporous structure, and we note that each puncture in the graft leaves a small hole due to material removal. In those areas, the ePTFE graft material was
Fig. 7. Luminal surface of reinforced Gore-tex ~ graft implanted for 1.5 months prior to removal because of clinical signs of infection. (A) Note presence of bacteremic colonization (BC) and ieukocytes (WBC) (original magnification X4000).
V A S C U L A R A C C E S S FOR HEMODIAL YS1S
N o 6 - 1992
TABLE VI.--Frequency of evidence of infection versus duration of implantation
Infected grafts Mean duration of implantation (months)
Noninfected grafts Mean duration of implantation (months)
Mean duration of implantation (months)
Light microscopy Gram strain +
SEM Leukocytes Bacteria Both
3 2 1
12.3 11.2 8.9
3 6 3
16.1 14.5 6.8
6 8 4
14.2 12.8 7.8
SEM - scanning electron microscopy In one case, both light and scanning electron microscopic evidence of infection were present in an ePTFE graft exhibiting clinical signs of infection.
seen to be discontinuous, and the spaces between fragments were filled with fibrous tissue. We believe that the lack of surrounding connective tissue, in addition to the material substitution in the wall, may result in development of a pseudoaneurysm directly above areas where the graft is repeatedly punctured. The formation of pseudoaneurysm is not due to degeneration of graft material such as that reported with bioprostheses. Indeed, ePTFE graft is recognized to have good mechanical properties and stability [211. Thrombosis is a common cause of graft failure reported with all types of arteriovenous grafts and especially ePTFE grafts. In the ePTFE grafts, the incidence of outflow obstruction requiring operative revision ranges from 10% to 34% [4,22-24]. Immediate postoperative failures are due to technical factors, or thrombosis caused by mishandling of the graft by dialysis personnel, hemostatic banding, or unintentional obstruction by a sleeping patient. When these are excluded, the cases of thrombosis which occur are usually the consequence of progressive thickening and fibrous hyperplasia of the graft vein anastomosis . In addition, accumulation of organized laminar clot at frequently used sites of puncture may occur, leading to diffuse regions of stenosis. This is due to regenerative fibrosis resulting in uneven flow surface and narrowing of the lumen. The turbulence and blood flow disturbances which occur at these sites may also lead to late thrombosis. Another common cause of the prosthesis failure is infection. Expanded PTFE grafts, as any other foreign surfaces, are known to have limited resistance of bacteremic colonization [25,26]. The risk of infection is increased by contamination both during the initial surgical procedure and during repeated puncture of the graft for hemodialysis. Indeed, its incidence correlates, in general, with the completeness of the tissue ingrowth following graft insertion [27,28]. However, although infection remains an important problem, emphasizing the need for rigorous use of aseptic techniques, it does not always
result in the loss of the graft. In our study, 28% of the grafts were removed because of infection, but we noted that the criteria of infection were absent in 64% of the cases. As the incidence of bacteria shown with the Gram stain was lower than anticipated (21.4% of the grafts explanted because of clinical infection), we can deduce that the infection reaction was not uniformly distributed over the entire length of the graft. Since we found infection to be located preferentially on the external part of the graft, we may assume that the propagation of the infection is slow because of the need for the destruction of the capsules. However, it is possible that areas free of infection might have been selected for investigations. Some of the grafts were probably infected by direct contamination at the time of insertion, but the more likely route was during percutaneous cannulation. Indeed, if the infection spread from the luminal surface after exposure to hematogenous seeding, its coverage would be more homogenous on the flow surface of the graft. In addition, it may be possible that iatrogenic folds on the luminal surface induced by needle injuries may offer a preferred site for bacteremic colonization not seen by microscopic observation.
CONCLUSION The ideal material for secondary angioaccess procedure should be nonthrombogenic, conformable and easy to handle. It should also allow ingrowth of perigraft tissue and be readily available. Expanded PTFE graft is known to have many advantages as a material for secondary angioaccess in patients on maintenance hemodialysis. However, disruption of the graft wall caused by repeated needle puncture can lead to loss of graft integrity, resulting in the development of pseudoaneurysms at the site of puncture. In addition, we believe that, since the prosthesis lies superficially, an inflammatory reaction caused by repeated cannulation may be misidentified as infection, with resultant graft
VASCULAR ACCESS FOR HEMODIAL YS1S
loss, even though the graft may actually not be infected. Although it is still not the ideal alternative material, we believe that the ePTFE graft remains the best secondary vascular access for the hemodialysis patient because of its good mechanical properties and stability and reasonably good resistance to infection.
ACKNOWLEDGMENTS This work has been supported by the Medical Research Council of Canada. The following hospitals collaborated in this study: in USA, University Health Center, Burlington, Vermont (M. Ricci); in Canada, Toronto General Hospital, Toronto, Ontario (K.W. Johnston), Royal Victoria Hospital, Montreal, Quebec (J.L. Meakins), Health Sciences Center, Winnipeg, Manitoba (A. Downs, R. Guzman), St-Franqois d'Assise Hospital, Quebec City, Quebec (C. Rouleau, C. Gosselin), H6tel-Dieu Hospital, Quebec City, Quebec (P. Roy, G. Laroche, R. Charrois); in Italy, Ospedale S. Chiara, Trento (C. Picetti); and in France, Clinique Bizet (P. Bonnaud), H6tel-Dieu St-Martin (P. Hallade, S. Contard) and CHU Nice (G. Avril, M. Batt, P. LeBas). The authors would like to extend their gratitude to Michel Marois and Paul-Emile Roy for help and guidance. The technical assistance of Suzanne Bourassa, Marieile Corriveau, Richard Couture and Karen Horth is gratefully acknowledged. We are also indebted to the attending staff of the operating rooms and the pathology laboratories who participated in harvesting the grafts. Finally, the collaboration of Barbara Boyce, Don Lass and Gary Warns of W.L. Gore a n d A s s o c i a t e s is a l s o g r e a t l y a p p r e c i a t e d .
REFERENCES 1. BRESCIA M J, CIMINO JE, APPEL K, et al. Chronic hemodialysis using venipuncture and a surgically created arteriovenous fistula. N Eng! J M e d 1966;275:1089-1092. 2. PALDER M J, KIRKMAN RL, WHITTEMORE AD, et al. Vascular access for hemodialysis: patency rates and results of revision. Ann Surg 1985;202:235-239. 3. KERLALIAN GM, ROEDERSHEIMER LR, ARBAUGH JJ, et al. Comparison of autogenous fistula versus expanded polytetrafluoroethylene graft fistula for angioaccess in hemodialysis. A m J Surg 1986;152:238-243. 4. WlNSETT OE, WOLMA FJ. Complications of vascular access for hemodialysis. South Med J 1985;78:513-517. 5. GIRARDET RE, HACKETT RE, GOODWlN M J, et al. Thirteen months experience with the saphenous vein graft arteriovenous fistula for maintenance hemodialysis. Trans A m Soc Artif lntern Organs 1970;16:285-291. 6. GUIDOIN R, ROY PE, BONNAUD P, et al. Stripped veins for secondary vascular access in hemodialysis. Pathologic study of the grafts after surgical excision. J Mal Vasc 1985;10:331-342. 7. BREMS J, CASTANEDA M, GARVIN PJ. A five-year mum
ANNALS OF VASCULAR SURGERY
experience with the bovine heterograft for vascular access. Arch Surg 1986;121:941-944. 8. MAY J, TILLER D, JOHNSON J, et al. Saphenous vein arteriovenous fistula in regular dialysis treatment. N Engl J Med 1969;280:770-771. 9. CHIMITZ JL, YOKOYAMA T, BOWER R. Self-healing prosthesis for arteriovenous fistula in man. Trans Am Soc Artif lntern Organs 1972;18:452-457. 10. DUNN I, FRUMKIN E, FORTE R. Dacron velour vascular prosthesis for hemodialysis. Proc Clin Dial Transplant Forum 1972;2:85. 11. DARDIK H, IBRAHIM 1M, DARDIK I. Arteriovenous fistula constructed with modified human umbilical cord vein graft. Arch Surg 1976;111:60-62. 12. VOLDER JG, KIRKHAM RL, KOLFF WJ. A-V shunts created in new ways. Trans Am Soc Artif Intern Organs 1973;19:38---42. 13. KIDSON IG. Arterial prostheses. Br J Hosp Med 1983;30: 248-254. 14. SOYER T, LEMPINEN M, COOPER B, et al. A new venous prosthesis. Surgery 1972;72:864-872. 15. MATSUMOTO H, HASEGAWA T, FUSE K, et al. A new vascular prosthesis for a small caliber artery. Surgery 1973; 74:519-523. 16. ANDERSON CB, ETHERIDGE GE, SICARD CA. One hundred polytetrafluoroethylene vascular access grafts. Dial Transpl 1980;47:145-151. 17. TELLIS VA, KOHLBERG WI, BHAT DJ, et al. Expanded polytetrafluoroethylene graft fistula for chronic hemodialysis. Ann Surg 1979;189:101-105. 18. SECCIA M, BUCCIANTI P, CHIARUGI M, et al. The PTFE graft as alternative blood access in the surgical management of long-term haemodialysis. Life Sup Syst 1984 ;2:63-70. 19. SABANAYAGAM P, SCHWARTZ AB, SORICELLI RR, et al. A comparative study of 402 bovine heterografts and 225 reinforced expanded PTFE grafts as AVF in the ESRD patient. Trans Am Soc Artif lntern Org 1980;26:88-91. 20. CHARARA J, GUIDOIN R, GILL F, et al. Morphologic assessment of ePTFE graft wall damage following hemodialysis needle punctures. J Appl Biomat 1990;1:27%287. 21. HANEL KC, MC CABE C, ABBOTT WM, et al. Current PTFE grafts: a biomechanical, scanning electron and light microscopic evaluation. Ann Surg 1982;195:456-463. 22. JENKINS AML, BUIST TAS, CLOVER SD. Medium-term follow up of forty autogenous vein and forty polytetrafluoroethylene (Gore-tex) grafts for vascular access. Surgery 1980 ;88:667-672. 23. MUNDA R, FIRST MR, ALEXANDER JW, et al. Polytetrafluoroethylene graft survival in hemodialysis. JAMA 1983; 249:21%222. 24. BUTLER HG, BAKER LD, JOHNSON JM. Vascular access in chronic hemodialysis: polytetrafluoroethylene (PTFE) versus bovine heterograft. Am J Surg 1977;136:791-793. 25. SHAH PM, ITO K, CLAUSS RH, et al. Expanded microporous polytetrafluoroethylene (PTFE) grafts in contaminated wounds: experimental and clinical study. J Trauma 1983;23:1030-1033. 26. GOEAU-BRISSONNIERE O, LEPORT C, GUIDOIN R, et al. Experimental colonization of an expanded polytetrafluoroethylene vascular graft with Staphylococcus aureus: a quantitative and morphologic study. J Vasc Surg 1987;5: 743-748. 27. MORGAN AP, DAMIN G J, LAZARUS JM. Failure modes in secondary vascular access for hemodialysis. ASAIO J 1978 ;1:44--48. 28. BHAT DJ, TELLIS VA, KOHLBERG WI, et al. Management of sepsis involving expanded polytetrafluoroethylene grafts for hemodialysis access. Surgery 1980;87:445-450.