Dacron, and to a lesser extent Teflon, has become the major synthetic grafting material. Unlike nylon, Ivalon, and Vinyon-N which lose their tensile strength after implantation, Dacron and Teflon remain essentially unchanged even after long periods. Teflon is generally less reactive than Dacron. Poor reactivity may be a more or a less desirable property since it is associated with less tissue reaction, less incorporation of tissue into the graft, and less fibroblastic invasion.

Grafts may be fabricated as either a woven or knitted yarn. They can also be manufactured with a velour surface. Braided yarns have fallen from use as they tend to be quite bulky and difficult to handle; fabric prostheses are now made from multifilament yarn. In woven grafts, the threads are interlaced in a simple over-and-under pattern whereas a knitted fabric is made with threads that are looped to form a continuous interlocking chain. In general, the woven fabrics are less porous and stiffer than the knitted fabrics.

Velour is a variant of the Dacron fabric prosthesis developed in the later 1960s. The velour surface has loops of yarn extending upward at right angles to the fabric surface, giving the surface a velvety, plush texture. The velour finish can be made on the external, internal, or both surfaces of the graft. The velour configuration may possibly be important in providing a ‘trellis’ for the graft healing process.

An alternative to the fabric protheses is expanded Teflon or polytetrafluoroethylene which unlike the Dacron prostheses, is not a textile graft. Expanded polytetrafluoroethylene is a fluorocarbon polymer manufactured by forcibly expanding Teflon through a heating and stretching process. The result is a material consisting of solid nodes with interconnecting small fibrils. The length of the fibril determines the pore size which has been standardized to approximately 30 &mgr;m. Fluorine atoms in Teflon impart a highly electronegative surface charge to the graft, making it hydrophobic: this property is claimed to play a role in resistance to thrombosis. Early clinical experience with the expanded polytetrafluoroethylene graft was disappointing: aneurysm dilation developed in a number of grafts. The graft has since been reinforced by either a thin skin of polytetrafluoroethylene, thickened walls, or application of an external support coil.

An alternative to the fabric or polymeric vascular graft is the use of biological materials as a conduit. Bovine heterografts were the first biological graft developed for human implantation. These grafts stemmed from the work of Rosenberg who in 1956 described a process using the enzyme ficin that resulted in the production of pure collagen tube which was then tanned. The tanning process stabilizes the collagen tube by chemically cross-linking the collagen fibrils with dialdehyde starch. The chemically modified heterografts were prone to aneurysmal degeneration and, like the homografts, they were abandoned. The bovine heterograft was, however, a prototype for another tanned prosthesis: the human umbilical cord vein allograft stabilized with glutaraldehyde.

Early clinical experience with the human umbilical cord vein gave disappointing results. The grafts underwent degenerative changes early, with subsequent thrombosis and aneurysmal dilation. Although a glutaraldehyde stabilized and external Dacron mesh reinforced human umbilical vein was used for a while, degradation occurs within about 5 years.

No single graft is suitable for every clinical situation. The choice of arterial substitute depends upon the diameter of the native vessel being replaced or bypassed, the anatomical location, whether a joint is being crossed, the presence of infection, the length of graft needed, and whether or not an autogenous substitute is an option.

The size of the vessel to be bypassed is obviously important in selecting the best graft. Since no autograft is of sufficient size to replace the aorta, prosthetic grafts are always necessary. Dacron grafts in either a woven or knitted configuration are the primary substitute, although polytetrafluoroethylene grafts are also available for aortic grafting. Knitted grafts are more pliable, softer, and easier to sew then are woven grafts. They are, however are more porous than woven grafts and must be rigorously preclotted before use to seal the interstices of the graft. During most operations employing porous, knitted grafts, 50 to 60 ml of blood is withdrawn from a convenient vessel prior to administration of systemic heparin and used to flush the prosthesis repeatedly until the interstices fail to pass blood. Knitted grafts impregnated with collagen do not need preclotting and are now available.

The choice of grafts for aortic replacement is arbitrary; all have good and fairly equivalent long-term patency. Special characteristics of each, however, makes one variety more favourable under certain circumstances. At the Massachusetts General Hospital and at Oxford Dacron grafts are the primary choice for aortic reconstruction, although expanded polytetrafluoroethylene is occasionally used, but usually in the smallest diameters. A knitted Dacron graft is ideal for elective aortic surgery because of its handling characteristics. However, with recent improvements in weaving technology, more woven grafts have been used in these institutions. In other circumstances, especially when blood loss is critical, such as with thoracic aortic surgery and ruptured aortic aneurysms, a woven Dacron graft is clearly the better choice.

Results have been good when using Dacron for arterial replacement of the aortofemoral system because of their high flow rate. Five-year patency rates approach 90 per cent for aortobifemoral bypass. By 10 years the patency rates fall to approximately 70 per cent. In a long-term study of nearly 1000 patients who underwent bypass for aortoiliac occlusive disease, Crawford et al. reported 10, 15, and 20-year patency rates of 79 per cent, 70 per cent, and 56 per cent respectively.

So-called extra-anatomic bypass grafting for aortoiliac disease gained popularity for the treatment of high risk patients who were considered to be unsuitable candidates for a major intra-abdominal operation. The two most common types of extra-anatomic bypass grafting are the axillofemoral and the femorofemoral bypasses. Both Dacron and expanded polytetrafluoroethylene have been used with similar success and patency rates. Patency rates after axillofemoral bypass range from 40 to 90 per cent at 5 years; the major determinant of patency is the status of the peripheral run-off. Femorofemoral bypass has patency rates comparable to those of aortofemoral grafting: 79, 68, and 60 per cent at 5, 10, and 15 years. At the Massachusetts General Hospital the femorofemoral bypass graft is used not only for the high risk patient, but also for the management of unilateral iliac occlusion in young patients, especially men, since avoiding para-aortic dissection eliminates the risk of sexual dysfunction. On the other hand, axillofemoral grafting is rarely used except in patients with aortic graft infection.

The choice of grafting material for infrainguinal repairs is less straightforward than it is with aortic surgery. The material and the construction of the graft play a major role in determining the long-term patency. The anatomic location of the distal anastomosis and the status of the distal run-off are very important elements to consider when deciding which graft to use.

Undoubtedly the best patency and limb salvage rates are found when autogenous saphenous vein is used. In at least 20 per cent of patients, however, the saphenous vein is of inadequate length or calibre, or has been surgically removed for coronary artery bypass grafting or management of varicose veins. In general, veins less than 3 mm in diameter are inadequate for use as bypass conduits, and an alternative must be chosen. Other autogenous conduits such as the short saphenous veins, contralateral greater saphenous vein, or arm veins can be used. If these are also inadequate or unaccessible, a prosthetic graft must be used.

Currently expanded polytetrafluoroethylene and human umbilical vein grafts are the two prostheses more frequently used. In certain cases a composite or segmental graft can be used to allow more distal grafting when the vein is of insufficient length: such grafts make use of available vein as the distal segment of the bypass. The composite graft is joined to a more proximal segment of prosthetic graft, thereby producing a suitable length of conduit for distal revascularization. We do not often use a true composite graft for femoropopliteal grafting because currently available prothetic grafts perform as well or better. When mixed graft materials are used, it is usually as a sequential graft with the distal anastomosis to two or more segments.

Textile prostheses such as Dacron have markedly inferior patency rates in bypass to the popliteal artery compared to saphenous vein; 5-year patency results for bypass in the femoropopliteal position for saphenous vein and Dacron are 72 and 20 per cent respectively. Because of consistently poor results, Dacron grafts are no longer used for this type of procedure.

Without doubt expanded polytetrafluoroethylene has become the most commonly used synthetic conduit for infrainguinal reconstruction. Many papers have been published in recent years describing use of this material for distal reconstruction, but clearly defined indications for its use are lacking.

In a recent 6-year, prospective, multicentre, randomized comparison of autogenous saphenous vein and expanded polytetrafluoroethylene, Veith et al. demonstrated autologous saphenous vein to be superior for femoropopliteal bypasses, particularly when they cross the knee. By 4 years in the above-knee position, the differences between vein and expanded polytetrafluoroethylene were statistically significant, with saphenous vein showing a patency rate of 68 per cent, compared with 38 per cent for synthetic grafts. In the below-knee position, saphenous vein was also superior, with patency rates of 76 per cent and 54 per cent respectively. In the tibioperoneal location, saphenous vein grafts showed patency rate of 49 per cent, compared to 12 per cent for expanded polytetrafluoroethylene. The study failed to support the routine preferential use of expanded polytetrafluoroethylene grafts for either femoropopliteal or more distal bypasses. Flinn et al. reported an improved patency rate of 37 per cent when expanded polytetrafluoroethylene grafts were placed into the tibioperoneal location with the administration of postoperative warfarin.

Despite the many studies reporting significantly better patency and limb salvage rates following femoropopliteal bypass with saphenous vein some authors prefer expanded polytetrafluoroethylene, especially when placed above the knee. We use such synthetic grafts instead of vein only under specific circumstances and only in the above-knee position. Expanded polytetrafluoroethylene is rarely taken to the below-knee popliteal artery and virtually never to the tibial vessels. We favour an above-knee expanded polytetrafluoroethylene when operative time is important, in unstable or high-risk patients or in those requiring a femoropopliteal bypass in addition to a more extensive proximal inflow procedure such as an aortobifemoral bypass. Angiographic evidence of good run-off must be present to satisfy predictions of patency of the graft. Less than two-vessel run-off and the presence of diabetes reduces the 2-year patency of polytetrafluoroethylene grafts from 70 to 30 per cent.

Using expanded polytetrafluoroethylene to save the saphenous vein in case future coronary artery bypass grafting is necessary is unwarranted. Natural history and retrospective studies have shown that only a small percentage of patients who present with symptoms of peripheral vascular disease actually ever require coronary artery bypass grafting at a later time.

The difference in the patency rates of synthetic material and autogenous saphenous vein anastomosed to below-knee vessels may be partly due to the length of graft needed to reach the below-knee arteries, the technical difficulties in anastomosing the graft to a small artery, and to the difference in the elastic properties of the prosthesis and the host artery. Interposing a cuff of vein between an expanded polytetrafluoroethylene prosthesis and the artery produces patency rates of 91 per cent and 72 per cent for femoropopliteal and femorotibial grafts, respectively. This cuff of vein may decrease the compliance mismatch between host artery and prosthetic graft, but it is still not apparent whether it will improve long-term patency or reduce the development of neointimal hyperplasia. Decreased juxta-anastomotic neo-intimal hyperplasia is present in vein cuffed grafts in dogs. We have had only limited experience with the technique, since we so rarely use prostheses in these patients.

The glutaraldehyde-treated human umbilical vein is the most commonly used alternative to polytetrafluoroethylene for infrainguinal grafting. A comparison of these two materials found no differences for above-knee anastomoses. However, patency rates and limb salvage were better for human umbilical vein in below-knee anastomoses and in poor run-off situations. Other studies have also found that below-knee human umbilical vein grafts tend to have slightly better patency than expanded polytetrafluoroethylene, but human umbilical vein has the disadvantage of being technically difficult to handle. One study reported a 42 per cent dilation rate and a 57 per cent rate of aneurysm formation in patients with human umbilical vein grafts implanted for more than 2 years. Because of the appreciable incidence of thrombosis and aneurysmal degeneration we have not implanted an umbilical graft for several years.

Graft failure is a major problem facing the vascular surgeon: patients are often more symptomatic after graft failure than they were before the bypass procedure. In a retrospective review of the outcome of failed femoropopliteal grafts, two-thirds of patients with failed grafts initially undertaken for claudication were no worse following graft occlusion. However, 24 per cent showed worsened ischaemia when the graft failed. A considerable number of patients were thereby converted to a limb-threatened status. Since all prosthetic grafts available are prone to complications, the vascular surgeon must understand the aetiologies of graft failure to enable him to reduce the risk of failure to a minimum.

The causes of graft failure can best be divided into four time periods (Table 1) 177. Acute graft failures (within 48 h) are usually secondary to technical errors in the creation of the anastomosis, a retained unlysed valve cusp, or placement of the graft at a site of poor inflow or outflow. Intraoperative angiography or angioscopy can provide detailed assessment of the anastomosis, and if an error such as an intimal flap or stenosis at the toe of the graft is recognized, the anastomosis can be repaired or revised. Reinspection of the preoperative angiogram must include assessment of the quality of the inflow vessels as well as the outflow. Areas of proximal stenosis revealed by angiography should be evaluated for haemodynamic significance by measuring intraluminal pressure gradients across the narrowed lumen. Significant blocks should be corrected before a more distal bypass procedure is performed. The blocks required proximal bypass or a balloon angioplasty.

Graft failure occurring between 2 days and 12 weeks after surgery is usually secondary to a heightened graft thromboreactivity. Thromboreactivity exists with all grafts or reconstructions but varies in intensity and duration and is governed by host factors (coagulability and blood flow), as well as by inherent properties of the graft (surface thrombogenicity and compliance). There may be a ‘thrombotic threshold velocity’ required to maintain graft patency and which reflects thromboreactivity. Velocities below a given level for any graft material will lead to thrombosis and closure of the graft.

Graft failure up to 18 months after operation is usually due to intimal hyperplastic lesions. Anastomotic intimal hyperplasia is commonly greater at the downstream or outflow anastomosis, and also occurs within the body of vein grafts. Its cause is unknown, but it is a very complex problem. Two major theories involve the response to injury hypothesis and the response of the vessel wall to new conditions of physical stress.

Graft failure after 18 months postoperatively is most often due to progression of atherosclerosis at or beyond the distal anastomosis. As discussed earlier the concept of distal progression of disease as a cause of late graft failure has been used by some as an argument in favour of using prosthetic material for femoropopliteal bypass, saving the saphenous vein for reconstruction of a more distal outflow vessel in those patient who develop further problems.

Structural failures are rare in modern day fabric prostheses. Some of the earlier grafts were flawed by friability, inability to hold sutures, and graft dilation; the knitted variety were more likely to fail. Complications associated with dilation include bleeding through graft interstices, fibre breakdown with resultant holes and tears, deposition of mural thrombus with possible graft occlusion or distal embolism, and development of anastomotic aneurysms. Almost all fabric prostheses dilate after implantation. An obligatory 15 to 20 per cent dilation occurs with knitted grafts, and this should not be viewed as a complication. In a small percentage of grafts, true aneurysmal dilation occurs. The risk of graft failure associated with dilation is unknown. In a study of 32 patients with knitted Dacron grafts to an average duration of 175 days no part of a graft dilated more than 94 per cent. Although woven grafts with interlocking yarns have little or no inherent stretch, knitted fabrics have much more stretch because of their looped structure. The loops straighten in the line of greatest stress, leading to dilation. Life-time followup of patients with these grafts is required. Advanced degeneration warrants replacement of the graft.

An anastomotic false aneurysm results from a partial or complete separation of the prosthetic graft from the host artery. Although false aneurysms can occur at any site, they are most frequent at the common femoral artery, with a frequency rate of approximately 3 per cent. The aetiology of anastomotic false aneurysms is multifactorial. In the distant past, fragmentation of silk sutures was associated with many false aneurysms; now that non-absorbable, synthetic suture is used, suture failure is rarely a cause. In the majority of cases the sutures are intact, remaining attached to the graft with the tear occurring along the host artery. The cause of the tear may be due to atherosclerotic degenerative changes in the host artery wall. Endarterectomy at the site of the anastomosis weakens the anastomotic site and could be a predisposing factor leading to the formation of false aneurysms. Other contributing factors include a compliance mismatch between graft and host artery, infection, improper suturing technique, and tension on the suture line. Complications of false aneurysms include rupture, thrombosis, and embolism. In general, an anastomotic false aneurysm should be surgically repaired when diagnosed. In the elderly or high risk patient small false aneurysms (<2 cm) can be left untreated but require close monitoring. Any sign of expansion mandates repair.

Graft infection is one of the most devastating complications of arterial reconstruction. Although infection does not necessarily lead to graft thrombosis, it is a failure of the bypass procedure. The incidence ranges from 1 to 2 per cent for series reporting aortic prosthesis and up to 6 per cent for femoropopliteal grafts. Major morbidity and mortality is associated with the development of graft infection, and the entire infected graft must usually be removed to control the problem, followed by revascularization by an extra-anatomical route such as the obturator foramen.

The perfect graft material does not exist, although a number of characteristics have been suggested (Table 2) 178. The graft must be available in a variety of sizes and lengths to accommodate aortic or distal tibial vessel bypasses. It should be easy to handle and suture. The graft must be non-fraying, pliable, and be able to conform to the artery to which it is being anastomosed. It should be durable, strong, and last the life expectancy of the patient without degeneration. The graft should be biocompatible, allowing healing with a non-thrombogenic surface; a graft that becomes completely lined with endothelium would be optimal. The prosthesis should match the viscoelastic properties of the host artery to which it is to be anastomosed. It should have a low associated infection rate, and a high, long-term patency rate. Unfortunately, no one arterial substitute, autogenous or prosthetic has all of these properties. An ideal graft however would have a perfect patency rate with no thrombotic occlusions, would not have pathological neo-intima, and would not develop infection, dilate, become aneurysmal, or otherwise breakdown and rupture.

When replacing or bypassing an arterial segment with a vascular graft, one hopes that the graft will behave like the normal artery, providing non-turbulent, pulsatile flow. Unfortunately this is not always the case. A number of properties of the graft must be controlled in order to provide the most durable conduit, such as the diameter and the length in order to provide adequate flow to the distal arterial tree. Flow through a graft is best considered by the Poiseuille equation: Equation 22

where Q is the flow (cm³ s), P&sub1;− P&sub2; represents the pressure gradient (dynes/cm²) between two points separated by the distance L, r (cm) is the radius of the tube, and n the coefficient of viscosity of the fluid blood in poise (dyne s cm²). Thus, graft diameter and length are the major determinants of resistance to steady flow.

Impedance is the equivalent to resistance in oscillating systems. Blood flowing through blood vessels is of course moved by pulsatile flow rather than by steady flow. As intraluminal pressure increases with systole, the arterial wall stretches and then returns to baseline during diastole. Thus, the distensibility or compliance of the vessel plays an additional important role in the impedance to pulsatile flow. Compliance is defined as the fractional change in diameter per unit change in pressure: Equation 23

The less compliant a vessel, the greater will be its impedance. Experimentally, compliant grafts are more likely to remain patent than stiff, non-compliant grafts. The consequences of a mismatch in viscoelastic properties (i.e. compliance) may therefore explain the inferior performance and decreased patency of small and medium sized grafts when a rigid graft is anastomosed to a compliant host vessel.

It was hoped that prosthetic grafts would develop an endothelial lined surface with flow characteristics of the host artery. Initial observations in animals gave very encouraging results since implanted fabric prostheses ‘healed’ completely, with the production of a viable endothelial cell lined neointima. Prosthetic grafts in human beings do not develop such a neointimal lining; instead the grafts are lined by fibrin or a pseudointima. Subsequent clinical work has shown that man has a limited ability to organize the fibrin layer and that the ability to ‘heal’ the graft with an intimal lining is confined only to a small zone of pannus ingrowth adjacent to the ends of the graft.

Shortly after implantation, a 1 mm layer of fibrin forms on the inner surface of a fabric prosthesis. This layer is important in small diameter conduits and ultimately leads to occlusion of blood vessels less than 5 mm in diameter. The fibrin layer than becomes organized by ingrowth of fibroblasts arising from the bloodstream and most importantly, from ingrowth through the interstices of the graft wall. This organization is apparent within a few days after graft implantation and is completed within a few weeks. Capillaries traverse the graft interstices to nurture the neoinitimal lining.

Much work has gone into to determining the ideal porosity of textile grafts and to promote quicker healing and ingrowth on tissue into the prostheses. Porosity has been standardized in terms of the volume of filtered water passed per minute through 1 cm² of fabric at a water pressure equivalent to 120 mmHg. Grafts must be minimally porous at implantation to prevent massive haemorrhage but yet have maximum porosity for healing. Most knitted grafts have a porosity between 1200 and 1900 ml/cm².min. The tight woven grafts have porosity as low as 50 ml/cm².min. A newly developed velour graft acts a trellis for cellular ingrowth and leads to firm graft adherence to surrounding tissues.

Abbott WM, Bouchier-Hayes DJ. In: Dardik H, ed. Graft Materials in Vascular Surgery. Chicago: Year Book Medical Publishers, 1978: 59–78.

Abbott WM, et al. Effect of compliance mismatch in vascular graft patency. J Vasc Surg 1987; 5 376–82.

Birinyi LK, Douville EC, Lewis SA, Bjornson HS, Kempczinske RF Increased resistance to bacteremic graft infection after endothelial cell seeding. J Vasc Surg 1987; 5: 193–7.

Brewster DC, Darling RC. Optimal methods of aortoiliac reconstruction. Surgery 1978; 84: 739–48.

Brewster DC, LaSalle AJ, Robison JG, Strayhorn EC, Darling RC Factors affecting patency of femoropopliteal bypass grafts. Surg Gynecol Obstet 1983; 157: 437–42.

Brewster DC, Lasalle AJ, Robison JG, Strayhorn EC, Darling RC Femoropopliteal graft failures. Arch Surg 1983; 118: 1043–7.

Crawford ES, Bomberger RA, Glaeser DH, Salch SA, Russell WC Aortoiliac occlusive disease: Factors influencing survival and function following reconstructive operation over a twenty-five year period. Surgery 1981; 90: 1055–67.

Dale WA. Arterial grafts: 1900–1978. In: Dardik H, ed. Graft Materials in Vascular Surgery. Chicago: Year Book Medical Publishers, 1978: 3–14.

Dardik H, et al. A decade of experience with the glutaraldehyde-tanned human umbilical cord vein graft for revascularization of the lower limb. J Vasc Surg 1988; 7: 336–46.

Eickhoff JH, et al. Four years' results of a prospective, randomized clinical trial comparing polytetrafluoroethylene and modified human umbilical vein for below-knee femoropopliteal bypass. J Vasc Surg 1987; 6: 506–11.

Flinn WR, et al. Improved long-term patency of infragenicular polytetrafluoroethylene grafts. J Vasc Surg 1988; 7: 685–90.

Freischlag JA, Moore WS. Infection in prosthetic vascular grafts. In: Rutherford RB, ed. Vascular Surgery. 3rd edn Philadelphia: W. B. Saunders, 1989: 510–21.

Hasson JE, et al. Mural degeneration in the glutaraldehyde-tanned human umbilical vein graft: Incidence and implications. J Vasc Surg 1986; 4: 243–50.

Houser LS Hashmi FH Jaeger VJ, Chawla SK, Brown L, Kemler RL. Should the greater saphenous vein be preserved in patients requiring arterial outflow reconstruction in the lower extremity? Surgery 1984; 95: 467–72.

Julien S, et al. Biologic and Structural evaluation of 80 surgically excised HUV grafts. Can J Surg 1989; 32: 101–7.

Lasalle AJ, et al. Femoropopliteal composite bypass grafts: current status. Surgery 1982; 92: 36.

Leather RP, Shah DM, Karomody AM. Infrapopliteal arterial bypass for limb salvage: increased patency and utilization of the saphenous vein used ‘in situ’. Surgery 1981; 90: 1000–8.

LoGerfo FW, Quist WC, Nowak MD, Cranshaw HM, Haudenschild CC. Downstream anastomotic hyperplasia: a mechanism of failure in Dacron arterial grafts. Am Surg 1983; 197: 479–83.

Miller, A, et al. Routine intraoperative angioscopy in lower extremity revascularization. Arch Surg 1989; 124: 604–8.

Nunn C, Carter MM, Donohue MT, Hudgins PC. Postoperative dilation of knitted Dacron aortic bifurcation grafts. J Vasc Surg 1990; 12: 291–7.

Ottinger LW, Darling RC, Wirthlin LS, Linton RR. Failure of ultralightweight knitted Dacron grafts in arterial reconstruction. Arch Surg 1976; 111: 145–9.

Ortenwall P, Wadenvik A, Risberg B. Reduced platelet deposition of seeded versus endothelial segments of expanded polytetrafluoroethylene. J Vasc Surg 1989; 10: 374–80.

Ortenwall P, Wadenvik H, Kutti J, Risberg B. Endothelial cell seeding reduces thrombogenicity of Dacron grafts in humans. J Vasc Surg 1990; 11: 403–10.

Prendiville EJ, Yeager A, O'Donnell TF, et al. Longterm results with the above-knee popliteal expanded polytetrafluoroethylene graft. J Vasc Surg 1990; 11: 517–24.

Quinones-Baldrich WJ, Martin-Paredero V, Balier JD, Busuttil RW, Machleder HI, Moore WS. Polytetrafluoroethylene grafts as the first-choice arterial substitute in femoropopliteal revascularization. Arch Surg 1984; 119: 1238–43.

Robison FG, Brewster DC, Abbott WM, Darling RC. Femoropopliteal and tibioperoneal artery reconstruction using human umbilical vein. Arch Surg 1983; 118: 1039–42.

Rosenberg N. Physiologic factors in arterial grafting. In: Dardik H, ed. Graft Materials in Vascular Surgery. Chicago: Year Book Medical Publishers, 1978: 15–28.

Rutherford RB. High risk patient with aortoiliac disease. In: Brewster DC, ed. Common Problems in Vascular Surgery. Chicago: Year Book Medical Publishers, 1989: 150–6.

Rutherford RB, et al. Factors affecting the patency of infrainguinal bypass. J Vasc Surg 1988; 8 236–46.

Satiani B. False aneurysms following arterial reconstruction. Surg Gynecol Obstet 1981; 152: 357–63.

Sauvage LR., Berger KE, Wood SJ, Yates SG, Smith JC, Mansfield PB. Interspecies healing of porous arterial prosthesis. Arch Surg 1974; 109: 698–705.

Shue WB, et al. Prevention of vascular prosthetic infection with an antibiotic-bonded Dacron graft. J Vasc Surg 1988; 8: 600–5.

Sterpette AV, Schultz RD, Feldhaus RJ, Pertz DJ. Seven year experience with polytetrafluoroethylene as above-knee femoropopliteal bypass graft. Is it worth while to preserve the autogenous saphenous vein? J Vasc Surg 1985; 2: 907–12.

Suggs WD, Henriques HF, DePalma RG. Vein cuff interposition prevents juxta-anastomotic neointimal hyperplasia. Ann Surg 1988; 207: 717–23.

Szilagyi DE, Elliott JR, Smith RF, Reddy DJ, McPharlin A. Thirty year survey of the reconstructive surgical treatment of aorto iliac occuluvive disease. J Vasc Surg 1986; 3: 421–35.

Whittemore AD, Clowes AW, Couch NP, Mannick JA. Secondary femoropopliteal reconstruction. Ann Surg 1981; 193: 35–42.

Veith FJ, 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: 104–14.