Operative and perioperative management of infected arteriovenous grafts


Vascular graft infections are a particularly troublesome complication for dialysis patients, many of whom are in an already immunocompromised state. The objective of this review is to detail the risk factors, etiology, diagnosis, perioperative and operative management of vascular graft infections.

J Vasc Access 2017; 18(1): 13 - 21

Article Type: REVIEW



Ehsan Benrashid, Linda M. Youngwirth, Leila Mureebe, Jeffrey H. Lawson

Article History


Financial support: No research funds were utilized in the composition or publication of this manuscript.
Conflict of interest: JHL is a paid consultant, and recipient of research funds from Cryolife Inc. (Kennesaw, GA USA). JHL has additionally been a recipient of speaker honoraria and consulting fees from Cryolife Inc., W.L. Gore and Associates, Inc (Newark, DE USA), Proteon Therapeutics, Inc. (Waltham, MA USA), and Humacyte, Inc. (Morrisville, NC).

This article is available as full text PDF.

Download any of the following attachments:

Introduction and scope of the problem

Hemodialysis (HD) is dependent on functional vascular access, whether through a central venous catheter (CVC), arteriovenous graft (AVG), or a native arteriovenous fistula (AVF). Although HD access practice patterns in the USA have varied over time, the incidence of patients on maintenance dialysis via AVG has slowly decreased over the past decade, with a concomitant rise in the number of patients dialyzing through AVF (1). This is in part due to initiatives set forth by the National Kidney Foundation (NKF), and the concept that AVG- and CVC-based modalities for HD access are at dramatically increased risk of infection as contrasted with AVF. Due to age, multiple comorbidities, malnutrition, and the uremic state present in patients with end-stage renal disease (ESRD), this cohort of individuals are immunocompromised, and are at increased risk for infectious complications, regardless of access type (2, 3). The magnitude of access infection is reinforced by the finding that patients on HD have a several-fold higher rate of annual mortality due to sepsis when compared to the general non-dialysis population (4). Thus, vascular access device infection, particularly for AVG (which must be surgically removed), is a potentially life-threatening complication in these patients. Additionally, it is reported that infection is second only to graft thrombosis in terms of causes for morbidity and access failure in the AVG patient cohort, occurring in 3%-35% of patients (5).

The vascular access specialist is oftentimes called upon to intervene when patients present with access infections, especially when prosthetic material is present. As the bacteriology and management of the infected CVC differs from that of the AVG, the scope of this discussion will be limited to the diagnosis, perioperative, and operative management of infected AVG, serving as a decision-making guide to aid the vascular access surgeon in the identification and management of this medically burdensome problem, which can place patients at risk for being without adequate access for days, weeks, or even months.

Bacteriology and initial antibiotic coverage

Prior to tackling the problem of graft-related infection, the vascular access specialist should have knowledge of the most common organisms implicated in AVG infections, in order to tailor appropriate initial antibiotic coverage. The majority of bacteria implicated in graft infections are Gram-positive organisms, with Staphylococcus species the most frequently cited infectious organism, with ranges of culture positivity reported as high as 70%-90% (of all infections) in several studies (6-7-8). In particular, Staphylococcus aureus and S. epidermidis species are most frequently implicated as they are resident skin flora, which can contaminate graft sites through improper cleansing or inadequate preparation of the skin before, during, and after access creation; or during cannulation for HD (6, 8). Although less common, species of Enterococcus and Clostridium difficile have also been implicated as organisms responsible for graft infection, with some authors advocating for the investigation of sources of gastrointestinal (GI) tract contamination in these cases (6, 7, 9).

Gram-negative organisms appear to be responsible for a much smaller overall percentage of graft infections, but have been notable in prior reports for seeding the graft from GI tract infections and outbreaks from contaminated HD machines (10). There appears to be an increased incidence and risk of Gram-negative graft infection when the lower extremity is utilized as the site of access placement (11). According to this same retrospective study of AVG infections, the most frequent Gram-negative organisms causing graft infections were Escherichia and Enterobacter species, with Citrobacter, Pseudomonas, Serratia, and Proteus species occurring with less frequency. There is likewise the thought that many Gram-negative AVG infections may be polymicrobial (12). Although graft infections with Pasteurella species are rare, owing to the location of the majority of AVG in the extremities and the susceptibility of patients to domestic animal bites in the upper extremities, these have been reported in the literature (13).

Increased susceptibility to Gram-negative graft infection has been noted in HIV-positive patients, and thus a low threshold for “atypical” organisms should be maintained during the treatment of these patients, especially in the setting of a low CD4 cell count (<200) (10). HIV-positive states appear to significantly predispose ESRD patients to AVG infection itself, as reported by Brock et al, with AVG infection rates of 43% in patients with AIDS, 36% in patients with HIV but without symptoms, and 15% in HIV-negative patients (14). Although these organisms all appear to be rare and infrequently reported in the literature, the clinician must assess individual patient risk factors and presentation on a case-by-case basis in order to begin appropriate therapy. Thus, owing to the severity of the conditions and potential for major morbidity or mortality in the HD population, early empiric coverage of both Gram-positive and -negative organisms is advocated, especially if signs of sepsis are present.

Our preferred management strategy for suspected AVG infection does not take into account systemic signs such as fever or hemodynamic instability (indicators of sepsis) prior to the initiation of empiric coverage of both Gram-positive and -negative organisms during the initial presentation. This is typically performed with renally dosed parenteral antibiotics, and empiric coverage is often started early in a patient’s stay, as subclinical infection may have been present for greater than 24 hours. Additionally, antibiotics are easily administered in the emergency department or perioperative setting, and do not typically delay necessary interventions. Blood and urine cultures should be drawn as quickly as possible in patients with fever and overt graft infection, prior to the administration of antibiotics. Needle aspiration or swabbing of open wounds or exposed graft outside of the operating room setting is not recommended; as such, antibiotic administration should not be delayed in order to perform these measures. Although initiation of antibiotics is an important step in the treatment of these patients, operative debridement should not be delayed in those who are suspected to be harboring infection with a particularly virulent organism, as these types of infection can place the vascular anastomosis at an increased risk of disruption.

Our institutional bias is to initiate therapy with vancomycin in combination with piperacillin/tazobactam, as both are widely available, and can be easily dosed to accommodate renal insufficiency, or at the very minimum, administered with HD. The combination of these two antibiotics typically provides adequate Gram-negative and -positive coverage, including methicillin-resistant strains of S. aureus (MRSA) and P. aeruginosa until further culture and sensitivity data can be obtained.

In the case of a reported vancomycin allergy or intolerance, the infusion rate may be slowed and the drug administered after premedication with an antihistamine agent (for intolerance patients only) or a penicillin (or other beta-lactam) or alternatively linezolid (if penicillin-allergic) may be chosen for administration. Linezolid may also be utilized in cases where patients have previously been documented to harbor vancomycin-resistant Enterococci (VRE) colonization or infection, which is oftentimes screened for in intensive care units upon admission. Clindamycin may be used in the case of infections where MRSA species are suspected and the patient has a reported vancomycin and penicillin allergy. Ultimately, when choosing antibiotics, the surgeon must remain mindful of the incidence and possible exposures to MRSA in their local patient population, with potential exposures to this organism in local dialysis centers remaining high. If the probability of MRSA exposure is low, then a more potent Gram-positive antibiotic may not initially be required.

In the case of requirement for Gram-negative coverage and the presence of a penicillin allergy, piperacillin/tazobactam may not be tolerated, and utilization of renally dosed gentamicin, a fluoroquinolone (such as cipro- or levofloxacin), aztreonam, or a third-/fourth-generation cephalosporin is preferred. However, the more “exotic” the antibiotic choice, typically the higher the associated cost, which should also be kept in mind when tailoring antibiotic therapy.

The use of antifungal agents is not necessary as fungal graft infections are exceedingly rare in the literature, but interestingly enough reported as a cause for what appears to be recurrent graft infection and thrombosis (15). It should be reiterated that the recommendations above should not be used as treatment guidelines. Instead, the access specialist should be mindful of their local hospital and dialysis facility’s microbiome and bacterial sensitivities.

Identification and diagnosis

The key to the management of graft infection lies in the rapid diagnosis and identification of graft infection so that definitive management may be provided to the patient. If the problem remains inadequately diagnosed or diagnosed in an untimely fashion, remote infectious complications or sequelae may result or be the initial presenting feature of AVG infection. Remote complications include, for example, brain, lung, or vertebral abscesses, septic arthritis, osteomyelitis, or endocarditis (16).

Patients will often be referred to local emergency departments or clinics through a dialysis access center. Causes for concern should be prompted by: overlying erythema, fluctuance or a mass around the graft site, rapid development of an aneurysm or pseudoaneurysm, pus expressed from the wound, exposure of graft material, draining sinuses or fistulae in the area of the graft, or skin necrosis at the dialysis needle cannulation entry point (Figs. 1 and 2). Early graft infection may in fact present with thrombosis or bleeding at the cannulation site, although these may also be signs of outflow obstruction or other technical issues. Regardless, the presence of these two factors in particular should prompt further investigation into the etiology behind AVG dysfunction. Of more concern are signs of systemic illness – or sepsis – including hypotension (the presence of which may be due to recent HD), leukocytosis, mental status changes, tachycardia, and fevers.

Clinical signs of graft infection vary in presentation and severity. (A) Erythema (outlined area) overlying the tunneled graft site. (B) Skin erosion with purulent material around the graft. (C) Infected pseudoaneurysmal graft segment with overlying erythema expanding past the margin of the pseudoaneurysmal segment, coupled with thinning of the skin and pinpoint areas of bleeding and skin necrosis; these can oftentimes be signs of early graft infection in less dramatic cases. (D) Exposed graft material near the axilla following infection presenting with breakdown of suture material, necrosis of the subcutaneous tissue, and areas of infection located in the peri-graft area. Note that the axilla and femoral regions represent more frequently infected sites for arteriovenous graft placement.

Patient presenting with localized graft infection who underwent admission with subsequent subtotal graft excision and reconstruction with additional prosthesis in a “jump” fashion. (A) Necrosis and skin ulceration present in the vicinity of the graft material. (B) Postoperative result of subtotal graft excision, with white arrows demonstrating areas of anastomotic revision, and black arrow demonstrating area from which isolated, limited infected graft was excised and allowed to heal by secondary intention. Entire new “jump” segment delineated by skin markings. (C) Two (2) month postoperative image in which patient’s wounds have fully healed and access is ready for cannulation.

Owing to the severity of the problem, it is recommended that blood cultures be drawn, especially if the patient has signs of systemic illness, as these patients are at heightened risk for significant complications following an episode of bacteremia. After determination of systemic illness, stratification of the patient for extent of graft excision, and operative planning has begun, it should be determined if the patient has a superficial, local, subtotal, or total graft infection, as well as the timing of the infectious process. As others have suggested, it should be determined whether an infection is related to the surgical graft implantation procedure itself (within 30 days, “wound-related”) or related to repetitive needling for HD access (usually greater than 30 days postoperatively) (17, 18). The chronicity and onset of infection will be impacted by the rapidity with which the graft adheres to the subcutaneous tissues and serves as a physical barrier to spread of early infection, effectively blocking loco-regional spread of the organism (19).

If there is any clinical uncertainty in terms of whether or not there is a graft infection present, several imaging adjuncts are available in terms of aiding diagnosis. These remain important diagnostic tools as graft infections are not always clinically apparent, as evidenced by the report of an indolent subclinical infection discovered on re-intervention for thrombosis or access revision (20).

A readily available bedside test that may be used in serial progression to examine graft flow, extent of infection, and for the presence of pseudoaneurysms and peri-graft fluid collections is duplex ultrasonography (12, 21). Ultrasonography is a useful adjunct aiding diagnosis, but there may be difficulty in differentiating fluid collections that are infectious in etiology, versus hematomas or seromas. In the case of an unclear or nonspecific infection, some advocate for the utilization of radiolabeled indium-111 or technetium-99m leukocyte nuclear medicine scans to further localize the infectious process. The utility of these examinations have been questioned due to: (i) difficulty obtaining these studies; (ii) inadequate timing in obtaining these studies; and (iii) false positives secondary to the inflammatory response and hematoma resulting from thrice-weekly cannulation (12, 21, 22). Although frequently employed in examining abdominal or inguinal vascular prosthesis infection, use of computed-tomography (CT) for examination of extremity vascular HD prosthesis infections is not routine. However, if obtained, expected findings would include stranding and air-fluid collections in the peri-graft area, which could represent a false positive from a recent cannulation hematoma. Owing to the length of time required to perform and intermittent difficulty in obtaining magnetic resonance-based imaging, this modality is rarely employed in diagnosis.

If present, certain clinical factors may obviate the need to perform any imaging – for instance, frank purulence from the wound or exposure of graft material. In these cases we advocate for employing empiric antibiotic therapy and urgent operative intervention.

Operative management

After the diagnosis is confirmed, operative timing determined, and treatment with appropriate antibiotics initiated, the extent of the procedure is predicated on the amount of graft incorporation into the subcutaneous tissue, the extent of the peri-graft infectious process, and the clinical status of the patient. Although imaging can be a useful adjunct in determining these features, the direct visualization of the spread, nature, and extent of the infection along the graft tunnel during intervention cannot be replaced. Blindly operating in a field in which infection is far from certain or merely suspected is not advocated, as the introduction of skin flora to the graft site could occur and result in infection of an otherwise sterile graft.

Importantly, there remain differences in the opinion and strategies for managing graft infections, but the overarching goals of treatment are to adequately and appropriately treat the infectious process and preserve future access options (5, 17-18-19-20, 23-24-25-26-27-28). Approaches are defined by the extent of graft excision (i.e., total or subtotal) with or without subsequent interposition (23). Not infrequently, patients may present in the early postoperative period (within 7-10 days) with swelling and erythema around the graft insertion site. While swelling may be standard within the early postoperative period due to the physiological changes that occur as a result of graft implantation and potential ultrafiltration syndrome (as well as possible seroma or hematoma), the clinician must ensure that a distal pulse or arterial signal is present and the patient still has normal movement of the affected extremity. Percutaneously sampling a suspicious hematoma or seroma is not advised as the inoculation of skin flora into an otherwise sterile space may occur. Erythema around the graft insertion site is more concerning and suspicious for infection, and should be managed with at a minimum outpatient antibiotic administration and close follow-up, as well as avoidance of AVG cannulation. However, especially within the first 48 hours after implantation, a certain degree of erythema may occur as a result of manipulation of the skin, and possible allergic reactions to sterile dressings. Early graft infections are thought to typically result from breaches in aseptic technique and may necessitate graft excision, and duplex ultrasonography may be useful in determining peri-graft fluid collections that are loculated or have other concerning characteristics mandating excision.

In general, there are several well-described operative approaches to dealing with graft infection. They include: complete or total graft excision (TGE) with or without brachial artery ligation, subtotal graft excision (SGE), selective graft excision with subsequent graft interposition, and local incision and drainage with adjuncts that are modifications thereof. General consideration includes the fact that the access specialist should be aware that some form of temporary dialysis is mandatory for those in which the AVG cannot be salvaged, as the graft and access site are totally removed. A temporary (non-tunneled) dialysis catheter can be placed after appropriate antibiotics are initiated, even for grafts that are not removed in toto, so that the patient may dialyze independent of the infected operative site and allow time for the new prosthesis to incorporate into the subcutaneous tissue. Additionally, in order to tailor appropriate and subsequent antibiotic therapy, Gram stain, aerobic and anaerobic cultures, and cultures for fungi and acid-fast bacilli should be obtained at the time of operation, with the excised graft material and adjacent tissue itself sent to pathology and the microbiology lab for further examination. At the time of operation, incision size should not limit complete and thorough surgical examination of the graft tunnel, as any residual infection will result in failure of salvage attempts, if deemed appropriate.

Total graft excision is appropriate in cases where the majority of the graft and graft tunnel is involved in the infectious process, and the graft is not incorporated into the subcutis (i.e., a “free floating” graft). TGE is also preferred if: the anastomosis is involved or under immediate threat, particularly virulent organisms are implicated, the patient is clinically septic, or multiple abscesses are present (5, 17, 19, 20, 23, 25, 27). These represent situations in which the graft and access location cannot be salvaged, and temporary dialysis catheters cannot be avoided. Technical preference in performing upper extremity TGE involves the utilization of a tourniquet, which is inflated to 250 mmHg after initial exsanguination of the limb. To access the anastomoses, incisions are made directly over these sites. Intraoperative ultrasound can be used if the patient’s body habitus prevents adequate palpation of the graft tract or determination of where the anastomoses lie. Operative treatment involves clamping and cutting the graft proximally and distally, with initial attention at the arterial anastomosis, which is taken down and then repaired primarily, followed by subsequent takedown of the venous anastomosis, which is also repaired primarily. It is not routine to use vein patch angioplasty after takedown of the arterial anastomosis, as it requires having to prep an additional operative site and harvest vein segment, which has its own morbidity. This is due to the concern that vein patch angioplasty may be at risk of rupture in the setting of residual infection, although its safe use has been described (5, 19). A counter-incision is utilized near the midpoint of the graft to remove it from the tunnel, which can be difficult to do in the setting of dense graft adhesions. As other authors have proposed, several small incisions can be utilized to complete graft excision (20), although this may limit the adequacy of the mandatory debridement and drainage. After generous irrigation, drainage, and debridement of the infected tunnel tract, the wound is left open, packed, and allowed to heal via secondary intention. Incisions over the anastomotic areas are closed primarily. The patient is given thorough instruction on how to appropriately care for their wound, with home health nursing implemented on discharge if the patient is uncomfortable or unable to do so. Iodine-soaked or impregnated wound packing is not frequently employed, as it may inhibit proliferation of cells involved in wound healing, such as fibroblasts (29). The major limitation of TGE is that the utilization of the affected inflow and outflow tracts is no longer an option for future access.

In the face of dense adhesions, the peri-anastomotic dissection becomes more arduous and time consuming, and the operative risk for bleeding or neurovascular injury is magnified in the face of a thickened inflammatory rind. Some authors have thus advocated SGE, in which graft excision and tissue debridement is performed, but a small residual over-sewn cuff of prosthetic graft material is left behind at the arterial anastomosis. This effectively leaves the native arterial lumen intact, and the site is devoid of a patch that may break down in the face of residual infection and cause hemorrhage (5, 19, 20, 30). It should be emphasized that SGE does not serve as an access salvage procedure; instead it serves as an option to potentially avoid the major morbidity associated with vein patch angioplasty, subsequent blowout and neurovascular injury, which can be hazards while performing exposure in inflamed and infected fields. Additionally, leaving remnant prosthesis behind in patients that are floridly septic is not recommended. Several authors have compared their individual case series utilizing TGE versus SGE – as they are performed for similar indications. Walz et al (19) reported no increased risk of recurrent infection in the subtotal excision group, while Schild et al reported a 17% incidence of recurrent infections that were attributable to residual prosthetic stump infection (from prior SGE) (18). Interestingly, these authors subsequently changed their practice paradigm and now prefer TGE for all cases of AVG infection that would otherwise be candidates for either procedure. Supporting this notion, despite utilizing a less rigorous definition of partial graft excision to include interventions for access infection that left any portion of residual graft behind, Schutte et al (27) reported a significantly higher rate of complications – including sepsis, recurrence of local infection, and hemorrhage – in the SGE cohort. The authors additionally reported that 18.7% of SGE eventually were noted to require a second surgical procedure for infection, and nearly 50% of the SGE reintervention group required eventual TGE. Technically, there is little difference from what has been described for TGE, aside from over-sewing of the graft remnant with either 4-0 or 5-0 polypropylene suture after adequate drainage, debridement, irrigation, and graft excision.

Although rare, authors have also reported the utilization of brachial artery ligation in conjunction with TGE (26, 28). Tan and colleagues (28) report utilization of this technique in the setting of AVG complications, particularly that of uncontrolled local sepsis, often presenting with pseudoaneurysm. In this report, the authors considered this technique only in the context of the inability to perform an access salvage procedure or for prior salvage failure. Ligations were performed in both the proximal (upper arm) and distal (near elbow) brachial artery, with 10% of all procedures having ischemic complications. It should be noted that only very proximal ligations were associated with ischemic complications. Schanzer et al (26) also prefer TGE in the context of infectious complications involving AVG anastomoses. The authors cite the practicality of the procedure in patients that are critically ill and have densely scarred wounds in which dissection at the area of the anastomosis would be too tedious and potentially dangerous, with primary or vein patch repair vulnerable to rupture in the presence of residual infection. In order to mitigate this risk for vein patch rupture or “blowout” in the scenario of residual infection, the authors cite the strategy of control and ligation of the brachial artery after TGE with primary repair of both the arteriotomy and venotomy. Of course, TGE with brachial artery ligation procedures are likewise limited as they further preclude ipsilateral access, and place the arm at risk for ischemic complications.

As described, AVG patients have limitations in terms of further access options owing to anatomy and other features, hence many have preferred graft salvage, which was first described by Bhat et al in 1980 (31). AVG preservation with interposition of new graft material in an uninfected field allows the surgeon to obviate the need for temporary dialysis catheter placement, as the graft may be cannulated almost immediately postoperatively (17). This should be performed only on an already well-incorporated graft segment, which requires thorough communication with the dialysis staff. This likewise can be achieved by indicating the permissible cannulation site with a non-permanent skin marker, the direction of flow, and inflow and outflow sites. Limited graft excision with interposition is performed for non-toxic patients who present with well-localized infections that do not involve the anastomosis. Local incision and drainage with subsequent partial graft excision of the infected portions is preferentially performed through an incision separate from any in which new access is placed or tunneled. A preferred technique for SGE with interposition is to first perform the bypass in a new tunnel created in a separate incision away from the infected portion of the graft. After creation of the new interposition or bypass, sterile dressings and drapes should be placed on this wound and the infected graft segment addressed via direct incision, debridement, and limited graft excision. This can be done in a single or staged fashion – to allow for inflammation to subside, as described by Raju (25).

It should be noted that there is inherent morbidity that must be both considered by the clinician and explained to the patient when performing partial excision with bypass. This has been examined given the preference of vascular access specialists for graft salvage and the issues that arise by attempting to preserve as much graft material as possible. Patients who underwent partial versus total graft excisions have been determined to be prone to reoperation (including TGE) and recurrent infection (19, 27). Taylor et al (17) were able to salvage 60% of all infected grafts, with no difference in polytetrafluoroethylene (PTFE) or bovine xenografts. Deneuville also found that 33% of those patients with infected AVG that underwent reconstructive procedures subsequently developed recurrent infection (24).

Additional adjuncts for use as conduit in partial graft excision with reconstruction are xenografts, available as ProCol® (CryoLife, Inc., Kennesaw, GA USA), a bioprosthesis constructed of bovine mesenteric vein; or Omniflow® II (LeMaitre Vascular, Inc., Burlington, MA USA), a biosynthetic vascular graft derived from ovine collagen around a polyester mesh scaffold (32-33-34-35-36). However, xenograft use for HD access in the setting of infection is limited.

However, a biologic, allogenic conduit noted for use in patients at risk of infectious access complications is CryoVein® (CryoLife, Inc., Kennesaw, GA USA), a cryopreserved saphenous or femoral vein (37, 38). In a study utilizing this vein in a similar cohort to the partial excision with interposition group, Matsuura et al (37, 38) demonstrated a lack of recurrent infection when implanted “in proximity” to sites of infection, with eventual cannulation for HD within 14 days of placement. Data to the contrary was presented by Bolton et al (39) in a series of patients with active graft infections, where an astounding 65% major graft-related complication rate was observed, including an alarmingly high rate of repeat infection and graft thrombosis, with a long-term follow up patency rate of 25% (39). Utilization of xenograft or allograft is thought to allow the surgeon to avoid catheter-based modalities for HD access and the harvesting of saphenous or femoral vein, which prolongs operative time, may not be available, and has long-term implications in this patient cohort, who are at risk of requiring subsequent procedures for peripheral vascular and coronary disease, which may require saphenous vein harvest. Unfortunately, there remain concerns over the use of cadaveric human conduits in those patients that are potential transplant recipients, as these grafts have been shown to induce allosensitivity as measured by panel reactive antibody (PRA) values, leading some to conclude that transplantation is precluded by those who received these grafts (40). Emerging technologies in which allogenic grafts created from human endothelial cells and fibroblasts have demonstrated a marked lack of induction of PRA values, which would not preclude these patients from potential transplantation (41).

Rarely, patients with AVG have undergone local wound management strategies in order to salvage the graft. Negative-pressure therapy (VAC) has been utilized in conjunction with aggressive debridement and antibiotics to salvage HD prostheses after initial flap coverage with continued dialysis through the prosthesis and without subsequent infection (42). As described in a series by Dosluoglu et al (43), exposed PTFE grafts used for arterial reconstruction were aggressively debrided and managed with VAC closure without subsequent infectious complications. There remain a lack of prospective trials comparing attempted salvage therapy with “conservative” local debridement and VAC therapy, so it is recommended that negative-pressure devices are optimally employed as adjuncts to treatment after adequate soft tissue coverage of exposed graft material.

Additional groups for consideration are those presenting with thrombosed AVG that have not been cannulated for some time presenting with systemic illness and verified bacteremia. Although data is limited, Ayus and Sheikh-Hamad presented their retrospective findings of silent graft infections in patients with thrombosed AVG (22). The authors noted that radiolabeled leukocyte scanning was able to localize the source of infection to the thrombosed graft in a bulk of patients with fevers of unknown origin, revealing that thrombosed AVG should be high on the list of infectious sources in HD patients. In this scenario, TGE should be performed only after complete infectious workup rules out other sources of illness. However, with improvements in imaging resolution and the sensitivity of modern imaging, an expensive and time-consuming nuclear medicine study can be avoided with ultrasound or cross-sectional CT imaging.

A point of controversy is the utilization of endovascular techniques to treat infectious access graft complications. Although other authors note that utilization of endovascular exclusion techniques for pseudoaneurysms are an effective means to prevent impending rupture and other complications (44, 45), there remains the fact that any endovascular prosthesis may itself become secondarily infected. Additionally, the area that is actively infected is not completely addressed until it is debrided. Many endovascular devices utilized in the process of exclusion of pseudoaneurysms are covered (“stent grafts”), which, given the presence of graft material, are at increased risk of infectious complications. A recent analysis has demonstrated that endoluminal graft exclusion for the specific indication of pseudoaneurysm placed patients at significantly increased risk of subsequent AVG infection versus bare or covered stents placed for other reasons (42.1% vs 18.2%), with 16.3% of AVG patients receiving stents having to undergo graft excision due to infection (46). It is generally recommended to avoid the utilization of endovascular techniques for any situation in which infection is suspected in dialysis access grafts, including pseudoaneurysm.

Other experimental techniques yet to be verified in clinical settings have been attempted in animal models, such as the peri-graft implantation of antibiotic-impregnated beads PTFE graft infection (47). Regional administration of antibiotics appeared to decrease biofilm concentration, without major harm to the animal. Rifampin-impregnated grafts have also been reported as feasible and safe in animal models and clinical scenarios of graft infection (48-49-50-51-52-53), although these grafts do not seem to have gained substantial ground in the vascular access community.

Postoperative care

It is important to initiate antibiotic therapy early in the presentation, with subsequent tailoring of antibiotics based on intraoperative Gram stain and cultures. Assistance regarding duration of antibiotic therapy can be delineated by colleagues in infectious diseases, although authors typically describe use of intravenous antibiotics for two to six weeks (depending on blood culture positivity), which can be administered with dialysis (7, 54). In cases of cellulitis without definitive abscess or graft involvement, antibiotics can be administered orally for 14 days, with follow-up scheduled at this time.

With regards to wound care, the patient should be notified early in the course that they will require aggressive wound therapy, possibly with the assistance of home health providers. Of utmost importance is the vigilance of the vascular access specialist in postoperative surveillance, in order to monitor wound and access-related complications, allowing for reintervention in prompt fashion. One should be mindful of the necessity to inform the patient and dialysis access center to notify the vascular access provider of any changes in the wound. Nontunneled catheters should be exchanged for tunneled catheters when blood cultures are negative if total graft excision is performed. Further access placement in TGE and SGE patients can be considered one month postoperatively.

Emerging technologies and future directions

The emergence of three-dimensional (3D) printing and refinement of tissue engineering strategies makes this a particularly exciting time for vascular access specialists. With these novel technologies, physicians are entering an era where the production of blood vessels tailored to the patient that are non-immunogenic is possible. This is particularly important in those being considered for renal transplantation, as well as those at risk for infection (55). With the potential of these “living” bioscaffolds to be seeded with endothelial cells and remodel in a manner to native vessels, physicians are on the cusp of a non-immunogenic, non-thrombotic, infection-resistant, and readily available bypass and HD access conduit.


Infection of arteriovenous grafts remain a source of major morbidity and a limitation in their use. This is problematic as these devices are often the only remaining option for patients without adequate vasculature for native arteriovenous fistulae, or who have otherwise exhausted other access options. The most common organisms implicated in infection of AVG are Staphylococcus species, owing to improper hygiene and/or prep prior to access creation or cannulation. Early diagnosis is key in preventing possible septic and remote infectious complications. An aggressive response to post-access implantation infection is warranted, with immediate empiric antibiotic coverage followed by urgent operative debridement and excision of prosthetic graft material if indicated. Given the nature of the anatomic factors for AVG placement in the recent era of improved preoperative imaging and adherence to guidelines, a salvage-first strategy should be attempted if at all feasible. The key to preventing infectious complications of access is the prevention and surveillance of these patients in a multidisciplinary fashion, by nephrologists, dialysis centers, and vascular access specialists.


Financial support: No research funds were utilized in the composition or publication of this manuscript.
Conflict of interest: JHL is a paid consultant, and recipient of research funds from Cryolife Inc. (Kennesaw, GA USA). JHL has additionally been a recipient of speaker honoraria and consulting fees from Cryolife Inc., W.L. Gore and Associates, Inc (Newark, DE USA), Proteon Therapeutics, Inc. (Waltham, MA USA), and Humacyte, Inc. (Morrisville, NC).
  • 1. Pisoni RL Zepel L Port FK Robinson BM Trends in US vascular access use, patient preferences, and related practices: an update from the US DOPPS practice monitor with international comparisons. Am J Kidney Dis 2015 65 6 905 915 Google Scholar
  • 2. Ryan SV Calligaro KD Dougherty MJ Management of hemodialysis access infections. Semin Vasc Surg 2004 17 1 40 44 Google Scholar
  • 3. Jaber BL Bacterial infections in hemodialysis patients: pathogenesis and prevention. Kidney Int 2005 67 6 2508 2519 Google Scholar
  • 4. Sarnak MJ Jaber BL Mortality caused by sepsis in patients with end-stage renal disease compared with the general population. Kidney Int 2000 58 4 1758 1764 Google Scholar
  • 5. Anderson JE Chang AS Anstadt MP Polytetrafluoroethylene hemoaccess site infections. ASAIO J 2000 46 6 S18 S21 Google Scholar
  • 6. Minga TE Flanagan KH Allon M Clinical consequences of infected arteriovenous grafts in hemodialysis patients. Am J Kidney Dis 2001 38 5 975 978 Google Scholar
  • 7. Lew SQ Kaveh K Dialysis access related infections. ASAIO J 2000 46 6 S6 S12 Google Scholar
  • 8. Tabbara MR O’Hara PJ Hertzer NR Krajewski LP Beven EG Surgical management of infected PTFE hemodialysis grafts: analysis of a 15-year experience. Ann Vasc Surg 1995 9 4 378 384 Google Scholar
  • 9. Akoh JA Patel N Infection of hemodialysis arteriovenous grafts. J Vasc Access 2010 11 2 155 158 Google Scholar
  • 10. Lafrance JP Rahme E Lelorier J Iqbal S Vascular access-related infections: definitions, incidence rates, and risk factors. Am J Kidney Dis 2008 52 5 982 993 Google Scholar
  • 11. Harish A Allon M Arteriovenous graft infection: a comparison of thigh and upper extremity grafts. Clin J Am Soc Nephrol 2011 6 7 1739 1743 Google Scholar
  • 12. Padberg FT Jr Calligaro KD Sidawy AN Complications of arteriovenous hemodialysis access: recognition and management. J Vasc Surg 2008 48 5 Suppl 55S 80S Google Scholar
  • 13. Schneider JR White GW Dejesus EF Pasteurella multocida-infected expanded polytetrafluoroethylene hemodialysis access graft. Ann Vasc Surg 2012 26 8 1128 e15-1128 e17 Google Scholar
  • 14. Brock JS Sussman M Wamsley M Mintzer R Baumann FG Riles TS The influence of human immunodeficiency virus infection and intravenous drug abuse on complications of hemodialysis access surgery. J Vasc Surg 1992 16 6 904 910 discussion 911-912. Google Scholar
  • 15. Lee HK Park JH Kim ES Kim HS Han SY Arteriovenous thrombosis due to Aspergillus in a patient with hemodialysis. J Vasc Access 2013 14 2 204 Google Scholar
  • 16. Sexton DJ Vascular access infections in patients undergoing dialysis with special emphasis on the role and treatment of Staphylococcus aureus. Infect Dis Clin North Am 2001 15 3 731 742, vii Google Scholar
  • 17. Taylor B Sigley RD May KJ Fate of infected and eroded hemodialysis grafts and autogenous fistulas. Am J Surg 1993 165 5 632 636 Google Scholar
  • 18. Schild AF Simon S Prieto J Raines J Single-center review of infections associated with 1,574 consecutive vascular access procedures. Vasc Endovascular Surg 2003 37 1 27 31 Google Scholar
  • 19. Walz P Ladowski JS Partial excision of infected fistula results in increased patency at the cost of increased risk of recurrent infection. Ann Vasc Surg 2005 19 1 84 89 Google Scholar
  • 20. Ryan SV Calligaro KD Scharff J Dougherty MJ Management of infected prosthetic dialysis arteriovenous grafts. J Vasc Surg 2004 39 1 73 78 Google Scholar
  • 21. Legout L D’Elia PV Sarraz-Bournet B et al. Diagnosis and management of prosthetic vascular graft infections. Med Mal Infect 2012 42 3 102 109 Google Scholar
  • 22. Ayus JC Sheikh-Hamad D Silent infection in clotted hemodialysis access grafts. J Am Soc Nephrol 1998 9 7 1314 1317 Google Scholar
  • 23. Ceppa EP Sileshi B Beasley GM Lawson JH Surgical excision of infected arteriovenous grafts: technique and review. J Vasc Access 2009 10 3 148 152 Google Scholar
  • 24. Deneuville M Infection of PTFE grafts used to create arteriovenous fistulas for hemodialysis access. Ann Vasc Surg 2000 14 5 473 479 Google Scholar
  • 25. Raju S PTFE grafts for hemodialysis access. Techniques for insertion and management of complications. Ann Surg 1987 206 5 666 673 Google Scholar
  • 26. Schanzer A Ciaranello AL Schanzer H Brachial artery ligation with total graft excision is a safe and effective approach to prosthetic arteriovenous graft infections. J Vasc Surg 2008 48 3 655 658 Google Scholar
  • 27. Schutte WP Helmer SD Salazar L Smith JL Surgical treatment of infected prosthetic dialysis arteriovenous grafts: total versus partial graft excision. Am J Surg 2007 193 3 385 388 discussion 388. Google Scholar
  • 28. Tan YM Tan SG Emergency ligation of the brachial artery for complications of vascular access. Br J Surg 2005 92 2 244 245 Google Scholar
  • 29. Thomas GW Rael LT Bar-Or R et al. Mechanisms of delayed wound healing by commonly used antiseptics. J Trauma 2009 66 82 90 discussion 90-91. Google Scholar
  • 30. Gifford RR Management of tunnel infections of dialysis polytetrafluoroethylene grafts. J Vasc Surg 1985 2 6 854 858 Google Scholar
  • 31. Bhat DJ Tellis VA Kohlberg WI Driscoll B Veith FJ Management of sepsis involving expanded polytetrafluoroethylene grafts for hemodialysis access. Surgery 1980 87 4 445 450 Google Scholar
  • 32. Hatzibaloglou A Velissaris I Kaitzis D Grekas D Avdelidou A Kiskinis D ProCol vascular bioprosthesis for vascular access: midterm results. J Vasc Access 2004 5 1 16 18 Google Scholar
  • 33. Widmer MK Aregger F Stauffer E et al. Intermediate outcome and risk factor assessment of bovine vascular heterografts used as AV-fistulas for hemodialysis access. Eur J Vasc Endovasc Surg 2004 27 6 660 665 Google Scholar
  • 34. Bozoglan O Mese B Eroglu E et al. Which prosthesis is more resistant to vascular graft infection: polytetrafluoroethylene or Omniflow II biosynthetic grafts? Surg Today 2015 Epub ahead of print. Google Scholar
  • 35. Palumbo R Niscola P Calabria S et al. Long-term favorable results by arteriovenous graft with Omniflow II prosthesis for hemodialysis. Nephron Clin Pract 2009 113 2 c76 c80 Google Scholar
  • 36. Töpel I Betz T Uhl C Wiesner M Bröckner S Steinbauer M Use of biosynthetic prosthesis (Omniflow II®) to replace infected infrainguinal prosthetic grafts—first results. Vasa 2012 41 3 215 220 Google Scholar
  • 37. Matsuura JH Johansen KH Rosenthal D Clark MD Clarke KA Kirby LB Cryopreserved femoral vein grafts for difficult hemodialysis access. Ann Vasc Surg 2000 14 1 50 55 Google Scholar
  • 38. Matsuura JH Rosenthal D Wellons ED Castronovo CS Fronk D Hemodialysis graft infections treated with cryopreserved femoral vein. Cardiovasc Surg 2002 10 6 561 565 Google Scholar
  • 39. Bolton WD Cull DL Taylor SM et al. The use of cryopreserved femoral vein grafts for hemodialysis access in patients at high risk for infection: a word of caution. J Vasc Surg 2002 36 3 464 468 Google Scholar
  • 40. Benedetto B Lipkowitz G Madden R et al. Use of cryopreserved cadaveric vein allograft for hemodialysis access precludes kidney transplantation because of allosensitization. J Vasc Surg 2001 34 1 139 142 Google Scholar
  • 41. Wystrychowski W McAllister TN Zagalski K Dusserre N Cierpka L L’Heureux N First human use of an allogeneic tissue-engineered vascular graft for hemodialysis access. J Vasc Surg 2014 60 5 1353 1357 Google Scholar
  • 42. Vallet C Saucy F Haller C Meier P Rafoul W Corpataux JM Vacuum-assisted conservative treatment for the management and salvage of exposed prosthetic hemodialysis access. Eur J Vasc Endovasc Surg 2004 28 4 397 399 Google Scholar
  • 43. Dosluoglu HH Schimpf DK Schultz R Cherr GS Preservation of infected and exposed vascular grafts using vacuum assisted closure without muscle flap coverage. J Vasc Surg 2005 42 5 989 992 Google Scholar
  • 44. Shah AS Valdes J Charlton-Ouw KM et al. Endovascular treatment of hemodialysis access pseudoaneurysms. J Vasc Surg 2012 55 4 1058 1062 Google Scholar
  • 45. Shemesh D Goldin I Zaghal I Berelowitz D Verstandig AG Olsha O Stent graft treatment for hemodialysis access aneurysms. J Vasc Surg 2011 54 4 1088 1094 Google Scholar
  • 46. Kim CY Guevara CJ Engstrom BI et al. Analysis of infection risk following covered stent exclusion of pseudoaneurysms in prosthetic arteriovenous hemodialysis access grafts. J Vasc Interv Radiol 2012 23 1 69 74 Google Scholar
  • 47. Keeling WB Myers AR Stone PA et al. Regional antibiotic delivery for the treatment of experimental prosthetic graft infections. J Surg Res 2009 157 2 223 226 Google Scholar
  • 48. Bandyk DF Novotney ML Johnson BL Back MR Roth SR Use of rifampin-soaked gelatin-sealed polyester grafts for in situ treatment of primary aortic and vascular prosthetic infections. J Surg Res 2001 95 1 44 49 Google Scholar
  • 49. Gao H Sandermann J Prag J Lund L Lindholt JS Rifampicin-soaked silver polyester versus expanded polytetrafluoro-ethylene grafts for in situ replacement of infected grafts in a porcine randomised controlled trial. Eur J Vasc Endovasc Surg 2012 43 5 582 587 Google Scholar
  • 50. Goëau-Brissonnière O Javerliat I Koskas F Coggia M Pechère JC Rifampin-bonded vascular grafts and postoperative infections. Ann Vasc Surg 2011 25 1 134 142 Google Scholar
  • 51. Javerliat I Goëau-Brissonnière O Bruneval P Coggia M Experimental study of a new vascular graft prebonded with antibiotic: healing, toxicity, and antibiotic retention. Ann Vasc Surg 2007 21 5 603 610 Google Scholar
  • 52. Javerliat I Goëau-Brissonnière O Sivadon-Tardy V Coggia M Gaillard JL Prevention of Staphylococcus aureus graft infection by a new gelatin-sealed vascular graft prebonded with antibiotics. J Vasc Surg 2007 46 5 1026 1031 Google Scholar
  • 53. Sacar M Goksin I Baltalarli A et al. The prophylactic efficacy of rifampicin-soaked graft in combination with systemic vancomycin in the prevention of prosthetic vascular graft infection: an experimental study. J Surg Res 2005 129 2 329 334 Google Scholar
  • 54. Akoh JA Vascular access infections: epidemiology, diagnosis, and management. Curr Infect Dis Rep 2011 13 4 324 332 Google Scholar
  • 55. Manson RJ Unger JM Ali A Gage SM Lawson JH Tissue-engineered vascular grafts: autologous off-the-shelf vascular access? Semin Nephrol 2012 32 6 582 591 Google Scholar



  • Division of Vascular Surgery, Department of Surgery, Duke University Medical Center, Durham, North Carolina - USA

Article usage statistics

The blue line displays unique views in the time frame indicated.
The yellow line displays unique downloads.
Views and downloads are counted only once per session.

No supplementary material is available for this article.