Hemodialysis Vascular Access Dysfunction: Molecular Mechanisms and Treatment


  • Akira Mima

    Corresponding author
    1. Department of Nephrology, Graduate School of Medicine, Institute of Health Biosciences, University of Tokushima, Tokushima, Japan
    Search for more papers by this author

Dr Akira Mima, Department of Nephrology, Graduate School of Medicine, Institute of Health Biosciences, University of Tokushima, Tokushima 770-8503, Japan. Email: akiramima@clin.med.tokushima-u.ac.jp


Hemodialysis vascular access complications are one of the main causes associated with an increase in morbidity and hospitalization in hemodialysis (HD) patients. The most common cause of vascular access dysfunction is venous stenosis as a result of venous neointimal hyperplasia within the peri-anastomotic region (arteriovenous [AV] fistula) or at the graft-vein anastomosis. There have been few studies regarding effective therapeutic interventions for HD vascular dysfunction at the present time, despite the magnitude of the clinical problem. This review will focus initially on the pathology and pathogenesis of HD vascular access dysfunction in the setting of both native AV fistula and polytetrafluoroethylene (PTFE) graft, then experimental and clinical therapies that could potentially be used in the setting of HD vascular access dysfunction.

A well functioning vascular access is of critical importance to the hemodiaysis (HD) patient; however, HD vascular access dysfunction remains one of the most important causes of morbidity in the HD patient population (1). HD vascular access dysfunction is often caused by progressive development of neointimal hyperplastic stenosis (2). These lesions increase intra-access resistance and thereby decrease blood flow, leading to thrombosis, the primary cause of HD vascular access failure. Despite a reasonable understanding of the pathogenesis and pathology of neointimal hyperplasia in the setting of HD vascular access dysfunction, there have been few effective therapeutic interventions to manage HD vascular access stenosis because of the lack of molecular mechanism that lead to development of neointimal hyperplasia in HD patients.


There are two most widely-used forms of permanent HD vascular access: the native arteriovenous fistulae and arteriovenous grafts. Native arteriovenous fistulae, either at the wrist (radiocephalic), elbow (brachiocephalic) or upper arm (brachiobasilic) transpositions are the preferred mode of permanent HD access due to their superior long term survival and lack of infectious complications (3). The pathology of late venous stenosis in native fistulae is similar to that of venous stenosis in the setting of polytetrafluoroethylene (PTFE) dialysis grafts; venous stenosis is primarily due to venous neointimal hyperplasia (4). In arteriovenous (AV) grafts, venous stenosis occurs most commonly at the graft-artery anastomosis and occurs more frequently than previously thought (5). Previous reports showed that this is characterized by smooth muscle cell/myofibroblast migration and proliferation, microvessel formation and migration and extracellular matrix deposition (4). Intimal hyperplasia occurs in discrete phases and is under the control of growth factors; platelet derived growth factor (PDGF) most likely plays a particularly important role in the initial migration of vascular smooth muscle cells from the media to the intima. Basic fibroblast growth factor (bFGF) plays crucial roles for neoangiogenesis within the intima as well as the replication of vascular smooth muscle cells. Transforming growth factor-β (TGF-β) is critical in extra cellular matrix (ECM) production and secretion by vascular smooth muscle cells leading to narrowing of the vessel lumen (6–8). In vivo, inhibition of each of these growth factors can be shown to reduce the development of intimal hyperplasia (9).

Polytetrafluoroethylene dialysis access grafts are a poor second choice for permanent HD vascular access mainly due to increasing the risk of infectious complications (10). It has become apparent that in the vast majority of cases, graft thrombosis occurs after development of a stenosis at or distal to the venous anastomosis (11). Venous stenosis in PTFE dialysis access grafts is due to venous neointimal hyperplasia, which is characterized by the presence of smooth muscle cells, myofibroblasts, and microvessels within the venous neointima (12).

The upstream events in the pathogenesis of venous neointimal hyperplasia in the setting of dialysis grafts and fistulae include (i) hemodynamic stress at the graft-vein or artery-vein anastomosis as a result of a combination of low shear stress over a chronic course (13); (ii) the PTFE graft functions as a foreign body and results in a cellular inflammatory reaction and production of macrophages, which then release a variety of inflammatory cytokines; (iii) dialysis needle injury and turbulence induced by needle placement during the dialysis sessions, leading to an increase in the proinflammatory cytokines (14); (iv) the uremia in HD patients can cause endothelial dysfunction and predispose to venous neointimal hyperplasia (15); and (v) repeated angioplasties causing further endothelial injury (16).

The downstream events are essentially a response to endothelial and smooth muscle cell injury secondary to the upstream events, resulting in the migration of smooth muscle cells from the media to the intima, eventually forming neointimal hyperplasia (17,18). Also, it is important to emphasize that the venous stenosis and venous neointimal hyperplasia that occurs in arteriovenous fistula and PTFE dialysis access grafts is a far more aggressive lesion, with a poorer response to therapy, as compared with arterial stenosis, especially in the setting of peripheral vascular disease. Following vascular injury a large number of downstream events are responsible for smooth muscle cell/fibroblast migration and proliferation, microvessel formation and the production of ECM components (2,17,18).

Endothelial dysfunction

It is reported that endothelial dysfunction, caused by the pathways of inflammation or oxidative stress has been recognized in HD patients (19,20). Endothelial dysfunction is followed by pre-existing venous neointimal hyperplasia, medial hypertrophy and radial artery intima-media thickening (21,22).

Endothelial vascular function is compromised because of changed production of vasodilator and vasoconstrictor substances, particularly nitric oxide (NO). Dysfunction of endothelium-dependent vasodilatation characterizes an atherosclerotic state (23). In addition, it likely plays a casual role in atherogenesis because the underlying changes in vascular intracellular signal transduction and in secretion of endothelium-derived vasoprotective factors predispose to atherosclerosis processes by causing decreased NO production (24). It has been shown that asymmetrical dimethylarginine (ADMA) is an endogenous inhibitor of NO and induces renal injury and endothelial dysfunction (25). Importantly, it has been reported that there is a correlation between elevated levels of ADMA at the time of percutaneous transluminal angioplasty of an initial AV fistula and the risk of a recurrent AV fistula stenosis in HD patients (26).

Increased oxidative stress

Free radicals are significantly unstable molecules containing one or more unpaired elections in atomic or molecular orbitals (27). These molecules, more generally known as reactive oxygen species, along with reactive nitrogen species, are constantly produced in physiological conditions. Recent studies suggest that matrix metalloproteinase (MMP)-mediated vascular remodeling in response to hemodynamic conditions could be modulated by interplay between reactive nitrogen and oxygen species, which can lead to local oxidative stress (17,28). Studies of flow-induced remodeling of rabbit arteriovenous fistulae suggested that the effect of NO might be exerted via modulation of MMPs expression. Ex vivo comparison of human saphenous vein grafts in simulated arterial vs. venous conditions indicated that arterial conditions stimulate MMPs expression and activation, likely via control of the vessel wall's redox state (29). Also, a previous report showed upregulation of MMPs in stenotic and thrombotic AV grafts or AV fistulas (30). Also, it is reported that co-localization of oxidative stress markers with inflammatory cytokines such as TGF-β and PDGF within the neointima of stenotic AV grafts and fistulas (31).

Besides, in patients with CKD, the balance between pro- and antioxidant capacities is shifted towards a state of increased oxidative stress, which has been implicated in the causation of endothelial dysfunction (32). Also, oxidative stress has been linked to several surrogate markers of atherosclerosis in patients with CKD and anti-oxidant therapies have been shown to decrease cardiovascular events in CKD patients (33). Thus, further studies will be needed to evaluate antioxidant therapies for the prevention of AV fistula and graft stenosis.


Inflammation is prevalent in patients with CKD and worse as the CKD progresses towards ESRD (34). Inflammation is a physiological response and contributes to defense mechanisms in living organisms. Inflammatory response must be precisely regulated because both deficiencies and excesses have been linked to morbidity and mortality (35). Supporting this, uremic mice showed increases of inflammation, which developed a greater magnitude of neointimal hyperplasia at the arteriovenous anastomosis as compared with non-uremic animals in a mouse model of AV fistula stenosis (36).

In addition, a large number of cytokines such as PDGF, bFGF, TNF-α, TGF-β, insulin-like growth factor-1 (IGF-1) and IL-6 were increased in the HD patients with AV fistula (18). Inflammatory cells are an important source of MMPs, which are basement membrane-disrupting molecules and could induce HD access dysfunction. Activated macrophages secrete cytokines that upregulate MMPs gene expression in vascular cells (37). Further, nuclear factor-κB (NF-κB) dependent MMPs are inducible by oxidants and proinflammatory cytokines (38). In addition, other groups showed that possible linkages between the presence of macrophages, cytokines and the magnitude of neointimal hyperplasia and venous stenosis within stenotic AV fistulas (39). These findings support that effective therapy for vascular stenosis in HD patients needs to neutralize the oxidative stress and inflammation (Fig. 1).

Figure 1.

The toxins and signaling pathways contributing to hemodialysis (HD) vascular access dysfunction. bFGF, basic fibroblast growth factor; IGF-1, insulin-like growth factor-1; IL-6, interleukin-6; MMP, matrix metalloproteinase; NF-κB, nuclear factor-κB; PDGF, platelet derived growth factor; TGF-β, transforming growth factor-β; TNFα, tumor necrosis factor-α.

Alternative possibilities for the pathogenesis

Although the traditional paradigm for the pathogenesis of neointimal hyperplasia has been thought to be due to the migration of cells from media, more recently a number of studies have suggested a role for the migration of adventitial cells into the intima where they contribute to final neointimal volume (40). The role of bone marrow progenitor cells in the pathogenesis of arteriovenous stenosis, however, remains unclear, with experimental studies describing a role for progenitor cells within neointimal microvessels (41).


Cutting balloon angioplasty

Recently, balloon angioplasty has been recognized as a standard therapy for HD vascular dysfunction. However, restenosis is mainly due to neointima formation, which is caused primarily by the effects of smooth muscle cell proliferation and migration (42). It is reported that in an animal model of carotid artery dilatation by balloon angioplasty, the first step in allowing smooth muscle cell proliferation from the tunica media to the intima is the occurrence of internal elastic lamina rupture (43).

Cutting balloon angioplasty, a newer modification, features several atherotomes, mounted longitudinally on the outer surface of a noncompliant balloon. Few studies have been done to compare the use of cutting balloon angioplasty against its conventional counterpart. However, a randomized controlled trial showed that patients with HD vascular stenosis revealed that there was no difference in the 6-month primary patency rates in the target lesion or the entire vascular access circuit (44).


A liquid form of nitrous oxide enters the balloon and becomes a gas, which simultaneously inflates the balloon and freezes the adjacent vascular tissue. After freezing, cultured smooth muscle cell's viability was significantly reduced, despite the exact mechanisms involved being not well known (45). A recent report has shown that cryoplasty increased the intervention-free interval of five HD grafts with rapidly recurring stenosis from 3 weeks to 16 weeks (46). In contrast, Gray et al. found low anatomical success rates after cryoplasty of dialysis accesses (47). Thus, further clinical trials will be needed to evaluate cryoplasty for HD vascular access stenosis.

Systemic pharmacological agents

It has been reported that a number of pharmacologic agents have been examined for their ability to prevent access stenosis by inhibiting smooth muscle cell proliferation and improving endothelial dysfunction; aspirin (48) and dipyridemole (49) have been shown to prevent stenosis and thrombosis in vascular access survival, but were terminated early because of increased risk of bleeding (50). The National Institutes of Health-sponsored Dialysis Access Consortium (DAC) clopidogrel study is the largest multicenter randomized controlled trial of antiplatelet therapy on fistula outcomes. The investigators carefully considered, however, that there are several problems in this study: 61% of AV fistulas were not suitable for dialysis, even though the patient risk profile was relatively low. Further, patient criteria for an AV fistula in this study were not defined. Thus, a comparison of outcomes with other studies is difficult due to the lack of standardized definitions for fistula primary failure, failure to mature, and now dialysis suitability. Such standardized definitions will be needed. Therefore, some groups suggest that clopidogrel cannot be recommended for use to reduce early AV fistula thrombosis and improve its suitability (51).

In addition, there are a large number of currently available drugs that are potent inhibitors of smooth muscle cell proliferation and migration, and endothelial cell progenitors such as sirolimus (52), rosiglitazone (53), angiotensin receptor blockers (54), angiotensin converting enzyme inhibitors (55) and statins (56). A recent study showed that fish oils have unique biologic properties that positively influence the incidence of graft thrombosis, and therefore, represent a possible treatment strategy for the prevention of access thrombosis (57). The protective actions of calcium channel blockers on HD vascular access suggest that inhibition of platelet aggregation or increases of antioxidant activities might be another possible therapeutic target (58) (Table 1).

Table 1. Clinical trials of therapeutic agents for arteriovenous [AV] fistula or AV graft
Study (drug)Treatment planOutcome
  1. CI, confidential interval; OR, odds ratio.

The Dialysis Outcomes and Practice Patterns Study (DOPPS) (48)Not describedA 16% reduction in the adjusted relative risk for loss of primary graft patency and a 30% (P < 0.001) lower risk of loss of cumulative graft patency (95% CI, 0.42 to 0.95).
Ticlopidine (59)Ticlopidine (250 mg × 2/day)Thrombosis rate 12% in intervention vs 19% in control.
Clopidogrel plus aspirin (50)Clopidogrel (75 mg/day) and aspirin (325 mg/day)Significantly increased risk of bleeding and would likely not result in a reduced frequency of graft thrombosis.
DAC clopidogrel study (Clopidogrel) (60)Clopidogrel (75 mg/day [loading 300 mg/day])Fistula thrombosis occurred in 12%, compared with 20% of participants assigned to receive placebo (OR, 0.01, 95% CI, 0.00–0.15; P = 0.018)
DAC study (Dipyridamole plus aspirin) (61,62)Dipyridamole (200 mg × 2/day) and aspirin (25 mg × 2/day)Favored active treatment over placebo (OR, 0.77, 95% CI, 0.19–3.19; P = 0.03)
Calcium channel blocker (58)Not describedFavored active treatment over no therapy group (OR, 1.1, 95% CI, 1.27–20.5; P < 0.05)
Fish oil (57)Fish oil (4 g/day) vs. control oil (4 g/day)The primary patency rates at one year were 76%, compared with 15% for the control group (OR, 0.07, 95% CI, 0.01–0.49; P < 0.03)

Radiation therapy

The efficacy of endovascular radiation therapy for the prevention and treatment of vascular stenosis through the inhibition of cellular proliferation has been demonstrated (63). Experimental angioplasty models showed significant reductions in luminal stenosis with both endovascular and external vascular radiation therapy (64), and these findings have been confirmed in clinical studies of radiation therapy for coronary restenosis (65). Recently, endovascular radiation in patients with PTFE graft stenosis demonstrated a beneficial effect on primary patency of the graft (66).

Cellular and gene therapies

Cellular and gene therapies could become an effective means of local therapy for neointimal hyperplasia in HD fistulae or grafts (67). Endothelial cell loaded gel foam wraps that were placed around the graft-vein anastomosis or the gene transfer of endothelial NO synthase (eNOS) (68), hepatocyte growth factor (HGF) (69), cyclin-dependent kinase inhibitors (70), and Edifoligide (E2F) (18) could reduce neointimal hyperplasia. Szmitko et al. have reported an antibody against CD34 to recruit endothelial progenitor cells (71). Other groups have tried to increase the number of circulating endothelial progenitor cells using therapies such as granulocyte colony-stimulating factor to achieve similar outcomes (72).

Drug eluting stents

It is well known that the Akt/mTOR pathway is a key signal in cell hypertrophy or migration, which is inhibited by sirolimus (73). Recently, sirolimus and paclitaxel drug eluting stents have demonstrated extreme reductions in restenosis rates (74); however, there are currently no clinical studies on the use of these stents in the clinical settings of HD vascular access. Here, we should state that long-term follow-up studies from the first patients who received the sirolumus stents demonstrate minimal neointimal hyperplasia (75). Therefore, whether these drug-eluting stents will be effective in the setting of HD vascular access is still controversial. Thus, further studies are needed to conclude whether drug eluting stents for HD access dysfunction are efficacious or not.

Perivascular drug delivery

There have been a number of studies in experimental models of arteriovenous graft stenosis that demonstrate the efficacy of perivascular paclitaxel using different modes of drug delivery (76,77). It has been reported that a significant reduction in neointimal hyperplasia has been achieved when paclitaxel eluting wraps are placed around the graft-vein anastomosis in animal models of arteriovenous graft stenosis using PTFE grafts coated with paclitaxel (77,78). Lastly, other groups have demonstrated the efficacy of paclitaxel loaded perivascular polymers in an animal model with arteriovenous graft stenosis (79). Further work is clearly needed to better define the indication of paclitaxel loaded perivascular polymers.


Recently, the National Kidney Foundation K/DOQI guideline recommended routine prospective HD vascular access to detect blood access thrombosis and failure (80). According to this guideline, when access blood flow is less than 600 mL/min or has decreased by more than 25%, intervention is recommended. K/DOQI also recommends that access blood flow should be measured monthly and static dialysis venous pressure should be measured every 2 weeks (80). The guidelines emphasize the importance of using trend analysis to guide referral decisions rather than relying on a single measurement. Appropriate access blood flow evaluation was associated with a significantly reduced relative risk of thrombosis (81,82); however, this surveillance did not increase the likelihood of vascular access survival (83).


In order to advance the field further, we need to better understand both the clinical and experimental pathways that result in venous neointimal hyperplasia using the combination of advances in cellular and molecular pathobiology, biomaterials, and drug delivery techniques.