Preventing AVF thrombosis: the rationale and design of the Omega-3 fatty acids (Fish Oils) and Aspirin in Vascular access OUtcomes in REnal Disease (FAVOURED) study

The incidence and prevalence of severe chronic kidney disease (CKD) is increasing due to ageing of the population and co-morbid conditions such as diabetes. Haemodialysis is the chosen renal replacement therapy for 70% of patients with end-stage kidney disease (ESKD) in Australia. However, effective haemodialysis is critically dependent upon obtaining and maintaining repeated access to the circulation. Delivering the required blood flow (>300 ml/minute) necessary for haemodialysis has long been referred to as the "Achilles' heel" of dialysis. Vascular access options include native arterio-venous fistula (AVF), synthetic arterio-venous graft (AVG) or central venous catheter (CVC). AVF is universally acknowledged as the optimal access option with the best long-term patency, lowest cost and lowest infection rate [1‐3] and is the most prevalent access used in Australasia (75%)[4]. Because vascular access related surgical procedures and complications of vascular access represent a high proportion of all admissions in the ESKD population and are a major economic burden for health care providers[5], various international initiatives have made this a focus of quality care improvements[6].

Several studies confirm the risk of all cause death with artificial vascular access devices, compared with AVF as the reference group, is around 1.5 for AVG and 3 for CVC, with similar rates for infectious mortality [2, 3, 7]. The complications associated with AVG and CVC use are sepsis, especially staphylococcal, vascular malfunction (thrombosis) and death. The increased risk of death is directly attributable to sepsis and the need for interventions to restore patency which in themselves enhance the risk of infection. Furthermore, in comparison with AVF, CVC and AVG have a significantly increased frequency of interventions to maintain patency (assisted patency).

Therefore, while the clinical imperative is to establish AVF in as many patients as possible, primary failure because of early thrombosis and failure of maturation are the major impediments to clinical success[8]. Reports in the literature from the 1970s and 1980s generally describe primary failure rates of between 10 – 20%. A recent meta-analysis examining the primary failure rate of AVF estimated the pooled primary failure rate at 15.3% [9]. However, this meta-analysis included studies back to the 1960s and omitted studies published after the meta-analysis. More contemporary publications are less encouraging with reports of the primary failure rate between 20 – 54%[1]. Consistent with these findings of a high early failure rate, 50% of new patients commence haemodialysis in Australia and New Zealand with a CVC, due to a failed or inadequate AVF[10]. The higher rate of primary failure in the more recent era may reflect changing patient factors – such as older patients and a higher prevalence of diabetes.

Primary failure usually occurs as a result of one of two processes; 1) thrombosis, which usually occurs within weeks of the procedure; 2) inadequate size of the artery or maturation of the vein[8]. Strategies to reduce primary failure rates may therefore include pre-operative identification of unsuitable anatomy by the use of ultrasound, improved surgical technique and pharmacological interventions designed to prevent vessel occlusion (thrombosis).

Early thrombosis is defined as thrombosis within the first 30 days post-operatively[11]. There have been a number of small trials evaluating pharmacological agents aimed at reducing the early thrombosis rate. These trials utilised the anti-platelet agents, aspirin, sulphinpyrazone and ticlopidine and varied in size from as few as 5 patients up to 261 patients. Andrassey et al (n = 92) compared patients given aspirin 500 mg/day for 4 weeks with placebo: the thrombosis rate in the aspirin group was 4% compared with the control group of 24%, OR 0.15 (0.03, 0.73)[12]. Another study [13] with a smaller sample size (n = 68), did not demonstrate a benefit from aspirin. In a study using ticlopidine [14], fistula thrombosis rates for the ticlopidine versus placebo groups were 12% versus 19% (p = 0.10) respectively. Studies using sulphinpyrazone showed variable results, but were underpowered with the largest study enrolling only 36 patients[15]. Pooled data for ticlopidine suggests a reduction in thrombosis rate from 25% to 12% (p < 0.001)[14]. Thus although there is some data to suggest that anti-platelet agents may increase the primary patency of AVF, the limited evidence base and the uncertainty regarding the choice of agent has not supported the widespread use of anti-platelet agents in the prevention of AVF thrombosis. There is a strong clinical need to conduct an adequately powered clinical trial to address the important question of efficacy of anti-platelet therapy in the prevention of early thrombosis of AVF. Since the inception and launch of the FAVOURED study, a US study has been published examining the effect of anti-platelet agent clopidogrel on early AVF thrombosis[16]. We have addressed the findings of this study and their implications for FAVOURED under Discussion below.

Two anti-platelet agents have been chosen to explore their prevention of early thrombosis in AVFs, low dose aspirin and omega-3 fatty acids. Aspirin has been chosen as it has well-established anti-platelet effects, a suggestion of efficacy in the limited studies to date and because it is both inexpensive and an agent familiar to clinicians. Its use as an anti-platelet agent is well established in clinical practice for other purposes relating to thrombosis prevention particularly in patients with established cardiovascular disease. Aspirin's anti-platelet effect is mediated by the inhibition the platelet enzyme cyclo-oxygenase (COX) resulting in blockade of the synthesis of the pro-aggregatory vasoconstrictor, thromboxane A2 (TxA2).

Omega-3 fatty acids have a number of biological effects, which make them an attractive agent for the prevention of vascular access thrombosis. These include the inhibition of platelet aggregation, anti-inflammatory effects and anti-proliferative actions. Omega-3 fatty acids and aspirin both affect the balance between the pro-aggregatory and vasoconstrictor effects of TxA2 and the anti-aggregatory and vasodilator effects of prostacyclin (PGI2) but the mechanism of their effect is different. Omega-3 fatty acids have a weaker inhibitory effect on TxA2 level, their effect mediated by reducing the availability of arachidonic acid (AA), a precursor of TxA2. In addition, omega-3 fatty acids leads to an increase in PGI3 formation (anti-aggregatory and vasodilatation effects equipotent to PGI2)[17]. The anti-inflammatory effect of omega-3 fatty acids is mediated via a reduction in leukotriene and cytokine production. Theoretically, a combination of aspirin and omega-3 fatty acids should result in a more favourable effect on platelet aggregation than either agent used alone.

One study has examined the effects of low-dose aspirin in combination with omega-3 fatty acids on whole blood eicosanoid production[17]. This study demonstrated an additive effect of the combination on TxA2 (40% with aspirin alone vs. 62% with the combination) and a smaller reduction in the concentrations of PGI2 and PGI3 compared with aspirin alone. There has only been one study exploring the efficacy of omega-3 fatty acids in the prevention of vascular access thrombosis. This study involved the use of omega-3 fatty acids as a single agent and only explored its use in AVG, not AVF. Importantly, the study demonstrated a substantial reduction in thrombosis at 1 year in patients with AVG[18]. There are other effects of omega-3 fatty acids, which may be particularly beneficial in patients with CKD. Omega-3 fatty acids has been shown to result in an improvement in lipid profile [18‐20], to reduce BP and heart rate[21], and to attenuate inflammatory responses [22] and oxidative stress[23]. Other postulated but as yet unproven benefits in the CKD population include a reduction in inflammatory response, cardiovascular mortality and a reduction in uraemic pruritis[24].

The relationship between aspirin use and bleeding is a potential concern for clinicians in patients with CKD. Even in the general population, the use of a chronic low dose of aspirin doubles the risk of serious gastrointestinal bleeding [25] and the theoretical risk may be higher in patients with CKD because of the presence of uraemic induced impairment of haemostasis. There have been published studies[26, 27], showing a significant elevation in bleeding times in patients on haemodialysis treatments administered a single dose of aspirin, but surprisingly little evidence-based clinical trial data in this population. This has more recently been explored in the UK-HARP-I study. In this factorial design study involving simvastatin and aspirin as active therapies, allocation to the aspirin group was not associated with an excess of major bleeds (2% in patients on aspirin vs. 3% in patients not on aspirin), but there was a three-fold increase in minor bleeding[28]. Importantly though, in this study, patients received aspirin for 12 months, whereas our study has a much shorter period on therapy of 3 months. In addition, patients on dialysis undergoing renal transplantation have successfully received pre-operative aspirin for the prevention of graft thrombosis without an increased risk of major bleeding [29, 30].

Omega-3 fatty acids appear to be well tolerated, even in high doses, with gastrointestinal complaints particularly nausea, vomiting, diarrhoea and non-specific discomfort being the most often reported. In studies looking at the effect of omega-3 fatty acids on bleeding time it has not been found to be significantly prolonged [31]. In studies that have been done thus far in the CKD cohort, there have been no clear effects demonstrated in relation to platelet aggregation and bleeding times. There has only been one report of a serious bleeding event in a single patient in an uncontrolled study[32]. Furthermore, it has been reported that administration of omega-3 fatty acids can protect the gastric mucosa against aspirin-induced injury, the postulated mechanism being that omega-3 fatty acids counteracts the effect of aspirin on the decrease in prostacyclin (combined effect of PGI2 and PGI3)[33]. This is supported by clinical data suggesting that omega-3 fatty acids decreases gastric erosions and ulcers caused by aspirin or alcohol[34]. Thus the combination of aspirin and omega-3 fatty acids may be expected to be additive in terms of efficacy in vascular access thrombosis, but to have a lower risk of bleeding complications than the use of aspirin alone.

We have chosen an oral route of administration for both aspirin and omega-3 fatty acids in this study. The choice of 100 mg for aspirin is based on this dose having adequate anti-platelet aggregation properties and is the dose most commonly used in patients with cardiovascular disease. A 100 mg dose is readily available. The dose of omega-3 fatty acids (4 g daily) has been shown to be well tolerated and improve cardiovascular risk factors in patients at increased risk of cardiovascular disease [35‐39]. The marine derived omega-3 fatty acids (commonly known as fish oils) capsules chosen (Omacor^©), are commercially available and provide the highest concentration of omega-3 fatty acids per gram of oil. The oral absorption of both agents is excellent and thus alternative routes of administration are not required.

The design of a factorial study is based on the hypothesis that aspirin and omega-3 fatty acids will both be effective therapies and that the combination of aspirin and omega-3 fatty acids will be additive, but not synergistic. In relation to safety concerns, although both aspirin and omega-3 fatty acids prolong bleeding time, the combination may be expected to be safer than when aspirin is used alone.

Ethics approval for the Omega-3 fatty acids (Fish Oils) and Aspirin in Vascular access OUtcomes in REnal Disease (FAVOURED) trial has been obtained from several local Human Research Ethics Committees and will be obtained in all participating centres prior to study initiation and patient enrolment. The study will be performed in accordance with the 2000 Edinburgh, Scotland Revision of the Declaration of Helsinki, the National Health and Medical Research Committee (NHMRC) Statement on Human Experimentation, Joint NHMRC/AVCC Statement and Guidelines on Research Practice, applicable ICH guidelines and the Therapeutic Goods Administration (TGA). The trial has been registered with Australia & New Zealand Clinical Trials Register (ACTRN12607000569404).

This multicentre Australian and New Zealand study has been designed to determine whether the outcome of surgery to create AVF can be improved by the use of aspirin and/or omega-3 fatty acids. Vascular access failure is a significant problem for patients receiving dialysis because it increases their risk of infection, hospitalisation and death. The economic costs of the increased hospitalisations arising from inadequate or artificial vascular access are considerable. Effective therapies to prevent failure of primary AVF are therefore warranted. The therapies nominated in the FAVOURED study are widely available and inexpensive.

It is an appropriate time to study the effects of anti-platelet agents in the prevention of AVF thrombosis. The number of patients requiring dialysis is growing and contemporary studies of the natural history of vascular access outcomes show an increase in the incidence of primary failure, which is most likely related to an increasing prevalence of older patients with comorbidities such as diabetes. At the time of inception of the FAVOURED study, there were no adequately powered studies examining the efficacy of anti-platelet agents in the prevention of AVF thrombosis. Subsequent to FAVOURED being launched in Australia and New Zealand, a US group has published a placebo-controlled parallel arm study examining the prevention of early thrombosis in AVFs using Clopidogrel (an antiplatelet agent) with study treatment commencing one day post surgery with a primary outcome of patency at 6 weeks [16]. There was a 12.2% thrombosis rate in the Clopidogrel arm vs. 19.5% in the control arm: RR of 0.63 (0.46–0.97) p = 0.018, adjusted for interim analyses. Suitability of the AVF for dialysis failed in 63% of the clopidogrel group and in 60% of the placebo group (61% overall). There were no differences in adverse events and, surprisingly, bleeding events were very low in both trial arms. The conclusions drawn by the investigators were that although Clopidogrel is effective in safely preventing primary AVF thrombosis, it was ineffective at increasing functional patency and therefore cannot be recommended.

The implications this trial for the FAVOURED Study include the following:

1. The high functional failure rate is especially concerning and much higher than expected. This may reflect aspects related to surgical (e.g. small number of procedures performed by individual surgeons) and nursing care (needling skills for AVF) in the US and/or patients demographics. Furthermore, significant variation between USA and both European and Australian clinical practice could explain the high US study failure and unsuitability rates. In this context, the generalisability of this American study to other countries is uncertain and justifies additional trial confirmation in other countries;

2. Earlier administration of the active agents before surgery (as in FAVOURED) rather than after may lead to an even greater efficacy in relation to the primary thrombosis rate;

3. The rational for trialing two agents in a factorial design remains valid. Omega-3 fatty acids may be of an additional benefit to aspirin by improving vascular endothelial function and smooth muscle relaxation, which may enhance maturation and greater attainment of functional patency;

4. The low bleeding rates and safety data with Clopidogrel are reassuring in the context of uncertainty regarding the safety of anti-platelet therapy in renal failure, given that anecdotally Clopidogrel is considered to be associated with a higher risk of bleeding than aspirin;

5. The greater duration of follow up (minimum of 1 year) in the FAVOURED Study will allow greater ascertainment of the secondary outcome measure with relation to functional patency. This acquires even greater importance in view of the Clopidogrel Study result.

If the trial demonstrates a positive effect of either or both agents, this will support routine use of these agents in clinical practice and lead to a widespread reduction in thrombosis and an increase in functioning AVF. This has significant patient benefits and health system benefits in reduced costs associated with repeat hospitalisations. If the trial demonstrates no effect of the agents, this will suggest that platelet aggregation is not the major mechanism for thrombosis in this population and that efforts to improve AVF patency would then need to be redirected to other mechanisms associated with access failure, such as vascular anatomical function.