Outcomes of Patients Referred for Arteriovenous Fistula Construction: A Systematic Review

Chronic Kidney Disease (CKD) affects 10-16% of the US population and its incidence is rising due to increasing prevalence of associated risk factors. Renal replacement therapy is required to treat late stage CKD and hemodialysis is the preferred modality for many patients. Vascular access is required for hemodialysis and arteriovenous fistulas (AVF) are currently the gold standard. This review intended to collate current knowledge on AVF outcomes regarding both the patient and fistula. Scopus and Medline were utilized to identify relevant literature. Inclusion and exclusion criteria were applied to narrow search results. Among CKD patients, 33.5-77.4% require a central venous catheter (CVC) before dialysis through a fistula. Many patients (33-51%) use a CVC regardless of AVF creation due to fistula immaturity or failure. There are large variations in AVF creation policies internationally; 16% of American hemodialysis patients use a fistula compared to 72% of German patients. Primary patency and primary AVFs' failure ranges from 60-70% and 20-26%, respectively. AVFs reduce morbidity and mortality in CKD. At present, too many patients are receiving hemodialysis through a CVC. Inadequate referral times for AVF creation can lead to fistula immaturity or failure in the intervention. Many countries are lagging behind recommended AVF creation rates published by the Kidney Disease Outcomes Quality Initiative. There is a paucity of literature concerning when a patient should be referred for AVF creation. It is paramount to have better predictive outcome measures and more clarity as to when patients will benefit from an AVF.


Introduction
Chronic Kidney Disease (CKD) is a pathologic condition resulting in a progressive decline of kidney function. It currently affects 7-12% of individuals globally, while its incidence is rapidly rising. [1][2][3][4] The disease involves structural pathology such as nephron loss and fibrosis, which result in decreased glomerular filtration. 5 These insults to the kidney contribute to systemic complications of CKD, including fluid and electrolyte abnormalities, anaemia, mineral-bone disorder, metabolic acidosis, hyperuricaemia, hypertension, dyslipidemia, cardiovascular disease and endocrine dysfunction. 5 Severity of the disease can be divided into five stages depending on estimated glomerular filtration rate (eGFR), with substantial loss of kidney function and end-stage renal disease (ESRD) comprising stages IV and V, respectively. [6][7][8] Renal replacement therapy is necessary once a patient has progressed to ESRD. Kidney transplants are ideal for renal replacement; however, they are not widely accessible. Consequently, hemodialysis (HD) is often the modality of choice for patients in ESRD. 9 The process of HD clears the blood of uremic toxins using a series of pumps, membranes and dialysates. Patients undergoing this long-term therapy require permanent vascular access placement, as suggested by the Kidney Disease Outcomes Quality Initiative (KDOQI). 1,8,10 The three main types of vascular access include arteriovenous fistulae (AVF), arteriovenous grafts (AVG) and central venous catheters (CVC). 8 The current gold-standard for HD access is the formation of an AVF, as it is clinically reported to have better patient outcomes with reduced morbidity and improved survival. 8,[10][11][12][13] AVFs are established in the forearm through surgical anastomosis of a relatively small, peripheral artery with a larger subcutaneous vein. 14 According to the KDOQI, the optimal timing for AVF creation is 6 months before cannulation, however this maturation can be affected by factors, such as age and gender. 8,11 In the case where a patient's vascular integrity does not support a fistula, an AVG can be implemented. Studies indicate that AVGs have more drawbacks compared to AVFs. These include higher infection susceptibility potentially resulting in sepsis, reduced patency, and greater risk of complications ultimately leading to repeated interventions and diminished survival. 15 CVCs, on the other hand, are generally employed prior to AVF or AVG maturation, when immediate initiation of HD is required. Infection and thrombosis among other life threatening complications, are persisting impediments. 16 For instance, Lee et al. noted that catheter-related bacteremia was present in half of all HD patients studied 6 months after CVC implantation. 17 Furthermore, HD patients with CVCs have a 3.43-fold increase in relative mortality risk compared to patients with an AVF. [17][18][19] Therefore, AVFs are the gold standard method to attain vascular access in patients undergoing HD, as they result in fewer complications when compared to CVCs and AVGs. 8,[10][11][12][13] However, to better maintain access sites and retain the integrity of the vasculature, radiocephalic AVFs (RCAVF) are the recommended option by the KDOQI, and are associated with improved patient survival. 12,20 In cases where the creation of a RCAVF is not feasible due to poor vasculature brachiocephalic, brachiobasilic, and brachiobrachial AVFs can be created. 21 In spite of the clear benefits provided by AVFs for HD patients, they do carry some drawbacks and can potentially pose a risk for serious complications requiring hospitalization. 22 The reported complications surrounding fistulas include aneurysm development, stenosis of the vein, dialysis-associated steal syndrome (due to ischemia), thrombosis, and infection. 22 Additionally, the primary failure rates of AVF formation and maturation are approximated at 23% and 20-60%, respectively. 1,[23][24][25] Indeed, AVFs are a source of patient morbidity, however, they still remain the principal type of vascular access for HD compared to AVGs or CVCs. 8,21 To ensure its success, the timing of AVF creation relative to HD must be considered. As there exists a long lag time between AVF formation and usage, fistulas can be created in a pre-emptive manner to circumvent potential CVC access, if HD is required. Even so, if the AVF is not needed or not used for access by the patient, the surgical procedure to create the fistula can result in unwarranted patient distress. 1 Moreover, the average maturation time for an AVF falls between 148 to 285 days, with a 75% successful cannulation rate at 16 weeks (112 days) postsurgery. 1,25,26 Taking the aforementioned factors into account, the temporally sensitive nature of this complex therapeutic intervention must be considered at the time of consultation. The time-course for formation of the fistula needs to be appropriately managed from initial patient referral to when the mature AVF will be needed. As the incidence of CKD rises, more patients will require renal replacement therapy and the creation of an AVF for HD. Therefore, it is imperative that we analyze and examine current practices to determine the best course of action.

Objectives
This review will analyze the literature on arteriovenous fistulas used for vascular access to determine the current paradigms on: 1) Outcomes regarding the patient after AVF creation i. The proportion of patients who end up on dialysis ii. The number of patients who require a central line due to AVF immaturity or failure iii. Whether the patient receives a transplant and avoids dialysis or dies before dialysis commencement iv. Estimated prognosis between AVF and non-AVF patients 2) Outcomes regarding the fistula itself i. Primary AVF patency, secondary AVF patency ii. Primary and secondary AVF failure iii. Fistula maturation times 3) Predictive factors for AVF outcomes i. Sex, morbidity, lifestyle, site of fistula ii. Vein diameter, arterial flow rate.

Search Strategy
An electronic search was conducted using the databases of Medline (PubMed) and Scopus. The search identified publications pertaining to the objectives and research question of the current study. In total, 10 articles were identified between the two database searches (Figure 1).

Study Eligibility:
The inclusion and exclusion criteria were generated a priori. Studies were included if they addressed populations who underwent AVF creation for HD access, if primary patency or AVF maturation were mentioned in the abstract and if the authors reviewed the outcomes of their subjects after AVF creation. The selected manuscripts all present new data published in its first report and are not review papers.
Exclusion criteria specified studies carried out on non-human species, articles not available in English, and articles that were not available as Open Access. Additionally, papers addressing exclusively elderly or adolescent populations were removed. Finally, articles were excluded if the study evaluated endovascular fistula creation, focused primarily on anaesthetic technique or on steal syndrome, or if the study addressed AVF interventions and revisions rather than primary creation, it was excluded.

Results
All of the papers identified by the search methods were assessed for their quality and validity using the Evidence-based librarianship (EBL) critical appraisal tool. 27,28 The abbreviated results are displayed in Table 1 and the extended appraisal can be viewed in Appendix A. All of the selected articles were deemed valid when applying the EBL criteria. 27 The methodology of all the papers identified to be eligible for review were adequate and minimized bias that the individual study may be susceptible to.

Patient Outcomes
A retrospective study looking at HD patients in Canada showed that 27% patients had at least one AV access created in a study population of 17,183 ( Table 2). 29 Of the patients who had an AVF created, 65% were able to cannulate it for HD, while 33% had to resort to a CVC. 29 Of the patients who were referred late to a nephrologist, 8% had an AV access created as opposed to 39% who were referred early. 29 Prior to AVF creation, CVC use occurred in 35.7-77.4% of patients. 13,21,26,30,31 A study investigating the natural history of AVFs noted that 66% of vascular access procedures were to create an AVF, with 33% of these being RCAVFs. 26 Four studies reported that 48-75% of AVFs were being used for HD at the time of follow-up, while 37-51% of them were abandoned. 1,25,26,30,32 A study on pre-emptive AVF creation showed 49% of patients ended up on HD during their 10 month follow-up with 65% of these patients being dialyzed through their AVF. 1 This paper also reported that 23% of the patients never used their viable AVF for HD. 1 A paper which assessed vascular access use since the dialysis outcomes and practice patterns study, found that 27% of Canadian and16% American HD patients were being dialyzed through an AVF. 13 This is in contrast to 72% of German and 69% of Japanese HD patients who utilized an AVF. 13 The proportion of HD patients using an AVG was highest in America at 15%, followed by Sweden at 9%. 13

AVF Outcomes
Four studies reported an average AVF maturation time of 148-285 days with a cumulative functional 1-year patency of 60-70%. 1,21,26,32 Primary failure was recorded in 3 studies and occurred in 20-26% of cases. 25,26,31 In one study, AVFs failed to mature 20% of the time. 30 Primary assisted 1-year patency was only measured in one study, and it was found to be 93-100%. 21 Complications occurred in 21.2% of AVFs and 54% of interventions occurred before maturation was achieved. 31 RCAVF were shown to have a higher primary failure rate but better overall survival than brachiocephalic and brachiobasilic fistulas. 25 When the AVFs were allowed to mature for 10 and 16 weeks, they had a 50% and 75% survival respectively upon cannulation. 25

Pre-Operative Vasculature Status & AVF Outcomes
Dageforde et al. showed that minimum vein diameter is associated with lower risk for AVF failure. 30 Veins < 2.7 mm in diameter had > 33% failure to mature at 6 months. 30 A patent upper arm cephalic vein was shown to improve primary patency, secondary patency and maturation in patients undergoing RCAVF creation. 33 RCAVFs with arterial flow rates < 50 mL/min were shown to have a 7 fold increase in failure rate. 32 The flow rate was also shown to be a more sensitive marker than vein diameter when assessing failure to mature. 32

Multivariate Analysis on AVF Outcome Predictors
Cox regression analysis associated female gender, being on dialysis at the time of AVF creation, and diabetes with worse AVF survival. 25 Dageforde et al. also showed that preoperative dialysis was associated with higher risk of AVF failure. 30 The study by Wilmink et al. also demonstrated that females were associated with higher primary failure and longer maturation times. 25 One study found that an age ≥ 65 years was an independent predictor of secondary AVF patency. 33 Those less likely to have an AVF created were females, as well as patients with a high number of comorbidities. 29 Gender was shown to be unassociated with primary or secondary patency by Schinstock

Discussion
This systematic review intends to combine recent investigations on outcomes of patients who are referred for AVF creation. The outcomes were subdivided into those that pertained to patient prognosis and those that measured the success of the AVF itself. The findings of this review suggest that patients are not being referred at an adequate time for AVF creation based on current KDOQI guidelines as 33.5-77.4% of patients are requiring a CVC use before they are using their AVF for HD. 8 Even once an AVF was made, 33-51% of the patients still ended up on a CVC for HD due to AVF immaturity or failure. 1,25,26,29,30,32 There were vast differences between countries in regards to the uptake of the KDOQI recommendations for AVF implementation. These guidelines suggest that 65% of HD patients should be using a fistula for their HD sessions. 8,29 Only 27% and 16% of HD patients from Canada and USA, respectively, were being dialyzed through an AVF. 13 In contrast, 72% of German and 69% of Japanese HD patients were using an AVF. 13 These stark contrasts between nations may reflect local policies regarding healthcare or surgical preference, as 15% of American HD patients were using AVGs, while the next highest country was Sweden with 9%. 13 The level of access to vascular surgeons may also impact results across regions. This is especially applicable to the Canadian and American healthcare systems which have some of the longest wait times between AVF consultation and AVF creation. 13,34 Encouragingly, the use of AVFs is on the rise in most countries while the use of AVGs is on the decline 13 Between 1996 and 2007, the largest changes occurred in the US where the use of AVFs jumped from 24% to 47%, while AVG usage dropped from 58% to 29% in HD patients. 13 Despite the numerous drawbacks of an AVF procedure, it is still the best option of the available AV access modalities, and the literature supports its utility in terms of patient prognosis over AVGs and CVC. 8,35  RCAVFs with arterial flow less than 50ml/min failed to mature 7 times more often than those with higher flow rates (P<0.001) Radial artery volume flow < 50ml/min is a more sensitive measure for fistula failure to mature than mean vessel diameter of < 2.7mm 69% of the AVFs were functionally patent, 60% of the AVFs achieved primary patency 45% of AVFs failed to mature and were abandoned  24 They reported a primary patency of 60% at 1 year and 51% at 2 years. Secondary patency was 71% and 64% at one and two years, respectively. 24 The pooled primary failure from the same study was 23%. 24 The current study found primary patency to be 60-70% while primary failure ranged from 20-26%. Secondary patency was difficult to estimate due to variations in reporting between the different publications. However, the abandonment rate for the fistulas ranged from 37-51%, which leads to the conclusion that the secondary patency may be lower in this review.
The success of AVFs can be predicted by multiple factors but the most accurate methods reviewed in this study are preoperative arterial flow rate and minimum vein diameter measurements. 30,32 Both of these parameters had high predictive capability compared to factors such as sex, age, morbidity status, lifestyle factor and fistula site. 30,32 Other than arterial flow rate and vein diameter, there was discordance between the publications as to whether other factors correlated with AVF outcomes or not.
All publications included in this review were deemed valid using the EBL critical appraisal tool. Using this method, a threshold of ≥ 75% of the specified criteria was necessary for validity in each individual section and cumulatively, in the individual articles. The lowest overall score was 86%, which was calculated for 3 different studies. 13,21,33 The main area of methodological concern arose from population validity. Three studies were found to have assessed a poorly representative population, as all of them attained a score of 66%. Problematic areas for the population validity were: a small sample size, lack of clearly defined exclusion and inclusion criteria, and no randomization of subjects in comparative studies. For the rest of the validity calculation, all studies were found to be diligently designed other than the pervasive theme of ambiguity regarding whether the investigators played a role in delivering a service to the target population or not.
The articles included in this study were stringently examined using a standardized appraisal tool, reducing bias in calculating the validity of the selected publications. The limitations of this study are that there were only two databases used to identify publications to be included in the review. The database search was only carried out by one investigator which leaves the possibility for selection and reporting bias. This review is limited by an English availability filter used in the database search. The free-full text filter may have removed potentially relevant articles.
Future studies should be directed to the application of preoperative vasculature assessment for prediction of AVF outcomes. Further investigation into the reasons for late referrals to nephrology and vascular surgery in ESRD patients would also be beneficial. Finally, examining the rate of eGFR decline in patients to try and make standardized recommendations for when to refer them for AVF consultation based on their diminishing renal function could provide important and topical data.

Conclusions
The gold standard for vascular access is still the AVF. With the aging global population, there will be an increasing demand for dialysis, which necessitates better standardization regarding patient referral for AVF creation. There are large variations in vascular access use between countries, despite HD patients faring much better when being dialyzed through an AVF as opposed to AVGs or CVCs. A concerted effort is required to try and meet the KDOQI guidelines for timely vascular access creation, improved AVF function and enhanced patient survival. Are data collection methods clearly described?
If a face-to-face survey, were inter-observer and intraobserver bias reduced?