In 2011, USRDS reported that 51 % of hemodialysis patients were dialyzing with a catheter at day 91 of treatment. According to US DOPPS, 19–38 % of patients were dialyzing with a CVC in 2013 [29]. FFBI has set a goal to decrease catheter use to <10 % for patients on HD longer than 90 days. In fact, in recent years, the FFBI has transitioned to the Fistula First Catheter Last (FFCL) Workgroup Coalition “to focus on the development of tools and resources to help dialysis facilities and clinicians reduce catheters and increase AV fistula rates in hemodialysis patients” [19]. Catheter use is associated with significant morbidity, mortality, and cost. A major complication of catheter use is catheter-related bacteremia and the attendant risks of hematogenous spread causing complications such as endocarditis, septic emboli, and osteomyelitis. The cumulative risk of an episode of catheter-related bacteremia is close to 50 % in the first 6 months of use, and each hospitalization for catheter-related bacteremia costs around $23,000 [30, 31]. One study reports a threefold increased mortality in patients dialyzing through catheters compared to AVFs [7]. In one large cohort of almost 80,000 patients, changing from a catheter to a fistula or graft significantly improved patient survival, with a 30 % decrease in risk of death in prevalent hemodialysis patients [24]. With respect to impact on future vascular access, Rayner et al. found prior catheter use was associated with a significantly increased risk of fistula failure [32].

Many factors contribute to the increased use of catheters in incident hemodialysis patients [33]. While many point to delayed nephrology referral, as shown above, even among patients followed by a nephrologist for a year, 60 % initiate hemodialysis with a catheter in place. Some posit that attempting fistula placement in the vast majority of patients has the potential to increase catheter use, compromise vasculature for future vascular accesses, and necessitate more interventions for salvaging the existing access and creating a new vascular access [26, 34–36].

While a functioning fistula is the gold standard of vascular access and is associated with the best outcomes, AVF may not be the optimal choice for all patients [37]. For instance, AV fistulas may not be the best choice for patients who are older and have multiple comorbidities, shorter life expectancy, or unsuitable vessels. In such cases, AV grafts may be a more appropriate HD access and may translate into less catheter use [38]. In the 2006 guidelines for vascular access, the KDOQI Work Group recognized that the “fistula first at all costs” approach may not be the optimal approach for all patients [10]. Many now agree that a universal policy of fistula first may not be appropriate for all incident patients and, instead, providers should take a patient-centered approach in determining the optimal vascular access. Factors affecting the reduced number of working fistulas at dialysis start and contributing to increased catheter time, as discussed below, include (1) inadequate timing of vascular access placement, (2) fistula nonmaturation, (3) inadequate fistula surveillance postoperatively, and (4) inadequate reimbursement for vascular access procedures.

Establishing a functional AV fistula takes time. There are a number of steps involved in vascular access placement: referral to surgery, surgical evaluation, scheduling the surgery, time for maturation, and the possibility of a need for a salvage procedure to achieve usability [39].

Even among those patients followed by a nephrologist, the above process is often not initiated with sufficient time to ensure patients initiate hemodialysis with a mature fistula. KDOQI encourages educating patients with a glomerular filtration rate (GFR) less than 30 ml/min/1.73 m² on all modalities of kidney replacement therapy, so that timely referral can be made and a permanent dialysis access placed, when indicated. Both KDOQI and the Society for Vascular Surgery (SVS) recommend that an AVF should be placed at least 6 months in advance of the anticipated need to start hemodialysis [10, 12]. This timing allows for adequate maturation, as well as potential revisions or placement of a new vascular access when an access fails to mature.

A complicating factor in timely vascular access creation is the difficulty in accurately predicting the rate of progression of kidney failure, especially in cases of acute-on-chronic kidney injury where patients need to initiate dialysis urgently [33]. Further, many patients resist permanent access placement, hoping their kidney function will stabilize with improved blood pressure and glycemic management [33].

Regarding surgical planning, KDOQI recommends duplex ultrasound of the upper extremity arteries and veins. Routine preoperative vessel mapping has not consistently translated into improved fistula maturation rates. Preoperative mapping is associated with an increase in fistula placement in several observational studies, but is not necessarily associated with improved maturation [40]. Patel et al. reported increased fistula creation from 61 to 73 % but decreased maturation rate from 73 to 57 % after implementing preoperative vascular ultrasounds [41]. In another study, radiocephalic fistulas constructed with veins less than 2.0 mm had a primary patency of 16 % at 3 months compared with 76 % with veins greater than 2.0 mm [42]. Wong et al. reported that when the radial artery or cephalic vein diameter was <1.6 mm, fistulas did not mature [43]. Peterson et al. found that older age, female gender, and forearm location were associated with a significantly higher risk of primary fistula failure despite adequate preoperative vessel size [44]. Most studies support a minimum vein diameter of 2.5 mm and artery diameter of 2 mm for successful fistula creation.

There are no randomized controlled trials comparing anatomic order with respect to access construction. Both SVS and KDOQI recommend that the first access should be placed as far distally as possible to preserve proximal sites for future accesses. Per KDOQI, “good surgical practice makes it obvious that when planning permanent access placement, one should always consider the most distal site possible” [10]. In patients with small vessels, some advocate for the placement of a forearm AV graft to mature upper arm veins, which both enables a future successful upper arm AVF and provides a functioning access without the need for catheter use.

In 1966, Cimino and Brescia reported that 13 of their 16 patients were dialyzed successfully using their radiocephalic fistulas. These accesses were cannulated with 14-gauge needles on postoperative day 1 and used for as long as 15.5 months in some patients. In contrast to hemodialysis patients today, these were young nondiabetic patients with an average age of 43 years, and all but one had chronic glomerulonephritis [3]. In contrast to fistulas placed decades ago where primary failure ranged from 10 to 24 % [40, 45–49], fistulas placed today have reported primary failure rates ranging from 30 to 60 % [41, 44, 50–52], and primary patency rates are lower at 40–70 % [53]. These high failure rates and low primary patency rates have largely been attributed to changing patient demographics and comorbidities [36, 41, 44, 51].

Factors associated with failure to mature include diabetes mellitus, peripheral vascular disease, congestive heart failure, advanced age, and female gender [54, 55]. The highest failure rates are reported in older and female patients, Hispanics and African-Americans, and patients with cardiovascular disease and forearm fistulas [56–59].

The NIH-funded Dialysis Access Consortium (DAC) clopidogrel study is the largest randomized controlled trial evaluating fistula outcomes. It examined the effects of clopidogrel on AVF thrombosis and suitability for dialysis use in newly created fistulas. Dialysis suitability was defined as the ability of the AVF to support a dialysis treatment with two needles at a blood flow rate of ≥300 ml/min or greater than eight hemodialysis sessions during a 30-day period [60]. While clopidogrel significantly improved primary patency with reduction in early fistula thrombosis by 37 % (mainly in forearm fistulas), it did not improve fistula maturation. High nonmaturation rates were observed in both groups, despite 75 % of patients undergoing preoperative vascular mapping: 61.8 % in the clopidogrel group and 59.5 % in the placebo group. The authors comment that “our finding of a beneficial effect of clopidogrel on fistula patency but not on suitability is important to the evolving understanding of the pathophysiology of fistula maturation…and suggests that early patency is necessary but not sufficient for fistula maturation” [60].

In a study of the natural history of AVFs, only 11 % of AVFs matured without the need for intervention, while 36 % of fistulas required at least one intervention [61]. Similarly, other studies report that approximately one-third of all AVFs require an intervention to facilitate maturation [62–64]. In the DAC clopidogrel study, only a small percentage of fistulas underwent angioplasty or surgical revision to aid in maturation. Authors query whether more procedures to promote maturation would have translated to a beneficial effect of clopidogrel on fistula suitability.

The mechanisms underlying AVF maturation are complicated and remain poorly understood. Three main biologic reasons for nonmaturation are failure of arterial dilation, failure of venous dilation, and accelerated neointimal hyperplasia [65]. Given our limited understanding of the complex vascular remodeling involved in fistula creation, the National Institute of Diabetes and Digestive and Kidney (NIDDK) Diseases is sponsoring a multicenter prospective cohort study evaluating patients undergoing fistula surgery. This Hemodialysis Fistula Maturation (HFM) study is ongoing and will study patients pre-, intra-, and postoperatively to assess vascular anatomy and biology, clinical attributes, and processes of care with the goal of identifying modifiable predictors of fistula maturation [66].

A mature, functional fistula is defined as one that has adequate blood flow to support dialysis and is large enough for successful repetitive cannulation [10, 65]. Studies show that an increase in blood flow and vein diameter occurs soon after fistula creation. In one study, blood flow increased from 20.9 ± 1.1 ml/min in the radial artery to 174 ± 13.2 ml/min in the AVF 10 min after the anastomosis was created [67]. Other investigators documented blood flows of 539 ± 276 ml/min on postoperative day 1 and 848 ± 565 ml/min 1 week following fistula creation [68]. Robbin et al. found increases of venous diameter to >4 mm and blood flow >500 ml/min at 4 weeks that did not significantly change in subsequent months. Vein diameter ≥4 mm and access blood flow ≥500 ml/min were associated with a 95 % likelihood that the access will be usable for dialysis [69]. On balance, adequate blood flow and access diameter are achieved within 4–8 weeks of creation [68–70], and fistulas not achieving these benchmarks are unlikely to mature to support dialysis [43, 69].

The KDOQI Work Group suggests the so-called rule of 6 s to describe characteristics of a mature or functional fistula: the access has a blood flow greater than 600 ml/min, a diameter greater than 0.6 cm, and a depth of approximately 0.6 cm from the skin surface [10]. Both physical examination and ultrasonography are useful tools for assessing fistula maturation and early AVF failure. In one study where experienced nurses examined fistulas for maturation, their overall accuracy of prediction was 80 % [69]. The two most common causes of fistula nonmaturation – juxta-anastomotic stenosis and the presence of accessory veins – can be identified on physical examination [71]. Both the SVS and KDOQI Work Group guidelines recommend further investigation to identify “potentially remediable anatomic lesions” if a fistula is not maturing adequately at 6 weeks [10, 12]. Some estimate that at least 80 % of nonmaturing AV fistulas can be salvaged after intervention on an underlying lesion [71, 72].

Regarding timing of cannulation, Rayner et al. showed that the median time to first AVF cannulation differed among countries, ranging from <28 days in Japan and Italy to as long as 98 days in the USA. Fistulas used within the first 2 weeks after fistula creation were two times more likely to fail than those cannulated after 14 days [32]. There was no increased risk of failure in fistulas accessed after 14 days, and failure was not significantly different among any of the cannulation interval groups greater than 14 days. This prolonged time to fistula use in the USA increases the likelihood of catheter dependence, as patients often need to initiate dialysis before fistulas are ready for cannulation.

A major barrier to increasing fistula use among incident hemodialysis patients is inadequate coverage for predialysis vascular access placement. In 2010, a clinical technical expert panel, convened by CMS to make recommendations about ways to decrease vascular access-related infections, posited that improving reimbursement for vascular access would ultimately reduce the prevalence of catheters and catheter-related costs [73, 74]. It identified a number of financial and regulatory barriers to timely AVF placement and recommended changes to Medicare reimbursement for vascular access placement, including (1) earlier disbursement of Medicare benefits for vascular access procedures for the uninsured, (2) full payment when fistulas and catheters are placed on the same day in hospitalized patients, and (3) payment for access surgery when patients are hospitalized to initiate hemodialysis. The panel argued that changing reimbursement for vascular access will, not only, motivate providers to place more timely vascular access, but will improve patient outcomes and reduce the high costs associated with dialyzing with a central venous catheter. Potential annual cost savings are estimated at close to a $1 billion [73].

Lok et al. reported that AV grafts were more likely to be placed in high-risk patients, yet cumulative survival was similar to those lower-risk patients who received AVFs [36]. Patients with grafts were more likely to be female, diabetic, and black. Comparing cumulative patency between fistulas and grafts, they found the primary failure rate for AVFs was 40 %, two times greater than for grafts. Fistulas demonstrated better cumulative patency than grafts, but when primary failures were included in the access survival analysis, cumulative survival was similar between both forms of vascular access. Lee et al. reported similar patency findings between AV grafts and AV fistulas when primary AVF failures were excluded and actually observed superior graft compared to fistula survival within the first 18 months of access creation [75]. Based on these findings, judicious use of AV grafts may afford similar cumulative patency compared with AVFs while reducing exposure to the risks associated with catheter use.

An advantage of graft placement is grafts can often be cannulated within 2 weeks of creation. In fact, KDOQI recommends not placing an AVG earlier than 3–6 weeks before initiation of hemodialysis because of the high risk of venous outflow stenosis which can occur anytime after placement [10]. The downside to grafts is, once in use, they require twice as many interventions to maintain patency [36]. Maintaining long-term graft patency requires 2.4- to 7.1-fold higher frequency of salvage procedures, including angioplasty, thrombectomy, and surgical revision [40].

While fistulas have superior cumulative patency to grafts and require fewer interventions to maintain patency, they are associated with higher primary failure rate, more interventions to achieve maturation, and longer catheter dependence [75]. As such, many argue they may not be the optimal vascular access for all hemodialysis patients, especially elderly patients. Patients with lower likelihood of fistula maturation may benefit from having an AVG placed upfront [58].

Of the 382,029 prevalent hemodialysis patients in 2012, approximately 80 % were over age 50 and about a third were ≥70 years. Over the last decade, the prevalence of patients on hemodialysis has increased 31 % among patients between the ages of 65 and 74 years and 48 % in those ≥75 years [21]. In one study looking at outcomes in octogenarians, 89 % initiated hemodialysis with a tunneled catheter, and 56 % of patients died within 180 days of dialysis start. Among the patients who died, 70 % had a fistula placed that was never used [76]. De Silva et al. found similar survival outcomes in octogenarians and nonagenarians whether an AVF or AVG was placed predialysis. Further, among the octogenarians, patients were 77 % more likely to initiate dialysis with a catheter if an AVF was in place [77]. Lok et al. found that patients >65 years have a two times greater fistula nonmaturation rate compared with younger patients [58]. Given such findings, Tamura et al. proposed a conceptual framework to guide decision-making regarding the choice of vascular access in older patients with ESRD that takes into account life expectancy, the benefits and harms of competing strategies, and patients’ preferences [78].

Vascular access is the lifeline for patients requiring hemodialysis. Delayed vascular access placement is associated with significant patient morbidity and mortality and an increased number of inpatient hospitalizations [26]. It takes time and coordination to achieve a permanent vascular access. Encouraging timely placement of an arteriovenous fistula remains the goal for suitable patients and necessitates coordination of care among many providers: primary care providers, nephrologists, vascular surgeons, and dialysis staff. Many posit that a dedicated vascular access program with an appointed vascular access coordinator is critical for ensuring an integrated, multidisciplinary approach to vascular access care [79]. Successful vascular access creation and maintenance depends on timely referral to vascular surgery, close monitoring and surveillance postoperatively (especially in the first 6 weeks), early intervention for nonmaturation (when indicated), and expert cannulation to ensure access preservation.