Clinical Practice Guidelines and Clinical Practice Recommendations
2006 Updates
Hemodialysis Adequacy
Peritoneal Dialysis Adequacy
Vascular Access



8.1 Goals of access placement:

8.2 The primary access failure rates of HD accesses in the following locations and configurations should not be more than the following:

8.3 Access complications and performance:

8.3.4 Cumulative patency rate of TCCs: Not specified. (B)

8.4 Efficacy of corrective intervention:
The rate of certain milestones after correction of thrombosis or stenosis should be as follows:


HD access failure is a major cause of morbidity and mortality for patients on HD therapy.7,8,14-17 Expenditures for reconstituting patency are substantial and increasing.2,8,9,12 Throughout this document, methods and recommendations have been proposed to improve vascular access results. These include:

  1. Establishment of QA programs that track access complication rates and outcomes:
    1. Formation of VATs
  2. Improvement of the skill set of staff:
    1. Physical examination of the accesses
    2. Cannulation techniques among staff
    3. Aseptic techniques
  3. Increasing the percentage of patients with native or primary AVFs by implementing the FFBI. Key portions of the program include the following:
    1. Early identification and referral of patients with progressive kidney disease to nephrologists, allowing access construction well in advance of the need for HD
    2. Protection of veins
    3. Adequate artery and vein evaluation by using DDU and/or angiography
    4. Reevaluation for a native AVF after every access failure
  4. Periodic monitoring of accesses to detect hemodynamically significant stenoses before thrombosis:
    1. Expeditious referral of patients for appropriate angioplasty or surgical revision after the detection and characterization of stenoses
    2. Documentation of functional improvement in access function after corrective intervention
  5. Improved catheter care.

The following Clinical Outcome goals are target suggestions for measuring improvement in performance.


Goals of Access Placement (CPG 8.1)
Data should be updated periodically and methods should be identified to increase the rate of AVF placement. Flow charts should be developed and root-cause analysis should be carried out to identify barriers to fistula placement, causes for excessive thrombosis rates, and reasons for excessive catheter-related infection.

These goals are greater than those previously recommended in the KDOQI Vascular Access Guidelines.20,348 They represent the goals expected by CMS, which has set the target for fistula prevalence of 65% by 2009. Although there has been slow improvement in fistula rates since implementation of the FFBI, rates have increased only slowly (NVAII, www.fistulafirst.org; last accessed 2/20/2006). The Work Group believes that with the reimbursement of DDU procedures and early referral of patients by nephrologists for access evaluation and constructions, rates will improve. In some cases, this will require the use of brachial artery level constructions. An increase in percentage of native AVFs is accomplished best by early determination of the patient's preferred dialysis modality while dialysis therapy initiation is still months away (see CPG 1) because primary AVFs ideally need an extended period of 1 to 6 months to mature. However, those entering the CKD stage 5 program with inadequate or no prior medical care for CKD will continue to blunt the impact of such efforts.

These goals are achievable.37,38,88,569 A primary AVF using the cephalic vein confers the best permanent access with the fewest complications (see CPG 2). Native accesses have the best 4- to 5-year patency rates and require fewer interventions compared with other access types. In many patients, a previous native or synthetic access produces dilation of arm veins, permitting construction of a new primary AV access at a site not previously available.

Catheter usage presents a conundrum. On one hand, catheters provide access that is immediately available; on the other hand, complications are high.180,359 Blood flow frequently is inadequate, thrombolytics frequently are required, and the infection rate is an order of magnitude higher than with grafts or fistulae. Cuffed catheters are associated with lower BFRs compared with grafts and fistulae. As a result, long-term catheter use without appropriate adjustments in treatment duration can compromise dialysis adequacy. Compromise of dialysis adequacy is associated with increased morbidity and mortality. Systemic and local infections occur more frequently with cuffed catheters and account for some of the excess mortality associated with this access type. Finally, long-term catheter access is associated with a risk for central venous stenosis development, which can preclude the establishment of a permanent vascular access for HD (see CPG 2). The initial success, ease of use, and painless access to the patient's blood offered with a dialysis catheter may foster reluctance in some patients to accept more permanent access options with a fistula or graft despite the greater risk for infection and inadequate dialysis associated with long-term permanent catheter use. Patients should be educated on these issues and strongly encouraged to allow creation of a fistula for permanent access whenever possible.

When a catheter must be used either initially or to bridge the patient to the next permanent access, “time-urgency” for initiating/continuing HD therapy with a permanent access does not justify substitution of a graft for a fistula because cuffed catheters are an effective means of bridging the longer time necessary for primary AVF maturation.148,178,184,200 Although catheters can be used for long-term dialysis,187,189 they should be reserved for patients with comorbid conditions limiting life expectancy, those with systolic hypotension in whom attempts to create/maintain a permanent access have met with failure, and those in whom all available sites for fistula or graft (including chest-wall loop grafts) have been exhausted or are not feasible.

The Primary Access Failure Rates (CPG 8.2)
Primary failure is defined as the inability to use the graft at 30 days or obtain sufficient blood flow from the catheter within the first week after insertion. By proposing these goals for 30-day primary failure rates for various graft configurations, the Work Group does not wish to imply that upper-arm grafts should be elected over forearm grafts solely on the basis of these recommended primary failure rates. The Work Group encourages the creation and maintenance of access sites as distally as possible to preserve more proximal veins for future access options. For example, a forearm straight or a brachial loop graft may be used to develop a vein for fistula construction. The Work Group realizes that in some instances, a dialyzer BFR of 300 mL/min might be excessive and produce disequilibrium during the first or second treatment of a very uremic patient. However, by the third treatment, this should not be an issue, and a limit of 1 week is set to determine that the catheter can deliver adequate blood flow.

Primary access failure is considered failure of patency within the first 30 days after placement. Primary failure of dialysis AVGs is caused by technical problems or selection of inappropriate vessels (artery or vein). Neointimal hyperplasia is unlikely to be so virulent as to cause access graft failure within 30 days of construction. It is the Work Group's opinion that the primary failure rate reflects a center effect that is influenced by surgical access construction, patient demographics, adequacy of workup (see CPG 1 and CPG 5), comorbidities, and graft loss caused by premature cannulation and hematoma formation. Primary failure rates of dialysis AVGs at the same anatomic sites vary depending on whether the grafts are the primary, secondary, or tertiary access. The rates provided are derived from the published literature for first graft accesses constructed in a general HD population.24-26,65,67,92,423,570

Failure of a graft before use reflects surgical construction problems. Prosthetic bridge graft survival is decreased in patients with diabetes, even at 30 days, and may be affected adversely by increasing age in patients without diabetes.571 Patient demographics, characteristics, and comorbidities may differ across centers and explain some of the center effect. Each center should monitor its performance, recognizing the influence of some demographic factors, but tracking its own problems in access construction and use (see CPG 8.1.1). Marked deviations from the recommended patency rate should invoke a multidisciplinary evaluation of possible factors and their modification.

A modern properly placed catheter (see CPG 2.4 and CPG 7.1) can deliver more than 300 mL/min at a prepump pressure of –250 mm Hg in adults. A catheter that cannot deliver a flow of 300 mL/min is not being run at a sufficient negative pressure, is improperly positioned, or is dysfunctional for some other reason. Because blood flow with time is a major determinant of adequacy of dialysis, the cause must be determined quickly and corrected. The Work Group believes that catheter blood flow is an indicator of quality in a program. Data for performance should be collected and analyzed to improve quality and protect the patient from underdialysis.

Access Complications and Performance (CPG 8.3)
The Work Group believes that it cannot provide a reasonable estimate of expected cumulative patency of dialysis catheters. The use of cuffed catheters as permanent vascular access is discouraged, except in particular patient groups (see CPG 3).

The current national average rate of thrombosis of dialysis AVGs can only be approximated because there is no mandatory reporting. It is likely to be greater than the overall rate (all permanent accesses) of approximately 0.8 episodes/patient-year at risk10,29 because these rates include the much lower rate of thrombosis of fistulae. In grafts, rates varying from 0.5 to almost 2 episodes/graft-year at risk have been reported in the absence of surveillance or monitoring programs.10,29, 374,572 The rate of graft thrombosis is determined largely by the presence of unrecognized hemodynamically significant stenosis.10,266,572 Six published studies showing the value of surveillance reported baseline thrombosis rates varying from 0.5 to 0.8 episodes/graft-year at risk, which then decreased by 43% to 67% to rates of 0.2 to 0.4 episodes/graft-year.10,322,343,352,373,374 Implementation of surveillance techniques should reduce stenosis and make a rate of 0.5 achievable, even in programs with greater than average rates of thrombosis Therefore, dialysis grafts should be monitored/undergo surveillance to permit early detection of hemodynamically significant stenosis with the goal of reducing the thrombosis rate to a maximum of 0.5 thrombosis/y for AVGs.

PTA is performed to dilate a stenotic lesion within a vascular access or its draining vein. Adequacy of the procedure is measured best by the duration of effect: ie, duration of subsequent patency until either another PTA is required for recurrence of stenosis or thrombosis occurs. A number of observational studies showed that a greater fraction of grafts remained free of interventions or thrombosis if the AVG was patent at the time of intervention (see CPG 6). The fraction of AVGs free of further intervention or thrombosis ranged from 71% to 85% among 4 studies if PTA was performed preemptively compared with only 33% to 63% if PTA was performed after thrombectomy of the graft.373,374 After PTA of stenoses associated with nonthrombosed AVGs, published series consistently reported a 6-month primary (unassisted) patency rate of 40% to 50% (see CPG 6). The duration of effect after thrombectomy and correction of stenosis is considerably shorter. The 4-month criteria are meant to foster 2 processes: (1) preemptive PTA, and (2) assessment of the adequacy of the intervention (PTA or surgery) because inadequate correction typically is manifested by thrombosis or recurrence of the lesion within weeks.

The rate of fistula thrombosis is much less than that of grafts. Fistulae have the lowest rate of thrombosis,57 require the fewest interventions,57,58 and provide longer survival of the access.3,57,58 For native fistulae, access events are only 14% to 33% of those observed in grafts.3,57,58 Therefore, a fistula thrombosis rate that is half that of grafts should result in at least a 1-year longer access survival in well-functioning dialysis programs.

Infectious complications of accesses are a leading cause of morbidity and mortality in dialysis patients. The current national combined infection rates for permanent accesses for local and bacteremic infections are calculated to be 1% to 4% for primary AVFs and 11% to 20% for AVGs during their expected periods of use.4,16,232,543,573-577 Significant variance among dialysis centers is noted.232,543,575,577 Rates of 1% and 10% are the lower end of the published ranges and will demand more attention to aseptic technique (see CPG 3) by some centers.

The catheter infection rate is highly variable532,578,579 and clearly depends on the duration of use.156,211,533 At 2 weeks of catheterization, the incidence of infection of noncuffed central catheters generally is less than 8%.580 One study reported a bacteremia rate of less than 5% in cuffed catheters used less than 3 months and a 50% removal rate for cuffed catheter infection at 12 months of use.156 Other factors include being an incident patient, changing from 1 vascular access to another, and poor patient hygiene.238 The National Nosocomial Infections Surveillance data show that national surveillance of health care–associated infections combined with an intervention prevention program can reduce infection rates, reduce morbidity and mortality, and improve patient safety.581 Establishment of such health care–associated infection surveillance and prevention systems in countries throughout the world should be a priority.

The Work Group's recommendations are significantly less than the experiences of some centers. The Work Group believes infection rates can be decreased significantly through meticulous attention to detail and, in the case of catheters, following the recommendations in CPG 3 and CPG 7. Catheter infection rates can be decreased to less than 1.5 episodes/1,000 days by paying scrupulous attention to the hub,247 a rate that is significantly less (<5%) than the 10% rate proposed. Infection rates also can be decreased by paying attention to skin preparation at the time of placement,582 appropriate use of topical antibiotics,578,583 and use of nonocclusive dressings.584 Programs with high infection rates should consider the importance of nurse and patient training585 (see CPG 3.5).

Complications related to the insertion of TCCs depend on operator skill. Cuffed catheters can be inserted with reference to anatomic landmarks, with or without ultrasound guidance,151,579,586-588 but always with the use of fluoroscopy to verify proper positioning of the catheter tip (see CPG 2.4). Cuffed catheters can be placed by nephrologists, surgeons, or radiologists. Cumulative complication rates less than 5% are obtained routinely without ultrasound guidance.151,579 A recommended complication rate less than 2% is less than values reported in the literature. However, published results are based on procedures obtained without benefit of ultrasound guidance. The RR for complication decreased 5-fold with the use of ultrasound.587 In the Work Group's opinion, rates of 1% should be obtainable in almost all centers and should be the goal.

Double-lumen cuffed catheters are used as both temporary access while a permanent access is maturing and as permanent access in patients who have exhausted other options. This variation in intended use creates significant variation in catheter survival rates. A study reported a median cumulative catheter survival rate of 18.5 months; 65% of silicone dual-lumen catheters survived 1 year.151,587 Conversely, another group reported a 1-year cumulative patency of 30%.579 Another study using 2 single-lumen Silastic catheters (with the majority serving to bridge a period until permanent access was established) reported an average catheter survival of 57 days.152 Others reported a 50% catheter survival rate at 12.7 months156 and median survival period of 289 days.186 Finally, 1 study reported an 80% survival rate at 1 year,589 no doubt in part the result of an all-cause infection rate less than 2 episdes/1,000 days.

Numerous studies reported 1-year patency rates of grafts between 63% and 90%.24,25,67,590 One report described an overall average patency rate of 70%.4 Many investigators reported patency rates at 2 and 3 years, as well.4,24,25,67,73 Outflow obstruction, followed by thrombosis, accounts for the majority of AVG failures. The Work Group believes that prospective surveillance and monitoring (see CPG 4) may improve this reported experience despite the aging of the population and increasing percentage of patients with diabetes or peripheral vascular disease. Thus, cumulative patency targets for grafts of 70% at 1 year, 50% at 2 years, and 50% at 3 years should be achievable. Because fistulae have a lower thrombosis rate, their cumulative survival should be greater. Despite the current problems with maturation and early failure, the Work Group believes that rates comparable to those in Europe can be achieved.3,87,591,592

With respect to grafts, there now has been sufficient time to assess the effect of the efforts made since the previous guidelines, at which the time the Work Group recommended that the primary failure rate of AVFs not be used as an indicator of quality. This was done for fear that during the learning curve of fistulae construction, patients with more complex vascular anatomy (ie, patients at greater risk for failure) might be discouraged. The Work Group recommended that primary failure of native AVFs be examined in dialysis centers as part of their QA/CQI vascular access programs. Since then, many studies documented the superior patency (with lower thrombotic rates) of fistulae compared with grafts.3,37,57,570,593-598 The median patency of 3 years is based on current data and may improve if we can improve cannulation skills.