ESRD, and the maintenance of hemodialysis access, is a tremendous public health problem that has reached almost epidemic proportions in the United States. In 2010, there were 594,374 prevalent (prevalence = patient count at a single point in time) and 116,946 incident (incidence = new patients over a time interval) cases of ESRD.¹ Notably, this represents more than tenfold increase from 1980, and it has been estimated that this trend will continue, with estimated prevalent and incident counts of 784,613 and 150,772, respectively by 2020 (Fig. 73-1).² The majority of patients with ESRD are on hemodialysis (hemodialysis 65%, transplantation 30%, peritoneal dialysis 5%).¹ Approximately 17,000 kidney transplants were performed in 2010, but the number of patients awaiting transplant has continued to increase and has outstripped the number of available organs. Indeed, more than 87,000 patients in the United States were awaiting a kidney transplant in 2010, with the median time on the transplant list of 1.71 years.¹ This gap between the supply and demand of kidney transplants is expected to continue to widen. The majority of patients initiated dialysis in 2010 (i.e., first dialysis session) using a catheter as the sole access (catheter 81%, fistula 16%, graft 3%).¹ However, claims data suggest that there was some reduction in catheter usage within the first 4 months (catheter 53%, fistula 17%, graft 3%, unknown 27%).¹

Morbidity and Mortality

The mortality and morbidity associated with ESRD and maintaining access are significant. The unadjusted 1-year mortality rate for hemodialysis patients in the United States was 22% as reported from the Dialysis Outcomes and Practice Patterns Study (DOPPS) with the incident mortality rate highest within the first 120 days after initiation.^3,4 The majority of the early deaths are related to cardiovascular and infectious causes, with the presence of a central venous catheter and hypoalbuminemia identified as predictors.^1,5 The mean life expectancy for all patients with ESRD in the United States is 5.8 years and, predictably, varies by age, gender, race, and renal replacement therapy.² Notably, the life expectancies for patients 50 to 54 years and 80 to 84 years of age are 6.3 and 2.3 years, respectively.² Patients whose initial ESRD treatment is a transplant are 2.3 times more likely to survive 5 years (survival probability: transplant 0.74, hemodialysis 0.33, peritoneal dialysis 0.33).¹ Patients undergoing dialysis (both hemodialysis and peritoneal) are admitted to the hospital approximately twice per year whereas those with transplants are admitted once.¹ These admissions account for a mean 12.1 hospital inpatient days per year for the dialysis patients and 5.5 days for those with transplants.¹ The corresponding principal admission diagnoses are dialysis-related infection > bacteremia/septicemia > pneumonia > cellulitis.¹ Patients on hemodialysis are more likely to be admitted with each of these diagnoses than those on peritoneal dialysis or with a transplant.¹ The annual event (i.e., removal, replacement with catheter, replacement with internal device) and complication (i.e., infection, sepsis) rates for hemodialysis catheters exceed those for autogenous and prosthetic AV accesses, with reported values of 0.7 and 2.1 episodes/year, respectively.¹ The event (i.e., replacement with prosthetic access, replacement with autogenous access, replacement with catheter, revision, removal) and complication (i.e., infection, sepsis, declot, angioplasty) rates are lowest for patients with autogenous accesses (event rate 0.25 vs. 0.13, complication rate 0.77 vs. 0.44) although the values are comparable to rates for those with prosthetic accesses.¹

Cost

The Medicare costs associated with ESRD are staggering. Although ESRD accounts for only 1.3% of the general Medicare population, it accounted for 7.5% ($25.7 billion) of the general Medicare costs in 2010.¹ Notably, Medicare is the primary payer for patients with ESRD in the United States and is responsible for more than 80% of the patients.¹ The Medicare expenditure per patient with ESRD in 2006 was greatest for patients on hemodialysis (hemodialysis $71,889, peritoneal dialysis $53,327, transplant $24,951).² Among those patients undergoing hemodialysis, the expenditure (patient per year) was greatest for those with catheters (catheters $77,093, grafts $71,616, fistulae $53,470).²

National Kidney Foundation

Foremost among the initiatives and guidelines is the National Kidney Foundation Kidney Disease Outcomes Quality Initiative (KDOQI) Clinical Practice Recommendations for Vascular Access.⁶ The recommendations were originally published in 1997 as the Dialysis Outcome Quality Initiative (DOQI) and then updated in 2000 and 2006 as KDOQI to reflect the expanded emphasis on quality improvement for kidney disease. The KDOQI recommendations were developed through the use of an evidence-based approach encompassing an exhaustive, critical review of the literature by a panel of experts. The stated goals included increasing the number of autogenous AV accesses and detecting access dysfunction prior to thrombosis. The KDOQI recommendations comprised a series of individual guidelines (Box 73-1) that address the various components of access care, including the selection of a specific access configuration, access placement goals, and the management of complications. These guidelines emphasize autogenous AV access, as might be suspected by the overarching goals, with the generic recommendations that autogenous access should be used before prosthetic access and that dialysis catheters should be avoided. The expert panel defined the autogenous radial-cephalic, brachial-cephalic, and brachial-basilic AV accesses as “preferred” and stated that the prosthetic AV access was “acceptable.” The latest version of the KDOQI guidelines established a 65% goal for autogenous access among prevalent dialysis patients with a catheter rate of less than 10% in the absence of a maturing autogenous access. The expert panel stated that the patency rate for autogenous access should be more than 3 years by life-table analysis with thrombosis and infectious rates of less than 0.25 episodes/patient-year at risk and less than 1% over the lifetime of the access, respectively. Similarly, the patency rates for prosthetic accesses should be more than 2 years with a thrombosis rate of less than 0.5 episodes/patient-year and an infection rate of less than 10% over the lifetime of the access.

The National Vascular Access Improvement Initiative, or Fistula First Breakthrough Initiative, was a Centers for Medicare and Medicaid Services (CMS) project, initially spanning 2003 through 2006, that was designed to increase the proportion of hemodialysis patients using autogenous AV access as the primary mode of dialysis.⁷ The stated mission of the Fistula First was to maximize autogenous access in all suitable patients, minimize catheter use, and avoid all types of complications. These laudable goals were implemented through the ESRD networks using a series “change concepts” (Box 73-2). Notably, there were 11 initial “change concepts,” with two others added later. A variety of different tools was used to implement these concepts, including multimedia resources and provider-specific education and feedback. The initial autogenous AV access goals of the Fistula First Initiative were identical to those of the original DOQI (incident 50%, prevalent 4), but were raised to 66% prevalence rate by 2009. A later report suggests that these efforts have been largely successful across the United States as reflected by the trend in autogenous and prosthetic accesses (Fig. 73-2).⁷ In 2012, however, the leadership of the Fistula First suggested that the focus of the initiative should likely be “catheter last,” given the high prevalence of dialysis catheter use, particularly among incident patients.⁸

The Centers for Medicare and Medicaid Services has implemented a Quality Improvement Program or pay-for-performance initiative that places up to 2% of dialysis center payments at risk.⁹ This program was implemented in 2012 for the payment year 2014 and is designed to cut reimbursement for dialysis centers not reaching the defined benchmarks. It is based on six quality improvement measures (three clinical and three reporting), including one for the type of vascular access with the access measure based on the percentages of dialysis patients using an autogenous AV access and those using a tunneled catheter.

Society for Vascular Surgery

The Society for Vascular Surgery also assembled a multidisciplinary group and commissioned a systematic review of the literature to examine the timing of referral for vascular access, the type of access to be placed, and the role of access surveillance. The multidisciplinary group then formulated seven practice guidelines.¹⁰ Similar to KDOQI and Fistula First guidelines, the Society for Vascular Surgery guidelines emphasized early referral to an access surgeon, the use of autogenous access, and the role of access surveillance and monitoring. However, the systematic review that constituted the basis for the recommendations emphasized the limitations in the supporting evidence. The researchers concluded that “low quality evidence from inconsistent studies with limited protection against bias shows that autogenous access for hemodialysis is superior to prosthetic access.”¹¹ Similar statements qualifying the recommendations for early referral and access surveillance were published in the guidelines. Indeed, there has been some resistance or “pushback” with regard to several of the KDOQI guidelines, as reflected by the evolution from “fistula first” to “catheter last” and the focus on a “functional access” rather than an autogenous access. Hiremath et al¹² performed a decision analysis examining the strategy of early referral as recommended by the guidelines and concluded that a “wait strategy” may be associated with greater quality-adjusted life expectancy.

General Considerations

The requirements for an ideal AV access include a flow rate sufficient for effective dialysis, excellent long-term patency, minimal access-related complications, and a cosmetic appearance acceptable to patients. Unfortunately, no access type fulfills all of these requirements. A mature autogenous AV access satisfies most of these criteria and is likely the optimal access choice, as emphasized by the national guidelines and initiatives previously detailed. Indeed, most developed countries outside the United States have already achieved the KDOQI targets.¹³ However, it is important to emphasize that the ultimate goal is a functional access with minimal complications, not necessarily an autogenous access.

Patient-Specific Considerations

The goal for the individual patient is to establish the most functional access with good long-term patency and minimal complications. As outlined previously, this is an autogenous AV access for most patients. However, a variety of concerns should be addressed in this complex decision process about the most appropriate access for a specific patient, including age, gender, life expectancy, anatomy, comorbidities, likelihood of success (i.e., autogenous maturation rate, long-term prosthetic access patency), urgency of dialysis, and patient preference. In most cases, the question about the most appropriate access can be distilled down to whether the patient is suitable for an autogenous AV access and whether the access is likely to be successful. The important distinction between the emphasis on a functional access and an autogenous access is illustrated by the prospective, longitudinal study by Solesky et al,²² who tracked outcome after access creation. They were able to achieve the national target rate for autogenous access of 66%, but a significant portion of the dialysis support over the study period was provided by catheters, with a 27% catheter prevalence rate. This experience led them to conclude that the current algorithms are problematic and rely heavily on catheters.

Decision Making

The critical question remains as to how individual predictors can or should be incorporated into the selection of the most appropriate access choice for an individual patient. Such a choice incorporates some potentially modifiable (e.g., smoking, specific choice of artery/vein, obesity) and nonmodifiable (e.g., age, gender, presence of peripheral vascular disease) factors. Lok et al⁵⁷ have developed a scoring system to predict autogenous access maturation. They identified age greater than 65 years (2 pts), peripheral vascular disease (3 pts), coronary artery disease (2.5 pts), and white race (−3 pts) as significant predictors. In their model, patients were given 3 points as a baseline and then the individual scores for the various comorbidities were summed. Successful maturation was found to correlate with the point total, in that patients in the highest-scoring group, with a score higher than 9, had a failure rate of 69% (Fig. 73-4). Notably, a 70-year-old African-American woman (age 2 pts, race 0, baseline 3 pts) with an above-knee amputation (peripheral veno-occlusive disease 3 pts) and coronary artery disease (CAD 2.5 pts) would have a score of 10.5, corresponding to a 31% successful maturation rate. Interestingly, Lilly et al⁵⁸ used data from the Centers for Medicare and Medicaid Services to examine the Lok model and concluded that the model was not all that predictive, emphasizing that many “high-risk” patients still had successful autogenous accesses. However, their validation study was limited by the administrative database, the lack of the access history, and the inability to validate the data.

Lok et al⁵⁷ have incorporated their scoring system into an initiative entitled “achieve appropriate access” (personal communication, 2012). Their comprehensive approach includes selecting the most appropriate patient on the basis of their scoring system, comorbidities, prior access history, and stage of chronic kidney disease. It involves selecting the most appropriate access for a patient with the least amount of complications by selecting the most appropriate vessels. Furthermore, it involves diligent follow-up with appropriate interventions, all within the context of a multidisciplinary team offering the full range of access options and encompassing experienced surgeons.

The role of prosthetic accesses in this patient-centric decision algorithm about the optimal access choice was further defined by Cull (Table 73-1)⁵⁹ and reflects many of the issues already identified. In his thoughtful discussion, this writer emphasizes the balance between urgency of dialysis and the likelihood of successful autogenous access maturation as part of the clinical algorithm for selecting an autogenous or prosthetic access. Lacson et al⁶⁰ have developed an alternative approach, focusing on a “catheter last” algorithm that incorporates expanded use of peritoneal dialysis, use of early cannulation prosthetic accesses, judicious use of autogenous accesses, careful use of existing accesses, and aggressive follow-up with monitoring and surveillance.