Utilization, patency, and complications associated with vascular access for hemodialysis in the United States

Utilization, patency, and complications associated with vascular access for hemodialysis in the United States Isibor J. Arhuidese, MD, MPH,a,b Babak J. Orandi, MD, PhD, MSc,a,c Besma Nejim, MD, MPH,a and Mahmoud Malas, MD, MHS,a Baltimore, Md; Tampa, Fla; and San Francisco, Calif

ABSTRACT Background: This study examines the utilization and outcomes of vascular access for long-term hemodialysis in the United States and describes the impact of temporizing catheter use on outcomes. We aimed to evaluate the prevalence, patency, and associated patient survival for pre-emptively placed autogenous fistulas and prosthetic grafts; for autogenous fistulas and prosthetic grafts placed after a temporizing catheter; and for hemodialysis catheters that remained in use. Methods: We performed a retrospective study of all patients who initiated hemodialysis in the United States during a 5-year period (2007-2011). The United States Renal Data System-Medicare matched national database was used to compare outcomes after pre-emptive autogenous fistulas, preemptive prosthetic grafts, autogenous fistula after temporizing catheter, prosthetic graft after temporizing catheter, and persistent catheter use. Outcomes were primary patency, primary assisted patency, secondary patency, maturation, catheter-free dialysis, severe access infection, and mortality. Results: There were 73,884 (16%) patients who initiated hemodialysis with autogenous fistula, 16,533 (3%) who initiated hemodialysis with prosthetic grafts, 106,797 (22%) who temporized with hemodialysis catheter prior to autogenous fistula use, 32,890 (7%) who temporized with catheter prior to prosthetic graft use, and 246,822 (52%) patients who remained on the catheter. Maturation rate and median time to maturation were 79% vs 84% and 47 days vs 29 days for pre-emptively placed autogenous fistulas vs prosthetic grafts. Primary patency (adjusted hazard ratio [aHR], 1.26; 95% confidence interval [CI], 1.25-1.28; P < .001) and primary assisted patency (aHR, 1.36; 95% CI, 1.35-1.38; P < .001) were significantly higher for autogenous fistula compared with prosthetic grafts. Secondary patency was higher for autogenous fistulas beyond 2 months (aHR, 1.36; 95% CI, 1.32-1.40; P < .001). Severe infection (aHR, 9.6; 95% CI, 8.86-10.36; P < .001) and mortality (aHR, 1.29; 95% CI, 1.27-1.31; P < .001) were higher for prosthetic grafts compared with autogenous fistulas. Temporizing with a catheter was associated with a 51% increase in mortality (aHR, 1.51; 95% CI, 1.48-1.53; P < .001), 69% decrease in primary patency (aHR, 0.31; 95% CI, 0.31-0.32; P < .001), and 130% increase in severe infection (aHR, 2.3; 95% CI, 2.2-2.5; P < .001) compared to initiation with autogenous fistulas or prosthetic grafts. Mortality was 2.2 times higher for patients who remained on catheters compared to those who initiated hemodialysis with autogenous fistulas (aHR, 2.25; 95% CI, 2.21-2.28; P < .001). Conclusions: Temporizing catheter use was associated with higher mortality, higher infection, and lower patency, thus undermining the highly prevalent approach of electively using catheters as a bridge to permanent access. Autogenous fistulas are associated with longer time to catheter-free dialysis but better patency, lower infection risk, and lower mortality compared with prosthetic grafts in the general population. (J Vasc Surg 2018;68:1166-74.) Keywords: Hemodialysis access; Dialysis access; End-stage renal disease; Chronic kidney disease; Arteriovenous fistula; Arteriovenous graft; Autogenous fistula; Prosthetic graft; Hemodialysis catheter; Dialysis catheter; Permacath

National initiatives, such as Fistula First, and clinical guidelines, such as those promulgated by the Kidney Disease Outcomes Quality Initiative, have proliferated to minimize morbidity and to maximize the duration of access function for patients with end-stage renal disease

From the Division of Vascular Surgery, Johns Hopkins Medical Institutions, Baltimorea; the Division of Vascular Surgery, University of South Florida, Tampab; and the Division of Transplant Surgery, University of California San Francisco, San Francisco.c Author conflict of interest: none. Correspondence: Mahmoud Malas, MD, MHS, FACS, Professor of Surgery, Johns Hopkins School of Medicine, Professor of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Director of the Center for Research Excellence and Surgical Trials (CREST), The Johns Hopkins Hospital, 4940 Eastern Ave, A 547, Baltimore, MD 21401 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2018.01.049

1166

(ESRD).1,2 Accordingly, it has been recommended that autogenous fistulas be considered the preferred initial access for hemodialysis in patients with ESRD, followed by prosthetic grafts and finally by hemodialysis catheters.3 These recommendations stem from numerous studies that have demonstrated an increased risk of sepsis, hospitalization, central venous stenosis, and mortality with catheters as well as increased cost.4-10 The National Kidney Foundation recommends that patients with stage 4 chronic kidney disease undergo evaluation for permanent access placement 6 months before their anticipated dialysis start date.1 This is to allow maturation, a process reported to be as long as 4 months for autogenous fistulas and 4 weeks for prosthetic grafts, and to avoid the need for a catheter as a bridge to maturation. Despite international recommendations and guidelines backed by strong evidence, 80% of incident hemodialysis patients initiate hemodialysis with a catheter rather than with an autogenous fistula or prosthetic graft in the United States.11,12 Indeed,

Journal of Vascular Surgery

Arhuidese et al

1167

Volume 68, Number 4

in many countries, this trend may be increasing. For example, from 1996 to 2007, the proportion of hemodialysis patients using a catheter in Spain tripled.13 In France, it doubled.13 The need to reverse the high prevalence of catheter use, nationally and globally, remains urgent. The prompt creation of autogenous fistulas or prosthetic grafts after prior hemodialysis catheter placement is a rational step to limiting the risk associated with catheters. However, the risks attributable to prior catheter use on outcomes of autogenous fistulas and prosthetic grafts are largely unknown. In the past, there was no reimbursement from the Centers for Medicare and Medicaid Services (CMS) for the creation of autogenous fistulas or prosthetic grafts until at least 30 days after the diagnosis of ESRD. This practice changed through the extensive efforts of the leadership of the Fistula First Initiative, among others.14 Changes in payment policies for hemodialysis-related care that incentivize efficiency, durability, and low cost have also recently been made by the CMS.15 These necessitate objective examination of the performance and durability of hemodialysis access alternatives. In this study, we performed a 5-year audit of hemodialysis access in the United States. Our objective was to compare patterns of utilization of autogenous fistulas and prosthetic grafts or catheters for hemodialysis, access maturation, attainment of catheter-free dialysis, patency, imaging, and mortality in a population-based cohort of patients. We also estimate the risks attributable to prior and persistent use of hemodialysis catheters.

METHODS A retrospective analysis of all patients in the United State Renal Data System (USRDS) who initiated hemodialysis between January 1, 2007, and December 31, 2011, was performed. The USRDS maintains a prospective database of all ESRD patients receiving renal replacement therapy in the United States. Annual reports published since 1988 appear at usrds.org and provide information on epidemiology and mortality among other parameters.16 The USRDS maintains a robust database of every ESRD patient by integrating patient-specific data on hospitalization and costs from the CMS, Centers for Disease Control and Prevention, United Network for Organ Sharing, and ESRD networks. The Johns Hopkins Institutional Review Board approved this study, and the need for consent of the individual patient was waived. The USRDS database contains data on patients’ initial hemodialysis access type and demographic and medical characteristics obtained through CMS Form 2728, End Stage Renal Disease Medical Evidence Report. This form is filled out at the treatment facility by trained medical personnel. Data on arteriovenous access creation, interventions, and complications such as stenosis, thrombosis, and infection necessitating excision in predialysis patients were obtained from a cohort of Medicare beneficiaries

ARTICLE HIGHLIGHTS d

d

d

Type of Research: Retrospective study of data from the United States Renal Data System-Medicare matched national database Take Home Message: In over 470,000 patients who initiated hemodialysis, catheter use as a bridge to permanent access was associated with higher mortality, higher infection, and lower patency. Initiating hemodialysis with autogenous fistula was associated with better patency and survival and lower infection rate. Recommendation: Whenever possible, the authors recommend initiating hemodialysis with autogenous fistula and abandoning the approach of electively temporizing with hemodialysis catheters.

who progressed to ESRD and were subsequently captured in the USRDS database. Follow-up data were obtained from the linked Medicare claims database. The USRDS database also contains data on the death of patients, collected from CMS Form 2746, ESRD Death Notification Form, that is filled out by providers to notify Medicare whenever an ESRD patient dies and matched with the Social Security Death Index. The cohort was divided into patients who initiated hemodialysis with autogenous fistula, prosthetic grafts, or hemodialysis catheters. Patients who initiated hemodialysis with a catheter were further stratified into those who converted from catheter to fistula or graft (converts) and those who persisted on catheters for hemodialysis. The type of arteriovenous access created after initiation with a catheter was identified by Current Procedural Terminology (CPT) codes for autogenous fistulas (36818, 36819, 36820, 36821, and 36825) and prosthetic grafts (36830). Among patients who initiated hemodialysis with a catheter, the total duration of catheter exposure was computed as the interval between catheter placement (CPT code: 36557, 36558) and removal (CPT code: 36589). The time to catheter-free dialysis was computed as the interval between autogenous fistula or prosthetic graft placement and catheter removal in a patient who initiated hemodialysis with a catheter. The ends of these intervals were identified by catheter removal not associated with a new catheter placement within 7 days and preceded by autogenous fistula or prosthetic graft placement, thus incorporating all catheter exchanges within the interval. Primary patency, primary assisted patency, and secondary patency were defined in accordance with published standards.17 The following interventions signified the event times used for computing patencies: angioplasty (CPT code: 35476), stenting (CPT code: 37205, 75960), thrombectomy (CPT code: 36831, 36870), and surgical revision (CPT code: 36832, 36833). The incidence of severe infection necessitating arteriovenous

1168

Arhuidese et al

Journal of Vascular Surgery October 2018

access excision (CPT code: 35903) was also examined. Conduit excision was considered to be hemodialysis access related if the excision was performed within 3 days of a subsequent catheter or graft placement, given the time needed for alternative access after excision of an infected conduit. The access-related imaging studies evaluated were ultrasound-based vascular access and flow studies (CPT code: 90940, 93990) and fistulography (CPT code: 36145, 36147, 75790, 75791). Angioplasties were linked to fistulography, and only those occurring on the same day as fistulography were considered hemodialysis access related. To minimize misclassification bias, patients with more than one autogenous fistula or prosthetic graft created on the same day were excluded (n ¼ 589). Statistical methods. Descriptive analyses of the study groups were performed using c2 and analysis of variance as appropriate. Kaplan-Meier analysis, log-rank tests, and univariable and multivariable logistic and Cox regression analyses were employed to evaluate the outcomes adjusting for baseline characteristics, and inferences were made from the risk-adjusted analyses. Riskadjusted spline analysis was applied to identify significant points of inflection in outcomes over time. Patients who initiated hemodialysis with a catheter and subsequently underwent autogenous fistula or prosthetic graft placement but did not achieve catheter-free dialysis were considered to have failed to mature the first autogenous fistula or graft. Failure was deemed to have occurred on the date a subsequent access (fistula, graft) was placed or by the time the patient was expected to achieve catheter-free dialysis. The expected time to catheter-free dialysis was computed on the basis of the probability of attaining catheter-free dialysis per the patient’s characteristics. The expected time to catheterfree dialysis for patients who failed to achieve catheterfree dialysis was calculated as the median time to catheter-free dialysis for patients within the same centile of probability scores. Eligible patients were censored on the date of death, kidney transplantation, or conversion to peritoneal dialysis or at the end of the study (December 31, 2011). Statistical models were built on predictive variables from univariate analyses, prior literature, guidance of likelihood ratio tests, and Akaike information indices with a goal to achieve model parsimony. All analyses were performed using Stata 14.1 statistical software (StataCorp, College Station, Tex), and a P < .05 was considered statistically significant.

RESULTS This study comprised 476,926 patients who initiated hemodialysis during the study period. Of these, 73,884 (16%) initiated hemodialysis with autogenous fistulas, 16,533 (3%) initiated hemodialysis with prosthetic grafts, and 386,509 (81%) initiated hemodialysis with catheters (Table I). Of those who initiated hemodialysis with a

catheter, 106,797 (28%) patients converted to autogenous fistulas, whereas 32,890 (9%) converted to prosthetic grafts within the study period, leaving 246,822 (63.8%) patients who continued hemodialysis with catheters. The majority of patients who used autogenous fistulas or prosthetic grafts for hemodialysis did so after prior catheter use (61%). Mean follow-up was 21.9 months (standard deviation [SD], 15.3 months; median, 19.1 months; interquartile range [IQR], 9.1-32.6 months) for autogenous fistulas, 20.2 months (SD, 15.2 months; median, 16.8 months; IQR, 7.5-30.5 months) for prosthetic grafts, and 19.3 months (SD, 16.0 months; median, 15.0 months; IQR, 5.9-30.0 months) for patients in the catheter persistent group. There were 39,921 patients in the subcohort of predialysis Medicare beneficiaries who progressed to ESRD, and sensitivity analyses showed no difference in patency for these patients compared with the non-Medicare predialysis patients. Of all patients who received a pre-emptive autogenous fistula, 78.8% used that autogenous fistula for hemodialysis compared to 84.3% for prosthetic grafts (P < .001). The median maturation time was 47 days for autogenous fistulas and 29 days for prosthetic grafts (P < .001), and an average of 0.1 interventions were performed per autogenous fistula and prosthetic graft before use (P ¼ .1). Among patients who initiated hemodialysis with a catheter, the median time from initiation with catheter to permanent access creation was 94 days for autogenous fistulas and 122 days for prosthetic grafts (P < .001). The incidence of catheter-free dialysis was 61.9% for autogenous fistula and 71.4% for prosthetic grafts at 6 months (P < .001). By 1 year, the incidence was 79.2% and 77.2%, respectively (P < .001; Fig 1). The median time to catheter-free dialysis was 4.2 months for autogenous fistulas and 2 months for prosthetic grafts (P < .001). Adjusted Cox analyses confirmed the longer time to access maturation (adjusted hazard ratio [aHR], 1.63; 95% confidence interval [CI], 1.58-1.68; P < .001) and catheterfree dialysis for autogenous fistulas compared with prosthetic grafts (aHR, 1.42; 95% CI, 1.39-1.44; P < .001). The mean number of interventions performed before access maturation was 0.09 (95% CI, 0.09-0.10) for autogenous fistulas and 0.10 (95% CI, 0.10-0.11) for prosthetic grafts (P ¼ .01). The median duration of total catheter exposure was 244 days for autogenous fistula recipients and 201 days for prosthetic graft recipients (P < .001). The unadjusted Kaplan-Meier estimates of primary patency comparing autogenous fistulas to prosthetic grafts were 43.3% vs 31.5% at 1 year, 28.2% vs 14.2% at 3 years, and 20.3% vs 9.3% at 5 years (P < .001; Fig 2). Primary assisted patency comparing autogenous fistulas with prosthetic grafts was 53.3% vs 39.7% at 1 year, 42.0% vs 22.7% at 3 years, and 34.9% vs 16.3% at 5 years (P < .001; Table II; Fig 3). Secondary patency at 1 year, 3 years, and 5 years for autogenous fistulas vs prosthetic grafts was 59.5% vs 56.8%, 50.3% vs 41.7%, and 43.9% vs 33.2%,

Journal of Vascular Surgery

Arhuidese et al

1169

Volume 68, Number 4

Table I. Characteristics of patients who initiated hemodialysis with autogenous fistulas and prosthetic grafts, patients who converted from hemodialysis catheter to autogenous fistulas or prosthetic grafts, and patients who persistently utilized hemodialysis catheters in the United States: 2007-2011

Characteristic

Autogenous Prosthetic Autogenous Prosthetic graft Catheter fistula initiates graft initiates fistula converts converts persistent (n ¼ 73,884 [16%]) (n ¼ 16,533 [3%]) (n ¼ 106,797 [22%]) (n ¼ 32,890 [7%]) (n ¼ 246,822 [52%])

P valuea

64.1 6 15.1

65.5 6 14.9

64.7 6 15.3

66.8 6 14.9

61.6 6 16.7

<.001

35.4

55.4

42.4

57.2

43.1

<.001

White

56.2

43.0

52.8

42.7

51.4

Black

26.3

41.4

28.1

40.9

27.9

Hispanic

11.1

9.9

13.9

11.5

15.2

Other

6.4

5.8

5.2

4.9

Body mass index

29.3 6 13.0

29.3 6 8.0

29.3 6 7.9

29.3 6 8.3

28.8 6 7.9

.1

Diabetes mellitus

52.1

55.7

56.2

57.9

51.2

<.001

Hypertension

88.7

87.4

86.4

86.6

82.7

<.001

Coronary artery disease

21.0

21.0

22.6

22.2

19.4

<.001

Peripheral artery disease

12.5

14.6

14.8

15.5

12.9

<.001

8.4

11.3

9.7

11.6

8.9

<.001

25.3

29.5

36.0

37.2

32.1

<.001

Chronic obstructive pulmonary disease

7.2

7.9

10.3

10.1

9.4

<.001

Cancer

6.7

6.9

7.0

7.2

8.0

<.001

Immobility

2.9

5.8

6.6

9.1

8.6

<.001

Active smoking

5.9

5.7

6.6

5.9

6.4

<.001

Age at initial hemodialysis access creation, years Female sex Raceb

Stroke or transient ischemic attack Congestive heart failure

<.001

5.5

Categorical variables are presented as percentages. Continuous variables are presented as means 6 standard deviation. a P value refers to the statistical test of the difference across the entire group. b Race was self-reported by the patient and documented by staff of the health facility where the diagnosis of end-stage renal disease was made.

respectively (P < .001; Fig 4). The mean number of interventions required to achieve these patency rates were 1.0 (95% CI, 1.01-1.03; median; 0; IQR, 0-1) for autogenous fistulas and 2.4 (95% CI, 2.36-2.43; median, 1; IQR, 0-3) for prosthetic grafts (P < .001). The proportion of patients who required surgical or endovascular interventions to achieve these patency rates was 34.5% for autogenous fistulas and 52.4% for prosthetic grafts (P < .001). In the multivariable Cox regression analyses, primary patency was 26% higher for autogenous fistulas compared with prosthetic grafts (aHR, 1.26; 95% CI, 1.25-1.28; P < .001). Initiation of hemodialysis with a catheter and subsequent conversion to either an autogenous fistula or a prosthetic graft was associated with a 60% decrease in primary patency (aHR, 0.31; 95% CI, 0.31-0.32; P < .001) compared to patients who initiated hemodialysis with either an autogenous fistula or a prosthetic graft. Primary patency was 3.4 times higher for autogenous fistula initiates compared to autogenous fistula converts (aHR, 3.49; 95%CI, 3.44-3.54; P < .001) and 2.4 times higher for prosthetic graft initiates compared with prosthetic graft converts (aHR, 2.43; 95% CI, 2.38-2.49; P < .001). Compared to

patients who initiated hemodialysis with an autogenous fistula, primary patency was 42% lower for prosthetic graft initiates (aHR, 0.58; 95% CI, 0.57-0.60; P < .001); 72% lower for autogenous fistula converts (aHR, 0.28; 95% CI, 0.28-0.29; P < .001); and 75% lower for prosthetic graft converts (aHR, 0.25; 95% CI, 0.24-0.25; P < .001). Primary assisted patency remained higher for autogenous fistulas compared with prosthetic grafts (aHR, 1.36; 95% CI, 1.35-1.38; P < .001). Overall, there was no significant difference in secondary patency for autogenous fistulas compared with prosthetic grafts (aHR, 0.99; 95% CI, 0.98-1.01; P ¼ .83). Considering the duration of conduit use, secondary patency was higher for prosthetic grafts compared with autogenous fistulas within the first 2 months (aHR, 1.18; 95% CI, 1.14-1.25; P < .001). Beyond that time, secondary patency was 36% higher for autogenous fistulas relative to prosthetic grafts (aHR, 1.36; 95% higher CI, 1.32-1.40; P < .001). Overall, an average of 1.4 imaging studies were performed per autogenous fistula per year (SD, 1.1; median, 2.4) compared with 2.5 per prosthetic graft per year (SD, 1.7; median, 3.7; P < .001). Of these, the average use of angiography

1170

Arhuidese et al

Journal of Vascular Surgery October 2018

Fig 1. Time to catheter-free dialysis comparing autogenous fistulas and prosthetic grafts in patients who initiated dialysis with a catheter and thereafter converted to autogenous fistula or prosthetic graft. The incidence of catheter-free dialysis was higher for prosthetic grafts than for autogenous fistulas at 6 months (71.4% vs 61.9%; P < .001). Standard error did not exceed 5% at any point of follow-up.

Fig 2. Unadjusted Kaplan-Meier estimates of primary patency comparing autogenous fistulas and prosthetic grafts. Primary patency was superior in autogenous fistulas compared to prosthetic grafts in the long term (log-rank, P < .001). Standard error <1% at all points of follow-up.

studies was 1.1 per autogenous fistula per year (SD, 0.9; median, 1.9) and 2.2 per prosthetic graft per year (SD, 1.5; median, 3.3; P < .001). The average number of ultrasound studies performed was 0.3 per autogenous fistula per year (SD, 0.2; median, 0.5) and 0.3 per prosthetic graft per year (SD, 0.2; median, 0.4; P ¼ .1). Arteriovenous access infection necessitating excision occurred in 925 (0.5%) fistulas and 2491 (5.0%) grafts (P < .001), equivalent to an incidence rate of 4 per 1000 person-years for autogenous fistulas and 40 per 1000 person-years for prosthetic grafts. Prosthetic grafts were associated with a 9.6-fold increase in severe infection (aHR, 9.6; 95% CI, 8.86-10.36; P < .001) compared with autogenous fistulas. Initiation with a catheter and subsequent conversion to either an autogenous fistula

or a prosthetic graft was associated with a 2.3-fold increase in severe infection (aHR, 2.3; 95% CI, 2.15-2.54; P < .001). Absolute all-cause mortality during the study period was 29.5% for autogenous fistulas, 36.7% for prosthetic grafts, and 42.2% for patients persistently dialyzing through a catheter (P < .001). This corresponds to an incidence death rate of 161 per 1000 person-years for autogenous fistulas, 218 per 1000 person-years for prosthetic grafts, and 262 per 1000 person-years for the catheter persistent group (P < .001). Kaplan-Meier estimates of patient survival comparing autogenous fistulas, prosthetic grafts, and persistent hemodialysis with catheters were 84.0%, 78.2%, and 69.6% at 1 year; 72.3%, 64.4%, and 58.8% at 2 years; 62.0%, 53.6%, and 51.2% at 3 years; 53.5%, 44.9%, and 45.2% at 4 years; and 50.0%, 41.2%, and 42.9% at 5 years (P < .001; Fig 5). Prosthetic grafts were associated with 29% increase in mortality compared to autogenous fistulas (aHR, 1.29; 95% CI, 1.27-1.31; P < .001). Temporizing with a catheter and subsequent conversion to arteriovenous access was associated with a 51% increase in mortality (aHR, 1.51; 95% CI, 1.48-1.53; P < .001) compared to patients who initiated hemodialysis with autogenous fistulas or prosthetic graft. Compared to patients who initiated hemodialysis with an autogenous fistula, the risk-adjusted relative hazard of mortality was 32% higher for patients who initiated hemodialysis with prosthetic graft (aHR, 1.32; 95% CI, 1.28-1.36; P < .001); 45% higher for patients who temporized with a catheter then converted to autogenous fistula (aHR, 1.45; 95% CI, 1.42-1.47; P < .001); 79% higher for patients who temporized with a catheter and converted to prosthetic graft (aHR, 1.79; 95% CI, 1.75-1.83; P < .001); and 125% higher for patients who persistently dialyzed via catheter for the duration of the study (aHR, 2.25; 95% CI, 2.21-2.28;

Journal of Vascular Surgery

Arhuidese et al

1171

Volume 68, Number 4

Table II. Patency comparing autogenous fistulas and prosthetic grafts Patency, % (95% CI) Patency a

Primary

Primary assisteda Secondarya

Access type

1 year

2 years

3 years

4 years

5 years

Autogenous fistula Prosthetic graft

42.7 (42.4-42.9)

34.1 (33.9-34.4)

27.7 (27.4-28.0)

22.8 (22.5-23.1)

19.9 (19.4-20.4)

31.5 (31.1-32.0)

20.5 (20.1-20.9)

14.1 (13.7-14.6)

10.8 (10.4-11.3)

Autogenous fistula

9.3 (8.7-9.9)

53.3 (53.0-53.5)

46.8 (46.6-47.1)

42.0 (41.7-42.2)

37.7 (37.3-38.0)

34.9 (34.3-35.4)

Prosthetic graft

39.7 (39.2-40.2)

29.3 (28.8-29.7)

22.7 (22.2-23.3)

18.9 (18.4-19.5)

16.3 (15.5-17.2)

Autogenous fistula

59.5 (59.2-59.7)

54.2 (54.0-54.5)

50.3 (50.0-50.5)

46.6 (46.2-46.9)

43.9 (43.3-44.4)

Prosthetic graft

56.8 (56.3-57.2)

48.1 (47.6-48.6)

41.7 (41.2-42.3)

37.0 (36.3-37.7)

33.2 (32.2-34.3)

CI, Confidence interval. a P < .001 for the overall comparison of autogenous fistula vs prosthetic graft.

Fig 3. Unadjusted Kaplan-Meier estimates of primary assisted patency comparing autogenous fistulas and prosthetic grafts. Primary assisted patency was superior in autogenous fistulas compared to prosthetic grafts in the long term (log-rank, P < .001). Standard error <1% at all points of follow-up.

Fig 4. Unadjusted Kaplan-Meier estimates of secondary patency comparing autogenous fistulas and prosthetic grafts. Secondary patency was superior in autogenous fistulas compared with prosthetic grafts in the long term (log-rank, P < .001). Standard error <1% at all points of follow-up.

P < .001). There were 2.8 renal transplantations per 100 hemodialysis patient-years (autogenous fistulas, 2.2; prosthetic grafts, 1.3; catheter persistent, 2.8; P < .001), whereas 2.5% of the study population converted to peritoneal dialysis during the study period (autogenous fistulas, 2.3%; prosthetic grafts, 2.2%; catheter persistent, 2.7%; P < .001). There were no occurrences of stenting to remedy a failed or failing hemodialysis access without concomitant angioplasty.

significantly higher for autogenous fistula recipients than for prosthetic graft recipients, with the latter group undergoing more interventions to maintain patency (1.0 vs 2.4 average number of interventions). Access abandonment (loss of secondary patency) was higher for fistulas within the first 2 months, but the trend reversed thereafter. Infections requiring excision of the arteriovenous access were 10 times more likely with prosthetic grafts compared with autogenous fistulas. After risk adjustment, mortality was lowest for fistula initiates followed by prosthetic graft initiates, fistula converts, graft converts, and catheter persistent patients in increasing order. The risk of death was 29% higher for prosthetic graft recipients compared with autogenous fistula recipients (aHR, 1.29; 95% CI, 1.27-1.31; P < .001). Prior catheter use for hemodialysis was associated with a 250% and 140% increase in loss of patency, 130% increase in infection, and 79% and 120% increase in mortality for autogenous fistulas and prosthetic grafts, respectively. Most persons who initiate hemodialysis in the United States do so with a catheter, and the concept of catheters as a bridge to permanent access is not

DISCUSSION Hemodialysis dependence confers a high-risk health status compared with the general population, and the type of access used for hemodialysis influences survival in these patients.16,18 In this national cohort of patients initiating hemodialysis during a 5-year period in the United States, the majority of patients who received arteriovenous access did so after prior catheter use (61%). The population-based maturation rate of preemptively placed autogenous fistulas is high despite a longer maturation time relative to prosthetic grafts. Primary and primary assisted patencies were all

1172

Arhuidese et al

Journal of Vascular Surgery October 2018

Fig 5. Kaplan-Meier estimates of patients’ overall survival comparing autogenous fistula, prosthetic graft, and catheter persistent groups. Standard error is <1% at all time points. AVF, Arteriovenous fistula; AVG, arteriovenous graft.

uncommon.12,19 The results from this study demonstrate that the disadvantage of hemodialysis catheters persists despite conversion to autogenous fistula or prosthetic graft. This novel finding is significant, given the high prevalence of catheters. It undermines the approach of electively utilizing catheters as a bridge to permanent access and supports the national Fistula First recommendations and other campaigns for pre-emptive fistula creation. Multiple factors including but not limited to central vein stenosis and bouts of sepsis likely contribute to the worse outcomes associated with prior catheter use. It can be argued that the disadvantage associated with catheters and prosthetic grafts relative to fistulas is driven by a selection bias in which patients who receive fistulas are in better health than patients who receive prosthetic grafts or catheters. However, the relatively worse outcomes associated with prosthetic grafts and hemodialysis catheters in this study and others persist despite adjustment for patient risk factors, duration of nephrologist care, and general well-being.20,21 The interval between initiation of hemodialysis with a catheter and permanent access placement as well as the time to catheter-free dialysis relative to access maturation time that we have shown represent avoidable exposure to the deleterious impact of hemodialysis catheters. Avoiding catheter use and limiting these prolonged periods of catheter exposure are potent opportunities for outcomes improvement at population levels. It is remarkable that 79% of pre-emptive fistulas placed in patients who progressed from pre-ESRD to ESRD

were used for dialysis. This is in tandem with the proportion of patients who initiated hemodialysis with a catheter, underwent autogenous fistula placement, and achieved catheter-free dialysis in this study. This is also in line with a recent multicenter randomized trial that reported maturation rates ranging from 54% to 87%, depending on anatomic location, flow, and vessel diameter criteria.22 Recent policy changes have facilitated reimbursement for pre-emptively placed fistulas and grafts.14 It is anticipated that these changes will improve arteriovenous access creation in predialysis patients. The arteriovenous access patency rates reported in this large, national study are consistent with those from other studies that have reported overall median patency for autogenous fistulas ranging from 3 to 5 years, whereas that of prosthetic grafts was 1 to 2 years.20,21,23-29 In this study, we report secondary patency rates of 50% at 3 years for autogenous fistulas and 42% for prosthetic grafts. These outcomes also fall within the recommended limits set by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative.1 Given the size and the scope of this study’s data source, the patency rates reported here reflect an aggregate of the “real-world” results of hemodialysis access surgery. Arteriovenous access failure and the number of interventions performed to restore patency were significantly higher for prosthetic grafts than for autogenous fistulas. The higher intervention rate for prosthetic grafts has important implications relating to cost in contemporary practice. The results from this study will serve as a useful

Journal of Vascular Surgery

Arhuidese et al

1173

Volume 68, Number 4

benchmark to evaluate the impact of recent and anticipated health care policy changes on hemodialysis access-related outcomes.15 Our assessment of arteriovenous access-related interventions and complications in predialysis patients was based on a cohort of Medicare beneficiaries who progressed to ESRD. This study does not include nonMedicare predialysis patients or those who received pre-emptive arteriovenous access but did not progress to ESRD. Nonetheless, the cohort of patients who progress to hemodialysis is a closed one, and the proportion of patients who used their pre-emptive autogenous fistulas or prosthetic grafts at hemodialysis is informative. Because of data coding constraints, our report of severe infections excludes those that did not require excision, therefore underestimating the overall incidence of arteriovenous access-related infections. Nevertheless, the incidence of such severe infections leading to loss of access is an important outcome to report. We are unable to differentiate surveillance vs diagnostic access-related imaging and planned second-stage revisions vs those performed for failed or failing accesses. We also acknowledge a potential ascertainment bias with regard to arteriovenous access patency in the predialysis vs postdialysis initiation periods. This study is limited in its retrospective nature, and it does not offer a randomized comparison of autogenous fistulas and prosthetic grafts. As such, there is some inherent bias in the selection of patients because prosthetic grafts are often offered to patients whose veins preclude them from receiving autogenous fistulas. The study is also limited in the granularity of clinical details available in the USRDS. We cannot account for factors that might affect patency, such as conduit quality, medication use, and biologic or synthetic graft subtypes, and the precise cause of access failure or death and the surgeon’s experience and skill. Strengths of this study include its robust sample size, made possible by the magnitude and comprehensive nature of the USRDS, and its longitudinal nature. This populationbased study comprehensively delineates clinically relevant outcomes associated with hemodialysis access, and the results should inform the expectations of patients and their providers in considering options for hemodialysis access.

CONCLUSIONS Temporizing catheter use is associated with higher mortality, higher infection, and lower patency, thus undermining the highly prevalent approach of electively using catheters as a bridge to permanent access. Autogenous fistulas are associated with longer maturation time but better patency, lower infection risk, and lower mortality compared to prosthetic grafts. There is an enduring national need to initiate hemodialysis with pre-emptive fistulas and promptly convert from catheters to fistulas as it becomes anatomically

permissible to mitigate the undesirable impact of incident and persistent catheter use. Improvement in the number and dialysis access-related skills of surgeons is a perennial avenue for better outcomes, given the persistent need for pre-emptive fistula creation. The data reported here have been supplied by the United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as official policy or interpretation of the U.S. government.

AUTHOR CONTRIBUTIONS Conception and design: IA, BO, MM Analysis and interpretation: IA, BO, BN, MM Data collection: Not applicable Writing the article: IA, BO, MM Critical revision of the article: IA, BO, BN, MM Final approval of the article: IA, BO, BN, MM Statistical analysis: IA Obtained funding: Not applicable Overall responsibility: MM

REFERENCES 1. Vascular Access Work Group. Clinical practice guidelines for vascular access. Am J Kidney Dis 2006;48(Suppl 1):S248-73. 2. Tordoir J, Canaud B, Haage P, Konner K, Basci A, Fouque D, et al. EBPG on vascular access. Nephrol Dial Transplant 2007;22(Suppl 2):ii88-117. 3. ESRD NCC. Fistula First Catheter Last. Available at: http:// esrdncc.org/ffcl/for-ffcl-professionals/. Accessed April 21, 2016. 4. Lok CE, Foley R. Vascular access morbidity and mortality: trends of the last decade. Clin J Am Soc Nephrol 2013;8: 1213-9. 5. Dhingra RK, Young EW, Hulbert-Shearon TE, Leavey SF, Port FK. Type of vascular access and mortality in U.S. hemodialysis patients. Kidney Int 2001;60:1443-51. 6. Pastan S, Soucie JM, McClellan WM. Vascular access and increased risk of death among hemodialysis patients. Kidney Int 2002;62:620-6. 7. Moist LM, Trpeski L, Na Y, Lok CE. Increased hemodialysis catheter use in Canada and associated mortality risk: data from the Canadian organ replacement registry 2001-2004. Clin J Am Soc Nephrol 2008;3:1726-32. 8. Ng LJ, Chen F, Pisoni RL, Krishnan M, Mapes D, Keen M, et al. Hospitalization risks related to vascular access type among incident US hemodialysis patients. Nephrol Dial Transplant 2011;26:3659-66. 9. Astor BC, Eustace JA, Powe NR, Klag MJ, Fink NE, Coresh J, et al. Type of vascular access and survival among incident hemodialysis patients: the choices for healthy outcomes in caring for ESRD (CHOICE) study. J Am Soc Nephrol 2005;16: 1449-55. 10. Manns B, Tonelli M, Yilmaz S, Lee H, Laupland K, Klarenbach S, et al. Establishment and maintenance of vascular access in incident hemodialysis patients: a prospective cost analysis. J Am Soc Nephrol 2005;16:201-9. 11. Malas MB, Canner JK, Hicks CW, Arhuidese IJ, Zarkowsky DS, Qazi U, et al. Trends in incident hemodialysis access and mortality. JAMA Surg 2015;150:441-8. 12. Collins AJ, Foley RN, Gilbertson DT, Chen SC. The state of chronic kidney disease, ESRD, and morbidity and mortality

1174

Journal of Vascular Surgery

Arhuidese et al

October 2018

13.

14.

15.

16.

17.

18.

19.

20.

21.

in the first year of dialysis. Clin J Am Soc Nephrol 2009;4(Suppl 1):S5-11. Ethier J, Mendelssohn DC, Elder SJ, Hasegawa T, Akizawa T, Akiba T, et al. Vascular access use and outcomes: an international perspective from the dialysis outcomes and practice patterns study. Nephrol Dial Transplant 2008;23: 3219-26. Solid CA, Carlin C. Timing of arteriovenous fistula placement and Medicare costs during dialysis initiation. Am J Nephrol 2012;35:498-508. Centers for Medicare and Medicaid Services. Comprehensive ESRD care model. Available at: https://innovation.cms.gov/ initiatives/comprehensive-esrd-care/. Accessed April 21, 2016. U.S. Renal Data System (2015). USRDS 2015 annual data report: atlas of chronic kidney disease and end-stage renal disease in the United States. Available at: https://www.usrds. org/2015/view. Accessed September 9, 2016. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M Jr, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002;35: 603-10. Arhuidese IJ, Obeid T, Hicks C, Qazi U, Botchey I, Zarkowsky DS, et al. Vascular access modifies the protective effect of obesity on survival in hemodialysis patients. Surgery 2015;158:1628-34. Allon M, Daugirdas J, Depner TA, Greene T, Ornt D, Schwab SJ. Effect of change in vascular access on patient mortality in hemodialysis patients. Am J Kidney Dis 2006;47: 469-77. Grubbs V, Wasse H, Vittinghoff E, Grimes BA, Johansen KL. Health status as a potential mediator of the association between hemodialysis vascular access and mortality. Nephrol Dial Transplant 2014;29:892-8. Ravani P, Palmer SC, Oliver MJ, Quinn RR, MacRae JM, Tai DJ, et al. Associations between hemodialysis access type and

22.

23.

24.

25.

26.

27.

28.

29.

clinical outcomes: a systematic review. J Am Soc Nephrol 2013;24:465-73. Robbin ML, Greene T, Cheung AK, Allon M, Berceli SA, Kaufman JS, et al. Arteriovenous fistula development in the first 6 weeks after creation. Radiology 2016;279:620-9. Lok CE, Sontrop JM, Tomlinson G, Rajan D, Cattral M, Oreopoulos G, et al. Cumulative patency of contemporary fistulas versus grafts (2000-2010). Clin J Am Soc Nephrol 2013;8:810-8. Huijbregts HJ, Bots ML, Wittens CH, Schrama YC, Moll FL, Blankestijn PJ, et al. Hemodialysis arteriovenous fistula patency revisited: results of a prospective, multicenter initiative. Clin J Am Soc Nephrol 2008;3:714-9. Peterson WJ, Barker J, Allon M. Disparities in fistula maturation persist despite preoperative vascular mapping. Clin J Am Soc Nephrol 2008;3:437-41. Schinstock CA, Albright RC, Williams AW, Dillon JJ, Bergstralh EJ, Jenson BM, et al. Outcomes of arteriovenous fistula creation after the Fistula First Initiative. Clin J Am Soc Nephrol 2011;6:1996-2002. Maya ID, O’Neal JC, Young CJ, Barker-Finkel J, Allon M. Outcomes of brachiocephalic fistulas, transposed brachiobasilic fistulas, and upper arm grafts. Clin J Am Soc Nephrol 2009;4:86-92. Arhuidese I, Reifsnyder T, Islam T, Karim O, Nejim B, Obeid T, et al. Bovine carotid artery biologic graft outperforms expanded polytetrafluoroethylene for hemodialysis access. J Vasc Surg 2017;65:775-82. Hirth RA, Turenne MN, Woods JD, Young EW, Port FK, Pauly MV, et al. Predictors of type of vascular access in hemodialysis patients. JAMA 1996;276:1303-8.

Submitted Jun 12, 2017; accepted Jan 22, 2018.