Share this article on:

Outcomes of Vascular Access Creation Prior to Dialysis: Building the Case for Early Referral

Weber, Catherine L.*; Djurdjev, Ognjenka†; Levin, Adeera*†; Kiaii, Mercedeh*

doi: 10.1097/MAT.0b013e31819f635c
Kidney Support

A 2-year single institution experience of the successes and complications of arteriovenous fistula (AVF) creation before dialysis initiation is reported. Study cohort: all patients who underwent AVF creation before need for dialysis (AVF group, n = 125). “Control” group: all patients with a sustained glomerular filtration rate (GFR) ≤25 ml/min and eligible for AVF creation during the same time period, but did not have an access placed before dialysis (no AVF group, n = 198). Median nephrology contact time was >12 months in both groups and mean eGFR at creation (AVF group) was 12 ml/min. In the AVF group, 72% underwent a successful first AVF creation, 11% suffered AVF thrombosis, and 17% had a nonmaturing AVF before need for dialysis. Sixty-six percent (n = 23) of these latter patients underwent a second AVF creation and 48% were mature at dialysis initiation. During the study period, 70% (n = 88, AVF group) and 61% (n = 121, no AVF group) commenced dialysis. Impressively, 72% (n = 66) of the AVF group used a mature fistula as their first dialysis access. A snapshot of “access in use” at the 6-month mark of dialysis revealed the AVF group had a higher number of patients using an AVF (81% vs. 44%) and a lower number with catheters (19% vs. 56%, respectively, p = 0.001). In conclusion, the success rate of early AVF creation is reasonable and complications when identified can be remedied without the need for a catheter, thus ultimately maximizing the use of AVF in dialysis patients.

From the *Division of Nephrology, Department of Medicine, University British Columbia; and †British Columbia Provincial Renal Agency, Vancouver, British Columbia, Canada.

Submitted for consideration July 2008; accepted for publication in revised form October 2008.

Reprint Requests: Dr. Catherine Weber, St. Paul’s Hospital, 1081 Burrard Street, Providence Building, 6th Floor, Room 6010A, Vancouver, British Columbia, Canada, V6Z 1Y6. Email:

An arteriovenous fistula (AVF) is the preferred vascular access for hemodialysis patients. Compared with synthetic grafts or central venous catheters (CVC), an AVF is associated with a longer lifespan, lower hospitalization and infection rates, and lower mortality.1–4 Current Kidney Foundation Dialysis Outcome and Quality Initiative (K-DOQI) guidelines suggest that an AVF should be placed at least within 6 months of anticipated need.5 Despite these guidelines, minimum rates of prevalent AVF use in Canadian and US hemodialysis units [>60% (Canadian Society of Nephrology),6 >65% (KDOQI)5] are still not being met.7,8 This is paralleled by an alarming increase in catheter use in many centers.8

Late referral to a nephrologist has been cited as a barrier to timely AVF creation.9 Although DOPPS II demonstrates that approximately 80% of patients in Canadian centers are followed by a nephrologist for >4 months before dialysis start, late referral to nephrology is still a major issue in many centers.7 It may be that there are delays related to both patient and physician factors. Although guidelines cite creating an AVF within 6 months of anticipated need, concomitant co-morbidity and stability of renal function at low levels lead to uncertainty as to ideal timing of AVF creation. It is possible that fear of primary AVF failure in this group may also influence the decision to refer.10 Additionally, lack of timely access to a surgeon is an issue.7 Finally, the necessity to undergo further diagnostic investigations and additional endovascular or surgical procedures of AVF complications may delay AVF creation.

Ideally, early AVF placement would allow adequate time for AVF maturation and for complications to be identified and remedied before dialysis to avoid the need for a CVC and its attendant complications. Furthermore, with a focus on considering a fistula first (Fistula First Initiative), incident and prevalent rates of AVF use can be maximized. The purpose of this analysis is to describe the “real life” outcomes of early AVF creation and specifically, both the successes and complications (failure to mature, thrombosis, ischemia, infection, steal syndrome) in a cohort of chronic kidney disease (CKD) patients undergoing AVF creation before the need for dialysis. This data may provide helpful information to clinicians and patients regarding assessment of risks and benefits of early AVF creation.

Back to Top | Article Outline

Subjects and Methods

A study cohort was established using all patients followed by Nephrologists at a single large tertiary care institution in Vancouver, British Columbia, Canada between the years 2003-2005 inclusive who underwent access creation before the need for dialysis (AVF group). The subjects in the AVF group are the main focus of this article, and are the cohort from which the primary outcome measure is derived. A second cohort or “control group” was established from the same patient population who did not undergo AVF creation prior to the need for dialysis (no AVF group). To be included in this “control group,” a patient had to have a sustained glomerular filtration rate (GFR) ≤25 ml/min (CKD stages 4 and 5) during the same time period, and thus be “eligible” to undergo AVF creation. Data is collected prospectively in all these patients, and was obtained from the Provincial Renal Agency database (PROMIS) and supplemental chart review; each category was >99% complete. All patients in the PROMIS database sign a consent form allowing access to their information for statistical and quality improvement purposes. This study and analysis is part of Continuing Quality Improvement of the Vascular Access Program at St. Paul’s Hospital, ethics approval was therefore not required.

Demographic data collected included age, sex, ethnicity, etiology of chronic kidney disease, presence of diabetes mellitus (Type I or II) and a history of cardiac and/or vascular disease defined as angina, myocardial infarction, percutaneous coronary intervention or bypass surgery, congestive heart failure, peripheral vascular disease, or stroke. Data pertaining to the placement of vascular access included referral date for surgery, date of surgery, location of access placement [radiocephalic (RC), brachiocephalic (BC), transposed brachiobasilic (BB) AV fistula or synthetic graft (PTFE)], number created per patient, and eGFR at time of incident AVF creation.

Patients were followed until dialysis initiation or the end of the study (April 2006). In the AVF group, those who initiated dialysis during the study period were followed for an additional 6 weeks to capture any dysfunctional AVF discovered on needling in the hemodialysis unit. There were three additional nonmaturing (NM) AVF discovered during this 6 week time period, and the results of these patients are included with the 21 patients with a NM AVF discovered before dialysis initiation (see Table 2).

Our primary outcome measure is the occurrence of the following vascular access complications in the AVF group, a) failure to mature (AVF had not matured in 3 months by clinical examination), b) thrombosis of the access (clinical absence of a thrill or bruit ± fistulogram confirmation), c) ischemic monomelic neuropathy (clinical diagnosis ± nerve conduction studies), d) infection (clinical diagnosis and documented prescription of antibiotics), and e) steal syndrome (clinical diagnosis ± angiogram). Our secondary outcome measure is type of access in use (AVF, CVC) at 6 months in those who initiated dialysis during the course of the study (in both the AVF group and no AVF group).

Descriptive statistics are presented as mean with standard deviation or median with interquartile range, depending on the underlying distribution. The cumulative probability of experiencing an AVF complication as a function of time since creation was estimated using the Kaplan-Meier method. Patients who started dialysis without AVF complications were treated as censored observations 6 weeks post dialysis start. Patients who ended follow-up for other reasons, or had no complications or dialysis start, were censored at study exit or on April 01, 2006.

Continuous and categorical variables were compared across AVF complication types using ANOVA or Kruskal-Wallis and χ2 tests, respectively. The logistic regression model was used to identify patient characteristics associated with AVF complications. The following variables were entered into model: age, gender, race, diabetes, history of CV disease, duration of CKD follow-up, location of CKD follow-up, eGFR level at AVF creation, and AVF location. We used backward selection to eliminate variables that were not statistically significant (p-value <0.05).

Back to Top | Article Outline


Patient Characteristics

During the years 2003-2005, a total of 125 patients followed by a group of Nephrologists at a single large tertiary care institution were referred to the Vascular Access Clinic and had a native fistula (AVF) created or synthetic graft (PTFE) placed before need for dialysis (AVF group). One hundred ninety-eight patients with a sustained GFR ≤25 ml/min and “eligible” for access creation did not undergo AVF placement before need for dialysis for indeterminate reasons (no AVF group). Of the access’ created in the AVF group, 122 patients (98%) received an AVF and 3 (2%) a PTFE graft. As the number of grafts was small they were considered as part of the AVF group. Table 1 describes the demographics of the cohorts. Note that the AVF group was younger (66 ± 15 years vs. 69 ± 13 years, p = 0.04) and had a lower burden of cardiac and/or vascular disease (42% vs. 64%, p < 0.01) than those who did not receive an AVF before dialysis initiation. Median Nephrology contact time was statistically different between the two groups (18 vs. 14 months, p < 0.01), however, both had greater than 1 year of follow-up. Fifty eight percent of the AVF group were male, 54% Caucasian, and 42% Asian. The most common etiologies of CKD were diabetic nephropathy (26%) and hypertensive nephrosclerosis (29%). A significant proportion of patients were diabetic (44%). Mean eGFR at time of creation was 12 ml/min.

Back to Top | Article Outline

Arteriovenous Fistula Location

Of incident AVF created, 46% (n = 57) were forearm fistulae (RC) and 54% (n = 68) had an upper arm (BC or transposed BB) AVF. Compared with an upper arm AVF, those who received a forearm AVF were significantly younger (62 ± 16 years vs. 69 ± 13 years, p = 0.01), more likely to be male (68% vs. 50%, p = 0.04), had a lower burden of both diabetes (33% vs. 54%, p = 0.02)and cardiac and/or vascular disease (32% vs. 50%, p = 0.04), and longer Nephrology contact time (median 25 months vs. 13 months, p < 0.01) The odds of receiving an upper arm AVF were higher if you were older (OR 1.16, CI: 1.02-1.32), female (OR 2.45, CI: 1.12-5.34), or diabetic (OR 2.26, CI: 1.04-4.89).

Back to Top | Article Outline

Flow of Patients

Figure 1 outlines the flow of patients in the AVF group during the course of the study prior to dialysis initiation. Overall, 72% (n = 90) had an uncomplicated AVF course and 28% (n = 35) suffered complications resulting in AVF failure prior to need for dialysis: 17% (n = 21) NM AVF and 11% (n = 14) AVF thrombosis (Th). Please note that three patients with an uncomplicated AVF course in Figure 1 were discovered to have an NM AVF within the first 6 weeks of dialysis commencement. As such, these patients are included in the NM AVF group instead of the no AVF complications category in Table 2.

Of those who suffered AVF thrombosis (n = 14, Figure 1), 79% occurred within 6 months of access creation and 21% occurred 6-12 months after creation (please note that time of complication is not depicted in Figure 1); all except one patient underwent a 2nd AVF creation. Sixty-two percent of NM AVF were identified in the first 6 months and the remainder 6-12 months after creation. In contrast to those with AVF thrombosis, less than half (n = 10) with a NM AVF underwent a 2nd AVF creation. Of all patients undergoing a 2nd AVF’s creation (n = 23), 48% started dialysis successfully with their 2nd AVF.

The overall rates of success we observed with early AVF creation (irregardless of whether they had one or two AVF’s created during this time period) are as follows: 50% started dialysis with a mature AVF, 20% started dialysis with a catheter as the AVF was not yet ready to needle (or not present in those who did not undergo 2nd AVF creation), 22% did not start dialysis by study end, and had a healthy AVF (i.e., they were not diagnosed with AVF thrombosis or nonmaturation during the course of the study), 4% had a healthy AVF, but it was not needed as they went on to other modalities of renal replacement therapy or died, and 3% suffered thrombosis or nonmaturation of their second AVF and did not undergo a third attempt during the study.

Figure 2 depicts probability of complication-free (i.e., thrombosis and not maturing) AVF survival in the AVF group before dialysis initiation. We observed a 6-month probability of 80% and 12 month probability of 62% complication-free AVF survival. Table 2 depicts the demographic and clinical differences between patients depending on complication status in the AVF group during the entire course of the study, including the first 6 weeks of dialysis in those who commenced dialysis during the study period. In a logistic regression model including all significant variables (p < 0.05) in Table 2, we found that younger age (OR 0.79 per 5 years, 0.65-0.97) was an independent predictor of AVF Th, and a history of cardiac and/or vascular disease (OR 4.27, 1.52-12.00) was predictive of a NM AVF. Other serious complications experienced by patients in the AVF group before dialysis initiation are depicted in Table 3.

Interestingly, not all the cohort commenced dialysis: 70% of the AVF group and 61% of the no AVF group started dialysis (p = 0.14) during the course of the study for an overall rate of 65%. Seventy-two percent (n = 66) of the AVF group used a mature AVF as their first dialysis access, in contrast to none of the patients in the no AVF group. At the 6-month mark of dialysis, the AVF group had a much higher number of patients using an AVF use compared with the no AVF group whose access was created after dialysis initiation (81% vs. 44%); CVC use was 19% vs. 56%, respectively (p-value = 0.001, Figure 3).

Back to Top | Article Outline


We have described outcomes of an organized vascular access program at a single tertiary care institution in Canada with specific focus on the occurrence of complications leading to AVF failure (thrombosis, failure to mature) and other serious complications which do not necessarily lead to loss of the AVF (ischemic monomelic neuropathy, steal syndrome, and infection) before the access is used for dialysis. Importantly, not all patients with AVF creation require dialysis over a 2 year follow up period; in those that do, the creation of a first AVF in accordance with current recommendations leads to 70% success rate and a 30% complication rate, most of which occurs prior to dialysis start. More importantly we have described that with early AVF creation, these complications are remedied without the need for a catheter, and that this strategy maximizes intermediate term (6 month) AVF use in dialysis patients as compared with fistulas created after dialysis start.

Thirty percent of the AVF cohort experienced primary AVF failure, with a slight predominance of nonmaturation relative to thrombosis before dialysis initiation over the 2-year period of our study. A study by Ernandez found that independent predictors of primary failure included forearm AVF, female gender, level of surgical expertise, and diabetes mellitus.11 In the present study, the only independent predictor of AVF thrombosis was younger age at time of AVF creation, however, approximately 70% of these patients had a lower arm AVF; cardiac and/or vascular disease was the only independent predictor of a NM AVF. We described that the majority of those with AVF thrombosis, but less than half of those with a NM AVF underwent a second AVF creation, with good outcomes. The reasons that so few with NM AVF went for a second procedure are unknown, however, other investigators have demonstrated good outcomes in similar patient groups.12–14 Similar to other reports, we found a low occurrence of arterial steal, infection, and ischemic monomelic neuropathy for AVFs.4,15–18

Back to Top | Article Outline

Other Observations

It remains that a significant proportion of patients are not referred for access creation despite requiring dialysis. In the present study, a similar percentage of patients in the no AVF group (61% vs. 70%, respectively, p = 0.14) started dialysis during the period of observation and did not undergo early AVF creation; perhaps they were not referred due to MD or patient failure to appreciate the need for this. As demonstrated in Figure 2, at the 6-month mark of dialysis the no AVF group had an unacceptably high rate of CVC use (56%). Other published data defines the important and problematic consequences: those patients starting HD with a CVC and remaining on a CVC at 3 months (CVC/CVC) have the highest mortality compared with CVC/AVF, with the lowest risk observed in those with AVF/AVF.19,20 In addition, people with a high burden of cardiovascular disease, a similar demographic to this cohort, who are referred late for vascular access are at high risk for primary AVF failure.21

In the AVF group, mean eGFR was 12 ml/min at time of vascular access creation. Although we observed excellent rates of AVF use in this cohort at the 6-month mark of dialysis (81%) meeting both K-DOQI and CSN prevalent AVF recommendations, there was still a moderately high rate of CVC use (19%, K-DOQI: <10% prevalent). Current guidelines suggest that an AVF should be created at least 6 months before anticipated need (KDOQI). The data presented herein suggests that an AVF might be created at least 1-year ahead of anticipated need. This is to address the fact that the majority of AVF thrombosis occurred within the first 6 months, however, approximately 50% of the NM AVF were diagnosed 6-12 months after creation. Additionally, with early creation, there is adequate time to create a second AVF if needed. We were able to identify some patient factors (i.e., history of cardiac and/or vascular disease, age) that might influence the decision as to timing of referral for vascular access creation given the successes and complication rates described.

Finally, it is interesting that approximately 2 of every 3 patients (65%) that underwent early AVF creation and/or had a sustained GFR ≤25 ml/min at the beginning of the study started dialysis within 2 years. In contrast, one of every three patients did not start dialysis, even though a proportion (AVF group) had an access created early. It would be interesting to investigate these latter patients further, so that a clinical/demographic profile of those who do not progress can be established, and AVF creation can be potentially deferred in these patients.

There are several limitations to our study. First, given the cohort, nonrandomized design of this study, a true comparison group does not exist. To mitigate this we used contemporaneous controls that were “eligible” to undergo AVF placement as a comparison group. Second, at the time of the study, no information regarding reasons for nonreferral to the Vascular Access Clinic was provided. Thus, patient versus MD versus other reasons for nonreferral are unknown. Future studies will attempt to discern factors related to early and late referral for vascular access creation. Third, we had access to patient information recorded in the database or chart, but may have missed some minor complications not recorded in this source. This is a minor limitation. We are confident all serious complications were captured as the PROMIS database was cross-referenced to the hospital database and the patient’s hospital chart.

In conclusion, we have identified the importance of early access creation with 70% success and identification and treatment of complications mostly before dialysis start. Future research is required to identify and develop strategies to decrease the variation in referral practices (which are likely both patient and MD related) so as to achieve the best outcomes in a “fistula-first” environment.

Back to Top | Article Outline


The results presented in this article have not been published previously in whole or part, except in abstract format at the annual meeting of the American Society of Nephrology, San Diego, CA, 2006.

Back to Top | Article Outline


1. Astor BC, Eustace JA, Powe NR, 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 16: 1449–1455, 2005.
2. Dhingra RK, Young EW, Hulbert-Shearon TE, et al: Type of vascular access and mortality in U.S. hemodialysis patients. Kidney Int 60: 1443–51, 2001.
3. Ishani A, Collins AJ, Herzog CA, Foley RN: Septicemia, access and cardiovascular disease in dialysis patients: The USRDS Wave 2 study. Kidney Int 68: 311–8, 2005.
4. Zibari GB, Rohr MS, Landreneau MD, et al: Complications from permanent hemodialysis vascular access. Surgery 104: 681–686, 1988.
5. Foundation NK: KDOQI Clinical practice guidelines for vascular access. Am J Kidney Dis 48: S176–S247, 2006.
6. Ethier JH, Lindsay RM, Barre PE, et al: Clinical practice guidelines for vascular access. Canadian Society of Nephrology. J Am Soc Nephrol 10: S297–S305, 1999.
7. Mendelssohn DC, Ethier J, Elder SJ, et al: Haemodialysis vascular access problems in Canada: Results from the Dialysis Outcomes and Practice Patterns Study (DOPPS II). Nephrol Dial Transplant 21: 721–728, 2006.
8. Ethier J, Mendelssohn DC, Elder SJ, et al: Vascular access use and outcomes: An international perspective from the dialysis outcomes and practice patterns study. Nephrol Dial Transplant 23: 3219–3226, 2008.
9. Arora P, Obrador GT, Ruthazer R, et al: Prevalence, predictors, and consequences of late nephrology referral at a tertiary care center. J Am Soc Nephrol 10: 1281–1286, 1999.
10. Miller PE, Tolwani A, Luscy CP, et al: Predictors of adequacy of arteriovenous fistulas in hemodialysis patients. Kidney Int 56: 275–280, 1999.
11. Ernandez T, Saudan P, Berney T, et al: Risk factors for early failure of native arteriovenous fistulas. Nephron Clin Pract 101: 39–44, 2005.
12. Dixon BS, Novak L, Fangman J: Hemodialysis vascular access survival: Upper-arm native arteriovenous fistula. Am J Kidney Dis 39: 92–101, 2002.
13. Lok CE, Oliver MJ, Su J, et al: Arteriovenous fistula outcomes in the era of the elderly dialysis population. Kidney Int 67: 2462–2469, 2005.
14. Weyde W, Letachowicz W, Kusztal M, et al: Outcome of autogenous fistula construction in hemodialyzed patients over 75 years of age. Blood Purif 24: 190–195, 2006.
15. Kizilisik AT, Kim SB, Nylander WA, Shaffer D: Improvements in dialysis access survival with increasing use of arteriovenous fistulas in a Veterans Administration medical center. Am J Surg 188: 614–616, 2004.
16. Revanur VK, Jardine AG, Hamilton DH, Jindal RM: Outcome for arterio-venous fistula at the elbow for haemodialysis. Clin Transplant 14: 318–322, 2000.
17. Lok CE, Oliver MJ: Overcoming barriers to arteriovenous fistula creation and use. Semin Dial 16: 189–196, 2003.
18. Ballard JL, Bunt TJ, Malone JM: Major complications of angioaccess surgery. Am J Surg 164: 229–232, 1992.
19. Oliver MJ, Rothwell DM, Fung K, et al: Late creation of vascular access for hemodialysis and increased risk of sepsis. J Am Soc Nephrol 15: 1936–1942, 2004.
20. Ortega T, Ortega F, Diaz-Corte C, et al: The timely construction of arteriovenous fistulae: A key to reducing morbidity and mortality and to improving cost management. Nephrol Dial Transplant 20: 598–603, 2005.
21. Ravani P, Brunori G, Mandolfo S, et al: Cardiovascular comorbidity and late referral impact arteriovenous fistula survival: A prospective multicenter study. J Am Soc Nephrol 15: 204–209, 2004.
Copyright © 2009 by the American Society for Artificial Internal Organs