Vascular Access Thrombosis in France: Incidence and Treatment Patterns

Vascular Access Thrombosis in France: Incidence and Treatment Patterns Nirvana Sadaghianloo,1,2 Elix
ene Jean-Baptiste,1,2 Mohammed S. Islam,2,3 Alan Dardik,4 Serge Declemy,1,2 and Reda Hassen-Khodja,1,2 Nice, France; and New Haven, Connecticut

Background: Vascular access thrombosis lacks the implementation of a treatment algorithm at large scale, involving all the actors. We aimed to determine a better understanding of the current practice patterns around vascular access thrombosis in France, with 4 axes: incidence, surveillance protocol, treatment, and time to treatment. Methods: A comprehensive survey of all the nephrologists staffing all hemodialysis centers in France during April 2013 included 266 of 269 (99%) centers, treating 27,798 patients with arteriovenous fistula or graft. Results: In 104 centers treating 11,088 patients, there were 905 documented episodes of vascular access thrombosis (8.8%) in 1 year; in the other 162 centers that supplied a range of events, the mean incidence was 8.4%. Use of in-line access flow monitoring as part of surveillance program was not correlated with better outcome compared with Doppler ultrasound (thrombosis: 7.9% vs. 10%, respectively, P ¼ 0.09). Fifty-three percent of centers referred the patients to a vascular surgeon and 32% to an interventional radiologist (2% to urologist and 13% variable referral depending on the case complexity). Time to treatment was <24 hr in 58% and <48 hr in 91% of the centers; treatment >48 hr (9%) occurred mainly in rural zones (P ¼ 0.04). The specialty of the treating physician did not influence time to treatment (P > 0.05). Conclusions: In France, vascular access thrombosis rate is low and not influenced by surveillance protocol type. Most patients can receive timely treatment by vascular surgeons or interventional radiologists.

INTRODUCTION Patients with end-stage renal disease requiring dialysis have an incidence of 349 per million population 1 Department of Vascular Surgery, University Hospital of Nice, Nice, France. 2

University of NicedSophia Antipolis, Nice, France. Department of Nephrology and Dialysis, University Hospital of Nice, Nice, France. 4 Department of Surgery, Yale University Medical School, New Haven, CT. Correspondence to: Reda Hassen-Khodja, MD, Service de Chirurgie Vasculaire, H^opital Saint-Roch, Centre Hospitalier Universitaire de Nice, 5 Rue Pierre Devoluy, Nice 06000, France; E-mail: [email protected] 3

Ann Vasc Surg 2015; 29: 1203–1210 http://dx.doi.org/10.1016/j.avsg.2015.02.008 Ó 2015 Elsevier Inc. All rights reserved. Manuscript received: November 6, 2014; manuscript accepted: February 26, 2015; published online: May 21, 2015.

(pmp) in the United States,1 and 149 pmp in France (Table I).2 In this growing number of patients, approximately 90% choose hemodialysis (Table I), and therefore, require a functional vascular access. Despite practice guidelines and the Fistula First initiative to increase arteriovenous (AV) fistula use,3,4 AV accesses remain imperfect with overall poor patency rates and multiple reinterventions to maintain or reestablish functionality.5,6 Thrombosis is the ultimate complication of AV accesses, and we previously showed that delaying treatment of an access thrombosis can lead to access loss and unnecessary temporary catheter placement.7 However, the incidence of thrombosis among the hemodialysis population is not properly recorded by national registries. The influence of a standardized surveillance program has been evaluated in local studies, but not at large scale, and is still controversial.8,9 After diagnosis of thrombosis, the algorithm of 1203

1204 Sadaghianloo et al.

Annals of Vascular Surgery

Table I. ESRD population in USA and France Data

USA (2011)1

France (2012)2

Population

311,582,564

65,585,857

ESRD Dialysis Hemodialysis (% of dialysis patients) Primary diagnostic: diabetes, n (%)a Hypertension (% of dialysis patients)a

Prevalence

Rate pmp

Incidence

Rate pmp

Prevalence

Rate pmp

Incidence

Rate pmp

615,899 430,273 395,656 (92)

1,901 1,329 280

115,643 112,788 103,744 (92)

357 349 322

73,491 40,983 38,275 (93)

1,127 629 587

10,048 9,706 8,681 (89)

154 149 133

185,020 (43)

582

48,735 (43)

154

8,737 (21)b

134

2,164 (22)

33

121,322 (28)

382

31,292 (28)

99

9,301 (23)

143

2,520 (26)

39

ESRD, end-stage renal disease. a Among dialysis patients, at treatment initiation. b Probably underestimates refers to diabetes as primary diagnosis for ESRD, but represents only 54% of all patients with ESRD and diabetes (prevalence count, 14,999; rate pmp, 230; 42% of incident ESRD patients).

care still depends on local collaborations between the multiple physicians treating vascular access thrombosis (VAT) and suffers from disparity. Many nephrologists complain of a lack of communication and reactivity in the treatment of VAT,10 leading, in some countries, to develop their own surgical and interventional skills.11,12 Therefore, we first determined the incidence of VAT in a closed but large geographical locality and evaluated the role of surveillance protocols on VAT rates. Then, we tried to determine a better understanding of the nephrologist’s practice patterns and the reactivity of the treating physicians (vascular surgeons [VSs], interventional radiologists [IRs], and urologists) after diagnosis of VAT.

METHODS During April 2013, we used 3 public Web-based directories13,14 to reach 269 hemodialysis centers based in the entire countries of France and Monaco, by phone. Centers in overseas French departments were excluded from the survey but present in the governmental registries. Patients treated with home hemodialysis, peritoneal dialysis, and pediatric patients were also excluded. Data from freestanding dialysis centers and satellite units were included in the data provided by the main center they were referred. For each center, a standardized questionnaire, referring to calendar year 2012, was read to a senior nephrologist to determine the number of hemodialysis patients, percentage of permanent catheter usage, incidence of VAT, surveillance program

(in-line access flow monitoring and/or Doppler ultrasound), specialty of the referral physician consulted to treat VAT, and usual time between diagnosis and treatment. For the incidence of VAT, when precise data were not available from a local database, we asked to provide a range of events (0e5, 5e10, 10e15, and 15e20). When data were not immediately available, we gave time to the nephrologist to gather precise information from database and call them later. Six local centers were audited. We cross-checked the thrombosis rate with the number of procedures for surgical or radiologic thrombectomy performed in 2012, based on the governmental database of procedure codings. Centers were divided into 2 categories based on their location, rural, and urban areas (French departments with, respectively, less or more than 1 million inhabitants). We also compared the different types of centers: those in public university hospitals (UNIV), in public nonuniversity hospitals (PUB), and private centers (PRIV). Individual patients undergoing hemodialysis were not contacted; because no human patients were contacted, and all data were recorded without any patient specific information, no informed consent was obtained, and the institutional ethic committee deemed that review of the study was not necessary. Statistical analysis was performed with StatView software (SAS Institute, Inc., Cary, NC; version 5.0). Continuous variables were expressed as mean ± standard deviation and compared using unpaired Student’s t-test. Qualitative variables were analyzed using chi-squared comparison. For centers that provided a range of events, we used

Type of center

University hospital

Nonuniversity hospital

Private center

37 Centers

117 Centers

112 Centers

0.08 ± 0.05, n ¼ 53

Variable Vascular access thrombosis, incidence ± standard deviation Documented data, n ¼ 104 centers 0.11 ± 0.07, 0.08 ± 0.05, n ¼ 18 n ¼ 33 Data from estimated range of events, n ¼ 162 centers

0.10 ± 0.08, n ¼ 19

a

Statistical significance.

0.07 ± 0.04, n ¼ 59

Urban areas

P

130 Centers

136 Centers

P

UNIV.PUB: P ¼ 0.04a PUB.PRIV: P ¼ 0.86 UNIV.PRIV: P ¼ 0.03a UNIV.PUB: P ¼ 0.23 PUB.PRIV: P ¼ 0.04a UNIV.PRIV: P ¼ 0.03a

0.09 ± 0.04, n ¼ 48

0.09 ± 0.06, n ¼ 56

0.80

0.09 ± 0.05, n ¼ 82

0.08 ± 0.06, n ¼ 80

0.25

16 (14) 83 (71) 18 (15)

2 (2) 69 (62) 41 (37)

0.0003a

14 (11) 91 (70) 25 (19)

7 (5) 84 (62) 45 (33)

0.02a

75 29 11 2

53 42 16 1

0.06

87 29 11 3

55 58 22 1

0.0006a

(64) (25) (9) (2)

26 (22) 66 (56) 103 (88)

(47) (38) (14) (1)

51 (46) 70 (63) 107 (96)

0.0009a 0.30 0.11

(67) (22) (9) (2)

38 (29) 73 (56) 114 (88)

(40) (43) (16) (1)

51 (38) 81 (60) 129 (95)

0.15 0.57 0.04a

Vascular access thrombosis in France 1205

Temporary catheter placement: centers (%) Systematic 3 (8) On demand 23 (62) Never 11 (30) Specialty of the referral physician: centers (%) Vascular surgeons 14 (38) Interventional radiologists 16 (43) Mixed 6 (16) Urologists 1 (3) Time to treatment: centers (%) Immediate treatment 12 (32) Treatment <24 hr 18 (49) Treatment <48 hr 33 (89)

0.09 ± 0.05, n ¼ 84

Rural areas

Vol. 29, No. 6, August 2015

Table II. Incidence of vascular access thrombosis, temporary catheter placement, specialty of the referral physician and time to treatment: comparison between type of centers and areas

1206 Sadaghianloo et al.

Annals of Vascular Surgery

Fig. 1. Distribution of participating hemodialysis centers. (A) Type of hemodialysis centers. (B) Distribution of patients among centers. (C) Distribution of permanent catheter (C) and AV access among hemodialysis patients.

the mean of the range to estimate the incidence. A value of P < 0.05 was considered to indicate statistical significance. Maps were made with Geoclip O3 Web-based software.15

RESULTS Incidence of VAT: Results from the Survey Among the 269 dialysis centers we reached, 266 centers (99%) treating 32,461 hemodialysis patients participated in the survey. Three centers refused to participate. There were 37 UNIV (3,681 patients), 117 PUB (15,656 patients), and 112 PRIV (13,124 patients; Fig. 1A). The distribution of patients between types of center was similar (Fig. 1B). Centers reported a mean rate of permanent catheter use of 14%. Approximately 27,798 (86%) dialysis patients were treated with a permanent AV access (fistula or graft; Fig. 1C). Seventy-three centers provided an estimate of AV grafts use, with a mean of 8% of their AV accesses. Among the participating centers, 104 centers treating 11,088 patients provided precise data based on their local database, with 905 documented VAT in 2012. Their mean incidence rate of thrombosis was 8.8 VAT per 100 hemodialysis patients. In the other 162 centers that could only provide a range of events, the mean incidence rate of VAT was estimated at 8.4 VAT per 100 hemodialysis patients. UNIV reported significantly more thrombosis than PUB (P ¼ 0.04) and PRIV (P ¼ 0.03). Comparison of the VAT rate between rural and urban areas did not show any significant differences (Table II).

Surveillance Programs Among 250 centers that provided their surveillance protocol, our survey distinguished 3 main protocol patterns among dialysis centers: (1) clinical examination and in-line access flow monitoring (ILAFM; n ¼ 83; 33.2%), (2) clinical and Doppler ultrasound examination (n ¼ 59; 23.6%), and (3) clinical examination, ILAFM, and Doppler ultrasound (n ¼ 77; 30.8%). Other centers were using clinical examination alone (n ¼ 21; 8.4%), ILAFM and Doppler ultrasound (n ¼ 5; 2%), ILAFM alone (n ¼ 3; 1.2%), or Doppler ultrasound alone (n ¼ 2; 0.8%). We detailed the surveillance program of the 98 centers that provided both documented VAT rates and surveillance protocols in Table III. There was no significant difference between VAT rates in centers with type 1 protocol or type 2 protocols (8% vs. 10%, respectively, P ¼ 0.21), nor between type 2 and type 3 protocols (10% vs. 7.9%, P ¼ 0.18) or between type 1 and type 3 protocols (P ¼ 0.89). Nephrologist’s Attitude toward VAT Once thrombosis was diagnosed, the attitude toward the use of a temporary catheter differed among centers. Twenty-one centers (8%) reported placing a temporary catheter systematically. They were predominantly located in rural areas (11% vs. 5%, P ¼ 0.02, Table II). Seventy centers (26%) claimed that they never used temporary catheters. They were mainly located in urban areas (33% vs. 19%, P ¼ 0.02, Table II). Patients were referred to a VS, an IR, or a urologist for declotting and revision of the access, with choice of physician specialty referred to dependent on the center’s network and timeliness of the

Vol. 29, No. 6, August 2015

Vascular access thrombosis in France 1207

Table III. Vascular access surveillance programs Clinical examination (n ¼ 37)

Clinical examination + Doppler ultrasound (n ¼ 53)

In-line access flow monitoring

n

%

n

%

In-line dialysance (Fresenius) Ultrasound dilution (Transonic) Fresenius AND Transonic Total (Fresenius OR Transonic) No in-line surveillance

10 11 5 26 11

27.0 29.7 13.5 70.3 29.7

14 11 4 29 24

26.4 20.8 7.5 54.7 45.2

Based on 98 centers providing both documented VAT rates and surveillance programs; 2 centers were using Transonic only and 6 centers Doppler ultrasound only.

Fig. 2. Distribution of hemodialysis centers in France. Base map gray scale represents the density of population. Circle’s size is proportional to center’s size. (A) In red, centers referring VAT to vascular surgeons (VASC)

exclusively (53%). (B) In blue, centers referring VAT to interventional radiologists (RAD) exclusively (33%). (C) In Paris and immediate suburbs, 10% of centers refer VAT to VS and (D) 71% refer VAT to IR.

referring physician’s response. There was not any report of referral to an interventional nephrologist. In 142 centers (53%), patients were referred exclusively to VS (Fig. 2A). These centers were mainly located in rural areas (67% of all centers in rural areas vs. 40% of centers in urban areas,

P ¼ 0.0006, Table II). In 87 centers (33%), patients were referred exclusively to IR (Fig. 2B). These centers were predominantly located in urban areas (43% of all centers in urban areas vs. 22% of centers in rural areas), particularly in Paris and its immediate suburbs, where they represented 71% of the

1208 Sadaghianloo et al.

centers (Fig. 2C, D). In 33 centers (12%), patients were referred to a specialist depending on the assumed case complexity (Supplementary Fig. 1A). Only 4 centers (2%) worked with urologists trained to vascular access care (Supplementary Fig. 1B). Reaction of the Treating Physicians Eighty-nine centers (33%) reported having access to a specialist immediately after diagnosis of VAT. These centers represent 46% of PRIV, 32% of UNIV, and 22% of PUB (P ¼ 0.0009, Table II). In 154 centers (58%), they claimed having the patient treated in 24 hr. These centers represent up to 63% of PRIV. Overall, in 243 centers (91%), patients had access to treatment within 48 hr after diagnosis of VAT. The remaining 23 centers (9%), where treatment beyond 48 hr still routinely occurred, were predominantly in rural areas (P ¼ 0.04). The specialty of the treating physician did not influence time to treatment (P > 0.05). Around VAT: Data from Existing Registries The prevalence of hemodialysis patients on the last calendar day of 2012 was 36,151.16 The number of thrombectomy procedure codings in 2012 in France was 4,688, including open surgical thrombectomy (n ¼ 3,411; 72.8%), percutaneous mechanical thrombectomy (n ¼ 355; 7.6%), percutaneous thromboaspiration (n ¼ 873, 18.6%), and percutaneous thrombolysis (n ¼ 49, 1%). With these data, we could calculate a rate of treated VAT of 13%. However, these numbers also include procedures for early thrombosis, meaning that some patients were not on hemodialysis yet. It was impossible, from these registries, to determine the number of procedures performed for hemodialysis patients only.

DISCUSSION This comprehensive survey of dialysis centers in France provides important new data, for example, an annual incidence of 8.8 VAT per 100 hemodialysis patients. As a comparison, the incidence obtained by dividing the number of codings for thrombectomy procedures by the prevalent number of hemodialysis patient was 12.8%. This is probably overestimating the real incidence, because codings included thrombectomy for early thrombosis (either immediate thrombosis after creation or thrombosis before the use of the AV access), so they were not always relevant to the hemodialysis population. On

Annals of Vascular Surgery

the other hand, our survey can underestimate the real incidence because one can assume that centers providing documented data may be centers with good follow-up, where thrombosis are less likely to happen. It is very difficult to compare our data with published literature because the incidence of thrombosis is not available for a large unselected population, and surgical literature generally reports patency rates for vascular access rather than incidence of specific events. We found only 1 recent monocentric Spanish study that reported 268 episodes of VAT with an incidence rate of 0.10 VAT per patient year (10 VAT per 100 hemodialysis patients they treated).17 The incidence of thrombosis has probably decreased in developed countries after implementation of the Fistula First initiative. As a comparison, the Hemodialysis study,18 which included 1,846 patients between 1995 and 2000, was reanalyzed by Chang et al.19 to assess VAT. After exclusion of catheter-based vascular accesses, patients with <1-year follow-up, and patients without assessment of their vascular accesses, 1,426 patients presented with 2005 thromboses in a median follow-up period of 3.1 years. More than 40% of their cohort were using AV grafts. From these data, we estimate an incidence rate of 0.45 VAT per patient year. Establishing the incidence of VAT is a first step toward implementation of an optimized care protocol for VAT, as recommended in the National Kidney Foundation - Dialysis Outcome Quality Initiative (NKF-DOQI) guidelines for vascular access.3,6,20,21 Although recommended by the NKF-DOQI guidelines, surveillance programs leading to preventive intervention are still debated.3,9 Our study found 3 main patterns of surveillance, but none of them showed a significant benefit in terms of decreasing the rate of VAT. However, we do believe that monitoring and surveillance should be implemented for early diagnosis and treatment of dysfunctional accesses. We were amazed that French national registries could not provide the surveillance protocols and correlate them with dysfunction events. After diagnosis of VAT, the referral pattern of the nephrologist depended on the availability of the referral physician who could treat the patient. In addition, temporary catheters were more likely to be used in rural areas where access to VS may be easier than access to IR. On the contrary, in urban areas, and especially in Paris area, IRs were the preferred specialists treating VAT. These important data raise the question of the role of various specialists in the care of VAT. Although there is no consensus to prefer surgical or percutaneous

Vol. 29, No. 6, August 2015

approaches in declotting vascular accesses,22e24 the ability to perform balloon angioplasty of stenosis in the access is a critical component of VAT treatment algorithms. Overall, the best specialist seems the one who is more reactive to the patient’s request. We also showed that the VAT rate was higher in university hospitals than in PRIV or nonuniversity hospitals. We believe that this increased rate of VAT may reflect the characteristics of the dialysis population cared for in tertiary centers or the distribution of VS and IR in these centers; however, this may also reflect a selection bias. Time to treatment was consistent with European best practice guidelines4 for more than 90% of the centers, which require patient treatment within 48 hr. However, immediate and within 24-hr treatment was not yet a standard for most centers and specialists. These data suggest the need for the next generation of nationwide organization with written recommendations and protocols, particularly in rural areas, to facilitate timely access to a treating physician.21 This study has several limitations. First, it was based on a telephone survey, and although we reached a senior nephrologist to answer the questionnaire, we only verified the data by directly checking their files or electronic databases in 6 centers. Second, some centers did not use databases or their data were incomplete, leading to lack of precision in some data. However, for our main end point, comparison between the actual and estimated incidence rates showed that precise and estimated data were very similar, and data extrapolated from national registries were also similar. Finally, our lack of detailed patient-specific data regarding VAT (comorbidities and fistula history) may preclude specific details of particular patterns. For example, the VAT rate may depend on the AV access age and type, but the design of the study did not allow the collection of this information. However, our high rate of response (99%) yielded homogenous data with precise rates of VAT.

CONCLUSIONS In conclusion, we report a low rate of VAT among adult patients undergoing hemodialysis, in a country where the Fistula First initiative has been implemented for a long time (less than 10% grafts). Various surveillance programs are followed throughout the country, but none of them was correlated with an improved VAT rate. Patients with VAT undergo treatment within the time recommended by the European guidelines; however, treatment is referred to different specialties

Vascular access thrombosis in France 1209

depending on the area of referral. The referral pattern of the nephrologist seems to depend on the availability of the referral physician. Quality and outcome improvement cannot be determined without analyzing data, and our study also highlights the need for the record of all AV access events in national registries.

The authors thank the 266 nephrologists who participated to the survey and helped us understand their practice pattern. The authors also thank Dr. Yves Gendreike (University Hospital of Nice), who provided data on procedure codings.

SUPPLEMENTARY DATA Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.avsg.2015.02. 008. REFERENCES 1. U.S. Renal Data System. USRDS 2013 Annual Data Report: Atlas of End-Stage Renal Disease in the United States. Bethesda, MD: National Institute of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2013. 2. Hannedouche T, Brianc¸on S, Lassalle M, et al. Chapter 2: 2012 ESRD Incidence Rates, in REIN-Rapport annuel 2012. Saint-Denis la Plaine, France: Agence de la Biomedecine. pp 37e83. Available from: http://www.agence-bio medecine.fr/Le-programme-REIN; 2014. Accessed November 2014. 3. National Kidney Foundation’s. KDOQI 2006 Vascular Access Guidelines. Am J Kidney Dis 2006;48(Suppl 1):S177e322. 4. Tordoir J, Canaud B, Haage P, et al. EBPG on vascular access. Nephrol Dial Transplant 2007;22(Suppl 2):S88e117. 5. Schwab SJ. Hemodialysis vascular access: the Achilles’ heel remains. Kidney Int 2007;72:665e6. 6. Riella MC, Roy-Chaudhury P. Vascular access in haemodialysis: strengthening the Achilles’ heel. Nat Rev Nephrol 2013;9:348e57. 7. Sadaghianloo N, Jean-Baptiste E, Gaid H, et al. Early surgical thrombectomy improves salvage of thrombosed vascular accesses. J Vasc Surg 2014;59:1377e84. 8. Paulson WD, Moist L, Lok CE. Vascular access surveillance: case study of a false paradigm. Semin Dial 2013;26:281e6. 9. Polkinghorne KR. Vascular access surveillance: time to end the controversy. Semin Dial 2013;26:257e9. 10. Hakim R, Himmelfarb J. Hemodialysis access failure: a call to action. Kidney Int 1998;54:1029e40. 11. Chan MR. Interventional nephrology: what the nephrologist needs to know about vascular access. Clin J Am Soc Nephrol 2013;8:1211e2. 12. Ponce P, Carvalho T, Messias H, Neves F, Lisbon VACINT. Assessing the approach to a thrombosed AV graft. Semin Dial 2014;27:518e21. 13. Societe de Nephrologie. Annuaire des centres de dialyse. Available from: http://www.soc-nephrologie.org/epublic/ centres/. Accessed August 2014. 14. Ligue Rein et Sante de France. Annuaires de centres de dialyse. Available from: http://www.centrededialyse.com/ index.php. Accessed August 2014.

1210 Sadaghianloo et al.

15. Geoclip - Intuitive and smart cartography. Available from: http://www.geoclip.fr/fr/index.php. Accessed August 2014. 16. Kolko A, Lassalle M, Vigneau C. Chapter 5: Clinical Characteristics and Care Indicators for Dialysis Patients, in REINRapport annuel 2012. Saint-Denis la Plaine, France: Agence de la Biom edecine. pp 175e212. Available from: http:// www.agence-biomedecine.fr/Le-programme-REIN; 2014. Accessed November 2014. 17. Jimenez-Almonacid P, Gruss E, Lasala M, et al. Economic repercussions of implementing a protocol for urgent surgical repair of thrombosed arteriovenous fistulae. Nefrologia 2014;34:377e82. 18. Eknoyan G, Beck GJ, Cheung AK, et al. Effect of dialysis dose and membrane flux in maintenance hemodialysis. N Engl J Med 2002;347:2010e9. 19. Chang TI, Paik J, Greene T, et al. Intradialytic hypotension and vascular access thrombosis. J Am Soc Nephrol 2011; 22:1526e33.

Annals of Vascular Surgery

20. Flu H, Breslau PJ, Krol-van Straaten JM, et al. The effect of implementation of an optimized care protocol on the outcome of arteriovenous hemodialysis access surgery. J Vasc Surg 2008;48:659e68. 21. Allon M, Bailey R, Ballard R, et al. A multidisciplinary approach to hemodialysis access: prospective evaluation. Kidney Int 1998;53:473e9. 22. Huber TS, Buhler AG, Seeger JM. Evidence-based data for the hemodialysis access surgeon. Semin Dial 2004;17: 217e23. 23. Sidawy AN, Spergel LM, Besarab A, et al. The Society for Vascular Surgery: clinical practice guidelines for the surgical placement and maintenance of arteriovenous hemodialysis access. J Vasc Surg 2008;48(Suppl 5): S2e25. 24. Tordoir JH, Bode AS, Peppelenbosch N, et al. Surgical or endovascular repair of thrombosed dialysis vascular access: is there any evidence? J Vasc Surg 2009;50:953e6.