Surveillance of arteriovenous hemodialysis access: A systematic review and meta-analysis

Article Outline

• Abstract
• Methods
  • Eligibility criteria
  • Study identification
  • Data collection
  • Statistical analysis
• Results
  • Study identification
  • Methodologic quality
  • Meta-analysis
  • Subgroup analyses
• Discussion
  • Our findings
  • Limitations and strengths of this review
  • Implications for research and practice
• Conclusions
• Author contributions
• References
• Copyright

Objectives

Hemodialysis centers regularly survey arteriovenous (AV) accesses for signs of dysfunction. In this review, we synthesize the available evidence to determine to what extent proactive vascular access monitoring affects the incidence of AV access thrombosis and abandonment compared with clinical monitoring.

Methods

We searched electronic databases (MEDLINE, EMBASE, Cochrane CENTRAL, Web of Science, and SCOPUS) and sought references from experts, bibliographies of included trials, and articles that cited included studies. Two reviewers independently assessed trial quality and extracted data. We used random effects meta-analysis to estimate the pooled relative risk (RR) and 95% confidence interval (CI) across studies and conducted subgroup analyses to explain heterogeneity. The I² statistic was used to assess heterogeneity of treatment effect among trials.

Results

Nine studies (1363 patients) compared a strategy of surveillance vs clinical monitoring. A vascular intervention to maintain or restore patency was provided to both groups if needed. Surveillance followed by intervention led to a nonsignificant reduction of the risk of access thrombosis (RR, 0.82; 95% CI, 0.58-1.16; I² = 37%) and access abandonment (RR, 0.80; 95% CI, 0.51-1.25; I² = 60%). Three studies (207 patients) compared the effect of vascular interventions vs observation in patients with abnormal surveillance result. Vascular interventions after an abnormal AV access surveillance led to a significant reduction of the risk of access thrombosis (RR, 0.53; 95% CI, 0.36-0.76) and a nonsignificant reduction of the risk of access abandonment (RR, 0.76; 95% CI, 0.43-1.37).

Conclusion

Very low quality evidence yielding imprecise results suggests a potentially beneficial effect of AV access surveillance followed by interventions to restore patency. This inference, however, is weak and will require randomized trials of AV access surveillance vs clinical monitoring for rejection or confirmation.

Patients with deteriorating renal function and end-stage renal disease require vascular access that is safe, reliable, and associated with minimal complications. Although autogenous arteriovenous (AV) access represents most noncatheter vascular access in the world, >50% of some United States dialysis patients with AV access receive a prosthetic access.¹

Hemodynamically significant outflow stenosis leading to thrombosis is the most common cause of prosthetic access abandonment. Early observational studies suggested that correction of this stenosis could prevent thrombosis and prolong access longevity.², ³ Investigators have since conducted randomized controlled trials (RCTs) testing whether serial measurements of blood flow vs usual clinical monitoring would increase the longevity of AV access. Some studies showed limited effect of surveillance on the incidence of thrombosis or access longevity, whereas others suggested surveillance could decrease complications and reduce hospitalizations for prosthetic access dysfunction.², ³, ⁴ These mixed research results notwithstanding, the National Kidney Foundation, the Canadian Society of Nephrology, and Caring for Australians with Renal Impairment are among several professional organizations that recommend frequent, regular surveillance with physical examination and some form of serial access flow measurements.⁵, ⁶, ⁷

Seeking to provide guidance in this area, the Society for Vascular Surgery (SVS) formed a multispecialty committee to formulate evidence-based clinical practice guidelines for the care of patients who have a vascular access used for chronic hemodialysis. Upon their request and to guide the formulation of these guidelines, we conducted a systematic review of the literature to identify and summarize the best available evidence about the efficacy of access surveillance.

Back to Article Outline

Methods 

The report of this protocol-driven systematic review was approved by the SVS and adheres to the Quality of Reporting of Meta-analyses (QUOROM) standards for reporting systematic reviews of RCTs and reporting Meta-analyses of Observational Studies in Epidemiology (MOOSE).⁸, ⁹ Whenever possible, we used the nomenclatures and definitions as published in the “Recommended Standards for Reports Dealing with Arteriovenous Hemodialysis Accesses” by the SVS.¹⁰

Eligibility criteria 

Eligible studies were RCTs and cohort studies that compared a group that underwent active monitoring (the periodic evaluation of the vascular access by using tests that may involve special instrumentation and for which an abnormal test result suggests the presence of dysfunction) with a control group in which patients received usual clinical monitoring and underwent an interventional procedure only if they developed clinical indications of access dysfunction. Usual clinical monitoring could include physical findings of persistent swelling of the arm, presence of collateral veins, prolonged bleeding after needle withdrawal, or altered characteristics of pulse or thrill in a graft.⁷

In addition, we included as a separate group studies in which all participants had abnormal surveillance results and were then randomly allocated to either vascular intervention (angioplasty or surgical revision) or usual clinical monitoring. The former studies provide evidence about the efficacy of surveillance in general, whereas the latter studies provide evidence regarding the efficacy of vascular interventions in patients with abnormal surveillance results. Hence, randomization was a necessary inclusion criterion only for the efficacy of intervention studies. We included studies that measured the outcomes of interest (thrombosis and access abandonment) regardless of their language, sample size, access type (autogenous or prosthetic), or duration of patient follow-up.

Study identification 

An expert reference librarian (P. J. E.) designed and conducted the electronic search strategy with input from study investigators with expertise in conducting systematic reviews. To identify eligible studies, we searched electronic databases (MEDLINE, EMBASE, Cochrane CENTRAL, Web of Science, and SCOPUS) through March 2007 and actively monitored the literature for newer publications thereafter. We also sought references from experts, bibliographies of included trials, and the Institute for Scientific Information (ISI) Science Citation Index for publications that cited included studies (the strategy is available from the authors upon request).

References were uploaded in a Web-based software package developed for systematic review data management (SRS, TrialStat Corp, Ottawa, Ontario, Canada). Paired reviewers working independently screened all abstracts and titles for eligibility. References that were deemed potentially relevant were retrieved in full text and uploaded for full text evaluation against eligibility criteria. The chance-adjusted inter-reviewer agreement (κ statistic) for study eligibility was 0.78. Disagreements were resolved by consensus (the two reviewers discussed the study and reached a consensus) and by arbitration (a third reviewer adjudicated the study) when disagreement continued.

Data collection 

Two reviewers (E. T. C. and A. Z. R.) working independently and using a standardized form extracted descriptive, methodologic, and outcome data from all eligible studies. We attempted to contact authors of all included studies by e-mail to obtain missing data. One study was translated to English.

Statistical analysis 

For the meta-analyses, we pooled relative risks (RR) from each trial using the DerSimonian-Laird random effects model,¹¹ estimated the 95% confidence intervals (CIs) for each outcome, and calculated the I² statistic, which represents the proportion of variability across trials that is not attributable to chance or random error, but rather due to real differences in study design, populations, or interventions.¹² I² values of 25%, 50%, and 75% indicate low, moderate, and high heterogeneity, respectively. We used StatsDirect 2.5.4 software (StatsDirect Statistical Software Ltd, England; 2005) for analysis.

To explain possible heterogeneity, we planned to conduct subgroup analyses based on patients' gender, age, the presence of diabetes mellitus, access type (autogenous vs prosthetic), study design (RCT vs cohort study), and surveillance method (duplex ultrasound imaging, intra-access flow, or static venous dialysis pressure). We then tested for effect-subgroup interactions,¹³ with α = 0.05.

Back to Article Outline

Results 

Study identification 

Our search and selection procedures are depicted in Fig 1. We found 14 eligible studies (12 original studies and 2 reanalysis of published data). Nine studies compared surveillance and intervention vs usual clinical monitoring and intervention in 1363 participants (mean sample size, 151) with a mean duration of 17 months (range 6-28 months).¹⁴, ¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹, ²⁰, ²¹, ²² The other three were trials of patients who had abnormal surveillance results and were randomly allocated to either vascular interventions (angioplasty or surgical revision) or usual clinical monitoring; included were 207 participants with a range of follow-up of 12 to 15 months.²³, ²⁴, ²⁵, ²⁶ Table I summarizes the characteristics of the included studies. Authors of six of the included studies responded to our queries for missing data.¹⁶, ¹⁷, ¹⁹, ²⁰, ²¹, ²⁶

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Flow chart shows how studies were selected.

Table I. Study characteristics

DUS, Duplex ultrasound; HD, hemodialysis; NR, not reported; PTA, percutaneous transluminal angioplasty; SVP, static venous pressure; SVPR, static venous pressure ratio; US, ultrasound.

We excluded some surveillance studies in which all patients underwent surveillance without a control group²⁷ or studies of efficacy of vascular interventions in which patients were partially randomized,²⁸ because they did not meet our inclusion criteria.

Methodologic quality 

The methodologic quality of the included studies is summarized in Table II. Reviewers had adequate chance-adjusted agreement in judging study quality (κ = 0.73). Overall, studies had poor reporting of allocation concealment and blinding. Although blinding of patients and care providers is often not feasible in surveillance trials, blinding of data collectors and outcome assessors is possible and highly desirable. The proportion of patients lost to follow-up was <10% in seven of 12 (58%) of the trials, and the funding source was nonprofit in seven of 12 (58%).

Table II. Study quality

NR, Not reported.

Meta-analysis 

Pooling results from the nine studies that compared surveillance and intervention vs usual clinical monitoring and intervention (Fig 2) demonstrates that access surveillance led to a nonsignificant reduction in the incidence of access thrombosis (7 studies: RR, 0.82; 95% CI, 0.58-1.16; I² = 37%), and access abandonment (6 studies: RR, 0.80; 95 % CI, 0.51-1.25; I² = 60%). Pooling results from the three studies of patients who had abnormal surveillance results and were randomly allocated to either vascular intervention or usual clinical monitoring (Fig 3) demonstrates a statistically significant reduction of the incidence of thrombosis (RR, 0.53; 95% CI, 0.36-0.76; I = 0%). Nevertheless, there was no significant difference in the incidence of access abandonment (3 studies: RR, 0.76; 95% CI, 0.43-1.37; I² = 70%).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  Meta-analysis of the effect of surveillance on access thrombosis and abandonment. The vertical line indicates no treatment effect; squares and horizontal lines, point estimates, and associated 95% confidence intervals (CIs) for each study; diamonds, random-effects pooled relative risks. RR, Relative risk.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 3. 

  Meta-analysis of the effect of vascular interventions vs observation in patients with abnormal surveillance results. The vertical line indicates no treatment effect; squares and horizontal lines, point estimates and associated 95% confidence intervals (CIs) for each study; diamonds, random-effects pooled relative risks. RR, Relative risk.

Subgroup analyses 

Subgroup analyses are summarized in Table III. We found no significant effect-subgroup interactions based on access type (prosthetic vs autogenous), study design, or surveillance method (P < .05 for both outcomes of access thrombosis and abandonment). We found insufficient data to conduct the other planned subgroup analyses.

Table III. Subgroup analyses

Subgroup	Studies, No.a	RRb (95% CI)	P (interaction test)
Access type
Access thrombosis
Autogenous	1	1.52(0.48-4.85)	.43
Prosthetic	6	0.94(0.72-1.23)
Access abandonment
Autogenous	0	NA	NA
Prosthetic	5	0.74(0.45-1.23)
Study design
Access thrombosis
Randomized trials	6	0.90(0.68-1.20)	.06
Observational trials	1	0.24(0.06-0.85)
Access abandonment
Randomized trials	5	0.74(0.45-1.23)	.28
Observational trials	1	1.36(0.52-3.72)
Surveillance method
Access thrombosis
Duplex ultrasound	5	0.67(0.36-1.24)	.25
Intra-access flow	2	1.10(0.62-1.96)
Static venous dialysis pressure	0	NA
Access abandonment
Duplex ultrasound	3	0.70(0.48-1.02)	.98
Intra-access flow	5	0.91(0.53-1.57)
Static venous dialysis pressure	1	1.60(0.79-3.99)

CI, confidence interval; NA, nonapplicable or incalculable; RR, relative risk.

Back to Article Outline

Discussion 

Our findings 

Our systematic review demonstrated a trend toward benefit in the rate of thrombosis and access survival in patients who underwent active access surveillance compared with patients who had usual clinical monitoring.

Two of the included studies that compared a policy of surveillance with that of usual clinical monitoring demonstrated a significant reduction of access abandoment¹⁴ and thrombosis.²⁰ The first, by Malik et al,¹⁴ had methodologic questions regarding randomization of patients, intention to treat analysis, and widely varying times of ultrasound Doppler surveillance. The second study, by Roca-Tay et al,²⁰ followed patients with predominantly autogenous AV access (89%) and reported an unusually high rate of thrombosis in the usual clinical monitoring group of 17% compared with the group undergoing AV access surveillance.²⁰ This study must be interpreted with caution because it is a preliminary report of a 5-year cohort study, and a full report has not been published.

In those patients who underwent an endovascular or surgical procedure for a known stenosis and then were followed up with access surveillance or usual clinical monitoring, the incidence of thrombosis was decreased; however, overall graft life was equivalent between the two groups, perhaps due to restenosis shortly after the vascular intervention.²⁹

Limitations and strengths of this review 

The inferences from this review are limited by the very low quality of the evidence available in the published literature. The quality of evidence supporting surveillance to prolong access survival is downgraded³⁰ due to methodologic limitations of the primary studies (lack of bias protection measures), imprecision (the CIs are wide enough to include important benefit and harm from surveillance), and inconsistency (60% of the heterogeneity of the pooled estimate was not attributed to chance).

Other limitations of the review are related to publication and reporting bias. Publication bias refers to the preferential and prompt publication of significant results in indexed and prominent journals; reporting bias refers to the publication of the outcomes on which the intervention had a significant effect without publication of nonsignificant results.³¹ Generally, the overall effect of these biases is to create an impression in favor of a particular technology. Because our review is inconclusive owing to the imprecise and negative studies, these biases appear not as problematic in this field, but their effect cannot be assessed reliably³² and remains a limitation of any systematic review. Our search for published and unpublished studies, although thorough, does not completely exclude this possibility.

The strengths of this review are derived from having a focused question, an explicit protocol of review with eligibility criteria for studies, and a thorough and systematic search strategy. We also used measures to decrease reviewer bias, such as having independent reviewers select, evaluate, and extract evidence in duplicate with adequate reproducibility, and contacting authors of the primary studies to obtain complete and accurate data.

Implications for research and practice 

The implications for practice are discussed in the accompanying clinical practice guideline. Outcomes other than access thrombosis and survival have also been reported in the literature and may guide the decision about surveillance. A recent reanalysis of one of the studies included in this report suggested that although surveillance by duplex ultrasound imaging combined with preemptive angioplasty did not prolong prosthetic access life, reductions occurred in hospitalization rates and costs as well as the total cost of access-related care.⁴ However, this study was small, and cost-effectiveness analyses in general tend to be highly affected by publication bias.³³ Further studies to determine both the relative efficacy and the financial benefit of AV access surveillance vs clinical monitoring are needed to strengthen the inferences about the relative efficacy of this intervention.

Back to Article Outline

Conclusions 

Very low-quality evidence suggests that serial surveillance of asymptomatic arteriovenous hemodialysis access for the detection and treatment of stenosis may reduce the risk of thrombosis and prolong access survival more than usual clinical monitoring, but these comparisons were not statistically significant. Large multicenter clinical trials of surveillance vs clinical monitoring appear necessary to establish whether surveillance is cost effective in patients with AV access for hemodialysis.

Back to Article Outline

Author contributions 

Back to Article Outline

References 

1. Pisoni RL, Young EW, Dykstra DM, Greenwood RN, Hecking E, Gillespie B, et al. Vascular access use in Europe and the United States: results from the DOPPS. Kidney Int. 2002;61:305–316
   • View In Article
   • MEDLINE
   • CrossRef
2. Schwab SJ, Raymond JR, Saeed M, Newman GE, Dennis PA, Bollinger RR. Prevention of hemodialysis fistula thrombosis (Early detection of venous stenoses). Kidney Int. 1989;36:707–711
   • View In Article
   • MEDLINE
   • CrossRef
3. Besarab A, Sullivan KL, Ross RP, Moritz MJ. Utility of intra-access pressure monitoring in detecting and correcting venous outlet stenoses prior to thrombosis. Kidney Int. 1995;47:1364–1373
   • View In Article
   • MEDLINE
   • CrossRef
4. Dossabhoy NR, Ram SJ, Nassar R, Work J, Eason JM, Paulson WD. Stenosis surveillance of hemodialysis grafts by duplex ultrasound reduces hospitalizations and cost of care. Semin Dial. 2005;18:550–557
   • View In Article
   • MEDLINE
   • CrossRef
5. Churchill D. Canadian clinical practice guidelines (Canadian Society of Nephrology). J Am Soc Nephrol. 1999;10(Suppl 13):S287–S321
   • View In Article
6. Caring for Australians With Renal Impairment (CARI guidelines. Australia Kidney Foundation and Australia New Zealand Society of Nephrology). 2000;
   • View In Article
7. NKF-KDOQI Clinical Practice Guidelines and clinical practice recommendations for Vascular Access, update 2006. Anonymous Am J Kidney Dis. 2006;48(Suppl 1):S177–S247
   • View In Article
8. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement (Quality of Reporting of Meta-analyses). Lancet. 1999;354:1896–1900
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (86 KB)
   • CrossRef
9. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting (Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group). JAMA. 2000;283:2008–2012
   • View In Article
   • MEDLINE
   • CrossRef
10. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg. 2002;35:603–610
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (60 KB)
    • CrossRef
11. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188
    • View In Article
    • MEDLINE
    • CrossRef
12. Higgins J, Thompson S, Deeks J, Altman D. Measuring inconsistency in meta-analyses. BMJ. 2003;326:219
    • View In Article
    • CrossRef
13. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. [see comment] BMJ. 2003;326:219
    • View In Article
    • CrossRef
14. Malik J, Slavikova M, Svobodova J, Tuka V. Regular ultrasonographic screening significantly prolongs patency of PTFE grafts. Kidney Int. 2005;67:1554–1558
    • View In Article
    • MEDLINE
    • CrossRef
15. Mayer DA, Zingale RG, Tsapogas MJ. Duplex scanning of expanded polytetrafluoroethylene dialysis shunts: Impact on patient management and graft survival. Vasc Surg. 1993;27:647–658
    • View In Article
16. Moist LM, Churchill DN, House AA, Millward SF, Elliott JE, Kribs SW, et al. Regular monitoring of access flow compared with monitoring of venous pressure fails to improve graft survival. J Am Soc Nephroly. 2003;14:2645–2653
    • View In Article
17. Plantinga LC, Jaar BG, Astor B, Fink NE, Eustace JA, Klag MJ, et al. Association of clinic vascular access monitoring practices with clinical outcomes in hemodialysis patients. Nephron Clin Prac. 2006;104:151–159
    • View In Article
18. Polkinghorne KR, Lau KKP, Saunder A, Atkins RC, Kerr PG. Does monthly native arteriovenous fistula blood-flow surveillance detect significant stenosis–a randomized controlled trial. Nephrol Dial Transplant. 2006;21:2498–2506
    • View In Article
    • MEDLINE
    • CrossRef
19. Robbin ML, Oser RF, Lee JY, Heudebert GR, Mennemeyer ST, Allon M. Randomized comparison of ultrasound surveillance and clinical monitoring on arteriovenous graft outcomes. Kidney Int. 2006;69:730–735
    • View In Article
    • MEDLINE
    • CrossRef
20. Roca-Tey R, Samon Guasch R, Ibrik O, Garcia-Madrid C, Herranz JJ, Garcia-Gonzalez L, et al. [Vascular access surveillance with blood flow monitoring: a prospective study with 65 patients]. Nefrologia. 2004;24:246–252
    • View In Article
    • MEDLINE
21. Sands JJ, Jabyac PA, Miranda CL, Kapsick BJ. Intervention based on monthly monitoring decreases hemodialysis access thrombosis. ASAIO J. 1999;45:147–150
    • View In Article
    • MEDLINE
    • CrossRef
22. Ram SJ, Work J, Caldito GC, Eason JM, Pervez A, Paulson WD. A randomized controlled trial of blood flow and stenosis surveillance of hemodialysis grafts. Kidney Int. 2003;64:272–280
    • View In Article
    • MEDLINE
    • CrossRef
23. Dember LM, Holmberg EF, Kaufman JS. Randomized controlled trial of prophylactic repair of hemodialysis arteriovenous graft stenosis. Kidney Int. 2004;66:390–398
    • View In Article
    • MEDLINE
    • CrossRef
24. Martin LG, MacDonald MJ, Kikeri D, Cotsonis GA, Harker LA, Lumsden AB. Prophylactic angioplasty reduces thrombosis in virgin ePTFE arteriovenous dialysis grafts with greater than 50% stenosis: Subset analysis of a prospectively randomized study. J Vasc Interv Radiol. 1999;10:389–396
    • View In Article
    • Abstract
    • Full-Text PDF (1830 KB)
    • CrossRef
25. Lumsden AB, MacDonald MJ, Kikeri D, Cotsonis GA, Harker LA, Martin LG. Prophylactic balloon angioplasty fails to prolong the patency of expanded polytetrafluoroethylene arteriovenous grafts: results of a prospective randomized study. J Vasc Surg. 1997;26:382–390discussion 390-2
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (1886 KB)
    • CrossRef
26. Tessitore N, Lipari G, Poli A, Bedogna V, Baggio E, Loschiavo C, et al. Can blood flow surveillance and preemptive repair of subclinical stenosis prolong the useful life of arteriovenous fistulae? (A randomized controlled study). Nephrol Dial Transplant. 2004;19:2325–2333
    • View In Article
    • MEDLINE
    • CrossRef
27. Smits JHM, Van Der Linden J, Hagen EC, Modderkolk-Cammeraat EC, Feith GW, Koomans HA, et al. Graft surveillance: venous pressure, access flow, or the combination?. Kidney Int. 2001;59:1551–1558
    • View In Article
    • MEDLINE
    • CrossRef
28. Tessitore N, Mansueto G, Bedogna V, Lipari G, Poli A, Gammaro L, et al. A prospective controlled trial on effect of percutaneous transluminal angioplasty on functioning arteriovenous fistulae survival. J Am Soc Nephrol. 2003;14:1623–1627
    • View In Article
    • MEDLINE
    • CrossRef
29. Chang CJ, Ko PJ, Hsu LA, Ko YS, Ko YL, Chen CF, et al. Highly increased cell proliferation activity in the restenotic hemodialysis vascular access after percutaneous transluminal angioplasty: implication in prevention of restenosis. Am J Kidney Dis. 2004;43:74–84
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (408 KB)
    • CrossRef
30. Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, et al. Grading quality of evidence and strength of recommendations. BMJ. 2004;328:1490
    • View In Article
    • CrossRef
31. Sinha M, Montori VM. Reporting bias and other biases affecting systematic reviews and meta-analyses: a methodological commentary. Expert review of pharmacoeconomics and outcomes research. 2006;6:603–611
    • View In Article
32. Lau J, Ioannidis JPA, Terrin N, Schmid CH, Olkin I. The case of the misleading funnel plot. BMJ. 2006;333:597–600
    • View In Article
    • CrossRef
33. Bell CM, Urbach DR, Ray JG, Bayoumi A, Rosen AB, Greenberg D, et al. Bias in published cost effectiveness studies: systematic review. BMJ. 2006;332:699–703
    • View In Article
    • CrossRef