A Meta-analysis of Stent Placement vs. Angioplasty for Dialysis Vascular Access Stenosis Ning Fu,* Emily Joachim,† Alexander S. Yevzlin,* Jung-Im Shin,‡ Brad C. Astor,*‡ and Micah R. Chan* *Division of Nephrology, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, †Division of General Internal Medicine, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, and ‡Department of Population Health Sciences, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
ABSTRACT Dysfunction of arteriovenous fistulas (AVF) and arteriovenous grafts (AVG) contribute significantly to morbidity and hospitalization in the dialysis population. Despite advances in endovascular techniques, the incidence of vascular access stenosis remains problematic. Currently, the role of endovascular stent placement in the treatment of vascular access stenosis is poorly defined. This metaanalysis compares the primary patency rates of stenotic vascular access treated with stent placement vs. angioplasty. We searched Medline for English language publications from 1980 through December 2013, along with national conference proceedings and reference lists of all included publications. Inclusion criteria were a measure of primary patency, secondary patency, or access dysfunction. Studies were excluded if they were not in English or if they included pediatric patients. Ten studies with a total of 860 subjects met the inclusion criteria, including six experimental studies and four observational
studies. There was significantly higher overall primary patency in those receiving stent placement than in those treated with angioplasty (pooled relative risk [RR] = 0.79; 95% confidence interval [CI]: 0.65–0.96). The estimate did not differ by study design. The effect of treatment differed significantly (p = 0.001) by the type of stents used, however. In studies including nitinol stents (six studies, 678 patients), 6-month patency was significantly better for stent placement than angioplasty (pooled RR = 0.67; 95% CI: 0.54–0.84), whereas there was no significant differences between stent placement and angioplasty in those studies using bare metal stents exclusively (four studies, 182 patients; pooled RR = 1.09; 95% CI: 0.91–1.32). There was significant heterogeneity between studies (I2 = 70.6%; p < 0.0001). Our results suggest that stent placement may confer an advantage over balloon angioplasty in primary patency of dialysis access stenoses.
Establishing and maintaining adequate vascular access continues to be a major challenge in dialysis. While arteriovenous fistulas (AVF) and grafts (AVG) are the preferred method of long-term access for hemodialysis, complications including thrombosis or stenosis often limit their duration of use. Significant resources are put into ensuring patency of AVF and AVG (1,2). Surgical interventions to extend the lifetime of AVF and AVG were largely supplanted in the 1980s by percutaneous transluminal angioplasty (PTA). While PTA provided a presumed improvement over surgical intervention, early studies dem-
onstrated a 1-year primary patency following PTA in AVF ranging between 31% and 64% (3–5). Even more concerning was a further decline in 1-year patency from a first angioplasty to a second angioplasty (41–22% respectively) (3). Similarly, poor rates of patency following PTA have been observed in AVG, with overall survival of less than 50% at 3 years (6–10). The use of endovascular stents for the treatment of failed dialysis access was first described in 1988 (10,11), and initially, it was used largely as a salvage technique in failed angioplasty or early recurrent stenosis. Initial studies with endovascular stents failed to show improved patency compared to angioplasty. However, as stent technology has advanced, several recent studies have demonstrated superior results using stents for failing dialysis grafts and fistulas (12–14). Increasing attention has been drawn toward the use of stents as a primary method of managing failed dialysis access. Haskal et al. first demonstrated in a multicenter, randomized trial, that primary placement of stents at the venous anastomosis of AVGs had significantly better
Portions of manuscript presented in abstract form at the National Kidney Foundation Spring Clinical Meeting 2014, Las Vegas, NV. Address correspondence to: Micah R. Chan M.D., MPH, University of Wisconsin Hospitals and Clinics, 3034 Fish Hatchery Road, Suite B, Madison, WI 53713, or e-mail: mr. [email protected]. Seminars in Dialysis—2014 DOI: 10.1111/sdi.12314 © 2014 Wiley Periodicals, Inc. 1
2
Fu et al.
patency vs. angioplasty alone (38% vs. 20%, p = 0.008) over a follow-up period of 6 months (13). Furthermore, preliminary results from the RENOVA postapproval, multicenter, randomized trial demonstrated a significant improvement over angioplasty on 12- and 24-month primary patency of venous anastomotic lesions in AVGs (15). Previous studies have important methodological limitations, including small sample sizes, retrospective design (14), and inclusion of AVG venous anastomosis only (13). Thus, the role of endovascular stent placement in the treatment of thrombosed and stenotic AVF and AVG remains controversial. To date, there has been no comprehensive review of the available literature. The objective of this study was to perform a meta-analysis of the available studies comparing primary patency after endovascular stent placement vs. angioplasty in AVF and AVG. Methods Literature Sources and Search Strategy We conducted a computerized search of the literature for clinical studies that addressed stent placement and angioplasty in dialysis access published during the period 1980 through December 2013 using MEDLINE. We used combinations of terms related to dialysis (dialysis, CKD, chronic kidney disease, end-stage renal disease, and ESRD) and vascular access (arteriovenous fistula, AVF, arteriovenous graft, and AVG). This strategy also combined four exploded Medical Subject Headings (stent placement, angioplasty, patency, and dialysis).
Fig. 1. Literature search strategy.
Further, we hand-searched the reference lists of all articles identified in the search as well as those of review articles. We also reviewed the abstracts of the American Society of Nephrology and National Kidney Foundation annual scientific meetings from 2008 through 2013. Study Selection Inclusion criteria were a measure of primary patency, secondary patency, or access dysfunction. It was decided, a priori, to exclude any of the following categories of articles: wrong topic, reviews, pediatric studies, and non-English language studies. From among 668 citations, we reviewed all relevant articles (Fig. 1). Of those, 72 were abstracts or articles appropriate for detailed review. Two independent authors (NF, EJ) reviewed each article from the initial search to determine if inclusion criteria were met. Disagreement was resolved by consensus and, when necessary, with arbitration from a third author (MRC). Data Abstraction Of the 72 studies reviewed in detail, 51 were eliminated due to non-English studies, case reports, pediatric studies, or reviews. Of the remaining 21 studies that assessed access patency, six were eliminated due to inestimable time factors, and one study with data unavailable. Finally, 10 studies were retained for further analysis. The data from each included article were abstracted by two authors (NF, EJ) using a standardized form
3
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY
which included patient number, demographics, control and study group information, study design and primary outcome of interest. Disagreements were resolved through consensus or third reviewer (MRC).
evidence. All studies that were identified provided second-tier evidence except the four randomized trials (13,19–21). Results
Statistical Analysis Kappa scores were used to assess interobserver agreement between the two reviewers. Analyses were performed for all studies and after stratification by study design (experimental or observational) and stent type (nitinol or bare metal exclusively). The outcomes were pooled using random effects models. Publication bias was assessed using funnel plots. Heterogeneity across studies was assessed using the I2 index. Stata/MP13 (Stata Corporation, College Station, Texas, USA) was used for all analyses. p-values < 0.05 were considered statistically significant. Assessment of Study Quality We defined studies as truly experimental only in instances in which the study and data collection were explicitly planned at the time of collation of the study cohort (16). Only in such circumstances would study follow-up procedures be expected to be conducted for the purposes of the study question. Thus, post hoc analysis of a prospective study cohort was considered observational. We assessed study quality by using a modification of the US Preventive Services Task Force criteria described by Fletcher et al. (17,18). These criteria assist in grading the internal validity of studies based on evidence-based guidelines. By these standards, properly designed randomized controlled trials constitute the highest tier evidence; well-designed nonrandomized trials, case–control, or cohort analytic studies provide second-tier
The literature search strategy is summarized in Fig. 1. The agreement between the two reviewers for articles, as well as citations excluded and articles identified, was excellent (j = 0.96; 95% confidence interval [CI]: 0.90–0.99). We identified 10 studies with a total of 860 subjects between 1992 and 2012. Six studies were classified as experimental, including four randomized (13,19–21) and two nonrandomized (6,22), that proceeded with intervention. The observational group comprised four studies (14,23–25) observing patients who had undergone intervention without a priori treatment assignment (Table 1). Vascular access type, stent type, lesions sites, and outcomes of interest are listed in Table 2. Seven of 10 studies compared stent placement vs. angioplasty for the treatment of stenotic AVGs. Most (8 of 10) studies focused on graft-vein anastomosis and venous outflow stenoses, while two studies included central venous stenoses (21,25). A total of six studies (678 patients) included nitinol stents, whereas four studies (182 patients) used bare metal stents exclusively. Primary Patency by Treatment Primary patency was lower in patients receiving angioplasty than in those treated with stent placement (RR = 0.79; 95% CI: 0.65–0.96; p = 0.02) (Fig. 2). The pooled relative risk was markedly similar in observational studies (RR = 0.78; 95% CI: 0.62–0.96; p = 0.02) and experimental studies (RR = 0.77; 95% CI: 0.55–1.09, p = 0.14).
TABLE 1. Summary of included studies Source
Primary location
Number of patients
Karnabatidis et al., 2012 (6) Haskal et al., 2010 (13)
Rion, Greece
35
Maryland, US
190
Vogel et al., 2005 (22)
California, US
60
Hoffer et al., 1997 (19)
Washington, US
37
Beathard et al., 1992 (20)
Texas, US
58
Quinn et al., 1995 (21)
Oregon, US
87
Chan et al., 2008 (14)
Wisconsin, US
Maya et al., 2006 (23) Kakisis et al., 2011 (24) Bakken et al., 2007 (25)
Study design
Cases treated years
Mean age (years, stent vs. PTA)
Male, number (%, stent vs. PTA)
04/2009–01/2011
67.814.3
45.7%
─
61.814.6 vs. 59.813.6 5313 vs. 5915
34% vs. 39%
211
Experimental, nonrandomized Experimental, randomized Experimental, nonrandomized Experimental, randomized Experimental, randomized Experimental, randomized Observational
Alabama, US Athems, Greece
48 61
Observational Observational
04/1999–03/2005 02/2007–12/2010
New York, US
73
Observational
1995–2003
06/2001–03/2003 01/1994–08/1996
32% vs. 42.9%
─
50.919.0 vs. 54.521.7 ─
29.4% vs. 15% ─
over 3 years
58 (26 - 84)
47%
01/2005–06/2007
62.518.1 vs. 63.914.7 5317 vs. 5813 73(39–84) vs. 72(40–84) 5715 vs. 5718
67.1% vs. 67.5% 28% vs. 38% 49% vs. 39% 35% vs. 45%
4
Fu et al. TABLE 2. Summary of vascular access type, lesion site, stent type, and access outcome of included studies
Source
Proportion of AVG
Karnabatidis et al. (6)
100%
Haskal et al. (13)
100%
Vogel et al. (22)
100%
Stenoses at graft-vein anastomosis and venous outflow
Nitinol—SMART nitinol
Hoffer et al. (19)
100%
Bare metal—Wallstent
Beathard et al. (20)
100%
Quinn et al. (21)
99%
Stenoses at or within 3 cm of the vein-graft junction or in the peripheral outflow vein Stenotic lesions at graft-vein anastomosis All sites of venous stenoses and occlusions
Chan et al. (14)
47%
Nonthrombosed, noncentral venous stenoses
Nitinol—SMART nitinol
Maya et al. (23)
100%
Stenoses at venous anastomosis after thrombectomy
Mixed—SMART, Protege, Wallstent, Fluency
Kakisis et al. (24)
100%
Nitinol—absolute nitinol
Bakken et al. (25)
100%
Stenoses at venous anastomosis of prosthetic brachial-axillary accesses after thrombectomy Stenoses at central venous sites
Lesion sites
Stent
Recurrent anastomotic and/or venous outflow stenoses in prosthetic AVGs Venous anastomotic stenosis
Six-month patency was significantly better for stent placement than angioplasty in studies that included nitinol stents (pooled RR = 0.67; 95% CI: 0.54–0.84) (Fig. 3). In contrast, there were no significant differences between stent placement and angioplasty in those studies using bare metal stents exclusively (pooled RR = 1.09; 95% CI: 0.91–1.32). Two studies (21,25) that used stents in central venous stenosis showed a RR above 1. Six of seven studies with a RR below 1 focused on anastomostic and venous outflow stenoses. Test of Homogeneity There was significant overall heterogeneity across studies with respect to the primary patency rate outcomes (I2 = 70.6%; p < 0.0001). The observational studies had moderate heterogeneity, while the experimental studies had significant heterogeneity. Due to a limited number of studies with significant differences in patient risk factors across study populations, the use of meta-regression to assess potential sources of heterogeneity was not undertaken. There was no publication bias identified throughout the studies using funnel plots.
Outcome of interest
Nitinol—Fluency Plus
Primary patency rates (6 and 12 months)
Nitinol—Flair
Primary patency rates of the treatment area and the entire vascular access circuit (6 months); incidence of binary restenosis and other adverse events Primary patency rates (3, 6, and 12 months); lower postprocedural midgraft systolic pressures ratio; complication rates Primary patency rates (30, 60, 180, and 360 days); secondary patency rates Primary patency rates (DE) (30, 60, 90, 180, and 360 days) Primary patency rates (60, 180, and 360 days), secondary patency rates Primary assisted patency rates (30, 90, and 180 days); access flow and urea reduction ratio Immediate postprocedure graft systemic pressure ratio; primary patency rates (30, 90, and 180 days); assisted or secondary patency rates Primary patency rates (3, 6, and 12 months); primary-assisted and secondary patency rates Primary patency rates (30 days and 12 months); assisted primary patency rates
Bare metal—Gianturco Bare metal—Gianturco, Palmaz, Wallstent
Mixed—Wallstent except one SMART stent
Discussion There are considerable challenges facing arteriovenous access creation, maintenance, and salvage (26–28). Although stent placement is the standard of care in percutaneous coronary and peripheral arterial interventions, its role in dialysis access has been debated since the first clinical trials, some of which are included in our analyses. The first rigorous clinical trial to study stent placement in dialysis access was conducted by Beathard et al. (20). Although this study included only 58 patients, the results demonstrated no significant difference between stent vs. angioplasty in any parameter before treatment, after intervention, or patency rates with 90-, 180-, and 360-day survival of 85% vs. 79%, 72% vs. 64%, and 17% vs. 28%, respectively (p > 0.07). In the last decade, nitinol and SMART (shape memory alloy recoverable technology) stents have been used more, which have advantages over bare metal stents in that there is better wall contact, radial strength, and superelasticity (29–31). Vogel, et al. showed in their prospective, nonrandomized trial of 60 patients, a significant decrease in restenosis in the stent group as compared with angioplasty (7% vs. 16%; p = 0.001) and an
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY
5
Fig. 2. Six-month primary patency rates comparing experimental and observational studies.
Fig. 3. Six-month primary patency rates comparing different stent types.
improvement in mean primary graft patency (5.6 vs. 8.2 months; p = 0.05) (22). The observational trials in our pooled analysis all used nitinol stent technology and demonstrated a significant advantage of stent placement as compared to angioplasty alone in 6-month primary patency. More recently, the largest and most robust clinical trial ever published in dialysis access intervention showed a significant advantage of the ePTFE stent graft vs. angioplasty for the primary treatment
of venous anastomotic lesions of AVGs (13). The FLAIR Pivotal trial randomized 190 patients to PTA alone or PTA/Flair Stent Graft. At 6 months, the incidence of patency of the access circuit was significantly greater in the stent-graft group than in the balloon angioplasty group (38% vs. 20%, p = 0.008). The angioplasty group had more instances of restenosis (p < 0.001) and the incidence of freedom from subsequent interventions at 6 months was significantly greater in the stent-graft
6
Fu et al.
group than in the balloon angioplasty group (32% vs. 16%, p = 0.03). The postapproval RENOVA trial examined 270 patients randomized to angioplasty alone vs. Flair Stent Graft at the venous anastomosis of AVGs (15). The interim analysis of the study showed 12-month access patency of 24% vs. 11% (p = 0.007) and 24-month patency of 9.5% vs. 5.5% (p = 0.011) in the stent graft vs. angioplasty groups, respectively. Stenosis requiring reintervention was also greater in the angioplasty group as compared to the stent group (82% vs. 60%; p < 0.001) over 12 months. More recently, Yevzlin et al. presented interim results of the RESCUE trial which was a multicenter, randomized trial of angioplasty vs. stent grafts for in-stent restenosis (32). These results similarly showed superiority of stent grafts as compared to angioplasty over 6 months of postintervention lesion patency (65% vs. 10%, p < 0.001). When AVF and AVG were analyzed separately, the advantage of stent grafts persisted. The interim results of the latter and former trials were not included in the pooled analysis because the followup of patients likely are not complete. The current body of evidence, however, suggests that stent grafts may have an advantage even more durable than uncovered nitinol stents. However, even with all of the compelling evidence pointing toward stents as more than a salvage technique, it must be tempered by the fact that care should be taken not to jeopardize the vein for a future AVF placement and that option must be considered before placing a stent. This study has many limitations. First, we identified significant heterogeneity, implying a significant variation in effect across studies. We recognize this as an inherent deficiency given the varying types of stents used in the studies, type of access (AVF, AVG), and lesions sites. We employed the random effects model that considers intrinsic random between study variance of effect besides variance due to sampling error or within-study variance (33,34). Moreover, differences between studies in key methodological issues, such as variation in types of stents, can be expected to lead to differences in the observed effects. Statistical heterogeneity due to such methodological variation may not necessarily imply that the actual intervention effect differs (35). Moreover, only four randomized clinical studies were included. The predominance of nonrandomized studies could lead to bias due to unrecognized differences between groups. Conclusion In conclusion, our meta-analysis suggests significantly higher primary patency rates in stenotic AV access receiving stent placement vs. angioplasty. Due to the limited and heterogeneous data currently available, it supports that stent placement, especially with nitinol uncovered stents and covered stent grafts, may extend access patency compared to
angioplasty alone in the treatment of graft-vein anastomotic and venous outflow stenoses. Whether there are significant benefits of stent placement or angioplasty in arteriovenous fistulas or secondary patency requires further robust studies. Further investigation in regard to differences in rates of infection, number of interventions, and cost analyses are warranted as secondary outcomes in future trials. Overall, these findings have significant implications for future design of clinical trials in stent placement in dialysis access and the delivery of dialysis-related services. Disclosures No conflict of interest reported by authors. References 1. United States Renal Data System. Annual data report, 2013. Available at http://www.usrds.org/adr.htm, accessed June 22, 2013 2. Feldman HI, Kobrin S, Wasserstein A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 7:523–535, 1996 3. Kanterman RY, Vesely TM, Pilgram TK, Guy BW, Windus DW, Picus D: Dialysis access grafts: anatomic location of venous stenosis and results of angioplasty. Radiology 195:135–139, 1995 4. Glanz S, Gordon DH, Butt KM, Hong J, Lipkowitz GS: The role of percutaneous angioplasty in the management of chronic hemodialysis fistulas. Ann Surg 206:777–781, 1987 5. Beathard GA: Percutaneous transluminal angioplasty in the treatment of vascular access stenosis. Kidney Int 42:1390–1397, 1992 6. Karnabatidis D, Kitrou P, Spiliopoulos S, Katsanos K, Diamantopoulos A, Christeas N, Siablis D: Stent-grafts versus angioplasty and/or bare metal stents for failing arteriovenous grafts: a cross-over longitudinal study. J Nephrol 26:389–395, 2013 7. Bittl JA: Catheter interventions for hemodialysis fistulas and grafts. JACC Cardiovasc Interv 3:1–11, 2010 8. Vascular Access Work Group: Clinical practice guidelines for vascular access. Am J Kidney Dis 48(Suppl):S248–S273, 2006 9. Schwab SJ: Vascular access for hemodialysis. Kidney Int 55: 2078–2090, 1999 10. Zollikofer CL, Antonucci F, Stuckmann G, Mattias P, Bruhlmann WF, Salomonowitz EK: Use of the Wallstent in the venous system including hemodialysis-related stenoses. Cardiovasc Intervent Radiol 15:334–341, 1992 11. Gunther RW, Vorwerk D, Klose KC, Bohndorf K, Kistler D, Mann H, Sieberth HG: Self-expanding stents for the treatment of a long venous stenosis in a dialysis shunt: case report. Cardiovasc Intervent Radiol 12:29–31, 1989 12. Schwab SJ: Hemodialysis vascular access: the Achilles’ heel remains. Kidney Int 72:665–666, 2007 13. Haskal ZJ, Trerotola S, Dolmatch B, Fletcher KE, Underwood W 3rd, Davis SQ, Mangrulkar RS, McMahon LF Jr, Saint S: Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med 362:494–503, 2010 14. Chan MR, Bedi S, Sanchez RJ, Young HN, Becker YT, Kellerman PS, Yevzlin AS: Stent placement versus angioplasty improves patency of arteriovenous grafts and blood flow of arteriovenous fistulae. Clin J Am Soc Nephrol 3:699–705, 2008 15. Saad TF: Update on RENOVA trial. Presented at the 10th Annual Scientific Meeting of the ASDIN, Phoenix AZ, 2014 16. Szklo M, Nieto J: Epidemiology: Beyond the Basics. Gaithersburg, MD: Aspen Publishers, Inc, 2000 17. Harris RP, Helfand M, Woolf SH, Lohr N, Mulrow CD, Teutsch SM, Atkins D, Methods Work Group, Third US Preventive Services Task Force: Current methods of the US Preventive Services Task Force: a review of the process. Am J Prev Med 20:21–35, 2001 18. Fletcher KE, Underwood W, Davis SQ, Mangrulkar RS, McMahon LF Jr, Saint S: Effects of work hour reduction of residents’ lives. A systematic review. JAMA 294:1088–1100, 2005 19. Hoffer EK, Sultan S, Herskowitz MM, Daniels ID, Sclafani SJA: Prospective randomized trial of a metallic intravascular stent in hemodialysis graft maintenance. J Vasc Interv Radiol 8:965–973, 1997 20. Beathard GA: Gianturco self-expanding stent in the treatment of stenosis in dialysis access grafts. Kidney Int 43:872–877, 1993
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY 21. Quinn SF, Schuman ES, Demlow TA, Standage BA, Ragsdale JW, Green GS, Sheley RC: Percutaneous transluminal angioplasty versus endovascular stent placement in the treatment of venous stenoses in patients undergoing hemodialysis: intermediate results. J Vasc Interv Radiol 6:851–855, 1995 22. Vogel PM, Parise C: Comparison of SMART stent placement for arteriovenous graft salvage versus successful graft PTA. J Vasc Interv Radiol 16:1619–1626, 2005 23. Maya ID, Allon M: Outcomes of thrombosed arteriovenous grafts: comparison of stents vs angioplasty. Kidney Int 69:934–937, 2006 24. Kakisis JD, Avgerions E, Giannakopoulos T, Moulaskakis K, Papapetrou A, Liapis CD: Balloon angioplasty vs nitinol stent placement in the treatment of venous anastomotic stenoses of hemodialysis grafts after surgical thrombectomy. J Vasc Surg 55:472–478, 2012 25. Bakken AM, Protack CD, Saad WE, Lee DE, Waldman DL, Davies MG: Long-term outcomes of primary angioplasty and primary stenting of central venous stenosis in hemodialysis patients. J Vasc Surg 45:776–783, 2007 26. National Vascular Access Improvement Initiative. http://www.fistulafirst.org/, accessed October 30, 2013 27. Chan MR: Interventional nephrology: what the nephrologist needs to know about vascular access. Clin J Am Soc Nephrol 8:1211–1212, 2013
7
28. Al-Jaishi AA, Oliver MJ, Thomas SM, Lok CE, Zhang JC, Garg AX, Kosa SD, Quinn RR, Moist LM: Patency rates of the arteriovenous fistula for hemodialysis: a systematic review and meta-analysis. Am J Kidney Dis 63:464–78, 2014 29. Clark TW: Nitinol stents in hemodialysis access. J Vasc Interv Radiol 15:1037–1040, 2004 30. Roy-Chaudhury P, Sukhatme VP, Cheung AK: Hemodialysis vascular access dysfunction: a cellular and molecular viewpoint. J Am Soc Nephrol 17:1112–1127, 2006 31. Duda SH, Wiskirchen J, Tepe G, Bitzer M, Kaulich TW, Stoeckel D, Claussen CD: Physical properties of endovascular stents: an experimental comparison. J Vasc Interv Radiol 11:645–654, 2000 32. Yevzlin AS: Update on RESCUE trial. Presented at the 10th Annual Scientific Meeting of the ASDIN, Phoenix AZ, 2014 33. Sutton AJ, Abrams Kr, Jones DR, Sheldon TA, Song F: Methods for Meta-Analysis in Medical Research. Somerset, NJ: John Wiley and Sons, 2000 34. Egger M, Smith GD, Phillips AN: Meta-analysis: principles and procedures. BMJ 315:1533–1537, 1997 35. Higgins JPT, Green S: Analysing and presenting results. Cochrane Handbook for Systematic Reviews of Interventions.5.0.1. Available at http://www.cochrane.org/resources/handbook/hbook.htm, accessed December 8, 2013
ABSTRACT Dysfunction of arteriovenous fistulas (AVF) and arteriovenous grafts (AVG) contribute significantly to morbidity and hospitalization in the dialysis population. Despite advances in endovascular techniques, the incidence of vascular access stenosis remains problematic. Currently, the role of endovascular stent placement in the treatment of vascular access stenosis is poorly defined. This metaanalysis compares the primary patency rates of stenotic vascular access treated with stent placement vs. angioplasty. We searched Medline for English language publications from 1980 through December 2013, along with national conference proceedings and reference lists of all included publications. Inclusion criteria were a measure of primary patency, secondary patency, or access dysfunction. Studies were excluded if they were not in English or if they included pediatric patients. Ten studies with a total of 860 subjects met the inclusion criteria, including six experimental studies and four observational
studies. There was significantly higher overall primary patency in those receiving stent placement than in those treated with angioplasty (pooled relative risk [RR] = 0.79; 95% confidence interval [CI]: 0.65–0.96). The estimate did not differ by study design. The effect of treatment differed significantly (p = 0.001) by the type of stents used, however. In studies including nitinol stents (six studies, 678 patients), 6-month patency was significantly better for stent placement than angioplasty (pooled RR = 0.67; 95% CI: 0.54–0.84), whereas there was no significant differences between stent placement and angioplasty in those studies using bare metal stents exclusively (four studies, 182 patients; pooled RR = 1.09; 95% CI: 0.91–1.32). There was significant heterogeneity between studies (I2 = 70.6%; p < 0.0001). Our results suggest that stent placement may confer an advantage over balloon angioplasty in primary patency of dialysis access stenoses.
Establishing and maintaining adequate vascular access continues to be a major challenge in dialysis. While arteriovenous fistulas (AVF) and grafts (AVG) are the preferred method of long-term access for hemodialysis, complications including thrombosis or stenosis often limit their duration of use. Significant resources are put into ensuring patency of AVF and AVG (1,2). Surgical interventions to extend the lifetime of AVF and AVG were largely supplanted in the 1980s by percutaneous transluminal angioplasty (PTA). While PTA provided a presumed improvement over surgical intervention, early studies dem-
onstrated a 1-year primary patency following PTA in AVF ranging between 31% and 64% (3–5). Even more concerning was a further decline in 1-year patency from a first angioplasty to a second angioplasty (41–22% respectively) (3). Similarly, poor rates of patency following PTA have been observed in AVG, with overall survival of less than 50% at 3 years (6–10). The use of endovascular stents for the treatment of failed dialysis access was first described in 1988 (10,11), and initially, it was used largely as a salvage technique in failed angioplasty or early recurrent stenosis. Initial studies with endovascular stents failed to show improved patency compared to angioplasty. However, as stent technology has advanced, several recent studies have demonstrated superior results using stents for failing dialysis grafts and fistulas (12–14). Increasing attention has been drawn toward the use of stents as a primary method of managing failed dialysis access. Haskal et al. first demonstrated in a multicenter, randomized trial, that primary placement of stents at the venous anastomosis of AVGs had significantly better
Portions of manuscript presented in abstract form at the National Kidney Foundation Spring Clinical Meeting 2014, Las Vegas, NV. Address correspondence to: Micah R. Chan M.D., MPH, University of Wisconsin Hospitals and Clinics, 3034 Fish Hatchery Road, Suite B, Madison, WI 53713, or e-mail: mr. [email protected]. Seminars in Dialysis—2014 DOI: 10.1111/sdi.12314 © 2014 Wiley Periodicals, Inc. 1
2
Fu et al.
patency vs. angioplasty alone (38% vs. 20%, p = 0.008) over a follow-up period of 6 months (13). Furthermore, preliminary results from the RENOVA postapproval, multicenter, randomized trial demonstrated a significant improvement over angioplasty on 12- and 24-month primary patency of venous anastomotic lesions in AVGs (15). Previous studies have important methodological limitations, including small sample sizes, retrospective design (14), and inclusion of AVG venous anastomosis only (13). Thus, the role of endovascular stent placement in the treatment of thrombosed and stenotic AVF and AVG remains controversial. To date, there has been no comprehensive review of the available literature. The objective of this study was to perform a meta-analysis of the available studies comparing primary patency after endovascular stent placement vs. angioplasty in AVF and AVG. Methods Literature Sources and Search Strategy We conducted a computerized search of the literature for clinical studies that addressed stent placement and angioplasty in dialysis access published during the period 1980 through December 2013 using MEDLINE. We used combinations of terms related to dialysis (dialysis, CKD, chronic kidney disease, end-stage renal disease, and ESRD) and vascular access (arteriovenous fistula, AVF, arteriovenous graft, and AVG). This strategy also combined four exploded Medical Subject Headings (stent placement, angioplasty, patency, and dialysis).
Fig. 1. Literature search strategy.
Further, we hand-searched the reference lists of all articles identified in the search as well as those of review articles. We also reviewed the abstracts of the American Society of Nephrology and National Kidney Foundation annual scientific meetings from 2008 through 2013. Study Selection Inclusion criteria were a measure of primary patency, secondary patency, or access dysfunction. It was decided, a priori, to exclude any of the following categories of articles: wrong topic, reviews, pediatric studies, and non-English language studies. From among 668 citations, we reviewed all relevant articles (Fig. 1). Of those, 72 were abstracts or articles appropriate for detailed review. Two independent authors (NF, EJ) reviewed each article from the initial search to determine if inclusion criteria were met. Disagreement was resolved by consensus and, when necessary, with arbitration from a third author (MRC). Data Abstraction Of the 72 studies reviewed in detail, 51 were eliminated due to non-English studies, case reports, pediatric studies, or reviews. Of the remaining 21 studies that assessed access patency, six were eliminated due to inestimable time factors, and one study with data unavailable. Finally, 10 studies were retained for further analysis. The data from each included article were abstracted by two authors (NF, EJ) using a standardized form
3
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY
which included patient number, demographics, control and study group information, study design and primary outcome of interest. Disagreements were resolved through consensus or third reviewer (MRC).
evidence. All studies that were identified provided second-tier evidence except the four randomized trials (13,19–21). Results
Statistical Analysis Kappa scores were used to assess interobserver agreement between the two reviewers. Analyses were performed for all studies and after stratification by study design (experimental or observational) and stent type (nitinol or bare metal exclusively). The outcomes were pooled using random effects models. Publication bias was assessed using funnel plots. Heterogeneity across studies was assessed using the I2 index. Stata/MP13 (Stata Corporation, College Station, Texas, USA) was used for all analyses. p-values < 0.05 were considered statistically significant. Assessment of Study Quality We defined studies as truly experimental only in instances in which the study and data collection were explicitly planned at the time of collation of the study cohort (16). Only in such circumstances would study follow-up procedures be expected to be conducted for the purposes of the study question. Thus, post hoc analysis of a prospective study cohort was considered observational. We assessed study quality by using a modification of the US Preventive Services Task Force criteria described by Fletcher et al. (17,18). These criteria assist in grading the internal validity of studies based on evidence-based guidelines. By these standards, properly designed randomized controlled trials constitute the highest tier evidence; well-designed nonrandomized trials, case–control, or cohort analytic studies provide second-tier
The literature search strategy is summarized in Fig. 1. The agreement between the two reviewers for articles, as well as citations excluded and articles identified, was excellent (j = 0.96; 95% confidence interval [CI]: 0.90–0.99). We identified 10 studies with a total of 860 subjects between 1992 and 2012. Six studies were classified as experimental, including four randomized (13,19–21) and two nonrandomized (6,22), that proceeded with intervention. The observational group comprised four studies (14,23–25) observing patients who had undergone intervention without a priori treatment assignment (Table 1). Vascular access type, stent type, lesions sites, and outcomes of interest are listed in Table 2. Seven of 10 studies compared stent placement vs. angioplasty for the treatment of stenotic AVGs. Most (8 of 10) studies focused on graft-vein anastomosis and venous outflow stenoses, while two studies included central venous stenoses (21,25). A total of six studies (678 patients) included nitinol stents, whereas four studies (182 patients) used bare metal stents exclusively. Primary Patency by Treatment Primary patency was lower in patients receiving angioplasty than in those treated with stent placement (RR = 0.79; 95% CI: 0.65–0.96; p = 0.02) (Fig. 2). The pooled relative risk was markedly similar in observational studies (RR = 0.78; 95% CI: 0.62–0.96; p = 0.02) and experimental studies (RR = 0.77; 95% CI: 0.55–1.09, p = 0.14).
TABLE 1. Summary of included studies Source
Primary location
Number of patients
Karnabatidis et al., 2012 (6) Haskal et al., 2010 (13)
Rion, Greece
35
Maryland, US
190
Vogel et al., 2005 (22)
California, US
60
Hoffer et al., 1997 (19)
Washington, US
37
Beathard et al., 1992 (20)
Texas, US
58
Quinn et al., 1995 (21)
Oregon, US
87
Chan et al., 2008 (14)
Wisconsin, US
Maya et al., 2006 (23) Kakisis et al., 2011 (24) Bakken et al., 2007 (25)
Study design
Cases treated years
Mean age (years, stent vs. PTA)
Male, number (%, stent vs. PTA)
04/2009–01/2011
67.814.3
45.7%
─
61.814.6 vs. 59.813.6 5313 vs. 5915
34% vs. 39%
211
Experimental, nonrandomized Experimental, randomized Experimental, nonrandomized Experimental, randomized Experimental, randomized Experimental, randomized Observational
Alabama, US Athems, Greece
48 61
Observational Observational
04/1999–03/2005 02/2007–12/2010
New York, US
73
Observational
1995–2003
06/2001–03/2003 01/1994–08/1996
32% vs. 42.9%
─
50.919.0 vs. 54.521.7 ─
29.4% vs. 15% ─
over 3 years
58 (26 - 84)
47%
01/2005–06/2007
62.518.1 vs. 63.914.7 5317 vs. 5813 73(39–84) vs. 72(40–84) 5715 vs. 5718
67.1% vs. 67.5% 28% vs. 38% 49% vs. 39% 35% vs. 45%
4
Fu et al. TABLE 2. Summary of vascular access type, lesion site, stent type, and access outcome of included studies
Source
Proportion of AVG
Karnabatidis et al. (6)
100%
Haskal et al. (13)
100%
Vogel et al. (22)
100%
Stenoses at graft-vein anastomosis and venous outflow
Nitinol—SMART nitinol
Hoffer et al. (19)
100%
Bare metal—Wallstent
Beathard et al. (20)
100%
Quinn et al. (21)
99%
Stenoses at or within 3 cm of the vein-graft junction or in the peripheral outflow vein Stenotic lesions at graft-vein anastomosis All sites of venous stenoses and occlusions
Chan et al. (14)
47%
Nonthrombosed, noncentral venous stenoses
Nitinol—SMART nitinol
Maya et al. (23)
100%
Stenoses at venous anastomosis after thrombectomy
Mixed—SMART, Protege, Wallstent, Fluency
Kakisis et al. (24)
100%
Nitinol—absolute nitinol
Bakken et al. (25)
100%
Stenoses at venous anastomosis of prosthetic brachial-axillary accesses after thrombectomy Stenoses at central venous sites
Lesion sites
Stent
Recurrent anastomotic and/or venous outflow stenoses in prosthetic AVGs Venous anastomotic stenosis
Six-month patency was significantly better for stent placement than angioplasty in studies that included nitinol stents (pooled RR = 0.67; 95% CI: 0.54–0.84) (Fig. 3). In contrast, there were no significant differences between stent placement and angioplasty in those studies using bare metal stents exclusively (pooled RR = 1.09; 95% CI: 0.91–1.32). Two studies (21,25) that used stents in central venous stenosis showed a RR above 1. Six of seven studies with a RR below 1 focused on anastomostic and venous outflow stenoses. Test of Homogeneity There was significant overall heterogeneity across studies with respect to the primary patency rate outcomes (I2 = 70.6%; p < 0.0001). The observational studies had moderate heterogeneity, while the experimental studies had significant heterogeneity. Due to a limited number of studies with significant differences in patient risk factors across study populations, the use of meta-regression to assess potential sources of heterogeneity was not undertaken. There was no publication bias identified throughout the studies using funnel plots.
Outcome of interest
Nitinol—Fluency Plus
Primary patency rates (6 and 12 months)
Nitinol—Flair
Primary patency rates of the treatment area and the entire vascular access circuit (6 months); incidence of binary restenosis and other adverse events Primary patency rates (3, 6, and 12 months); lower postprocedural midgraft systolic pressures ratio; complication rates Primary patency rates (30, 60, 180, and 360 days); secondary patency rates Primary patency rates (DE) (30, 60, 90, 180, and 360 days) Primary patency rates (60, 180, and 360 days), secondary patency rates Primary assisted patency rates (30, 90, and 180 days); access flow and urea reduction ratio Immediate postprocedure graft systemic pressure ratio; primary patency rates (30, 90, and 180 days); assisted or secondary patency rates Primary patency rates (3, 6, and 12 months); primary-assisted and secondary patency rates Primary patency rates (30 days and 12 months); assisted primary patency rates
Bare metal—Gianturco Bare metal—Gianturco, Palmaz, Wallstent
Mixed—Wallstent except one SMART stent
Discussion There are considerable challenges facing arteriovenous access creation, maintenance, and salvage (26–28). Although stent placement is the standard of care in percutaneous coronary and peripheral arterial interventions, its role in dialysis access has been debated since the first clinical trials, some of which are included in our analyses. The first rigorous clinical trial to study stent placement in dialysis access was conducted by Beathard et al. (20). Although this study included only 58 patients, the results demonstrated no significant difference between stent vs. angioplasty in any parameter before treatment, after intervention, or patency rates with 90-, 180-, and 360-day survival of 85% vs. 79%, 72% vs. 64%, and 17% vs. 28%, respectively (p > 0.07). In the last decade, nitinol and SMART (shape memory alloy recoverable technology) stents have been used more, which have advantages over bare metal stents in that there is better wall contact, radial strength, and superelasticity (29–31). Vogel, et al. showed in their prospective, nonrandomized trial of 60 patients, a significant decrease in restenosis in the stent group as compared with angioplasty (7% vs. 16%; p = 0.001) and an
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY
5
Fig. 2. Six-month primary patency rates comparing experimental and observational studies.
Fig. 3. Six-month primary patency rates comparing different stent types.
improvement in mean primary graft patency (5.6 vs. 8.2 months; p = 0.05) (22). The observational trials in our pooled analysis all used nitinol stent technology and demonstrated a significant advantage of stent placement as compared to angioplasty alone in 6-month primary patency. More recently, the largest and most robust clinical trial ever published in dialysis access intervention showed a significant advantage of the ePTFE stent graft vs. angioplasty for the primary treatment
of venous anastomotic lesions of AVGs (13). The FLAIR Pivotal trial randomized 190 patients to PTA alone or PTA/Flair Stent Graft. At 6 months, the incidence of patency of the access circuit was significantly greater in the stent-graft group than in the balloon angioplasty group (38% vs. 20%, p = 0.008). The angioplasty group had more instances of restenosis (p < 0.001) and the incidence of freedom from subsequent interventions at 6 months was significantly greater in the stent-graft
6
Fu et al.
group than in the balloon angioplasty group (32% vs. 16%, p = 0.03). The postapproval RENOVA trial examined 270 patients randomized to angioplasty alone vs. Flair Stent Graft at the venous anastomosis of AVGs (15). The interim analysis of the study showed 12-month access patency of 24% vs. 11% (p = 0.007) and 24-month patency of 9.5% vs. 5.5% (p = 0.011) in the stent graft vs. angioplasty groups, respectively. Stenosis requiring reintervention was also greater in the angioplasty group as compared to the stent group (82% vs. 60%; p < 0.001) over 12 months. More recently, Yevzlin et al. presented interim results of the RESCUE trial which was a multicenter, randomized trial of angioplasty vs. stent grafts for in-stent restenosis (32). These results similarly showed superiority of stent grafts as compared to angioplasty over 6 months of postintervention lesion patency (65% vs. 10%, p < 0.001). When AVF and AVG were analyzed separately, the advantage of stent grafts persisted. The interim results of the latter and former trials were not included in the pooled analysis because the followup of patients likely are not complete. The current body of evidence, however, suggests that stent grafts may have an advantage even more durable than uncovered nitinol stents. However, even with all of the compelling evidence pointing toward stents as more than a salvage technique, it must be tempered by the fact that care should be taken not to jeopardize the vein for a future AVF placement and that option must be considered before placing a stent. This study has many limitations. First, we identified significant heterogeneity, implying a significant variation in effect across studies. We recognize this as an inherent deficiency given the varying types of stents used in the studies, type of access (AVF, AVG), and lesions sites. We employed the random effects model that considers intrinsic random between study variance of effect besides variance due to sampling error or within-study variance (33,34). Moreover, differences between studies in key methodological issues, such as variation in types of stents, can be expected to lead to differences in the observed effects. Statistical heterogeneity due to such methodological variation may not necessarily imply that the actual intervention effect differs (35). Moreover, only four randomized clinical studies were included. The predominance of nonrandomized studies could lead to bias due to unrecognized differences between groups. Conclusion In conclusion, our meta-analysis suggests significantly higher primary patency rates in stenotic AV access receiving stent placement vs. angioplasty. Due to the limited and heterogeneous data currently available, it supports that stent placement, especially with nitinol uncovered stents and covered stent grafts, may extend access patency compared to
angioplasty alone in the treatment of graft-vein anastomotic and venous outflow stenoses. Whether there are significant benefits of stent placement or angioplasty in arteriovenous fistulas or secondary patency requires further robust studies. Further investigation in regard to differences in rates of infection, number of interventions, and cost analyses are warranted as secondary outcomes in future trials. Overall, these findings have significant implications for future design of clinical trials in stent placement in dialysis access and the delivery of dialysis-related services. Disclosures No conflict of interest reported by authors. References 1. United States Renal Data System. Annual data report, 2013. Available at http://www.usrds.org/adr.htm, accessed June 22, 2013 2. Feldman HI, Kobrin S, Wasserstein A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 7:523–535, 1996 3. Kanterman RY, Vesely TM, Pilgram TK, Guy BW, Windus DW, Picus D: Dialysis access grafts: anatomic location of venous stenosis and results of angioplasty. Radiology 195:135–139, 1995 4. Glanz S, Gordon DH, Butt KM, Hong J, Lipkowitz GS: The role of percutaneous angioplasty in the management of chronic hemodialysis fistulas. Ann Surg 206:777–781, 1987 5. Beathard GA: Percutaneous transluminal angioplasty in the treatment of vascular access stenosis. Kidney Int 42:1390–1397, 1992 6. Karnabatidis D, Kitrou P, Spiliopoulos S, Katsanos K, Diamantopoulos A, Christeas N, Siablis D: Stent-grafts versus angioplasty and/or bare metal stents for failing arteriovenous grafts: a cross-over longitudinal study. J Nephrol 26:389–395, 2013 7. Bittl JA: Catheter interventions for hemodialysis fistulas and grafts. JACC Cardiovasc Interv 3:1–11, 2010 8. Vascular Access Work Group: Clinical practice guidelines for vascular access. Am J Kidney Dis 48(Suppl):S248–S273, 2006 9. Schwab SJ: Vascular access for hemodialysis. Kidney Int 55: 2078–2090, 1999 10. Zollikofer CL, Antonucci F, Stuckmann G, Mattias P, Bruhlmann WF, Salomonowitz EK: Use of the Wallstent in the venous system including hemodialysis-related stenoses. Cardiovasc Intervent Radiol 15:334–341, 1992 11. Gunther RW, Vorwerk D, Klose KC, Bohndorf K, Kistler D, Mann H, Sieberth HG: Self-expanding stents for the treatment of a long venous stenosis in a dialysis shunt: case report. Cardiovasc Intervent Radiol 12:29–31, 1989 12. Schwab SJ: Hemodialysis vascular access: the Achilles’ heel remains. Kidney Int 72:665–666, 2007 13. Haskal ZJ, Trerotola S, Dolmatch B, Fletcher KE, Underwood W 3rd, Davis SQ, Mangrulkar RS, McMahon LF Jr, Saint S: Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med 362:494–503, 2010 14. Chan MR, Bedi S, Sanchez RJ, Young HN, Becker YT, Kellerman PS, Yevzlin AS: Stent placement versus angioplasty improves patency of arteriovenous grafts and blood flow of arteriovenous fistulae. Clin J Am Soc Nephrol 3:699–705, 2008 15. Saad TF: Update on RENOVA trial. Presented at the 10th Annual Scientific Meeting of the ASDIN, Phoenix AZ, 2014 16. Szklo M, Nieto J: Epidemiology: Beyond the Basics. Gaithersburg, MD: Aspen Publishers, Inc, 2000 17. Harris RP, Helfand M, Woolf SH, Lohr N, Mulrow CD, Teutsch SM, Atkins D, Methods Work Group, Third US Preventive Services Task Force: Current methods of the US Preventive Services Task Force: a review of the process. Am J Prev Med 20:21–35, 2001 18. Fletcher KE, Underwood W, Davis SQ, Mangrulkar RS, McMahon LF Jr, Saint S: Effects of work hour reduction of residents’ lives. A systematic review. JAMA 294:1088–1100, 2005 19. Hoffer EK, Sultan S, Herskowitz MM, Daniels ID, Sclafani SJA: Prospective randomized trial of a metallic intravascular stent in hemodialysis graft maintenance. J Vasc Interv Radiol 8:965–973, 1997 20. Beathard GA: Gianturco self-expanding stent in the treatment of stenosis in dialysis access grafts. Kidney Int 43:872–877, 1993
META-ANALYSIS OF STENT PLACEMENT VS. ANGIOPLASTY 21. Quinn SF, Schuman ES, Demlow TA, Standage BA, Ragsdale JW, Green GS, Sheley RC: Percutaneous transluminal angioplasty versus endovascular stent placement in the treatment of venous stenoses in patients undergoing hemodialysis: intermediate results. J Vasc Interv Radiol 6:851–855, 1995 22. Vogel PM, Parise C: Comparison of SMART stent placement for arteriovenous graft salvage versus successful graft PTA. J Vasc Interv Radiol 16:1619–1626, 2005 23. Maya ID, Allon M: Outcomes of thrombosed arteriovenous grafts: comparison of stents vs angioplasty. Kidney Int 69:934–937, 2006 24. Kakisis JD, Avgerions E, Giannakopoulos T, Moulaskakis K, Papapetrou A, Liapis CD: Balloon angioplasty vs nitinol stent placement in the treatment of venous anastomotic stenoses of hemodialysis grafts after surgical thrombectomy. J Vasc Surg 55:472–478, 2012 25. Bakken AM, Protack CD, Saad WE, Lee DE, Waldman DL, Davies MG: Long-term outcomes of primary angioplasty and primary stenting of central venous stenosis in hemodialysis patients. J Vasc Surg 45:776–783, 2007 26. National Vascular Access Improvement Initiative. http://www.fistulafirst.org/, accessed October 30, 2013 27. Chan MR: Interventional nephrology: what the nephrologist needs to know about vascular access. Clin J Am Soc Nephrol 8:1211–1212, 2013
7
28. Al-Jaishi AA, Oliver MJ, Thomas SM, Lok CE, Zhang JC, Garg AX, Kosa SD, Quinn RR, Moist LM: Patency rates of the arteriovenous fistula for hemodialysis: a systematic review and meta-analysis. Am J Kidney Dis 63:464–78, 2014 29. Clark TW: Nitinol stents in hemodialysis access. J Vasc Interv Radiol 15:1037–1040, 2004 30. Roy-Chaudhury P, Sukhatme VP, Cheung AK: Hemodialysis vascular access dysfunction: a cellular and molecular viewpoint. J Am Soc Nephrol 17:1112–1127, 2006 31. Duda SH, Wiskirchen J, Tepe G, Bitzer M, Kaulich TW, Stoeckel D, Claussen CD: Physical properties of endovascular stents: an experimental comparison. J Vasc Interv Radiol 11:645–654, 2000 32. Yevzlin AS: Update on RESCUE trial. Presented at the 10th Annual Scientific Meeting of the ASDIN, Phoenix AZ, 2014 33. Sutton AJ, Abrams Kr, Jones DR, Sheldon TA, Song F: Methods for Meta-Analysis in Medical Research. Somerset, NJ: John Wiley and Sons, 2000 34. Egger M, Smith GD, Phillips AN: Meta-analysis: principles and procedures. BMJ 315:1533–1537, 1997 35. Higgins JPT, Green S: Analysing and presenting results. Cochrane Handbook for Systematic Reviews of Interventions.5.0.1. Available at http://www.cochrane.org/resources/handbook/hbook.htm, accessed December 8, 2013
