Letters Criticize Study Supporting Stent Grafts for Dialysis-Access Key Points:
* NEJM paper spurs critical letters regarding positive stent graft results in patients with failing dialysis-access grafts
* Letters question choice of endpoints, follow-up protocol
* Study findings need independent confirmation
By Caitlin E. Cox
Wednesday, May 19, 2010
In reaction to a paper in the February 11, 2010, issue of the New England Journal of Medicine showing the effectiveness of stent graft placement in patients with kidney disease who have failing dialysis-access grafts, several letters in the May 20, 2010, issue of NEJM raise numerous methodological problems with the study. The correspondence authors question the clinical relevance of the study’s endpoints and the possible effects of its follow-up protocol.
John A. Bittl, MD, of the Ocala Heart and Vascular Institute (Ocala, FL), told TCTMD in a telephone interview that the letters appropriately “raised concerns about trial design and questioned whether the results of the study were protocol driven.” Moreover, some interventional nephrologists have been made uneasy by financial ties between the device manufacturer and the research, he reported.
Better Outcomes with Stent Graft
For the study, researchers led by Ziv J. Haskal, MD, of the University of Maryland Medical Center (Baltimore, MD), randomized 190 hemodialysis patients who had venous anastomotic stenosis to receive balloon angioplasty with or without stent graft placement. Based on intention to treat, procedural success was significantly better in the stent graft group. While primary patency of the treatment area and access circuit were similar between the 2 groups at 2-month follow-up, stent grafting produced significantly better patency by 6 months. Freedom from reintervention and binary restenosis were also improved at 6 months.
The graft is a self-expanding nitinol stent covered in carbon-impregnated ePTFE (Flair Endovascular Stent Graft, Bard Peripheral Vascular, Tempe, AZ). It improves flow dynamics and acts as a barrier against neointimal tissue growth, Dr. Haskal told TCTMD in a telephone interview in February, adding that its implantation does not require interruption of dialysis. The device was approved for marketing by the US Food and Drug Administration based on the trial’s findings in July 2007, prior to publication.
Doubts Over Study Design
Michael Allon, MD, of the University of Alabama at Birmingham (Birmingham, AL), writes in his letter that the lower frequency of restenosis at 6 months in patients treated with stent grafts rather than angioplasty alone (40% vs. 77%; P < 0.001) may be “attributable to the scheduled angiograms at 2 and 6 months after enrollment, which may not have been obtained on the basis of clinical indications.”
Oculostenotic reflex seems a likely explanation for the paper’s main finding, Dr. Bittl confirmed. “Figure 1 in the original article shows this abrupt drop in the patency rates in the control group. That coincided in time with this protocol-mandated angiographic, unblinded 60-day follow-up assessment,” he said, adding that that the investigators did not handle the issue completely in their NEJM rebuttal.
Dr. Allon commented that the “ultimate goal” of angioplasty performed preemptively before clinical symptoms develop should not be to treat stenosis but rather to prevent graft thrombosis and improve graft survival. On the contrary, when Dr. Haskal and colleagues performed their mandatory angiographic follow-up, they found no difference in thrombosis (33% with stent grafts vs. 21% with balloon angioplasty alone; P = 0.1). “Unfortunately, the authors did not report on cumulative graft survival,” Dr. Allon added.
Graft loss, in which a stent graft becomes unusable and must be surgically replaced, is indeed a clinically relevant endpoint and perhaps the most important outcome to measure, Dr. Bittl concurred. “I hate to say it but you can keep working on a specific graft or fistula for a patient 10 or 20 times. As long as that particular access is still functioning, that patient’s prognosis is better. But if you have to move to a new site in a dialysis patient, you then start consuming vein and will actually shorten that patient’s survival time on dialysis,” he said, emphasizing that graft loss is a critical measure that seems to be missing both from the study and the investigators’ reply.
Cost, Benefits Must Be Balanced
Dr. Bittl, too, was concerned about the trend toward higher rates of thrombosis with stent grafts. Only 2 or 3 more cases, or a slightly larger study population, would have translated into a significant difference, he stressed, adding, “When you have access thrombosis, it doubles the procedure cost and reduces long-term patency, so it’s a big issue.”
In another letter, John P. Killen, MB, BS, Murty Mantha, MD, and Richard Baer, MB, BS, of Cairns Base Hospital (Cairns, Australia), also express qualms about the lack of difference in thrombosis rates. Moreover, given the aging dialysis population, a cost-benefit analysis taking into account baseline characteristics and prognosis would be especially relevant for this indication, they advise.
Cost is indeed relevant, Dr. Allon concludes, stating that without proven benefits with respect to clinically meaningful endpoints such as graft thrombosis and longevity, “routine deployment of graft stents for stenosed arteriovenous grafts is expensive and cannot be recommended.”
Researchers Vouch for Clinical Relevance
But the study investigators “strongly disagree” with the notion that the study shows no proven benefit for clinical endpoints. Dr. Haskal and Scott Trerotola, MD, of the Hospital of the University of Pennsylvania (Philadelphia, PA), respond to the letters by emphasizing the finding that freedom from subsequent intervention at 7 months was significantly better after stent graft placement than after balloon angioplasty alone (32% vs. 26%; P = 0.03).
“Having fewer painful, costly additional interventions is very important to patients and no doubt to payers. Similarly, doubling of access-circuit patency and higher technical success rates are important,” they write. “Indeed, a cost analysis will take into account the angioplasty costs related to a higher rate of interventions, including added professional and technical fees, lost days at work, ineffective or interrupted dialysis, and added procedural costs.”
Yet the trialists also concede that longer access life is an important endpoint. “We have since seen many anecdotal cases in which anastomoses that were treated with a stent graft have remained patent for years after therapy, suggesting a potential opportunity to view grafts as far longer-term tools in appropriate patients,” they report.
Baseline Bias?
Two other letters, meanwhile, point out baseline factors that could have influenced results.
Jan Malik, MD, PhD, and Vladimir Tuka, MD, PhD, of Charles University in Prague (Prague, Czech Republic), question whether the 2 patient groups could have had different rates of angioplasty performed prior to study inclusion. Moreover, only 18% of balloon angioplasty patients and 25% of stent graft patients showed declines in access flow at baseline, meaning that these patients perhaps had only borderline stenosis.
Faruk Tomak, MD, and Bernd Krüger, MD, of Ruhr University Bochum (Bochum, Germany), and Bernard K. Krämer, MD, of Mannheim University Hospital (Mannheim, Germany), further question whether the mode of anticoagulation could have been a factor.
Although the use of drugs such as oral anticoagulant agents, platelet inhibition, and heparin were similar at baseline, “no information is provided regarding the use of anticoagulants before, during, and immediately or long after the intervention,” they write in their letter. Any difference might be an important confounder, they say, because dipyridamole plus aspirin has been shown to improve patency after placement of a new arteriovenous hemodialysis graft and clopidogrel has been found to prevent fistula thrombosis after new fistula creation.
The study investigators reassure that rates of previous intervention were equivalent between the 2 treatment groups but do not directly address the low prevalence of patients with declines in access flow at baseline. As for anticoagulant and antiplatelet protocols, the “use of these medications was wholly at the discretion of the investigators,” they respond.
Financial Ties Cloud Study’s Reception
Dr. Bittl also related to TCTMD how the paper had been received since its publication. “Some interventional nephrologists who I’ve spoken with have concerns about the conflicts of interest and industry ties of the investigators,” he said. “In their society meetings, it’s pretty apparent that there are conflicts.” Not only was the study supported by device manufacturer Bard Peripheral Vascular but, based on the NEJM disclosures, Dr. Haskal receives lecture fees and Dr. Trerotola receives consulting fees from the company. Most of the remaining coauthors also receive consulting and/or lecture fees from Bard. One served as an expert witness about the device’s “healing characteristics.”
Currently, the RENOVA study is evaluating the Flair Endovascular Stent Graft through 24-month follow-up. It will provide both graft-life endpoints and assess patency function, Drs. Haskal and Trerotola report.
While such a study would be valuable, it would still have the same potential conflicts, because it is being done by the same investigators, Dr. Bittl countered. “An independent confirmation would be welcome,” he said, reporting that the Gore Viabahn Endoprosthesis (WL Gore and Associates, Flagstaff, AZ) is also being tested for this indication.
Sources:
1. Haskal ZJ, Trerotola S, Dolmatch, B, et al. Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med. 2010;362:494-502.
2. Malik J, Tuka V. Stent graft or balloon angioplasty alone for dialysis-access grafts [letter to the editor]. N Engl J Med. 2010;362:1938-1939.
3. Tokmak F, Krämer BK, Krüger B. Stent graft or balloon angioplasty alone for dialysis-access grafts [letter to the editor]. N Engl J Med. 2010;362:1939.
4. Allon M. Stent graft or balloon angioplasty alone for dialysis-access grafts [letter to the editor]. N Engl J Med. 2010;362:1939.
5. Killen JP, Mantha M, Baer R. Stent graft or balloon angioplasty alone for dialysis-access grafts [letter to the editor]. N Engl J Med. 2010;362:1939-1940.
6. Haskal ZJ, Trerotola S. Stent graft or balloon angioplasty alone for dialysis-access grafts [author reply]. N Engl J Med. 2010;362:1940.
Disclosures:
* The study was supported by Bard Peripheral Vascular.
* Dr. Haskal reports receiving consulting fees from WL Gore and Associates, lecture fees from Bard Peripheral Vascular, and holding stock in AngioDynamics.
* None of the letter authors report any relevant conflicts of interest.
* Dr. Bittl reports no relevant conflicts of interest.
Related Story:
* Stent Grafts a Better Fix for Failing Dialysis-Access Grafts
http://www.tctmd.com/Show.aspx?id=88670http://www.tctmd.com/show.aspx?id=90080
I Hate Dialysis Message Board
Author
