#NephJC Chat
Tuesday Sep 8 9 pm Eastern
Wednesday Sep 9 9 pm IST
Wednesday Sep 9 9 pm BST
N Engl J Med. 2020 Aug 20;383(8):733-742.
doi: 10.1056/NEJMoa1914617.
Drug-Coated Balloons for Dysfunctional Dialysis Arteriovenous Fistulas
Robert A Lookstein, Hiroaki Haruguchi, Kenneth Ouriel, Ido Weinberg, Lanyu Lei, Stephanie Cihlar, Andrew Holden, IN.PACT AV Access Investigators
PMID: 32813949
Introduction
Following observations from traumatic arteriovenous fistulas caused by mechanical injuries, Dr. Brescia, Cimino and Hurwich first described the creation of a radial cephalic AVF in the forearm forming a side-to-side anastomosis between the cephalic vein and radial artery in 1966; thereby revolutionizing care for hemodialysis patients. The AVF provided an AV access with higher flow rate, better patency and longer survival. See this year’s NephMadness scouting report for more on the historical aspects and developments.
From Brescia et al, NEJM 1966
…of all AVFs placed between June 2014 and May 2016, 38.9% failed to mature sufficiently for use in dialysis…
While present day best practices for HD access vary from patient to patient, observational studies have shown that AVFs, compared to central venous catheters and arteriovenous grafts, have longer patency and access survival, and fewer complications including infection, stenosis and thrombosis. Despite the introduction of new fistula surgical techniques and locations over the last fifty-four years, AVFs are mired by high primary failure rates. In fact, USRDS data shows that of all AVFs placed between June 2014 and May 2016, 38.9% failed to mature sufficiently for use in dialysis (USRDS 2018, PDF link). It has also been shown that AVFs that require multiple interventions to assist with maturation have poor long term access patency (Harms et al. J Vasc Surg 2016). Additionally, the high the failure rate of dialysis arteriovenous access due to vascular stenosis remains pervasive, contributing to significant morbidity and mortality; including but not limited to missed or inadequate treatments, hospitalization, incident catheter use, infectious complications and death (USRDS 2010, PDF link).
Vascular stenosis occurs due to endothelial and smooth muscle injury resulting in neointimal hyperplasia. Percutaneous transluminal angioplasty (PTA) is the standard treatment for AV access stenosis, but unfortunately it causes vessel wall trauma further contributing to intimal hyperplasia and restenosis. In this scenario, the intervention itself contributes to the pathophysiologic process implicated in AVF stenosis. Not surprisingly, balloon angioplasty is associated with poor patency rates and high-reintervention rates. In a systematic review and meta-analysis, the primary AVF patency rates post-balloon angioplasty were 50% at 6 months, and 20% at 12 months (Wee et al. J Vasc Surg 2019).
More recent data shows that the annual vascular access costs in the United States are approximately $2.8 billion annually (Thamer et al, AJKD 2018). Vascular access related complications and resultant interventions cause significant morbidity and mortality. Therefore, undoubtedly there is a need for innovation in this arena to improve AVF outcomes.
In addition to standard balloon angioplasty, endovascular stents have shown promise in improving patency for some anatomic locations in AV accesses, such as the AVG venous anastomosis. (Haskal et al. NEJM 2010, Vesely et al. J Vasc Surg 2016). However, balloon angioplasty remains the standard of care and the first line treatment for most AV access stenosis.
The percutaneous endovascular AVF (endoAVF) techniques utilizing the WavelinQ and Ellipsys systems introduced over the last 5 years have demonstrated promising results, yielding lower AVF primary failure rates and lower reintervention rates. (NEAT & PIVOTAL trials; Hull et al. J Vasc Interv Radiol 2018, Lok et al. AJKD 2017, Yang et al. J vasc Access 2017). Unfortunately, not all patients fulfill the criteria required to qualify for endovascular AVF creation. Additionally, endovascular AVF outcomes have not been directly compared to surgical AVFs. Hence, endovascular AVFs provide additional vascular access options for patients but are not a replacement for surgical AVFs. See the 2020 NephMadness scouting report for more on this topic.
To combat neointimal hyperplasia and restenosis, the concept of a drug coated balloon (DCB) was introduced two decades ago with the first animal trial beginning in 2002. Paclitaxel crystals, derived from a cytotoxic agent that blocks the cell cycle during mitosis, are loaded onto the DCB and delivered directly to the vessel wall during balloon deployment. The crystalline formulation has a half-life of weeks to months, delivering prolonged vessel exposure to the drug, inhibiting the smooth muscle remodeling and fibrosis that contributes to stenosis. The first human trial of DCB began in 2003. DCBs have been used successfully in the treatment of coronary and peripheral artery disease for the revascularization of stenotic lesions (Scheller et al, NEJM 2006, Rosenfield et al. NEJM 2015).
The efficacy data for DCB use for stenotic lesions in dialysis vascular access has been mixed. Multiple single-center and small randomized control trials have demonstrated safe and effective use of DCB for AVF stenosis at 6 and 12 months post-intervention. See summary studies and outcomes of DCB use in dialysis vascular access below (Wee et al. J vasc Surg. 2019).
From Wee et al, J Vasc Surgery 2019
From Wee et al, J Vasc Surgery 2019
However, a large RCT published in 2018, failed to demonstrate a difference in outcomes with DCB use versus standard balloon angioplasty, fueling skepticism regarding the role of DCB for AVF stenosis, see visual abstract below (Trerotola et al. CJASN 2018).
From Tretola et al, CJASN 2018
Some have suggested that the lower Paclitaxel dose in DCB balloons (2.2 μg per mm²) used in the CJASN study could be the plausible explanation for the negative study results. However, due to conflicting trial results, it was clear that a large RCT was needed to settle the debate regarding DCB use in AVF, which brings us to the study under discussion today.
The Study
Methods
This study was a prospective, global, multicenter, single-blinded 1:1 randomized controlled trial comparing outcomes between DCB (Medtronic) and standard non-drug-coated balloons in the treatment of new (not previously treated) or nonstented restenotic AVF lesions (up to 100mm in length). There were 29 sites involved; 204 in the US, 112 in Japan and 14 in New Zealand. Due to macroscopic differences between the balloons, investigators and research coordinators were aware of treating assignments.
Study population
Inclusion criteria:
> 21 years old
New or nonstented restenotic native AVF with at least 50% stenosis
AVF required to be created at least 60 days before index procedure
AVF must have been used for dialysis for at least 8 of 12 sessions during a 4 week period, ensuring fistula maturity. (Important: as they did not include AVF with primary failure - meaning AVF that failed to mature)
Lesions up to 100mm in length in AVFs
Exclusion criteria:
History of or current access-circuit thrombosis
History of previous stent in the arteriovenous circuit
Patients with hemodynamically significant central venous stenosis
Procedure:
Trial Device: Medtronic AV drug-coated balloon impregnated with 3.5μg per mm² of Paclitaxel.
Predilation: All eligible participants underwent predilation with high-pressure balloon matching the reference vessel diameter utilizing 1:1 sizing. Duration of balloon inflation and successful predilation was judged by the interventionist. However, residual stenosis was expected to be no more than 30% of vessel diameter. At this point, patients were enrolled in the study.
Balloon inflation:
Drug-coated balloon: A length was selected to exceed the target lesion by 10mm at either end of the stenotic lesion to ensure full coverage. If more than one balloon was required, a 10mm overlap was required.
Standard balloon: Interventionalists followed individual device instructions for use.
Residual restenosis: If residual stenosis remained >30% or flow limiting dissection was present, additional dilation with standard balloon was performed.
Follow-up: Follow-up assessments were performed at 30 days, 3 months and 6 months after index procedure. Duplex was performed at 30-day and 6 month time points. Angiographic follow-up was performed at any point in time if clinically indicated.
Endpoints:
Primary effectiveness: Target lesion primary patency as defined by freedom from clinically driven target lesion revascularization or access-circuit thrombosis at 6 months after the index procedure. A clinically driven target lesion revascularization was determined if
the target lesions had at least 50% stenosis or the diameter of the vessel in the setting of clinical of physiology abnormalities to suggest access dysfunction OR
At Least 70% stenosis.
Primary safety: Any serious adverse event involving the AVF access circuit within 30 days after the index procedure.
Statistical Analysis:
Power: The study calculated that planned enrollment of 330 participants would provide 92% power for SUPERIORITY for the primary effectiveness end point, assuming 60% of the DCB group and 40% standard group would meet the endpoint with a 15% attrition rate at 6 months. Simultaneously, it would provide 80% power to show NON-INFERIORITY for the primary safety endpoint, assuming that 5% would have serious adverse events and a 2% attrition rate at 30 days. Only after the 2 primary endpoints were significant would a number of secondary endpoints be tested with a hierarchical ordering.
Funding: From the paper, “Medtronic sponsored the trial and owns the data. Data analyses were performed by Medtronic and the Baim Institute for Clinical Research (formerly the Harvard Clinical Research Institute). The authors had unrestricted access to the data; one academic author not employed by the sponsor and two authors employed by the sponsor wrote the manuscript”
Results
There were 330 participants, 170 randomized to the DCB group and 160 to the standard balloon group.
Figure 1 from Lookstein et al, NEJM 2020
Table 1 from Lookstein et al, NEJM 2020
Demographics and baseline characteristics are shown in Table 1.
Baseline participant dialysis characteristics revealed the distribution of stenotic lesions to be radiocephalic 50%, brachiocephalic 37% and brachiobasilic 10%. The most common presenting clinical symptoms were decreased blood flow, abnormal physical findings (thrill, murmur, arm swelling, etc.) and elevated venous pressures. Not surprisingly, close to 75% of participants had undergone previous peripheral arteriovenous access circuit revascularizations. (Table S3)
Table S3 from Lookstein et al, NEJM 2020
Baseline lesion characteristics were similar between groups. Of them, approximately 70% were considered restenotic. The most common target lesions were identified in the venous outflow (32%) and anastomosis (25.5). (Table S4)
Table S4 from Lookstein et al, NEJM 2020
Procedural characteristics were similar between the groups. However, the DCB group had a greater length of balloon.
Primary effectiveness endpoint: Primary patency
At 6 months, the primary patency was 82% in the DCB group and 60% in the standard-ballon group with a risk difference of 23% (CI 12.8-32.8; p<0.001). These conclusions were consistent when the effect of missing data was considered.
Figure 2A from Lookstein et al, NEJM 2020
Primary safety endpoint:
Considering a serious adverse event involving the AVF circuit within 30 days from index procedure, the DCB was found to be noninferior to the standard balloon. These conclusions remained consistent when the effect of missing data was considered.
Table 3 from Lookstein et al, NEJM 2020
Key secondary endpoints:
The primary patency of the entire dialysis circuit (arterial inflow to venous outflow) during the 6 month follow-up period was 73% in the DBG group and 48% of the standard-balloon group with a risk difference of 25% (CI 95% 14.6-35.9; <0.001). Finally, over the course of 12 months, mortality rate was 9.4% in the DCB group and 9.6% in the standard-ballon group (P=0.93).
Figure 2B from Lookstein et al, NEJM 2020
Discussion
In 2003, CMS and ESRD networks implemented the ‘Fistula First Breakthrough Initiative’ which set a target for 40% AVF use in prevalent HD patients. In 2005, the target of 40% AVF use in prevalent HD patients was attained, and CMS established a new goal of 66% AVF use by 2009. Currently, 62.8% of prevalent HD hemodialysis patients use an AVF in the United States. (USRDS 2018)
More recently, a greater emphasis has been placed on choosing the Right Access for the Right Patient as AVF may not be the best AV access for all hemodialysis patients. Having said that, based on observational studies, AVFs are associated with better outcomes compared to AVGs or central venous catheters.
Unfortunately, high AVF primary failure rate and the need for repeated interventions to maintain AVF patency remains a barrier for widespread use of AVF among HD patients. PTA causes endothelial injury resulting in vascular restenosis. The rate of restenosis following conventional balloon angioplasty is up to 50% within 6 months. The DCBs sought to address the fibrotic remodeling in the vessel through direct introduction of an anti-proliferative agent (Paclitaxel) into the vessel wall through the deployment of a drug coated balloon. Following the initial negative RCT trial reported in the CJASN 2018 in which DCBs failed to show a benefit when compared to the the standard-balloon angioplasty in AVFs, this present study opted to evaluate the effectiveness and safety of DCBs at a higher Paclitaxel dose (3.5μg per mm² compared to the 2.2μg per mm² in the CJASN study). This trial demonstrated that a DCB provided superior primary patency compared to standard balloon angioplasty at 6 month follow-up (82% vs. 59.5%). Furthermore, primary safety outcomes of the DCB intervention was noninferior to standard balloon angioplasty.
The 6-month outcomes from this trial are promising and confirm the short-term patency benefit and safety of the DCB device. However, the long-term data compiled from Paclitaxel-coated devices is mixed. While no significant difference in mortality was observed between the DCB intervention and the standard-balloon angioplasty at 12 months follow-up, longer-term follow-up is necessary to provide a more complete risk-benefit profile in hemodialysis patients. The DCB use in peripheral artery disease have also shown improved patency compared to conventional balloon angioplasty, however, the long-term safety data from DCB use in those studies have shown a higher mortality associated with DCB use at 3 and 5 year follow -up. Therefore, it remains to be seen if there is higher risk of mortality with DCB use in HD patients. This unanswered question will be a big deciding factor in widespread adoption of DCB use for AVF stenosis (Katsanos, J Am Heart Assoc 2018).
The main strength of this trial is that it was a relatively large, multicenter, multinational randomized controlled trial.
Limitations: It was impossible to conduct a double-blind trial due to the macroscopic appearances of the DCB and standard balloon. Outcomes are limited to 6 month follow-up. Central vein, in-stent and arteriovenous graft stenotic lesions were excluded from the trial.
All AVF were functional and being used for dialysis at the time of intervention, hence AVF that failed to mature (primary failure) were not included in the study.
CONCLUSION
Are drug coated balloons the solution to a 54 year-old problem? Likely not, but they are a step in the right direction. Due to the physiologic nature of the hemodialysis AVF and AVG circuit, endothelial injury and stenotic remodeling will likely never go away. Furthermore, until we are able to innovate procedures that reduce the initial endothelial trauma inflicted at the time of AVF creation, and the subsequent vascular remodeling and fibrosis, AVF stenosis will remain a pervasive problem. While long-term follow-up for patency and safety is needed, this DCB study offers a promising alternative to the age-old angioplasty