Drug-eluting balloon (DEB) versus plain old balloon angioplasty (POBA) in the treatment of failing dialysis access: A prospective randomized trial.

OBJECTIVE: To compare the efficacy of angioplasty using drug-eluting balloons (DEB) compared with plain old balloon angioplasty (POBA) to reduce the rate of restenosis.
METHODS: This prospective, single-centre, single-blinded, 1:1 randomized, clinical trial enrolled patients that had primary or restenotic lesions in native upper extremity arteriovenous (AV) fistulas or at the graft-venous anastomosis. Patients were randomized to angioplasty with a POBA or a DEB. The primary effectiveness endpoints were freedom from target lesion revascularization (TLR) and functional status of access circuit at 12 months.
RESULTS: A total of 42 (28 male, 14 female; age range, 42-83 years) patients were enrolled. Patients were followed for 12 months. No significant differences were detected between the POBA and DEB groups regarding total number of TLR procedures (31 versus 36, respectively), freedom from TLR (3 versus 4, respectively) and functional status of the access circuit at 12 months (14 of 20 patients [70%] versus 17 of 22 patients [77%], respectively).
CONCLUSION: This clinical trial did not demonstrate any significant differences between DEB angioplasty and standard balloon angioplasty when treating dysfunctional haemodialysis access.

Introduction

In accordance with the Kidney Disease Outcomes Quality Initiative guidelines, patients in haemodialysis care frequently receive an upper extremity arteriovenous (AV) fistula or graft,[1-3] in conjunction with the idea of choosing the ‘right access at the right time for the right patient’. With continuous regular use, these circuits will develop significant stenosis.[4,5] Stenoses of AV fistulas and grafts are a common problem that often compromise optimal dialysis. Haemodialysis patients are usually intensely monitored and when a malfunction of the access is noticed, further investigations and interventions are undertaken. Currently, the standard treatment is balloon angioplasty with standard, or often, high-pressure balloons. Unfortunately, there are limitations with this treatment because primary patency is limited and the rate of restenosis is high.[6-12] Studies show an approximate 50% reintervention rate within 6 months with plain old balloon angioplasty (POBA).[6,8,10,13-15] One option that may reduce the number of repeat angioplasties and extend treatment intervals is to use drug-eluting balloons (DEB). These balloons are coated with an antiproliferative drug, paclitaxel; and have documented effects on reducing the rate of restenosis when treatments are performed in other vascular territories.[16-26] Two randomized clinical trials,[27,28] as well as many smaller series,[29-49] have been published on this topic reporting conflicting results. Whereas treatment is considered safe,[27,50,51] some studies have shown positive effects on primary patency and reduced number of reinterventions,[27,31,33,35-37,39,40,45,47,52] whereas others have not.[28,29,32,41,44,53,54]

This prospective randomized clinical trial compared the efficacy of angioplasty using DEB compared with POBA to reduce the rate of restenosis in patients that had primary or restenotic lesions in native upper extremity AV fistulas or at the graft-venous anastomosis.

Patients and methods

Study population

This prospective, single-centre, single-blinded, 1:1 randomized, clinical trial, evaluated angioplasty with an Advance® 18 PTX Drug-Eluting PTA Balloon Dilatation Catheter (paclitaxel-coated balloon; Cook Medical, Bloomington, IN, USA) compared with an Advance® 18LP – Low Profile PTA Balloon Dilatation Catheter (identical, non-coated, standard balloon; Cook Medical). The study enrolled patients in the Department of Thoracic and Vascular Surgery, Skåne University Hospital, Malmö, Sweden between March 2014 and July 2017. The inclusion criteria were as follows: (i) adult >18 years; (ii) active dialysis with mature upper extremity dysfunctional haemodialysis access; (iii) primary stenosis or nonstented restenosis in a native AV fistula or at graft-venous location; (iv) target vessel diameter 3–8 mm. The exclusion criteria were as follows: (i) thrombosed access; (ii) in-stent restenosis; (iii) pregnancy: (iv) ≤18 years. The study participants were treated with percutaneous transluminal angioplasty (PTA) using the study balloon as described below.

The trial protocol was approved by the regional research ethics committee in Lund, Sweden (Dnr: 2012/305). In error, the trial was not prospectively registered, but has been registered retrospectively at ClinicalTrials.gov (registration no. NCT05173857). The trial was conducted in accordance with the Declaration of Helsinki and all participants provided written informed consent. The reporting of this study conforms with the CONSORT statements.

Definition of dysfunctional access

Dysfunctional access was defined in two ways in this study. First, clinically-evident dysfunctions that occur when there are repeated problems with performing the haemodialysis procedure. The reasons for this type of dysfunctional access can be multifactorial including inflow problems, with low flow and corresponding low quality (k/tV) dialysis and recirculation, or outflow problems, with high venous pressures and corresponding prolonged bleeding times at the end of the haemodialysis session. Some circuits also have additional problems with accessing the circuits or leakage. It is not uncommon for this type of dysfunction to result from a combination of problems. Secondly, there was an option for treatment as defined by a possible pre-emptive strategy, i.e. ‘physiological’ or ‘subclinical’ malfunction. In these circumstances, circuits did not display any clinical problems, but were identified by large access flow reductions (>50%), i.e. reductions in the volume flow between two consecutive registrations. Some patients with flow reductions, approaching the ‘low flow window’ and at risk for thrombosis (<500 ml/min), were also further investigated with duplex ultrasonography as described below.

All patients identified as having dysfunctional circuits were examined with duplex ultrasonography (Philips iU22 system; Philips Healthcare B.V, Best, the Netherlands) so that significant stenoses (>50%) could be detected before any endovascular treatment was scheduled. An access circuit stenosis was judged significant if >50% binary stenosis was subsequently detected on the actual angiogram (Siemens ArtisQ; Siemens Healthcare GmbH, Erlangen, Germany). Access circuit extension was defined from the juxta-anastomotic arterial inflow to the axillary vein outflow before the thoracic outlet region.

Randomization procedure

Trial participants were randomized in a 1:1 ratio using a blinded envelope method (POBA or DEB) in two blocks of 25 + 25 tickets. The patients were randomized in blocks to compensate for inclusion skewness if the study inclusion was extended over time. The staff at the angiographic suite performed this randomization procedure after the lesion was crossed with a guidewire. Only the treating physician and staff, not the patients, were aware of the choice of treatment (i.e. single-blinded).

PTA procedure

The lesion was crossed with a guidewire and then after randomization the lesion underwent a direct dilatation with one of the study balloons. The balloon was matched to the reference vessel diameter in an approximate 1:1 sizing. Inflation time was set to 90 s. If the primary dilatation was judged unsuccessful, further treatment was initiated at the discretion of the operator. In the case of multiple stenoses in the circuit, all of the stenoses received the same treatment according to the patient’s study allocation. On the rare occurrence of a long lesion, when more than one angioplasty balloon was used, the trial protocol emphasized the importance of adequate dilatation and treatment of the overlapping regions and minimization of the risk for skipping regions, in order to optimize drug delivery.

Follow-up procedure

The prespecified trial protocol dictated a follow-up period of 12 months. At the haemodialysis unit, monthly recordings of volume flow (Transonic HD03 Hemodialysis Monitor; Transonic Systems Inc., Ithaca, NY, USA) were undertaken as part of routine care. If large flow reductions were noticed (i.e. >50% as prespecified), a duplex ultrasound examination was performed (Philips iU22 system; Philips Healthcare). Ultrasound was also performed when clinical haemodialysis access related problems were detected regardless of volume flow measurements. When duplex ultrasound indicated a significant stenosis or restenosis, corresponding to clinical or functional issues, the patient was scheduled for a new endovascular procedure. No regularly timed prespecified investigations were performed, other than volume flow measurements, as was the standard clinical practice.

Primary and secondary endpoints

The prespecified primary endpoints were as follows: (i) freedom from target lesion revascularization (TLR) at 6 and 12 months; (ii) freedom from access circuit revascularization at 6 and 12 months; (iii) functional access at 12 months. The prespecified secondary endpoints were as follows: (i) time to first TLR; (ii) survival at 12 months; (iii) procedural- and access-related complications; (iv) procedural technical performance. Endpoint censorship occurred in following situations: (i) deceased patient; (ii) access thrombosis; (iii) surgically ligated access after placement of a new access or a change of renal replacement strategy, i.e. usually a kidney transplant.

Statistical analyses

The initial null hypothesis was that PTA with DEB performed at least 50% better than PTA with POBA, regarding primary patency and freedom from reintervention during 12 months of follow-up. This calculation specified an alpha level of 5% and a power of 90% to show clinical superiority. The prespecified enrolment was set to 50 patients in each group according to pretrial calculations. At the time of the trial set up, previous research was investigated to provide an idea of a reasonable cut-off value regarding efficacy.

All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp., Armonk, NY, USA). Normal distribution was not assumed. Continuous data are presented at median (interquartile range) and compared using Mann–Whitney U-test. Categorical data are presented as n of patients (%) and compared using χ2-test and Fisher’s exact test. Kaplan–Meier survival curves were used to present time-to-event data and any differences compared with log-rank test. The survival data are presented with values ± standard error in %. A P-value <0.05 was considered statistically significant.

Results

A total of 42 patients (28 male, 14 female; age range, 42–83 years) were included in the trial (Figure 1). The inclusion period lasted for 40 months and during this time 301 patients in total were treated for dysfunctional haemodialysis access with an endovascular method. For different reasons, mostly administrative shortcomings, 14 patients had signed written informed consent but failed inclusion on the day of the intervention. Two patients were included twice and only the primary inclusion was ratified in the post-trial analysis. Four patients were actually treated for outflow in-stent restenosis and were excluded from the post-trial analysis according to the study protocol. Of the 42 patients finally included in the trial, 22 patients were assigned to treatment with a DEB and 20 with a POBA. All treatments were performed at the same institution. The baseline demographic and clinical characteristics of the study population are shown in Table 1. Other than the side of access (P = 0.041), there were no significant differences between the two groups in terms of age, sex, medical comorbidities, access circuit history and use of antiplatelet/anticoagulant drugs before or after the index intervention. The majority of patients in both groups had a previous endovascular treatment for a dysfunctional AV fistula (17 of 20 patients [85%] in the POBA group versus 17 of 22 patients [77%] in the DEB group).

Figure 1.

Flow diagram showing the progress through enrolment, randomization and analysis of patients that underwent percutaneous transluminal angioplasty (PTA) with plain old balloon angioplasty or drug-eluting balloon. AV, arteriovenous.

Table 1.

Baseline demographic and clinical characteristics of patients (n = 42) that underwent percutaneous transluminal angioplasty with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB).

Characteristic	POBA n = 20	DEB n = 22
Age, years	62 (57–77)	68 (59–76)
Sex, male	13 (65)	15 (68)
Diabetes mellitus	9 (45)	11 (50)
Hyperlipidaemia	13 (65)	12 (55)
Hypertension	16 (80)	17 (77)
Cardiovascular disease, cardiac or cerebral	11 (55)	8 (36)
First access	11 (55)	17 (77)
Previous endovascular treatment	17 (85)	17 (77)
Median number of previous endovascular treatments	2 (1–6)	2 (1–8)
Previous surgical treatment	2 (10)	2 (9)
Side of access, left	13 (65)	20 (91)*
Old access, >24 months	9 (45)	11 (50)
Antiplatelet therapy, aspirin	12 (60)	12 (55)
Anticoagulant therapy	1 (5.0)	2 (9)

Data presented as median (interquartile range) or n of patients (%).

*P = 0.041 between-group comparison; all other between-group comparisons were not significant (P ≥ 0.05); continuous data were compared using Mann–Whitney U-test and categorical data were compared using χ2-test and Fisher’s exact test.

With regard to AV fistula and lesion characteristics (Table 2), there were no significant differences between the two groups, although there were three dialysis grafts included in total, all in the POBA group. There was a small difference, although not significant, regarding lesion location, with more proximal vein lesions in the DEB group compared with the POBA group. There was also a higher frequency of lower arm fistulas in the DEB group, although this was not significant. The median lesion length was similar in the POBA and DEB groups (46 mm versus 36 mm, respectively). The need for procedural complementary retreatment with high pressure balloon or standard PTA was similar in both groups; and 13 of 20 patients (65%) in the POBA group versus 13 of 22 (59%) patients in the DEB group received this adjunctive treatment. One patient in the DEB group received a rescue stent for suboptimal angioplasty and this circuit were later surgically ligated due to arm oedema.

Table 2.

Arteriovenous (AV) fistula and lesion characteristics of patients (n = 42) that underwent percutaneous transluminal angioplasty (PTA) with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB).

	POBA n = 20	DEB n = 22
Graft	3 (15)	0 (0)
Proximal vein fistula	7 (35)	5 (23)
Distal vein fistula	10 (50)	17 (77)
Clinical inflow dysfunction	3 (15)	5 (23)
Clinical outflow dysfunction	3 (15)	3 (14)
Reduction of volume flow >50%	4 (20)	4 (18)
Other access related problems, not classified as inflow or outflow dysfunction	10 (50)	11 (50)
Proximal vein lesion	7 (35)	14 (64)
Puncture zone lesion	6 (30)	8 (36)
AV anastomosis lesion	7 (35)	12 (55)
Graft-venous lesion	3 (15)	0 (0)
Venous outflow lesion	9 (45)	8 (36)
Venous outflow treated	8 (40)	6 (27)
Arterial inflow treated	2 (10)	3 (14)
Post-PTA treatment	11 (55)	11 (50)
High pressure balloon angioplasty	3 (15)	3 (14)
Total lesion length, mm	46 (24–81)	36 (31–62)

Data presented as median (interquartile range) or n of patients (%).

No between group comparisons were significant (P ≥ 0.05); continuous data were compared using Mann–Whitney U-test and categorical data were compared using χ2-test and Fisher’s exact test.

The median size of the angioplasty balloon was 6 mm in both groups (Table 3). The maximal percentage stenosis pretreatment and post-treatment were similar in both groups. The clinical procedural success was 100% in both groups, while the corresponding radiological procedural success was rated lower, but not significantly different between the POBA and DEB groups (11 of 20 patients [55%] versus 15 of 22 patients [68%], respectively).

Table 3.

Arteriovenous fistula treatment characteristics of patients (n = 42) that underwent percutaneous transluminal angioplasty with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB).

	POBA n = 20	DEB n = 22
Technical procedural success	20 (100)	22 (100)
Radiological procedural success	11 (55)	15 (68)
Pretreatment maximal stenosis, outflow referencea	70 (55–80)	71 (48–79)
Pretreatment maximal stenosis, inflow referencea	61 (50–68)	59 (50–70)
Post-treatment maximal stenosis, outflow referencea	43 (29–60)	44 (28–56)
Post-treatment maximal stenosis, inflow referencea	30 (8–42)	17 (0–25)
Treatment balloon diameter, mm	6 (5–7)	6 (5–7)

Data presented as median (interquartile range) or n of patients (%).

aCalculation was undertaken with both outflow and inflow access diameter as baseline reference.

No between group comparisons were significant (P ≥ 0.05); continuous data were compared using Mann–Whitney U-test and categorical data were compared using χ2-test and Fisher’s exact test.

Nine patients in total had multiple, separate stenoses in the access circuit and all stenoses received the same allocated treatment. Five patients received treatment with DEB and four patients with POBA. In total, 53 stenoses were treated. Regarding these 53 locations, no significant differences in TLR procedures were noted (17 of 25 lesions [68%] in the POBA group versus 17 of 28 lesions [61%] in the DEB group).

No patients were lost to follow-up. Flow measurement data were partially incomplete in 19 patients. One patient suffered a thrombosed access in the follow-up period, without successful treatment, and consequently there were no secondary patent circuits.

In terms of the primary endpoints, there were no significant differences regarding freedom from TLR or access circuit primary patency at 12 months between the two groups (Table 4) (Figures 2 and 3). The proportion of patients with functional access at 12 months was similar between the two groups. Two circuits were lost, one in each group.

Table 4.

Primary and secondary endpoints of patients (n = 42) that underwent percutaneous transluminal angioplasty with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB).

	POBA n = 20	DEB n = 22
Mortality at 12 months	2 (10)	3 (14)
Access-related serious adverse event at 12 months	0 (0)	0 (0)
Access circuit thrombosis at 12 months	1 (5)	0 (0)
Any access-related complication at 12 months	4 (20)	1 (5)
Freedom from TLR at 12 monthsa	3 (16 ± 9)	4 (24 ± 10)
Freedom from TLR at 6 monthsa	10 (54 ± 11)	10 (48 ± 11)
Access circuit primary patency at 12 monthsa	2 (11 ± 7)	4 (24 ± 10)
Access circuit primary patency at 6 monthsa	8 (44 ± 11)	10 (48 ± 11)
Functional access at 12 months	14 (70)	17 (77)
Number of TLR interventions at 12 months	1 (1–2)	1 (0–3)
Number of access circuit interventions at 12 months	2 (1–3)	2 (1–3)
Time to first TLR, days	125 (57–261)	140 (6–294)
Number of duplex scans at 12 months	3 (1–5)	3 (1–5)

Data presented as median (interquartile range) or n of patients (%).

aLife table data are presented with numbers and (% ± SE).

No between group comparisons were significant (P ≥ 0.05); continuous data were compared using Mann–Whitney U-test and categorical data were compared using χ2-test and Fisher’s exact test.

TLR, target lesion revascularization.

Figure 2.

Kaplan–Meier curve analysis showing cumulative freedom from target lesion revascularization (TLR) in patients (n = 42) that underwent percutaneous transluminal angioplasty (PTA) with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB). The corresponding table presents the number of patients at risk (% ± SE) at certain time-points and P-values for each time-point. The colour version of this figure is available at: http://imr.sagepub.com.

Figure 3.

Kaplan–Meier curve analysis showing cumulative freedom from access circuit reintervention in patients (n = 42) that underwent percutaneous transluminal angioplasty (PTA) with plain old balloon angioplasty (POBA) or drug-eluting balloon (DEB). The corresponding table presents the number of patients at risk (% ± SE) at certain time-points and P-values for each time-point. The colour version of this figure is available at: http://imr.sagepub.com.

In terms of the secondary endpoints, the median time to TLR was similar in both groups (Table 4). There were no access-related serious adverse events and no difference in mortality at 12 months between the two groups. The median number of interventions and the number of TLR-specific interventions were similar in both groups.

Discussion

This current prospective, single-blinded, randomized clinical trial detected no significant differences between DEB and standard POBA in the treatment of dysfunctional haemodialysis circuits. Superior primary patency, longer retreatment intervals or a higher degree of freedom from TLR were not demonstrated when using drug-eluting technology.

The benefits of deposition of antiproliferative drugs (in this case paclitaxel) in certain angioplasty settings have been described for some time. For example, there is a considerable amount of scientific evidence supporting a possible anti-restenotic effect in various vascular territories.[16-26,28,30,31,33,35-47,49,56] However, these procedures have not been fully adopted as standard care and safety issues have been discussed in recent years. In particular, a meta-analysis from 2018 suggested an increase in late mortality in the cohort receiving drug-eluting therapy, so the use of this technology was heavily questioned. Mortality differences have not been evident in more recent publications on the use of antiproliferative drugs in the endovascular treatment of surgery-created arteriovenous fistulas.[27,50] A similar safety conclusion was drawn from a recent publication from the SWEDEPAD registry, which evaluated patients treated for lower limb ischaemia with drug-eluting devices. Drug-eluting treatment is currently considered to be safe in terms of late mortality issues.[27,50,51]

Regarding the local anti-restenotic effects after endovascular treatment of haemodialysis access, publications to date have shown conflicting results.[8,27,29,31-33,35-37,39-41,44,45,47,52-54] No definite signs of superior efficacy were demonstrated for a long period of time.[8,29,32,41,44,53,54] The recent large IN.PACT AV Access multicentre randomized trial showed significantly better target lesion primary patency at 6 months when patients were treated with drug-eluting balloon angioplasty. These results represent an important advance in the field of endovascular treatment of dysfunctional AV fistulas. When analysing subgroups at 2 years in the randomized Lutonix AV trial, there was an observed positive effect with significantly longer retreatment intervals when using DEB. Despite this, the trial failed to reach its primary endpoint of a superior target lesion primary patency at 6 and 24 months for DEB, although superiority could be shown at 9 and 12 months. The recently published multicentre prospective randomized PAVE study, which enrolled 212 patients, did not show any difference between standard treatment with POBA and DEB angioplasty when treating dysfunctional AV access.

There were no significant differences in the baseline demographic and clinical characteristics between the two groups in the current study. Although there was a significantly higher proportion of patients with AV fistulas on the left arm in the DEB group compared with the POBA group, there was no obvious theoretical explanation for this difference and it was presumably the result of pure chance. There was a small difference, although not significant, regarding lesion location, with proximal vein lesions being more frequent in the DEB group compared with the POBA group. Based on this difference and in conjunction with the limited number of patients enrolled, it could be argued that those in the DEB group had a worse outcome as shown by a previous study. This aforementioned study showed that the location of the main treated lesion close to the arteriovenous anastomosis may be a predictor of poorer long-term patency. A higher frequency of lower arm fistulas in the DEB group, although not significant, may have an implication on the interpretation of data due the limited number of patients.

There have been similar, small, randomized studies regarding AV fistulas treated with DEB or POBA that did not show any differences in TLR or patency,[32,41,60] as shown in the current study. In contrast, there are also similar small studies that demonstrated significant differences between the two groups.[14,36,38] A retrospective study from our institution in 2014 that analysed 159 patients treated between 2008–2009 reported a primary patency after standard PTA of 61% at 6 months and 42% at 12 months. In comparison with this, the overall effect of PTA in the current study was inferior at 12 months, with an overall primary patency rate of 14% (six of 42 patients). These current findings were at the lower end of the wide range of results from similar studies presenting POBA primary patency rates of 5–55% at 12 months.[8,31,33,37,39-41,45]

The radiological success shown in the current study was not optimal although similar in both groups. The reason is that the post-interventional analysis with detailed measurements was sometimes not performed in line with the procedure. When viewing all of the images in the post-trial analysis, some of the treated stenoses, judged and treated correctly at the intervention, still had low grade residual stenoses (>30%). The figures based on the more relevant inflow reference diameters showed median (interquartile range) residual stenosis values of 30% (8–42) in POBA versus 17% (0–25) in DEB.[8-42] There were not enough patients to perform a relevant subgroup analysis regarding this issue.

This current study had several limitations. First, the heterogenicity regarding lesion location might have caused potential bias, but this could not be significantly shown with the limited number of study patients. Secondly, the technical procedure protocol with direct angioplasty might be a possible mechanism for suboptimal DEB results. The standard procedure currently implements a strategy with predilatation and vessel preparation, with the frequent use of high-pressure balloons, before drug delivery. The reasons for treatment with direct angioplasty in the current trial was the limited scientific knowledge concerning DEB performance at the time of the study initiation and the intention to simplify treatment as much as possible. This study was planned in the early era of drug-eluting angioplasty for dysfunctional haemodialysis access, and at that time, complete knowledge regarding the technical performance of these new balloons had not been established. In this situation, the current study opted for a strategy with direct PTA because study protocols with similar features were being used in ongoing studies at the time.[33,38,56,61,62] In the end, most patients in the current study received post-treatment angioplasty due to suboptimal primary PTA, with another standard PTA balloon or high-pressure PTA balloon, at the discretion of the performing interventionalist. Consequently, 13 of 20 patients (65%) in the POBA group versus 13 of 22 (59%) patients in the DEB group got this adjunctive treatment to finally receive a good clinical angioplasty result. Thirdly, a considerable number of AV fistula treatments were performed outside of the study protocol during the trial’s extended timeframe. All of these factors could introduce potential bias, although obvious signs of such systemic bias could not be detected in the post-trial analysis. Finally, the study became underpowered as it was stopped before the planned inclusions were achieved, making investigation of the pretrial hypothesis suboptimal. The reason for stopping the trial was a slow inclusion rate and a company-initiated withdrawal of the product from the market for financial and company-related structural reasons. There were no safety or efficacy issues behind this decision. There are other similar studies that have shown significant differences in treatment efficacy with a similar number of randomized participants. Although the study product was withdrawn, the active substance (paclitaxel) is still widely used in the clinic for the treatment of both dysfunctional haemodialysis fistulas and lesions in other vascular territories.

In conclusion, this current randomized trial demonstrated no clinically significant advantages with DEB angioplasty compared with POBA when treating stenoses in dysfunctional upper extremity AV fistulas.