Author + information
- Received December 28, 2015
- Revision received February 5, 2016
- Accepted February 11, 2016
- Published online May 23, 2016.
- Salvatore Cassese, MD, PhDa,∗ (, )
- Gjin Ndrepepa, MDa,
- Francesco Liistro, MDb,
- Fabrizio Fanelli, MDc,
- Sebastian Kufner, MDa,
- Ilka Ott, MDa,
- Karl-Ludwig Laugwitz, MDd,e,
- Heribert Schunkert, MDa,e,
- Adnan Kastrati, MDa,e and
- Massimiliano Fusaro, MDa
- aDeutsches Herzzentrum München, Technische Universität München, Munich, Germany
- bCardiovascular and Neurologic Department, San Donato Hospital, Arezzo, Italy
- cUnit of Vascular and Interventional Radiology, Department of Radiological Sciences, Sapienza University, Rome, Italy
- d1. Medizinische Klinik und Poliklinik, Klinikum Rechts der Isar, Technische Universität München, Munich, Germany
- eDeutsches Zentrum für Herz-Kreislauf Forschung, partner site Munich Heart Alliance, Munich, Germany
- ↵∗Reprint requests and correspondence:
Dr. Salvatore Cassese, Deutsches Herzzentrum München, Technische Universität München, Lazarettstrasse, 36, Munich, Germany.
Objectives The aim of this study was to perform a meta-analysis of randomized trials investigating the outcomes of patients undergoing percutaneous revascularization with drug-coated balloons (DCBs) for atherosclerotic disease of the infrapopliteal arteries.
Background The role of DCBs as revascularization therapy for infrapopliteal arteries represents a matter of ongoing controversy.
Methods Scientific databases were searched for randomized trials of DCB therapy for atherosclerotic disease of the infrapopliteal arteries. The primary efficacy and safety outcomes were target lesion revascularization and amputation, respectively. Secondary outcomes were death, major adverse events, Rutherford class 5 or 6, and late lumen loss.
Results A total of 641 patients enrolled in 5 trials received DCBs (n = 378) or control therapy (uncoated balloons or drug-eluting stents; n = 263). Median follow-up duration was 12 months. Patients treated with DCBs had risk for target lesion revascularization (risk ratio: 0.71; 95% confidence interval [CI]: 0.47 to 1.09; p = 0.12), amputation (risk ratio: 1.01; 95% CI: 0.65 to 1.58; p = 0.95), death (risk ratio: 1.14; 95% CI: 0.71 to 1.82; p = 0.59), major adverse events (risk ratio: 0.92; 95% CI: 0.59 to 1.43; p = 0.70), and Rutherford class 5 or 6 (risk ratio: 0.87; 95% CI: 0.46 to 1.62; p = 0.65) comparable with that of patients treated with control therapy. Lesions treated with DCBs showed lower late lumen loss (weighted mean difference −0.41; 95% CI: −0.79 to −0.03; p = 0.04) compared with those treated with control therapy.
Conclusions In comparison with uncoated balloons or drug-eluting stents, the treatment of infrapopliteal arteries with DCBs is associated with similar clinical outcomes and favorable angiographic efficacy at 1-year follow-up. Further studies in larger numbers of patients are still needed to definitively address the role of DCB technology in this setting. (Drug-coated balloon for revascularization of tibialpedal arteries: a meta-analysis of randomized trials; CRD42015029283)
In patients with clinically relevant atherosclerotic disease of the infrapopliteal arteries, endovascular treatment with uncoated balloon angioplasty has a first-line recommendation because of the high percentage of acute success and the relatively low cost (1,2). However, the scarce durability of uncoated balloon angioplasty and the recent development of antiproliferative drug-based therapies have prompted their use in this vascular bed (3).
On one hand, coronary drug-eluting stents (DESs) have been found to be safe and effective for the treatment of focal lesions of the infrapopliteal arteries (4). However, the typically diffuse pattern of atherosclerotic disease of below-the-knee vasculature restricts the use of balloon-expandable stents to certain patient and lesion subsets. On the other hand, balloon catheters coated with paclitaxel have emerged as a valuable treatment option in patients with peripheral artery disease (3). In particular, treatment of the femoropopliteal arteries with drug-coated balloons (DCB) has been associated with angiographic and clinical efficacy superior to that of uncoated balloon catheters (5). Notwithstanding this, the role of DCBs for revascularization of stenotic or occluded infrapopliteal arteries represents a matter of ongoing controversy. Indeed, randomized trials evaluating the performance of DCBs versus other endovascular technologies in this field have reported inconsistent results: some trials demonstrated benefit (6,7), others equivalence (8,9), and others potential harm (10).
Against this background, we performed a meta-analysis of randomized trials investigating the clinical and angiographic outcomes associated with DCB-based revascularization in patients with atherosclerotic disease of the infrapopliteal arteries.
Search strategy and selection criteria
We searched MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, scientific sessions abstracts, and relevant Web sites (www.cardiosource.org, www.clinicaltrialresults.org, www.escardio.org, www.tctmd.com, and www.theheart.org), without restrictions by language or publication status. The references listed in all eligible reports were checked to identify further citations. The final search was performed in November 2015. Search terms included the following keywords and corresponding Medical Subject Headings: “infrapopliteal artery,” “below the knee,” “balloon angioplasty,” “drug-coated (-eluting) balloon,” “trial,” and “randomized trial”. Inclusion criteria were: 1) revascularization of the infrapopliteal arteries with DCBs; 2) randomized design; 3) intention-to-treat analysis; and (4) a minimum 6-month follow-up period. Exclusion criteria were: 1) vessels other than the infrapopliteal arteries treated with DCBs; 2) devices other than DCBs, uncoated balloons, or DESs used; and 3) duplicated data.
Data collection and assessment of risk for bias
Two investigators (S.C. and G.N.) independently assessed publications for eligibility at the title and/or abstract level, with disagreements resolved by a third investigator (M.F.). Studies that met the inclusion criteria were selected for further analysis. Freedom from bias was evaluated for each study by the same investigators, in accordance with the Cochrane Collaboration method (11). No formal quality score adjudication was performed (12).
The primary efficacy outcome of the present study was target lesion revascularization (TLR). The primary safety outcome was amputation. Secondary outcomes were death, major adverse events (MAEs), Rutherford class (RC) 5 or 6, and late lumen loss (LLL). Additional outcomes of interest were clinically driven TLR, major amputations, and incomplete wound healing. All endpoints were evaluated at the longest available follow-up according to definitions of original protocols. If further details were required, we contacted the principal investigators of original studies.
Risk ratios (RRs) and weighted mean differences (WMD) with 95% confidence intervals (CIs) were used as summary statistics and were derived for comparison of DCBs versus pooled uncoated balloons and DESs (the control therapy). The Mantel-Haenszel random-effects model (DerSimonian and Laird) was used to calculate pooled RRs for categorical variables, while the inverse-variance random-effects model served to calculate pooled mean difference for continuous variables. The Breslow-Day chi-square test and the I2 statistic were used to test heterogeneity across the studies: I2 values of <25%, 25% to 50%, and >50% indicated low, moderate, or high heterogeneity, respectively (11). The restricted maximum likelihood method (tau2) took into account the occurrence of residual heterogeneity.
For the primary outcomes, we performed: 1) a visual estimation of a funnel plot, as well as statistical tests to evaluate the possibility of publication bias (13–15); 2) an influence analysis, in which meta-analysis estimates are computed omitting 1 study at a time; and 3) a trial sequential analysis, in which meta-analysis sample-size calculations are combined with the threshold of statistical significance (16). A sensitivity analysis evaluated the extent to which several covariates—the proportion of male sex, or patients with diabetes, the proportion of patients with chronic kidney disease or reporting critical limb ischemia (CLI) at admission, the length of the treated lesions, and the percentage diameter stenosis—might have influenced the risk estimates for the primary efficacy outcome. Additionally, the sources of moderate to high statistical heterogeneity in the risk estimates were investigated with a sensitivity analysis according the aforementioned covariates.
Statistical analysis was performed using RevMan version 5.3 (The Cochrane Collaboration, Copenhagen, Denmark), Stata version 11.4 (StataCorp, College Station, Texas) and TSA version 0.9 beta. This study was registered with PROSPERO (CRD42015029283) and conducted in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (17).
The process of trial selection is summarized in Online Figure 1. A total of 5 trials, all with full-length reports, were selected (6–10). In 1 trial, patients with multilevel atherosclerotic disease (involving the femoropopliteal and/or infrapopliteal arteries) were randomly assigned to DCBs versus uncoated balloons (7). For the present analysis, the principal investigator of this trial agreed to share data from those patients and lesions randomized to angioplasty of the infrapopliteal arteries with DCBs versus uncoated balloons. Ultimately, a total of 641 participants (378 randomized to DCB and 263 to control therapy) with 892 treated lesions (527 randomized to DCB and 365 to control therapy) were studied.
The main characteristics of the trials included are described in Online Table 1. Two of 5 trials had multicenter designs (9,10). Patients with evidence of moderate to severe stenoses (≥50% in diameter and ≥30 mm in length) or occlusion of the infrapopliteal arteries and with symptoms ranging from severe claudication to ischemic ulcerations or gangrene were randomly allocated to revascularization with DCB versus control therapy. One trial enrolled only patients with diabetes and CLI (6). Patients with acute limb ischemia, untreated inflow or outflow lesions, aneurysms, or previous stenting of target lesions or surgery of target vessels were excluded. In case of significant inflow lesions of the target limb, these had to be concomitantly treated in order to proceed to randomization. In those trials comparing DCBs with uncoated balloons, a few patients required bailout stenting because of suboptimal initial results after angioplasty (6,7,9,10). In the trial comparing DCBs with DESs, patients allocated to stent implantation received limus-eluting coronary stents (8).
The clinical characteristics of trial participants typically matched those of patients with advanced-stage disease of the infrapopliteal arteries (Table 1). Among trials, patients had a median age of 71.5 years (71.3 to 72.5 years), were predominantly men, had a high frequency of diabetes mellitus, and, in nearly 20% of cases, had impaired renal function. The mean proportion of patients with CLI was 99.6% (78% to 100%). Overall, lesions treated were intermediate in length (median: 121.0 mm; interquartile range [IQR]: 101.5 to 129.0 mm), and median percentage diameter stenosis was 86.0% (IQR: 83.9% to 88.0%). Patients allocated to DCB had a median lesion length of 113.1 mm (IQR: 101.5 to 129.0 mm) and a median percentage diameter stenosis of 85.0% (IQR: 83.9% to 85.3%). Patients allocated to control therapy had a median lesion length of 127.0 mm (IQR: 115.0 to 128.6 mm) and a median percentage diameter stenosis of 86.8% (IQR: 86.6% to 88.0%). Roughly one-third of lesions treated were completely occluded, and nearly 20% of lesions had severe calcifications on baseline angiography. The duration of dual-antiplatelet therapy after revascularization ranged between 1 and 6 months (median: 1 month; IQR: 1 to 3 months).
An overview of main clinical endpoint definitions among the trials included is reported in Online Table 2. In all trials, the primary endpoints predominantly consisted of angiographic measures of efficacy at 6- to 12-month follow-up. Standard medical therapy was prescribed to all patients irrespective of assigned treatment. The risk for bias among studies is reported in Online Table 3.
Of those randomized, 614 patients (95.8%) were available for assessment of outcomes of interest. Median follow-up duration was 12 months.
TLR, the primary efficacy outcome of the present analysis, occurred in 139 patients (23.6%) (Figure 1). The risk for TLR was not significantly different between patients treated with DCBs and those treated with control therapy (18.3% vs. 29.1%; RR: 0.71; 95% CI: 0.47 to 1.09; p = 0.12; I2 = 39%; p for heterogeneity [phet] = 0.16). The trial sequential analysis revealed the inadequacy of sample size accumulated for this outcome (Online Figure 2A). Clinically driven TLR occurred in 122 patients (20.7%); the risk for clinically driven TLR was not significantly different between patients treated with DCBs and those treated with control therapy (15.7% vs. 25.9%; RR: 0.73; 95% CI: 0.46 to 1.16; p = 0.18; I2 = 40%; phet = 0.16).
Amputation, the primary safety outcome of the present analysis, occurred in 86 patients (14.6%) (Figure 2). Patients treated with DCB versus control therapy had comparable risk for amputation (13.3% vs. 14.9%; RR: 1.04; 95% CI: 0.70 to 1.54; p = 0.86; I2 = 0%; phet = 0.47). The trial sequential analysis revealed the inadequacy of sample size accumulated for this outcome (Online Figure 2B). Major amputation occurred in 34 patients (5.7%); DCB versus control therapy led to a similar risk for major amputation (6.4% vs. 4.3%; RR: 1.30; 95% CI: 0.58 to 2.88; p = 0.53; I2 = 0%; phet = 0.48).
Death occurred in 68 patients (11.5%) (Figure 3). No significant difference in terms of risk for death was found with DCB therapy in comparison with control therapy (11.4% vs. 10.6%; RR: 1.14; 95% CI: 0.71 to 1.82; p = 0.59; I2 = 0%; phet = 0.91).
MAEs occurred in 209 patients (37.2%) (Figure 4) (data available for 562 patients [95.4%] enrolled in 4 trials [6,7,9,10]). No significant difference in terms of risk for MAEs was found with DCB therapy in comparison with control therapy (29.9% vs. 38.5%; RR: 0.92; 95% CI: 0.59 to 1.43; p = 0.70). There was high heterogeneity for this risk estimate (I2 = 67%, phet = 0.03).
A total of 330 patients (62%) still reported RC 5 or 6 at clinical surveillance (Figure 5) (data available for 533 patients [90.4%] enrolled in 3 trials [6,7,10]). No significant difference in terms of risk for RC 5 or 6 was found in patients treated with DCB therapy in comparison with control therapy (66.0% vs. 55.6%; RR: 0.87; 95% CI: 0.46 to 1.22; p = 0.65). There was high heterogeneity for this risk estimate (I2 = 73%, phet = 0.03). The occurrence of incomplete wound healing was reported for 107 limbs (26.2%; data available for 408 limbs treated in 3 trials [6,8,10]). No significant difference in terms of risk for incomplete wound healing was found with DCB therapy in comparison with control therapy (24.5% vs. 28.7%; RR: 0.84; 95% CI: 0.45 to 1.58; p = 0.60; I2 = 67%; phet = 0.05).
Among those lesions included in the current analysis, 482 (54.0%) were available for quantitative analysis after a median of 12 months. The proportion of lesions available for angiographic surveillance among studies varied between 25.6% and 100%. Lesions treated with DCBs showed significantly lower LLL (range 0.51 to 1.15 vs. 0.54 to 2.00 mm; WMD: −0.41; 95% CI: −0.79 to −0.03; p = 0.04) compared with control therapy (Figure 6). There was high heterogeneity for this risk estimate (I2 = 87%, phet < 0.001). The direction of the point estimate did not change when excluding the trial (8) in which the control group received DESs (WMD: −0.53; 95% CI: −1.00 to −0.06; p = 0.03; I2 = 89%; phet < 0.001).
Small study effects, influence, and sensitivity analyses
A funnel-plot distribution of primary outcomes was derived from the standard error of the logarithm RR plotted against the RR of TLR and amputation, respectively (Online Figure 3). In addition to visual estimation of the funnel plot, statistical tests addressed the possibility of publication bias for primary outcomes. Additionally, the influence analysis demonstrated that no single study significantly altered the summary RR for TLR and amputation, respectively (Online Figure 4).
We addressed whether the pre-specified covariates might have influenced the risk estimates for the primary efficacy outcome as well as for those outcomes, which displayed moderate to high heterogeneity in the main analyses. Interestingly, there was no modification of risk estimates for TLR, MAEs, and LLL according to sex (p for interaction [pint] ≥ 0.59), according to proportion of patients with diabetes (pint ≥ 0.25) or renal insufficiency (pint ≥ 0.09) or CLI (pint ≥ 0.15), and according to the lengths of treated lesions (pint ≥ 0.09). However, there was significant interaction between the percentage diameter stenosis at baseline and the risk for TLR (pint = 0.005), the risk for MAEs (pint = 0.02), and LLL (pint < 0.001). In addition, the baseline diameter stenosis and lesion length significantly modified the risk estimates for RC 5 or 6 at clinical surveillance (pint = 0.014 and pint = 0.015, respectively).
We conducted a meta-analysis of randomized trials to investigate the performance of DCBs in patients with atherosclerotic disease of the infrapopliteal arteries. The main findings are that: 1) DCBs have similar clinical efficacy; and 2) favorable angiographic outcomes in comparison with uncoated balloons or DESs at 1-year follow-up, but 3) available evidence is insufficient to definitively assess the role of DCBs compared with uncoated balloons or DESs in this setting.
The use of balloon catheters coated with antirestenotic drugs in other vascular districts (3) has encouraged investigations of DCB angioplasty for long-segment infrapopliteal disease. Initial data (6,18) displayed favorable outcomes with DCBs versus uncoated balloons in patients with advanced-stage disease of the infrapopliteal arteries at 1-year follow-up. However, these results were not confirmed in 2 subsequent multicenter randomized trials (9,10). Notably, 1 of these latter randomized trials suggested potential harm associated with the use of a specific DCB platform, which was withdrawn from the market (10).
Given the lack of incontrovertible evidence regarding the role of DCBs in this setting and considering the number of ongoing randomized trials aiming at address this issue (NCT01870401, NCT02563535, NCT02279784, and NCT02137577), we conducted a meta-analysis to study the performance of DCBs in patients with occlusive infrapopliteal atherosclerotic disease. This report highlights a number of important issues.
Risk for TLR after DCB angioplasty: Is the evidence robust?
In the present analysis, DCBs versus uncoated balloons or DESs showed comparable risk for TLR at 1-year follow-up, with a significant interaction between this risk estimate and the degree of baseline stenosis. The most important drawback of endovascular therapies in the infrapopliteal vascular territory is the scarce durability of acute results (1). Interestingly, the large majority of trials included in this analysis contained sufficient power only for angiographic or composite clinical outcomes, thus supporting the necessity of a meta-analysis to investigate the risk for TLR after DCB angioplasty. We show for the first time that even after pooling >600 patients with infrapopliteal artery disease, the impact on TLR of DCBs needs to be further studied, because the available sample size accounts for <50% of that required to address a measurable effect of DCBs for this endpoint. Larger randomized trials in this setting remain of paramount importance to determine whether the efficacy and safety of DCBs best other revascularization therapies, particularly in those cases with huge plaque burden.
Risk for amputation after DCB angioplasty: Is there a real concern?
In the present analysis, including predominantly patients with CLI, DCB therapy did not heighten the risk for amputation or incomplete wound healing. Although the sample size precludes a firm conclusion regarding these outcomes, the results are in contrast to recent evidence (10). Up to one-half of patients with infrapopliteal artery disease may undergo amputation at various levels within 1 year after the diagnosis, especially if CLI is present at baseline (19). Therefore, in these patients, the care of wounds appears at least as important as the preservation of straightforward distal perfusion (20). In support of this argument, reports of those centers with institutionally standardized wound care programs showed positive outcomes with DCB therapy, even in advanced stages of infrapopliteal disease (6,18). In contrast, a recent multicenter randomized trial found opposite results (10), likely because of baseline clinical imbalances between groups and lack of homogeneous wound care management across participating centers (21). Remarkably, only 1 trial among those included in this analysis described wounds according to the Texas wound classification system (6). Against this practice, a recent report from the Academic Research Consortium emphasizes the importance of standardized classification and care of wounds in future trials investigating revascularization of the infrapopliteal arteries (22).
Superior antirestenotic efficacy of DCB angioplasty: Is it sufficient?
In this study, we observed lower LLL at angiographic surveillance after infrapopliteal revascularization with DCBs versus uncoated balloons or DESs. The favorable angiographic outcomes associated with DCBs seem due to the inhibition of neointimal proliferation and positive vessel remodeling (8). However, in patients with advanced-stage atherosclerotic disease of the infrapopliteal vessels, positive angiographic results do not per se improve clinical outcomes (e.g., lower amputations, complete ulcer healing) unless aggressive wound care management is adopted (23). Moreover, the scarce participation in invasive surveillance among trials included, due to pre-existing comorbidities and poor functional status, somewhat challenges the use of angiographic metrics of device efficacy in this setting (23). Finally, the 11.5% overall 1-year mortality rate observed in this analysis is consistent with previous reports dealing with similar populations (4,24) and underscores the importance of identification and modification of general atherosclerotic risk factors to prolong overall survival regardless of revascularization strategies (25).
First, this meta-analysis was based on aggregate data and shared the flaws of the original trials. Second, different devices were studied in the original trials, though the influence analyses discarded a possible imbalanced contribution of a single trial on risk estimation for primary outcomes. Third, one of the devices investigated in this report is no longer available on the market. However, DCB technology for patients with infrapopliteal disease has not been abandoned, and the number of ongoing randomized trials investigating this technology reinforces the value of this meta-analysis. Fourth, protocol-mandated surveillance angiography may have magnified differences in absolute proportion of revascularizations across groups. However, the risk for repeat revascularization driven by worsening of functional status rather than by oculostenotic reflex did not reveal discordant results. Fifth, although patients with CLI typically have multilevel disease (26), no data were available regarding the outcomes of inflow lesions in participants undergoing revascularization of infrapopliteal arteries. Sixth, the median follow-up period was limited to 12 months; longer term follow-up would certainly be desirable in assessing the late clinical performance of revascularization strategies. Finally, we did not perform a cost-effectiveness analysis evaluating the economic impact of DCBs compared with uncoated balloons and DESs for the infrapopliteal arteries.
The results of our meta-analysis suggest that in patients with infrapopliteal artery disease, percutaneous intervention with DCBs compared with uncoated balloons or DESs displays similar clinical efficacy and favorable angiographic outcomes at 1-year follow-up. Further studies with extended follow-up in larger number of patients are still required to definitively address the role of DCBs in this setting.
WHAT IS KNOWN? The role of DCB therapy for atherosclerotic disease of the infrapopliteal arteries remains a matter of ongoing controversy.
WHAT IS NEW? We conducted a meta-analysis of randomized trials enrolling patients with advanced stages of atherosclerotic disease of infrapopliteal arteries. DCB therapy compared with uncoated balloon or DES therapy was associated with similar clinical outcomes and favorable angiographic efficacy at 1-year follow-up. Notwithstanding this, the limited number of patients and the lack of standardized wound care management in the majority of included trials preclude definitive assumptions regarding the comparative efficacy of these endovascular therapies.
WHAT IS NEXT? Further studies in larger numbers of patients receiving standardized wound care management and longer follow-up are still required to disclose the role of DCBs in atherosclerotic disease of the infrapopliteal arteries.
For supplemental tables, figures, and references, please see the online version of this article.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- critical limb ischemia
- drug-coated balloon
- drug-eluting stent(s)
- late lumen loss
- major adverse event(s)
- Rutherford class
- risk ratio
- target lesion revascularization
- Received December 28, 2015.
- Revision received February 5, 2016.
- Accepted February 11, 2016.
- American College of Cardiology Foundation
- European Stroke Organization,
- Tendera M.,
- Aboyans V.,
- et al.
- Anderson J.L.,
- Halperin J.L.,
- Albert N.,
- et al.
- Fusaro M.,
- Cassese S.,
- Ndrepepa G.,
- et al.
- Cassese S.,
- Byrne R.A.,
- Ott I.,
- et al.
- Liistro F.,
- Porto I.,
- Angioli P.,
- et al.
- Fanelli F.,
- Cannavale A.,
- Corona M.,
- Lucatelli P.,
- Wlderk A.,
- Salvatori F.M.
- Siablis D.,
- Kitrou P.M.,
- Spiliopoulos S.,
- Katsanos K.,
- Karnabatidis D.
- Zeller T.,
- Beschorner U.,
- Pilger E.,
- et al.
- Zeller T.,
- Baumgartner I.,
- Scheinert D.,
- et al.
- Higgins J.P.,
- Thompson S.G.,
- Deeks J.J.,
- Altman D.G.
- Sterne J.A.,
- Egger M.,
- Smith G.D.
- Schmidt A.,
- Piorkowski M.,
- Werner M.,
- et al.
- Tsai T.T.
- Laird J.R.,
- Armstrong E.J.
- Patel M.R.,
- Conte M.S.,
- Cutlip D.E.,
- et al.
- Thukkani A.K.,
- Kinlay S.