Author + information
- Received December 22, 2014
- Revision received March 19, 2015
- Accepted March 26, 2015
- Published online July 1, 2015.
- Carlos A. Gongora, MD∗,
- Masahiko Shibuya, MD∗,
- Jeffrey D. Wessler, MD†,
- Jenn McGregor, BS∗,
- Armando Tellez, MD‡,
- Yanping Cheng, MD∗,
- Gerard B. Conditt, RCIS∗,
- Greg L. Kaluza, MD, PhD∗ and
- Juan F. Granada, MD∗,‡∗ ()
- ∗Cardiovascular Research Foundation-Skirball Center for Innovation, Cardiovascular Research Foundation, Orangeburg, New York
- †Columbia University Medical Center, New York
- ‡Alizee Pathology, Thurmont, Maryland
- ↵∗Reprint requests and correspondence:
Dr. Juan F. Granada, CRF-Skirball Center for Innovation, Cardiovascular Research Foundation, 8 Corporate Drive, Orangeburg, New York 10962.
Objectives This study sought to compare the effect of paclitaxel-coated balloon (PCB) concentration on tissue levels and vascular healing using 3 different PCB technologies (In.Pact Pacific = 3 μg/mm2, Lutonix = 2 μg/mm2 and Ranger = 2 μg/mm2) in the experimental setting.
Background The optimal therapeutic dose for PCB use has not been determined yet.
Methods Paclitaxel tissue levels were measured up to 60 days following PCB inflation (Ranger and In.Pact Pacific) in the superficial femoral artery of healthy swine (18 swine, 36 vessels). The familial hypercholesterolemic swine model of superficial femoral artery in-stent restenosis (6 swine, 24 vessels) was used in the efficacy study. Two weeks following bare-metal stent implantation, each in-stent restenosis site was randomly treated with a PCB or an uncoated control balloon (Sterling). Quantitative vascular analysis and histology evaluation was performed 28 days following PCB treatment.
Results All PCB technologies displayed comparable paclitaxel tissue levels 4 h following balloon inflation. At 28 days, all PCB had achieved therapeutic tissue levels; however, the In.Pact PCB resulted in higher tissue concentrations than did the other PCB groups at all time points. Neointimal inhibition by histology was decreased in all PCB groups compared with the control group, with a greater decrease in the In.Pact group. However, the neointima was more mature and contained less peri-strut fibrin deposits in both 2-μg/mm2 PCB groups.
Conclusions Compared with the clinically established PCB dose, lower-dose PCB technologies achieve lower long-term tissue levels but comparable degrees of neointimal inhibition and fewer fibrin deposits. The impact of these findings in restenosis reduction and clinical outcomes needs to be further investigated.
Early clinical trials demonstrated the clinical efficacy of a paclitaxel-coated balloon (PCB) in the femoropopliteal arterial territory (1–3). The original coating formulation used in these studies (iopromide as a carrier at a dose of 3 μg/mm2) was highly crystalline, achieved high tissue levels over time, and produced a high number of coating particles following balloon inflation (4–7). Newer generation PCB technologies have reduced both paclitaxel concentration and coating crystallinity aiming to decrease vascular toxicity and the production of coating microparticles. However, although the pharmacokinetic profiles and clinical efficacies of these first-generation technologies are well documented (1,8,9), the biological effects of these technological changes on clinical performance are still unknown. Recent studies suggest that not only efficient tissue transfer, but also sustained drug retention play a critical role in the clinical efficacy of PCB technologies (5,7,10). Therefore, understanding the impact of reducing paclitaxel dose on vascular healing is critical for the successful development of future local drug delivery technologies. In this study, we aimed to study the biological effect of paclitaxel concentration and coating type on drug tissue levels and neointimal inhibition using several experimental methodologies.
Paclitaxel-coated balloon description
The following PCB technologies were used in this study: 1) Ranger (2 μg/mm2; Boston Scientific, Marlborough, Massachusetts); 2) In.Pact (3 μg/mm2; Medtronic, Dublin, Ireland); and 3) Lutonix (2 μg/mm2; BARD, Tempe, Arizona). Each treatment device uses different coating techniques and excipients. The Ranger PCB employs an acetyl tributyl citrate as the excipient. The In.Pact PCB uses urea as the excipient. The Lutonix PCB uses a non-polymer-based polysorbate/sorbitol as the excipient. The control device used was an uncoated balloon (Sterling, Boston Scientific). All balloons used in all the studies were 5 to 6 × 40 mm. The self-expanding stents used in the efficacy study were 5 to 6 × 20 mm (EverFlex, eve3 Endovascular, Plymouth, Minnesota).
In vitro coating particulate formation study
Coating characterization (n = 3 per group) was performed using an in vitro coating particulate testing bench top model. Phosphate-buffered saline (pH 7.4) at 37°C was circulated at 150 ml/min through a glass iliofemoral model using a contralateral approach. An in-line 47-mm diameter black polycarbonate filter (5.0-μm pore size) was used to collect downstream particulate. Each PCB (6.0 × 40 mm) was inserted into individual 5.5-mm inner diameter Tecothane tubes (Lubrizol Advanced Materials, Inc., Cleveland Ohio) through a new introducer sheath, tracked (2 min), and inflated per their nominal label for 45 s. After deflation, the devices were withdrawn from the model and the sheath was flushed with 20 ml of media. Filters were collected, dried, and then imaged (25×) using a Clemex Vision Particle Analyzer (Longueuil, Quebec, Canada) with Leica DM4000M microscope (Leica Microsystems, Buffalo Grove, Illinois). All particles were highlighted with green color for better visualization. The particulate images were further analyzed for the size and number (Vision PE, version 5.0 software, Clemex); the size of a particle was estimated by converting the number of neighboring pixels with similar color associated with a particle to an equivalent diameter. The average number of particulates above 300 μm in size was plotted to illustrate the different levels of systemic particles for the devices.
Pharmacokinetic study in femoral arteries
Arterial paclitaxel concentrations were determined following PCB delivery (Ranger or In.Pact) in the femoral territory of swine. The results were compared with well-established historical data published for the Lutonix PCB technology (11). Arterial tissue concentrations of paclitaxel were measured at 4 h and 1, 7, 21, 45, and 60 days. A total of 3 animals per time point (18 animals) were included and received PCB delivery in every femoral artery (36 vessels). At sacrifice, individual arteries were harvested, homogenized, extracted, and quantified by mass spectrometer for arterial paclitaxel concentration.
Hypercholesterolemic swine model
A total of 6 familial hypercholesterolemic swine obtained from the University of Wisconsin, Department of Animal Sciences were used in this study. The genotypic and clinical features of this model have been already published (12,13). For this study, animals were maintained on a low-grade cholesterol pig chow in order to increase the cholesterol levels and to accelerate the disease process. In this study, younger animals were used (∼8 months) as based on previous studies demonstrated the accelerated progression of neointimal proliferation following vascular injury despite the absence of significant atherosclerotic burden at this age (14,15).
In-stent restenosis model and treatment
Figure 1 describes the study design. The study was approved by the Institutional Animal Care and Use Committee of the Skirball Center for Innovation. All animals received standard care in accordance with the Institute of Laboratory Animal Resources guidelines (16). The diet, animal preparation, and procedural methodologies followed in this protocol have been previously published (7,17). Following vessel sizing by quantitative vascular analysis (QVA), oversized balloon injury (1.25:1.0 balloon-to-artery ratio) was applied to the target area. Each balloon inflation was followed by the deployment of a self-expanding bare-metal stents in each injured segment. After 14 days, all in-stent restenosis segments were treated with either a PCB or uncoated control balloon for a total of 60 ± 3 s (target of 1.2:1.0 balloon-to-artery ratio). Four weeks later (day 42), terminal angiography of all treated sites was performed, followed by termination and histological analysis of treatment sites.
QVA was performed at 3 different time points for the angiographic analysis by experienced angiographic technicians previously qualified by the Cardiovascular Research Foundation core lab. On-line QVA at day 0 was performed at the Advantage Workstation (AW 4.3-07, GE Healthcare, Waukesha, Wisconsin). QVA analysis at other time points was done using CAAS software (Medis Medical Imaging Systems Inc., Leiden, the Netherlands). All QVA analysis methodology used in this study has been already published (5,7). Percentage diameter stenosis at pre-procedure (day 14) and termination (day 42) was calculated as: (1 – [minimal lumen diameter /reference vessel diameter] × 100%) and the late lumen loss was calculated as a difference between minimal lumen diameter at day 42 and day 14.
The histological analysis of this study was conducted in a blind fashion by an independent pathology laboratory (Alizee Pathology, LLC, Thurmont, Maryland). Following euthanasia, arterial samples were harvested and processed following previously standardized methodologies (11). All slides were stained with hematoxylin and eosin and elastic trichrome. For histomorphometry, the cross-sectional areas (external elastic lamina, internal elastic lamina, stent area, and lumen area) of each section were measured. Neointimal thickness was calculated as: [(internal elastic lamina diameter – lumen diameter)/2] × 1,000. Neointimal area was derived from internal elastic lamina area – lumen area. Qualitative histological assessment on inflammation and healing was performed using standardized score systems (11).
Peri-strut fibrin deposition evaluation
Digital images of vascular cross sections were captured using an Olympus VS120 slide scanner (10×, Olympus America, Center Valley, Pennsylvania). Each digital image was masked in Adobe Photoshop (version 14, San Jose, California) to obscure image background and nonintimal tissue. Using Image-Pro Plus (version 7, Media Cybernetics, Rockville, Maryland), the masked images were segmented in a 24-bit hue-saturation-intensity model for fibrin and fibrous tissue and the area of tissues in each hue segment was measured in microns. The hue segmentation was then used to create a mask displaying all pixels in either segment as black, and the mask area was measured for a baseline area value. The percentages of fibrin and fibrous tissue were calculated with the formula: (segment area/baseline area) × 100.
Statistical analysis was conducted using the statistical software SigmaStat (version 3.5, Systat Software, San Jose, California). Parametric data (angiographic and histomorphometric) were analyzed by 1-way analysis of variance (ANOVA) if the data were normally distributed and the variances homogeneous. Otherwise data were analyzed by Kruskal-Wallis ANOVA. When the ANOVA was significant, post-hoc testing was conducted comparing the 3 individual DCB technologies versus the control technology using the Holm-Sidak test. Nonparametric data (qualitative histopathology scores) were analyzed with Kruskal-Wallis 1-way ANOVA. If the ANOVA was significant, post-hoc testing was conducted comparing the 3 individual DCB technologies versus control technology using the Dunn test. Differences were only considered significant when the calculated p value was <0.05.
Particulate formation and paclitaxel tissue levels
In the bench top study, the Ranger PCB had fewer observable particles than either Lutonix or In.Pact Pacific (Figure 2A). In the quantitative analysis, the average number of large particles (<300 μm) was approximately 6× to 8× lower than for both Lutonix and In.Pact Pacific (Figure 2B).
Figure 3 shows the arterial paclitaxel tissue concentration profiles of each PCB over time. All PCB technologies displayed comparable levels of acute tissue transfer following balloon inflation. At 24 h, paclitaxel tissue levels decreased to ∼20 ng/mg in both Ranger and In.Pact groups versus ∼5 ng/mg in the Lutonix group (11). At 7 days, both Ranger and In.Pact displayed a decrease in tissue levels of ∼50%. At the same time point, the tissue levels in the Lutonix group had already dropped down to ∼1 to 2 ng/mg. In contrast, paclitaxel tissue levels were still several fold higher than this level at 21 days in the Ranger and In.Pact PCB groups and slowly decreased to ∼1 to 2 ng/mg at 30 days in both groups.
In-stent restenosis efficacy study
The QVA analysis is summarized in Table 1. Angiographic variables at baseline and at 2 weeks before PCB treatment were comparable among all tested groups. Minimum lumen diameter was unchanged immediately after treatment with PCB or control balloon. At day 42, terminal angiographic analysis demonstrated reduction in percentage of diameter stenosis among all PCB-treated vessels when compared with the uncoated control. Angiographic late lumen loss was lowest in the Ranger balloon (0.51 ± 0.42 mm) versus In.Pact (0.65 ± 0.74 mm) and Lutonix (0.91 ± 0.3 mm). Representative angiographic images from days 14 and 42 are shown in Figure 4.
A summary of histological data is presented in Table 2. When compared with the control balloon group, all PCB groups demonstrated significantly greater lumen areas and levels of neointimal inhibition as determined by reduced neointimal thickness and percentage of area stenosis. The neointima maturity score was the lowest in In.Pact compared with the other DCB and the uncoated control balloon. The percentage and distribution of peri-strut fibrin deposits is shown in Figure 5. Compared with other PCB groups, the In.Pact PCB group displayed a higher number of struts covered by fibrin.
PCB technologies are emerging as the therapy of choice for peripheral vascular interventions. The pharmacokinetic profile of PCB relies on the properties of the coating. In turn, coating properties are the result of the type of excipient and manufacturing process used to develop the coating. In general, crystalline coatings, compared with amorphous coatings, result in prolonged tissue level retention; however, crystalline coatings are limited by heterogeneous drug tissue uptake and neointimal inhibition (10). The first-generation PCB technology used a crystalline coating formulation (3 μg/mm2) and showed significant decrease in angiographic restenosis among patients undergoing percutaneous intervention of the superficial femoral artery (1). Despite its initial clinical success, technology acceptance was hampered by the unpredictable pharmacokinetic profile and potential for downstream particle embolization (4). Newer-generation PCB aimed to improve these technical limitations by either modifying the degree of coating crystallinity or by decreasing the total paclitaxel concentration loaded on the surface of the balloon (2,5). To date, the impact of these technological changes on tissue pharmacokinetics and clinical efficacy are still under investigation.
In the early stages of PCB development, the safety profile of this first-generation technology was questioned due to the high amount of coating particulate produced following balloon inflation. At the present time, although the potential impact of coating particulate dislodgement on microvessel embolization remains a theoretical concern, a clinical below-the–knee trial (18) using a specific PCB technology demonstrated increased amputation rates raising questions about embolization as a potential mechanism for these outcomes. In our study, there was a ∼30% (Lutonix) to 90% (Ranger) reduction in the number of large particles in the 2-μg/mm2 technology versus the 3-μg/mm2 technology (In.Pact). These differences are important as large particles have the potential to occlude microvessels downstream following balloon inflation. However, although these findings are theoretically appealing, their clinical relevance is unknown.
The tissue pharmacokinetic study also demonstrated that other variables besides total paclitaxel concentration determine long-term tissue retention in PCB technologies. In this study, we demonstrated that a lower-dose PCB, compared with the clinically established PCB dose, has the potential to achieve long-term tissue levels of paclitaxel. However, important differences were seen at the early time points in all 3 technologies. Whereas paclitaxel tissue levels are comparable among all 3 devices following balloon inflation, there is a sharp drop in the Lutonix tissue levels at 24 h compared with tissue levels for the other 2 PCB tested. Also, although the pharmacokinetic curve of the In.Pact and Ranger technologies were comparable over time, the In.Pact balloon achieved higher tissue levels after 7 days of initial delivery. This is an important finding as it shows that a comparable surrogate of clinical efficacy (tissue levels) could be demonstrated at lower paclitaxel balloon concentrations; therefore, there is potential to demonstrate comparable positive clinical outcomes as clinical data become available.
Our study also demonstrated that histology-based neointimal inhibition was comparable among all PCB technologies regardless of excipient type or paclitaxel density. The higher density PCB (3 μg/mm2), compared with the lower density PCB (2 μg/mm2), showed slightly higher levels of neointimal inhibition but also reduced neointimal maturity and higher fibrin deposition, suggesting that reduced paclitaxel density leads to slightly lower neointimal inhibition but better healing profiles. Although only the Ranger and Lutonix PCB can be directly compared (as they use the same paclitaxel densities and thus differ by only 1 variable), each coating excipient demonstrated similar levels of neointimal inhibition. This data is important as it shows that the differences seen in tissue levels in the early time points may not significantly impact the late development of neointimal inhibition in the experimental setting.
The technological developments in the PCB field continue to evolve. The ideal PCB has been defined as achieving a uniform coating that minimizes the quantity of particles released and provides maximal drug retention with minimal vascular toxicity. Although important differences exist among all the coating technologies, our study suggests that the pharmacological behavior and tissue effects of the lower-dose PCB concepts are comparable to the clinically proven standard dose balloon. This is an important finding as it continues to expand our knowledge in regard to the therapeutic thresholds of local drug delivery devices in the peripheral vascular territory.
First, the experimental nature of the study may limit the translation of our findings to the human condition. Second, we used the 28-day follow-up time point for the evaluation of efficacy. It is possible that some of the differences seen in tissue levels among all devices could reflect important differences at later time points not evaluated in this study. Finally, the lack of a significant atherosclerotic burden found in this model takes away important biological components known to be critical for the success of local drug delivery devices in humans (e.g., calcium).
Lower dose PCB technologies achieved lower long-term tissue levels but comparable degrees of neointimal inhibition and fewer peri-strut fibrin deposits than did the clinically established PCB dose. The impact of these findings in restenosis reduction and clinical outcomes needs to be further investigated.
WHAT IS KNOWN? The ideal PCB dose for peripheral vascular interventions has not been established yet. In this study we compared the biological effect on restenosis and healing of a standard dose PCB with 2 lower-dose PCB concepts.
WHAT IS NEW? The data showed that lower-dose PCB technologies achieved lower long-term tissue levels but comparable degree of neointimal inhibition and fewer fibrin deposits than did the clinically established PCB dose.
WHAT IS NEXT? The study established the potential boundaries of clinical safety and efficacy of PCB technologies.
The Skirball Center for Innovation has performed research work for all the technologies included in this study. This study was partially funded by a research grant provided by Boston Scientific Corporation. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- analysis of variance
- paclitaxel coated balloon
- quantitative vascular analysis
- Received December 22, 2014.
- Revision received March 19, 2015.
- Accepted March 26, 2015.
- 2015 American College of Cardiology Foundation
- Liistro F.,
- Grotti S.,
- Porto I.,
- et al.
- Werk M.,
- Langner S.,
- Reinkensmeier B.,
- et al.
- Buszman P.P.,
- Tellez A.,
- Afari M.E.,
- et al.
- Milewski K.,
- Afari M.E.,
- Tellez A.,
- et al.
- Granada J.F.,
- Stenoien M.,
- Buszman P.P.,
- et al.
- Granada J.F.,
- Milewski K.,
- Zhao H.,
- et al.
- National Research Council
- Zeller T.,
- Baumgartner I.,
- Scheinert D.,
- et al.