Abstract
Objectives:
The clinical efficacy and safety of drug-coated balloon (DCB) angioplasty in patients with coronary in-stent restenosis (ISR) has been demonstrated. The objective of this article is to provide comparative cost efficacy data for DCB angioplasty in various countries based on the original methodology of the Medical Technologies Evaluation Programme (MTEP) at the National Institute for Health and Clinical Excellence (NICE) in 2010.
Study design:
Published and unpublished Health Technology Assessment (HTA) reports were evaluated for comparison in selected countries. Furthermore, a systematic review of economic evaluations of DCB angioplasty versus standard treatments (uncoated balloon angioplasty or drug-eluting stent implantations) was conducted.
Methods:
National cost efficacy data were evaluated using Markov state transition models which were adapted to fit each country’s device and procedure related costs. The clinical input for adverse events was defined with two relevant trials for in-stent restenosis of bare metal stents (BMS-ISR) and of drug-eluting stents (DES-ISR).
Results:
In the UK, Germany, Switzerland, South Africa, Japan and Brazil, DCB angioplasty is cost-effective when compared with drug-eluting stents to treat either BMS-ISR or DES-ISR.
Conclusions:
DCB angioplasty ought to be the preferred treatment option for patients with BMS-ISR and DES-ISR from the payers’ point of view.
Introduction
In general, innovations are rapidly adopted in the field of interventional cardiology. In Europe, bare metal stents (BMS) in the 1990s and drug-eluting stents (DES) in 2002 were quickly introduced based on an adequate reimbursement with sufficient clinical evidence. The first drug-coated balloon (DCB) with acceptable clinical evidence to treat BMS in-stent restenosis (BMS-ISR) was first reviewed by the Medical Technologies Advisory Committee (MTAC) at the National Institute for Health and Clinical Excellence in 2010 [NICE, 2010]. MTAC concluded that there was sufficient evidence to support the claim of cost savings associated with the investigated paclitaxel-coated balloon (Paccocath®, DCB Technology). This NICE recommendation had a global impact which extended to mainland Europe, South Africa, Brazil and Japan.
This article reviews the clinical evidence for paclitaxel-coated balloon angioplasty, the Health Technology Assessments (HTAs) in various countries, and the real world hurdles to obtain reimbursement in the interest of patients and payers. Its objectives were to provide an overview of: (1) the clinical evidence for DCB; (2) state transition model parameters in various countries; and to (3) summarize cost savings per country for this ‘non implant’ DCB angioplasty.
Methods
Based on the original 2010 NICE recommendation Markov BMS-ISR state transition model [NICE, 2010], various adaptations were made to fit either country-specific cost data or other extended indications such as DES-ISR.
The clinical data used in the review were based on clinical trials conducted in accordance with Good Clinical Practice (ethics approval, patient consent). All clinical data were previously published in peer-reviewed journals. Ethical approval of this research was not necessary since cost efficacy data were derived from randomized controlled trials obtained from the literature.
Results
Systematic literature review
A MEDLINE search was conducted with the search terms listed in Table 1. Since DCBs were referred to as drug-eluting balloons (DEBs) in early studies and recently described as drug-coated balloons (DCBs), both search terms were used. As of 26 November 2014, a total of 157 publications referenced these search terms (Figure 1). Excluded from our cost efficacy analyses were meta-analyses, literature reviews or comments on original studies if they did not reference any cost efficacy data. Furthermore, only paclitaxel-coated balloon (PCB) studies for the coronary application were considered.
Search terms and hit sensitivity (search date 26 November 2014).

Overview of the systematic literature review with four references on cost-effectiveness (right side of the diagram, search date 26 November 2014).
Overall, there were 36 publications that reported outcomes of the original Paccocath® Technology (37 clinical, 2 preclinical) whereas 12 literature references described the results of other DCB technologies in patients with coronary artery disease (DIOR®, Freepac®, paclitaxel/BTHC). The initial NICE evaluation [NICE, 2010] was adopted in Germany [Bonaventura et al. 2012; Dorenkamp et al. 2013] and Brazil [Wainstein et al. 2013] to study cost-effectiveness in patients with BMS-ISR. One additional Swiss DCB cost efficacy study [Diehm and Schneider, 2013] was found in the peripheral field and was therefore not further explored.
Clinical evidence
DCB angioplasty demonstrated substantial evidence for the treatment of coronary lesions. Most evidence was generated with the Paccocath® Technology (SeQuent® Please, B. Braun Melsungen AG) in patients with DES-ISR, BMS-ISR and de novo lesions.
The majority of the trials had angiographic endpoints such as late lumen loss (LLL) which was indicative for the suppression of intimal hyperplasia. LLL was consistently low in patients treated with DCB angioplasty (paclitaxel–iopromide matrix coating). Also favorable clinical outcomes more relevant for HTAs could be derived from these trials. Overall, LLL was significantly reduced compared with their controls (uncoated balloon, paclitaxel-eluting stent). DCB angioplasty to treat BMS-ISR [Scheller et al. 2006, 2008, 2012; Unverdorben et al. 2009] as well as DES-ISR [Habara et al. 2011, 2013; Rittger et al. 2012; Byrne et al. 2013; Xu et al. 2014; Alfonso et al. 2014] delivered angiographic and clinical benefits. With the exception of the RIBS V trial [Alfonso et al. 2014], which had minimal lumen diameter (MLD) as the primary endpoint, all of these studies met the primary endpoint.
In addition, there are five studies which investigated the clinical efficacy and safety of DCBs in patients with de novo lesions. These studies included diabetics [Ali et al. 2011] and patients with small vessel disease [Unverdorben et al. 2010] and bifurcation lesions [Mathey et al. 2011; López Mínguez et al. 2014]. Moreover, a DCB trial with and without the additional implantation of stents coated with endothelial progenitor cells [Wöhrle et al. 2011] was conducted. Generally, favorable LLL in all of these studies were reported. However, there are only three angiographic endpoint trials in de novo small vessel disease patients. These are the single armed PEPCAD I trial [Unverdorben et al. 2010], the PICCOLETTO trial [Cortese et al. 2010] and the BELLO study [Latib et al. 2012]. DCB catheters with a paclitaxel-spacer matrix (Paccocath® or Freepac®) delivered consistently low LLL (<0.20 mm) if used without additional stents (‘DCB only’ concept) in de novo lesions. Non paclitaxel-spacer matrix DCBs did not show efficacy in de novo lesions [Cortese et al. 2010]. Kleber and colleagues proposed treatment recommendations on how to use the DCB in various lesion subsets [Kleber et al. 2013]. Large-scale postmarket surveillance registries have been conducted in a mélange of indications (BMS-ISR, DES-ISR, de novo) [Wöhrle et al. 2012] or in dedicated small vessel cohort studies with the intention to use the ‘DCB only’ approach [Zeymer et al. 2014].
Kaul and colleagues studied the combination of DCB and BMS in de novo lesions [Kaul et al. 2013]. They investigated the procedural sequence DCB followed by BMS versus BMS followed by DCB and demonstrated that the sequence of the device use did not impact on the angiographic or clinical outcomes. Moreover, Kleber and coworkers observed positive remodeling, i.e. a lumen enlargement following stent-free DCB angioplasty in selected patients with de novo lesions [Kleber et al. 2014].
HTAs
Upon market approval in 2009, a first HTA was initiated at NICE [NICE, 2010]. The original NICE Markov state transition model (Dr William Valentine, Ossian Consulting, Basle, Switzerland) was based on the 12-month PEPCAD II results and was populated with UK cost data (Figure 2). A number of HTAs were performed in various settings to accommodate for deviating cost data on a country basis (Table 2). A simplified overview of indications, comparators, state transition Markov model time horizons and cost savings based on assumed device costs are listed in Table 3.

Simplified state-transition Markov model for patients with BMS-ISR lesions treated with DCB or PES.
Market approval, clinical recommendations and reimbursement.
BMS-ISR, bare metal stent in-stent restenosis; DCB, drug-coated balloon; DES-ISR, drug-eluting stent in-stent restenosis; DRG, diagnosis related group; ESC, European Society of Cardiology; NICE, National Institute for Health and Clinical Excellence; VAT, value added tax.
Paclitaxel-coated balloon (PCB) cost savings in various countries.
BMS-ISR, bare metal stent in-stent restenosis; DCB, drug-coated balloon; DES-ISR, drug-eluting stent in-stent restenosis; DES, drug-eluting stent; GBP, British pound; ZAR, South African rand; EUR, euro; CHF, Swiss franc; BRL, Brazilian real; JPY, Japanese yen.
Discussion
Our research revealed that there are cost savings associated with DCB angioplasty in various countries and healthcare systems. In the following discussion, we would like to point out some peculiarities in selected countries.
The French reimbursement system in the interventional cardiology product range is based on implants. Consequently, non implant devices such as DCB catheters are by definition not applicable for a reimbursement category despite very favorable clinical and cost efficacy data. Within the Haute Autorité de Santé (HAS), the so-called CNEDIMPTS (Committee for the Evaluation of Medical Devices) is the entry point for new products for which reimbursement is requested. In case there is adequate evidence to support reimbursement, CNEDIMPTS, CEPS (Committee to Evaluate the Economic Value of a Medical Device) and the applicant needs to decide to conduct postmarket clinical studies before a reimbursement price is defined on a reimbursement list (Liste des Produits et Prestations Remboursables).
In Germany, hospitals receive funds in a system of dual financing. Infrastructural investments are financed with state budgets while operating costs are paid by public and private health insurance companies. In 2000, the Statutory Health Insurance Reform Act was the starting point for a diagnosis-related group (DRG) ‘package price’ for the entire procedure to cover all operating costs in German hospitals. Dedicated funds for innovative products that are not covered by the DRG system during market entry phase were used to ease the transition of DCB angioplasty into a more formal reimbursement. This allowed the reimbursement gap between product availability and DRG reimbursement to be covered. Bonaventura and colleagues give additional details of the German system [Bonaventura et al. 2012].
In South Africa there are currently no regulations regarding approval for medical devices. However the Medicines Control Council do require that combination medical devices be registered under the Medicines Control Act. Due to the lengthy process of such registration in South Africa so called ‘Section 21’ approval is granted as an interim measure. Once products are registered they are subject to annual price increases governed by what is known as single exit pricing (SEP). Reimbursement by the health insurance companies (Funders) is based on the DES as a comparator.
Prior DES HTAs
Sensitivity analyses revealed that the cost and duration of nongeneric platelet aggregation inhibitors (e.g. clopidogrel) have an important impact on savings. Despite the wish for a shortened dual antiplatelet treatment (DAPT), the European Society of Cardiology (ESC) guidelines initially recommended 12 months for all patients receiving DES. Our literature search did not yield any cost efficacy analysis for the use of sirolimus-eluting stents (SES) which required costly DAPT such as nongeneric clopidogrel prior to the availability of generic versions. When compared with BMS implantations that require only 1 month of postprocedural DAPT, the cost savings in the case of SES with 12 months of DAPT were significantly overestimated. With a view to improved patient outcomes there is also the option to implant bioresorbable scaffolds which, with the exclusion of the platinum radiopaque markers, promise to be fully absorbed. Given technical challenges, such as large crossing profiles and bulky stent struts, the principal question of how long costly DAPT needs to be given remains unanswered. Moreover, the arrival of new but more costly platelet aggregation inhibitors, such as ticagrelor or prasugrel, requires recalculation of the costs in challenged patient subgroups.
Clinical benefits of ‘non stent’ interventions
DCB angioplasty has been increasingly used in coronary interventions in the last decade. Pioneering work by Speck, Scheller and colleagues [Scheller et al. 2004; Speck et al. 2006] was the starting point to establish DCBs in the ESC guidelines [Windecker et al. 2014] and to obtain a NICE recommendation in the setting of BMS-ISR. Early work in small vessels [Unverdorben et al. 2010] and bifurcations [Mathey et al. 2011] suggested that the late loss remained very acceptable around 0.2 mm and drove the concept into the mainstream of de novo lesion percutaneous coronary intervention (PCI). There are sufficient data on the use of DCB in the setting of ISR, small vessels and to a lesser extent bifurcations, but as yet there is only one randomized study in de novo PCI compared with a standard of care stent strategy. Registry data are, however, very encouraging with low major adverse cardiovascular event (MACE) rates and impressive target lesion revascularization (TLR) and target vessel restenosis (TVR) rates [Wöhrle et al. 2012; Zeymer et al. 2014].
There are many potential advantages to the use of DCBs. A major issue in mainstream PCI is the length of DAPT, which is fraught with bleeding complications and impeded by the potential need for elective prognostic surgery and threatened by early inappropriate cessation with consequent stent thrombosis. DCB-only PCIs require only 1 month of DAPT in de novo and ISR procedures if no additional stents are implanted and if there is no outstanding requirement for DAPT from prior DES implantations. The only unanswered question related to the use of a DCB-only strategy is DAPT duration in the acute coronary syndrome (ACS) setting, where 12 months is still recommended. However, this recommendation is currently based on expert consensus and is perhaps not relevant to the entire ACS population currently treated.
The treatment of ISR has been a long-standing problem in PCI that now appears to be largely resolved with the advent of DCBs. With a DCB strategy, good acute results are achieved without the creation of a stent-in-stent effect and with acceptable ‘re-re-stenosis’ rates. In the case of re-ISR, further re-PCI is more likely to be performed and be successful after DCB treatment of previous ISR as there is no concern over multilayering of stents, thus avoiding either brachytherapy or coronary artery bypass grafting (CABG). Recurrent re-ISR may be treated with multiple dose delivery (personal communication B. Scheller, K. Bonaventura; personal experience S. Eccleshall).
The use of a DCB-only strategy also has the advantage of resolving very complex vessel anatomy and subsequent stent strategy. This is perhaps seen best in the setting of bifurcations and long lesions with aneurysmal segments. In the former, a ‘two stent strategy’ is sometimes favored, but the choice of strategy may not be clear or well evidenced. This can result in multiple stent layers or poorly opposed struts with subsequent re-stenosis or concerns of thrombosis. Similarly, isolated distal branch disease (i.e. Medina class 0,0,1 or 0,1,0 bifurcations) can be treated in a simplified manner with the DCB-only approach. In the case of aneurysms, it can be extremely difficult to reliably achieve full strut apposition again with concerns over thrombosis risk. This raises the more interesting prospect of de novo DCB PCI without implant and thus, theoretically, no long-term thrombosis risk and the maintenance of vessel geometry and endothelial integrity and function, i.e. the ‘no metal’ PCI concept.
HTAs and Markov state transition modeling are needed in patients with de novo lesions. From the clinical viewpoint, an alignment of clinical evidence, proven cost efficacy and adequate reimbursement seems appropriate to offer new and meaningful technologies to patients with coronary artery disease.
Conclusion
There were cost savings associated with the use of DCB angioplasty in patients with either BMS-ISR or DES-ISR in all researched countries. Future HTAs have to include costly comedications, which may vary depending on the chosen treatment strategy. In addition, costs for the device, procedure and for adverse events have to be considered for more accurate assessments in the real clinical setting. The NICE recommendation did have a global impact on reimbursement agencies both across Europe and elsewhere.
Footnotes
Acknowledgements
We thank Dr William Valentine (Ossian Consulting, Switzerland) for his skillful implementation of Markov state transition models which were applied for the healthcare settings in the UK, Switzerland and Japan. In addition, we would like to mention Dr Deon Olivier (Quark Healthcare Consulting, South Africa) and Dr Vanessa Teich (MedInsight, São Paulo, Brazil) for their respective contributions.
Funding
All external consulting fees to compute cost savings for DCB angioplasty were paid by B. Braun Melsungen AG, with the exception of the MTG1 recommendation which was funded by NICE. The state-transition Markov models for the UK, Switzerland, South Africa, Brazil and Japan were funded by B. Braun Melsungen AG or its subsidiaries.
Conflict of interest statement
M.W. is a full-time employee in the Medical Scientific Affairs department of B. Braun Melsungen AG.
