Are drug-coated balloons cost effective for femoropopliteal occlusive disease? A comparison of bare metal stents and uncoated balloons

Is the drug-coated balloon cost effective with regard to the rate of target lesion revascularization of the femoropopliteal artery in comparison with the bare metal stent and the uncoated balloon?

Mainly based on the results of published meta-analyses, we performed a cost–effectiveness analysis of drug-coated balloons versus bare metal stents and uncoated balloons. The literature review was divided into two parts. In the first part, we retrieved existing systematic reviews with meta-analyses and health technology assessment (HTA) reports to assess the effectiveness of drug-coated balloons versus bare metal stents and uncoated balloons for the femoropopliteal artery. To obtain the most recent data, our research was limited to the last 4 years (from January 2011 to October 2015). Databases used were MEDLINE (via PubMed), ScienceDirect, Centre for Reviews and Dissemination, the Cochrane Library and the websites of the Canadian Agency for Drugs and Technologies in Health (CADTH; Canada) and the Institut national d'Excellence en Santé et Services Sociaux (INESSS; Canada). The search was performed in English and in French. The keywords used were ‘drug coated balloon’ and ‘femoropopliteal’. In PubMed and ScienceDirect, for example, this gave the following search strategy: drug-coated balloon AND femoropopliteal. Inclusion criteria were: patients treated for de novo stenosis and their outcomes at follow-up, including target lesion revascularization (primary outcome), restenosis, limb amputation, mortality and restenosis or reintervention risk factors (secondary outcomes); paclitaxel drug-coated balloons; studies comparing drug-coated balloons to bare metal stents or uncoated balloons. Exclusion criteria were: narrative reviews and systematic reviews that compared drug-coated balloons with only covered stents. In the second part, given that we found no results with regard to risk factors associated with restenosis or reintervention following the first treatment for a femoropopliteal lesion, we performed another literature review. The methodology used was the same as mentioned earlier but included narrative reviews and primary studies as an inclusion criteria. The second literature search was not restricted in terms of publication date, but we retained only the most recent publications. Considering that this literature review was performed to complete the first one as regard to risk factors, this one was conducted up to retrieve a sufficient number of articles to provide enough information on this topic. As a consequence, this second review is not a systematic one, but a narrative one. The keywords used were ‘femoropopliteal’, ‘restenosis’, ‘occlusion’ and synonyms of ‘risk factor’, such as ‘predictors’, ‘risk stratification’, ‘association with outcomes’ and ‘impact on outcomes’. The quality of the systematic reviews retrieved was assessed with the AMSTAR tool (a measurement tool to assess systematic reviews) [9]. The level of scientific evidence associated with each primary study was assessed using the grid developed by Hailey et al. [10].

For the cost–effectiveness analysis, we collected all data about the equipment used, the number of health professionals involved and the duration of interventions in angiography during 1 month (June 2014). From the 182 interventions performed, 11 were related to a stenosis of the femoropopliteal artery (conjoint interventions with the iliac artery were excluded). These 11 interventions were considered representative of interventions performed during a regular month by the two surgeons who performed 80% of these interventions in our institution. This allowed us to calculate the cost of an intervention for a de novo stenosis and to determine the average number of stents and balloons used in each procedure. Precisely, we calculated the generic cost for a representative patient in our institution (e.g., healthcare professionals, medical furniture), to which we added the specific cost of the device used (i.e., cost of drug-coated balloons, uncoated balloons and bare metal stents). Then, these data were combined with data on clinical effectiveness and restenosis risk factors retrieved in the literature. The perspective of the analysis is that of the Quebec health network. The difference in costs for an intervention and then a reintervention after a failed revascularization between the drug-coated balloons, on the one hand, and the bare metal stents or the uncoated balloons, on the other hand, was calculated on the basis of 100 patients. The primary outcome in this study was the target lesion revascularization rate. This allowed us to calculate the cost per avoided reintervention for failed revascularization. The analysis period was for 2 years. The choice of this period was based on the fact that it is relatively well-documented in comparison to longer follow-up periods but also by the fact that it can cover most of the episodes of restenosis (i.e., more than half of reinterventions occur during the first 2 years) [11,12]. A sensitivity analysis was performed considering an interval of several values regarding the relative efficacy of drug-coated balloons compared with bare metal stents or uncoated balloons. Similarly, several simulations were performed by considering several groups of patients eligible for drug-coated balloons. These patient groups were determined according to the most frequent risk factors of reintervention following the installation of a bare metal stent. Considering that the evaluation period was short, we decided to do not use a discount rate.

Following our research protocol, we found three meta-analyses for clinical effectiveness and ten primary studies for restenosis risk factors. Regarding clinical effectiveness, we found a network meta-analysis including 2532 patients from 16 studies [3], a meta-analysis including 1464 patients from 11 studies [13] and a meta-analysis from the National Institute for Health Research including 290 patients from three studies [2]. These three meta-analyses obtained 11/11 on the AMSTAR tool. Regarding the primary studies on restenosis risk factors, various designs were used and the degree of evidence could be considered as moderate (i.e., mostly retrospective studies and no one was randomized).

All studies identified refer to risk factors for restenosis or primary patency [11,12,14–20]. The study of Soga et al. [21] also provided results for the risk of target vessel reintervention. Among these studies, the TASC II classification of femoral and popliteal lesions (Trans-Atlantic Inter-Society Consensus Document on Management of Peripheral Arterial Disease) was systematically reported as a significant single risk factor. Generally, classes A and B are compared with classes C and D. However, in some cases, only class D is compared, either with respect to class C, or to classes A to C. In all cases, it clearly appears that classifications C and D are associated with a higher risk of restenosis or reintervention [11,12,19–21]. Lesions of 15 cm or more or multiple lesions with a sum of more than 15 cm are therefore subject to greater risk with a risk ratio of approximately 2.4. Then, other factors often appear to be significant, such as diabetes mellitus [12,15], chronic total occlusion [12,14], critical limb ischemia [12,14], renal failure [12,21] and female gender [11,12,17]. The risk ratio of these factors usually varies between 1.2 and 1.8. Other factors appear more rarely to be significant: hyperlipidemia [20], poor below-the-knee runoff [12], specific genetic characteristics [15,18], arterial calcification [12], chronic heart failure [12] and patient's age [17].

Given the high prevalence of these risk factors in populations treated for stenosis or occlusion of the femoropopliteal artery in these studies (almost 50% for the TASC II classification C or D, hyperlipidemia, diabetes mellitus, chronic total occlusion and poor below-the-knee runoff; about a third for female gender, renal failure, critical limb ischemia, arterial calcification; nearly one sixth for chronic heart failure and certain genetic characteristics), it is very likely that a person not belonging to class C or D of the TASC II classification has one or more other risk factors for restenosis. Therefore, patients belonging to class C or D and having two or three other risk factors are considered as having a very high level of risk.

The methodology used in this study restricted the literature search on the effectiveness of drug-coated balloons to the last 4 years and limited the search to meta-analyses and health technology assessment agencies reports. As a consequence, some recent data from primary studies may not have been considered in this analysis. Indeed, the three meta-analyses identified, those of Katsanos et al. [3], Fusaro et al. [13] and Simpson et al. [2], reported that their last literature search was conducted in January 2013, November 2012 and May 2011, respectively. However, recent primary studies on this topic do not show very different results from these meta-analyses. Tepe et al. [22] and Rosenfield et al. [23], for example, indicate no difference as regard to amputation and mortality at 12 months between drug-coated balloon and uncoated balloon. The IN.PACT study by Tepe et al. [22] indicates a significant reduction in TLR (2.4 vs 20.6%), whereas the LEVANT 2 study by Rosenfield et al. [23] indicates a nonsignificant reduction in TLR (12.3 vs 16.8%), which they explain by the very low TLR rate in the uncoated balloon group as compared with previous studies. As for these recent primary studies, studies included in the three meta-analyses were randomized. A limit of the included studies is their short-term follow-up (less than 2 years) and the small number of subjects did not allow for subgroup analysis. In this regard, it would have been appropriate to perform sub-analyses for certain classes of technologies and even patients. Indeed, because the search for primary studies conducted by Katsanos et al. [3], Fusaro et al. [13] and Simpson et al. [2] was spread over several years, the results generated by these authors did not take into account improvements to the technology. Katsanos et al. [3] indicate, for example, the presence of differences in the design and geometry of the stents used in the various studies reviewed. However, we believe that we have limited this type of bias using a literature research methodology that was limited to the last 4 years. In this way, this evaluation took into account more recent data by eliminating systematic reviews that may put too much emphasis on outdated technologies. For example, some less efficient technologies that are no longer used for revascularization of the femoropopliteal artery, such as stent ‘tantalum’ and ‘balloon-expendable stainless steel stent’ were not included in the meta-analyses of Katsanos et al. [3], Fusaro et al. [13] and Simpson et al. [2]. The data for these technologies, however, were included in some less recent meta-analyses published in 2008 and 2009 [4,24,25]. As regard to a longer follow-up with drug-coated balloon, a very recent study by Laird et al. [26] indicates that the efficacy of the drug-coated balloon is maintained at 2 years as compared with the uncoated balloon (TLR of 9.1 vs 28.3%).

Some differences between the meta-analyses reviewed could explain the differences in the target lesion revascularization rates observed in Table 1. Specifically, in the meta-analysis of Katsanos et al. [3], they observed a rate of 8% for the drug-coated balloon, while this percentage was 20 and 14% in the meta-analyses of Fusaro et al. [13] and Simpson et al. [2], respectively. The meta-analysis of Katsanos et al. [3] is the most recent and therefore it is the only one to have included the ‘DEBELLUM’ study [27]. However, the work of Fanelli et al. [27] has generated very low target lesion revascularization rates (i.e., 6%), which could have pulled down the estimate generated in the meta-analysis of Katsanos et al. [3] For their part, the work of Fusaro et al. [13] also identified a study that was not present in the meta-analysis of Katsanos et al. [3]; the BIOLUX P-1 study, which has not yet been published but was presented as part of a conference [28]. Although it was not possible to obtain the data from this study for the target lesion revascularization rates, these data could also have influenced the final estimate generated in the meta-analysis of Fusaro et al. [13]

Regarding the methodology used in the cost–effectiveness analysis, we note two limitations in the calculation of the costs per intervention. First, it should be noted that the revascularization of the femoropopliteal artery requires a number of medical devices that varies greatly from one patient to another. Thus, for each category of intervention, we calculated the mean number of devices. For example, one intervention with the drug-coated balloon requires on average 1.25 drug-coated balloons, together with 1.5 uncoated balloons. However, these means were obtained from a limited number of interventions (11 cases analyzed), which limits the accuracy of the results. The second limitation about the costs per intervention concerns the risk factors of reintervention. In our model, we calculated an average cost per procedure, depending on the technology used, regardless of the TASC II classification. However, classes C and D are associated with longer lesions that may require longer devices. According to the price differences associated with the length of the devices, the calculated intervention costs may thus be different for one or the other technology. However, after examination of these various prices in our institution, we believe that these differences may only have a marginal impact on our results.

This study assessed whether the use of drug-coated balloons for revascularization of the femoropopliteal artery was cost effective compared with bare metal stents and uncoated balloons. First, our literature search indicated that the use of drug-coated balloons was associated with lower rates of target lesion revascularization compared with uncoated balloons. The results of two meta-analyses also indicated that target lesion revascularization rates for drug-coated balloons were lower or equivalent to those with bare metal stents. However, our study also indicates a substantial heterogeneity in clinical results. In a cost–effectiveness analysis this turns to uncertainty that we dealt with a sensitivity analysis. The cost–effectiveness analysis indicated that the use of drug-coated balloons was generally cost effective compared with bare metal stents, particularly for patients at high risk of recurrence. In addition, drug-coated balloons are cost effective in comparison with uncoated balloons if the target lesion revascularization rate associated with uncoated balloons is very high and for the most at-risk patients (TASC II C or D). In short, in a Canadian context where the use of drug-coated balloons are three-times more expensive than uncoated balloons and slightly more expensive than bare metal stents, our analysis indicated that the use of drug-coated balloons in a context of high risk of restenosis and reintervention was associated with greater savings. For patients at low risk of restenosis and reintervention, uncoated balloons remained the optimal choice.

The authors acknowledge C Angers, A Banko, TB Bui, SK Bédard, C Bellemare, A Breton, C Garon, CS Giguère, M Lachance, M Robillard, A Turcotte and J Verreault for their helpful comments and support.

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.