Are Drug-Eluting Stents Cost-Effective?

Abstract

HomeCirculationVol. 114, No. 16Are Drug-Eluting Stents Cost-Effective? Free AccessArticle CommentaryPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessArticle CommentaryPDF/EPUBAre Drug-Eluting Stents Cost-Effective?It Depends on Whom You Ask Jason Ryan, MD, MPH and David J. Cohen, MD, MSc Jason RyanJason Ryan From the Division of Cardiology, Beth Israel Deaconess Medical Center, Boston, Mass (J.R.), and Saint Luke’s Mid America Heart Institute, Kansas City, Mo (D.J.C.). Search for more papers by this author and David J. CohenDavid J. Cohen From the Division of Cardiology, Beth Israel Deaconess Medical Center, Boston, Mass (J.R.), and Saint Luke’s Mid America Heart Institute, Kansas City, Mo (D.J.C.). Search for more papers by this author Originally published17 Oct 2006https://doi.org/10.1161/CIRCULATIONAHA.105.546010Circulation. 2006;114:1736–1744Since their commercial introduction in 2003, drug-eluting stents (DES) have rapidly altered the management of coronary artery disease. Before their development, the percutaneous management of coronary artery disease was performed predominantly by implantation of bare metal stents (BMS) made of either surgical stainless steel or metal alloys. Although such stents represented a considerable advance over balloon angioplasty alone, they remained limited by restenosis resulting from neointimal proliferation. As a result, ≈15% to 20% of patients treated with BMS required ≥1 repeat revascularization procedure within the 6 to 12 months after stent implantation.1 Despite numerous attempts at systemic pharmacotherapy, device modification, and even use of ionizing radiation, the rate of restenosis after BMS implantation remained largely unaffected.Response by Eisenberg p 1744Over the past 5 years, effective DES have become the first device to substantially reduce the incidence of restenosis after stent implantation. By delivering high concentrations of either antiproliferative or immunomodulatory compounds directly to the site of arterial injury and by controlling this delivery through polymer-based drug release, both sirolimus- and paclitaxel-eluting stents have safely and effectively inhibited the proliferative process that results in in-stent restenosis. In pivotal clinical trials, both sirolimus- and paclitaxel-eluting stents have reduced rates of angiographic restenosis by 70% to 90% compared with conventional BMS designs, with parallel reductions in the need for clinically driven target vessel revascularization (TVR).2–4 As a result, in April 2003, DES were approved for use in clinical practice in the United States. Within 9 months of their introduction, DES made up 35% of all stent implantations in the United States,5 and their use has increased rapidly since that time. At our own institution, DES comprised >85% of all stents implanted during the past year, and national estimates are that >90% of all percutaneous coronary intervention (PCI) procedures currently performed in the United States involve ≥1 DES.Given current procedural volumes (>1 million PCI procedures were performed in the United States in 2004)6 and costs (DES are generally 3 to 4 times more expensive than BMS), the rapid growth of this technology has raised important concerns about cost from both a hospital and a societal perspective.7,8 Because annual increases in US healthcare expenditures consistently outpace inflation,9 there is increasing interest in formally evaluating the economic impact of new technologies both before and immediately after their introduction. The aim of the present report is to summarize the economic consequences of current DES use and to explore the possible future impact of DES on US healthcare expenses.What Is the Clinical and Economic Impact of Restenosis?The major benefit of DES is a reduction in the rate of restenosis after PCI. Therefore, to understand the clinical and economic impact of DES, the overall scope and consequences of coronary restenosis must first be appreciated. Unlike native coronary artery stenoses in which plaque rupture and in situ coronary thrombosis can lead to acute myocardial infarction or death, in-stent restenosis is largely a nonfatal condition.10 Luminal narrowing after BMS implantation occurs from neointimal hyperplasia and vascular remodeling mainly at the stent margins.11 This process causes gradual, progressive loss of arterial diameter, resulting most commonly in recurrent myocardial ischemia and anginal symptoms. Given the gradual nature of the disease process, progression to myocardial infarction or sudden cardiac death is very rare.10Several clinical studies confirm the relatively benign long-term prognosis of patients who experience coronary restenosis. Weintraub and colleagues12 followed-up >3300 patients treated with successful balloon angioplasty who subsequently underwent angiographic restudy to assess for restenosis. Both unadjusted and adjusted comparisons of 6-year survival failed to demonstrate any excess mortality among those patients with restenosis. Similar insights may be derived from randomized comparisons of PCI with bypass surgery.13,14 Although rates of restenosis requiring repeat revascularization are consistently higher among patients treated with PCI in these trials (both balloon angioplasty and BMS implantation), in general, these studies have demonstrated little to no difference in long-term survival. These observations thus confirm that restenosis itself and its subsequent treatment generally do not increase mortality—at least among the vast majority of patients who currently undergo PCI procedures. One may thus infer that, although DES may dramatically reduce restenosis rates for patients treated with PCI, they are unlikely to have a major impact on long-term mortality.On the other hand, quality of life (QOL) is clearly affected by restenosis. For example, in the Optimum Percutaneous Coronary Angioplasty With Routine Stent Strategy (OPUS-1) trial of universal versus provisional bare metal stenting, patients without restenosis had less frequent angina, fewer physical limitations, and improved QOL scores at a 6-month follow-up compared with patients with restenosis.15 Similarly, in the Stent Primary Angioplasty for Myocardial Infarction (Stent-PAMI) trial, significant differences in QOL scores at 6 months were observed in BMS patients compared with those who received balloon angioplasty16; these differences were driven predominantly by lower rates of clinical and angiographic restenosis associated with BMS implantation. Although no studies to date have directly compared QOL scores between DES and BMS patients, one can reasonably assume that lower rates of restenosis associated with DES will result in improved QOL—at least in the short to intermediate term. There are no data, however, to suggest that these benefits will persist beyond the first year of follow-up once the restenosis process has completed its course and any necessary additional revascularization procedures have been performed.Economic Impact of RestenosisTo fully characterize the economic impact of restenosis on the US healthcare system, one must consider both the frequency of this adverse event and the cost of each episode. The cost of restenosis is not a single value, however, and varies considerably depending on both the setting of care and, in particular, on the underlying patient population. Much of the available data on the cost and frequency of restenosis within the US healthcare system are derived from clinical trials. For example, in the Enhanced Suppression of Platelet Receptor Glycoprotein IIb/IIIa Using Integrilin Trial (ESPRIT) of patients undergoing predominantly elective coronary stenting, the mean cost for each TVR event (which occurred in 14% of the study cohort) was $11 913.17 Thus, the economic burden of restenosis within the ESPRIT population was $1675 (ie, $11 913×14%). This value represents the potential savings to the healthcare system that could be derived from a hypothetical intervention that completely eliminated restenosis for this population. Other patient populations have higher costs associated with restenosis. For example, among patients undergoing PCI for in-stent restenosis in the Gamma-1 trial, the overall economic burden of restenosis was ≈5-fold higher than in ESPRIT, reflecting both the higher frequency of restenosis in this challenging patient population and the higher treatment costs per episode of restenosis.18 The economic burden of restenosis is similarly high among patients undergoing multivessel PCI.19Given the substantial variability in both the frequency and cost of restenosis across differing populations of PCI patients, it is clear that population-based data are required to fully characterize the economic burden of restenosis within the US healthcare system. We recently published such an analysis based on data from the Medicare program.20 Among US patients undergoing coronary stent implantation in 1999 (before the introduction of DES), the overall incidence of repeat revascularization in the first year after stent implantation was 16.9%, and the need for repeat revascularization was associated with excess costs (to the Centers for Medicare and Medicaid Services) of $19 074. By assuming, on the basis of previously published data, that 85% of procedures during this time were for treatment of restenosis,21 we estimated that the true frequency of TVR (ie, clinical restenosis) in the Centers for Medicare and Medicaid Services population was 14.4% and that the overall economic burden of restenosis to the US healthcare system during the BMS era was roughly $2500 per PCI patient.Economic Impact of DESThere are currently no population-based data on the overall clinical or economic outcomes of DES within the US healthcare system. In the absence of such data, the results of the aforementioned Medicare study can provide important preliminary insights into the potential economic impact of the introduction and widespread adoption of DES. When the sirolimus-eluting stent was first available commercially in the United States, its list price was about $3200 per stent. With increasing competition and volume discounts, many US centers currently pay roughly $2200 per DES (prices are even lower in Canada and Europe, where even more devices are routinely available). Compared with an average acquisition cost of approximately $600 per BMS, the incremental cost of each DES is currently about $1600 per stent. Because many PCI procedures require >1 stent (both to cover long lesions and to treat multiple lesions and vessels), this number does not represent the true incremental cost of DES use. Given current stent use of ≈1.6 per DES procedure,22,23 complete conversion of the current PCI population from BMS to DES would be predicted to increase costs by about $2500 per procedure.From a population perspective, however, this upfront increase in cost with DES would be offset to some degree by savings from reduced repeat revascularization procedures and other concomitant medically associated costs. To estimate the balance between these 2 effects, we have developed a computer-simulation model of PCI within the US healthcare system.24 Although the model originally was designed to reflect patients undergoing single-vessel PCI, we have recently updated the model to capture the full spectrum of PCI patients, including contemporary outcomes and costs for BMS based on published data from the Medicare program.20 Key model assumptions thus include the overall BMS clinical restenosis (TVR) rate of 14%, an incremental stent cost of $1600 per DES, and average use of 1.6 stents per PCI procedure.Figure 1 summarizes the results of this model in terms of the relationship between the restenosis relative risk reduction achieved by DES and the net 1-year cost difference between DES- and BMS-based strategies. Unless DES provide restenosis relative risk reductions >95% compared with BMS, our model projects that using DES will increase the aggregate cost of PCI within the US healthcare system. At current levels of efficacy (70% to 75% relative risk reduction),25 the net 1-year cost associated with replacing BMS with DES is estimated to be $600 per PCI patient. Given current procedural volumes of ≈1 million PCI procedures, we project that the overall impact of substituting DES for all US PCI patients is about $600 million in increased annual healthcare spending. Download figureDownload PowerPointFigure 1. Difference in 1-year cost for DES vs BMS as a function of TVR risk reduction for DES based on computer-simulation modeling.Distinction Between “Cost Saving” and “Cost-Effective”As we have shown, given current stent prices, use patterns, and outcomes, it is unlikely that drug-eluting stents will result in meaningful net cost savings for the current PCI population. This does not necessarily imply that the treatment is not worthwhile. In fact, many new medical procedures and drugs are associated with a net increase in healthcare expenditures. Under these conditions, formal cost-effectiveness analysis can be used to examine the relationship between the costs and clinical benefits of a medical device or procedure and to provide insight into those conditions under which the benefits of the device or procedure justify any long-term increase in costs.26 In formal economic analysis, the “value” of a new medical therapy is expressed in terms of an incremental cost-effectiveness ratio, which is calculated by dividing the net cost of the treatment under evaluation (relative to standard of care) by its net benefits (also relative to standard of care): equationDownload figureIn principle, costs are measured in monetary terms, and any valued clinical outcome may be used as a measure of health benefits. The standard approach, however, is to assess long-term health outcomes in terms of quality-adjusted life-years (QALYs),26 a metric that combines years of life and QOL into a single value. Each time interval in a given state of health is weighted by the “utility” of that health state; utility is a theoretical construct that represents an individual’s preference for that health state on a scale ranging from 0 to 1, where 1 represents perfect health and 0 represents death. Thus, years of life in good health yield more QALYs than years when health status is poor.Once a cost-effectiveness ratio is calculated, it typically is compared with cost-effectiveness ratios for other therapies in a “league table.” The threshold for determining whether a therapy is economically attractive varies with the available healthcare budget. In the United States, for example, cost-effectiveness ratios <$50 000 per QALY gained are generally viewed as favorable, and ratios between $50 000 and $100 000 per QALY gained are frequently considered to be in a “gray zone.” In contrast, cost-effectiveness ratios >$100 000 per QALY saved are generally viewed as economically unattractive.27 These standards do not necessarily apply to other healthcare systems, however. In particular, countries that spend considerably less on healthcare than the United States would appropriately have more stringent (ie, lower) thresholds.Although the use of QALYs as an outcome measure in cost-effectiveness analysis is widely accepted, several pragmatic issues limit the attractiveness of this end point for valuing treatments in which the principal benefit is preventing restenosis after PCI. Because there is no evidence that restenosis affects survival after PCI, one would not expect treatments with the sole benefit of reducing restenosis (such as DES) to improve population-level life expectancy. Furthermore, although restenosis is clearly associated with reduced QOL,27 empirical data as to the overall impact of restenosis on quality-adjusted life expectancy are limited.Given these limitations, several recent studies have used a disease-specific cost-effectiveness ratio: cost per repeat revascularization avoided.13,18,28,29 The advantages of this end point are that it is simple to measure, can be integrated easily into standard data collection for clinical trials or registries, and is readily interpreted by both clinicians and patients. The primary limitation of this end point is that it is specific to coronary revascularization and cannot be compared directly with cost-effectiveness ratios for other conditions or against cost-effectiveness analyses using different outcome measures. Thus, determination of an appropriate cost-effectiveness threshold may be challenging.Within a specific healthcare system, however, comparison with other established technologies that can prevent coronary restenosis may serve as a useful benchmark. For example, within the US healthcare system, several technologies with cost-effectiveness ratios <$10 000 per repeat revascularization avoided (eg, brachytherapy for in-stent restenosis, elective coronary stenting versus balloon angioplasty) have been widely adopted and are currently reimbursed by most third-party payers.18,29 These observations suggest that other therapies with cost-effectiveness ratios <$10 000 per repeat revascularization avoided may be considered reasonably attractive within the US healthcare system. Of note, several recent studies using the contingent valuation (ie, “willingness to pay”) technique suggest that at least in hypothetical scenarios, US patients undergoing cardiac procedures may assign even higher values to therapies that reduce restenosis.30Cost-Effectiveness Analyses of DESAs noted previously, because clinical trials are often performed relatively early in the development process for medical devices, they often represent a unique opportunity to study the cost-effectiveness of such devices under controlled conditions. The first prospective economic evaluation of drug-eluting stents was performed alongside the RAndomized study with the sirolimus-eluting Bx VELocity balloon-expandable stent in the treatment of patients with de novo native coronary artery lesions (Cypher; RAVEL) trial.4,31 In this trial, sponsored by Cordis, 238 patients undergoing PCI of a single, de novo coronary lesion were randomized to receive either sirolimus-eluting stents or a comparable BMS. Clinical outcomes demonstrated that the sirolimus-eluting stent completely eliminated both angiographic and clinical restenosis among this highly select patient population. Economic analysis from the Dutch healthcare perspective demonstrated that initial procedural costs were increased by 1284 , whereas follow-up costs were reduced to a similar extent. The net 1-year cost increment with DES was thus only 54 per patient.This study is of limited relevance to the US healthcare system, however, because both stent prices and the costs associated with repeat revascularization procedures differ substantially between the Netherlands and the United States. Moreover, the RAVEL economic analysis also is limited by the performance of protocol-mandated angiographic follow-up. This practice is well known to inflate rates of repeat revascularization procedures (and their associated costs) through the “oculostenotic reflex.”28,32 As a result, one should be particularly cautious in interpreting the results of economic analyses derived from clinical trials that incorporate a primary angiographic end point.The first study to examine the economic impact of DES from the perspective of the US healthcare system was performed alongside the SIRIUS trial (a Cordis-sponsored trial of the sirolimus-coated Bx Velocity stent [Cypher]).22 In this trial, 1058 patients undergoing single-vessel PCI for moderately complex coronary stenoses were randomized to receive either sirolimus-eluting stents or BMS and were followed up for both clinical and economic outcomes for 1 year. Initial hospital costs were approximately $2800 higher with the sirolimus-eluting stent compared with the BMS ($11 345 versus $8464; P<0.001). Much of this difference in initial costs was offset by lower follow-up costs ($5468 versus $8040; P<0.001), however, mainly because of a reduced requirement for repeat revascularization procedures. Thus, at 12 months, the DES strategy cost an average of $309 per patient more than the BMS strategy, yielding cost-effectiveness ratios of $1650 per repeat revascularization avoided and $27 500 per QALY gained (based on externally derived utility weights). These findings suggest that, for patients at moderate to high risk of restenosis (the population selected for the SIRIUS trial), sirolimus-eluting stents are reasonably cost-effective within the context of the US healthcare system. Of note, this study was conducted in 2004 and assumed an incremental cost of $2000 per DES versus BMS. If the study were to be repeated using 2006 US costs (with an incremental cost per DES of $1600), the resulting outcome would most likely be modest cost savings with the DES strategy.More recently, we completed an economic analysis based on the TAXUS-IV trial of paclitaxel-eluting stents versus BMS for patients undergoing planned single-vessel PCI.33 The general design of the TAXUS-IV trial and its economic analysis were similar to those of the SIRIUS trial. In TAXUS-IV, use of the paclitaxel-eluting stent increased initial hospitalization costs by $2028. This was partially offset by a reduction in 1-year follow-up costs of $1456 in the DES arm, so mean 1-year costs were $572 higher among DES patients ($14 583 versus $14 011; P<0.001). The overall cost-effectiveness ratio for the paclitaxel-eluting stent was $4678 per repeat revascularization avoided and $47 798 per QALY gained. In contrast to the previous studies, TAXUS-IV was the first randomized trial to incorporate a large, prespecified cohort of patients who were managed according to standard clinical practice without mandatory angiographic follow-up. In this 582-patient subset, economic outcomes were actually somewhat more favorable than those for the trial as a whole. The net 1-year cost increment for DES versus BMS in this subgroup was only $97 per patient, with a resulting cost-effectiveness ratio of $760 per repeat revascularization avoided and $5105 per QALY gained. Thus, these findings confirm that, for patients at moderate to high risk of restenosis who are undergoing single-vessel PCI, use of DES represents an economically attractive investment within the US healthcare environment.The most recent economic evaluation of DES versus BMS was performed alongside the BAsel Stent Kosten Effektivitäts Trial (BASKET).34 In BASKET, 826 “unselected” patients undergoing PCI were randomized to receive sirolimus-eluting stents, paclitaxel-eluting stents, or cobalt-chromium stents (BMS). In contrast to both SIRIUS and TAXUS-IV, BASKET enrolled a relatively diverse patient population, including patients undergoing treatment for acute myocardial infarction and patients undergoing multivessel PCI, and used a pragmatic “real-world” design without angiographic follow-up. At the 6-month follow-up, randomization to either DES was associated with a significant reduction in TVR. An economic analysis performed from the perspective of the Swiss healthcare system demonstrated that aggregate 6-month costs were increased by 905 with DES implantation (10 544 versus 9639 ). The incremental cost-effectiveness ratio was 18 311 per repeat revascularization avoided or 72 283 per QALY gained. Given that clinical restenosis after BMS implantation frequently manifests over a 9 to 12 months,21 however, it is unclear whether these relatively unfavorable results relate primarily to the relatively brief analytic time horizon or to other differences in the patient population and management of restenosis in European versus US practice.To overcome the limitations inherent in these trial-based economic analyses, we have developed a decision-analytic model to evaluate the cost-effectiveness of DES for patients undergoing PCI from the perspective of the US healthcare system. Although the original model was based on published data from the medical literature, we have recently updated the model assumptions to reflect population-based PCI outcomes and costs from the Medicare database as previously described.20 Key assumptions of the current PCI model thus include an average TVR rate for BMS of 14%,27 an incremental cost of $1600 per DES,29 and average stent use of 1.6 per procedure.35 Under these baseline conditions and assuming that DES reduce clinical restenosis by 70%,25 the model projects that 1-year aggregate costs with DES would be $691 per patient higher than with BMS, with an incremental cost-effectiveness ratio of $5422 per repeat revascularization avoided. Sensitivity analyses demonstrated that DES would be cost saving for patient populations in which the expected BMS TVR rate is >19% and economically attractive (at a threshold of $10 000 per repeat revascularization avoided) as long as BMS TVR rates were >11% (Figure 2). These results provide credible evidence that, at least on average, use of a DES rather than a BMS may be considered an economically attractive healthcare investment within the context of the US healthcare system. Download figureDownload PowerPointFigure 2. Disease-specific cost-effectiveness for ratios for DES as a function of BMS restenosis rate based on computer-simulation modeling.Further insights into the optimal patient population for DES implantation may be derived from statistical models to predict restenosis after BMS implantation. Several previous studies have demonstrated that smaller reference vessel diameter, greater lesion length, and the presence of treated diabetes mellitus are consistently associated with higher rates of angiographic and clinical restenosis after BMS implantation.21 By incorporating the predicted rates of clinical restenosis based on such a statistical model into our population-based economic outcomes model, we can estimate the projected cost-effectiveness of DES versus BMS for treatment of specific patient subsets (Table). If one considers a cost-effectiveness ratio <$10 000 per repeat revascularization avoided to be acceptable within the US healthcare system, our model suggests that DES are reasonably cost-effective for virtually all diabetic patients. On the other hand, despite their proven clinical benefits, for many nondiabetic patients with reference vessel diameters ≥3.5 mm, our model suggests that DES are not particularly efficient healthcare investments given the relatively low rates of clinical and angiographic restenosis seen with BMS implantation in these subsets. Whether appropriate guidelines and reimbursement policies can be developed to encourage such a tailored approach is currently unknown. Cost-Effectiveness Ratios for DES According to Patient and Lesion CharacteristicsVessel diameter, mmLesion Length, mm1015202530The cost-effectiveness ratio is reported in dollars per repeat revascularization avoided.Diabetic patients 2.5Cost savingCost savingCost savingCost savingCost saving 3.040821876Cost savingCost savingCost saving 3.513 46010 90213 45940822910 4.032 21425 51625 51616 58513 460Nondiabetic patients 2.5 957Cost savingCost savingCost savingCost saving 3.010 902696840821876 957 3.525 51620 49316 58510 9028771 4.055 65741 59141 59125 51625 516Hospital PerspectiveIt is important to note that our discussion thus far has focused on costs as assessed from a societal (or healthcare system) perspective. From these perspectives, the initial higher cost of DES implantation can be recouped, at least in part, through downstream cost savings resulting from a reduction in the need for subsequent revascularization procedures. The situation is quite different when viewed from the perspective of a typical US hospital, however, which is paid for each episode of care and does not realize any direct financial benefits related to improved long-term outcomes. From the perspective of the hospital, economic attractiveness is simply the difference between reimbursement and the cost of the episode of care. Although the Centers for Medicare and Medicaid Services have provided incremental hospital reimbursement (in the form of diagnosis-related group payments) for implantation of DES since their approval in 2003, the difference in reimbursement rates between DES and BMS implantation is approximately $1800,36 which is less than the difference in cost for the 2 procedures. Thus, regardless of the specific cost structure of a hospital, it is clear that profitability is diminished by using drug-eluting stents. Moreover, under the current diagnosis-related group–based reimbursement system, the more stents that are required to treat a specific patient, the greater the financial burden is that is imposed on hospitals by DES adoption.Insufficient third-party reimbursement accounts for only part of the financial challenge faced by hospitals resulting from the introduction of DES, however. Given the benefits of DES in reducing restenosis, hospitals face further loss of revenue resulting from the expected downstream reduction in the need for repeat revascularization procedures.Finally, hospitals have had to face an important loss of revenue resulting from the substitution of less remunerative DES procedures for bypass surgery—traditionally one of the best reimbursed and most profitable procedures for many tertiary hospitals. An analysis based on historical data from Duke University Medical Center projected that the introduction of DES (with 85% penetration over a 5-year period) would reduce overall hospital revenue by more than $5 million per year despite the availability of higher Medicare reimbursements for DES procedures—mainly because of erosion of coronary artery bypass grafting (CABG) volumes.7How Has the Introduction of DES Af