Cost-Effectiveness of Sirolimus-Eluting Stents for Treatment of Complex Coronary Stenoses
Results From the Sirolimus-Eluting Balloon Expandable Stent in the Treatment of Patients With De Novo Native Coronary Artery Lesions (SIRIUS) Trial
Background— Recently, sirolimus-eluting stents (SESs) have been shown to dramatically reduce the risk of angiographic and clinical restenosis compared with bare metal stent (BMS) implantation. However, the overall cost-effectiveness of this strategy is unknown.
Methods and Results— Between February and August 2001, 1058 patients with complex coronary stenoses were enrolled in the SIRIUS trial and randomized to percutaneous coronary revascularization with either a SES or BMS. Clinical outcomes, resource use, and costs were assessed prospectively for all patients over a 1-year follow-up period. Initial hospital costs were increased by $2881 per patient with SESs. Over the 1-year follow-up period, use of SESs led to substantial reductions in the need for repeat revascularization, including repeat percutaneous coronary intervention and bypass surgery. Although follow-up costs were reduced by $2571 per patient with SESs, aggregate 1-year costs remained $309 per patient higher. The incremental cost-effectiveness ratio for SES was $1650 per repeat revascularization event avoided or $27 540 per quality-adjusted year of life gained, values that compare reasonably with other accepted medical interventions. Under updated treatment assumptions regarding available stent lengths and duration of antiplatelet therapy, use of SESs was projected to reduce total 1-year costs compared with BMSs.
Conclusions— Although use of SESs was not cost-saving compared with BMS implantation, for patients undergoing percutaneous coronary intervention of complex coronary stenoses, their use appears to be reasonably cost-effective within the context of the US healthcare system.
Received January 4, 2004; revision received April 20, 2004; accepted April 22, 2004.
Over the past decade, a number of technical and pharmacological advances have dramatically improved the safety and durability of percutaneous coronary intervention (PCI). Foremost among these innovations has been the development of coronary stents. In addition to largely eliminating abrupt vessel closure and the need for emergency bypass surgery, stents have also led to substantial reductions in both angiographic and clinical restenosis compared with balloon angioplasty alone.1,2 Nonetheless, stenting remains limited by restenosis in 20% to 30% of “ideal” lesions, with rates approaching 50% in more complex clinical and anatomical subsets.3
See p 472
Recently, the sirolimus-eluting Bx VELOCITY stent (Cypher, Cordis Corp) has been shown to reduce the rates of angiographic and clinical restenosis compared with conventional bare metal stent implantation in simple and complex coronary stenoses.4,5 As a result, this device was the first drug-eluting stent approved for the prevention of coronary restenosis by the US Food and Drug Administration. Given the high cost of sirolimus-eluting stents (≈$2900 per stent), however, some have questioned their value in the current healthcare environment.6 To address this important issue, we performed a prospective economic evaluation alongside the SIRolImUS-Eluting Balloon Expandable Stent in the Treatment of Patients With De Novo Native Coronary Artery Lesions (SIRIUS) trial.
Patient Population and Treatment Protocol
Between February and August 2001, 1058 patients undergoing PCI of a single native coronary artery were enrolled in the SIRIUS trial at 53 US hospitals. Details of the study design have been described previously.5 Patients were eligible if they were undergoing planned PCI to a de novo lesion 15 to 30 mm in length located in a native coronary artery with a reference vessel diameter between 2.5 and 3.5 mm (by visual estimate). Key exclusion criteria included planned multivessel or staged PCI, myocardial infarction within the previous 48 hours, ostial lesion location, and involvement of a major side branch. The study protocol was approved by the institutional review board at each site, and each patient provided informed consent before enrollment.
Patients were randomized in a double-blind fashion to implantation of either conventional bare metal stents (control group, n=525) or sirolimus-eluting stents (sirolimus group, n=533), stratified by clinical site and the presence of diabetes mellitus. Stent implantation was performed with standard techniques and anticoagulation regimens. After the procedure, all patients remained on aspirin and received clopidogrel (300-mg loading dose followed by 75 mg daily) for ≥3 months. Both the patient and treating physicians remained blinded to treatment group throughout follow-up, even if repeat revascularization was required.
Assessment of In-Hospital Outcomes and Clinical Follow-Up
Case report forms documenting baseline characteristics, procedural details, and clinical outcomes during the initial hospitalization and 1-year follow-up period were completed by a research coordinator at each site. All end points were reviewed by an independent clinical events committee that was blinded to treatment assignment. To limit contamination of clinical outcomes by the performance of routine angiographic follow-up, only clinically indicated repeated revascularization procedures (and their associated costs) were included in the economic analysis. Repeat revascularization was considered clinically indicated if there was evidence of active myocardial ischemia by symptoms, provocative testing, or both.
Determination of Medical Care Costs
Medical care costs for the initial hospitalization and for the 1-year follow-up period were assessed by a combination of “bottom-up” and “top-down” methods as previously described.7
Cardiac Catheterization Laboratory Costs
Detailed resource use was recorded for each procedure, and the cost of each item was estimated on the basis of the mean hospital acquisition cost for the item in 2002. The acquisition costs of bare metal and sirolimus-eluting stents were estimated at $900 and $2900 per stent, respectively, on the basis of a national survey of US hospitals in September 2003 (personal communication, M. Landy, September 15, 2003). Costs of additional disposable equipment, overhead, and depreciation for the cardiac catheterization laboratory were estimated from the average cost per procedure at Beth Israel Deaconess Medical Center in 2002 and adjusted for actual procedure duration.
Other Hospital Costs
All other hospital costs were determined by top-down accounting methods. For 650 randomly selected patients, itemized bills were obtained for the initial hospitalization and any subsequent cardiovascular hospitalizations during the follow-up period. Hospital costs were determined by multiplying itemized hospital charges by the cost center–specific cost-to-charge ratio obtained from the Medicare cost report for the hospital.8 All costs were converted to 2002 dollars on the basis of the medical care component of the Consumer Price Index.
For the remaining 408 patients, nonprocedural hospital costs were imputed from a linear regression model developed from hospital admissions for which complete billing information was available (n=877). Independent variables for this model included length of stay, intensive care unit length of stay, bleeding complications, and revascularization procedures (model R2=0.68). Although cost data are frequently skewed, untransformed costs were used as the dependent variable for these analyses because of the underlying linear relationship between cost and length of stay and because model fit did not improve appreciably when log-transformed costs were used.
Outpatient services were assessed by patient self-report at 3-month intervals, and their costs were estimated from 2002 Medicare reimbursement rates. Outpatient medications were not tracked, except for thienopyridine therapy. Although both treatment groups received 3 months of clopidogrel, patients in the control group were assigned a cost for only 1 month of clopidogrel because this is the predominant practice after bare metal stent implantation.
Discrete data are reported as frequencies, and continuous data are reported as mean±SD. Discrete variables were compared by Fisher’s exact test. Normally distributed continuous variables were compared by Student’s t test, whereas nonnormally distributed data were compared by the Wilcoxon rank-sum test. Cost data are reported as both mean and median values and were compared by t tests. All statistical and cost-effectiveness analyses were performed according to the intention-to-treat principle.
Because the major benefit of the sirolimus-eluting stent was a reduction in the need for repeat revascularization, the primary end point for the cost-effectiveness analysis was the incremental cost per repeat revascularization avoided by sirolimus-eluting stenting compared with conventional stenting. This disease-specific cost-effectiveness ratio was calculated by dividing the difference in mean 1-year medical care costs for the 2 treatment groups by the difference in repeat revascularization rates over the same time frame.9,10
Stratified analyses of 1-year costs, clinical outcomes, and cost-effectiveness were performed according to prespecified subgroups defined by diabetic status, reference vessel diameter, and lesion length. To estimate the relationship between the bare metal stent clinical restenosis rate and the cost-effectiveness of sirolimus-eluting stents, we used logistic regression to predict the need for target lesion revascularization (TLR) in the control group. Independent variables for this model included diabetes, reference vessel diameter, lesion length, left anterior descending artery location, and prior CABG. We then examined the cost-effectiveness of sirolimus-eluting stents within strata defined by each patient’s predicted TLR rate.
Finally, we performed a secondary cost-effectiveness analysis in which the effectiveness measure was the difference in quality-adjusted life expectancy between the 2 treatment groups over the 1-year follow-up period. Because quality of life was not measured directly in this study, quality-adjusted survival for each treatment group was estimated on the basis of observed clinical outcomes using empiric health state–specific utility weights derived from 771 US PCI patients enrolled in the Stent-PAMI trial.9 For this cost-utility analysis, we assumed that there would be no differences in long-term survival or quality of life beyond the first year of follow-up. This assumption is reasonable because there were no significant differences in rates of death or myocardial infarction in the study population and because previous studies have demonstrated no association between coronary restenosis and long-term mortality.11 To estimate the uncertainty surrounding the cost-effectiveness ratios, we calculated bias-corrected CIs by the bootstrap method using 1000 repeated samplings of the study population.
Baseline clinical and angiographic characteristics were well matched for the 2 treatment groups (Table 1). The mean age was 62 years, ≈25% had diabetes, and 41% had multivessel coronary disease. The mean lesion length was 14.4 mm, and 13% of patients had lesion lengths >20 mm. The mean reference vessel diameter (by quantitative angiography) was 2.8 mm, and 28% of patients had reference vessel diameters <2.5 mm.
Initial Treatment Costs
Table 2 summarizes resource use and costs for the index revascularization procedures. Not surprisingly, given the blinded nature of the study, procedural resource use was virtually identical for the 2 treatment groups. An average of 1.4 stents per patient were implanted in both treatment groups. The difference in initial procedural costs was $2856 per patient (95% CI, 2643 to 3069; P<0.001) and was driven entirely by the higher acquisition cost for sirolimus-eluting stents compared with bare metal stents. Similarly, there were no significant differences in initial hospital events or resource use between the 2 treatment groups (Table 3). As a result, hospital costs for the index hospitalization were $2881 higher per patient for the sirolimus group compared with the control group (95% CI, 2533 to 3277; P<0.001), driven largely by the difference in initial revascularization procedure costs.
Follow-up Resource Utilization and Costs
Use of sirolimus-eluting stents was associated with substantial reductions in follow-up resource utilization and related healthcare costs (Table 4). In particular, the need for ≥1 repeat revascularization procedures was reduced by 52% in the sirolimus group compared with the control group (13.7% versus 28.4%; P<0.001), with similar relative benefits in the need for bypass surgery and repeat PCI procedures. These benefits were driven primarily by a 15% absolute reduction in the need for clinically driven TLR (4.9% versus 20.0%; P<0.001). The absolute reduction in the number of revascularization procedures performed (19.4 per 100 patients treated) was even greater because patients in the sirolimus group were less likely to require multiple revascularization procedures during follow-up (2.4% versus 5.7%; P=0.01). The number of hospital admissions and non-protocol–related diagnostic catheterization procedures were also significantly lower in the sirolimus group.
As a result, mean follow-up medical care costs were $2571 per patient lower in the sirolimus group compared with the control group (95% CI, 1338 to 3797; P<0.001; Figure 1). Although these cost savings were substantial, they did not fully offset the higher cost of the initial revascularization procedures. Aggregate 1-year medical care costs remained $309 per patient higher for the sirolimus group compared with the control group (95% CI, $976 less to $1594 more; P=0.64).
In our base-case analysis, the incremental cost-effectiveness ratio for sirolimus-eluting stenting was $1650 per repeat revascularization avoided (Table 5). Bootstrap simulation demonstrated that the cost-effectiveness ratio for sirolimus-eluting stenting remained <$10 000 per repeat revascularization avoided in 98.2% of simulations (Figure 2). In our secondary analysis (in which individual utility values were assigned to the SIRIUS population according to empirically derived weights from the Stent-PAMI trial), the cost-utility ratio for sirolimus-eluting stenting was $27 540 per quality-adjusted life-year (QALY) gained, and it remained <$50 000 per QALY in 63.2% and <$100 000 per QALY in 87.1% of bootstrap simulations.
Because the longest stent length available in the SIRIUS trial was 18 mm but lengths up to 33 mm are currently available, we performed a sensitivity analysis to estimate the impact of longer stents on our findings. We assumed that among patients with lesion lengths 15 to 25 mm who required 2 study stents, 50% would have been treated with a single longer stent. Similarly, we assumed that among patients with lesion lengths >25 mm who required 3 study stents, 50% would have been treated with 2 longer stents. Under these assumptions, the mean number of stents per patient decreased from 1.4 to 1.3 (in both treatment groups), and the 1-year cost difference and cost-effectiveness ratio for the sirolimus-eluting stent fell to $136 per patient and $727 per repeat revascularization avoided, respectively (Table 5). If we further assumed that patients in both the sirolimus and control groups would be treated with 1 year of postprocedure clopidogrel, consistent with the recent results of the CREDO trial,12 use of sirolimus-eluting stents was projected to be cost-saving over the 1-year follow-up period.
Stratified analyses according to baseline patient and lesion characteristics are summarized in Figure 3. There were no significant interactions between treatment assignment and either the incidence of TLR at 1 year or costs for each of the prespecified subgroups (all P>0.05). Nonetheless, there were modest differences in the cost-effectiveness of sirolimus-eluting stents across the various strata. For example, sirolimus-eluting stents were economically dominant (ie, yielding better outcomes and lower costs) in patients with reference vessel diameters <2.5 mm or lesion lengths >20 mm (by operator assessment). On the other hand, sirolimus-eluting stents were somewhat less attractive for patients with reference vessel diameters >3.0 mm (cost-effectiveness ratio ≈$6000 per repeat revascularization avoided). In general, there was an inverse relationship between the predicted TLR rate with bare metal stents and the cost-effectiveness ratio for sirolimus-eluting stents. Nonetheless, the cost-effectiveness ratio for the sirolimus-eluting stent remained <$10 000 per repeat revascularization avoided within each of the subgroups examined.
This is the first prospective economic evaluation of drug-eluting coronary stents within the US healthcare system. In this clinical trial involving ≈1100 patients with complex coronary stenoses, we found that use of sirolimus-eluting stents increased initial hospital costs by >$2800 per patient compared with conventional bare metal stent implantation. Nonetheless, over the 1-year follow-up period, sirolimus-eluting stenting was associated with substantial reductions in a variety of morbid events, including rehospitalization (24 fewer events per 100 patients treated), repeat PCI (17 fewer events per 100 patients treated), and bypass surgery (1.7 fewer events per 100 patients treated) with associated reductions in follow-up medical care costs of >$2500 per patient. These cost savings, however, were insufficient to fully offset the higher initial cost of the sirolimus-eluting stent. Aggregate 1-year costs remained ≈$300 per patient higher for the sirolimus group, with an incremental cost-effectiveness ratio of ≈$1600 per repeat revascularization event avoided.
Although this disease-specific cost-effectiveness ratio cannot be compared directly with standard “league tables” that are generally expressed in terms of cost per QALY gained, several lines of reasoning suggest that sirolimus-eluting stents may be viewed as reasonably attractive, at least within the context of the US healthcare system. Both bare metal stents (versus conventional balloon angioplasty) and vascular brachytherapy (for treatment of in-stent restenosis) have cost-effectiveness ratios of ≈$10 000 per repeat revascularization avoided and are widely used and incrementally reimbursed within the US healthcare system.1,9,13 These data may be considered indirect evidence that the “shadow price” for technologies that reduce coronary restenosis is at least $10 000 per repeat revascularization avoided within the US healthcare system. Moreover, direct assessment of willingness to pay among potential PCI patients using a contingent valuation approach also suggests that a threshold of $5000 to $10 000 per repeated revascularization avoided may be acceptable for the US population.14 Finally, when we estimated quality-adjusted life expectancy for the trial population, the cost-utility ratio for the sirolimus-eluting stent was $27 000 per QALY gained, similar to the cost-effectiveness of other accepted medical interventions.15
Although our analysis suggested that sirolimus-eluting stents do not fully pay for themselves in the long run, it is possible that drug-eluting stents will achieve true cost savings in the near future. In the SIRIUS trial, only 8- and 18-mm stent lengths were available, and mean stent use averaged 1.4 stents per procedure. A conservative reanalysis of stent use considering the availability of stent lengths up to 33 mm (the longest sirolimus-eluting stent currently available in the United States) suggested that mean stent use might be reduced to 1.3 stents per procedure, at least for the SIRIUS population, thus lowering the initial cost increment by ≈$200 per patient. Moreover, in our baseline analysis, we made the conservative assumption that use of a sirolimus-eluting stent would increase the duration of clopidogrel treatment from 1 to 3 months. Because many PCI patients currently receive 9 to 12 months of clopidogrel treatment regardless of stent type, this assumption further biased our analysis against the sirolimus-eluting stent. Finally, with increasing competition, it is reasonable to expect gradual reductions in the incremental cost of drug-eluting stents relative to bare metal stents.
It is important to recognize that our analysis was performed from a societal perspective. Consideration of alternative perspectives might be expected to yield differing conclusions. In particular, economic evaluation of sirolimus-eluting stents from the perspective of a typical US hospital would be far less favorable because hospitals must bear the full acquisition cost of the stents, yet the incremental costs are not fully reimbursed under the current diagnosis-related group system. The fact that widespread use of drug-eluting stents will undoubtedly lead to reductions in repeat PCI procedures (for treatment of in-stent restenosis) and conversion of some highly remunerative bypass surgery procedures to less profitable multivessel drug-eluting stent procedures is likely to further compound the economic hardship to hospitals,16 -at least until reimbursement levels match procedural costs.
Our study has several important limitations. As with any clinical trial, the results of our study may not be generalizable to the full population of PCI patients. In particular, the SIRIUS trial included a large percentage of complex PCI lesions, including long lesions, smaller vessels, and diabetic patients. As a result, the observed clinical restenosis rate was relatively high compared with routine clinical practice.17 Although it is likely that the clinical benefits of sirolimus-eluting stents extend to lower-risk populations, the precise balance between lower incremental costs (because of a smaller number of stents) and smaller cost offsets for such patients is difficult to predict. Further studies, including disease-simulation models, may be useful in extending the results of our empiric study to other lower (and higher) -risk populations.
Second, the performance of mandatory angiographic follow-up in a subset of the study cohort may have increased the incidence of repeat revascularization during follow-up compared with standard clinical practice.1 To limit this bias, all repeat revascularization procedures were reviewed by an independent events committee to determine their clinical necessity, and any procedures (and their associated hospitalizations) judged to be clinically unnecessary were excluded from our analysis.
Finally, our analysis was limited to a 1-year follow-up period. Although longer-term follow-up would have been enlightening, it seems unlikely that the cost-effectiveness of sirolimus-eluting stents would be less favorable with longer follow-up. Previous studies suggest that the restenosis process is largely complete after 12 months,18 so later events are related predominantly to atherosclerosis progression, which would be unlikely to be altered by a drug-eluting stent. Moreover, exclusion of outpatient medication costs (except for clopidogrel) from our analysis may also have biased our analysis slightly against sirolimus-eluting stents. On the other hand, if serious complications such as delayed hypersensitivity or delayed stent thrombosis emerge in longer-term follow-up, their long-term health consequences could overwhelm the benefits of avoiding restenosis in the short term and render sirolimus-eluting stents less attractive on both clinical and economic grounds.9
In patients with complex stenoses at relatively high risk of restenosis, use of sirolimus-eluting stents reduced the rate of clinically significant restenosis by ≈75% with an associated cost offset of >$2500 per patient compared with bare metal stent implantation. Although 1-year costs were ≈$300 per patient higher with sirolimus-eluting stenting, its incremental cost-effectiveness ratio compares favorably with other accepted medical interventions. These findings suggest that sirolimus-eluting stents are reasonably cost-effective for the target population of the SIRIUS trial.
Study funding was provided in part by a grant from Cordis, Inc. Dr Bakhai was supported by grants from the Royal Brompton Hospital and Beth Israel Deaconess Medical Center.
Moses JW, Leon MB, Popma JJ, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med. 2003; 34: 1315–1323.
O’Neill WW, Leon MB. Drug-eluting stents: costs versus clinical benefit. Circulation. 2003; 107: 3008–3011.
Cohen DJ, Krumholz HM, Sukin CA, et al. In-hospital and one-year economic outcomes after coronary stenting or balloon angioplasty: results from a randomized clinical trial. Circulation. 1995; 92: 2480–2487.
Cohen DJ, Taira DA, Berezin RH, et al. Cost-effectiveness of coronary stenting in acute myocardial infarction: results from the Stent-PAMI trial. Circulation. 2001; 104: 3039–3045.
Cohen DJ, Cosgrove RS, Berezin RH, et al. Cost-effectiveness of gamma radiation for treatment of in-stent restenosis: results from the Gamma-1 trial. Circulation. 2002; 106: 691–697.
Weintraub WS, Ghazzal ZM, Douglas JS Jr, et al. Long-term clinical follow-up in patients with angiographic restudy after successful angioplasty. Circulation. 1993; 87: 831–840.
Seto TB, Cohen DJ. Cost-effectiveness of adjunctive intracoronary brachytherapy for the treatment of in-stent restenosis. Vascular Radiotherapy Monitor. 2001; 4: 9–15.
Greenberg D, Bakhai A, Neumann PJ, et al. Willingness to pay for avoiding coronary restenosis and repeat revascularization: results from a contingent valuation study. Health Policy. In press.
Mark DB, Hlatky MA. Medical economics and the assessment of value in cardiovascular medicine, part I. Circulation. 2002; 106: 516–520.
Williams DO, Holubkov R, Yeh W, et al. Percutaneous coronary intervention in the current era compared with 1985–1986: the National Heart, Lung, and Blood Institute Registries. Circulation. 2000; 102: 2945–2951.