Author + information
- Received January 6, 2006
- Revision received March 23, 2006
- Accepted March 28, 2006
- Published online October 3, 2006.
- Francesco Saia, MD, PhD⁎,⁎ (, )
- Giancarlo Piovaccari, MD†,
- Antonio Manari, MD‡,
- Andrea Santarelli, MD†,
- Alberto Benassi, MD§,
- Enrico Aurier, MD∥,
- Pietro Sangiorgio, MD¶,
- Fabio Tarantino, MD#,
- Giuseppe Geraci, MD⁎⁎,
- Giuseppe Vecchi, MD††,
- Paolo Guastaroba, MSc‡‡,
- Roberto Grilli, MD‡‡ and
- Antonio Marzocchi, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Francesco Saia, Catheterization Laboratory, Institute of Cardiology—University of Bologna, Policlinico S.Orsola-Malpighi (Pad 21), Via Massarenti, 9, 40138 Bologna, Italy.
Objectives We compared the clinical outcome of sirolimus-eluting stents (SES) and paclitaxel-eluting stents (PES) in a real-world scenario.
Background In selected patients, SES has been associated with lower late luminal loss than PES. Whether this emerging biological difference could translate into different clinical efficacy in daily practice is presently unknown.
Methods This analysis included 1,676 consecutive patients with de novo coronary lesions treated solely with drug-eluting stents (SES = 992; PES = 684). All patients were enrolled in a dynamic prospective registry comprising 13 hospitals. We assessed the cumulative incidence of major adverse cardiac events (MACE), defined as death, myocardial infarction (MI), and target vessel revascularization (TVR) during follow-up.
Results Overall, 29% of the patients had diabetes, 23% had prior MI, and 9% had poor left ventricular function. ST-segment elevation MI was diagnosed at admission in 12%. Multivessel intervention was performed in 16%. At 1-year follow-up, SES was associated with a reduced incidence of MACE (9.2% SES vs. 14.1% PES; p = 0.007) and TVR (5.0% SES vs. 10.0% PES; p = 0.0008) compared to PES. A propensity analysis with many clinical and angiographic variables was carried out to adjust for baseline differences. In this analysis, SES was associated with a 44% risk reduction of MACE (hazard ratio 0.56, 95% confidence interval 0.39 to 0.78) and a 55% reduction of TVR (hazard ratio 0.45, 95% confidence interval 0.29 to 0.70). This result was consistent across most subgroups tested. Similar rates of death and MI were observed in the 2 treatment groups.
Conclusions In this large real-world population, SES improved 1-year clinical results as compared to PES.
A number of clinical trials have demonstrated a clear superiority of the sirolimus-eluting stent (SES) and polymer-coated paclitaxel-eluting stent (PES) over a bare metal stent for the prevention of restenosis and the need for further revascularization (1–5). In head-to-head comparisons, SES has been consistently associated with lower in-stent late luminal loss than PES (6–9). However, the importance of this angiographic end point as a surrogate of clinical events has been questioned.
Indeed, the interrelationship among angiographic late loss, binary restenosis, and clinical recurrence after coronary stent implantation has been incompletely evaluated, and this holds especially true in the drug-eluting stent era. However, although small differences in luminal late loss did not seem to be important in the setting of nearly ideal patients and lesions treated in randomized trials, they might be pivotal in more complex patients and lesions as treated in the real world. Accordingly, SES has been associated with a clinical advantage over PES only in trials enrolling patients at high risk of restenosis.
The present study was therefore conducted to compare the 1-year clinical outcome of patients treated with SES and PES in a large multicenter registry comprising patients with a broad variety of clinical characteristics and lesion characteristics.
Study design and patient population
The REAL registry (REgistro regionale AngiopLastiche dell’Emilia-Romagna) has been previously described (10). Briefly, the REAL is a large prospective web-based registry launched in July 2002 and designed to collect clinical and angiographic data of all consecutive percutaneous coronary interventions (PCIs) performed in a 4-million-resident Italian region. Thirteen public and private centers of interventional cardiology participate in data collection. The present study focuses on all patients who were treated exclusively with the Cypher SES (Cordis, Johnson and Johnson, Miami Lakes, Florida) or the Taxus PES (Boston Scientific, Natick, Massachusetts) for de novo coronary lesions between July 2003 and December 2004.
The REAL registry is based on current clinical practice; therefore, the local hospital ethics committees required only an ordinary written informed consent to coronary intervention, which was obtained from all patients. The protocol of the study is in accordance with the Declaration of Helsinki.
Procedures and post-intervention medications
Interventional strategy and device utilization, including drug-eluting stent type, were left to the discretion of the attending physicians. Periprocedural glycoprotein IIb/IIIa inhibitors and antithrombotic medications were used according to the operator’s decision and current guidelines. Lifelong aspirin was prescribed to all patients. At least 2-month ticlopidine (250 mg bid) or clopidogrel treatment (75 mg/day) was recommended to all patients treated with SES, whereas the same treatment was extended to at least 6 months for patients treated with PES.
Definitions and follow-up
The primary end point of the survey was the occurrence of major adverse cardiac events (MACE), defined as: 1) death (cardiac and non-cardiac); 2) non-fatal acute myocardial infarction (MI); and 3) target vessel revascularization (TVR). Myocardial infarction during follow-up was diagnosed by local cardiologists at the hospital of admission according to standard criteria (increased levels of troponin or creatinine kinase-MB fraction in association with chest pain and/or ischemic electrocardiographic changes). Target vessel revascularization was defined as any re-intervention (surgical or percutaneous) to treat a luminal stenosis occurring in the same coronary vessel treated at the index procedure, within and beyond the target lesion limits. Thrombotic stent occlusion was angiographically documented as a complete occlusion (Thrombolysis In Myocardial Infarction flow grade 0 or 1) or a flow-limiting thrombus (Thrombolysis In Myocardial Infarction flow grade 1 or 2) of a previously successfully treated artery. In addition, we defined as “possible stent thrombosis” the occurrence of acute MI in the territory of the vessel treated and unexplained sudden cardiac death. Lesion length and vessel reference diameter were visually estimated by the operators. Online quantitative coronary analysis was allowed if required by the attending physician.
Follow-up was obtained directly and independently from the Emilia-Romagna Regional Health Agency through the analysis of the hospital discharge records and the municipal civil registries. All repeat interventions during follow-up (either surgical or percutaneous) were prospectively collected from the single institutions and matched with the administrative data to adjust for eventual inconsistency. Hospital records were reviewed for additional information whenever deemed necessary. Specific queries were sent to the single institution to justify/correct discrepancies between administrative data, largely provided by independent cardiologists, and data derived from the web-based PCI database, compiled by the interventional cardiologists.
Continuous variables were expressed as mean ± SD and were compared using an unpaired Student ttest. Categorical variables were expressed as counts and percentages and the chi-square test was used for comparison. The cumulative incidence of adverse events was estimated according to the Kaplan-Meier method and compared by the log-rank test. Because of the observed differences in baseline characteristics between the treatment groups, a propensity score analysis was carried out by use of a logistic regression model for treatment with SES versus PES.
This analysis included a number of clinical, angiographic, and procedural variables, such as age, gender, Charlson comorbidity index, diabetes mellitus, hypercholesterolemia, prior angioplasty, prior MI, prior coronary artery bypass graft, low (<35%) left ventricular ejection fraction, diagnosis at admission (ST-segment elevation MI, unstable angina/non–Q-wave MI, stable coronary disease), target vessel, left main stenting, number of lesions treated, reference vessel diameter, total lesion length, ostial lesion, chronic total occlusion, and bifurcation.
The logistic model by which the propensity score was estimated showed good predictive value (C-statistic = 0.741) and calibration characteristics by the Hosmer-Lemeshow test (p = 0.56). The score was then incorporated into subsequent proportional-hazards models as a covariate. Cox proportional hazards models were used to assess relative risk of adverse events in subgroups of patients. There were 3 hospitals using only 1 type of stent (2 SES = 255 patients; 1 PES = 130 cases). Such a strong relationship between center and treatment made problematic the inclusion of the variable “center” in the propensity score. However, to rule out a possible bias, we performed a sensitivity analysis excluding the 3 centers using only 1 type of stent and adjusting the comparison between SES and PES through a propensity score that included as covariates the other hospitals. All analyses were performed with the SAS 8.2 system (SAS Institute, Cary, North Carolina).
The authors had full access to the data and take responsibility for its integrity. All authors have read and agreed to the manuscript as written.
In the study period, 2,539 patients enrolled in the registry received at least 1 drug-eluting stent to treat de novo lesions (of which 1,556 received SES and 983 PES). Patients who received both types of stents and patients also treated with bare-metal stents were excluded. Therefore, there were 1,676 eligible patients (SES, n = 992; PES, n = 684) with 2,130 lesions (SES, n = 1,175; PES, n = 955). Baseline and procedural characteristics of this population are shown in Tables 1 and 2.⇓⇓Overall, 29% of the patients had diabetes, 23% had prior MI, and 9% had poor left ventricular function. ST-segment elevation MI was diagnosed at admission in 12%, whereas multi-vessel PCI was performed in 16%. Few, though remarkable, differences were observed between the two treatment groups. Prior MI and poor left ventricular ejection fraction were more frequently observed in the SES group, which showed a higher Charlson comorbidity index as well (Table 1). The SES was also used more than PES in patients admitted with ST-segment elevation MI (SES 14.3% vs. PES 9.6%; p = 0.004) and in the treatment of the left anterior descending coronary artery (60.4% SES vs. 51.5% PES; p < 0.0001), type C lesions (37.1% SES vs. 20.8% PES; p < 0.0001), bifurcations (21.5% SES vs. 15.8% PES; p = 0.0008), and chronic total occlusion (12.8% SES vs. 7.5% PES; p = 0.0002).
Conversely, the PES was more used to accomplish multi-vessel interventions (11.6% SES vs. 22.2% PES; p < 0.0001). Accordingly, the average number of lesions treated (1.2 ± 0.4 SES vs. 1.4 ± 0.7 mm PES; p < 0.0001) and total lesion length were greater in the PES group. However, in the SES group, individual lesions were longer (18.9 ± 9.3 mm SES vs. 16.8 ± 8.2 mm PES; p = 0.0001) and reference vessel diameter smaller (2.8 ± 0.4 mm SES vs. 2.9 ± 0.4 mm PES; p = 0.0001) than in the PES group. Complete procedural success was achieved in >99% of the procedures in both groups.
Median follow-up was 296 days (range 90 to 639 days). The 1-year cumulative incidence of death (2.5% SES vs. 2.7% PES; p = 0.44) and MI (3.8% SES vs. 4.2% PES; p = 0.77) was similar in the 2 groups (Table 3).However, rates of TVR (5.0% SES vs. 10.0% PES; p = 0.0008) and target lesion revascularization (TLR) (3.4% SES vs. 6.9% PES; p = 0.006) were significantly lower in the SES group, as was the cumulative incidence of MACE (9.2% SES vs. 14.1% PES; p = 0.007). The incidence of angiographic stent thrombosis was 0.7% in both groups (p = 0.6). Possible stent thrombosis (sudden death or acute MI in the territory of the same vessel treated) occurred in 0.7% of the SES group and 0.8% of the PES group (p = 0.46).
Propensity score analysis
To adjust for differences in baseline clinical and angiographic characteristics, a propensity score analysis of the data was carried out as previously described. As shown in Figure 1,this analysis confirmed a similar incidence of death and MI between the 2 cohorts, and a lower incidence of TVR, TLR, and MACE in the SES group. The separate analysis performed after exclusion of the 3 centers using only 1 of the 2 stents gave similar results, although with reduced statistical power (MACE: hazard ratio [HR] 0.61, 95% confidence interval [CI] 0.40 to 0.93, p = 0.02; TVR: HR 0.59, 95% CI 0.33 to 1.03, p = 0.06).
The impact of SES and PES implantation on the risk of subsequent TVR in specific subsets is shown in Figure 2.Sirolimus-eluting stents were associated with a similar risk reduction across many subgroups. Not surprisingly, the benefit of SES appeared significantly more pronounced in small vessels (<2.5 mm, HR 0.13, 95% CI 0.03 to 0.68) and long lesions (>20 mm, HR 0.35, 95% CI 0.17 to 0.73). Conversely, in this study the effect of SES and PES in diabetic patients was similar (HR 0.97, 95% CI 0.45 to 2.10).
The major finding of this study is that in a real-world complex population, SES is associated with a lower risk of reinterventions compared to the PES. This result deserves some attention, because a very intense debate is ongoing about the relative performance of these 2 drug-eluting stents in clinical practice.
In randomized head-to-head comparisons, the SES has been consistently associated with superior suppression of neointimal hyperplasia compared to the PES (6–9). However, the reduction of in-stent and in-segment late loss was not always paralleled by a reduction of binary restenosis, need for repeat revascularization, and MACE. In fact, in the SIRTAX (Sirolimus-Eluting Versus Paclitaxel-Eluting Stents for Coronary Revascularization) trial the 9-month incidence of all MACE, TLR, and TVR was significantly lower in the SES than in the PES group, whereas in the larger multicenter REALITY trial, neither binary restenosis nor MACE and TVR rates were reduced by the SES as compared to the PES (7). Previous studies showed that small but clinically important differences in bare-metal stent performance might become more obvious as the patients’ type and lesions treated become more complex (11). This hypothesis also was postulated for comparison between drug-eluting stents (12). Maximal suppression of neointimal hyperplasia is likely to be particularly important, for example, in small vessels (13), which can accommodate less tissue inside the stent, and in diabetic patients, who exhibit an exaggerated neointimal proliferative response following PCI (14). Indeed, a clinical advantage of the SES over PES was observed in those trials focused on complex patients such as the ISAR-DESIRE (Intracoronary Stenting and Angiographic Results—Drug-Eluting Stents for In-Stent Restenosis) (6), a study comparing SES, PES, and balloon angioplasty for prevention of recurrences in patients with coronary in-stent restenosis; and in the SIRTAX trial (8), which enrolled a relatively complex group of patients. Thus, because increased patient and lesion complexity is reflected in everyday “real-world” PCI, this has to be considered the ideal scenario for comparison between competing drug-eluting stents.
In this context, the result of this large, multicenter, daily-practice registry supporting some clinical advantage of SES over PES is noteworthy. In other words, the angiographic superiority of SES may have a concrete impact on clinical outcome in the general population. The relationship between angiographic late loss and TVR in the drug-eluting stent era remains a large outstanding question. Whereas late loss has been shown to be monotonically related to restenosis risk even in drug-eluting stent studies (15), coronary stents result in large lumens with “room” to accommodate up to approximately 0.5 to 0.65 mm of tissue before the likelihood of clinical restenosis increases substantially (16). Hence, both SES and PES are well below this threshold, which can justify the similar clinical outcome observed in some randomized trials with selective inclusion criteria. Conversely, the 55% risk reduction of TVR observed with SES compared to PES in our registry indicates that small differences in late lumen loss may become relevant in complex patients and lesions. The results of our study are concordant with a recent meta-analysis of all head-to-head randomized trials (17), which showed that patients receiving SES had a significant lower risk of restenosis and TVR compared with those receiving PES.
Conversely, conflicting results emerged from other registries. A non-significant trend toward fewer revascularizations in SES-treated patients was observed in the RESEARCH (Rapamycin-Eluting Stent Evaluated At Rotterdam Cardiology Hospital) registry (18), whereas no differences between the two devices were noted in the large STENT (Strategic Transcatheter Evaluation of New Therapies) registry (C. Simonton, personal communication, Transcatheter Cardiovascular Therapeutics 2005) and in the Milan registry (19). However, it should be noted that in the RESEARCH registry a large proportion of patients treated with SES, mainly those with high-risk features (20), underwent routine angiography at follow-up (38%), and this was not the case for the PES group. Therefore, the negative impact of repeat angiography in SES-treated patients might have attenuated the difference between the 2 devices. On the other side, only preliminary 9-month results have been presented from the STENT registry, and this follow-up may be too short to detect differences in clinical restenosis rates (21). Finally, data from the Milan registry are important because they represent a really complex population, but a note of caution in interpretation of results is mandatory given the relatively limited number of patients enrolled (around 500).
Notably, in the REAL registry, the SES was found to be superior to the PES in most subgroups analyzed (Fig. 2). As expected, the positive effect of SES was more evident in patients with complex coronary anatomy such as those with long lesions and small vessels. Conversely, virtually equal results were observed in diabetic patients. With the inherent limitations of subgroup analyses, this might reflect a similar clinical efficacy of the 2 devices in these patients. Indeed, in the ISAR-DIABETES trial, SES reduced late loss and binary restenosis compared to the PES, but the reduction of clinical events was not statistically significant (9). In addition, diabetic patients are known to have a higher symptomatic threshold for angina pectoris, and in the REAL registry virtually all reinterventions can be considered clinically driven. Therefore, a number of diabetic patients may have developed silent restenosis, thus decreasing the chance to detect a clinical difference between stents in this subgroup. Accordingly, in a subgroup analysis of the RESEARCH registry, clinically driven TVR was not significantly reduced by SESs in comparison to bare-metal stents in diabetic patients, who did not undergo routine angiographic follow-up (22).
The REAL registry confirms once again the effectiveness of both SES and PES in the prevention of restenosis and new revascularizations. Indeed, the 1-year incidence of TVR in the 2 groups was remarkably low (5% to 10%), despite the complex characteristics of the population enrolled, substantiating the results of previous studies and closely resembling the results of the real-world SIRTAX trial (8).
A possibly different safety profile between the SES and the PES has been evoked by the REALITY trial. The investigators evaluated the number of acute and subacute stent thromboses within the first 30 days of the trial, detecting a 4-fold statistically significant increase in the number of stent thromboses in the patients who received the PES compared with those who received the SES (1.8% vs. 0.4%, p = 0.0196) (7). In our registry, we did not find a significant difference in the incidence of stent thrombosis over 1 year between the 2 stents (both angiographically documented and clinical stent thrombosis, including sudden death and acute MI in the same territory of the vessel treated), confirming previous analyses of data from randomized trials (23) and large registries (24). Both drug-eluting stents have been associated with delayed endothelialization and signs of persistent local inflammation, although the therapeutic window of PES may be somewhat narrower (25). Although a precise analysis of actual antiplatelet therapy in our registry was not performed, longer dual antiplatelet therapy prescribed by design to patients receiving PES may have contributed to limit and equalize the incidence of stent thrombosis.
This study suffers the obvious limitations of observational non-randomized studies. On the other hand, it carries important and complementary information derived from a real-world registry, given the inclusion of patients and lesions often excluded from randomized trials. The exclusion from this analysis of patients also treated with bare-metal stents did not give a precise picture of different utilization profiles of drug-eluting stents in the REAL registry. However, consistent with this report, in the entire REAL population (2,539 patients treated with drug-eluting stents in the same period), SES was used more than PES in higher risk lesions (longer lesions and smaller reference diameter); in more compromised patients; and in patients with acute MI, poor left ventricular function, chronic total occlusions, or bifurcations. The PES was preferred for left main treatment and multi-vessel intervention. This situation is not surprising, given the different body of evidence available for the 2 stents at the time of enrollment, different sizes of stents available, and the different market strategies of the 2 producing companies. Remarkably, in this larger population SES implantation was associated with a similar reduction of revascularizations and cardiac events, both in the unadjusted and in the propensity score-adjusted analyses (data not shown). Finally, although propensity analyses do not completely overcome the pitfalls of non-randomized comparisons assessing the effectiveness of health care interventions, they are known to be a valuable approach for taking adequately into account the potential confounding effect attributable to between-groups imbalances in case mix (26).
In this large and complex real-world population, the use of SES reduced the 1-year incidence of adverse cardiac events as compared with PES, mainly by decreasing the need for repeat revascularizations.
For the list of investigators, please see the online version of this article.
This study was supported by the Regional Health Agency of Emilia-Romagna, Bologna, Italy.
- Abbreviations and Acronyms
- major adverse cardiac events
- (acute) myocardial infarction
- percutaneous coronary intervention
- paclitaxel-eluting stent
- REgistro regionale AngiopLastiche dell’Emilia-Romagna
- sirolimus-eluting stent
- target lesion revascularization
- target vessel revascularization
- Received January 6, 2006.
- Revision received March 23, 2006.
- Accepted March 28, 2006.
- American College of Cardiology Foundation
- Colombo A.,
- Drzewiecki J.,
- Banning A.,
- et al.
- Edelman E.R.,
- Rogers C.
- Mehilli J.,
- Dibra A.,
- Kastrati A.,
- Pache J.,
- Dirschinger J.,
- Schomig A.
- Kornowski R.,
- Mintz G.S.,
- Kent K.M.,
- et al.
- Mauri L.,
- Orav E.J.,
- Kuntz R.E.
- Ellis S.G.,
- Popma J.J.,
- Lasala J.M.,
- et al.
- Ong A.T.,
- Serruys P.W.,
- Aoki J.,
- et al.
- Lemos P.A.,
- Hoye A.,
- Goedhart D.,
- et al.
- Cutlip D.E.,
- Chauhan M.S.,
- Baim D.S.,
- et al.
- Lemos P.A.,
- Serruys P.W.,
- Van Domburg R.T.,
- et al.,
- The rapamycin-eluting stent evaluated at Rotterdam Cardiology Hospital (RESEARCH) registry
- Moreno R.,
- Fernandez C.,
- Hernandez R.,
- et al.
- Ong A.T.,
- Hoye A.,
- Aoki J.,
- et al.
- Finn A.V.,
- Kolodgie F.D.,
- Harnek J.,
- et al.