Author + information
- Received July 1, 2008
- Revision received September 22, 2008
- Accepted September 22, 2008
- Published online February 24, 2009.
- Anne Kaltoft, MD, PhD⁎,⁎ (, )
- Lisette Okkels Jensen, MD, PhD†,
- Michael Maeng, MD, PhD⁎,
- Hans Henrik Tilsted, MD‡,
- Per Thayssen, MD, DMSci†,
- Morten Bøttcher, MD, PhD⁎,
- Jens Flensted Lassen, MD, PhD⁎,
- Lars Romer Krusell, MD⁎,
- Klaus Rasmussen, MD, DMSci‡,§,
- Knud Nørregaard Hansen, MD⁎,
- Lars Pedersen, MSc∥,
- Søren Paaske Johnsen, MD, PhD§∥,
- Henrik Toft Sørensen, MD, PhD, DMSci∥,¶ and
- Leif Thuesen, MD, DMSci⁎
- ↵⁎Reprint requests and correspondence:
Dr. Anne Kaltoft, Department of Cardiology, Aarhus University Hospital, Skejby, Brendstrupgaardsvej 100, 8200 Aarhus N, Denmark
Objectives This registry study assessed the safety and efficacy of the 2 types of drug-eluting stents (DES), sirolimus-eluting stents (SES) and paclitaxel-eluting stents (PES), compared with bare-metal stents (BMS).
Background Drug-eluting stents may increase the risk of stent thrombosis (ST), myocardial infarction (MI), and death.
Methods A total of 12,395 consecutive patients with coronary intervention and stent implantation recorded in the Western Denmark Heart Registry from January 2002 through June 2005 were followed up for 2 years. Data on death and MI were ascertained from national medical databases. We used Cox regression analysis to control for confounding.
Results The 2-year incidence of definite ST was 0.64% in BMS patients, 0.79% in DES patients (adjusted relative risk [RR]: 1.09; 95% confidence interval [CI]: 0.72 to 1.65), 0.50% in SES patients (adjusted RR: 0.63, 95% CI: 0.35 to 1.15), and 1.30% in PES patients (adjusted RR: 1.82, 95% CI: 1.13 to 2.94). The incidence of MI was 3.8% in BMS-treated patients, 4.5% in DES-treated patients (adjusted RR: 1.24, 95% CI: 1.02 to 1.51), 4.1% in SES-treated patients (adjusted RR: 1.15, 95% CI: 0.91 to 1.47), and 5.3% in PES-treated patients (adjusted RR: 1.38, 95% CI: 1.06 to 1.81). Whereas overall 2-year adjusted mortality was similar in the BMS and the 2 DES stent groups, 12- to 24-month mortality was higher in patients treated with PES (RR 1.46, 95% CI: 1.02 to 2.09). Target lesion revascularization was reduced in both DES groups.
Conclusions During 2 years of follow-up, patients treated with PES had an increased risk of ST and MI compared with those treated with BMS and SES. Mortality after 12 months was also increased in PES patients.
By reducing neointimal hyperplasia, drug-eluting stents (DES) are more effective than bare-metal stents (BMS) in preventing coronary restenosis (1–4), but may predispose to stent thrombosis through incomplete tissue coverage of the stent struts. Although the first randomized clinical trials comparing DES and BMS raised no safety concerns (5–7), long-term DES performance was questioned in studies reporting an increased risk of myocardial infarction (MI) and death in DES-treated patients (8–11).
An important question in the BMS versus DES controversy is the class effect issue of DES treatment, because there are major differences between the first commercialized DES, the Cypher sirolimus-eluting stent (SES) (Cypher, Cordis Corp., Johnson & Johnson, Warren, New Jersey), and the Taxus paclitaxel-eluting stent (PES) (Boston Scientific, Natick, Massachusetts) (5,8,10).
In an earlier publication (12), we observed an increased risk of stent thrombosis 12 to 15 months after implantation. In the present study, we extended the follow-up period to 2 years and addressed the possible DES class effect issue by reporting separate data on patients treated with SES or PES.
Setting and design
The cohort study's data originate from the WDHR (Western Denmark Heart Registry) and other national databases. The WDHR collects patient- and procedure-specific information on coronary interventions performed at the 3 coronary intervention centers in Western Denmark (Odense University Hospital, Aarhus University Hospital, Skejby, and Aarhus University Hospital, Aalborg). National databases include the Danish Civil Registration System, the National Registry of Causes of Deaths, and the National Patient Registry, covering the region's population (approximately 3 million inhabitants, 55% of the Danish population). A detailed description of these databases has been published previously (12). All patients in the cohort were followed up for 2 years.
Patients and procedures
We used the WDHR to identify all percutaneous coronary interventions (PCIs) between January 1, 2002, and June 30, 2005. For each patient we included only the first PCI procedure performed during the study period (the index procedure). The stents were selected at the operators’ discretion. The indication for the use of DES changed over time, from restenosis-prone patients/lesions at the beginning of the study period to an all-comers use. There was no difference in recommendations for the use of SES and PES. We excluded patients receiving balloon angioplasty without stent implantation or a combination of BMS and DES (n = 645, 4.9%). In the analyses of SES and PES, we excluded patients receiving both stent types (n = 48, 0.04%). Post-PCI antiplatelet regimens included lifelong acetylsalicylic acid (75 to 150 mg daily) and clopidogrel with a loading dose of 300 mg followed by 75 mg daily. Since November 2002, the recommended duration of clopidogrel treatment has been 12 months.
Study end points and definitions
Study end points were time to stent thrombosis, all-cause mortality, cardiac death, MI, and target lesion revascularization (TLR), as previously reported (12). We used the Academic Research Consortium (ARC) definition of stent thrombosis, with a modification for probable stent thrombosis (13,14).
We defined new MIs as hospitalization for MI occurring >28 days after the index PCI (15). We ascertained admissions and readmissions for MI (International Statistical Classification of Diseases and Related Health Problems-10th Revision, codes I21 to I21.9) from the Danish National Patient Registry (12) and deaths from the Civil Registration System. We used the original death certificates obtained from the National Registry of Causes of Deaths to classify deaths according to their underlying cause.
From the WDHR we ascertained TLR, defined as a repeat PCI or coronary artery bypass grafting, occurring within 2 years after the index stent implantation. We assessed all clinical end points occurring within 2 years of the index PCI. An expert committee reviewed relevant records and adjudicated the end points regarding stent thrombosis and cause of death.
We retrieved data from the WDHR on potential predictors of subsequent cardiovascular events, including patient characteristics, procedures performed, and lesion type. We obtained hospital diagnoses for each patient from the National Patient Registry from 1977 until the date of stent implantation, and computed the patients' Charlson Comorbidity Index scores, covering 19 major disease categories (16).
Data were more than 95% complete for patient and procedure characteristics and 100% complete for the clinical end points (stent thrombosis, death, and MI).
We estimated the cumulative incidence for each end point in the presence of competing risk. We used Cox proportional hazards regression to compute hazard ratios (HRs) as a measure of the relative risks (RRs) for each end point. Because the hazards were not proportional throughout the follow-up period, we computed the HR estimates within separate time windows, for which the proportionality assumption held. The HRs in these subanalyses reflected the risk among patients alive and at risk of a given end point at the start of each time period. In regression analyses, we controlled for age, sex, diabetes mellitus, PCI indication, and procedure time. To improve the precision of the estimates, we used the change-in-estimate method, in which we retained only those variables that changed HR estimates for an outcome by more than 10% (17). In the lesion-specific analyses (stent thrombosis and TLR), we also adjusted for stent length and reference vessel size.
We compared distributions of continuous variables with either the 2-sample ttest or the Mann-Whitney Utest, depending on whether data conformed to a normal distribution. We compared distributions of categorical variables using the chi-square test. We used SAS software version 9.13 (SAS Institute Inc., Cary, North Carolina) to analyze the data.
The study encompassed 12,395 consecutive patients with 17,152 lesions. Of these, 3,500 patients with 5,417 lesions received DES, either SES (n = 3,426, 63.2%) or PES (n = 1,991, 36.8%), and 8,847 patients with 11,730 lesions were treated with BMS. The patients' median age was 64 years (interquartile range 56 to 72 years), 16% of patients were older than 75 years, and 13% had diabetes mellitus. Indications for PCI were ST-segment elevation myocardial infarction (STEMI) (30%), non–STEMI/unstable angina (30%), stable angina (37.0%), and other (3%). Baseline patient, procedure, and lesion characteristics differed substantially between the DES and BMS groups, but were similar in the SES and PES groups (Tables 1 and 2).⇓
Definite stent thrombosis
The incidence of definite stent thrombosis was similar in the DES- and BMS-treated groups, with definite stent thrombosis occurring in 43 lesions in 43 patients treated with DES (2-year incidence 0.79%) and in 75 lesions in 75 patients treated with BMS (2-year incidence 0.64%, adjusted RR: 1.09, 95% confidence interval [CI]: 0.72 to 1.65). The incidence of acute, subacute, and late definite stent thrombosis also was similar in the 2 groups (Fig. 1A).Very late definite stent thrombosis occurred in 13 lesions in 13 patients in the DES group (2-year incidence 0.24%) and in 5 lesions in 5 patients in the BMS group (2-year incidence 0.04%, adjusted RR: 5.25, 95% CI: 1.84 to 14.98) (Table 3).None of the 118 cases of definite stent thrombosis occurred in saphenous vein grafts.
Analyzing the SES and PES groups separately revealed different risks of stent thrombosis. The SES group did not significantly differ from the BMS group with respect to definite stent thrombosis, whereas the PES group had a substantially increased risk.
Predictors of definite stent thrombosis
The risk of definite stent thrombosis was increased by STEMI (adjusted RR: 3.14, 95% CI: 1.93 to 5.10), stent length (adjusted RR per additional millimeter of stent length: 1.02, 95% CI: 1.00 to 1.05), and procedure time (adjusted RR per additional minute of procedure time: 1.01, 95% CI: 1.00 to 1.02).
Definite stent thrombosis and antiplatelet therapy
Among the 118 patients in whom definite stent thrombosis developed, 92 (78%) were treated with dual antiplatelet therapy (aspirin and clopidogrel) at the time of the thrombotic event. In the 18 patients with very late stent thrombosis, 3 patients (17%) were receiving dual antiplatelet therapy, 11 patients (61%) were treated with aspirin only, and 4 patients (22%) had discontinued both aspirin and clopidogrel.
Definite, probable, or possible stent thrombosis
Definite, probable, or possible stent thrombosis was found in 81 patients treated with DES (2-year incidence 2.28%) and in 218 patients treated with BMS (2-year incidence 2.46%, RR: 1.07, 95% CI: 0.82 to 1.39). After controlling for covariates, the risk of stent thrombosis (definite, probable, or possible) did not differ between the 2 groups (adjusted RR: 1.03, 95% CI: 0.79 to 1.34) (Table 4).
Within the DES group, definite, probable, or possible stent thrombosis was found in 38 patients treated with SES (2-year incidence 1.73%) and in 43 patients treated with PES (2-year incidence 3.31%). After controlling for covariates, the risk of stent thrombosis was not increased in the SES group compared with the BMS group, but there was an increased risk in the PES group compared with the BMS group (adjusted RR: 1.44, 95% CI: 1.03 to 2.01) (Table 4).
All-cause 2-year mortality was lower among DES-treated than BMS-treated patients (6.40% vs. 7.86%, log rank p = 0.0053), (Fig. 1B). After controlling for covariates, this difference disappeared (adjusted RR: 0.97, 95% CI: 0.83 to 1.13). Overall 2-year cardiac mortality was also lower among DES than among BMS patients (2.99% vs. 4.36%, log rank p = 0.0006). After adjustment, risk of cardiac death no longer differed in the 2 groups (adjusted RR: 0.87, 95% CI: 0.69 to 1.08). Noncardiac 2-year mortality was similar in the DES (3.24%) and BMS groups (3.47%), log-rank p = 0.46 (adjusted RR: 1.04, 95% CI: 0.82 to 1.33). During the second half of the follow-up period (12 to 24 months), all-cause mortality was higher among patients treated with DES than among those treated with BMS (adjusted RR: 1.33, 95% CI: 1.02 to 1.72). This increase was primarily driven by an increase in mortality in the PES group (adjusted RR: 1.46, 95% CI: 1.02 to 2.09) (Table 4).
The 2-year incidence of MI was similar in the 3 stent groups (DES 4.5% vs. BMS 3.8%, log-rank p = 0.1022). However, after controlling for covariates, the DES group had an increased risk of MI (adjusted RR: 1.24, 95% CI: 1.02 to 1.51), primarily because of an increase in MI events after 12 months of follow-up in the PES group (Fig. 1C, Table 4).
Target lesion revascularization occurred less frequently among DES-treated than among BMS-treated patients (2-year incidence 5.3% vs. 7.9%, log-rank p < 0.0001). After controlling for age, sex, clinical presentation, diabetes mellitus, stent length, and reference vessel size, the absolute risk reduction among patients in the DES compared with the BMS group was 44% (adjusted RR: 0.56, 95% CI: 0.49 to 0.65). In the SES group the risk reduction was 51% (adjusted RR: 0.49, 95% CI: 0.41 to 0.59), and in the PES group the risk reduction was 32% (adjusted RR: 0.68, 95% CI: 0.56 to 0.83) (Fig. 1D). Diabetes mellitus (adjusted RR: 1.43, 95% CI: 1.22 to 1.69), STEMI (adjusted RR: 1.35, 95% CI: 1.16 to 1.58), and stent length (adjusted RR per additional millimeter of stent length: 1.02, 95% CI: 1.01 to 1.03) increased the risk of TLR, whereas size of the reference vessel was inversely related to TLR risk (adjusted RR per decreasing millimeter of vessel diameter: 0.73, 95% CI: 0.65 to 0.81).
In this large population-based 2-year follow-up study, patients treated with either DES or BMS had similar rates of death and stent thrombosis, whereas the DES group had an increased risk of MI. During the last year of follow-up, rates of stent thrombosis, MI, and death were significantly increased in DES patients because of increased event rates in patients treated with PES. Further, clinically driven TLR was 39% lower in the DES compared with the BMS group.
The lack of adequately powered randomized clinical end point trials comparing BMS with DES prompted interventional cardiologists to assess the safety and efficacy of DES using registries and meta-analyses (18,19). Data from the Swedish Coronary Angiography and Angioplasty Registry (SCAAR) on PCI patients from 2003 and 2004, published in the New England Journal of Medicinein 2007, indicated that DES was associated with an increased risk of death and MI compared with BMS (8). More recent data from SCAAR, including patients registered in 2005 (reported at the 2007 meetings of the European Society of Cardiology and the Annual Scientific Symposium of Transcatheter Cardiovascular Therapeutics), showed no difference in overall mortality or risk of MI associated with the 2 types of stents, i.e., a pattern similar to our data. Coronary interventional revascularization therapy is organized differently in Sweden and Denmark, centralized in Denmark and decentralized in Sweden. Whereas use of DES was uniformly high in Denmark, use of DES ranged from 0.5% to 60% among Swedish centers (8). Also, PES was more widely used in Sweden than in our population. Importantly, the Swedish recommendations for dual antiplatelet therapy varied from 3 to 6 months for DES and 1 month for BMS, versus the uniform Danish recommendation of 12 months (8). Data from both registries showed a change in the slopes of stent thrombosis and MI curves over time in DES-treated patients. This occurred at 6 months in the SCAAR and at 12 months in the WDHR. This may be explained by the difference in duration of dual antiplatelet therapy.
Stettler et al. (20) published a network meta-analysis including 36 randomized trials on PES and SES, and found better safety and efficacy for SES compared with PES. In contrast to the present study, their analysis mainly included selected patients with relatively simple coronary artery lesions and patients with stable angina pectoris. Nevertheless, our population-based results, including both simple and complex coronary artery lesions and patients with stable and unstable coronary syndromes, were in line with this large meta-analysis.
The recently published Danish multicenter trial, the SORT OUT II (Danish Organization on Randomized Trials With Clinical Outcome) study, randomized 2,098 patients in a head-to-head comparison of SES and PES (21). Although the rates of death, MI, new revascularization, and ARC-defined stent thrombosis were higher in PES-treated patients than in SES-treated patients in the SORT OUT II study, the differences did not reach statistical significance.
The SES and the PES were the first widely used and until recently the only DES approved by the U.S. Food and Drug Administration. Apart from the drug released from the stents, there are marked differences in release kinetics and polymer coating. In the PES, the active drug, paclitaxel, is homogeneously distributed in the polymer. In the SES, sirolimus is placed in a base coating and released through a permeable top coating (22). In studies with angiographic follow-up, the late loss of the PES was about 0.3 mm, versus 0.1 mm for the SES (23). This difference in neointimal formation was found to be translated into an increased rate of new revascularization in studies using angiographic follow-up (24), meta-analysis (20), or as in the present study, registries.
It might be anticipated that some degree of late loss and optimal neointimal coverage of stent struts would be protective against late stent thrombosis, and consequently, that PES treatment would be associated with a lower risk of stent thrombosis compared with SES treatment. However, such a relationship has never been shown. In contrast, the meta-analysis from Stettler et al. (20) and the present study showed that PES, and not SES, was associated with an increased risk of late stent thrombosis. Therefore, the drug release kinetics, the drug itself, the polymer, or other characteristics of the PES are likely to predispose to stent thrombosis.
The SES and PES have been thought to behave identically from a clinical perspective. Until recently, DES has been considered as one stent type in registry studies (8,12), and the effect of DES was considered a class effect. The present study documents the importance of distinguishing between different types of DES. This may complicate stent treatment of patients, but it also raises the possibility that this first generation of DES may be optimized by modifying stent platform, drug, or polymer, thereby improving short-term and long-term results of future DES treatment.
The recommended duration of dual antiplatelet treatment was 12 months, but data on the actual duration of this treatment were available only for patients with definite stent thrombosis. Another concern is that we were unable to disaggregate MI events by area of infarction and thus could not include MI in the ARC definition of probable stent thrombosis.
According to the ARC definition of probable and possible stent thromboses, we categorized any unexplained death as stent thrombosis. However, it is unlikely that stent thrombosis caused all unexplained deaths.
In the analyses of stent thrombosis and mortality, we have reported the relative risk estimates at various left-truncated time periods, which might bias the point estimates. Therefore, the cumulative statistics at various right-censored time points are reported in Table 5.
During a 2-year follow-up period, patients treated with SES had similar risks of stent thrombosis and MI compared with patients treated with BMS, and those treated with PES had an increased risk of these outcomes. Death after 12 months was also higher in PES-treated patients. Clinically driven TLR was lower in both DES groups compared with the BMS group, with lowest rates in SES-treated patients. Our data indicate an overall better safety and efficacy profile of the SES than the PES, and show that the effect of DES treatment is not a class effect. This raises the possibility of improving future generations of DES.
Dr. Kaltoft has received lecture fees from Cordis and Eli Lilly. Dr. Jensen has received lecture fees from Abbott, AstraZeneca, Cordis, and Merck. Dr. Maeng has received lecture fees from Cordis and consultant fees from Novo Nordisk, Denmark, and Medtronic. Dr. Lassen has received lecture fees from Bristol-Myers Squibb, Nycomed, Boston Scientific, GlaxoSmithKline, and Cordis. Dr. Thuesen has received lecture fees from Medtronic, Abbott, Cordis, and Boston Scientific; received consulting fees from Abbott; is on the advisory board of Boston Scientific; and has received unrestricted research grant support from Medtronic, Abbott, Cordis, and Boston Scientific.
- Abbreviations and Acronyms
- Academic Research Consortium
- bare-metal stent(s)
- confidence interval
- drug-eluting stent(s)
- hazard ratio
- myocardial infarction
- percutaneous coronary intervention
- paclitaxel-eluting stent(s)
- relative risk
- sirolimus-eluting stent(s)
- ST-segment elevation myocardial infarction
- target lesion revascularization
- Western Denmark Heart Registry
- Received July 1, 2008.
- Revision received September 22, 2008.
- Accepted September 22, 2008.
- American College of Cardiology Foundation
- Moreno R.,
- Fernandez C.,
- Hernandez R.,
- et al.
- Pfisterer M.,
- Brunner-La Rocca H.P.,
- Buser P.T.,
- et al.
- Nordmann A.J.,
- Briel M.,
- Bucher H.C.
- Jensen L.O.,
- Maeng M.,
- Kaltoft A.,
- et al.
- Update to FDA Statement on Coronary Drug Eluting Stents. http:/www.fda.gov/cdrh/news/010407.html. Accessed December 2006.
- Tunstall-Pedoe H.,
- Kuulasmaa K.,
- Amouyel P.,
- Arveiler D.,
- Rajakangas A.M.,
- Pajak A.
- Marzocchi A.,
- Saia F.,
- Piovaccari G.,
- et al.
- Wessely R.,
- Schomig A.,
- Kastrati A.