Five-Year Follow-Up After Sirolimus-Eluting Stent Implantation

Journal of the American College of Cardiology Vol. 53, No. 17, 2009 2009 by the American College of Cardiology Foundation ISSN 0735-1097/09/$36.00 Published by Elsevier Inc. doi:10.1016/j.jacc.2009.01.050 CLINICAL RESEARCH Interventional Cardiology Five-Year Follow-Up After Sirolimus-Eluting Stent Implantation Results of the SIRIUS (Sirolimus-Eluting Stent in De-Novo Native Coronary Lesions) Trial Giora Weisz, MD,* Martin B. Leon, MD,* David R. Holmes, JR, MD, Dean J. Kereiakes, MD, Jeffrey J. Popma, MD, Paul S. Teirstein, MD, Sidney A. Cohen, MD, PHD, # Hong Wang, MD, MPH,# Donald E. Cutlip, MD,** Jeffrey W. Moses, MD* New York, New York; Rochester, Minnesota; Cincinnati, Ohio; Boston, Massachusetts; La Jolla, California; Philadelphia, Pennsylvania; and Warren, New Jersey Objectives Background Methods Results Conclusions The aim of this study was to examine the 5-year clinical safety and efficacy outcomes in patients enrolled in the SIRIUS (Sirolimus-Eluting Stent in De-Novo Native Coronary Lesions) trial. The SIRIUS trial was a double-blinded randomized study that demonstrated that sirolimus-eluting stents () significantly improved angiographic results (at 8 months) and clinical outcomes (at 9 and 12 months) compared with bare-metal stents (). Patients (n 1,058) with de novo native coronary artery lesions were randomized to either (n 533) or control (n 525) and were followed for 5 years. Between 1 and 5 years, additional clinical events were similarly distributed among the sirolimus and control groups. At 5 years, in sirolimus versus control patients, target lesion revascularization was 9.4% versus 24.2% (p 0.001) and major adverse cardiovascular events and target vessel failure rates were 20.3% versus 33.5% and 22.5% versus 33.5%, respectively (p 0.0001 for both). There were no significant differences in death, myocardial infarction, and nontarget lesion revascularization. No significant differences were observed in the cumulative incidence of stent thrombosis for sirolimus versus control patients with either protocol-derived (1.0% vs. 0.8%) or Academic Research Consortium definitions (3.9% vs. 4.2%). In patients with noncomplex coronary artery disease, clinical outcomes 5 years after implantation of continue to demonstrate significant reduction in the need for repeat revascularization, with similar safety (death and myocardial infarction) compared with, without evidence for either disproportionate late restenosis or late stent thrombosis. (J Am Coll Cardiol 2009;53:1488 97) 2009 by the American College of Cardiology Foundation Sirolimus-eluting stents () have shown significant improvement in angiographic outcomes, with reduced rates of restenosis and need for revascularization compared with baremetal stent () controls (1 5). These favorable clinical outcomes have resulted in the frequent use of drug-eluting stents during percutaneous coronary revascularization procedures over the past several years in the U.S. and around the world. However, reports of drug-eluting stent very late stent thrombosis, associated with possible increased mortality, have elicited long-term safety concerns (6 11). The SIRIUS (Sirolimus-Eluting Stent in De-Novo Native Coronary Lesions) multicenter, blinded, randomized From the *Center for Interventional Vascular Therapy, New York-Presbyterian Hospital, Columbia University, and Cardiovascular Research Foundation, New York, New York; Saint Mary s Hospital, Rochester, Minnesota; Ohio Heart Health Center, Cincinnati, Ohio; St. Elizabeth s Medical Center, Boston, Massachusetts; Scripps Clinic and Research Institute, La Jolla, California; University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; #Cordis (Johnson & Johnson), Warren, New Jersey; and the **Beth Israel Deaconess Medical Center and Harvard Clinical Research Institute, Boston, Massachusetts. Dr. Leon has served on the advisory board and received consulting fees from Cordis, Johnson & Johnson. Dr. Kereiakes has served on the scientific advisory boards of Boston Scientific Corp., Reva Medical, and Abbott Vascular, and served as consultant for Cordis, Reva Medical, Medtronic, Eli Lilly, and Daichi-Sanyo. Dr. Popma has received a research grant for study and served on the advisory board for Cordis, Johnson & Johnson. Dr. Teirstein has received research grants and consulting fees from Cordis, Johnson & Johnson, and served as consultant to and received honorarium from Boston Scientific, Medtronic, and Abbott. Drs. Cohen and Wang are employees of Cordis, Johnson & Johnson. Dr. Moses has received consulting fees from Cordis, Johnson & Johnson. Manuscript received January 6, 2008; revised manuscript received January 7, 2009, accepted January 12, 2009.

JACC Vol. 53, No. 17, 2009 April 28, 2009:1488 97 Weisz et al. 5-Year Follow-Up of the SIRIUS Trial 1489 trial examined the safety and efficacy of for up to 2 years and indicated significant reduction in both restenosis and target lesion revascularization (TLR) (3 5). This report extends the clinical follow-up of the original SIRIUS patient cohort to determine whether the early and midterm safety and efficacy of is maintained at 5 years. Methods Study design and eligibility criteria. The methods of the SIRIUS trial were previously reported (3). Patients enrolled in the study had a clinical history of angina and single coronary target lesions, 15 to 30 mm in length, in vessels 2.5 to 3.5 mm in diameter. Data collection and follow-up. Patients had clinical evaluations at 30 days and 6, 9, 12, 24, 36, 48, and 60 months. The treatment identity (sirolimus or ) remained blinded throughout the follow-up period to both patients and investigators. Complete data compliance for all 5-year follow-up end points was 93.6% for the group and 94.0% for the group. Study end points: repeat revascularization. This 5-year follow-up study focuses on late safety and effectiveness, namely clinical restenosis or TLR. This was defined as the need for clinically driven repeat revascularization of the target lesion. In the absence of objective criteria for ischemia, an in-lesion diameter stenosis by quantitative coronary angiography 70% was also considered of sufficient severity to justify repeat revascularization. The independent Clinical Events Committee blindly adjudicated all revascularization episodes. Other secondary end points. SIRIUS secondary clinical end points included non-tlr target vessel revascularization (non-tlr TVR), target vessel failure (TVF), and major adverse cardiac events (MACE). Non-TLR TVR was defined as any clinically driven repeat percutaneous intervention of the target vessel or bypass surgery of the target vessel for a lesion other than the target lesion. A non Q-wave myocardial infarction (MI) was defined as an increase in the creatine kinase level to more than twice the upper limit of the normal range accompanied by an increased level of creatine kinasemyocardial band in the absence of new Q waves on the electrocardiogram. Cardiac death was defined as death due to obvious cardiovascular events and procedure-related causes or any death in which a cardiac cause could not be excluded. Noncardiac death was defined as any death clearly not due to cardiac causes (e.g., malignancy). A TVF was defined as any TVR (either target vessel coronary bypass surgery or target vessel percutaneous coronary intervention), MI (Q-wave or non Q-wave), or cardiac death that could not be clearly attributed to a vessel other than the target vessel. Major adverse cardiac event was defined as a composite of death (cardiac and noncardiac), MI (Q-wave or non Q-wave), and any TLR (either target vessel coronary bypass surgery or target lesion percutaneous coronary intervention). Stent thrombosis. Early stent thrombosis, either acute (within 24 h) or subacute (between 24 h and 30 days), was Abbreviations and Acronyms ARC Academic Research Consortium bare-metal stent(s) MACE major adverse cardiac event(s) MI myocardial infarction OR odds ratio sirolimus-eluting stent(s) TLR target lesion revascularization TVF target vessel failure TVR target vessel revascularization defined according to protocol as angiographic documentation of target vessel occlusion or any death or MI occurring within 30 days that was not clearly related to causes other than stent occlusion. The protocol definition of late stent thrombosis was MI occurring 30 days after the index procedure and attributable to the target vessel and angiographic documentation (site-reported or by quantitative coronary angiography) of thrombus or total occlusion of the target site and freedom from an interim revascularization of the target vessel. For the purposes of this report, we used both the protocol-derived and the recently suggested Academic Research Consortium (ARC) (12) definitions of stent thrombosis. The ARC definitions consider distinct reportable time points: acute stent thrombosis (0 to 24 h after stent implantation); subacute stent thrombosis ( 24 h to 30 days); late stent thrombosis (30 days to 1 year); and very-late stent thrombosis (more than 1 year after stent implantation). The ARC definitions also recognize 3 categories of evidence in defining stent thrombosis: confirmed/definite, probable, and possible. Confirmed/ definite stent thrombosis was defined as an acute coronary syndrome with symptoms, electrocardiogram changes, or more than 2-fold elevation in creatine phosphokinase combined with either angiographic or pathologic confirmation of stent thrombosis. Probable stent thrombosis was defined as any unexplained death within the first 30 days or irrespective of the time after the index procedure, or any MI in the absence of an obvious cause that is related to documented acute ischemia in the territory (target vessel) of the implanted stent without angiographic confirmation of stent thrombosis. Possible stent thrombosis was defined as any unexplained death more than 30 days after the index procedure. The ARC definitions did not censor stent thrombosis events that occurred after an intervening revascularization of the target vessel. Statistical analysis. Analyses were performed on a modified intent-to-treat population; deregistered patients were not included in the analysis, because they received neither study treatment (3). Continuous variables are summarized as mean SD and were compared between treatment groups with the t test. Categorical variables are summarized as frequencies and percentages and were compared between treatment groups with the Fisher exact test. Differences and 95% confidence intervals (CIs) between the 2 comparison groups were also calculated. Event-free incidence or cumulative incidences of selected long-term outcomes were summarized with the Kaplan-Meier method and compared between treatment groups with log-

1490 Weisz et al. JACC Vol. 53, No. 17, 2009 5-Year Follow-Up of the SIRIUS Trial April 28, 2009:1488 97 rank tests. To identify the risk factors for the major long-term outcomes, a series of univariate and multivariate predictor analyses were performed; multivariate predictors were chosen with an entry criterion of 0.20 and with a stay criterion of 0.10. All tests are 2-sided with a significant level of 0.05. All Cumulative Clinical Events at 1- and 5-Year Clinical Follow-Up Table 1 Cumulative Clinical Events at 1- and 5-Year Clinical Follow-Up statistical analyses were performed with SAS software (version 8.2, SAS, Cary, North Carolina). The data were managed and analyzed exclusively by the Harvard Clinical Research Institute during the 5-year follow-up. The authors had full access to the data and take responsibility for its integrity. (n 533) (n 525) Difference (95% CI) p Value All events at 1 yr Death 7 (1.3) 4 (0.8) 0.6% ( 0.7% to 1.8%) 0.55 Cardiac 3 (0.6) 2 (0.4) 0.2% ( 0.6% to 1.0%) 1.00 Noncardiac 4 (0.8) 2 (0.4) 0.4% ( 0.5% to 1.3%) 0.69 MI (all) 16 (3.0) 22 (4.2) 1.2% ( 3.4% to 1.1%) 0.33 Q-wave 4 (0.8) 4 (0.8) 0.0% ( 1.1% to 1.0%) 1.00 Non Q-wave 12 (2.3) 18 (3.4) 1.2% ( 3.2% to 0.8%) 0.27 Death or any MI 23 (4.3) 25 (4.8) 0.4% ( 3.0% to 2.1%) 0.77 Death or Q-wave MI 11 (2.1) 8 (1.5) 0.5% ( 1.1% to 2.1%) 0.64 TLR 26 (4.9) 106 (20.2) 15.3% ( 19.2% to 11.4%) 0.001 TL CABG 5 (0.9) 9 (1.7) 0.8% ( 2.2% to 0.6%) 0.30 TL PCI 23 (4.3) 102 (19.2) 15.1% ( 18.9% to 11.3%) 0.001 All TVR 38 (7.1) 120 (22.9) 15.7% ( 19.9% to 11.5%) 0.001 TVR (non-tl) 20 (3.8) 34 (6.5) 2.7% ( 5.4% to 0.1%) 0.05 MACE 44 (8.3) 122 (23.2) 15.0% ( 19.3% to 10.7%) 0.001 TVF 52 (9.8) 130 (24.8) 15.0% ( 19.5% to 10.5%) 0.001 All events at 5 yrs Death 45 (8.4) 44 (8.4) 0.1% ( 3.3% to 3.4%) 1.00 Cardiac 22 (4.1) 19 (3.6) 0.5% ( 1.8% to 2.8%) 0.75 Noncardiac 23 (4.3) 25 (4.8) 0.4% ( 3.0% to 2.1%) 0.77 MI (all) 33 (6.2) 34 (6.5) 0.3% ( 3.2% to 2.7%) 0.90 Q-wave 8 (1.5) 6 (1.1) 0.4% ( 1.0% to 1.7%) 0.79 Non Q-wave 26 (4.9) 28 (5.3) 0.5% ( 3.1% to 2.2%) 0.78 Death or any MI 74 (13.9) 70 (13.3) 0.6% ( 3.6% to 4.7%) 0.86 Death or Q-wave MI 51 (9.6) 49 (9.3) 0.2% ( 3.3% to 3.8%) 0.92 TLR 50 (9.4) 127 (24.2) 14.8% ( 19.2% to 10.4%) 0.001 TL CABG 12 (2.3) 18 (3.4) 1.2% ( 3.2% to 0.8%) 0.27 TL PCI 43 (8.1) 121 (23.0) 15.0% ( 19.3% to 10.7%) 0.001 All TVR 88 (16.5) 160 (30.5) 14.0% ( 19.0% to 8.9%) 0.001 TVR (non-tl) 55 (10.3) 68 (13.0) 2.6% ( 6.5% to 1.2%) 0.21 MACE 108 (20.3) 176 (33.5) 13.3% ( 18.5% to 8.0%) 0.001 TVF 120 (22.5) 182 (34.7) 12.2% ( 17.6% to 6.8%) 0.001 All events between 1 and 5 yrs Death 38 (7.1) 40 (7.6) 0.5% ( 3.6% to 2.7%) 0.81 Cardiac 19 (3.6) 17 (3.2) 0.3% ( 1.9% to 2.5%) 0.87 Noncardiac 19 (3.6) 23 (4.4) 0.8% ( 3.2% to 1.5%) 0.53 MI (all) 17 (3.2) 12 (2.3) 0.9% ( 1.1% to 2.9%) 0.45 Q-wave 4 (0.8) 2 (0.4) 0.4% ( 0.5% to 1.3%) 0.69 Non Q-wave 14 (2.6) 10 (1.9) 0.7% ( 1.1% to 2.5%) 0.54 Death or any MI 51 (9.6) 45 (8.6) 1.0% ( 2.5% to 4.5%) 0.59 Death or Q-wave MI 40 (7.5) 41 (7.8) 0.3% ( 3.5% to 2.9%) 0.91 TLR 24 (4.5) 21 (4.0) 0.5% ( 1.9% to 2.9%) 0.76 TL CABG 7 (1.3) 9 (1.7) 0.4% ( 1.9% to 1.1%) 0.62 TL PCI 20 (3.8) 19 (3.6) 0.1% ( 2.1% to 2.4%) 1.00 All TVR 50 (9.4) 40 (7.6) 1.8% ( 1.6% to 5.1%) 0.32 TVR (non-tl) 35 (6.6) 34 (6.5) 0.1% ( 2.9% to 3.1%) 1.00 MACE 64 (12.0) 54 (10.3) 1.7% ( 2.1% to 5.5%) 0.38 TVF 68 (12.8) 52 (9.9) 2.9% ( 1.0% to 6.7%) 0.15 Values are number of cases (%). Events at 1 year were reported previously (4). bare-metal stent(s); CABG coronary artery bypass graft surgery; CI confidence interval; MACE major adverse cardiac event(s); MI myocardial infarction; PCI percutaneous coronary intervention; sirolimus-eluting stent(s); TL target lesion; TLR target lesion revascularization; TVF target vessel failure; TVR target vessel revascularization.

JACC Vol. 53, No. 17, 2009 April 28, 2009:1488 97 Weisz et al. 5-Year Follow-Up of the SIRIUS Trial 1491 Results As previously reported, both groups had similar baseline clinical and angiographic characteristics, procedural factors, and acute (in-hospital) results (3). Revascularization and clinical events. After 5 years, the significant differences between the and groups in TLR, TVR, TVF, and MACE were all maintained (Table 1, Fig. 1). During the follow-up from 1 to 5 years, in both groups there were similar increases in the rates of all clinical end points. Figure 1 Kaplan-Meier Event-Free Survival Curves Kaplan-Meier event-free survival curves at 2 years for sirolimus-eluting stents () and control stents (CS) for (A) death any myocardial infarction (MI) and death Q-wave myocardial infarction (QMI); (B) target lesion revascularization (TLR) and target vessel revascularization (TVR); and (C) major adverse cardiac events (MACE) and target vessel failure (TVF). The number of patients at risk each year is provided, equal to the number of patients for whom follow-up data were available at each time minus those with earlier events. LR log-rank test.

1492 Weisz et al. JACC Vol. 53, No. 17, 2009 5-Year Follow-Up of the SIRIUS Trial April 28, 2009:1488 97 The comparison of survival curves in freedom from death or MI showed no significant differences over the follow-up period (Fig. 1A). The event-free curves of TLR, TVR, MACE, and TVF stayed separated and parallel throughout the extended follow-up after 1 year (Figs. 1B and 1C). Various higher restenosis risk patient and lesion subgroups (including diabetes, left anterior descending coronary artery lesion location, small vessel size, and long lesion length) were examined (Fig. 2). In all cases, the significant reduction in TLR associated with compared with controls in these selected subgroups was maintained at 5 years. Multivariate analyses identified the following predictors of TLR: treatment with control stent (vs. ; odds ratio [OR]: 3.44, p 0.0001), diabetes mellitus (OR: 1.86, p 0.001), prior coronary artery bypass graft surgery (OR: 1.78, p 0.04), total numbers of stent implanted (OR: 1.70/mm stent-length increase, p 0.0001), left anterior descending coronary artery (vs. other locations; OR: 1.52, p 0.02), post-procedure in-stent minimal luminal diameter (/mm, OR: 0.52, p 0.003), and age (/year, OR: 0.98, p 0.04). Multivariate analyses identified the following independent predictors of death and MI: age (OR: 1.05, p 0.0001), stent length (OR: 1.04, p 0.002), smoking during the last year before procedure (OR: 1.94, p 0.005), congestive heart failure (OR: 2.06, p 0.02), reference vessel diameter (OR: 0.62, p 0.03), hypertension (OR: 1.55, p 0.047), prior MI (OR: 1.47, p 0.054), and post-procedure hospital length of stay (OR: 1.32, p 0.004). Stent thrombosis. Stent thrombosis episodes (according to either the protocol or ARC definitions) are reported as early events up to 30 days after the index procedure, late events from 30 days to 1 year, and very late events between 1 and 5 years for both the sirolimus-eluting and the control stent patients (Table 2). Early stent thrombosis was rare, irrespective of the definition. In the late time period between 30 days and 1 year after stent placement, with both types of definitions there were low rates of stent thrombosis with control stents and (0.6% vs. 0.2% with the protocol definition of stent thrombosis; 0.7% vs. 0.2% with the definite plus probable ARC definition). During the last 4 years of follow-up, the very late period (years 1 to 5), there were 3 events with the protocol definition of stent thrombosis but none with control stents (p 0.16). With the ARC definitions of very late definite stent thrombosis, there were 4 cases of definite stent thrombosis with and 2 with (0.8% and 0.4%, p 0.57). There was only 1 case of probable very late stent thrombosis in the group and none in the group and no significant difference in the rates of possible very late stent thrombosis between and (2.6% vs. 2.3%, p 0.74). After 5 years of follow-up (Table 2, Fig. 3), the overall cumulative incidence of protocol-defined stent thrombosis was not significantly different between and patients (1.0% vs. 0.8%; p 0.75). With ARC definitions compared with protocol definitions of stent thrombosis, there was an increase in the number of stent thrombosis events (to 41 from 9) due to a lack of censoring stent Figure 2 Odds Ratios for Target Lesion Revascularization at 5 Years for the Overall Population and by Subgroup Values represent event rates (%) in each arm by subgroup. Bars represent odds ratio point estimates and 95% confidence intervals. CI confidence interval; CS control stent; LAD treatment of lesions in the left anterior descending artery; sirolimus-eluting stent.

JACC Vol. 53, No. 17, 2009 April 28, 2009:1488 97 Weisz et al. 5-Year Follow-Up of the SIRIUS Trial 1493 Stent Thrombosis According to the Protocol and ARC Definitions (Up to 5 Years Follow-Up) Table 2 Stent Thrombosis According to the Protocol and ARC Definitions (Up to 5 Years Follow-Up) Early (0 30 Days) Late (31 Days 1 yr) Very Late (1 5 yrs) Overall Overall Difference (95% CI), p Value Very Late (1 5 yrs) Difference (95% CI), p Value* (n 525) (n 533) (n 525) (n 533) (n 525) (n 533) (n 525) (n 533) Protocol 1 (0.2) 1 (0.2) 1 (0.2) 3 (0.6) 3 (0.6) 0 5 (1.0) 4 (0.8) 0.6% ( 0.1% to 1.3%), 0.1556 0.2% ( 1.1% to 1.5%), 0.7478 ARC definitions Definite 1 (0.2) 0 1 (0.2) 3 (0.6) 4 (0.8) 2 (0.4) 6 (1.2) 5 (1.0) 0.4% ( 0.7% to 1.5%), 0.5672 0.2% ( 1.2% to 1.6%), 0.7745 Probable 0 2 (0.4) 0 1 (0.2) 0 1 (0.2) 0 4 (0.8) 0.2% ( 0.6% to 0.2%), 0.4871 0.8% ( 1.6% to 0.1%), 0.0446 Possible 0 0 1 (0.2) 1 (0.2) 13 (2.6) 11 (2.3) 14 (2.8) 12 (2.5) 0.3% ( 1.8% to 2.4%), 0.7378 0.3% ( 1.8% to 2.5%), 0.7040 Definite or probable 1 (0.2) 2 (0.4) 1 (0.2) 4 (0.7) 4 (0.8) 3 (0.7) 6 (1.2) 9 (1.8) 0.2% ( 0.9% to 1.3%), 0.8355 0.6% ( 2.2% to 1.1%), 0.4257 Any ARC events 1 (0.2) 2 (0.4) 2 (0.4) 5 (0.9) 17 (3.3) 14 (2.9) 20 (3.9) 21 (4.2) 0.5% ( 1.8% to 2.8%), 0.6410 0.3% ( 2.9% to 2.4%), 0.8520 See the Methods section for the various categories of stent thrombosis. Values are number of cases (%). For all calculations, event rates are cumulative incidences with Kaplan-Meier method. *Cumulative incidences with Kaplan-Meier method were used starting from 1 year, because no patient would have very late stent thrombosis before 1 year. ARC Academic Research Consortium; other abbreviations as in Table 1. Figure 3 Kaplan-Meier Curves Representing the 5-Year Cumulative Incidence Rates of Stent Thrombosis Kaplan-Meier curves representing the 5-year cumulative incidence rates of stent thrombosis (ST) (A) protocol-defined (B) with the Academic Research Consortium (ARC) definition. The number of patients at risk each year is provided, equal to the number of patients for whom follow-up data were available at each time minus those with earlier events. LR log-rank test; other abbreviations as in Figure 2. thromboses after repeat revascularization events and the more expansive categories of probable and especially possible late stent thrombosis. The definite category of the ARC-defined stent thrombosis events added only 1 case to each stent group compared with the protocol-based definition, but the probable category included 4 additional cases (all ) and the possible category included 26 additional cases (14 and 12 ). There were no significant differences between and in the overall

1494 Weisz et al. JACC Vol. 53, No. 17, 2009 5-Year Follow-Up of the SIRIUS Trial April 28, 2009:1488 97 frequency of stent thrombosis with ARC definitions (cumulative incidence 3.9% vs. 4.2%; p 0.85). Diabetes mellitus. A total of 279 patients (26.4%) had diabetes mellitus, and the baseline demographic data, lesion characteristics, and short-term outcomes of the diabetic patients have been reported previously (13). The long-term clinical outcomes of the patients with diabetes mellitus are summarized in Table 3. Diabetic patients generally had higher frequencies of clinical events compared with patients without diabetes in each treatment group (Table 3). In patients, diabetic patients had more frequent cardiac death than nondiabetic patients (9.9% vs. 4.2%, p 0.0004) and in patients, diabetic patients had more frequent TLR events (33.1% vs. 20.7%, p 0.0045). When comparing versus at 5 years in both diabetic and nondiabetic patients, there were significant reductions in the frequency of TLR, TVR, MACE, and TVF, but no significant differences were observed in the rates of death and MI. A nonsignificant increased rate of cardiac death with versus in diabetic patients was observed (9.9% vs. 4.7%; p 0.11). This was associated with a cluster of death events in the group in the first one-half of the third year after the index procedure (Fig. 4). In the diabetic subgroup, there were no significant differences in overall protocol-defined stent thrombosis after treatment (: 2 events [1.6%] vs. : 1 event [0.7%]; p 0.48), and there were also no significant differences in ARC definite plus probable stent thrombosis (: 3 events [2.6%] vs. : 4 events [2.9%]; p 0.85). Due to the increased cardiac death events during follow-up in the diabetic patients, there was a trend toward more frequent ARC possible stent thrombosis (: 9 events [7.5%] vs. : 4 events [3.0%]; p 0.09). The overall cumulative incidence of ARC-defined stent thrombosis was 9.9% (12 events) in patients and 5.9% (8 events) in patients (p 0.22). The results related to the diabetic patients in the SIRIUS trial should be interpreted with caution. The number of diabetic patients in this study was small, and further studies focused on this group of patients are needed. Discussion The SIRIUS trial is the largest randomized study (n 1,058) of patients with symptomatic coronary disease comparing treatment of the culprit lesion with with the predicate (3). Eight-month angiographic follow-up in 703 patients and 9-month clinical follow-up in all patients indicated a marked reduction in late lumen loss (in-stent: 1.00 mm vs. 0.17 mm, p 0.001), binary in-lesion restenosis (36.3% vs. 8.9%, p 0.001), and clinically driven TLR (16.6% vs. 4.1%, p 0.001) when comparing control with the. When clinical follow-up was extended to 12 and 24 months, the improvement in clinical restenosis (clinically driven TLR) was maintained after treatment (4,5). On the basis of these angiographic and clinical Cumulative Clinical Events at 5-Year Clinical Follow-Up, Versus, in Patients With and Without DM Table 3 Cumulative Clinical Events at 5-Year Clinical Follow-Up, Versus, in Patients With and Without DM DM Non-DM Difference (95% CI), p Value Non-DM vs. DM vs. CS DM vs. Non-DM DM vs. Non-DM (n 376) (n 402) (n 148) (n 131) Death 20 (15.3) 14 (9.5) 25 (6.2) 30 (8.0) 9.0% ( 15.6% to 2.5%), 0.01 1.5% ( 6.9% to 4.0%), 0.60 5.8% ( 1.9% to 13.6%), 0.15 1.8% ( 5.4% to 1.9%), 0.40 Cardiac 13 (9.9) 7 (4.7) 9 (2.2) 12 (3.2) 7.7% ( 13.0% to 2.4%), 0.001 1.5% ( 5.4% to 2.3%), 0.44 5.2% ( 1.0% to 11.4%), 0.11 1.0% ( 3.2% to 1.3%), 0.51 Noncardiac 7 (5.3) 7 (4.7) 16 (4.0) 18 (4.8) 1.4% ( 5.7% to 2.9%), 0.47 0.1% ( 4.0% to 4.1%), 1.00 0.6% ( 4.5% to 5.8%), 1.00 0.8% ( 3.7% to 2.1%), 0.60 MI (all) 9 (6.9) 12 (8.1) 24 (6.0) 22 (5.9) 0.9% ( 5.8% to 4.0%), 0.68 2.3% ( 7.3% to 2.7%), 0.33 1.2% ( 7.4% to 4.9%), 0.82 0.1% ( 3.2% to 3.4%), 1.00 Q-wave MI 3 (2.3) 0 (0.0) 5 (1.2) 6 (1.6) 1.0% ( 3.8% to 1.7%), 0.41 1.6% (0.3% to 2.9%), 0.19 2.3% ( 0.3% to 4.9%), 0.10 0.4% ( 2.0% to 1.3%), 0.77 Non Q-wave MI 7 (5.3) 12 (8.1) 19 (4.7) 16 (4.3) 0.6% ( 5.0% to 3.8%), 0.82 3.9% ( 8.7% to 1.0%), 0.09 2.8% ( 8.6% to 3.1%), 0.48 0.5% ( 2.4% to 3.4%), 0.86 Death or any MI 27 (20.6) 24 (16.2) 47 (11.7) 46 (12.2) 8.9% ( 16.5% to 1.3%), 0.01 4.0% ( 10.8% to 2.8%), 0.25 4.4% ( 4.7% to 13.5%), 0.36 0.5% ( 5.1% to 4.0%), 0.83 Death or Q-wave MI 22 (16.8) 14 (9.5) 29 (7.2) 35 (9.3) 9.6% ( 16.5% to 2.7%), 0.01 0.2% ( 5.7% to 5.4%), 1.00 7.3% ( 0.6% to 15.3%), 0.08 2.1% ( 6.0% to 1.8%), 0.30 TLR (clinically driven) 18 (13.7) 49 (33.1) 32 (8.0) 78 (20.7) 5.8% ( 12.2% to 0.7%), 0.06 12.4% ( 21.0% to 3.7%), 0.01 19.4% ( 29.0% to 9.8%), 0.001 12.8% ( 17.7% to 7.9%), 0.001 TL CABG 2 (1.5) 7 (4.7) 10 (2.5) 11 (2.9) 1.0% ( 1.6% to 3.6%), 0.74 1.8% ( 5.6% to 2.0%), 0.30 3.2% ( 7.2% to 0.8%), 0.18 0.4% ( 2.7% to 1.8%), 0.83 TL PCI 17 (13.0) 47 (31.8) 26 (6.5) 74 (19.7) 6.5% ( 12.7% to 0.3%), 0.03 12.1% ( 20.6% to 3.6%), 0.01 18.8% ( 28.2% to 9.3%), 0.001 13.2% ( 17.9% to 8.5%), 0.001 All TVR 29 (22.1) 58 (39.2) 59 (14.7) 102 (27.1) 7.5% ( 15.4% to 0.4%), 0.06 12.1% ( 21.1% to 3.0%), 0.01 17.1% ( 27.7% to 6.4%), 0.01 12.5% ( 18.1% to 6.8%), 0.001 TVR (non-tlr) 19 (14.5) 23 (15.5) 36 (9.0) 45 (12.0) 5.5% ( 12.2% to 1.1%), 0.10 3.6% ( 10.3% to 3.1%), 0.31 1.0% ( 9.4% to 7.4%), 0.87 3.0% ( 7.3% to 1.3%), 0.20 MACE 39 (29.8) 64 (43.2) 69 (17.2) 112 (29.8) 12.6% ( 21.3% to 4.0%), 0.01 13.5% ( 22.7% to 4.2%), 0.01 13.5% ( 24.7% to 2.3%), 0.03 12.6% ( 18.5% to 6.7%), 0.001 TVF 42 (32.1) 66 (44.6) 78 (19.4) 116 (30.9) 12.7% ( 21.5% to 3.8%), 0.01 13.7% ( 23.0% to 4.5%), 0.01 12.5% ( 23.8% to 1.2%), 0.04 11.4% ( 17.5% to 5.4%), 0.001 Values are number of cases (%). DM diabetes mellitus; other abbreviations as in Table 1.

JACC Vol. 53, No. 17, 2009 April 28, 2009:1488 97 Weisz et al. 5-Year Follow-Up of the SIRIUS Trial 1495 Figure 4 Kaplan-Meier Curves Representing the 5-Year Cumulative Incidence Rates of Death for Patients With DM Kaplan-Meier curves representing the 5-year cumulative incidence rates of death for patients with diabetes mellitus (DM). The number of patients at risk each year is provided, equal to the number of patients for whom follow-up data were available at each time minus those with earlier events. Abbreviations as in Figures 2 and 3. outcomes from the SIRIUS trial, the Food and Drug Administration approved the use of in April 2003. The 5-year SIRIUS follow-up analysis shows no evidence of disproportionate late clinical events with, including death, MI, stent thrombosis, and repeat revascularizations (Tables 1 and 2, Figs. 1 to 3). The major findings of this 5-year follow-up report involving patients treated with from the SIRIUS trial are: 1) clinical restenosis (with blinded adjudication of all clinically driven TLR events) shows maintained anti-restenosis efficacy; and 2) late complications such as stent thrombosis, MI, and death between 1 and 5 years were uncommon and occurred with similar frequency in sirolimus and groups. On the basis of previous experience with potent antiproliferative strategies to prevent in-stent restenosis, such as brachytherapy (14 16), concerns were expressed regarding the long-term durability of the anti-restenosis efficacy associated with. The 5-year results from the SIRIUS trial demonstrate that the significant differences in rates of revascularization, both TLR and TVR, were maintained without late catch-up phenomenon (Table 1, Fig. 1). With, the need for revascularization between 1 and 5 years was uncommon (TLR 1.1%; TVR 2.3%/year) and similar to control stents. At 5-year follow-up, the long-term clinically driven TLR was still markedly reduced (by 61%) from 24.2% in the control group to 9.4% in -treated patients (p 0.001). Thus, by using an in the SIRIUS trial, at 5 years, 149 repeat revascularization events were prevented for every 1,000 patients treated (Fig. 2). This demonstration of the potency and durability of the anti-restenosis effect of is clearly distinguished from previous strategies to prevent in-stent restenosis. No important safety differences were observed comparing versus control stents, examining the hard clinical end points of death (all and cardiac) and MI (all and Q-wave), alone or in combination. The mean annual frequency (after year 1) was 1% for both cardiac death and MI. These long-term events are comparable to previously published natural history reports in patients with coronary disease treated with percutaneous coronary revascularization modalities (17 19). Stent thrombosis. Stent thrombosis is a dramatic clinical event associated with MI and frequent mortality. Recent publications including long-term follow-up of randomized trials and registries as well as some meta-analyses have indicated that drug-eluting stents (both and paclitaxeleluting stents) might be associated with increased rates of very late stent thrombosis (after 1 year), compared with (7,9,11,20 22). For patients similar to those enrolled in the SIRIUS trial, the incremental very late stent thrombosis event rates have been small, approximately 2 patients/ year/1,000 patients treated. Therefore, rigorous late clinical follow-up of the blinded randomized trials using readjudicated standardized definitions becomes essential to ensure patient safety. The SIRIUS trial s 5-year follow-up confirms a low rate of very late stent thrombosis events associated with and (3 and 0 events, respectively; p 0.1556). However, after re-adjudicating all follow-up events with the ARC definitions, which both broadens the classification of events captured as stent thrombosis and no

1496 Weisz et al. JACC Vol. 53, No. 17, 2009 5-Year Follow-Up of the SIRIUS Trial April 28, 2009:1488 97 longer censors events after an intervening repeat revascularization, the differences between and are substantially diminished ( definite plus probable : 4 vs. 3 events, respectively) (Table 2). The cumulative 5-year frequencies of stent thrombosis applying the ARC definitions also indicated a similar occurrence of stent thrombosis in the and patients ( definite : approximately 0.2%/ year; and combined definite, probable, and possible : approximately 0.8%/year for both and patients). Thus, in the SIRIUS patient population constituting what is now considered on-label use in a carefully conducted, randomized, double-blind study with almost complete long-term follow-up, the use of did not lead to an increase in stent thrombosis. Diabetes mellitus. Diabetes mellitus is a well-known risk factor for increased restenosis and worse prognosis (death and MI) in patients with coronary artery disease treated with percutaneous transcatheter interventions (23 26). During follow-up, the diabetic patients enrolled in SIRIUS had more repeat revascularization events than the nondiabetic patients, in both the and treatment groups (Table 3). The time course of restenosis seemed unaltered, with the vast majority of events occurring in the first year. Importantly, despite an absolute increase in the frequency of events, the relative antirestenosis efficacy of was similar in diabetic versus nondiabetic patients (TLR decrease by was 59% in diabetic patients and 61% in nondiabetic patients, and the actual number of TLR events prevented/1,000 patients treated was actually greater in diabetic patients than in nondiabetic patients. A recent pooled analysis raised concerns that might adversely affect mortality in patients with diabetes (27). This pooled analysis examined the 4-year outcomes from 4 blinded, randomized studies comparing with, including the SIRIUS trial, which comprised 65% of the overall patient cohort. There were significant differences in long-term mortality that might be explained by an unexpectedly low frequency of deaths in the diabetic control stent population. In the SIRIUS 5-year follow-up analysis, there is a statistically nonsignificant higher overall noncardiac mortality in diabetic patients treated with versus (overall mortality: 15.3% vs. 9.5%, difference 5.8%, 95% CI: 1.9% to 13.6%; and cardiac mortality: 9.9% vs. 4.7%, difference 5.2%, 95% CI: 1.0% to 11.4%). The mortality differences were largely the result of a cluster of cardiac death events occurring in the first portion of year 3 (Fig. 4). Once again, there was an unexpectedly low frequency of cardiac deaths in diabetic patients ( 1%/ year), which contributed to the apparent differences. Follow-up from dedicated randomized trials in diabetic patients and large registries is ongoing and should further clarify this hypothesis-generating observation. Diabetes has previously been reported as a predictor for stent thrombosis in drug-eluting stent clinical studies (6 8,21). In the SIRIUS trial, the overall 5-year stent thrombosis frequencies (both protocol-defined and ARC definite and probable ) were not different in diabetic patients compared with nondiabetic patients and uninfluenced by versus treatment. Due to the aforementioned unexplained late increase in cardiac deaths with in diabetic patients, there was a likewise statistically nonsignificant increase in 5-year ARC possible stent thrombosis events ( vs. CS: 9.9% vs. 5.9%, p 0.22). Study limitations. The SIRIUS trial was the first large randomized trial evaluating the clinical outcomes of a new drug-eluting stent compared with a blinded control. As such, the patient population was restricted in complexity, and so-called off label anatomic categories that might result in higher clinical event rates were systematically excluded (6,7,28). The use of dual antiplatelet therapy (aspirin and clopidogrel) was only mandated by protocol for 3 months, and no long-term information on use of antiplatelet agents was collected. Diabetic patients were neither randomized nor stratified for enrollment in this study; thus, evaluations of outcomes in diabetic patients were not pre-specified and are post hoc analyses. Reported outcomes in diabetic patients are thus subject to the limitations of such an analysis. Conclusions The 5-year follow-up of the double-blinded, randomized SIRIUS trial comparing with confirms the long-term safety and efficacy of in patients with simple and medium complexity native coronary lesions. All antirestenosis efficacy parameters of reduced revascularization, all safety end points (death, MI, and stent thrombosis), and composite safety and efficacy end points (MACE and TVF) suggested maintained benefit of in reducing subsequent revascularization events without adversely affecting safety. Undoubtedly, because this study only represents a slice of the patient population currently treated with, additional long-term follow-up studies are essential to determine the safety and efficacy of in more complex patient groups and higher-risk anatomic subsets. Reprint requests and correspondence: Dr. Giora Weisz, Center for Interventional Vascular Therapy (CIVT), Columbia University Medical Center, Herbert Irving Pavilion, 161 Fort Washington Avenue, 5th Floor, New York, New York 10032. E-mail: gweisz@ crf.org. REFERENCES 1. Morice MC, Serruys PW, Sousa JE, et al. A randomized comparison of a sirolimus-eluting stent with a standard stent for coronary revascularization. N Engl J Med 2002;346:1773 80. 2. Fajadet J, Morice MC, Bode C, et al. Maintenance of long-term clinical benefit with sirolimus-eluting coronary stents: three-year results of the RAVEL trial. Circulation 2005;111:1040 4. 3. 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;349:1315 23. 4. Holmes DR Jr., Leon MB, Moses JW, et al. Analysis of 1-year clinical outcomes in the SIRIUS trial: a randomized trial of a sirolimus-eluting

JACC Vol. 53, No. 17, 2009 April 28, 2009:1488 97 Weisz et al. 5-Year Follow-Up of the SIRIUS Trial 1497 stent versus a standard stent in patients at high risk for coronary restenosis. Circulation 2004;109:634 40. 5. Weisz G, Leon MB, Holmes DR Jr., et al. Two-year outcomes after sirolimus-eluting stent implantation: results from the Sirolimus- Eluting Stent in de Novo Native Coronary Lesions (SIRIUS) trial. J Am Coll Cardiol 2006;47:1350 5. 6. Iakovou I, Schmidt T, Bonizzoni E, et al. Incidence, predictors, and outcome of thrombosis after successful implantation of drug-eluting stents. JAMA 2005;293:2126 30. 7. Ong AT, McFadden EP, Regar E, de Jaegere PP, van Domburg RT, Serruys PW. Late angiographic stent thrombosis (LAST) events with drug-eluting stents. J Am Coll Cardiol 2005;45:2088 92. 8. Daemen J, Wenaweser P, Tsuchida K, et al. Early and late coronary stent thrombosis of sirolimus-eluting and paclitaxel-eluting stents in routine clinical practice: data from a large two-institutional cohort study. Lancet 2007;369:667 78. 9. Camenzind E, Steg PG, Wijns W. Stent thrombosis late after implantation of first-generation drug-eluting stents: a cause for concern. Circulation 2007;115:1440 55, discussion 1455. 10. Lagerqvist B, James SK, Stenestrand U, Lindback J, Nilsson T, Wallentin L. Long-term outcomes with drug-eluting stents versus bare-metal stents in Sweden. N Engl J Med 2007;356:1009 19. 11. Nordmann AJ, Briel M, Bucher HC. Mortality in randomized controlled trials comparing drug-eluting vs. bare metal stents in coronary artery disease: a meta-analysis. Eur Heart J 2006;27:2784 814. 12. Mauri L, Hsieh WH, Massaro JM, Ho KK, D Agostino R, Cutlip DE. Stent thrombosis in randomized clinical trials of drug-eluting stents. N Engl J Med 2007;356:1020 9. 13. Moussa I, Leon MB, Baim DS, et al. Impact of sirolimus-eluting stents on outcome in diabetic patients: a SIRIUS (SIRolImUS-coated Bx Velocity balloon-expandable stent in the treatment of patients with de novo coronary artery lesions) substudy. Circulation 2004;109: 2273 8. 14. Waksman R, Ajani AE, White RL, et al. Five-year follow-up after intracoronary gamma radiation therapy for in-stent restenosis. Circulation 2004;109:340 4. 15. Leon MB, Teirstein PS, Moses JW, et al. Declining long-term efficacy of vascular brachytherapy for in-stent restenosis: 5-year follow-up from the gamma-1 randomized trial. Circulation 2004;110:III405. 16. Gruberg L, Caiati R, Aronson D, et al. Five-year clinical follow up after intracoronary radiation for the prevention of in-stent restenosis. J Invasive Cardiol 2006;18:494 8. 17. Influence of diabetes on 5-year mortality and morbidity in a randomized trial comparing CABG and PTCA in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI). Circulation 1997;96:1761 9. 18. Hoffman SN, TenBrook JA, Wolf MP, Pauker SG, Salem DN, Wong JB. A meta-analysis of randomized controlled trials comparing coronary artery bypass graft with percutaneous transluminal coronary angioplasty: one- to eight-year outcomes. J Am Coll Cardiol 2003;41: 1293 304. 19. Cutlip DE, Chhabra AG, Baim DS, et al. Beyond restenosis: five-year clinical outcomes from second-generation coronary stent trials. Circulation 2004;110:1226 30. 20. Pfisterer M, Brunner-La Rocca HP, Buser PT, et al. Late clinical events after clopidogrel discontinuation may limit the benefit of drug-eluting stents: an observational study of drug-eluting versus bare-metal stents. J Am Coll Cardiol 2006;48:2584 91. 21. Daemen J, Garcia-Garcia HM, Kukreja N, et al. The long-term value of sirolimus- and paclitaxel-eluting stents over bare metal stents in patients with diabetes mellitus. Eur Heart J 2007;28:26 32. 22. Lagerqvist B, James SK, Stenestrand U, Lindback J, Nilsson T, Wallentin L. Long-term outcomes with drug-eluting stents versus bare-metal stents in Sweden. N Engl J Med 2007;356:1009 19. 23. Kornowski R, Mintz GS, Kent KM, et al. Increased restenosis in diabetes mellitus after coronary interventions is due to exaggerated intimal hyperplasia. A serial intravascular ultrasound study. Circulation 1997;95:1366 9. 24. Seven-year outcome in the Bypass Angioplasty Revascularization Investigation (BARI) by treatment and diabetic status. J Am Coll Cardiol 2000;35:1122 9. 25. King SB III, Kosinski AS, Guyton RA, Lembo NJ, Weintraub WS. Eight-year mortality in the Emory Angioplasty versus Surgery Trial (EAST). J Am Coll Cardiol 2000;35:1116 21. 26. Abizaid A, Costa MA, Centemero M, et al. Clinical and economic impact of diabetes mellitus on percutaneous and surgical treatment of multivessel coronary disease patients: insights from the Arterial Revascularization Therapy Study (ARTS) trial. Circulation 2001;104: 533 8. 27. Spaulding C, Daemen J, Boersma E, Cutlip DE, Serruys PW. A pooled analysis of data comparing sirolimus-eluting stents with baremetal stents. N Engl J Med 2007;356:989 97. 28. Weisz G, Stone GW. Safety and efficacy of drug-eluting stents: on-label and off-label perspectives. Rev Cardiovasc Med 2008;9:46 61. Key Words: coronary disease y diabetes mellitus y restenosis y stents y thrombosis.