The superiority of everolimus-eluting stents (EES) over sirolimus-eluting stents (SES) for long-term clinical outcomes has not been yet firmly established. We conducted a systematic review and a meta-analysis of randomized controlled trials (RCTs) comparing EES directly with SES using the longest available follow-up data. We searched PubMed, the Cochrane database, and ClinicalTrials.gov for RCTs comparing outcomes between EES and SES and identified 13,434 randomly assigned patients from 14 RCTs. EES was associated with significantly lower risks than SES for definite stent thrombosis (ST), definite/probable ST, target-lesion revascularization (TLR), and major adverse cardiac events (MACE). The risks for all-cause death and myocardial infarction were similar between EES and SES. By the stratified analysis according to the timing after stent implantation, the favorable trend of EES relative to SES for ST, TLR, and MACE was consistently observed both within and beyond 1 year. The lower risk of EES relative to SES for MACE beyond 1 year was statistically significant (pooled odds ratio 0.77, 95% confidence interval 0.61 to 0.96, p = 0.02). In conclusion, the current meta-analysis of 14 RCTs directly comparing EES with SES suggested that EES provided improvement in both safety and efficacy; EES compared with SES was associated with significantly lower risk for definite ST, definite/probable ST, TLR, and MACE. The direction and magnitude of the effect beyond 1 year were comparable with those observed within 1 year.
The superiority of everolimus-eluting stent (EES; Xience V [Abbott Vascular, Santa Clara, California]/PROMUS [Boston Scientific, Natick, Massachusetts]) over sirolimus-eluting stent (SES; Cypher/Cypher select/Cypher select plus [Cordis Corporation, Johnson and Johnson, Warren, New Jersey]) is still unclear. There have been 3 reports of meta-analysis comparing EES directly with SES. The latest report including 11 randomized controlled trials (RCTs) with 12,869 patients demonstrated superiority of EES over SES in terms of definite stent thrombosis (ST) and repeat revascularization (defined as target-lesion revascularization [TLR] or target-vessel revascularization varying among trials) but failed to show the differences in major adverse cardiac events (MACE) and definite or probable ST between EES and SES. Since the publication of the last meta-analysis, a few new RCTs were reported and several RCTs included in the previous meta-analyses extended the follow-up duration, and beyond 1 year, outcomes became evaluable. In addition, the continuous hazards of SES have been postulated, and further analysis comparing EES and SES beyond 1 year is warranted. Therefore, we conducted a systematic review and a meta-analysis of 14 RCTs comparing EES directly with SES using the longest available follow-up data.
Methods
We searched all reported trials comparing EES with SES in patients with coronary artery disease, using the term “everolimus eluting stent,” “Xience,” “Promus,” “everolimus-eluting stent,” “sirolimus eluting stent,” “Cypher,” and “sirolimus-eluting stent.” We searched the US National Library of Medicine (Pubmed at http://www.pubmed.gov ), the US National Institutes of Health clinical trials registry ( http://www.clinicaltrials.gov ), and the Cochrane Central Register of Controlled trials ( http://www.mrw.interscience.weiley.com/cochrane/cochrane_clcentral_articles_fs.html ). From the gathered studies, RCT comparing EES and SES were extracted, and if there were several reports from the same RCTs, we selected the report providing the longest follow-up data for the trial. The last search was performed in May 2014. Each trial was evaluated by referring to the Cochrane Collaboration’s tool (recommendation for the qualification of RCT) for the adequacy of allocation concealment, performance of the analysis according to the intention-to-treat principle, and blind assessment of the outcomes of interest. Because this study used only published reports without individual patient information, the procedure of informed consent and institutional review board approval was not applicable.
The end points analyzed in this study included ST (definite and definite/probable), repeat revascularization (TLR, TLR equivalent, and target-vessel revascularization equivalent), all-cause death, myocardial infarction, and MACE. ST was defined according to the Academic Research Consortium definition. Academic Research Consortium definition was adopted in all trials except for one, which reported no event of ST. ST was further analyzed according to the timing (early ST within 30 days, late ST between 31 days and 1 year, and very late ST beyond 1 year), and study protocol (protocol-mandated clopidogrel duration and protocol-mandated angiographic follow-up).
The definitions and reporting of repeat revascularization varied across trials, 9 trials reported both TLR and target-vessel revascularization, 2 trials mentioned TLR only, and the other 3 trials stated target-vessel revascularization only ( Supplementary Table 1 ). We prespecified TLR as the primary outcome measure for repeat revascularization. For TLR, we used the reported risk estimates of TLR only based on the definition in each trial and excluded the trials not reporting the TLR outcome. For TLR equivalent, we selected the risk estimate of TLR if available and used the risk estimate of target-vessel revascularization as an alternative in case the risk estimate of TLR was not available. For target-vessel revascularization equivalent, we used the risk estimate of TLR as an alternative if target-vessel revascularization outcomes were not available.
Myocardial infarction and MACE were evaluated according to the definitions in each trial. The definition of myocardial infarction included periprocedural myocardial infarction in all trials except for 2 trials, in which it was not clearly stated whether they measured cardiac biomarker for detection of periprocedural myocardial infarction. The definitions of MACE varied slightly across trials but consistently included both safety and efficacy components ( Table 1 ).
| Study | Studies with follow-up >1-year | Studies with follow-up <=1-year | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| BASKET -PROVE | ISAR -TEST-4 | SORT OUT IV | RESET | SEA -CORP † | XAMI/APPENDIX-AMI † | EXCELLENT | LONG-DES III | ESSENCE -DIABETES | Song.HG et al. | Sakakibara.T et al. | CIBELES | |
| Number of Patients | 1549 | 1304 | 2774 | 3196 | 443 | 1602 | 1443 | 450 | 300 | 66 | 100 | 207 |
| EES: Number of Patients | 774 | 652 | 1390 | 1596 | 223 | 902 | 1079 | 224 | 149 | 34 | 50 | 106 |
| SES: Number of Patients | 775 | 652 | 1384 | 1600 | 220 | 700 | 364 | 226 | 151 | 32 | 50 | 101 |
| Age (years) | 66 | 67 | 64 | 69 | 64 | 63 | 63 | 63 | 63 | – ∗ | 66 | 64 |
| Male | 75% | 77% | 76% | 77% | 80% | 73% | 65% | 70% | 59% | 52% | 70% | 83% |
| Diabetes | 17% | 29% | 14% | 45% | 31% | 14% | 38% | 30% | 100% | 30% | 70% | 36% |
| Hypertension | 61% | 68% | 55% | 80% | 68% | 39% | 73% | 59% | 71% | 56% | 70% | 68% |
| Dyslipidemia | 63% | 65% | 71% | 75% | 54% | 42% | 76% | 57% | 38% | 79% | 32% | 71% |
| Smoker | 32% | 16% | 30% | 21% | 21% | 36% | 27% | 22% | 24% | 12% | 24% | 56% |
| Acute Coronary Syndrome | 65% | 40% | 42% | 18% | 76% | 67% | 52% | 42% | 42% | N/A | N/A | 44% |
| Acute Myocardial Infarction | 33% ‡ | 11% | 10% ‡ | 6% | N/A | 44% | 10% | N/A | 5% | N/A | N/A | N/A |
| Multivessel Disease | 43% | 86% | N/A | 47% | 45% | 52% | 52% | 56% | 55% | 30% | N/A | N/A |
| Bifurcation | 8% | 29% | 12% | 40% | 100% | 19% | 11% | 41% | N/A | 14% | 31% | 25% |
| Chronic Total Occlusion | 4% | 7% | 6% | 6% | N/A | N/A | 4% | N/A | N/A | N/A | N/A | 100% |
| Follow-up Period (months) | 24 | 36 | 24 | 36 | 36 | 24 | 12 | 12 | 12 | 12 | 12 | 12 |
| Clopidogrel Duration | 12 | ≥6 | 12 | Not mandated | 12 | 12 | 6 or 12 | ≥12 | ≥12 | ≥6 | N/A | ≥9 |
| Follow-up Angiography | No | Yes | No | No | No | No | Yes | Yes | Yes | Yes | Yes | Yes |
| Definition of Major Adverse Cardiac Events | Death, MI, TVR | Cardiac death, Target-vessel MI, TLR | Cardiac death, MI, Definite ST, TVR | Cardiac death, MI, Ischemia-driven TLR | Death, AMI, TVR | Cardiac death, MI, TVR | Cardiac death, Target-vessel related MI, Clinically indicated TLR | All-cause death, MI, Ischemia-driven TVR | Death, MI, Ischemia-driven TLR | Death, MI, TLR | All-cause death, Nonfatal MI, TLR | Death, MI, New target-vessel revascularization |
∗ Song HG et al reported median age of each groups (EES; 65 years, SES; 61 years) instead of mean age.
† Results of cooperative studies; SEA-CORP: the SEA-SIDE and CORpal trials, XAMI/APPENDIX-AMI: the XAMI and APPENDIX-AMI trials.
The baseline characteristics were collected from each reported RCT. We used the risk estimates for each event of interest from individual reported RCT. To calculate the risk estimates for each trial, we used the reported absolute number of patients with at least 1 event, and if the absolute numbers were not reported, we estimated the number of patients with at least 1 event from the reported cumulative incidence and the number of patients assigned to each group. When neither the absolute number nor the cumulative incidence for the outcome of interest was available, we used the reported Kaplan-Meier curve to count or to read the cumulative incidence of the outcome. The longest follow-up data for each event of interest were collected. Trials in which the event of interest was not observed in either study group were excluded from the analysis for that event. We also performed stratified analysis of each clinical end point according to the timing after stent implantation (within 1 year and beyond 1 year). The numbers of events for each clinical outcome within 1 year and beyond 1 year were separately collected from the reported RCTs.
We used the Cochrane test to assess heterogeneity among trials and calculated the I-squared statistic for quantification, with values <25% indicating low, 25% to 50% indicating moderate, and >50% indicating high heterogeneity. The Mantel-Haenszel method for fixed-effect model was used for the calculations of pooled odds ratio, unless heterogeneity exceeded moderate in each outcome. Publication bias was assessed by Begg’s funnel plot and Egger’s test.
Results were considered statistically significant at 2-sided p <0.05. Statistical analysis was performed using Stata software, version 13.1 (Stata Corp., College Station, Texas).
Results
From the 1,397 studies initially collected for the comparison between EES and SES, 14 RCTs were selected for the current meta-analysis ( Supplementary Table 1 ). The study flow chart was illustrated along with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement ( Supplementary Figure A ). The investigators of the 4 RCTs performed 2 co-operative studies to investigate long-term clinical outcomes. Because the co-operation was performed without exclusion of patients and the baseline trial designs were similar, we decided to include the results of the co-operative analysis. Therefore, a total of 16 reports from 14 RCTs were used in this meta-analysis.
Of 13,434 patients from the 14 RCTs, 7,179 patients (53.4%) were assigned to EES and 6,255 patients (46.6%) to SES. The trials with follow-up beyond 1 year were separately analyzed to assess the differences in long-term outcomes between EES and SES (follow-up ≤1 year, 7 trials with 3,191 patients; and follow-up >1 year, 7 trials with 10,243 patients). The main characteristics of the individual trials were summarized in Table 1 and Supplementary Table 1 . All trials were described as randomized, and characteristics regarding allocation concealment, blinding of participants/personnel/clinical outcome assessment, incomplete outcome data, selective reporting, and other sources of bias were included in Supplementary Table 1 . We judged all the results from the included reports to have quality sufficient for the meta-analysis.
The risk for definite ST was significantly lower for EES compared with SES ( Figure 1 ). The risk for definite or probable ST was also significantly lower for EES compared with SES ( Supplementary Figure B ). In the subgroup of trials with follow-up >1 year, the lower risk of EES relative to SES for definite and definite/probable ST was also significant ( Figure 1 , Supplementary Figure B ). The trends toward lower risk of EES relative to SES for definite and definite/probable ST were consistently seen regardless of the presence or absence of protocol-mandated clopidogrel duration and scheduled angiographic follow-up ( Figure 2 ). The lower risk of EES relative to SES for TLR was significant ( Figure 1 ). In the subgroup of trials with follow-up >1 year, the lower risk of EES relative to SES for TLR was also significant ( Figure 1 ). The significantly lower risk of EES relative to SES was also seen for TLR equivalent and target-vessel revascularization equivalent ( Supplementary Figure B ). The risks for all-cause death and myocardial infarction were not significantly different between EES and SES ( Figure 1 ). The risk for MACE was significantly lower for EES compared with SES ( Figure 1 ). In the subgroup of trials with follow-up >1 year, the lower risk of EES relative to SES for MACE was also significant ( Figure 1 ).
