The high rate of restenosis associated with percutaneous coronary intervention (PCI) procedures1,2 can be reduced with the implantation of metallic stents into the stenotic vessels.3,4 Initially, bare-metal stents (BMS) were introduced, which were found to substantially reduce the rate of restenosis (~20–30%). However, BMS were associated with a high rate of in-stent restenosis. The knowledge that neointimal formation can result in restenosis after stent implantation led to the development of drug-eluting stents (DES), which are now commonly used for PCI procedures.5–7 DES are metallic stents coated with antiproliferative agents. The drug in question is intended to reduce inflammation or inhibit cell proliferation. In randomised controlled trials, a significant reduction in restenosis, late lumen loss and target lesion revascularisation (TLR) has been noted with DES compared with BMS.8–11 The first-generation DES include the sirolimus-eluting Cypher™ stent (Cordis Corp, Miami, Florida) and the polymer-based paclitaxel-eluting Taxus® Express™ stent (Boston Scientific, Natick, Massachusetts). Data from large randomised clinical trials have demonstrated the efficacy and safety of paclitaxel-eluting stents (PES) and sirolimus-eluting stents (SES).10–12 However, a higher rate of predominantly late stent thrombosis (ST) has been reported with these DES compared with BMS,12 and this has been linked to the possibility of delayed and incomplete endothelialisation with the DES.13,14 Second- and third-generation DES have subsequently been developed with the aim of further improving the safety and efficacy profiles of the DES.
The polymer-based everolimus-eluting Xience V™ stent (Abbott Vascular, Santa Clara, CA, US) is one example of a second-generation DES. Pre-clinical data from a 14-day rabbit iliac model have shown a faster re-endothelialisation rate with everolimus-eluting stents (EES) versus sirolimus-, zatarolimus- or paclitaxel-eluting stents.15 In terms of clinical results, the SPIRIT series of randomised clinical trials have shown the EES to be safe and efficacious for the treatment of coronary artery disease (CAD).16–18 An analysis of the pooled two-year results from SPIRIT II and SPIRIT III showed that patients treated with EES experienced low rates of additional cardiac death, myocardial infarction (MI) and ST between one- and two-year follow-up.19
Furthermore, three-year follow-up results from the SPIRIT III trial have shown a significantly reduced rate (p=0.07–0.003) of major cardiac adverse events (MACE, defined as a composite of all death, MI, ischaemia-driven TLR [ID-TLR] by coronary artery bypass grafting [CABG] or PCI), ID-TLR and MI, plus a non-inferior rate of cardiac death, with the EES compared with PES.20 One-year follow-up results from the SPIRIT IV trial have shown a significantly reduced rate (p=0.004–0.0008) of target lesion failure (TLF, defined as a composite of cardiac death, target vessel MI or ID-TLR), ID-TLR, target vessel MI, Academic Research Consortium (ARC)-defined definite or probable ST and protocol-defined ST, plus a non-inferior rate of cardiac death, with EES versus PES.21 The safety and efficacy of EES in patients referred for PCI have been compared with a second-generation polymer-based PES (Taxus Liberté®, Boston Scientific, Natick, MA, US) by the all-comer, real-world design COMPARE trial. The Taxus Liberté stent is a second-generation stent that was developed because the first-generation Taxus Express stent was associated with a high incidence of peri-procedural MIs related to side-branch occlusion. The purpose of this article is to present the rationale and design of the COMPARE trial.
Rationale and Design of the COMPARE Trial
The COMPARE trial was recently completed. The main objective of COMPARE was to carry out a head-to-head comparison of the clinical outcomes achieved with the two second-generation DES (EES and PES) in a real-world population.22,23 The COMPARE trial was a European-based, physician-initiated, single-centre, prospective, randomised, open-label, all-comer study. Patients who were referred for PCI procedures at the centre were assessed for eligibility for the study. Patients were randomised on a 1:1 basis to either EES or PES. The primary end-point of the study was the composite end-point of all death, non-fatal MI and target vessel revascularisation (TVR) at one year. The secondary end-points included the combined end-point of cardiac death, non-fatal MI and ischaemic-driven TLR rate at 12-month follow-up, the combined end-point of all death, non-fatal MI and TVR rate at two, three, four and five years and the incidence of ARC-defined definite, probable or possible ST at 12 months, three and five years.
The physician-initiated COMPARE trial was supported by unrestricted research grants. In contrast to the majority of controlled trials, the objective was to evaluate the efficacy and safety of the two second-generation DES in a real-world setting with a more diverse and complex patient population. This would potentially provide data on treatment outcomes that are much closer to the results achieved in clinical practice rather than the outcomes measured by controlled trials. Essentially, COMPARE was a patient-based study rather than a device-based study. An all-comer real-world patient accrual method was utilised in the COMPARE trial. This accrual method not only provided a real-world patient population setting but also meant that there were few inclusion and exclusion criteria. All patients who were eligible for PCI and with a life expectancy of more than five years were included in this study. Excluded patients were those patients who had not received dual antiplatelet therapy for 12 months or those with cardiogenic shock at presentation. Patients accrued to the study included both males and females. Clinical presentations included ST-segment-elevation MI (STEMI) and non-STEMI patients as well as those with stable angina, unstable angina or silent ischaemia. It is particularly important to note that acute coronary syndrome patients were recruited, as these are often patients who require treatment with PCI. These patients were excluded from the SPIRIT IV trial. Furthermore, in the COMPARE trial the second-generation Taxus Liberté stent was used as the comparator stent, whereas in SPIRIT IV the first-generation Taxus Express stent was used. In total, 1,800 consecutive patients referred for PCI were enrolled in the study. Baseline characteristics were similar between the EES and PES groups.
The open-label design of the trial raises the question of whether there is the risk of introducing bias in assessing outcomes. However, there is expected to be little risk of bias as the study was not totally open-label and the characteristics of the trial design may be considered semi-blinded: the operator was only made aware of the type of stent at the time of implantation and the patients were totally unaware of the type of stent implanted. Furthermore, post-implantation, the clinical events were adjudicated by an independent clinical event committee while TVRs were analysed by an independent quantitative coronary angiography core laboratory.
The end-points utilised by the COMPARE trial differ from those of other trials evaluating the EES. The SPIRIT II and III trials involving EES utilised end-points designed to show the efficacy and safety of the device for a particular indication. These end-points included angiographic restenosis and the degree of neointimal formation or late lumen loss. These wend-points act as surrogate measures of clinical efficacy and provide more device-orientated data. By contrast, COMPARE utilised end-points that are intended to clinically reflect relevant outcomes. Accordingly, the objective of the COMPARE data is to provide relevant patient-orientated outcomes of the efficacy and safety of the tested devices.
The characteristics of the COMPARE study design, such as physician initiation, randomisation and all-comer real-world patient accrual, allowed for the comparison and assessment of the clinical efficacy and safety of the evaluated stents in a real-world, diverse, complex patient population. The COMPARE trial used end-points that can be considered more clinically relevant compared with the end-points used by some of the early SPIRIT trials, where only surrogate measures of clinical outcome were utilised. Preliminary data from the COMPARE trial have recently been presented that demonstrate superior efficacy and safety of the EES compared with PES in a real-world setting and appear consistent with data from the SPIRIT IV trial. Publication of the full results of COMPARE is eagerly anticipated, and will appear in The Lancet (online 8 January 2010).