|
|
|
Background Percutaneous coronary intervention (PCI) involving drug-eluting stents is increasingly used to treat complex coronary artery disease, although coronary-artery bypass grafting (CABG) has been the treatment of choice historically. Our trial compared PCI and CABG for treating patients with previously untreated three-vessel or left main coronary artery disease (or both).
Methods We randomly assigned 1800 patients with three-vessel or left main coronary artery disease to undergo CABG or PCI (in a 1:1 ratio). For all these patients, the local cardiac surgeon and interventional cardiologist determined that equivalent anatomical revascularization could be achieved with either treatment. A noninferiority comparison of the two groups was performed for the primary end point — a major adverse cardiac or cerebrovascular event (i.e., death from any cause, stroke, myocardial infarction, or repeat revascularization) during the 12-month period after randomization. Patients for whom only one of the two treatment options would be beneficial, because of anatomical features or clinical conditions, were entered into a parallel, nested CABG or PCI registry.
Results Most of the preoperative characteristics were similar in the two groups. Rates of major adverse cardiac or cerebrovascular events at 12 months were significantly higher in the PCI group (17.8%, vs. 12.4% for CABG; P=0.002), in large part because of an increased rate of repeat revascularization (13.5% vs. 5.9%, P<0.001); as a result, the criterion for noninferiority was not met. At 12 months, the rates of death and myocardial infarction were similar between the two groups; stroke was significantly more likely to occur with CABG (2.2%, vs. 0.6% with PCI; P=0.003).
Conclusions CABG remains the standard of care for patients with three-vessel or left main coronary artery disease, since the use of CABG, as compared with PCI, resulted in lower rates of the combined end point of major adverse cardiac or cerebrovascular events at 1 year. (ClinicalTrials.gov number, NCT00114972
[ClinicalTrials.gov]
.)
Percutaneous coronary intervention (PCI) was introduced in 1977.7 Experience with this approach, coupled with improved technology, has made it possible to treat increasingly complex lesions and patients with a history of clinically significant cardiac disease, risk factors for coronary artery disease, coexisting conditions, or anatomical risk factors.8,9 Several trials comparing PCI involving bare-metal stents with CABG in patients with multivessel disease (e.g., the Arterial Revascularization Therapies Study Part I [ARTS I], the Medicine, Angioplasty, or Surgery Study for Multivessel Coronary Artery Disease [MASS II; Current Controlled Trials number, ISRCTN66068876 [controlled-trials.com] ], the Argentine Randomized Study of Coronary Angioplasty with Stenting versus Coronary Bypass Surgery in Patients with Multiple Vessel Disease [ERACI-II], and the Angina with Extremely Serious Operative Mortality Evaluation [AWESOME]) showed similar survival rates but higher revascularization rates among patients with bare-metal stents at 5 years. Others have shown a significant long-term survival advantage with surgery (e.g., the Stent or Surgery [SOS] study).10,11,12 Studies comparing PCI involving drug-eluting stents with CABG have generally been smaller and nonrandomized.13,14,15,16,17,18,19,20,21,22,23,24
Data from randomized, controlled trials of drug-eluting stents as compared with bare-metal stents have shown significant reductions in the rate of repeat intervention, with similar rates of death and myocardial infarction.25 These improvements have led to expanded use of PCI in patients with complex coronary anatomical features, though most randomized trials comparing drug-eluting stents and bare-metal stents excluded such patients. According to current guidelines,26 CABG remains the treatment of choice for patients with severe coronary artery disease, including those with left main coronary artery disease and those with three-vessel disease. There is a lack of data from adequately powered randomized trials of PCI in such patients. Thus, PCI is being performed in this group without adequate support from evidence-based medicine and randomized clinical trials.27
In the Synergy between PCI with Taxus and Cardiac Surgery (SYNTAX) trial, we assessed the optimal revascularization strategy for patients with previously untreated three-vessel or left main coronary artery disease and defined the populations of patients for whom only one revascularization method will be effective.
Methods
Study Design
The SYNTAX trial is a prospective, clinical trial conducted in 85 sites and approved by the institutional review board at each participating center. The study had an "all-comers" design involving the consecutive enrollment of all eligible patients with three-vessel or left main coronary artery disease at sites in 17 countries in Europe and the United States. The study design has been described previously.28 Criteria for study and registry enrollment and outcome data are described in the Supplementary Appendix. The authors designed the study, as part of their role on the steering committee, in collaboration with the sponsor, Boston Scientific. The sponsor was involved in collection and source verification of the data, with oversight by an independent clinical events committee. The sponsor's biostatisticians performed the analyses; however, data analyses were verified independently by a statistician on the data and safety monitoring committee. The authors wrote the manuscript and vouch for the completeness and accuracy of the data gathering and analysis.
Selection and Randomization of Patients
A local interventional cardiologist and cardiac surgeon at each site prospectively evaluated eligible patients with previously untreated three-vessel coronary disease and those with left main coronary artery disease (alone or with one-, two-, or three-vessel disease). Inclusion and exclusion criteria are listed in the Methods section of the Supplementary Appendix. Patients in whom it was determined that equivalent anatomical revascularization could be achieved with either CABG or PCI involving Taxus Express paclitaxel-eluting stents (Boston Scientific) were randomly assigned to undergo one of the two treatment options by means of an interactive voice-responding system. Randomization was stratified at each site according to the presence or absence of left main coronary artery disease and medically treated diabetes (diabetes for which the patient was receiving oral hypoglycemic agents or insulin at the time of enrollment). Patients for whom only one treatment option was suitable were entered into a parallel, nested registry: the PCI registry for CABG-ineligible patients and the CABG registry for PCI-ineligible patients.
All diagnostic angiograms and electrocardiograms were reviewed by staff at an independent core laboratory (Cardialysis, Rotterdam, the Netherlands) who were unaware of the treatment assignments. Diagnostic angiograms were scored, according to the SYNTAX score algorithm,29 at the site and at the core laboratory. In addition, staff at an independent central chemistry laboratory (Covance, Indianapolis and Geneva) who were unaware of treatment assignments assessed selected variables.
The institutional review board at each site approved the protocol, and all patients provided written informed consent. The protocol and consent forms were consistent with the Food and Drug Administration's Guidance for Industry E6 Good Clinical Practice, the Declaration of Helsinki, the International Conference on Harmonisation, and all local regulations, as appropriate.
Revascularization and Pharmacologic Treatment
Patients were treated with the intention of achieving complete revascularization of all vessels at least 1.5 mm in diameter with stenosis of 50% or more, as identified by the local interventional cardiologist and cardiac surgeon. The surgical technique for CABG, the approaches used for stent implantation, and the postprocedure medication regimen were chosen according to local clinical practice. In patients who underwent PCI, antiplatelet medication was prescribed on the basis of the directions for use of the Taxus Express stent and local clinical practice. In most centers, thienopyridines were continued even after 6 months, with 71.1% of patients receiving them at 12 months. Aspirin was prescribed indefinitely for all patients who underwent randomization. Use of the standard of postintervention care was recommended.30 Procedural details are described in the Methods section of the Supplementary Appendix.
Primary End Point
The primary clinical end point was a composite of major adverse cardiac and cerebrovascular events (i.e., death from any cause, stroke, myocardial infarction, or repeat revascularization) throughout the 12-month period after randomization. An independent clinical events committee (including cardiologists, cardiac surgeons, and a neurologist; see the list in the Supplementary Appendix) adjudicated all primary clinical end points, staged procedures, and cases in which the sternum was reopened.
Statistical Analysis
The primary analysis was a noninferiority comparison of the two treatments for the primary end point of adverse binary cardiac or cerebrovascular events in all patients who underwent randomization (on an intention-to-treat basis). If the one-sided 95% upper confidence limit for the difference was less than the prespecified delta value (6.6%), PCI with the drug-eluting stents would be considered to be noninferior to CABG in the overall randomized cohort. The noninferiority margin was based on historical data (see the Methods section in the Supplementary Appendix). We calculated the means (±SD) for continuous variables in each of the two groups and compared them using Student's t-test. Binary variables are reported as counts and percentages, and differences between the two groups were assessed by means of the chi-square or Fisher's exact test. Cumulative event rates were estimated by means of the Kaplan–Meier method. In addition, the 12-month rates of major adverse cardiac or cerebrovascular events were analyzed on the basis of the SYNTAX score and compared with the use of a chi-square test. The SYNTAX score reflects a comprehensive anatomical assessment, with higher scores indicating more complex coronary disease; a low score was defined as 
22, an intermediate score as 23 to 32, and a high score as 
33 (see the Supplementary Appendix for details).
Results
Study Participants
From March 2005 through April 2007, a total of 4337 patients with previously untreated three-vessel or left main coronary artery disease (or both) were screened (Figure 1). After consideration by the local interventional cardiologist and cardiac surgeon and after written informed consent was obtained, 3075 patients (70.9%) were included in the study. Of these, 1800 patients were randomly assigned to undergo CABG (897 patients) or PCI with drug-eluting stents (903 patients) at sites in the United States (245 patients) and in Europe (1555 patients). The reasons for exclusion of the remaining 1262 patients are listed in Figure 1. Only one treatment option was suitable in 1275 patients (29.4%), who were enrolled in the nested registry for CABG (1077 patients) or PCI (198 patients).
|
150 mg per deciliter [1.7 mmol per liter]), and patients with reduced high-density lipoprotein cholesterol levels (<40 mg per deciliter [1.0 mmol per liter] for men or <50 mg per deciliter [1.3 mmol per liter] for women) were higher in the CABG group. A total of 38.8% of patients in the CABG group and 39.5% of those in the PCI group had left main coronary artery disease, with or without additional diseased vessels. Approximately 25% of patients had medically treated diabetes, of whom about one third required insulin. Moreover, nearly half the patients (45.8%) met the criteria for the metabolic syndrome.33 More than 20% of patients in both groups were considered to be at high surgical risk, on the basis of a European System for Cardiac Operative Risk Evaluation (euroSCORE)31 value of 6 or more (24.9% in the CABG group and 24.7% in the PCI group, P=0.94) and a Parsonnet score32 of 15 or more (20.2% and 20.5%, respectively; P=0.87).
|
The length of time between randomization and performance of the study procedure, the duration of the procedure, and the duration of the postprocedural hospital stay were significantly greater with CABG than with PCI (Table 1). A higher proportion of patients had complete revascularization after CABG than after PCI (63.2% vs. 56.7%, P=0.005). Medical management at discharge differed between the CABG and PCI groups: patients who underwent CABG received less pharmacologic treatment, whereas those who underwent PCI were consistently treated with antiplatelet medications (Table 2). In the CABG group, off-pump surgery was performed in 15.0% of patients, one or more arterial grafts were used in 97.3% of patients, and an average of 2.8 conduits and 3.2 distal anastomoses per patient were performed. In the PCI group, 14.1% of patients underwent staged procedures, 63.1% had at least one bifurcation or trifurcation treated, more than four stents on average were implanted per patient, and a third of patients had placement of stents with a total length of more than 100 mm.
|
Primary Outcome
Preprocedural rates of major adverse cardiac or cerebrovascular events were low and did not differ significantly between the two groups (0.9% in the CABG group and 0.3% in the PCI group, P=0.13) (Table 6 in the Supplementary Appendix), as was the case for in-hospital rates. The preprocedural rates of two of the individual components of the primary outcome, stroke and myocardial infarction, were similar in the two groups (Table 2 in the Supplementary Appendix). At 12 months, the incidence of major adverse cardiac or cerebrovascular events was lower in the CABG group (12.4%) than in the PCI group (17.8%, P=0.002) (Figure 2 and Table 3). Thus, the absolute difference in the 12-month rate of major adverse cardiac or cerebrovascular events between the two groups was 5.5 percentage points, with an upper 95% confidence interval of 8.3 percentage points. The results of an as-treated analysis were similar: the 12-month rate of major adverse cardiac or cerebrovascular events was 12.3% with CABG and 17.6% with PCI (P=0.002).
|
|
Secondary Outcome
The rate of repeat revascularization at 12 months was significantly higher among patients in the PCI group than among those in the CABG group (13.5% vs. 5.9%, P<0.001) (Table 3). Most patients who underwent repeat revascularization were treated with PCI rather than CABG. The rate of stroke was significantly higher with CABG than with PCI at 12 months, even though the two groups were well balanced with regard to carotid artery disease and other risk factors for stroke (Table 3). At 12 months, the two groups had similar rates of death from any cause or myocardial infarction and of the combined end point of death from any cause, stroke, or myocardial infarction (Table 3). The rate of death from cardiac causes was greater with PCI than with CABG (3.7% vs. 2.1%, P=0.05); the rate of death from noncardiac causes, although not significant, was higher with CABG (1.4% vs. 0.7%, P=0.13). The 12-month rates of symptomatic graft occlusion (in the CABG group) and stent thrombosis (in the PCI group) were similar (P=0.89) (Table 3).
Outcomes According to the SYNTAX Score
In the CABG group, the binary 12-month rates of major adverse cardiac or cerebrovascular events were similar among patients with low SYNTAX scores (0 to 22, 14.7%), those with intermediate scores (23 to 32, 12.0%), and those with high scores (33, 10.9%) (Figure 3). In contrast, in the PCI group, the rate of major adverse cardiac or cerebrovascular events was significantly increased among patients with high SYNTAX scores (23.4%) as compared with those with low scores (13.6%) or intermediate scores (16.7%) (P=0.002 for high vs. low scores; P=0.04 for high vs. intermediate scores) (Figure 3). There was a significant interaction between SYNTAX score and treatment group (P=0.01); patients with low or intermediate scores in the CABG group and in the PCI group had similar rates of major adverse cardiac or cerebrovascular events, whereas among patients with high scores, the event rate was significantly increased in the PCI group (Figure 3).
|
The subgroups of patients with left main or three-vessel coronary artery disease were prespecified, and the study had a statistical power of 80% for each subgroup. However, the overarching statistical test was a noninferiority assessment of data from all patients with either left main coronary artery disease or three-vessel coronary disease (or both). Since noninferiority was not proven in this cohort, specific information for each subgroup is of an observational nature and is hypothesis generating. The 12-month rate of major adverse cardiac or cerebrovascular events among patients with left main coronary artery disease was similar in the CABG and PCI groups (13.7% and 15.8%, respectively; P=0.44). Although the rate of repeat revascularization among patients with left main coronary artery disease was significantly higher in the PCI group (11.8%, vs. 6.5% in the CABG group; P=0.02), this result was offset by a significantly higher rate of stroke in the CABG subgroup of patients with left main coronary artery disease (2.7%, vs. 0.3% in the corresponding PCI subgroup; P=0.01). A total of 36.6% of patients with left main coronary artery disease also had three-vessel disease. A post hoc analysis of the rates of major adverse cardiac or cerebrovascular events in the subgroups of patients with left main coronary artery disease revealed a higher rate among those who also had two- or three-vessel disease than among those with left main coronary artery disease alone or in combination with one-vessel disease (Fig. 6 in the Supplementary Appendix).
The 12-month rate of major adverse cardiac or cerebrovascular events among patients with three-vessel disease in the absence of left main coronary artery disease was significantly increased in the PCI group as compared with the CABG group (19.2% vs. 11.5%, P<0.001) (Fig. 6 in the Supplementary Appendix), as was the rate of repeat revascularization (14.6% vs. 5.5%, P<0.001). The rate of death from any cause, stroke, or myocardial infarction in this subgroup was similar with PCI and CABG (8.0% and 6.6%, respectively; P=0.39).
Comparisons of data for the cohort with left main coronary artery disease and the cohort with three-vessel disease showed that the cohort with three-vessel disease had higher rates of previous myocardial infarction, diabetes, poor left ventricular ejection fraction, and lesions with adverse characteristics (lesions that were totally occluded, bifurcated, or long). In addition, the cohort with three-vessel disease had increased numbers of treated vessels and lesions per patient.
Discussion
The SYNTAX trial was designed to compare current surgical and percutaneous techniques in patients with three-vessel or left main coronary artery disease (or both). For the primary end point, the 12-month rate of major adverse cardiac or cerebrovascular events, the noninferiority of PCI as compared with CABG was not demonstrated; CABG proved to be superior. Therefore, the findings with regard to components of the primary end point and subgroup analyses can only be considered as hypothesis-generating. Rates of death and myocardial infarction at 1 year were similar between patients who underwent CABG and those who underwent PCI with drug-eluting stents, whereas the rate of stroke was increased in the CABG group and the rate of repeat revascularization was increased in the PCI group.
Rates of repeat revascularization at 12 months were low in the PCI group, given the high rates of several known predictors of restenosis: lesions characterized by bifurcation or trifurcation (>80%), multivessel disease (>60%), diabetes (>25%), and lesions that were long (>20 mm in length, 20%) or totally occluded (>25%). This rate of repeat revascularization is lower than the rates reported in previous comparative trials involving patients with less-complex clinical profiles and lesions.10 The increase in the rate of repeat revascularization with PCI as compared with CABG did not appear to translate into a significant overall increase in the rate of death or myocardial infarction, although longer-term follow-up is needed. The risk of repeat revascularization after PCI needs to be balanced against the invasiveness of CABG and the risk of stroke, as previously reported in a meta-analysis of 23 studies comparing CABG and PCI, in which procedure-related strokes were found to be more common after CABG (in 1.2% of patients, vs. 0.6% of those undergoing PCI; P<0.001), without a concomitant decrease in survival.34
Recently, concern has been expressed about the possibility of an increased risk of late stent thrombosis with drug-eluting stents. In the SYNTAX trial, most cases of stent thrombosis occurred within 30 days after the procedure, and the 12-month rate of stent thrombosis in the PCI group was similar to the rate of symptomatic graft occlusion in the CABG group. However, as described in the literature, stent thrombosis often has more serious consequences for patients (rate of death, approximately 30%; rate of myocardial infarction, >60%)35,36 than does graft occlusion, which often results only in angina leading to revascularization.
The use of antiplatelet medication was high among patients in the PCI group (with 71.1% receiving a thienopyridine at 12 months). There was an imbalance between the two groups with regard to general medical management apart from thienopyridine use. Thienopyridine therapy was not mandated beyond 6 months in the PCI group, since the study was designed to compare current CABG and PCI practices, including medication regimens. The low rate of stroke among patients who underwent PCI may have resulted from the use of highly effective dual-antiplatelet therapy, which prevents thromboembolic events; additional treatment with antiplatelet drugs might therefore benefit patients undergoing CABG. In addition, more patients in the CABG group than in the PCI group declined to participate after providing consent; in general, this imbalance was due to the greater invasiveness of CABG.
The SYNTAX score was designed to predict outcomes related to anatomical characteristics and, to a lesser extent, the functional risk of occlusion for any segment of the coronary-artery bed (as reflected by the Leaman score37). In our study, the raw SYNTAX score was predictive of outcomes in patients who underwent PCI. In particular, among patients in the PCI group with high SYNTAX scores, not only was the overall rate of major adverse cardiac or cerebrovascular events significantly increased, but also the rate of the composite components of death, stroke, and myocardial infarction was slightly raised (11.9%, vs. 7.6% in the CABG group; P=0.08). This finding suggests that a percutaneous approach should be avoided in patients with high SYNTAX scores. Similar results were reported after the stratification of patients with three-vessel disease in the ARTS II registry.38 Retrospective analysis of the ability of the SYNTAX score to predict outcome is currently being performed and is anticipated to be used to evaluate the relative weight of the individual score components. Additional validation of the score in other populations of patients is also needed. Outcomes in the surgical group of our randomized cohort were not influenced by the SYNTAX score.
The completeness of revascularization (i.e., whether all identified lesions were treated) was determined after the procedure by the investigator. The rate of complete revascularization was lower in both treatment groups in our study than in previous studies.14,15,21,23,39 This result is most likely due to a different definition of completeness of treatment used in the earlier trials and the more complex anatomical characteristics of the patients in our trial.
Although our study provides important information about current treatment of coronary artery disease, there are limitations. First, the 12-month follow-up period may not be sufficient to reflect the true long-term effect of CABG as compared with PCI with drug-eluting stents on cardiac-related health. However, our early results in terms of major adverse cardiac or cerebrovascular events are similar to those of a meta-analysis34 of trials comparing CABG and PCI with predominantly bare-metal stents. The meta-analysis showed that the rate of major adverse cardiac or cerebrovascular events was lower with CABG than with PCI and that patients who underwent CABG had fewer repeat revascularization procedures than patients who underwent PCI. After 5 years of follow-up, the meta-analysis did not show any significant differences in rates of survival between the CABG and PCI groups,34 although other studies have shown differences in mortality.10,11,12
Second, the use of medication differed between the groups in our study, reflecting variations in standard care of patients between the two treatment groups. Third, more patients withdrew, after randomization, from the CABG group than from the PCI group. Fourth, although randomization was conducted in a blinded manner, with clinicians and participants unaware of future treatment assignments, it was not possible to blind the performance of the treatment. Finally, the definition of myocardial infarction was based on a surgical definition (the finding of a new Q-wave on electrocardiography, in association with a value for the creatine kinase MB fraction that was five times the upper limit of the normal range), which may have resulted in less severe cases of myocardial infarction being overlooked.
In conclusion, the results of our trial show that CABG, as compared with PCI, is associated with a lower rate of major adverse cardiac or cerebrovascular events at 1 year among patients with three-vessel or left main coronary artery disease (or both) and should therefore remain the standard of care for such patients.
Supported by Boston Scientific.
Dr. Kappetein reports receiving consulting and lecture fees from Boston Scientific; and Dr. Feldman, consulting and lecture fees and grant support from Boston Scientific and consulting fees from Abbott. Mr. Bass and Mr. Van Dyck and Drs. Leadley and Dawkins report being full-time employees of, and owning equity in, Boston Scientific. No other potential conflict of interest relevant to this article was reported.
We thank the clinical events committee (V.L. Babikian, D. Birnbaum, T.P. Carrel, M. Gorman, C. Hanet, O.M. Hess, E.W.L. Jansen, L.J. Kappelle, P.G. Steg) and the data and safety monitoring committee (J.-P. Bassand, T. Clayton, D.P. Faxon, B.J. Gersh, J.L. Monro, S. Pocock, M.I. Turina) of the SYNTAX trial, Kristine Roy of Boston Scientific for her help in preparing the manuscript, and Peggy Pereda of Boston Scientific for her help with data analysis.
Source Information
From Erasmus University Medical Center Rotterdam, Rotterdam, the Netherlands (P.W.S., A.P.K., M.B.); Institut Cardiovasculaire Paris Sud, Massy, France (M.-C.M.); San Raffaele Scientific Institute, Milan (A.C.); Mayo Clinic, Rochester, MN (D.R.H.); Medical City Hospital, Dallas (M.J.M.); University Hospital Uppsala, Uppsala, Sweden (E.S.); Evanston Hospital, Evanston, IL (T.E.F.); Boston Scientific, Marlborough, MA (E.J.B., N.V.D., K.L., K.D.D.); and Herzzentrum Universität Leipzig, Leipzig, Germany (F.W.M.).
This article (10.1056/NEJMoa0804626) was published at NEJM.org on February 18, 2009.
Address reprint requests to Dr. Serruys at the Erasmus University Medical Center, Rotterdam, Gravendijkwal 230, 3015 CE Rotterdam, the Netherlands, or at p.w.j.c.serruys{at}erasmusmc.nl.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
