Author + information
- Received June 14, 2010
- Accepted June 27, 2010
- Published online October 1, 2010.
- Yanai Ben-Gal, MD⁎,
- Jeffrey W. Moses, MD⁎,
- Roxana Mehran, MD⁎,
- Alexandra J. Lansky, MD⁎,
- Giora Weisz, MD⁎,
- Eugenia Nikolsky, MD⁎,
- Michael Argenziano, MD⁎,
- Matthew R. Williams, MD⁎,
- Antonio Colombo, MD†,
- Philip E. Aylward, MB, ChB, PhD‡ and
- Gregg W. Stone, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Gregg W. Stone, Columbia University Medical Center, The Cardiovascular Research Foundation, 111 East 59th Street, 11th Floor, New York, New York 10022
Objectives The aim of this study was to evaluate outcomes of patients with moderate- and high-risk acute coronary syndromes (ACS) and multivessel coronary artery disease managed with percutaneous coronary intervention (PCI) versus coronary artery bypass grafting (CABG).
Background There is uncertainty about the preferred revascularization strategy for high-risk patients with multivessel disease.
Methods Among 13,819 moderate- and high-risk ACS patients enrolled in the ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy) trial, 5,627 had multivessel disease (including left anterior descending artery involvement) and were managed by PCI (n = 4,412) or CABG (n = 1,215). Propensity score matching was applied to adjust for differences in baseline clinical and angiographic characteristics, yielding a total of 1,056 patients (528 managed by PCI, and 528 managed by CABG).
Results Propensity-matched patients undergoing CABG had higher 1-month rates of stroke (1.1% vs. 0.0%, p = 0.03) and myocardial infarction (13.3% vs. 8.8%, p = 0.03), received more blood transfusions (40.3% vs. 6.3%, p < 0.0001) and more frequently developed acute renal injury (31.7% vs. 14.2%, p < 0.0001), whereas PCI was associated with higher rates of unplanned revascularization at both 1 month and at 1 year (0.8% vs. 5.2%, p < 0.0001; and 3.8% vs. 16.5%, p < 0.0001, respectively). There were no significant differences between the CABG and PCI groups in 1-month or 1-year mortality (2.5% vs. 2.1%, p = 0.69; and 4.4% vs. 5.7%, p = 0.58, respectively).
Conclusions In this propensity-matched comparison from the ACUITY trial, moderate- and high-risk patients with ACS and multivessel disease treated with PCI rather than CABG had lower rates of peri-procedural stroke, myocardial infarction, major bleeding, and renal injury, with comparable 1-month and 1-year rates of mortality, but more frequently developed recurrent ischemia requiring repeat revascularization procedures during follow-up. (Comparison of Angiomax Versus Heparin in Acute Coronary Syndromes [ACS]; NCT00093158)
- acute coronary syndrome
- coronary artery bypass grafting
- multivessel coronary disease
- percutaneous coronary intervention
An early invasive approach in patients with moderate- and high-risk acute coronary syndromes (ACS) (1) is recommended by the American Heart Association, the American College of Cardiology, and the European Society of Cardiology (2–4). Although more than 1.4 million persons are admitted to hospitals in the U.S. every year with ACS (5), relatively few studies have evaluated the optimal revascularization modality for such patients. Most prior studies have been modest in size and have reported comparable outcomes in patients with ACS treated with either percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG) (6). However, no recent studies have examined the contemporary outcomes of PCI or CABG in these high-risk patients.
Therefore we evaluated the outcomes of patients with moderate- and high-risk ACS and multivessel disease undergoing an early invasive management strategy according to revascularization modality from the large-scale randomized ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy) trial.
Patients and study protocol
The ACUITY trial (7) was a prospective open-label randomized multicenter trial that compared 3 different antithrombotic regimens for patients presenting with moderate- and high-risk ACS and treated with an early invasive management strategy. Clinical follow-up at 450 academic and community-based institutions in 17 countries was performed to determine 1-year outcomes for these patients. The study design and protocol as well as its principal results have been described previously in detail (7). In brief, 13,819 patients presenting in non–ST-segment elevation ACS were randomized to receive unfractionated heparin or enoxaparin with a glycoprotein (GP) IIb/IIIa inhibitor, bivalirudin plus a GP IIb/IIIa inhibitor, or bivalirudin alone. Per protocol, angiography was required in all patients within 72 h of randomization with subsequent triage to PCI, CABG, or medical management according to investigator discretion.
Inclusion criteria were patients ≥18 years of age or older with symptoms of ACS lasting ≥10 min within the preceding 24 h with 1 or more of the following criteria: new ST-segment depression or transient elevation ≥1 mm; elevated troponin I, troponin T, or creatine kinase-myocardial band; known coronary artery disease; or the presence of all 4 other Thrombolysis In Myocardial Infarction risk criteria (7). Exclusion criteria included acute ST-segment elevation myocardial infarction (MI) or shock; bleeding diathesis or major bleeding within 2 weeks; thrombocytopenia; creatinine clearance ≤30 ml/min; or recent administration of abciximab, warfarin, fondaparinux, fibrinolytic agents, bivalirudin, or >1 dose of low-molecular-weight heparin. The study was approved by institutional review boards or ethics committees at each center, and all patients provided written informed consent.
Aspirin was administered daily during the hospital stay (300 to 325 mg orally or 250 to 500 mg IV) and daily indefinitely in all patients after hospital discharge (≥75 mg). Although timing and initial dosing of clopidogrel were left to the discretion of the investigator, a loading dose of 300 mg or more was required within 2 h after PCI in all cases. A 5-day clopidogrel washout period was recommended before surgery for patients who received clopidogrel and in whom CABG was subsequently planned. Clopidogrel (75 mg daily) was recommended for 1 year in all patients undergoing PCI and at operator discretion in patients undergoing CABG. All patients were anticoagulated during CABG with unfractionated heparin, with dosing per standard institutional practice.
End points, definitions, and statistical analyses
Pre-specified major adverse cardiovascular events (MACE) were assessed at 1 month (end of the follow-up window 35 days) and 1 year (end of the follow-up window 395 days) after the procedure, consisting of all-cause death, MI, or unplanned revascularization for recurrent ischemia, and were adjudicated by a blinded Clinical Events Committee. An MI after PCI was diagnosed if any creatine phosphokinase-myocardial band (CPK-MB) was increased to ≥3× normal within 24 h after the procedure or >1× normal with the occurrence of new, significant (≥0.04 s) Q waves in ≥2 contiguous electrocardiographic leads. An MI after CABG was diagnosed if any CPK-MB was increased to ≥10× normal within 24 h after the procedure or to ≥5× normal with the occurrence of new, significant Q waves in ≥2 contiguous electrocardiographic leads. Criteria were adjusted if the CPK-MB was not normal before the procedure (7). Major bleeding was defined as intracranial or intraocular bleeding; access site hemorrhage requiring intervention; ≥5-cm diameter hematoma; reduction in hemoglobin of ≥4 g/dl without or ≥3 g/dl with an overt bleeding source; reoperation for bleeding; or blood product transfusion. The detailed definitions for the other components of MACE have been previously detailed (7).
For the present study we compared the outcomes of patients with multivessel coronary artery disease involving the left anterior descending (LAD) artery who underwent PCI or CABG. The ACUITY core laboratory performed quantitative coronary angiography (QCA) on all lesions in the coronary tree from 6,921 patients in a formal angiographic substudy (8). A diseased epicardial vessel was defined as a major vessel (LAD, left circumflex, or right coronary artery, including its branches) containing 1 or more lesions with a diameter stenosis by QCA of ≥30% (roughly equivalent to a diameter stenosis of ≥50% by visual estimation). In patients without QCA, a diseased epicardial coronary artery was defined by the investigator as a major coronary vessel containing 1 or more lesions with a visually estimated diameter stenosis of ≥50%. Multivessel disease was then defined as 2 or more diseased epicardial coronary arteries.
We further performed—due to substantial differences between the groups (Table 1)—a propensity score analysis that was matched for the following patient characteristics: age, sex, diabetes, hypertension, hyperlipidemia, previous MI, previous percutaneous transluminal coronary angioplasty, previous CABG, renal insufficiency, elevated cardiac biomarkers, ST-segment deviation ≥1 mm, Thrombolysis In Myocardial Infarction risk score, weight, left main disease, number of diseased vessels, and left ventricular ejection fraction. The probability (propensity score) that a patient would receive a PCI or undergo CABG, according to the pre-procedural variables, was determined with a saturated logistic regression model to take into account the different patient profiles. Patients from the PCI group were then matched with patients from the CABG group according to their propensity score, resulting in 528 pairs of propensity-matched patients (n = 1,056 total) in whom early and late outcomes were determined. The C-statistic for the logistic regression model that was used to calculate the propensity score matching for the 2 groups was 0.846.
Categorical variables were compared by chi-square or Fisher exact test. Continuous variables were compared by the nonparametric Wilcoxon rank sum test. Time-to-event data are displayed with Kaplan-Meier methodology and were compared with the log rank test.
Among 13,819 moderate- and high-risk ACS patients enrolled in the ACUITY trial between August 2003 and December 2005 who underwent early angiographic assessment, a total of 7,789 (56%) were managed by PCI and 1,539 (11%) were managed by CABG. Among these, 5,627 patients had multivessel coronary artery disease involving the LAD territory, with 4,412 and 1,215 patients treated with PCI and CABG, respectively. Baseline and angiographic characteristics of these patients varied significantly between the groups and are displayed in Table 1. Specifically, patients undergoing CABG compared with PCI were more often female and more likely to have diabetes, cardiac biomarker elevation, or ST-segment deviation, more extensive coronary artery disease (including left main disease), and a lower left ventricular ejection fraction. In contrast, patients undergoing PCI were heavier and more likely to have hypertension; hyperlipidemia; and prior MI, PCI, or CABG.
Results before propensity matching
Early (1-month) and late (1-year) unadjusted clinical outcomes of the study patients are shown in Table 2. Patients undergoing CABG compared with PCI had significantly higher unadjusted early rates of mortality (3.4% vs. 1.4%, p < 0.0001), stroke, MI, major bleeding, and acute renal injury and higher late (1 year) rates of death (5.7% vs. 4.1%, p = 0.002). Although repeat revascularization for ischemia was less common with CABG at both time periods by multivariable analysis without propensity score matching, 1-month and 1-year mortality were not significantly different according to revascularization modality (odds ratio [OR]: 1.52, 95% confidence interval [CI]: 0.91 to 2.53, p = 0.11; OR: 1.01, 95% CI: 0.68 to 1.50, p = 0.96 for CABG vs. PCI, at 1-month and 1-year, respectively). Coronary artery bypass grafting was, however, associated with a significant increase in the rate of early and late MI (OR: 3.01, 95% CI: 2.31, 3.93, p < 0.0001; OR: 2.06, 95% CI: 1.60 to 2.63, p < 0.0001, respectively). Conversely, CABG was associated with lower 1-month and 1-year rates of unplanned repeat revascularization for ischemia (OR: 0.38, 95% CI: 0.22 to 0.65, p = 0.0005; OR: 0.48, 95% CI: 0.35 to 0.65, p < 0.0001, respectively).
Results after propensity matching
The baseline clinical and angiographic characteristics of the 2 propensity-matched groups (528 pairs, n = 1,056 total) were balanced in all measured characteristics (Table 3). Among PCI-treated patients, stents were implanted in 92.8% of all patients, with drug-eluting stents (DES) used in 61.8% of stented patients. Mean number of vessels and lesions treated in the PCI group were 1.36 ± 0.59 and 1.83 ± 1.13, respectively. In the CABG group, the mean number of grafts/patient was 3.37 ± 1.07, and 13.1% of the patients were operated without use of cardiopulmonary bypass (off-pump CABG).
The mean duration of hospital stay was 13.0 ± 18.1 days for the CABG group versus 5.1 ± 6.8 days for the PCI group (p < 0.0001). This difference derived from a longer time both from admission to revascularization (5.6 ± 17.0 days vs. 1.2 ± 1.6 days, p < 0.0001) as well as from revascularization to hospital discharge (8.2 ± 7.6 days vs. 4.0 ± 6.6 days, p < 0.0001) in the CABG compared with the PCI group.
The early and late clinical outcomes of the propensity-matched groups of patients are shown in Table 4. At 1-month there was no significant difference in all-cause mortality (2.5% for CABG vs. 2.1% for PCI, p = 0.69), cardiac mortality, or MACE in patients managed with CABG versus PCI. The rates of MI, stroke, major bleeding, and acute renal injury at 1-month were higher in the CABG group; however, the rate of unplanned repeat revascularization for ischemia was higher in the PCI group. At 1 year the propensity-adjusted rates of total mortality (4.4% for CABG vs. 5.7% for PCI, p = 0.58) as well as for cardiac mortality did not differ between the 2 groups (Fig. 1). The 1-year rate of MI was no longer significantly different between the 2 groups. The rate of unplanned repeat revascularization remained significantly higher in the PCI group at 1 year, however. As a result of the increased rate of repeat revascularization, a strong trend was present for a greater rate of MACE at 1 year with PCI compared with CABG. There were no additional strokes after 30 days in the present study in either propensity-matched group. The outcomes of the multivariable analysis and propensity score analysis are provided in Figures 2 and 3, respectively.
The major findings of the present study—in which propensity score modeling from a large, contemporary, multicenter, prospective randomized trial was used to assess the comparative outcomes of PCI and CABG in patients with ACS and multivessel disease undergoing an early invasive strategy—are that selection of either revascularization procedure may be performed with comparable rates of survival at both 1 month and 1 year. However, the safety and efficacy profiles of PCI and CABG vary substantially, with PCI associated with lower rates of peri-procedural stroke, MI, major bleeding, and acute renal injury, whereas CABG is associated with greater freedom from both early and late recurrent ischemia requiring repeat revascularization procedures.
Selection of the optimal revascularization modality for patients with multivessel disease continues to be controversial. This is especially true in patients with ACS, because most prior comparative studies of CABG and PCI have enrolled mostly patients with stable ischemic heart disease (6). The safety and utility of CABG in patients with multivessel disease and ACS are well-established (9–11). Recent studies have also demonstrated that PCI may be safely performed in patients with multivessel disease with outcomes comparable to those expected in patients with single vessel disease, even in the setting of ACS (12–14). By reducing restenosis, DES have further improved the rates of event-free survival in patients with ACS undergoing PCI. In the TAXUS IV-SR (Treatment of De Novo Coronary Disease Using a Single Paclitaxel-Eluting Stent) trial, among 450 patients presenting with ACS, treatment with paclitaxel-eluting stents resulted in a 1-year target vessel revascularization rate of 3.9% compared with 16.0% in those patients treated with bare-metal stents (BMS) (15). A recent report from the Massachusetts State database in matched ACS patients undergoing PCI with DES compared with BMS demonstrated significantly lower 2-year cumulative rates of death, MI, or repeat revascularization procedures for patients treated with DES (16).
Limited randomized trial data (none of which are contemporary) are available to guide selection of the appropriate revascularization strategy for high-risk ACS patients with multivessel disease. A pooled analysis from 10 randomized trials (6) reported no significant difference in mortality among 2,653 patients with multivessel disease and ACS treated by CABG versus PCI (with balloon angioplasty or BMS only) after a median follow-up of 5.9 years (9.6% vs. 11.1%, respectively; hazard ratio: 0.95, 95% CI: 0.80 to 1.12).
None of the aforementioned trials used DES, let alone the routine use of contemporary adjunct pharmacologic therapies recommended in patients with ACS such as thienopyridines, GP IIb/IIIa inhibitors, or direct thrombin inhibitors. Surgical therapies and outcomes have also continued to improve. Recently, the SYNTAX (Synergy Between PCI With Taxus and Cardiac Surgery) trial randomized 1,800 patients with triple vessel and/or left main disease to paclitaxel-eluting stents and contemporary pharmacotherapy versus CABG (17). With follow-up to 1 year, PCI and CABG resulted in nonsignificantly different rates of death and MI. Fewer PCI patients had stroke (0.6% vs. 2.2%, p = 0.003), whereas despite the use of DES, fewer CABG patients required repeat revascularization (13.5% vs. 5.9%, p < 0.001). However, only 28.5% of the patients in the SYNTAX trial had unstable angina, an underpowered subgroup for which the event rates have not separately been reported.
In this regard, the results of the present study, representing a nonrandomized but propensity score-adjusted comparison of PCI versus CABG in 1,056 patients with multivessel disease and ACS, are for the most part consistent with the SYNTAX trial findings. Coronary artery bypass grafting compared with PCI (with DES used in 70.7% of patients) resulted in a higher incidence of major bleeding and peri-procedural stroke (as seen in most prior studies of CABG vs. PCI) (18), greater renal insufficiency (also consistent with some [19,20] but not all  prior studies), and a lower rate of early and late recurrent ischemia necessitating repeat revascularization procedures, with comparable mortality. The lower 1-month rate of MI observed in the present study with PCI was not present in the SYNTAX trial or most earlier studies, however, and by 1 year this difference was no longer significant.
Several limitations of the present study should be noted. The ACUITY trial was not a randomized trial of revascularization modalities in ACS, and even propensity score matching cannot correct for unmeasured confounders. Important covariates (mainly procedural) such as vessel diameter, lesion length, and morphology were not included in the present analysis. These parameters, which affect restenosis, recurrent angina, and reintervention rates among patients undergoing PCI, are less important in the surgical group (17). Detailed CABG data, including the exact types of grafts used and vessels treated and completeness of revascularization, were incomplete. “Off-pump” CABG (infrequently used in the present series) has been purported to reduce reperfusion injury and “no reflow” in patients with ACS, thereby improving outcomes (22). However, a recent randomized trial reported reduced graft patency and event-free survival with off-pump compared with on-pump CABG (23). The relative rates of MI after PCI versus CABG depend on the specific definitions used. Frequently, different biomarker thresholds are employed for the 2 revascularization modalities, as were pre-specified in the ACUITY trial. Finally the follow-up period of 1 year might not be sufficient to evaluate the long-term effects of PCI versus CABG.
The present large-scale, contemporary study confirms prior reports demonstrating comparable 1-year survival and MI rates with PCI and CABG in patients with multivessel disease and ACS requiring revascularization. This finding reaffirms the legitimacy of both percutaneous intervention and surgical revascularization in these high-risk patients. Beyond survival, offsetting considerations that might affect the choice of PCI versus CABG include peri-procedural complications (particularly stroke, bleeding, renal injury, and a longer initial hospital stay) with CABG versus reduced long-term freedom from recurrent ischemia and repeat revascularization procedures with PCI. However, these conclusions are based on the current data available from randomized and nonrandomized trials. Larger contemporary studies with longer-term follow-up are required to fully characterize the relative safety and efficacy profiles of PCI and CABG in patients with ACS.
Dr. Ben-Gal has received consulting fees from Radbiomed-VGS. Dr. Mehran has received grant support and lecture fees from The Medicines Company. Dr. Moses has received consulting fees from Cordis and Johnson and Johnson. Dr. Mehran has received support from Abbott, AccessClosure, Accumetrics, AstraZeneca, BMS/Sanofi-Aventis, Cardiva, Cordis, Endothelix, Gilead, Guerbet, The Medicines Co., St. Jude, and Therox. Dr. Lansky has received research support from The Medicines Company. Dr. Williams has received consulting fees from Edwards Life Science. Dr. Stone has served on the scientific advisory boards for Boston Scientific and Abbott Vascular, and is a consultant to The Medicines Company. All other authors report that they have no relationships to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome(s)
- bare-metal stent(s)
- coronary artery bypass grafting
- confidence interval
- creatine phosphokinase-myocardial band
- drug-eluting stent(s)
- left anterior descending
- major adverse cardiovascular events
- myocardial infarction
- odds ratio
- percutaneous coronary intervention
- quantitative coronary angiography
- Received June 14, 2010.
- Accepted June 27, 2010.
- American College of Cardiology Foundation
- Braunwald E.,
- Antman E.M.,
- Beasley J.W.,
- et al.
- Braunwald E.,
- Antman E.M.,
- Beasley J.W.,
- et al.
- Bertrand M.E.,
- Simoons M.L.,
- Fox K.A.,
- et al.
- Graves E.J.,
- Kozak L.J.
- Aoki J.,
- Lansky A.J.,
- Mehran R.,
- et al.
- Shishehbor M.H.,
- Lauer M.S.,
- Singh I.M.,
- et al.
- Moses J.W.,
- Mehran R.,
- Nikolsky E.,
- et al.
- Szczech L.A.,
- Best P.J.,
- Crowley E.,
- et al.