Author + information
- Received December 18, 2000
- Revision received March 27, 2001
- Accepted April 6, 2001
- Published online July 1, 2001.
- Douglass A Morrison, MD, FACC*,
- Gulshan Sethi, MD, FACC,
- Jerome Sacks, PhD,
- William Henderson, PhD,
- Frederick Grover, MD, FACC,
- Steven Sedlis, MD, FACC,
- Rick Esposito, MD,
- Kodangudi Ramanathan, MD, FACC,
- Darryl Weiman, MD,
- Jorge Saucedo, MD,
- Tamim Antakli, MD,
- Venki Paramesh, MD,
- Stuart Pett, MD,
- Sarah Vernon, MD, FACC,
- Vladimir Birjiniuk, MD,
- Frederick Welt, MD,
- Mitchell Krucoff, MD, FACC,
- Walter Wolfe, MD, FACC,
- John C Lucke, MD,
- Sundeep Mediratta, MD,
- David Booth, MD, FACC,
- Charles Barbiere, CCRN,
- Daniel Lewis, MD, FACC,
- for the Investigators of the Department of Veterans Affairs Cooperative Study 385 the Angina With Extremely Serious Operative Mortality Evaluation (AWESOME)
- ↵*Reprint requests and correspondence:
Dr. Douglass A. Morrison, Department of Medicine and Radiology, University of Arizona SAVAHCS, 3601 South Sixth Avenue, Tucson, Arizona 85723
Percutaneous coronary intervention (PCI) and coronary artery bypass graft surgery (CABG) are being applied to high-risk populations, but previous randomized trials comparing revascularization methods have excluded a number of important high-risk groups.
This five-year, multicenter, randomized clinical trial was designed to compare long-term survival among patients with medically refractory myocardial ischemia and a high risk of adverse outcomes assigned to either a CABG or a PCI strategy, which could include stents.
Patients from 16 Veterans Affairs Medical Centers were screened to identify myocardial ischemia refractory to medical management and the presence of one or more risk factors for adverse outcome with CABG, including prior open-heart surgery, age >70 years, left ventricular ejection fraction <0.35, myocardial infarction within seven days or intraaortic balloon pump required. Clinically eligible patients (n = 2,431) underwent coronary angiography; 781 were angiographically acceptable; 454 (58% of eligible) patients consented to random assignment between CABG and PCI.
A total of 232 patients was randomized to CABG and 222 to PCI. The 30-day survivals for CABG and PCI were 95% and 97%, respectively. Survival rates for CABG and PCI were 90% versus 94% at six months and 79% versus 80% at 36 months (log-rank test, p = 0.46).
Percutaneous coronary intervention is an alternative to CABG for patients with medically refractory myocardial ischemia and a high risk of adverse outcomes with CABG.
Coronary artery bypass graft (CABG) surgery is associated with relief of symptoms and enhanced survival relative to medical therapy in selected patients with coronary artery disease (CAD) (1–10). Percutaneous coronary intervention ([PCI] can include stents or atherectomy, in addition to balloon angioplasty) is associated with relief of symptoms relative to medical therapy in selected patients (11–14). Direct comparison of PCI versus CABG suggests that long-term survival is comparable, albeit with more frequent repeat revascularization, among patients randomized to PCI (15–24). In some cases, better symptom relief has been observed among patients randomly assigned to CABG (15–24).
All previous randomized comparisons of PCI with CABG have excluded patients with prior heart surgery, on-going or very recent myocardial infarction (MI) and severe left ventricular dysfunction, defined as left ventricular ejection fraction (LVEF) <0.35 (15–24). No previous revascularization trial has specified medically refractory ischemia as an inclusion criterion (1–24). The nine reported randomized comparisons of PCI with CABG did not include the use of either stents or glycoprotein IIb/IIIa receptor-blocker drugs, both of which are currently commonly used with PCI (15–24).
Although it is difficult to define simple criteria for medically refractory ischemia that apply to diverse populations, it is clear that there are patients for whom medical therapy fails to control their symptoms (25). Similarly, there are patients with medically refractory ischemia who are turned down by surgeons for CABG, based on the perception of prohibitively high operative risk (26). In 1988, we began offering PCI, without surgical standby, to select patients with medically refractory ischemia who had been refused CABG (27). As experience with “salvage angioplasty” increased, the question was raised whether some patients at increased risk of adverse outcomes with CABG might be better served with PCI, even though CABG remained an option (28).
We proposed a clinical trial to compare the long-term survival of patients with medically refractory unstable ischemia and a high risk of adverse outcome with CABG, randomly allocated between a CABG strategy and a PCI strategy (26,28). From the Veterans Affairs (VA) Continuous Improvement in Cardiac Surgery Program database, we identified five risk factors associated with increased 30-day postoperative mortality. The risk factors were: 1) prior heart surgery; 2) age >70 years; 3) LVEF <0.35; 4) intraaortic balloon pump (IABP) before surgery; and 5) MI <7 days before CABG (29,30). This paper presents the three-year survival results of the randomized clinical trial.
The Angina With Extremely Serious Operative Mortality Evaluation (AWESOME) trial began enrollment in February 1995 and completed enrollment on March 31, 2000. Over the five-year period, 16 VA Medical Centers participated, with 11 sites enrolling throughout the entire period and five sites enrolling for periods of two to three years. Methodological issues have been presented in a design manuscript published by the Society for Clinical Trials (28).
Patient screening and accrual
Patient screening took place in a five-step process (28). The first three steps consisted of identifying clinically eligible patients who had medically refractory myocardial ischemia and one or more of the five risk factors. Myocardial ischemia was defined as one or more of the following: 1) rest angina with reversible electrocardiographic changes; 2) rest angina in a patient with prior MI or angiographically confirmed CAD; 3) recurrent rest angina; 4) rest angina within seven days of an acute MI; or 5) unstable angina that had been stabilized medically, with a subsequent positive provocative test for ischemia. Medically refractory ischemia was defined as ischemia that persisted despite a regimen that included aspirin or intravenous heparin and at least one of the following: heart rate <60 beats/min and systolic blood pressure <120 mm Hg, enough antianginal drug so that resting heart rate remains at <70 beats/min or systolic blood pressure <150 mm Hg or IABP required to stabilize or significant contraindication to both beta-adrenergic blocking agents and calcium channel blocking agents (25,26). The definition of medically refractory ischemia had to be broad enough to safely include very elderly patients, patients with severe degrees of ventricular dysfunction and patients with advanced degrees of pulmonary, renal or hepatic comorbidity (25,26). To be considered a high risk, the patient had to have one or more of the five risk factors associated with increased 30-day operative mortality (29,30).
Specific exclusions from the randomized trial included any of the following: single-vessel circumflex disease, unprotected (ungrafted) >50% left main stenosis, no graftable or dilatable vessels or comorbidity likely to limit the patient’s life to a greater extent than his (her) coronary disease or angioplasty within six months.
Patients who met the three clinical eligibility criteria (myocardial ischemia, medically refractory, high-risk of adverse outcomes with CABG) underwent coronary angiography, which was reviewed by both a surgeon and an interventional cardiologist. Clinically eligible patients who had coronary anatomy deemed acceptable for random assignment between revascularization strategies were approached for informed consent to participate in the trial.
A total of 22,662 patients were screened, and 20,231 were excluded because they did not meet all three clinical requirements. Among the screened patients, 7,278 did not meet the myocardial ischemia criteria; 5,783 did not meet the medically refractory criteria, and 10,030 did not have at least one of the high-risk factors. The sum of these three exclusion groups is greater than the 22,662 screened because of overlap between exclusion groups. A total of 2,431 patients met all three criteria and was clinically eligible. After coronary angiography had been reviewed by both interventional cardiologist and surgeon, a total of 781 (32%) was angiographically acceptable to both operators as candidates for random allocation of revascularization method. These patients were approached for informed consent, and 454 (58%) consented to a randomized choice of revascularization.
Randomization was under the direction of the study biostatistician at the Hines VA Cooperative Studies Program Coordinating Center. Patients were stratified by hospital, age (<70, >70 years) and prior heart surgery (yes, no). The co-chairmen and local site investigators were blinded from the study results throughout the course of the study.
The primary outcome of the trial was survival. Death records were completed on each patient who died during the course of the trial by the research nurse coordinators. Patient records were also matched with VA death records recorded in the VA Beneficiary Identification and Record Locator Subsystem database to check survival information. Unstable angina, repeat hospitalization, repeat catheterization, repeat CABG and repeat PCI were obtained from the follow-up visits and rechecked against the Patient Treatment File, an independent hospital discharge database maintained by the Department of the VA. As an outcome, unstable angina was defined as angina that is either new onset, progressive or occurring at rest (25–28).
The baseline comparability of the cohorts randomized to PCI and CABG is assessed by chi-square tests for proportions and Student ttests for continuous data. Kaplan-Meier survival estimates were used to estimate the primary end point, and the survival estimates were used to generate survival plots. The log-rank test provided a global statistical significance test for the CABG and PCI survival plots. The secondary end points of combined survival and freedom from unstable angina and combined survival and freedom from unstable angina and repeat revascularization were evaluated by comparing Kaplan-Meier plots and log-rank tests. All analyses are based on cohorts specified by the intention-to-treat principle.
Sample size and power calculations were discussed in the methods manuscript (28). All sample size calculations were two-sided. The final sample size has a power of 80% to detect a 9% survival difference at 36 months between CABG and PCI (82% vs. 73%) based on a two-sided test with a type I error of 5%.
Baseline comparability of CABG and PCI cohorts
Table 1presents selected baseline clinical and angiographic variables for the CABG and PCI groups. Coronary artery bypass graft surgery and PCI profiles are similar for all baseline variables except for angiographic thrombus. This single statistically significant difference among all the baseline comparisons is well within what would be expected by sampling variation.
CABG and PCI methods
The numbers of distal anastamoses performed were relatively stable at 2.9 across the five years of the study. The use of arterial conduits increased somewhat with average use over the five years. Left internal mammary use increased from 57% in 1995 to 78% in 1999/2000 with an overall rate of 70%. Right internal mammary use increased from 1.4% in 1995 to 4.8% in 1999/2000 with an overall rate of 3.4%, and radial artery use increased from 1.4% in 1995 to 4.0% in 1999/2000 with an overall rate of 2.8%.
Use of stents increased from 26% of cases in 1995 to 88% of cases in 1999/2000 with an overall rate of 54%. Similarly, use of glycoprotein IIb/IIIa receptor blocking agents as adjuncts to PCI increased from 1% in 1995 to 52% in 1999/2000 with an overall rate for the study of 11%. The use of intraaortic balloon counter-pulsation decreased from 21% of PCI cases in 1995 to 10% in 1999/2000. There was no use of cardiopulmonary support or heart-lung bypass as an adjunct for PCI throughout the course of the study. By 1998 there was no further use of directional atherectomy.
Table 2compares the CABG and PCI short-term outcomes. Of the 232 patients randomized to receive CABG, 227 (98%) were revascularized; 215 (93%) received only CABG; 12 (5%) received PCI (two in addition to CABG), and 5 (2%) were not revascularized. The in-hospital mortality of the CABG cohort was 4% (n = 8), and the 30-day mortality was 5%. Stroke, renal failure requiring dialysis and mediastinitis were each seen in three patients (1%). Of the 222 patients randomized to PCI, 221 (99.5%) were revascularized; 213 (96%) received only PCI; 8 (3.5%) received CABG (seven after PCI), and one patient (0.5%) died before a scheduled PCI. In-hospital mortality in the PCI cohort was 1% (n = 2), and the 30-day mortality was 3%. Stroke was 1% (n = 2), and renal failure requiring dialysis was 2% (n = 4) in the PCI cohort. Within 30 days of revascularization, none of the CABG patients underwent repeat CABG, but four patients (2%) underwent PCI. Within 30 days of revascularization, five (2%) of the PCI patients underwent repeat PCI, and nine (4%) underwent CABG.
Table 3compares the 36-month survival, survival free of unstable angina and survival free of unstable angina and repeat revascularization for the cohorts randomly allocated between CABG and PCI.
Figure 1presents the five-year Kaplan-Meier survival curves for the CABG and PCI cohorts. The log-rank test for global differences between the two survival curves is not statistically significant (p = 0.46). The CABG and PCI number of patients (N) and the percentage surviving for each time period are shown.
Figure 2presents the five-year Kaplan-Meier survival curves for the combined end point of survival free of unstable angina. Coronary artery bypass graft surgery survival free of unstable angina is generally greater than the corresponding PCI survival rate, but the PCI rate is approximately 90% of the CABG rate. The log-rank test for global differences between the two curves is not statistically significant (p = 0.16).
The CABG and PCI number (N) of patients and the percentage surviving free of unstable angina for each time period are shown.
Figure 3presents the five-year Kaplan-Meier curves for the combined end points of survival free of unstable angina or repeat revascularization. Coronary artery bypass graft surgery survival free of unstable angina and repeat revascularizations is generally greater than the corresponding PCI rates, and the log-rank test for global differences between the two curves is statistically significant (p = 0.001). The CABG and PCI number (N) of patients and the percentage surviving free of unstable angina or repeat revascularizations for each time period are shown.
The AWESOME study was designed to compare long-term survival between PCI and CABG in patients with medically refractory ischemia and increased risk of adverse outcomes with CABG (26,28). The primary result of this trial is comparable survival of patients randomly allocated between CABG and PCI. The additional finding that PCI survival free of unstable angina was within 90% of the CABG survival free of unstable angina also supports the concept that PCI is an alternative for these patients. The finding of more frequent repeat revascularization with a PCI strategy is consistent with all previous CABG versus PCI comparisons and reflects the differences in restenosis and completeness of revascularization, which are implicit in the two methods (15–24).
AWESOME patient population compared with previous revascularization trials
The contribution this trial makes to the comparison of CABG and PCI derives primarily from its focus on high-risk patients (31). This is the only trial, to our knowledge, that compares CABG and PCI in patients with myocardial ischemia recruited because they were both medically refractory and had risk factors for adverse outcomes with CABG (15–24,26,28,31). The mean age of AWESOME randomized trial patients of 67 years is at least five years older than the mean age of patients randomized in the Emory Angioplasty Surgery Trial (EAST) (62 years), Bypass Angioplasty Revascularization Investigation (BARI) (61 years) and the other PCI versus CABG trials (15–22). Similarly, the mean LVEF of AWESOME patients (0.45) is >0.10 lower than the mean values for any previous randomized trial comparing CABG with PCI (15–22). Prior CABG (31% of AWESOME patients) has been an exclusion criterion from all previous trials (15–22). Most previous trials have excluded patients who had an MI within seven days (33% of AWESOME patients) (15–22).
Medically refractory ischemia and processes of care
For this trial we attempted to push categories of medicines toward physiologic heart rate and blood pressure targets (25,26). The problem with a “one-size-fits-all” definition is that there is a different risk of lowering heart rate and blood pressure in patients with normal LVEF, versus severely reduced LVEF, and in patients who are >80 years as opposed to <60 years old (25). Nonetheless, to our knowledge, this is the first revascularization trial to attempt to limit enrollment to patients who had failed any predefined medical regimen. Largely as a result of the attempt to define medically refractory angina, a large proportion of AWESOME patients were receiving aspirin, heparin, nitrates and beta-blockers. Of the patients coming to diagnostic catheterization before enrollment in the AWESOME randomized trial, 88% were receiving aspirin; 62% were receiving heparin; 90% were receiving nitrates; 76% were receiving beta-blockers, and 31% were receiving calcium channel blockers. This is in contrast with aspirin use (45% and 49%), heparin use (31% and 32%), beta-blocker use (50% and 34%) and calcium channel blocker use (76% and 76%) in BARI and EAST, respectively (18,21,22). The data supporting nearly routine use of statin lipid-lowering agents and angiotensin-converting enzyme inhibition were not available when AWESOME was being planned.
Comparison of revascularization processes of care
With regard to surgical processes of care, 70% of AWESOME surgical patients received left internal mammary grafts; 3.2% received right internal mammary grafts, and 2.9% received radial artery grafts. By contrast, BARI and EAST surgical patients received internal mammary grafts in 84% and 86% of cases, respectively (18,21).
With regard to percutaneous interventional processes of care, 54% of AWESOME patients undergoing PCI received stents with an increase from 26% in 1995 to 88% in 1999/2000. Overall, 11% of AWESOME patients undergoing PCI received glycoprotein IIb/IIIa receptor blocking agents with an increase from 1% in 1995 to 51% in 1999/2000. Neither stents nor glycoprotein IIb/IIIa receptor blocking agents were used in either BARI or EAST (18,21).
The most important limitation of this trial is not having a larger sample size from which to derive greater statistical power. Most of this limitation appears to derive from the large proportion of clinically eligible patients for whom the physicians felt one of the two revascularization choices was preferable and, therefore, would not permit random allocation. Physician clinical judgment favored nonrandom allocation for 1,650 patients (68% of the clinically eligible patients). The lack of a core laboratory for central adjudication of electrocardiograms was the major reason for not including myocardial infarction as a secondary outcome measure. The lack of a core laboratory for adjudication of coronary angiograms was the major reason for not investigating relative patency rates. The unavailability of large numbers of female patients is a recognized limitation of VA trials.
Percutaneous coronary intervention is an alternative to CABG for some patients with medically refractory myocardial ischemia and a high risk of adverse outcomes with CABG. Patients with medically refractory myocardial ischemia and one or more of these risk factors have a comparable three-year survival but a higher rate of repeat revascularization with initial PCI.
The authors acknowledge the memory of Dr. Stewart Scott, Principal Investigator of the Veterans Affairs Cooperative Medical versus Surgical Trial in Unstable Angina and the first Surgical Co-principal Investigator of AWESOME and all of the United States Veterans who participated in AWESOME. The authors also thank the medical and surgical housestaff of the 16 University affiliated Veterans Affairs Medical Centers who contributed to the care of all of our AWESOME patients. (Appendix)
Participants in veterans affairs cooperative study #385: awesome
Executive Committee:Douglass Morrison, MD; Stewart Scott, MD (deceased); Gulshan Sethi, MD; Jerome Sacks, PhD; William Henderson, PhD; Steven Sedlis, MD; Rick Esposito, MD; Dan Lewis, MD; Edward Folland, MD; David Talley, MD.
Data Monitoring Board:Spencer King, III, MD (chairman); Cindy Grines, MD; Sidney Levitsky, MD; Bruce Lytle, MD; Daniel Seigel, ScD
Death Review Committee:David Holmes, MD; Hartzell V. Schaff, MD.
Department of Veterans Affairs Headquarters Staff:John R. Feussner, MD, Chief Research and Development Officer; Steven Berkowitz, PhD, Assistant Director, Cooperative Studies Program; Joseph Gough, MA, Program Assistant, Cooperative Studies Program.
Participating Sites:Albuquerque, New Mexico; Ann Arbor, Michigan; Asheville, North Carolina; Dallas, Texas; Denver, Colorado; Durham, North Carolina; Lexington, Kentucky; Little Rock, Arkansas; Memphis, Tennessee; Minneapolis, Minnesota; New York, New York; Portland, Oregon; San Antonio, Texas; Tucson, Arizona; West Los Angeles, California; West Roxbury, Massachusetts.
☆ Supported by the Cooperative Studies Program of the United States Department of Veterans Affairs Research and Development Service. This study was presented in abstract form as a Late Breaking Trial at the Society for Cardiac Angiography and Interventions meeting, May 2001, Houston, Texas.
- Angina With Extremely Serious Operative Mortality Evaluation
- Bypass Angioplasty Revascularization Investigation
- coronary artery disease
- coronary artery bypass graft surgery
- Emory Angioplasty Surgery Trial
- intra-aortic balloon pump
- left ventricular ejection fraction
- myocardial infarction
- percutaneous coronary invervention
- United States Department of Veterans Affairs
- Received December 18, 2000.
- Revision received March 27, 2001.
- Accepted April 6, 2001.
- American College of Cardiology
- Kirklin J.W.,
- Akins C.W.,
- Blackstone E.H.,
- et al.
- Parisi A.F.,
- Khuri S.,
- Deupree R.H.,
- Sharma G.V.R.K.,
- Scott S.M.,
- Luchi R.J.
- Russell R.O.,
- Moraski R.E.,
- Kouchoukos N.,
- et al.
- Veterans Administration Coronary Artery Bypass Surgery Cooperative Study Group
- Coronary Artery Surgery Study (CASS) Principal Investigators and Associates
- Ryan T.J.,
- Bauman W.B.,
- Kennedy J.W.,
- et al.
- Lembo N.J.,
- King S.B. III.
- Rodriguez A.,
- Boullon F.,
- Perez-Balino N.,
- et al.
- BARI Investigators
- Frye RL, King SB, III, Sopko G, Detre KM. A symposium: multi-vessel PTCA versus CABG: baseline data from the Bypass Angioplasty Revascularization Investigation (BARI) and the Emory Angioplasty Surgery Trial (EAST). Am J Cardiol 1995;75:1c–59c.5.
- Morrison D.A.
- Morrison D.A.
- Morrison D.A.,
- Sacks J.,
- Barbiere C.C.,
- Sethi G.