Long-Term Outcome of Isolated Coronary Artery Bypass Surgery in Patients With Severe Left Ventricular Dysfunction
Background— Coronary artery bypass grafting (CABG) is indicated in patients with coronary artery disease and impaired ventricular function. However, earlier studies have suggested that prognosis of patients with severe left ventricular dysfunction is extremely poor. We used the APPROACH registry to derive contemporary estimates of prognosis associated with CABG for this high-risk patient population.
Methods and Results— The study group consisted of 7841 patients who had isolated CABG in the province of Alberta, Canada between 1996 and 2001. Patients with markedly reduced left ventricular function (ejection fraction [EF] <30%, Lo EF, n =430) were compared with those with moderate reduction in ventricular function (EF 30% to 50%, Med EF, n =2581) and those with normal left ventricular function (EF >50%, normal [Nl] EF, n=4830). The operative mortality was higher in the patient group with Lo EF (4.6%) compared with Med EF and Nl EF groups (3.4% and 1.9%, respectively, P<0.001). At 5 years, survival was 77.7% for Lo EF patients compared with 85.5% and 91.2% for Med EF and Nl EF patients, respectively (P<0.001). After controlling for other independent variables, the adjusted hazard ratio for death was 1.98 (95% CI, 1.49 to 2.62) for Lo EF relative to Nl EF. The mortality rate at 1 year was significantly lower for Lo EF patients who underwent CABG than it was for nonrevascularized Lo EF patients (risk-adjusted odds ratio, 0.36; 95% CI, 0.24 to 0.55).
Conclusions— In the modern era of cardiac surgery, CABG can be performed in Lo EF cases with an acceptable perioperative mortality risk. Our estimate of 5-year survival in this high-risk group is better than previously reported in the literature from earlier periods.
Medical therapy in patients with coronary artery disease (CAD) and advanced left ventricular dysfunction (LVD) carries a poor long-term survival.1,2 Coronary artery bypass grafting (CABG) in these patients is associated with improved survival compared with medical treatment. Historically, CABG in patients with LVD has been associated with high perioperative mortality.3,4 However, advances in surgical technique and myocardial protection have lead to improved outcomes, allowing CABG to now be a relatively safe procedure in select patients at high risk.5–7 The number of patients with advanced LVD undergoing CABG has increased in the past few years. Although surgical revascularization may be successful in the short-term, little is known about the long-term follow-up of patients with CAD and severe LVD who undergo “successful” CABG. Death in these patients may be caused by progression of heart failure, ventricular arrhythmias, and recurrent ischemia.
In the past decade, some investigators have evaluated the results of CABG in patients with LVD, but the number of patients studied was limited8–10 and tended to be single surgeon experiences.11–13 These studies indicate that angina class, coronary heart failure class, quality of life, and ventricular function do improve after CABG in patients with LVD, but they also report a very high short-term mortality risk, as well as rather poor longer-term survival after CABG. More data on a larger cohort of patients are needed to derive contemporary estimates of survival in patients with severe LVD undergoing CABG.
In the present study, the long-term results of CABG in 430 patients with severe LVD were studied using a prospectively compiled province-wide database. To our knowledge, this reflects the largest experience of long-term follow-up of patients with LVD undergoing CABG. The survival curves of patients with severe LVD (Lo ejection fraction [EF]) were compared with patients with moderate LVD (Med EF) or normal ventricular function (Nl EF) and hazard ratios derived to identify the prognostic relevance of Lo EF on long-term survival. To further characterize the prognosis associated with Lo EF, we also compared survival in Lo EF patients who underwent CABG with survival in Lo EF patients who were only treated medically.
All data were derived from the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH), a prospective data collection initiative that captures detailed clinical information on all patients undergoing coronary angiography in Alberta, Canada, a province with a population of ≈3.2 million people. Patients are enrolled into the registry at the time of angiography and are followed-up prospectively for outcomes, including subsequent revascularization and death. Details of this database have been previously described.14
At the time of coronary angiography, demographic data are collected, including age and sex, and the presence of the following medical conditions is determined: diabetes, peripheral vascular disease, chronic lung disease, cerebrovascular disease, congestive heart failure, hypertension, hyperlipidemia, liver or gastrointestinal disease, neoplastic disease, and previous CABG surgery, angioplasty, myocardial infarction, or thrombolytic therapy for myocardial infarction. The indication for angiography is categorized as stable angina, myocardial infarction (within 8 weeks of angiography), unstable angina, and other. The results of the coronary angiography, specifically left ventricular ejection fraction and coronary anatomy, are also recorded. In describing coronary anatomy, a vessel is labeled as “diseased” when it contains a lesion causing ≥50% stenosis. Patient survival and time from catheterization until death are ascertained through semiannual linkage to Alberta Vital Statistics records. The APPROACH study protocol was reviewed and approved by the Ethics Review Boards of the University of Calgary and the University of Alberta.
All patients whose EF was measured (at cardiac catheterization or by echocardiography) before surgery and who had isolated CABG in Alberta between January 1, 1996 and December 31, 2001 were eligible. Patients undergoing concomitant cardiac surgical procedures such as valve repair/replacement, automatic implantable cardioverter defibrillator implantation, aneurysmectomy, or other open heart procedures were excluded. Patients with markedly reduced LV function (EF <30%, Lo EF) were compared with those with moderate reduction in LV function (EF 30% to 50%, Med EF) and those with normal LV function (EF >50%, Nl EF). Death occurring within 30 days after CABG was classified as operative death.
Baseline categorical variables were compared among the 3 EF groups (Lo EF versus Med EF versus Nl EF) by means of the χ2 test. Overall survival after CABG was estimated by use of the Kaplan–Meier method. A Cox proportional hazards analysis was used to compare survival up to 7 years after CABG across EF groups while adjusting for other recorded variables, and hazard ratios were reported for survival in Lo EF and Med EF groups relative to the Nl EF group.
χ2 tests were also used to compare the baseline characteristics of CABG-treated versus medically treated Lo EF patients. We then performed logistic regression to compare the mortality rate at 1 year in CABG-treated versus medically treated patients while controlling for any potential differences in baseline characteristics. We extended this multivariable analysis to also control for the propensity of being selected for CABG. This was performed by first using logistic regression to model predictors of undergoing CABG surgery among all patients. The resulting model was used to calculate a probability (ie, propensity) of being selected for CABG that was included as a covariate in a our final multivariable analysis assessing the association between CABG surgery and mortality at 1 year.
The study sample consisted of 7841 patients who underwent isolated CABG, of whom 430 (5.5%) had known Lo EF preoperatively. These patients were compared with 2581 patients who had moderate EF reduction (EF 30% to 50%, and referred to as Med EF throughout this report) and 4830 patients with normal EF (EF >50%). Clinical variables are summarized in Table 1. The difference in age among groups was not significant. Patients with Lo EF had a significantly greater prevalence of adverse clinical risk factors, including peripheral vascular disease, renal insufficiency, diabetes mellitus, congestive heart failure, and previous myocardial infarction. They also had a higher prevalence of multivessel and left main coronary artery disease, and a tendency toward a more acute indication for surgical revascularization, such as recent myocardial infarction.
Operative mortality (ie, mortality within 30 days) in the Lo EF group was 4.4% versus 3.4% and 1.9%, respectively, in the Med EF and Nl EF groups (P<0.001).
Figure 1 shows the Kaplan–Meier survival curves for the 3 different EF groups. At 3 years, Kaplan–Meier estimated survival of Lo EF patients was 84.7% compared with 90.0% and 94.5% for Med EF and Nl EF patients, respectively (P<0.001 from log rank test). At 5 years, Kaplan–Meier estimated survival of LoEF patients was 77.7% compared with 85.5% and 91.2% for Med EF and Nl EF patients, respectively (P<0.001).
The crude (ie, unadjusted) hazard ratio for Lo EF relative to Nl EF was 2.75 (95% CI, 2.15 to 3.51) (Table 2). After controlling for all other known independent variables, the adjusted hazard ratio for Lo EF was 1.95 (95% CI, 1.49 to 2.55) relative to Nl EF.
Comparison of Surgical Revascularization Versus Medical Treatment
Recognizing that Lo EF patients have poorer survival than patients with Med EF or Nl EF, we also considered it relevant to compare survival in Lo EF patients who underwent CABG with survival of other Lo EF patients who did not undergo surgery. The 430 patients with Lo EF who underwent isolated CABG were thus compared with 1739 patients in the database with Lo EF who proceeded to cardiac catheterization but did not undergo revascularization for a combination of possible reasons. Table 3 demonstrates that medically treated patients with Lo EF were more likely to have a clinical diagnosis of congestive heart failure and to have had a previous CABG procedure. The profile of indications for catheterization differed somewhat between the 2 groups, and the extent of coronary artery disease was greater in the CABG group, with more 3-vessel disease and left main disease in surgically treated Lo EF patients.
The unadjusted mortality rates at 1 year in the surgical and medical treatment groups were 8.6% versus 14.3%, respectively (P=0.002). A multivariable logistic regression analysis controlling for differences between groups in the baseline characteristics shown in Table 3 reveals an odds ratio of 0.36 (95% CI, 0.24 to 0.55) for CABG-treated patients relative to medically treated patients. When we additionally controlled for the propensity to be selected for CABG surgery, the adjusted odds ratio for CABG relative to medical therapy was 0.26 (95% CI, 0.17 to 0.41).
The present study, encompassing the largest published follow-up of CABG and low EF, shows that patients with this syndrome have higher operative and postoperative long-term mortality rates than patients with milder or no decrease in EF. However, in Lo EF patients, the operative mortality rate, although increased, is not overwhelmingly so (4.4% versus 1.9%). An operative mortality rate of 4.4% is consistent with recently published data on the topic and is a marked improvement from previous reports indicating a prohibitive operative mortality in the range of 10% to 50%.3 The definition of low EF is somewhat variable across studies. To isolate patients with severe left ventricular dysfunction in this study, low EF was defined as an EF <30% and patients with ejection fractions of 30% to 35% were grouped with the medium EFs. Thus, we can conclude that the 430 study patients comprised a high-risk group of patients given the presence of multiple comorbidities combined with a marked reduction in ventricular function.
Contrary to other published series, this study does not reflect a single surgeon experience. The operative mortality of 4.4% and 5-year estimated survival of 77.7% was achieved on a regional basis, involving many different surgeons, 3 surgical centers, and uncontrolled techniques of myocardial protection and surgery (ie, no imposed care protocol). Therefore, we conclude that an operative mortality rate of <5% with 5-year survival close to 80% is likely to be achievable in other centers in the current era of cardiac surgery.
Late survival, although reduced when compared with patients with milder or no decrease in EF, remains high, with estimated 3- and 5-year survival of 84.7% and 77.7%, respectively, in patients with preoperative low EF undergoing isolated CABG. Thus, it is reassuring to know that once these sick patients are through the perioperative course, they continue on a survival path indicating a reasonably favorable prognosis. When compared with normal EF, however, patients with low EF had an adjusted hazard ratio of 1.95, indicating that EF is still a significant independent risk factor for long-term outcome after CABG. Death in these patients may be caused by progression of heart failure, ventricular arrhythmias, and recurrent ischemia.
Newer medical and surgical therapies are being targeted toward this high-risk subgroup of patients. The role of aneurysmectomy and LV remodeling surgery as an adjunct to CABG in this subgroup of patients is currently being investigated as one arm of a National Institutes of Health-sponsored multicenter study (STICH Trial). Recent evidence from the MADIT II trial15 indicates that patients with previous myocardial infarction and EF <30% derive a survival benefit at 20 months with prophylactic AICD implantation. Cell-based myocardial regeneration therapies (ie, stem cell transplantation16 and/or gene therapy17) may also assume a role in the management of such patients. Future studies will determine whether these adjunctive measures will lead to even further improvement in long-term survival in patients with coronary artery disease and severe left ventricular dysfunction. The results from this study serve as a good reference point to compare the results of adjuvant treatments versus isolated coronary artery surgery.
The survival comparison of Lo EF patients treated with CABG relative to those treated medically is also notable. It demonstrates that although the prognosis associated with Lo EF is generally poor (relative to patients with better EF), undergoing CABG is associated with better outcomes than those seen in Lo EF patients who are medically managed. This latter finding underlines the point that Lo EF alone should not be used as a criterion to turn patients away from CABG surgery in anticipation of poor surgical prognosis that earlier studies might have suggested. Our combined findings of acceptable short-term and long-term prognosis in Lo EF patients undergoing CABG and poor prognosis in Lo EF patients treated medically argue for an aggressive surgical approach to CAD patients with Lo EF.
Preoperative EF was not standardized for this study but rather was measured by different techniques in the context of usual care (ie, echocardiography, calculated planimetry on catheterization, and visual estimation at time of catheterization). EF was not available for every patient preoperatively; 280 patients were classified as “missing EF” and 470 patients were classified as “EF not done due to instability.” These 2 groups were left out of the eventual analysis, because we did not want to incorrectly classify cases with, for a variety of reasons, unmeasured EF values. A second limitation is that we do not have information on left ventricular volumes that can provide prognostic information over and above EF values. A third limitation is that we have focused on mortality and late survival as the outcomes of interest. Other outcomes such as resource use (eg, costs, readmissions for cardiovascular problems), recurrent nonfatal events, functional status, and quality of life also warrant future study. A final caveat is that the comparison of the CABG-treated and medically treated Lo EF patients needs to be viewed with some caution. Our multivariable risk adjustment and propensity analyses provide partial adjustment for baseline differences and selection factors that would otherwise preclude a meaningful comparison between groups. Even with such analyses, however, the CABG versus medical group comparison needs to be interpreted carefully. We certainly do not present this analysis to imply that all of the medically treated patients should have had a CABG procedure. The analysis does, however, demonstrate that the prognosis associated with Lo EF in a contemporary CABG surgery population is quite favorable relative to contemporary patients with Lo EF who are treated medically rather than with CABG (for a variety of reasons).
Despite these limitations and caveats, this study provides important insights into the prognosis of patients with low EF who undergo CABG. Patients with severe CAD and markedly reduced EF represent a high-risk group that can undergo isolated CABG safely. Long-term survival is statistically significantly reduced when compared with patients with milder or no decrease in EF, but 5-year survival is still very good relative to previous published reports on prognosis in such patients. Advances in adjuvant medical and surgical therapies may further improve survival in this high-risk group of patients.
Dr. Ghali is supported by a Government of Canada Research Chair and by a Health Scholar Award from the Alberta Heritage Foundation for Medical Research. Dr. Norris was support by fellowships from the Canadian Cardiovascular Outcomes Research Team (CCORT) and Tomorrow’s Outcome Researchers in Cardiovascular Health (TORCH) during the preparation of this manuscript. The APPROACH initiative was originally funded in 1995 by a grant from the Weston Foundation. The ongoing operation of the project is supported by Merck Frosst Canada Inc, Monsanto Canada Inc, Searle, Eli Lilly Canada, Guidant Corporation, Boston Scientific Ltd, Hoffmann–La Roche Ltd, and Johnson and Johnson Inc–Cordis. We appreciate the assistance of the Calgary Regional Health Authority and the Capital Health Authority in supporting online data entry by cardiac catheterization laboratory personnel. Members of the APPROACH project Clinical Steering Committee are, in Edmonton, Dr. Stephen Archer, Dr. Arvind Koshal, Dr. William Hui, Dr. Ross Tsuyuki, and Dr. Colleen Norris; and, in Calgary, Dr. L. Brent Mitchell, Dr. Andrew Maitland, Dr. Michael Curtis, Dr. Merril L. Knudtson, Dr. William A. Ghali, and Ms. Diane Galbraith.
APPROACH is supported by donations from a variety of industrial sponsors (see Acknowledgements). W.A.G. is supported by supported by a Government of Canada Research Chair and by a Health Scholar Award from the Alberta Heritage Foundation for Medical Research.
Bruschke AVG, Proudfit WL, Sones FM. Progress study of 590 consecutive nonsurgical cases of coronary disease followed 5–9 years. Circulation. 1973; 47: 1144–1163.
Scott SM, Deupree RH, Sharma GVRK, et al. VA study of unstable angina. 10-year results show duration of surgical advantage for patients with impaired ejection fraction. Circulation. 1994; 90 (Suppl 2): 120–3.
Alderman EL, Fisher LD, Litwan P, et al. Results of coronary artery surgery in patients with poor left ventricular function (CASS). Circulation. 1983; 68: 785–795.
Mickleborough LL, Mauyama H, Takagi Y et al. Result of revascularization in patients with severe left ventricular dysfunction. Circulation. 1995; 92 (suppl II): II73–II79.
Heilmann CA, Attmann T, Thiem A, et al. Gene therapy in cardiac surgery: intramyocardial injection of naked plasmid DNA for chronic myocardial ischemia. Eur J Cardiothorac Surg. 2003; 24: 785–793.