Short-term survival in patients with viability and ischemic left ventricular dysfunction appears improved by revascularization, but no randomized studies have shown a long-term benefit of revascularization in patients with a wide range of viability. Propensity analysis was used as a substitute for randomization in a study comparing the survival of revascularized and medically treated patients with ischemic dysfunction. Dobutamine echocardiography was performed in 274 patients with ischemic left ventricular dysfunction (mean ejection fraction 32%), with 32% having viability in ≥25% of the myocardium. Clinical, angiographic, and echocardiographic characteristics were comparable between treatment groups except for multivessel disease, hyperlipidemia, and the percentage of nonviable myocardium. A propensity score, reflecting the probability of receiving revascularization, was derived for each patient from baseline variables. After stratification by propensity scores, there were no differences between groups. Patients were followed for cardiac death. Revascularization was performed in 130 patients, and 144 were medically treated. There were 114 cardiac deaths (42%) over 4.5 years of follow-up. After propensity score adjustment, survival was better with revascularization (mean survival 5.9 vs 3.3 years, hazard ratio 0.42, 95% confidence interval 027 to 0.65, p <0.0001). Medical and device therapy during follow-up was similar between treatment groups, except that β-blocker use was more common in revascularized patients. After adjustment for β-blocker use and propensity score, survival remained better in revascularized patients (hazard ratio 0.47, 95% confidence interval 0.30 to 0.72, p = 0.0006). In conclusion, revascularization improves long-term survival in patients with ischemic left ventricular dysfunction and a wide range of viability.
Revascularization and medical therapies aimed at the reduction of ischemia and ventricular remodeling are the mainstays of treatment of patients with ischemic left ventricular (LV) dysfunction. Nonrandomized studies evaluating the impact of viability testing have demonstrated that revascularization reduces mortality in patients with viability and does not appear to worsen the prognosis of patients without significant viability. The favorable showing of revascularization in nonrandomized studies has been cautiously interpreted, given that factors influencing the selection of therapy and outcome were not equally distributed between revascularized and medically treated groups. Additionally, drug and device therapies were often not accounted for in these trials, and follow-up was generally limited to 2 to 3 years. As a result, several randomized trials have been initiated. The Surgical Treatment for Ischemic Heart Failure Trial (STICH) is comparing the outcomes of medical therapy, coronary artery bypass grafting (CABG), and CABG with ventricular restoration. The Heart Failure Revascularization Trial (HEART) is comparing survival in medically treated and revascularized patients with viability. The results of these trials have yet to be reported. With the hypothesis that revascularization may benefit a broad spectrum of patients with ischemic dysfunction, the purpose of this study was to compare long-term survival in medically treated and revascularized patients without limiting subjects to those with significant viability. Propensity score analysis was used to adjust for differences between treatment groups.
Methods
The study was approved by the institutional review board of Indiana University. The study group was selected from 1,370 patients with LV dysfunction (wall motion abnormalities at rest in ≥25% of the left ventricle and reduced ejection fractions) who underwent dobutamine echocardiography from June 1989 to 2003. Patients who had only low-dose examinations (maximum ≤10 μg/kg/min) and those with recent transmural infarctions were excluded. Two hundred seventy-four patients who had significant coronary artery disease (CAD) documented by angiography within 4 months of dobutamine echocardiography were included.
Dobutamine was administered beginning at a dose of 5 μg/kg/min for 3 minutes, followed by 10 μg/kg/min for 3 minutes. Thereafter, the dose was increased in 10 μg/kg/min increments every 2 to 3 minutes to a peak dose of 50 μg/kg/min. Echocardiograms in the standard parasternal and apical views were digitally stored at rest, at the end of the 5 and 10 μg/kg/min stages, and at peak dose. All echocardiograms were analyzed by an investigator blinded to the clinical and outcome data. Wall motion and thickening were assessed in 16 segments using a previously described scoring system. Dysfunctional segments exhibiting improvement in wall motion or thickening by 1 grade from rest to 5 or 10 μg/kg/min of dobutamine were considered viable. Segments that were akinetic or dyskinetic at rest and remained without improvement at 5 and 10 μg/kg/min were defined as nonviable. Segments with worsening of wall motion at any stage of the examination were defined as ischemic (except akinesia to dyskinesia). Wall motion scores (WMS) at rest, 10 μg/kg/min (designated as low dose), and peak dose were calculated.
The extent of viable myocardium was assessed on the basis of measures used in previous studies. The change in WMS from rest to 10 μg/kg/min and the percentage of viable and nonviable segments and the presence or absence of viable myocardium (contractile reserve) in ≥25% of the left ventricle were determined. Subjects were divided into 3 groups according to low-dose WMS: extensive viability (WMS ≤2.00), intermediate viability (WMS 2.00 to 2.49), and limited viability (WMS ≥2.50). Ejection fractions reported in this study were obtained from a regression equation (ejection fraction = −18.8[rest WMS] + 70.2; r = 0.753) derived by comparing ejection fraction measurements using Simpson’s method and WMS at rest in 140 patients with a wide spectrum of systolic function. Significant CAD was defined as ≥50% diameter narrowing of a major epicardial coronary artery or bypass graft determined by visual or caliper assessment by a cardiologist blinded to the results of follow-up.
The decision for medical therapy or revascularization was made by referring physicians. A patient was assigned to the revascularized group if the procedure was performed within 5 months of the dobutamine examination. Follow-up was primarily retrospective and conducted by telephone interview and review of clinic records. Examination of clinic records after treatment assignment was used to determine drug and device therapy received during follow-up. Cardiac death was the end point of the study, defined as death from myocardial infarction, congestive heart failure (CHF), or sudden death without an obvious noncardiac cause. Medical records and contact with next of kin were obtained for documentation of cause of death, and in rare cases, death certificates were used to determine causes of death.
SAS version 9.1.3 (SAS Institute Inc., Cary, North Carolina) was used for statistical analyses. Continuous variables are reported as mean ± SD. A p value <0.05 was considered significant. Normally distributed continuous variables were compared using independent Student’s t tests, and Wilcoxon’s rank-sum tests were used for non-normally distributed variables. Fisher’s exact tests were used to compare categorical variables. McNemar’s test was used to compare the frequency of medication use before and after treatment assignment.
Because of potential baseline differences between revascularized and medically treated patients, propensity score analysis was used. A multivariate logistic regression model was developed in which the assigned treatment was the dependent variable and 25 factors that were considered to affect treatment assignment were independent variables. These variables were age; gender; diabetes; hypertension; hyperlipidemia; smoking; family history of CAD; previous CABG; previous infarction; New York Heart Association heart failure (CHF) class; use of nitrates, calcium channel blockers, β blockers, and angiotensin-converting enzyme inhibitors; combined use of β blockers and angiotensin-converting enzyme inhibitors; extent of CAD; left anterior descending artery disease; the ejection fraction; rest, low-dose, and peak-dose WMS; rest to low-dose change in WMS; and the percentages of viable, nonviable, and ischemic myocardium.
The model was used to calculate a propensity score for each patient that reflected the probability of being assigned to revascularization on the basis of the 25 independent variables. Patients were stratified into quintiles on the basis of their propensity scores, and treatment group comparisons were made. If the propensity scores are similar between medically treated and revascularized patients within the quintiles, then the distribution of the independent variables are also similar between treatment groups within each quintile, and treatment group assignment can be considered random. Survival time for each treatment group was compared with Cox regression models. The models included the original propensity score as a covariate. Survival plots were derived from the Cox models. Covariates also included medications used in follow-up. At 3 and 5 years of follow-up, the annualized cardiac mortality rates for each treatment group were computed as the total estimated number of cardiac deaths divided by the number of patient-years of follow-up. The total estimated number of cardiac deaths for each group was the sum of all probabilities of cardiac death estimated from multivariate logistic regression models adjusted for propensity score and for β-blocker use in follow-up.
Results
The mean age was 60 ± 11 years, and 24% (n = 66) were women. Diabetes was present in 46% (n = 125), and 32% (n = 89) had class III or IV CHF. The mean ejection fraction by regression was 32 ± 7% and 26 ± 9% in those who had ventriculography. Left anterior descending artery disease was present in 95% (n = 259), and 96% (n = 262) had multivessel disease. Viability (contractile reserve) in ≥25% of the left ventricle was present in 32% (n = 88).
One hundred thirty patients (47%) were revascularized, and 53% (n = 144) were medically treated. Of those revascularized, 93% (n = 122) underwent CABG and 7% (n = 8) had percutaneous revascularization. Complete revascularization was achieved in 73% of revascularized patients. Arterial grafts were used in 110 patients (90%) who underwent CABG. Four patients with CABG had concomitant procedures: 1 mitral valve replacement, 1 aneurysm resection, 1 mitral valve repair and aneurysm resection, and 1 aortic valve replacement for moderate stenosis. Over a mean follow-up period of 4.5 ± 3.7 years, there were 114 cardiac deaths (42%), including 35% of revascularized patients (45 of 130) and 48% of medically treated patients (69 of 144). There were 2 perioperative deaths in patients who underwent CABG.
Table 1 lists the clinical characteristics of medically treated and revascularized patients. Revascularized patients were more likely to have hyperlipidemia and multivessel CAD. There were no significant differences between the 2 groups on 15 other clinical variables. Table 2 lists echocardiographic variables between the 2 groups. Medically treated patients had a slightly larger proportion of nonviable myocardium.
| Variable | Medically Treated | Revascularized | p Value |
|---|---|---|---|
| (n = 144) | (n = 130) | ||
| Age (years) | 61 ± 12 | 60 ± 10 | 0.586 |
| Women | 33 (23%) | 33 (25%) | 0.673 |
| Diabetes mellitus | 70 (49%) | 55 (42%) | 0.332 |
| Hypertension | 94 (65%) | 93 (72%) | 0.299 |
| Hyperlipidemia ⁎ | 53 (37%) | 78 (60%) | <0.001 |
| Smokers | 112 (78%) | 94 (72%) | 0.329 |
| Previous CABG | 32 (22%) | 23 (18%) | 0.369 |
| Previous infarction | 95 (66%) | 90 (69%) | 0.607 |
| NYHA class III or IV HF | 47 (33%) | 42 (32%) | 1.000 |
| Nitrates | 95 (66%) | 87 (67%) | 0.899 |
| Calcium channel blockers | 41 (29%) | 40 (31%) | 0.693 |
| β blockers | 54 (38%) | 58 (45%) | 0.269 |
| ACE inhibitors | 87 (60%) | 78 (60%) | 1.000 |
| β blockers and ACE inhibitors | 37 (26%) | 30 (23%) | 0.577 |
| Left anterior descending artery disease | 134 (93%) | 125 (96%) | 0.298 |
| Multivessel CAD | 134 (93%) | 128 (98%) | 0.037 |
⁎ Total cholesterol >200 mg/dl or treatment with lipid-lowering agent.
| Variable | Medically Treated | Revascularized | p Value |
|---|---|---|---|
| (n = 144) | (n = 130) | ||
| Rest wall score | 2.06 ± 0.39 | 2.01 ± 0.39 | 0.256 |
| Ejection fraction by regression | 32 ± 7% | 32 ± 7% | 0.264 |
| Low-dose wall motion scores | 1.98 ± 0.39 | 1.89 ± 0.41 | 0.112 |
| Peak-dose wall motion scores | 2.01 ± 0.40 | 1.93 ± 0.42 | 0.121 |
| Low-dose wall motion scores group | 0.093 | ||
| <2.00 | 70 (48%) | 80 (62%) | |
| 2.00–2.49 | 57 (40%) | 37 (28%) | |
| ≥2.50 | 17 (12%) | 13 (10%) | |
| Rest to low-dose change in score | −0.08 ± 0.19 | −0.11 ± 0.19 | 0.182 |
| Percentage viable myocardium at low dose | 16 ± 15% | 19 ± 14% | 0.191 |
| ≥25% viable myocardium at low dose | 42 (29%) | 46 (35%) | 0.271 |
| Percentage nonviable (akinetic) myocardium at low dose | 27 ± 19 | 22 ± 17 | 0.024 |
| Ischemia present | 108 (75%) | 104 (80%) | 0.386 |
| Percentage ischemic myocardium | 16 ± 16 | 18 ± 16 | 0.257 |
Patients were stratified into quintiles on the basis of propensity score. Table 3 lists p values for the comparison of clinical characteristics between treatment groups after stratification by propensity scores. There were no significant differences in clinical characteristics between treatment groups within each quintile. Table 4 list p values for the comparison of echocardiographic variables between treatment groups after stratification by propensity scores. There were no significant differences in variables between treatment groups.
| Variable | Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 |
|---|---|---|---|---|---|
| Medically treated | 44 | 36 | 27 | 29 | 8 |
| Revascularized | 10 | 19 | 28 | 26 | 47 |
| Age | 0.859 | 0.705 | 0.240 | 0.967 | 0.849 |
| Women | 1.000 | 0.401 | 0.245 | 0.367 | 0.423 |
| Diabetes | 0.728 | 1.000 | 0.593 | 1.000 | 1.000 |
| Hypertension | 0.723 | 0.779 | 1.000 | 0.721 | 0.669 |
| Hyperlipidemia | 0.667 | 0.334 | 0.573 | 1.000 | 1.000 |
| Smoking | 0.667 | 1.000 | 0.759 | 0.364 | 0.426 |
| Family history of coronary artery disease | 0.723 | 0.775 | 0.591 | 0.788 | 0.244 |
| Previous coronary artery bypass grafting | 1.000 | 0.749 | 0.469 | 1.000 | 0.267 |
| Previous infarction | 0.471 | 0.565 | 0.768 | 0.149 | 0.414 |
| NYHA class III or IV heart failure | 0.476 | 0.779 | 1.000 | 1.000 | 0.700 |
| Nitrates | 0.708 | 0.389 | 1.000 | 1.000 | 0.089 |
| Calcium channel blockers | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| β blockers | 0.728 | 1.000 | 0.573 | 0.789 | 0.276 |
| Angiotensin-converting enzyme inhibitors | 1.000 | 1.000 | 0.412 | 0.395 | 1.000 |
| β blockers and angiotensin-converting enzyme inhibitors | 0.459 | 1.000 | 0.547 | 1.000 | 1.000 |
| Left anterior descending artery disease | 0.601 | 1.000 | 1.000 | 1.000 | 0.272 |
| Multivessel disease | 0.902 | 0.770 | 1.000 | 0.721 | 0.272 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
