Author + information
- Stephen Sawada, MD, FACC*,†,*† (, )
- Ashutosh Bapat, MD*,†,
- Dev Vaz, MD*,†,
- Juan Weksler, MD*,†,
- Naomi Fineberg, PhD*,
- Adam Greene, MD, FACC*,†,
- Irmina Gradus-Pizlo, MD, FACC*,† and
- Harvey Feigenbaum, MD, FACC*,†
- ↵*Reprint requests and correspondence:
Dr. Stephen Sawada, 1801 North Senate Boulevard, M.P.C. II, Suite D4082, Indianapolis, Indiana 46202, USA.
Objectives We assessed the incremental long-term prognostic value of myocardial viability in surgically revascularized (CABG) patients with left ventricular (LV) dysfunction.
Background Clinical factors, medical therapy, the degree of LV dysfunction, and stress-induced ischemia may affect the relative prognostic value of myocardial viability.
Methods Patients with coronary disease and ventricular dysfunction (mean ejection fraction 33% by echocardiography, 25% by angiography) were studied with dobutamine echocardiography. Follow-up (mean −4.9 years) was obtained in 95 patients (85% triple-vessel disease) who underwent CABG.
Results The use of angiotensin-converting enzyme inhibitors, advanced heart failure, rest, low- and peak-dose wall motion scores were univariate predictors of cardiac death. The extent of contractile reserve and ischemia were not predictive. Low-dose score was the strongest multivariate predictor of death (p < 0.001, hazard ratio 6.7). A biphasic response predicted better survival (p = 0.045, hazard ratio 0.5). Five-year survival was better in those with extensive (low-dose score <2.00) versus intermediate (score 2.00 to 2.49) amounts of viable myocardium (p = 0.019). Patients with the least viability (score ≥2.5) had the worst outcome (p = 0.0001 vs. those with low-dose score <2.00; p = 0.05 vs. those with score 2.00 to 2.49). In stepwise multivariate analysis, low-dose score added incremental prognostic value (p = 0.024) to clinical information and rest score.
Conclusions Low-dose score, representing the extent of viable myocardium, has incremental prognostic value as a predictor of long-term outcome in CABG patients with LV dysfunction.
Patients with left ventricular (LV) dysfunction and advanced coronary artery disease (CAD) who have viable myocardium have better survival with revascularization compared with medical therapy (1–12). Less is known about the importance of myocardial viability to outcome within patient cohorts who are revascularized. The results of prior studies suggest that revascularized patients without viable myocardium have relatively low short-term mortality (4). In general, prior studies have treated viability as a binary variable, identifying patients as “nonviable” or “viable” based on certain threshold values (13). As a result, less is known about the impact of the extent of viable myocardium on the prognosis of revascularized patients. Additionally, the long-term impact of viability in revascularized patients is un-known because, in most studies, follow-up has been limited to two to three years (4). Investigation of the prognostic importance of the extent of viable myocardium may help to determine which patients derive the greatest survival benefit from revascularization.
Previous studies have also not assessed the overall impact of clinical factors, the degree of resting LV dysfunction, the extent of ischemia, and the completeness of revascularization on the prognostic value of viability in revascularized patients. In the clinical arena, assessment of the patients' history, medical regimen, and resting function constitute the initial steps in risk stratification. An evaluation for viable myocardium may follow, and the prognostic information from this assessment is added to the preexisting information.
The purpose of this investigation was to determine if the presence and extent of viable myocardium influences long-term prognosis in revascularized patients with LV dysfunction. We used several different criteria for viability to determine which, if any, would be predictive of outcome. We also sought to determine if the extent of viable myocardium added incremental prognostic value to assessment of clinical factors, the degree of resting dysfunction, the extent and severity of CAD, the extent of ischemia, and the completeness of revascularization. Patients undergoing coronary artery bypass grafting (CABG) were selected because in these subjects viability may have a greater impact on outcome compared with those who have percutaneous revascularization because of the greater potential benefit from more complete and long-lasting revascularization.
The study group was selected from patients with LV dysfunction (≥4 segments with resting wall motion abnormalities) who underwent low and peak-dose dobutamine echocardiography (DobE) between April 1991 and December 1998 at Indiana University hospitals. Of 618 subjects with LV dysfunction evaluated during this time period, 111 (18%) had CABG within five months of DobE. Thirteen patients who had DobE within one week after myocardial infarction were excluded. One patient who did not have complete low-dose imaging was excluded, and two patients who had ischemic events before undergoing CABG were also excluded. The 95 remaining patients comprised the study population. The investigation was approved by the institutional review board, and informed consent was obtained from all subjects.
Dobutamine was administered beginning at a dose of 5 μg/kg/min for 3 min followed by 10 μg/kg/min for 3 min. Thereafter, the dose was increased by 10 μg/kg/min increments every 2 to 3 min to a peak-dose of 50 μg/kg/min. End points for dobutamine infusion were the same as described in a previous study (12). Echocardiograms were recorded and digitally stored at rest, at the end of the 5 μg/kg/min and 10 μg/kg/min stages, and at peak dose.
Echocardiograms were analyzed by at least one investigator blinded to the clinical, angiographic, and stress testing data. Wall motion and thickening were assessed in 16 segments using a 6-grade scoring system: 1 = normal at rest, hyperkinetic with dobutamine; 2 = mildly hypokinetic with <5 mm inward systolic motion; 2.5 = severely hypokinetic with minimal inward motion and wall thickening; 3 = akinetic with absence of inward motion and wall thickening; 4 = dyskinetic with paradoxical outward motion; 5 = aneurysmal with diastolic thinning (<7 mm thickness).
When a disparity existed between wall motion and thickening of an individual segment, regional function was graded on the basis of wall thickening. Wall motion scores at rest, at 10 μg/kg/min (designated as low dose), and at peak dose were obtained by dividing the sum of individual segment scores by the number of segments scored.
A blinded observer measured ejection fraction (EF) using the apical four-chamber view and Simpson's method.
Previous studies have identified patients with viable myocardium based on the presence of contractile reserve, the improvement of wall thickening, and motion with low-dose dobutamine (9–11,14). In this study, several measures of viable myocardium used in prior studies were included. Improvement in wall motion score from rest to low dose of ≥0.20, and a biphasic response, were included as measures of viability (9,15). The percent of myocardium with contractile reserve at low dose was also determined. This was defined as the number of segments with improved wall motion scores at 10 μg/kg/min divided by the number of segments scored at this dose × 100%. The proportion of patients with contractile reserve in at least 25% of ventricular myocardium was also determined (9,16).
Additionally, the percent of ischemic myocardium was assessed. This was defined as the number of segments that had worsening of wall motion at any stage of the test divided by the number of segments scored at peak dose × 100%. Segments with worsening of wall motion from a score of 3 (akinetic) to 4 (dyskinetic) were not considered ischemic.
Agreement between two blinded observers for the presence of contractile reserve was determined. In 20 patients, the resting and low-dose wall motion scores were compared between blinded interpretations and the clinical (unblinded interpretations) made by a stress echocardiography expert at the time of performance of DobE to determine if assessment of viability by blinded readers was similar to that of clinical readers.
Seventy-two of 95 patients in the study had been referred for evaluation of viable myocardium. These subjects were prospectively followed. Retrospective follow-up was obtained in 23 additional patients who met the inclusion criteria for this study. Patients were contacted by telephone for follow-up. In cases of hospital readmission or death, medical records were obtained for documentation. Cardiac death was the end point of the study. It was defined as death due to intractable heart failure (CHF), myocardial infarction, ventricular arrhythmia, or sudden death within 1 h of symptom onset without an obvious noncardiac cause.
Cox proportional hazards model was used to identify univariate predictors of cardiac death with a p value <0.05 considered statistically significant. Multivariate analysis was conducted using all variables with p values <0.10 on univariate analysis. Step-wise multivariate analysis was performed using all clinical, angiographic, and echocardiographic variables with p < 0.10 on univariate analysis. Receiver operator characteristic curve analysis was used to determine potentially useful threshold values of echocardiographic variables for derivation of survival curves. Kaplan-Meier survival curves were constructed using five-year follow-up data and complete follow-up data. The curves were compared using the log-rank test. Linear regression was used to compare wall motion scores between blinded and clinical observers. Statistical analysis was performed with SPSS version 11.0 for Windows (SPSS Inc., Chicago, Illinois).
Forty-one (43%) of 95 patients presented with worsening CHF, and 30 (32%) presented with worsening angina. The remaining 24 patients had stable symptoms or were asymptomatic.
The primary indication for performance of DobE was assessment of viability and ischemia in 72 subjects. In the remaining 23 subjects, 14 were studied for evaluation of worsening angina or CHF, and 9 were studied for preoperative risk stratification before noncardiac surgery.
The mean age of the patients was 60 ± 9 years. There were 73 men (77%) and 22 women. Thirty-six patients (38%) had diabetes, and 31 (33%) had class III or IV CHF. Of the 95 patients, 51 (54%) were receiving angiotensin-converting enzyme inhibitors (ACEI), and 33 (35%) were receiving beta-blockers. The mean resting EF was 33 ± 10% by echocardiography and 25 ± 8% by nuclear or contrast angiography. The average extent of akinetic myocardium without contractile reserve was 24%. Triple-vessel disease (disease defined by ≥50% diameter stenosis) was present in 85%, and two-vessel disease in 15%.
Follow-up was obtained in all patients for a mean of 4.9 ± 2.9 years. Coronary artery bypass grafting was performed 25 ± 34 days after echocardiography. Revascularization of all significantly diseased vessels was performed in 75% of patients, and 88% had at least one internal mammary graft. Six patients had concomitant procedures (LV aneurysm resection = 2 patients; aortic valve replacement = 1 patient; aortic valve repair = 1 patient; mitral valve repair = 2 patients). Cardiac deaths occurred in 36 patients (38%), and 10 patients had noncardiac death.
There were two perioperative deaths (2.1%). One subject who had aneurysmectomy could not be weaned from cardiopulmonary bypass. The other subject had death due to uncontrollable ventricular arrhythmias. Both subjects had viable myocardium by criteria used in prior studies. The remaining 93 subjects were included in the analysis of factors predicting long-term outcome. Patients with noncardiac death were analyzed in the “survivors” group with their follow-up terminating at their date of death.
Univariate clinical and echocardiographic predictors of outcome
The clinical characteristics of survivors and patients with cardiac death during long-term follow-up are compared in Table 1. Advanced (class III or IV) CHF predicted cardiac death (p = 0.022); ACEI use was more frequent in those who died (p = 0.040). Angiotensin-converting enzyme inhibitor use was associated with worsening CHF at presentation (kappa = 0.24, p = 0.021). There was a trend towards higher cardiac mortality in men compared with women (p = 0.097) and a trend towards lower mortality in those receiving beta-blockers (p = 0.081).
Echocardiographic parameters in survivors and patients who had cardiac death are compared in Table 2. Resting wall motion score was a significant predictor of outcome, but EF was not. With low-dose dobutamine, only low-dose score was a significant predictor (p < 0.001). There was a trend towards a greater reduction in wall motion score from rest to low dose in survivors compared with those who died (p = 0.094). With high-dose dobutamine, peak-dose wall motion score was the only significant predictor. There was a trend towards a higher frequency of a biphasic response in survivors compared with those with cardiac death (p = 0.092).
Table 3shows that angiographic and revascularization parameters were similar in survivors and those who had cardiac death, and there were no significant predictors of outcome.
Variables included in multivariate analysis were CHF class; the use of ACEI; gender; beta-blocker therapy; resting-, low-, and peak-dose scores; change in score from rest to low-dose, and a biphasic response. Low-dose score was the most important independent predictor of cardiac death with a p value of <0.001. The hazard ratio for low-dose score was 6.7 (2.8 to 15.8). A biphasic response was an independent predictor of survival (p = 0.045) with a hazard ratio of 0.5 (0.2 to 0.99).
Survival analysis using low-dose scores
Receiver operator characteristic curve analysis demonstrated that the optimal threshold value of low-dose score for prediction of cardiac death was 1.95. In previous studies, wall motion score values of 2.00 and 2.50 have been found to be useful thresholds for identification of high-risk patients (12,17). For the purposes of this study, patients were assigned to three groups based on low-dose score (group 1 with extensive viable myocardium, score <2.00; group 2 with intermediate viability, score 2.0 to 2.49; group 3 with limited viability, score ≥2.5). The frequency of cardiac death was 24% (13/55) in group 1, 48% (14/29) in group 2, and 82% (9/11) in group 3 (p = 0.028 group 1 vs. 2; p = 0.079 group 2 vs. 3). The annual cardiac mortality was 21% in group 3 for the first 3 years of follow-up.
Figure 1shows survival curves at five years of follow-up and for the complete duration of the study with all 95 patients included in the analysis. At five years, there was significantly better survival in group 1 (extensive viability) compared with group 2 (p = 0.019) and group 3 (0.0001). There was a trend towards improved survival in group 2 compared with group 3 (p = 0.053).
Eight of 10 noncardiac deaths in the study population occurred in group 1, and a substantial proportion of patients in this group were censored after five years. Over the complete duration of the study, there was a trend towards better survival in group 1 versus group 2 (p = 0.104). Both groups 1 and 2 had better survival than patients with limited viability in group 3 (p < 0.0001 vs. group 1; p = 0.0074 vs. group 2).
Incremental value of low-dose score
The incremental value of low-dose score was tested in a stepwise multivariate analysis (Fig. 2). Clinical variables used in the first step were CHF class, the use of ACEI, the use of beta-blockers, and gender. In the second step, resting wall score added incremental value to clinical parameters (p = 0.003). In the third step, low-dose score added incremental prognostic value to the combination of clinical variables and rest score (p = 0.024). In this step, the inclusion of change in score from rest to low-dose did not improve the prognostic value of low-dose score alone. In the final step, peak-dose score and a biphasic response provided no additional prognostic value.
Interobserver variability in wall motion scores
The interobserver agreement between two blinded observers for myocardium with contractile reserve was 89%. Linear regression was used to compare clinical and blinded assessments for resting wall motion score and low-dose score. The correlation coefficient was 0.94 for resting score and 0.87 for low-dose score. There was 95% agreement between clinical and blinded readings for assignment of patients to the three groups with extensive, intermediate, and limited viability.
This investigation showed that: 1) low-dose wall motion score, reflecting the extent of viable myocardium, was the best predictor of five-year outcome; 2) low-dose score enabled classification of patients into high, intermediate, and low-risk groups for cardiac death; 3) except for a biphasic response, measures of contractile reserve shown in prior studies to be predictors of short-term outcome were not predictive of long-term outcome; 4) low-dose score added incremental prognostic value to clinical factors, and assessment of resting function; and 5) ischemia did not have additional prognostic value.
Short- and long-term prognostic value of viability: implications for selection of patients for surgical revascularization
Patients with the least amount of viable myocardium (group 3) did very poorly in spite of revascularization with 63% dying within three years. As discussed in a recent editorial, there is relatively limited information on the optimal management (medical therapy vs. revascularization) of patients with limited viable myocardium (13). The results of previous studies have shown no adverse effects of revascularization in this group. A recent meta-analysis found an annual mortality of 7.7% in revascularized patients and 6.2% in medically treated patients with limited viability (p = 0.23) (4). Our results indicate that there is a subset of patients with very limited viability in whom surgical revascularization is associated with a very poor prognosis.
In this investigation, the presence of extensive viable myocardium (group 1, low-dose score ≤2.00) conferred a prognostic advantage over those with intermediate viability (group 2) lasting at least five years. Beyond five years, the groups with extensive and intermediate viability maintained better survival compared with those with limited viability.
In previous studies with mixed patient populations undergoing percutaneous or surgical revascularization, the prognostic benefit of contractile reserve has been shown to last for at least three years after revascularization (10,11). The results of our study indicated that there is extended prognostic benefit of viable myocardium in surgically revascularized patients. The completeness of revascularization may have contributed to the long-term benefit of viability in this study. In the Coronary Artery Surgery registry, outcomes in patients with LV dysfunction with ≥3 grafts was improved compared with those with fewer grafts (18). In our study, nearly all patients had three or more grafts. Nearly all patients also had at least one arterial graft. The importance of arterial conduits to outcome in patients with significant viable myocardium is uncertain (19).
Incremental prognostic value of myocardial viability
In clinical practice, selection of high-risk patients for CABG is a process that takes into consideration multiple clinical factors, the degree of resting dysfunction, the presence and extent of inducible ischemia, and the extent of coronary disease, in addition to information on myocardial viability. The results of our study indicated that assessment of viability adds incremental long-term prognostic information. To date, information documenting the incremental prognostic value of viability has been limited. Pasquet and colleagues (20)have shown that contractile reserve adds prognostic value to clinical factors in predicting outcome. Medications were not included in their analysis. Their results showed that, after adjustment for diabetes and the extent of disease, contractile reserve remained predictive of three-year survival in revascularized patients.
Indicators of viability: contractile reserve versus low-dose score
In contrast with the results of previous studies, contractile reserve was a weak predictor of outcome (9–11,14,16). The reasons for the stronger predictive ability of low-dose wall motion score compared with that for contractile reserve are unclear. Improvement in the prognosis of patients with LV dysfunction may result from postrevascularization improvement in function of viable myocardium. Meluzin et al. (14)and Bax et al. (16)demonstrated that patients who had contractile reserve and improvement of EF with revascularization had lower rates of total cardiac events in two years of follow-up. However, there were no significant differences in survival between patients with and without significant contractile reserve because of the low number of deaths in both studies.
An investigation by Samady et al. (21)called into question the mechanism for improvement in prognosis of revascularized patients with LV dysfunction. In his study, outcomes were similar in revascularized patients with and without improvement in systolic function. Factors other than improvement in systolic function may account for some of the prognostic benefit of revascularization. Revascularization may reduce the incidence of sudden death and the size of future infarction, outcomes that may be unrelated to the presence of contractile reserve (22,23).
Low-dose wall motion scores incorporate information on the functional status of all regions of the myocardium. Dysfunctional regions with contractile reserve, nonviable regions without response to dobutamine, and normal regions all contribute to the overall wall motion score. Revascularization of normal regions may relieve ischemia and contribute to better outcome. Revascularization of regions without contractile reserve may still be beneficial by preserving some viable cells and preventing adverse ventricular remodeling. By incorporating information on the functional status of all regions of the LV, low-dose wall motion score may provide comprehensive prognostic information.
In this study, the use of low-dose score also permitted assessment of viability as a continuous variable and improved risk stratification over the simple classification of patients as viable or nonviable. The use of low-dose score is similar to employment of a viability index by Pagley et al. (24). They assessed perfusion of each LV segment using single photon emission computed tomography thallium images, and an average perfusion score for each patient was determined. Patients with higher viability scores (more normal perfusion) had better survival with revascularization than those with lower scores.
Prognostic value of stress-induced ischemia
In this study, the presence and extent of ischemia did not add prognostic value. The extent of ischemia was limited in both survivors and those who died, perhaps because both groups had more extensive amounts of nonviable myocardium than in populations enrolled in previous studies.
The results of limited studies suggest that the extent of preoperative ischemia may not have significant impact on the outcome of revascularized patients. The survival of revascularized patients with ischemia in a study by Smart et al. (25)was excellent, regardless of the extent of ischemia. Pasquet et al. (20)found that dobutamine-induced ischemia did not predict the outcome of patients within the revascularized group. Our results are similar and suggest that, in patients with relatively limited ischemia, myocardial viability is a more important factor. The relative completeness of revascularization in our study may have also played a role in limiting the prognostic impact of ischemia.
Patients were included from 1991 to 1998, a protracted period of time during which significant changes in medical therapy have taken place. Fewer patients in our study were receiving ACEI and beta-blockers than may be expected for a study initiated more recently. However, the frequency of ACEI use in our study (54%) was comparable to that for Medicare patients (55%) treated for CHF in the mid-1990s (26). The frequency of beta-blocker use in our study was also low (35%) but very similar to rates (30% to 35%) reported by the National Registry of Myocardial Infarction in patients treated for infarction complicated by CHF between the years of 1994 and 2000 (27).
Angiotensin-converting enzyme inhibitor use was unexpectedly associated with higher cardiac mortality. There was a significant association of ACEI use with worsening CHF at presentation, suggesting that ACEI use was another marker for the degree of CHF.
Revascularization was considered complete if all obstructed vessels were grafted. No postoperative assessment of perfusion or stress testing was performed.
A highly select patient population was enrolled. The results of this study are applicable to those with extensive CAD and advanced LV dysfunction who have CABG and may not be applicable to those undergoing percutaneous revascularization or to those treated medically.
The authors thank Amy Stewart for secretarial assistance.
☆ Supported by the Herman C. Krannert Fund and the Showalter Foundation, Indianapolis, Indiana.
- angiotensin-converting enzyme inhibitor
- coronary artery bypass grafting
- coronary artery disease
- congestive heart failure
- dobutamine echocardiography
- ejection fraction
- left ventricle/ventricular
- American College of Cardiology Foundation
- Alderman E.L.,
- Bourassa M.G.,
- Cohen L.D.,
- et al.
- Rigott J.D.,
- Kouchoukos N.T.,
- Oberman A.,
- Cutter G.R.
- Allman K.C.,
- Shaw L.J.,
- Hachamovitch R.,
- Udelson J.E.
- Eitzman D.,
- Ziad Al-Aouar Z.,
- Kanter H.L.,
- et al.
- William M.J.,
- Odabashian J.,
- Lauer M.S.,
- Thomas J.D.,
- Marwick T.H.
- Afridi I.,
- Grayburn P.A.,
- Panza J.A.,
- Oh J.K.,
- Zoghbi W.A.,
- Marwick T.H.
- Chaudry F.A.,
- Tauke J.T.,
- Alessandri R.S.,
- Vardi G.,
- Parker M.A.,
- Bonow R.O.
- Senior R.,
- Kaul S.,
- Lahiri A.
- Bonow R.D.
- Meluzin J.,
- Cerny J.,
- Frelich M.,
- et al.
- Sicari R.,
- Ripoli A.,
- Picano E.,
- et al.
- Bax J.J.,
- Poldermans D.,
- Elhendy A.,
- et al.
- Shina A.,
- Tajik A.J.,
- Smith H.C.,
- Lengyel M.,
- Seward J.B.
- Bell M.R.,
- Gersh B.J.,
- Schaff H.V.,
- et al.
- Lorussa R.,
- La Canna G.,
- Ceconi C.,
- et al.
- Pasquet A.,
- Robert A.,
- D'Hondt A.-M.,
- Dion R.,
- Melin J.A.,
- Vanoverschelde J.-L.J.
- Samady H.,
- Elefteriadis J.A.,
- Abbott B.G.,
- Mattera J.A.,
- McPherson C.A.,
- Wackers F.J.
- Crean P.A.,
- Waters D.D.,
- Bosch X.,
- Pelletier G.B.,
- Roy D.,
- Theroux P.
- Veenhuyzen G.D.,
- Singh S.N.,
- McReavey D.,
- et al.
- Pagley P.R.,
- Beller G.A.,
- Watson D.D.,
- Gimple L.W.,
- Ragosta M.
- Smart S.C.,
- Dionisopoulos P.N.,
- Knickelbine T.A.,
- Schuchard T.,
- Sagar K.B.
- Spencer F.A.,
- Meyer T.E.,
- Gore J.M.,
- Goldberg R.J.