Revascularization in patients with coronary artery disease, left ventricular dysfunction, and viability: a meta-analysis

Clinical Investigations

Acute Ischemic Heart Disease

Revascularization in patients with coronary artery disease, left ventricular dysfunction, and viability: A meta-analysis Jamieson MacDonald Bourque, MD,a Vic Hasselblad, PhD,b Eric J. Velazquez, MD,a Salvador Borges-Neto, MD,c and Christopher M. O’Connor, MD,a Durham, NC

Background The effects of viability status and treatment allocation on long-term mortality in patients with left ventricular dysfunction and coronary artery disease have not been determined. Several observational studies with significant limitations have addressed this issue, and a recent meta-analysis has attempted to combine these results to increase statistical power. However, the analysis did not test for an interaction between viability status and treatment type, and included extraneous studies. We provide an alternate meta-analysis of this primary literature, utilizing interaction statistical methodology on relevant data and factoring in multiple limitations. Methods

We examined papers from this prior meta-analysis examining viable and nonviable patients undergoing surgical or medical therapy. We determined an interaction odds ratio for each study and used an empirical Bayes random-effects model to obtain a combined interaction odds ratio that was tested for statistical significance. We compared our results against an interaction odds ratio we estimated from the primary studies included in the previous meta-analysis.

Results Nine relevant studies with 1244 patients and 172 events were identified that utilized all 4 treatment/viability subsets. The interaction odds ratio was 2.76 (P ⫽ .0176, 95% CI 1.19-6.38), 2.5 times lower than our estimated interaction odds ratio of 7.27 for the prior meta-analysis. Conclusions We found a markedly reduced but statistically significant interaction between viability status and treatment allocation. However, numerous limitations in the primary studies and the application of meta-analysis along with significant improvements in medical therapies render a randomized controlled trial necessary to reach a definitive conclusion to this critical question. (Am Heart J 2003;146:621–7.) Coronary revascularization likely provides improvement for only a subset of patients with left ventricular dysfunction and coronary artery disease and carries increased risk and mortality as the ejection fraction decreases.1,2 Increasing myocardial viability potentially identifies the cohort that will manifest the greatest improvement with revascularization. Viability levels can be identified

From the aDivision of Cardiology, Department of Internal Medicine, bDepartment of Biostatistics and Bioinformatics and the cDivision of Nuclear Medicine, Department of Radiology, Duke University Medical Center, Durham, NC. Supported by a grant from the Tom & Lynn Royster Foundation, Durham, NC, and NIH Research Fellowship Grant T5 GM08679-04, Bethesda, Md. Guest Editor for this manuscript was Peter B. Berger, MD, Mayo Clinic, Rochester, Minn. Submitted October 28, 2002; accepted April 14, 2003. Reprint requests: Jamieson M. Bourque, MD, DUMC Box 3356, Durham, NC 27710. E-mail: [email protected] © 2003, Mosby, Inc. All rights reserved. 0002-8703/2003/$30.00 ⫹ 0 doi:10.1016/S0002-8703(03)00428-9

through radionuclide perfusion imaging, 18F-fluorodeoxyglucose metabolic imaging, or dobutamine echocardiography. All 3 imaging methods identify viable myocardium with high sensitivity and specificity.3 The effects of viability status before revascularization on short-term outcomes, including ejection fraction, New York Heart Association (NYHA) class, myocardial infarction, and short-term mortality have been well described in observational studies.4-6 There have been several small cohort analyses that have examined the link between myocardial viability and long-term mortality. However, significant limitations in study design, event rates, methods, and other factors have reduced the ability of these studies to reach a definitive answer.7 Meta-analysis could increase the statistical power of these studies, and has been recommended by Sackett and Rosenberg as an appropriate tool for synthesizing research findings.8 However, there are several inherent flaws of this technique that can be amplified by defi-

American Heart Journal October 2003

622 Bourque et al

ciencies within the primary resources.9 The weaknesses in study inclusion criteria, statistical methodology, and overall quality present in these manuscripts must be addressed before consideration of whether a meta-analysis on this subject should impact current clinical practice. However, a meta-analysis on this topic has already been published that included 24 studies examining survival and viability status through thallium perfusion imaging, 18F-fluorodeoxyglucose metabolic imaging, or dobutamine echocardiography in patients with ischemic heart failure.10 The purpose of our analysis was to critically examine the interaction between treatment allocation and myocardial viability status and its effect on long-term mortality in this cohort of patients with left-ventricular dysfunction and significant coronary artery disease while considering the many limitations of the primary studies and the meta-analytic techniques to combine them.

Methods The Institutional Review Board at Duke University approved of this study. We carefully examined the analyses included in the previously published meta-analysis by Allman et al for characteristics including study type and design, inclusion and exclusion criteria, patient number, methods, length of follow-up, event rate, and final results. We extracted the articles from this cohort that had patients in all 4 relevant groups: those who underwent revascularization and those who received medical therapy only, each subdivided by the presence or absence of viable myocardium. Both percutaneous coronary intervention and coronary artery bypass grafting were included in the revascularization group. We determined the characteristics of these primary manuscripts, including study type, number of patients analyzed, mean age, mean ejection fraction, method of viability imaging utilized, and the criteria set for viability. Within each study, the patients were subdivided into 4 groups based on presence of viability and treatment received. Mortality rates (at time points that varied among studies) for each cohort were calculated. We identified the baseline characteristics of this cohort and then performed a meta-analysis, testing for an interaction between viability status and treatment type. We used the empirical Bayes random-effects model as described by Hedges and Olkin to combine the estimates of the interaction odds ratio.11 This method works for as few as 2 studies and has been used extensively in the literature. For each treatment, the odds of dying for patients with revascularization versus medical therapy were combined into an odds ratio, providing an estimate of the likelihood of dying with revascularization. This ratio was determined for the groups with viable and nonviable myocardium. The interaction odds ratio (ratio of the viable and nonviable odds ratios) was then computed, which provided the magnitude and direction of the interaction between viability and treatment allocation. The interaction odds ratio states that viability makes the effect of treatment allocation X times as likely to affect mortality. Finally, the combined interaction ratio was

compared against the null hypothesis (odds ratio ⫽ 1). We performed the same methodology on the entire cohort used in the Allman et al meta-analysis, yielding a combined interaction odds ratio for comparison with our findings.

Results We analyzed the 24 studies in the Allman et al metaanalysis and found 9 with patients in all 4 of the groups we have defined as necessary for data inclusion.10 These reports examined cohorts broken down by viability status and type of therapy.1,12-19 The remaining 15 manuscripts had missing data critical to the interaction methodology used. As shown in Table I, 6 studies did not report the numbers of patients or deaths for any of the groups.20-25 Two studies only reported this information for patients with viable myocardium.26,27 Gioia et al included patients treated medically in 1 analysis, while 3 looked exclusively at patients receiving revascularization.4,5,28,29 Two papers only examined viable patients undergoing revascularization without any control groups for comparison.30,31 Finally, 1 study had zero patients in the nonviable revascularization group.32 We analyzed the remaining 9 studies that included a total of 1244 patients with a median sample population of 129 (25th–75th quartiles 87-161). The characteristics of this pooled cohort are presented in Tables II and III and the results are given in Table IV. This cohort had a median follow-up period of 18 months (25th–75th quartiles 16-33) and a median ejection fraction of 34% (25th–75th quartiles 27%–38%). The total number of events was 172, with a median number of events per study of 14 (25th–75th percentiles 13-22). The prevalence of viability in this cohort was 57.1%. The estimate of the interaction odds ratio for each analysis and for the combination of the patients in all 9 studies is provided in Figure 1. The combined estimated interaction ratio for all 9 studies was 2.76 (95% CI 1.19-6.38), yielding a P value of .0176. We estimated the combined interaction odds ratio to be 7.27 for the Allman meta-analysis. No 95% CI is given because no statistical method can obtain proper results utilizing all of the included manuscripts.

Discussion The effects of viability status and treatment allocation on long-term mortality have been examined in multiple small analyses with many significant limitations. Although there are drawbacks with this approach, Allman et al used meta-analytic techniques to combine these primary studies and thus increase the power and the ability to reach results with more statistical significance. They concluded that there is a “strong association: between myocardial viability on noninvasive testing

American Heart Journal Volume 146, Number 4

Bourque et al 623

Table I. Characteristics of primary studies from the Allman cohort excluded from our meta-analysis Viable PCI/ CABG*

Author Thallium Cuocolo (’98) (24) Gioia (’95) (27) Petretta (’97) (22) Gioia (’96) (4) Pagley (’97) (28) FDG Dreyfus (’94) (30) Di Carli (’98) (23) Tamaki (’93) (20) Haas (’97) (31) Huitink (’98) (25) Beanlands (’98) (26) Echo Meluzin (’98) (29) Williams (’96) (21) Bax (’99) (5) Smart (’99) (32) Totals

Viable medical

Nonviable PCI/CABG

Nonviable medical

Age*

EF*

F/U duration (m)*

76 85 82 89 70

55 65 57 69 66*

38 30 ⬍40* 27 28

17 31 22 31 39

? 38 N/A N/A 33

? 6 ? N/A 6

? 47 N/A 38 N/A

? 16 N/A 22 N/A

N/A ? N/A N/A 37

N/A ? N/A N/A 15

N/A ? N/A 43 N/A

N/A ? N/A 11 N/A

50 93 158 34 53 85

58 68 60 63 61 62

23 25 46 28 51 26

18 46 23 15* 47 18

46 ? ? 34 ? 31

3 ? ? 1 ? 1

N/A ? ? N/A ? 14

N/A ? ? N/A ? 4

N/A ? ? ? ? N/A

N/A ? ? ? ? N/A

N/A ? ? N/A ? N/A

N/A ? ? N/A ? N/A

133 136 76 350 1570

58 67 61 61

35 30 28 30

20 16 19 18

29 ? 23 78 312

1 ? 1 7 26

N/A ? N/A 90 189

N/A ? N/A 44 86

104 ? 45 0 186

13 ? 5 0 33

N/A ? N/A 182 225

N/A ? N/A 15 26

No. pts

No. pts

Deaths

No. pts

Deaths

No. pts

Deaths

No. pts

Deaths

Pts, Patients; EF, ejection fraction; F/U, follow-up; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; FDG, Fluorodeoxyglucose; Echo, echocardiography; ?, value is not provided in primary manuscript; N/A, value has not been studied. *Data given as median if available. Mean provided otherwise.

Table II. Study design and baseline characteristics of nuclear viability studies on ischemic heart failure patients included in our meta-analysis Author Eitzman (’92) (12) Pasquet (’99) (19) Senior (’99) (18) Anselmi (’98) (17) Afridi (’98) (16) vom Dahl (’97) (15) DiCarli (’94) (14) Yoshida (’93) (13) Lee (’94) (1)

Study type

No. pts

Age*

EF*

Imaging type

Viability criteria

Retro Cohort Prosp Cohort Prosp Cohort Prosp Cohort Retro Cohort Prosp Cohort Prosp Cohort Prosp Cohort Retro Cohort

82 137 87 202 318 161 93 35 129

59 62 62* 59 64 57 65 54 62

34 35 25 33 27 45 25 44 38

FDG PET T1-201 R/R & Dob Echo Dob Echo Dob Echo Dob Echo 99m Tc-MIBI & 18F-FDG PET 18F-FDG PET Rb-82 PET 18F-FDG, Rb-82 PET

ⱖ1 Segments viable ⬎50% Mismatched Segments ⱖ5 Segments viable ⱖ1Segments viable ⱖ4 Segments viable ⱖ50% Mismatched myocardium† ⬎5% PET mismatch Any viability in at risk zones ⱖ1 Segments viable

Retro, Retrospective; Prosp, prospective; R/R, rest-redistribution; Dob, dobutamine; MIBI, sestamibi. *Data given as median if available, marked with asterisk. Mean provided otherwise. †In area of primary interest during study.

and improved survival after revascularization in patients with chronic CAD and LV dysfunction.”10 However, they based this conclusion on their finding that there was a significant difference in mortality by treatment allocation when viability was present, but none if it was absent. They analyzed 375 events for 3088 patients and stated that revascularization reduced mortality by 79.6% for patients with viability (16% vs 3.2%, ␹2 ⫽ 147, P ⬍.0001) while providing no benefit for those with no viability present (7.7% vs 6.2%, P ⫽

NS).10 Their analysis did not examine for a significant difference between the viable and nonviable patients, which was required to arrive at their conclusions. They concluded that there was a strong association, but did not perform any statistical testing to specifically test for an interaction or obtain CIs or a P value. This analysis requires that the interaction between viability and treatment modality be determined. We calculated an interaction odds ratio of the treatment effect for patients with viable versus nonviable myo-

American Heart Journal October 2003

624 Bourque et al

Table III. Study cohort summary characteristics Characteristics Age (y) Male (%) LVEF (%) Follow-up (months) Treatment Revascularization (%) Medical therapy (%) Viability status Viable (%) Revascularization (# pts) Medical therapy (# pts) Nonviable (%) Revascularization (# pts) Medical therapy (# pts)

Median

25th–75th Percentile

62 84 34 18

(59–62) (79–87) (27–38) (16–33)

49 51

(44–53) (47–56)

54 49 18 46 19 33

(54–64) (26–58) (16–32) (36–46) (14–24) (24–61)

LVEF, Left ventricular ejection fraction.

cardium and then combined these ratios using the empirical Bayes random-effects model, which tests directly for a significant difference between the 2 viability groups, providing CIs and a P value. The random-effects portion provides a partial correction for the heterogeneity that is surely present between the primary studies.11 In order to determine the interaction odds ratio for each included study, the odds of dying must be calculated for patients with and without viability receiving revascularization or medical therapy. Studies must be excluded if they do not have patients in all of these 4 groups. Allman et al based their conclusions on 24 studies. Rather than identifying a novel cohort of analyses, we chose to examine the primary literature identified by Allman et al so that we could compare our results directly to those described in their manuscript. However, only 9 of their 24 studies had patients in all 4 necessary cohorts, whereas the other studies provide insufficient information (Table I). We used these 9 studies in the empirical Bayes random effects model, which introduces a factor that attempts to correct for the heterogeneity present between the individual studies. Using this method, we obtained an interaction odds ratio of 2.76, which states that treatment allocation is 2.76 times as likely to affect the odds of dying in patients with viable myocardium as in those with nonviable tissue. Because the 95% CIs do not cross 1 (where the odds are equal irregardless of viability status) and the P value is ⬍.05, we have determined that there was a statistically significant interaction of viability and treatment allocation with regard to long-term mortality. This significance is impressive because there were only 41% as many patients as in the Allman et al co-

hort and only 46% as many events. Nonetheless, several important factors must be weighed when interpreting this result and applying it to current clinical practice. The magnitude of the interaction was ⬎2.5 times lower than our estimated results for the Allman cohort (7.27 vs 2.76). This large difference highlights both the decreased effect with more representative data, and the heterogeneity of results that can be obtained when we attempt to use observational data through various techniques in a meta-analysis. Numerous limitations and multiple levels of potential confounding complicate any definitive conclusion of the interaction between viability status and treatment effects on long-term mortality from observational data. A fundamental drawback to this analysis is the inclusion of cohort studies rather than randomized controlled trials. Cohort studies can be superior when randomized controlled trials are infeasible or unethical. However, carefully-performed, randomized, controlled trials are often more appropriate for meta-analyses looking at long-term outcomes due to both methods that are carefully controlled and to subject randomization, which equalizes baseline characteristics to reduce patient selection bias and minimize confounders. Allman et al’s 24 studies had almost certain patient selection biases. Physicians choose treatments based on the entire cache of information at their disposal, and even the most careful prognostic factor adjustment cannot fully correct for the referral pattern for nuclear testing.7 This referral bias likely explains the increased prevalence of viable myocardium (57.1%) in the manuscripts with patients in both the viable and nonviable groups compared to values reported in other studies (36%– 39%).33,34 These patients with a higher likelihood of viable myocardium were likely referred for viability imaging in greater proportions, a situation avoided with randomized trials. The higher levels of viable myocardium seen in this study cohort could confound the interaction between viability and treatment allocation. The referral pattern for nuclear perfusion imaging in patients with left-ventricular dysfunction and ⱖ1 vessel with at least a 75% stenosis favors patients with higher ejection fractions, lower mitral regurgitation grades, and a lower incidence of 3-vessel disease. It is possible that the beneficial effects attributed to viability actually reflect the bias against perfusion imaging for patients with more severe coronary artery disease and left-ventricular dysfunction.7 Moreover, the patients with viability present in the 9 studies examining all 4 cohorts received revascularization at a higher rate than those with nonviable tissue (47.8% vs 33.7%). It is likely that this disproportion indicates additional unidentified characteristics beneficial for survival in the revascularization population with viability present.

American Heart Journal Volume 146, Number 4

Bourque et al 625

Table IV. Follow-up and mortality in nuclear viability studies on patients with ischemic heart failure Viable PCI/CABG

Author Eitzman (’92) (12) Pasquet (’99) (19)* Senior (’99) (18) Anselmi (’98) (17) Afridi (’98) (16) Vom Dahl (’97) (15) DiCarli (’94) (14) Yoshida (’93) (13) Lee (’94) (1)

Viable medical

Nonviable PCI/CABG

Nonviable medical

F/U duration (m ⴞ SD)

No. pts

Deaths

No. pts

Deaths

No. pts

Deaths

No. pts

Deaths

12 33 ⫾ 10 40 ⫾ 17 16 ⫾ 11 18 ⫾ 10 29 ⫾ 6 13.6 36 ⫾ 4 17 ⫾ 9

26 58 31 64 85 57 26 20 49

1 5 1 4 5 2 3 2 4

18 16 32 52 119 14 17 5 21

6 4 10 4 24 2 4 0 3

14 36 6 25 30 27 17 4 19

0 9 3 4 5 2 1 2 1

24 27 18 61 84 63 33 6 40

2 7 8 6 17 7 6 3 15

* Results provided for thallium data only.

These confounding effects cannot be adequately resolved in observational literature. An additional significant limitation is that most of the studies analyzed were retrospective in nature, rendering any standardization and control over variable selection and quality impossible. Meta-analysis not only fails to address these flaws, but can accentuate them, combining multiple heterogeneous studies with inherent weaknesses. Other aspects of the design followed this trend of heterogeneity and limit our use of these studies. Both our analysis and that from Allman et al were unable to control for the imaging protocols or the definitions of viability, revascularization, and medical therapy that were used. Imaging protocols change the sensitivity and specificity of the tests significantly. The definition of viability varied in each study, with some utilizing viability indices and others using the number of affected segments. Some studies defined revascularization as coronary artery bypass grafting only, while others included both percutaneous coronary intervention and coronary artery bypass grafting. There was minimal documentation of medical therapy in these studies, and pharmacologic advances such as statin and ␤-blocker therapy were likely underutilized.35,36 It is possible that different conclusions would be reached if these critical criteria were all defined and applied uniformly. Analysis of these studies was also misleading due to the deficiencies of the selection criteria utilized within the individual studies. The population of interest included patients with left ventricular dysfunction and severe coronary artery disease. Despite this definition, 9 of the 24 studies included patients with relatively little or no LV systolic dysfunction (ejection fractions ⬎50%), while 13 of the 24 studies defined potentially noncontributory epicardial blockages (⬍70%) as signifi-

Figure 1

Graph showing the interaction odds ratio for each of the 9 relevant studies. Values ⱖ1 represent the presence of an interaction with a greater odds ratio of treatment allocation effect on mortality when myocardial viability is present. For values ⬍1 there is an interaction with a lesser odds. When the value equals 1, there is no interaction present. For each study, the 95% CI is represented by the horizontal line.

cant. Two studies looked at subjects who were postmyocardial infarction rather than limiting the analysis to those with heart failure. Moreover, the representation of women and minorities was insufficient, and several studies have shown differences in these groups with respect to course of disease and response to treatment.37 Full generalizability mandates greater in-

American Heart Journal October 2003

626 Bourque et al

clusion of these population subsets. All these limitations and the low event rates within each study weaken the accuracy of the study conclusions.10

5.

Conclusion A previous meta-analysis on the effects of viability status and treatment allocation on long-term outcome in patients with left ventricular dysfunction and significant coronary artery disease did not directly test for an interaction between treatment type and viability status, and included studies that did not provide sufficient data for this critical test. Nonetheless, our analysis utilizing interaction methodology shows a statistically significant interaction between these variables, with the difference in treatment between revascularization and medical therapy having the largest effect on longterm mortality in patients with viable myocardium. However, several limitations of the analysis and primary studies, the multiple levels of bias present, and the marked reduction in effect size from Allman et al’s results reduce the strength and potentially the direction of the derived conclusions. Meta-analysis of observational studies is insufficient to answer whether viability predicts improved long-term prognosis with revascularization versus medical therapy in this population. We lack enough data to modify our current clinical practice. A randomized, controlled trial would eliminate many of the limitations and biases imposed by observational analysis and thus should be undertaken. The Surgical Treatment for Ischemic Heart Failure (STICH) trial has started enrollment and will provide a rigorous answer to this question. The data currently available indicate a possible effect of viability on the interaction between treatment allocation and long-term mortality, with suggested improvement for those with viability treated with revascularization. However, the limitations of the current literature are sufficiently significant to render a definitive conclusion impossible until a large, multicenter, randomized-controlled trial such as STICH provides more rigorous results.

6.

7.

8. 9.

10.

11. 12.

13.

14.

15.

16.

References 1. Lee KS, Marwick TH, Cook SA, et al. Prognosis of patients with left ventricular dysfunction, with and without viable myocardium after myocardial infarction: relative efficacy of medical therapy and revascularization. Circulation 1994;90:2687–94. 2. vom D, Eitzman DT, al Aouar ZR, et al. Relation of regional function, perfusion, and metabolism in patients with advanced coronary artery disease undergoing surgical revascularization. Circulation 1994;90:2356 – 66. 3. Marin-Neto JA, Dilsizian V, Arrighi JA, et al. Thallium scintigraphy compared with 18F-fluorodeoxyglucose positron emission tomography for assessing myocardial viability in patients with moderate versus severe left ventricular dysfunction. Am J Cardiol 1998; 82:1001–7. 4. Gioia G, Milan E, Giubbini R, et al. Prognostic value of tomographic rest-redistribution thallium 201 imaging in medically

17.

18.

19.

20.

treated patients with coronary artery disease and left ventricular dysfunction. J Nucl Cardiol 1996;3:150 – 6. Bax JJ, Poldermans D, Elhendy A, et al. Improvement of left ventricular ejection fraction, heart failure symptoms and prognosis after revascularization in patients with chronic coronary artery disease and viable myocardium detected by dobutamine stress echocardiography. J Am Coll Cardiol 1999;34:163–9. Marwick TH, Zuchowski C, Lauer MS, et al. Functional status and quality of life in patients with heart failure undergoing coronary bypass surgery after assessment of myocardial viability. J Am Coll Cardiol 1999;33:750 – 8. Bourque JM, Velazquez EJ, Borges-Neto S, et al. A reappraisal of myocardial viability testing in patients with coronary artery disease and left ventricular dysfunction: can meta-analysis be trusted? HeartDrug 2002;2:204 –5. Sackett DL, Rosenberg WM. On the need for evidence-based medicine. J Public Health Med 1995;17:330 – 4. Verhagen AP, de V, de B, et al. The Delphi list: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol 1998;51:1235– 41. Allman KC, Shaw LJ, Hachamovitch R, et al. Myocardial viability testing and impact of revascularization on prognosis in patients with coronary artery disease and left ventricular dysfunction: a meta-analysis. J Am Coll Cardiol 2002;39:1151– 8. Cooper HM, Hedges LV. Statistical methods for meta-analysis. Orlando: Academic Press; 1985. p. 200 –3. Eitzman D, al Aouar Z, Kanter HL, et al. Clinical outcome of patients with advanced coronary artery disease after viability studies with positron emission tomography. J Am Coll Cardiol 1992;20: 559 – 65. Yoshida K, Gould KL. Quantitative relation of myocardial infarct size and myocardial viability by positron emission tomography to left ventricular ejection fraction and 3-year mortality with and without revascularization. J Am Coll Cardiol 1993;22:984 –97. Di Carli MF, Davidson M, Little R, et al. Value of metabolic imaging with positron emission tomography for evaluating prognosis in patients with coronary artery disease and left ventricular dysfunction. Am J Cardiol 1994;73:527–33. vom DJ, Altehoefer C, Sheehan FH, et al. Effect of myocardial viability assessed by technetium-99m-sestamibi SPECT and fluorine18-FDG PET on clinical outcome in coronary artery disease. J Nucl Med 1997;38:742– 8. Afridi I, Grayburn PA, Panza JA, et al. Myocardial viability during dobutamine echocardiography predicts survival in patients with coronary artery disease and severe left ventricular systolic dysfunction. J Am Coll Cardiol 1998;32:921– 6. Anselmi M, Golia G, Cicoira M, et al. Prognostic value of detection of myocardial viability using low-dose dobutamine echocardiography in infarcted patients. Am J Cardiol 1998;81:21– 8G. Senior R, Kaul S, Lahiri A. Myocardial viability on echocardiography predicts long-term survival after revascularization in patients with ischemic congestive heart failure. J Am Coll Cardiol 1999;33: 1848 –54. Pasquet A, Robert A, D’Hondt AM, et al. Prognostic value of myocardial ischemia and viability in patients with chronic left ventricular ischemic dysfunction. Circulation 1999;100:141– 8. Tamaki N, Kawamoto M, Takahashi N, et al. Prognostic value of an increase in fluorine-18 deoxyglucose uptake in patients with myocardial infarction: comparison with stress thallium imaging. J Am Coll Cardiol 1993;22:1621–7.

American Heart Journal Volume 146, Number 4

21. Williams MJ, Odabashian J, Lauer MS, et al. Prognostic value of dobutamine echocardiography in patients with left ventricular dysfunction. J Am Coll Cardiol 1996;27:132–9. 22. Petretta M, Cuocolo A, Bonaduce D, et al. Incremental prognostic value of thallium reinjection after stress-redistribution imaging in patients with previous myocardial infarction and left ventricular dysfunction. J Nucl Med 1997;38:195–200. 23. Di Carli MF, Maddahi J, Rokhsar S, et al. Long-term survival of patients with coronary artery disease and left ventricular dysfunction: implications for the role of myocardial viability assessment in management decisions. J Thorac Cardiovasc Surg 1998;116:997– 1004. 24. Cuocolo A, Petretta M, Nicolai E, et al. Successful coronary revascularization improves prognosis in patients with previous myocardial infarction and evidence of viable myocardium at thallium-201 imaging. Euro J Nucl Med 1998;25:60 – 8. 25. Huitink JM, Visser FC, Bax JJ, et al. Predictive value of planar 18F-fluorodeoxyglucose imaging for cardiac events in patients after acute myocardial infarction. Am J Cardiol 1998;81:1072–7. 26. Beanlands RS, Hendry PJ, Masters RG, et al. Delay in revascularization is associated with increased mortality rate in patients with severe left ventricular dysfunction and viable myocardium on fluorine 18-fluorodeoxyglucose positron emission tomography imaging. Circulation 1998;98:II51– 6. 27. Gioia G, Powers J, Heo J, et al. Prognostic value of rest-redistribution tomographic thallium-201 imaging in ischemic cardiomyopathy. Am J Cardiol 1995;75:759 – 62. 28. Pagley PR, Beller GA, Watson DD, et al. Improved outcome after coronary bypass surgery in patients with ischemic cardiomyopathy and residual myocardial viability. Circulation 1997;96:793– 800.

Bourque et al 627

29. Meluzin J, Cerny J, Frelich M, et al. Prognostic value of the amount of dysfunctional but viable myocardium in revascularized patients with coronary artery disease and left ventricular dysfunction. J Am Coll Cardiol 1998;32:912–20. 30. Dreyfus GD, Duboc D, Blasco A, et al. Myocardial viability assessment in ischemic cardiomyopathy: benefits of coronary revascularization. Ann Thoracic Surg 1994;57:1402–7. 31. Haas F, Haehnel CJ, Picker W, et al. Preoperative positron emission tomographic viability assessment and perioperative and postoperative risk in patients with advanced ischemic heart disease. J Am Coll Cardiol 1997;30:1693–700. 32. Smart SC, Dionisopoulos PN, Knickelbine TA, et al. Dobutamineatropine stress echocardiography for risk stratification in patients with chronic left ventricular dysfunction. J Am Coll Cardiol 1999; 33:512–21. 33. Inglese E, Brambilla M, Dondi M, et al. Assessment of myocardial viability after thallium-201 reinjection or rest-redistribution imaging: a multicenter study: the Italian group of nuclear cardiology. J Nucl Med 1995;36:555– 63. 34. al Mohammad A, Mahy IR, Norton MY, et al. Prevalence of hibernating myocardium in patients with severely impaired ischaemic left ventricles. Heart 1998;80:559 – 64. 35. Cleland JG. Improving patient outcomes in heart failure: evidence and barriers. Heart 2000;84(1 Suppl):i8 –10. 36. Horan M, Barrett F, Mulqueen M, et al. The basics of heart failure management: are they being ignored? Euro J Heart Failure 2000; 2:101–5. 37. Mendes D, Dilillo V, Czajkowski S, et al. Psychosocial characteristics after acute myocardial infarction: the ENRICHD pilot study. Enhancing Recovery in Coronary Heart Disease. J Cardiopulmonary Rehab 2000;21:353– 62.