Supplement Table 1: A listof exact search terms and detailed search strategies
1. viability.mp.
2. viable.mp.
3. exp Magnetic Resonance Imaging/ or Magnetic Resonance Imaging.mp.
4. exp Thallium/ or thallium.mp.
5. exp Technetium/ or Technetium.mp.
6. exp Tomography, Emission-Computed, Single-Photon/
7. exp Positron-EmissionTomography/or Positron-EmissionTomography.mp.
8. exp Echocardiography/ or echocardiography.mp.
9. or/1-8
10. exp Myocardial Revascularization/ or revascularization.mp.
11. exp Coronary Artery Bypass/or coronary artery bypass.mp.
12. exp Angioplasty, Transluminal, Percutaneous Coronary/ or angioplasty.mp.
13. or/10-12
14. exp Myocardial Infarction/ or myocardial infarction.mp.
15. exp Death/ or death.mp.
16. exp Mortality/ or mortality.mp.
17. or/ 14-16
18. exp cohort studies/
19. cohort$.tw.
20. controlled clinical trial.pt.
21. epidemiologic methods/
22. limit 21 to yr=1966-1989
23. exp case-control studies/
24. (case$ and control$).tw.
25. or/18-10,22-24
26. animal/ not human/
27. 25 not 26
28. and/9,13,17,27
Detailed search strategies
We searched Ovid MEDLINE, Ovid MEDLINE In-Process and other non-indexed citations, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, and conference proceedings from the American College of Cardiology and the American Heart Association through December 2008 using Medical Subject Headings (MeSH) and text words. The search was limited to human studies and combined with the cohort and case-control study filter.We also searched bibliographies of retrieved articles and relevant reviews. If the publications did not contain the full information necessary for meta-analysis, we wrote to authors of identified studies to obtain the missing information.
Supplement Table 2. Patient and Study Characteristics of Included Studies
Eitzman D et al.9 / 82 / 87 / 59(10) / 34(13) / 15 / FDG/NH3 / MM / NA
Yoshida K et al.10 / 35 / 62 / 54(10) / 44(12) / NR / FDG/Rb / FDG/Rb uptake / NA
Lee KS et al.11 / 129 / 79 / 62(11) / 38(11) / 26 / FDG/Rb / MM / NA
vom Dahl J et al.12 / 161 / 89 / 57(9) / 45(12) / 47 / FDG/Tc-99m / MM / NA
Di Carli MF et al.13 / 93 / 83 / 65(10) / 25(7) / 100 / FDG/NH3 / MM / 18 (16)
Zhang X et al.14 / 123 / 92 / 56(9) / 35(6) / 13 / FDG/Tc-99m / MM / 22
Desideri A et al.15 / 261 / 88 / 66(6) / 29(5) / NR / FDG/NH3 / MM / 20 (15)
Zhang X et al.16 / 70 / 94 / 57(10) / 36(8) / 17 / FDG/Tc-99m / MM / NA
Anselmi M et al.17 / 202 / 85 / 59(9) / 33(10) / 38 / LDDE / Improvement / NA
Afridi I et al.18 / 318 / 78 / 64(11) / 27(7) / 31 / HDDE / Improvement, worsening or a biphasic response / 25
Smart SC et al.19 / 168 / 63 / 65(12) / 31(7) / 14 / DASE / Worsening or a biphasic response / 31(17)
Senior R et al.20 / 87 / 90 / 62(10) / 25(9) / 10 / LDDE / Improvement or worsening / 33(30)
Chaudhry FA et al.21 / 80 / 80 / 64(12) / 27(7) / 28 / HDDE / Improvement / 46 (30)
Sicari R et al.22 / 307 / 82 / 60(10) / 28(6) / 33 / Dipiridamole / Improvement or worsening / 25
Sawada SG et al.23 / 139 / 82 / 59(10) / 32(10) / 4 / LDDE / Worsening / NA
Meluzín J et al.24 / 124 / 95 / 57(9) / 25(4) / 12 / LDDE / Improvement or worsening / 24(12)
Sicari R et al.25 / 425 / 85 / 61(9) / 28(6) / 37 / LDDE / Improvement or worsening / 38
Liao L et al.26 / 107 / 82 / 63(10) / 21(4) / 17 / HDDE / Worsening or a biphasic response / 31
Gioia G et al.27 / 85 / 75 / 65(11) / 30(10) / 0 / Tl-201 RR / Uptake ≥50% / 25 (25)
Cuocolo A et al.28 / 76 / 94 / 55(10) / 38(9) / 41 / Tl-201 RR / Uptake ≥50% / 34 (26)
Morse RW et al.29 / 37 / 78 / 62(12) / 30(9) / 0 / Tl-201 RR / Uptake ≥50% or RR / NA
Pasquet A et al.30 / 137 / 84 / 62(10) / 35(12) / 23 / Tl-201 RR/LDDE / Uptake ≥50% or RR / NA
Shapira I et al.31 / 40 / 82 / 64(12) / NR / 0 / Tl-201 RR / Uptake ≥50% or RR / 24
Sciagrà R et al.32 / 105 / 90 / 60(10) / 33(9) / 46 / Tc-99m with Nitrates / Nitrate augumentation / 36 (22)
Podio V et al.33 / 153 / 88 / 59(9) / 45(11) / 45 / Tl-201 RR / RR / NA
Senior R et al.34 / 56 / 89 / 64(9) / 25(9) / 13 / Tl-201 RR / Uptake ≥50% or RR / 42
Petrasinovic Z et al.35 / 55 / 90 / 58(9) / 43(10) / NR / Tl-201 RR / Uptake ≥50% or RR / NA
He ZX et al.36 / 78 / 89 / 55(10) / 38(8) / 0 / Tc-99m with Nitrates / Nitrate augumentation / NA
Acampa W et al.37 / 272 / 89 / 53(12) / 37(7) / 42 / Tc-99m at Rest / Uptake ≥55% / 36(20)
Data are expressed as mean (SD).
CR denotes contractile reserve; DASE, dobutamine-atropine stress echocardiography; Echo, echocardiography; EF, ejection fraction; FDG, fluorine-18 fluorodeoxyglucose; HDDE, high dose dobutamine echocardiography; LDDE, low-dose dobutamine echocardiography; MM, FDG-perfusion mismatch; NA, not applicable; NH3, nitrogen-13; NR, not reported; PCI, percutaneous coronary intervention; PET, positron emission tomography; Rb, rubidium-82; RI, reinjection; RR, rest-redistribution; SD standard deviation; Tc-99m, technetium-99m; and Tl-201, thallium-201.
Supplement Table 3. Quality Assessment of the Included Studies
Author / Prospective vs. retrospective / Consecutive sampling / Potential confounders / Adjustment for the confounders / Blind assessment of the test results / Appropriate sample sizeEitzman D et al.9 / Retrospective / No / Age, EF / No / Yes / No
Yoshida K et al.10 / Unclear / Yes / Unclear / No / Yes / No
Lee KS et al.11 / Retrospective / Unclear / Age, EF / Yes / Yes / No
vom Dahl J et al.12 / Prospective / Yes / No / No / Unclear / No
Di Carli MF et al.13 / Unclear / Yes / No / No / Yes / No
Zhang X et al.14 / Unclear / Unclear / Age / Yes / Unclear / No
Desideri A et al.15 / Prosepctive / Yes / No / Yes / Yes / Yes
Zhang X et al.16 / Retrospective / No / Age / Yes / Yes / No
Anselmi M et al.17 / Retrospective / Yes / No / No / Yes / Yes
Afridi I et al.18 / Unclear / Unclear / Age / Yes / Unclear / Yes
Smart SC et al.19 / Unclear / Yes / EF / Yes / Yes / Yes
Senior R et al.20 / Prospective / Yes / No / Yes / Yes / No
Chaudhry FA et al.21 / Prospective / Yes / No / Yes / Yes / No
Sicari R et al.22 / Prospective / Yes / No / Yes / Yes / Yes
Sawada SG et al.23 / Unclear / Unclear / Age / Yes / Yes / Yes
Meluzín J et al.24 / Unclear / Yes / No / No / Yes / No
Sicari R et al.25 / Prospective / Yes / Age / Yes / Unclear / Yes
Liao L et al.26 / Prospective / Yes / Age, EF / Yes / Yes / No
Gioia G et al.27 / Retrospective / No / Age / No / Unclear / No
Cuocolo A et al.28 / Prospective / Yes / No / Yes / Unclear / No
Morse RW et al.29 / Retrospective / No / No / Yes / Yes / No
Pasquet A et al.30 / Prospective / Yes / Unclear / Yes / Yes / No
Shapira I et al.31 / Retrospective / No / Unclear / No / Yes / No
Sciagrà R et al.32 / Retrospective / Unclear / Age / Yes / Yes / No
Podio V et al.33 / Retrospective / Yes / Age, EF / Yes / Yes / No
Senior R et al.34 / Prospective / Yes / No / Yes / Yes / No
Petrasinovic Z et al.35 / Unclear / Unclear / Unclear / No / Unclear / No
He ZX et al.36 / Retrospective / No / No / No / Yes / No
Acampa W et al.37 / Unclear / Unclear / No / Yes / Unclear / Yes
EF denotes ejection fraction.
The quality of studies was assessed independently by two reviewers (YI, JAC), using the following criteria:prospective follow up, consecutive sampling of their cohorts, blind interpretation of the test results, balance of potential confounders between two groups, adjustment for the confounders, and appropriate sample size.5 Discrepancies in quality assessment were resolved by consensus among the authors.
Ten studies (34%) were prospective in design and the remaining 19 were retrospective (or unclear). Sixteen studies (55%) described the consecutive sampling of their cohorts. Twenty one studies (72%) clearly reported blind assessment of the test results. Sixteen studies (55%) found the presence of significant confounding variables, such as age in eleven studies (38%) and EF in five studies (17%). Nineteen studies (66%) adjusted the result for potential confounding factors by the multivariate analysis. A post-hoc sample size calculation, to have 80% power to detect the difference of cardiac mortality between patients with revascularization (mortality of 7%) and patients with medical therapy (mortality of 28%) with 5% significance level, suggests that a sample size of 49 patients is needed in each of the study groups. Thus, eight studies (28%) were determined to have appropriate sample sizes.
Supplement Table4. Risk of Bias
Authors / Prospective follow up / Consecutive sampling / Balance of potential confounders / Adjustment for the confounders / Blind assessment of the test results / Appropriate sample sizeEitzman D et al.9 / * / * / * / * / *
Yoshida K et al.10 / * / * / * / *
Lee KS et al.11 / * / * / * / *
vom Dahl J et al.12 / * / * / *
Di Carli MF et al.13 / * / * / *
Zhang X et al.14 / * / * / * / * / *
Desideri A et al.15
Zhang X et al.16 / * / * / * / *
Anselmi M et al.17 / * / *
Afridi I et al.18 / * / * / * / *
Smart SC et al.19 / * / *
Senior R et al.20 / *
Chaudhry FA et al.21 / *
Sicari R et al.22
Sawada SG et al.23 / * / * / *
Meluzín J et al.24 / * / * / *
Sicari R et al.25 / * / *
Liao L et al.26 / * / *
Gioia G et al.27 / * / * / * / * / * / *
Cuocolo A et al.28 / * / *
Morse RW et al.29 / * / * / *
Pasquet A et al.30 / * / *
Shapira I et al.31 / * / * / * / * / *
Sciagrà R et al.32 / * / * / * / *
Podio V et al.33 / * / * / *
Senior R et al.34 / *
Petrasinovic Z et al.35 / * / * / * / * / * / *
He ZX et al.36 / * / * / * / *
Acampa W et al.37 / * / * / *
*suggests the presence of risk of bias.
Supplement Table 5: The pooled summary estimate after adjusting publication bias
Relative risks before adjustment / Relative risks after adjustmentPET / 0.30 (95%CI, 0.17-0.52; p<0.001) / 0.34 (95%CI, 0.18- 0.62; p<0.001)
Stress echocardiography / 0.29 (95%CI, 0.21-0.40; p<0.001) / 0.30 (95%CI, 0.22- 0.41; p<0.001)
SPECT / 0.30 (95%CI, 0.20-0.46; p<0.001) / 0.36 (95%CI, 0.24- 0.53; p<0.001)
PET denotes positron emission tomography; and SPECT, single photon emission computed tomography.
The trim-and-fill method, imputing the missing studies and recalculating the pooled summary estimate, did not change the pooled relative risks significantly.
Supplement Figure Legend
Supplement Figure 1
Title: The optimal threshold values for the presence of viability in the first model
Caption: The optimal threshold value for the presence of viability was estimated in the graph where x-axis represents the amount of viability and y-axis represents the relative risk for revascularization compared with medical therapy on cardiac mortality. It is assumed that there is linear relationship between the amount of viable myocardium and survival benefit of revascularization and that the difference in the amounts of viable myocardium between patients with and without viability is the interquartile range of those. The point where the exponential line, containing the upper confidence interval of the baseline relative risk at the mean values of viable myocardium and with the slope of log relative risk difference between patients with and without viability per 1.35 x SD in amount of viability, intersects the RR = 1 linerepresents the optimal threshold values in this model.
Supplement Figure 2
Title: The optimal threshold values for the presence of viability in the second model
Caption: In the same graph as in Supplement Figure 1, the point where the exponential line, containing the two points of the upper confidence intervals of relative risk estimates both in patients with viability and in patients without viability, intersects with the RR = 1 line represents the optimal threshold values in this model.
Supplement Figure 1
Supplement Figure 2