Myocardial perfusion (MP) imaging during stress myocardial contrast echocardiography (MCE) improves the detection of coronary artery disease (CAD). However, its prognostic value to predict cardiac events in patients with known or suspected CAD is still undefined.
A search was conducted for single- or multicenter prospective studies that evaluated the prognostic value of stress MCE in patients with known or suspected CAD. A database search was performed through June 2015. Effect sizes of relative risk ratios (RRs) with their corresponding 95% CIs were used to evaluate the association between the occurrence of total cardiac events (cardiac death, nonfatal myocardial infarction, coronary revascularization) and hard cardiac events (cardiac death and nonfatal myocardial infarction) in subjects with normal and abnormal MP measured by MCE. The Cochran Q statistic and the I 2 statistic were used to assess heterogeneity.
A comprehensive literature search of the MEDLINE, Google Scholar, Cochrane, and Embase databases identified 11 studies enrolling a total of 4,045 patients. The overall analysis of RRs revealed that patients with abnormal MP were at higher risk for total cardiac events compared with patients with normal MP (RR, 5.58; 95% CI, 3.64–8.57; P < .001), with low heterogeneity among trials ( I 2 = 48.15%, Q = 7.71, P = .103). Similarly, patients with abnormal MP were at higher risk for hard cardiac events compared with patients with normal MP (RR, 4.99; 95% CI, 1.75–14.32; P = .003), with significant heterogeneity among trials ( I 2 = 81.48%, Q = 21.59, P < .001).
The results of this meta-analysis suggest that MP assessment using stress MCE is an effective prognostic tool for predicting the occurrence of cardiac events in patients with known or suspected CAD.
Coronary artery disease (CAD) is the most common type of heart disease and a leading cause of death worldwide. In 2013 alone, cardiovascular diseases resulted in 17.3 million deaths worldwide, with ischemic heart disease being responsible for 8.14 million deaths (47.1%). Some patients with CAD are in chronic stable condition, but unstable CAD (e.g., acute myocardial infarction or unstable angina) can lead to acute coronary syndrome (ACS), a spectrum of conditions that are due to an abrupt reduction in coronary blood flow. Hence, the development and optimization of diagnostic and prognostic tools for CAD and ACS is of utmost clinical importance.
The detection of myocardial ischemia has high prognostic value in patients with known or suspected CAD. Myocardial ischemia may cause a wall motion (WM) abnormality, which can be detected using a stress echocardiography. WM analysis has been a well-documented technique for the detection of CAD, and the absence of WM abnormalities is associated with reduced risk for cardiac events. Myocardial contrast echocardiography (MCE) exploits gas-filled microspheres that are inert, remain within the vascular space, and possess an intravascular rheology similar to that of red blood cells. Contrast agent microbubbles improve image quality and are approved for clinical use to increase the accuracy of WM analysis during rest and stress echocardiography. Additionally, because of their biophysical properties, these contrast agents allow accurate and simultaneous assessment of WM and myocardial perfusion (MP), improving sensitivity for the detection of CAD. MP abnormalities can be found in patients with normal WM, and it was demonstrated that MP analysis may possess higher prognostic power than analysis of WM alone. In particular, abnormal MP is a stronger predictor of total cardiac events than abnormal WM, and abnormal MP is associated with a significantly worse outcomes, irrespectively of WM responses.
Several studies have evaluated the diagnostic and prognostic potential of MP imaging by stress MCE. However, the prognostic role of MP abnormalities during stress MCE for prediction of cardiac events is not fully established, and no relevant systematic review and/or meta-analysis has been reported. The aim of this study was to further evaluate the prognostic value of MP assessed during stress MCE in patients with known or suspected CAD and patients with suspected ACS.
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for systematic reviews of observational and diagnostic studies. We searched the published literature using the MEDLINE, Google Scholar, Cochrane, Embase databases with various combinations of following keywords: “coronary artery disease,” “myocardial ischemia,” “acute coronary syndrome,” “angina pectoris,” “myocardial infarction,” “myocardial,” “contrast,” “echocardiography,” “perfusion,” and “stress.” Reference lists in relevant publications were hand searched to identify additional eligible trials. The search analyzed original literature published up to June 30, 2015.
For this meta-analysis, we used the following inclusion criteria: (1) studies were prospective, (2) subjects were patients with known or suspected CAD and patients with suspected ACS, (3) MP was visually analyzed by experienced readers using stress MCE, and (4) the end point included total cardiac events (cardiac death, nonfatal acute myocardial infarction, or coronary revascularization) and hard cardiac events (cardiac death, nonfatal acute myocardial infarction). Recruited patients provided written informed consent agreeing to undergo periodic follow-up. Baseline data were obtained by stress MCE, and the incidence of cardiac events was monitored through follow-up visits. We excluded crossover studies, retrospective studies, letters, comments, editorials, case reports, publications in languages other than English, as well as studies that used resting MCE (without stress) for the assessment of MP.
Study Selection and Data Extraction
Data were extracted independently by two reviewers. A third reviewer was consulted in case of disagreements. We extracted data on study population (number, age, and gender of subjects in each group), study design, patients’ clinical status (diagnosis and left ventricular ejection fraction), the presence of coronary risk factors (diabetes mellitus, hypertension, dyslipidemia, and family history of CAD), and smoking status. Additionally, technical characteristics of MCE, criteria of abnormal MP, end point definition, and information regarding MP image analysis were retrieved from the articles.
Two independent reviewers evaluated the risk for potential biases in the included studies as proposed by Hayden et al . A Quality in Prognosis Studies tool was used, and the following criteria were assessed: (1) study participation, (2) study attrition, (3) prognostic factor measurement, (4) outcome measurement, (5) confounding measurement and account, and (6) analysis. A third reviewer arbitrated on disagreements. The included studies were assigned a value of “low risk,” “high risk,” or “unclear risk of bias” after review.
A total of 11 studies were analyzed. The primary outcome measure was either total cardiac events or hard cardiac events at the last follow-up visit after MCE. The effect size and relative risk for total or hard cardiac event occurrence between patients with abnormal and normal MP were extracted during full-text review either as hazard ratio (HR) or odds ratio (OR) with corresponding 95% CI. The combined effect and pooled relative risk ratio (RR) (including HR and OR) with 95% CI were calculated for all studies combined. RR > 1, RR < 1, and RR = 1 indicate that patients with abnormal MP might be, respectively, at higher, lower, or similar risk for total cardiac events or hard cardiac events compared with patients with normal MP. Assessment of the consistency (homogeneity) of the results across studies is essential for a meta-analysis. The test seeks to determine whether the included studies had variation in the true effect (heterogeneity), or whether the variation in findings was the result of random error (homogeneity). A χ 2 test for homogeneity was conducted, and the inconsistency index ( I 2 ) and Q statistic were determined. A random-effects model (DerSimonian-Laird method) was used when the I 2 statistic was >50%, and fixed-effect models (Mantel-Haenszel method) were used otherwise. Sensitivity analysis was conducted using a leave-one-out approach. A two-sided P value < .05 was considered to indicate statistical significance. Publication bias analysis was not performed, because the number of studies included in the meta-analysis was <10. Data were arranged in Microsoft Excel 2007 (Microsoft, Redmond, WA), and all analyses were performed using Comprehensive Meta-Analysis version 2.0 (Biostat, Englewood, NJ).
Basic Characteristics of Included Studies
Using the keyword-based search, we initially identified 852 publications. Exclusion of irrelevant articles left 34 studies for full-text review. Of these, four were nonrandomized trials, six did not use MCE in all participants, nine either did not provide MP data or the data were incomplete, and four did not report the rate of total or hard cardiac events. Thus, we identified 11 eligible publications. A flowchart describing selection of the trials for the analysis is presented in Figure 1 . The studies recruited a total of 4,045 patients, ranging from 51 to 1,252 patients per study. Demographics and clinical characteristics of the patients are summarized in Table 1 . Patients’ age ranged from 58.5 to 78 years.
|Study||Patients ( n )||Age (y), mean ± SD||Men (%)||Diagnosis of patients||Reduced LVEF (<50%) (%)||DM (%)||HT (%)||Prior revascularization (%)||History of AMI (%)||Family history of CAD (%)||Smokers (%)||Length of follow-up|
|Shah et al . (2015)||197||66 ± 11||73||Cardiology patients referred for evaluation of CAD||20||32||64||47||44||NA||24||17 ± 7 mo|
|Gaibazzi et al . (2013)||723||65 ± 11||61.6||Chest pain with suspected or known CAD||24||28||68||31||19||33||23||16.5 mo (range, 8–22 mo)|
|Mattoso et al . (2013)||227||58 ± 8||52||Known or suspected CAD||NA||34||89.9||11||18||NA||22.5||32 mo (range, 5 d to 6.9 y)|
|Gaibazzi et al . (2012)||1,252||66 ± 11||60||Known or suspected CAD||26||26||71||29||23||29||24||25 mo (range, 6–48)|
|Gaibazzi et al . (2011)||545||67 ± 11||58||Suspected ACS but nondiagnostic ECG findings and normal troponin levels||NA||24||77||31||28||22||25||Median, 12 mo|
|Dawson et al . (2009)||261||Median, 64 (range, 28–90)||50||Known or suspected CAD||23||40||76||NA||38||50||61||14 ± 5 mo|
|Miszalski-Jamka et al . (2009)||84||58.5 ± 9.7||80.9||Known or suspected CAD||NA (58.3 ± 8.2%)||14||73||61||20||55||39||48.3 ± 8.9 mo|
|Tsutsui et al . (2008)||399||78 ± 5 (range, 70–92)||47||Known or suspected CAD||16||26||75||30||19||NA||22||15 mo; but 21 mo for patients with no events|
|Jeetley et al . (2007)||148||61 ± 12||58||Suspected ACS but with nondiagnostic ECG and normal troponin levels||NA||24||70||18||18||NA||37||8 ± 5 mo|
|Basic et al . (2006)||51||60 ± 11||66.7||Known or suggested CAD||NA||18||57||6||NA||16||28||Mean, 29 (range, 6–39)|
|Tsutsui et al . (2005)||158||61 ± 13||51.3||Chest pain and possible ACS||NA (58 ± 11 %)||11||73||36||28||NA||43||Median, 16 mo (range, 6–46) ∗ ( n = 131)|
The characteristics of MCE, type of stress, criteria for abnormal MP, and definition of cardiac events for each individual study are summarized in Table 2 . For the stressors, four studies used dipyridamole (a vasodilator) as a stressor, and one study used dipyridamole or dipyridamole followed by atropine (a muscarinic antagonist). Dobutamine (a sympathomimetic drug) was used as a stressor in one study, two studies used dobutamine with or without atropine, and one study used adenosine (a vasodilator) as a stressor. One study used exercise (bicycle), and the study reported by Jeetley et al . used three different stressors, including treadmill exercise, dipyridamole, and low-dose dobutamine. The heterogeneity in the criteria for abnormal MP in the included studies is shown in Table 2 . For instance, in the study reported by Mattoso et al ., myocardial opacification was qualitatively analyzed by grading visually in the first four to five cardiac cycles after the flash impulse. Reduced myocardial opacification (grade 2) could represent either reduced intensity at all times or delayed appearance of contrast, either compared with other segments or requiring >2 sec for replenishment during stress. In contrast, in the study reported by Gaibazzi et al ., abnormal perfusion after dipyridamole stress was assigned if one or more segments were not fully replenished 1.5 sec after the end of the flash. Rest perfusion was deemed abnormal if replenishment occurred >4 sec after the flash impulse.
|Study||Characteristics of MCE||Criteria for abnormal MP||Definitions of the end points|
|No. of segments||Contrast agents||Stressors||Procedure||Total cardiac events||Hard cardiac events|
|Shah et al . (2015)||17||SonoVue||Dobutamine||Low–mechanical index real-time and triggered mode (end-systolic at every cardiac cycle)||ND; WM remains the cornerstone for the assessment of ischemia, and MP findings were used to assist in reporting WM in patients in whom WM was deemed equivocal||Cardiac death, nonfatal AMI, or coronary revascularization||Cardiac death and nonfatal AMI|
|Gaibazzi et al . (2013)||17||SonoVue||Dipyridamole||Low–mechanical index real-time and triggered mode (end-systolic at every cardiac cycle)||Same as Gaibazzi et al . (2011)||Death, nonfatal MI, and urgent revascularization||Death or nonfatal MI|
|Mattoso et al . (2013)||17||Definity or PESDA||Adenosine||Low–mechanical index real-time MCE||Myocardial opacification was graded visually as grade 1 = intense, grade 2 = reduced, and grade 3 = lack of opacification in the first four to five cardiac cycles after the flash impulse; reduced myocardial opacification (grade 2) could represent either reduced intensity at all times or delayed appearance of contrast either compared with other segments or requiring >2 sec during stress; it was defined as positive if one or more segments exhibited a new grade 2 or 3 defect in one or more myocardial segments at peak stress||Cardiac death, MI, and urgent coronary revascularization||ND|
|Gaibazzi et al . (2012)||17||SonoVue||Dipyridamole or dipyridamole-atropine||Low–mechanical index real-time and triggered mode (end-systolic at every cardiac cycle)||Same as Gaibazzi et al . (2011)||ND||Death or nonfatal MI|
|Gaibazzi et al . (2011)||17||SonoVue||Dipyridamole||Low–mechanical index real-time and triggered mode (end-systolic at every cardiac cycle)||Perfusion was defined as abnormal if the myocardium replenishment was delayed beyond 2 sec after the flash impulse at peak hyperemia; a perfusion defect was scored as fixed or reversible on the basis of its persistence or disappearance at recovery stage; “abnormal” MP in a patient was defined as the presence of either a reversible or a fixed defect in at least one myocardial segment||Cardiac death, MI, and urgent coronary revascularization||Death or nonfatal MI|
|Dawson et al . (2009)||14||Optison||Dipyridamole||Low–mechanical index real-time MCE||Abnormal MBF velocity was defined as incomplete myocardial opacification (reduced or absent) at a pulsing interval > q5beats at rest or > q2 beats during stress||ND||Death or nonfatal MI|
|Miszalski-Jamka et al . (2009)||17||SonoVue||Bicycle||Low–mechanical index real-time MCE||Myocardial contrast opacification was graded using a three-point scale (1 = normal, 2 = reduced, 3 = none) on the basis of relative assessment of myocardial contrast enhancement; a perfusion defect was regarded as present if peak stress myocardial contrast opacification was graded as reduced or none and/or peak stress myocardial contrast replenishment exceeded three cardiac cycles; MP was defined as abnormal when one or more segments exhibited reversible or fixed perfusion defects||Cardiac death, MI, and urgent coronary revascularization||ND|
|Tsutsui et al . (2008)||17||Optison or Definity||Dobutamine ± atropine||Low–mechanical index real-time MCE and triggered mode (end-systolic)||Same as Tsutsui et al . (2005)||ND||Death or nonfatal MI|
|Jeetley et al . (2007)||17||SonoVue||Dipyridamole or treadmill exercise; low-dose dobutamine was used to elicit hyperemia||Low–mechanical index real-time MCE and triggered mode (end-systolic)||Any myocardial segment that did not fill within 1–2 sec after administration of vasodilator stress was considered to demonstrate a reversible perfusion defect; a perfusion defect at rest that remained unchanged at stress was considered a fixed defect; a positive myocardial contrast echocardiogram was defined as the presence of a fixed defect with no previous AMI)or the presence of a reversible defect at stress in one or more myocardial segment||Cardiac death, nonfatal AMI, or coronary revascularization||Cardiac death and nonfatal AMI|
|Basic et al . (2006)||12||Optison or Definity||Dipyridamole||Low–mechanical index real-time and triggered mode (end-systolic at every cardiac cycle)||0 = with normal perfusion; 1 = with a mild reduction in perfusion of one myocardial segment; 2 = with a moderate to severe reduction in perfusion of one myocardial segment, or a mild reduction in perfusion of two or more myocardial segments; 3 = with a moderate to severe reduction in perfusion of at least two myocardial segments||Cardiac death, cardiac arrest, MI, hospitalization for unstable angina or congestive heart failure, or coronary revascularization||ND|
|Tsutsui et al . (2005)||17||Optison or Definity||Dobutamine ± atropine||Low–mechanical index real-time MCE||MPA was considered positive for ischemia when two or more contiguous segments failed to exhibit contrast enhancement at peak stress compared with other segments at the same depth in the same view, and compared with contrast enhancement in the same segments at baseline using a side-by-side image analysis||Cardiac death, nonfatal AMI, coronary revascularization, unstable angina||ND|
Table 3 summarizes outcomes for myocardial contrast echocardiographic image analysis, the number of cardiac events in patients with different outcomes, and multivariate analysis of MP and WM abnormalities for the risk for cardiac events. Data from the included studies suggested that patients with abnormal MP had a higher proportion (7.0%–59%) of cardiac events than those with normal MP (0.8%–10%). Patients with abnormal MP but normal WM had cardiac event rates ranging from 7.0% to 31.0%, and patients with abnormalities in both MP and WM had cardiac event rates ranging from 10.9% to 33.0%. The relative risks (HR and OR) derived from multivariate analysis were used for the following meta-analysis.
|Study||Outcome of image analysis||Cardiac event rate||Outcome of multivariate analysis of risk for cardiac event ∗|
|Normal MP ( n )||Abnormal MP ( n )||Normal WM ( n )||Abnormal WM ( n )||Normal MP||Abnormal MP||HR/OR for abnormal MP||HR/OR for abnormal WM|
|Shah et al . (2015)||111||86||52||20||Normal WM: 10%||Normal WM: 29% |
Abnormal WM: 28%
|HR, 4.41; 95% CI, 1.37–14.20; P = .02 (total cardiac events)||HR, 0.61; 95% CI, 0.20–1.84; P = .38 (total cardiac events)|
|Gaibazzi et al . (2013)||520||198||589||129||Normal WM: 2.1% |
Abnormal WM: 2.2%
|Normal WM: 15.9% |
Abnormal WM: 20.0%
|HR, 5.97; 95% CI, 3.02–11.78; P < .001 (hard cardiac events)||HR, 0.72; 95% CI, 0.36–1.46; P = .37 (hard cardiac events)|
|Mattoso et al . (2013)||142||85||172||55||NA||NA||P = .729 (total cardiac events)||P = .995 (total cardiac events)|
|Gaibazzi et al . (2012)||846||222||983||85||Normal WM: 2.0%||Normal WM: 7.0% |
Abnormal WM: 11.8%
|HR, 6.70; 95% CI, 1.89–23.77; P = .003 (hard cardiac events)||HR, 4.89; 95% CI, 1.89–23.77; P = .001 (hard cardiac events)|
|Gaibazzi et al . (2011)||350||195||444||101||Normal WM: 0.8%||Normal WM: 10.7% |
Abnormal WM: 10.9%
|HR, 10.54; 95% CI, 3.30–33.64; P < .001 (total cardiac events)||HR, 1.04; 95% CI, 0.44–2.44; P = .936 (total cardiac events)|
|Dawson et al . (2009)||175||86||NA||NA||Normal WM: 1.2%||Normal WM: 11.6% |
Abnormal WM: 26.0%
|OR, 23; 95% CI, 6–201; P < .001 (hard cardiac events)||NA|
|Miszalski-Jamka et al . (2009)||30||54||46||38||Normal WM: 7.0%||Normal WM: 31.0% |
Abnormal WM: 33.0%
|HR, 6.79; 95% CI, 2.02–22.82; P = .001 (total cardiac events)||NA|
|Tsutsui et al . (2008)||216||183||270||129||4.2%||25%||OR, 0.867; 95% CI, 0.372–2.018 (Wald χ 2 = 7.5); P = .006 (hard cardiac events)||OR, 0.215; 95% CI, 0.072–0.648 (Wald χ 2 = 1.98); P = .159 (hard cardiac events)|
|Jeetley et al . (2007)||118 (low risk)||27 (high risk)||NA||NA||7% (low risk/normal)||59% (high risk/abnormal)||OR, 20; 95% CI, 7–57; P < .001 (total cardiac events)||NA|
|Basic et al . (2006)||30||21||NA||NA||0%||10%||NA||NA|
|Tsutsui et al . (2005)||82||76||108||50||10%||31%||HR, 3.23; 95% CI, 1.23–8.49; P = .018 (total cardiac events)||HR, 0.79; 95% CI, 0.27–2.34; P = .67 (total cardiac events)|