Author + information
- Received January 16, 2007
- Revision received March 1, 2007
- Accepted March 12, 2007
- Published online July 24, 2007.
- Girish Dwivedi, MD, MRCP, DM,
- Rajesh Janardhanan, MD, MRCP,
- Sajad A. Hayat, MBChB, MRCP,
- John M. Swinburn, MD, MRCP and
- Roxy Senior, MD, DM, FRCP, FESC, FACC⁎ ()
- ↵⁎Reprint requests and correspondence:
Prof. Roxy Senior, Department of Cardiology, Northwick Park Hospital, Harrow, Middlesex, HA1 3UJ, Hon. Professor, Middlesex University, United Kingdom.
Objectives This study sought to determine whether residual myocardial viability determined by myocardial contrast echocardiography (MCE) after acute myocardial infarction (AMI) can predict hard cardiac events.
Background Myocardial viability detected by MCE has been shown to predict recovery of left ventricular (LV) function in patients with AMI. However, to date no study has shown its value in predicting major adverse outcomes in AMI patients after thrombolysis.
Methods Accordingly, 99 stable patients underwent low-power MCE at 7 ± 2 days after AMI. Contrast defect index (CDI) was obtained by adding contrast scores (1 = homogenous; 2 = reduced; 3 = minimal/absent opacification) in all 16 LV segments divided by 16. At discharge, 65 (68%) patients had either undergone or were scheduled for revascularization independent of the MCE result. The patients were subsequently followed up for cardiac death and nonfatal AMI.
Results Of the 99 patients, 95 were available for follow-up. Of these, 86 (87%) underwent thrombolysis. During the follow-up time of 46 ± 16 months, there were 15 (16%) events (8 cardiac deaths and 7 nonfatal AMIs). Among the clinical, biochemical, electrocardiographic, echocardiographic, and coronary arteriographic markers of prognosis, the extent of residual myocardial viability was an independent predictor of cardiac death (p = 0.01) and cardiac death or AMI (p = 0.002). A CDI of ≤1.86 and ≤1.67 predicted survival and survival or absence of recurrent AMI in 99% and 95% of the patients, respectively.
Conclusions The extent of residual myocardial viability predicted by MCE is a powerful independent predictor of hard cardiac events in patients after AMI.
Acute myocardial infarction (AMI) continues to be a significant public health problem in industrialized countries and an increasingly significant problem in developing countries (1). The estimated mortality, although declining, still remains high, and the risk of further cardiac complications such as reinfarction, sudden death, and heart failure for those who survive is substantial especially after thrombolysis (2). A proportion of patients manifest clinical features that are associated with high risk, such as postinfarct angina, heart failure, or hemodynamic instability, and these patients require an aggressive invasive management strategy. The remaining patients who are asymptomatic, however, have a broad spectrum of risk that is not clinically apparent, and need further investigation to elucidate the extent of this risk that is likely to influence subsequent management strategy. Resting left ventricular ejection fraction (LVEF) is a well-recognized and important marker of prognosis. However, in the early postinfarct period, the resting LVEF is likely to underestimate the true contractile potential of the ventricle because of uncoupling of viability and function (3). Myocardial viability determined by dobutamine stress echocardiography (DSE) and radionuclide perfusion imaging have been shown to predict outcome independent of LVEF after AMI (4–6).
Myocardial contrast echocardiography (MCE) is a relatively new bedside technique that can assess myocardial perfusion using microbubbles (7–9). Both animal as well as human studies have indicated that MCE can be used reliably to assess infarct size and hence myocardial viability (8–13). Unlike DSE, MCE can detect myocardial viability at rest, and unlike radionuclide perfusion imaging, MCE has no radiation burden and can be performed at the bedside. Recent studies have indicated that myocardial viability assessed by MCE can predict recovery of left ventricular (LV) function in patients after AMI (8,14). However, to date no study has investigated the value of myocardial viability determined by MCE in predicting prognosis after AMI in a population in which thrombolysis is a major form of acute reperfusion therapy.
Therefore, this study was conducted to determine whether residual myocardial viability determined by MCE after AMI can predict hard cardiac events over and above that predicted by clinical, electrocardiographic, biochemical markers and resting LV function parameters, which are frequently used clinical markers of prognosis.
Consecutive stable patients shortly after their first presentation with AMI were enrolled. The diagnosis of AMI was based on patients satisfying 2 of 3 criteria: 1) typical anginal pain lasting for >30 min; 2) persistent ST-segment elevation electrocardiogram; and 3) a creatine kinase measurement of >2 times the upper limit of normal. The Hospital Ethics Committee approved the protocol. All patients gave informed consent before entering the study.
Within 7 ± 2 days after AMI, all patients underwent simultaneous baseline transthoracic echocardiography and low-power MCE for the estimation of LVEF and LV volumes. The decision to proceed to coronary arteriography and revascularization was based on clinical grounds (medical comorbidity, suitability of coronary anatomy, patients preference, and so on) independent of MCE results.
Resting echocardiography was undertaken using tissue harmonic imaging (Hewlett-Packard Sonos-5500, Best, the Netherlands). The American Society of Echocardiography 16-segment LV model was used for analysis. Wall thickening was graded by transthoracic echocardiography using a 4-point scoring system (1 = normal, 2 = reduced, 3 = absent, 4 = systolic wall thinning). The LVEF, left ventricular end-diastolic volume (LVEDV), and left ventricular end-systolic volume (LVESV) were assessed using the modified Simpson biplane method (15).
The MCE was performed in the 3 standard apical (4-, 2-, and 3-chamber) views using a low-power technique at a mechanical index of 0.1. Background gains were set so that minimal tissue signal was seen, and the focus was set at the level of the mitral valve. For the first 40 patients, Optison (Amersham Health, Little Chalfont, United Kingdom) was given as a slow bolus injection of 0.3 to 0.7 ml followed by a saline flush over 20 s (8). The remaining 59 patients were studied using Sonovue (Bracco Research SA, Geneva, Switzerland); this was administered as an intravenous infusion at 50 to 70 ml/h using a VueJect (BR-INF 100, Bracco Research SA, Geneva, Switzerland) infusion pump. Machine settings were optimized to obtain the best possible myocardial opacification with minimal attenuation. Replenishment was observed for at least 15 cardiac cycles after high mechanical index (1.7) microbubble destruction (8).
A semiquantitative scoring system that has been previously validated in our laboratory (8) was used to assess contrast intensity after microbubble destruction; 1 = homogenous opacification, 2 = heterogenous opacification, 3 = minimal or absent contrast opacification. We have previously performed interobserver and intraobserver variability analysis using the above scores in these patients, and the results are 89% (k = 0.76) and 90% (k = 0.77) respectively using the same technique (13). Contrast defect index (CDI) was obtained by adding contrast scores of all segments and dividing by the total number of evaluable segments. The CDI represented the extent and intensity of residual infarction and thus residual myocardial viability. Percentages of segments with heterogeneous opacification (patchy infarction) and absent contrast opacification (complete infarction) also were obtained for each patient. An experienced reader blinded to the clinical and angiographic details performed analysis of echocardiographic data. The MCE and LV function data were assessed separately.
Selective coronary arteriography was performed with the Judkins approach. Coronary artery disease (CAD) was defined as a >50% luminal diameter narrowing of one or more major epicardial arteries or their major branches assessed qualitatively. The presence of CAD was determined in the left anterior descending (anterior) circulation and the right coronary artery and/or left circumflex (posterior) circulation. Mean coronary luminal diameter stenosis and number of coronary vessels showing significant CAD were calculated.
Outcome and follow-up
Study patients were followed up for at least 6 months in a dedicated clinic where a research fellow and a nurse reviewed study patients for events. The hospital data base was checked and the patient’s general practitioner was contacted in case of nonattendance. Subsequent follow-up data were collected by questionnaires returned from patients with an additional telephone call or hospital record review to further verify events wherever appropriate. The vital status of all screened patients was determined from the hospital patient information system at the termination of the study. Primary study end points were: 1) cardiac death (defined as sudden death caused by AMI or arrhythmia or heart failure); and 2) cardiac death or nonfatal AMI (typical chest pain and increased troponin I). However, if a patient suffered both of the events, then cardiac death was recorded as the primary event.
All categorical variables are expressed as proportions/percentages, and all continuous variables as mean ± SD, except those that are not normally distributed, which are presented as medians with 95% confidence intervals (CIs). Positively skewed variables were log transformed for further analysis. Differences between means were calculated using an unpaired ttest. Cox regression analysis was used to examine the effect of prognostic variables listed in Table 1on time to outcome. However, when all subjects in a particular group had the same outcome, then a Fisher exact test was used. To examine the joint effect of the explanatory variables, multivariate analysis was then performed using Cox models. A backward selection procedure was used to retain only the statistically significant variables. To limit the number of variables, only variables that gave a p value of <0.2 from the univariate analyses were considered for the multivariate analyses. A receiver-operator characteristic analysis was used to test the predictive accuracy of MCE parameters. Optimal cutoff was defined as the threshold where the sum of sensitivity and specificity was maximum. A p value of <0.05 was considered statistically significant.
Patients characteristics and outcome
A total of 99 patients were enrolled, out of which 4 patients were lost to follow-up. The remaining 95 patients were considered for the final analysis. Patient characteristics of these 95 patients are listed in Table 1. Of the 95 patients, 86 (87%) underwent thrombolysis. Mean LVEF was 50% ± 11%. Coronary arteriography was performed in 86 (87%) patients; 77 (87%) patients had significant CAD of the infarct-related artery, and 35 (40%) showed evidence of multivessel disease. At discharge, 65 (68%) patients had either undergone or were scheduled for revascularization independent of the MCE result. The MCE scores were obtained in all patients. The mean time from presentation to revascularization was 46 ± 64 days. The remaining patients, i.e. 30 (32%) patients, were not revascularized on medical grounds. The mean duration of follow-up was 46 ± 16 months. Adverse cardiac events occurred in 15 patients (16%), of whom 8 (8%) died and 7 (8%) had nonfatal AMI.
Prediction of cardiac mortality
Prediction of cardiac mortality is shown in Table 2.Of the prognostic variables outlined in Table 1, the univariable predictors of cardiac death were elevated CDI (hazard ratio [HR] 1.27, 95% CI 1.06 to 1.53, p = 0.006), LVESV (HR 1.25, 95% CI 1.04 to 1.51, p = 0.02), and LVEF (HR 0.46, 95% CI 0.26 to 0.89, p = 0.009). The results indicated that higher LVESV was associated with a greater risk of death at any time. It was found that a 10-U increase in LVESV resulted in the hazard (or risk) of death at any time increasing by 25%. Conversely, higher values of LVEF were associated with a decreased risk of death. A 10-U increase in LVEF resulted in the hazard of death decreasing by over half. Among the MCE parameters, CDI emerged as the strongest univariate predictor. Higher CDI values were associated with an increased hazard of death at any time. A 0.1-U increase in CDI resulted in the hazard of death at any time increasing by 27%. As other parameters of MCE, i.e., percentage of patchy infarction and complete infarction, are strongly interrelated with CDI, which is the strongest univariate predictor, CDI was entered in the multivariate model.
The multivariable predictors of cardiac death were increased age (HR 3.43, 95% CI 1.27 to 9.27, p = 0.02), mean coronary luminal diameter stenosis (HR 2.97, 95% CI 1.12 to 7.81, p = 0.03), and elevated CDI (HR 1.37, 95% CI 1.08 to 1.75, p = 0.01). A 0.1-U increase in CDI resulted in the hazard of death at any time increasing by 37%. A 10-unit increase in mean stenosis was associated with the hazard of death increasing almost threefold, whereas a 10-year increase in age was associated with the hazard of death at any time increasing over 3-fold.
Prediction of cardiac death or nonfatal AMI
Prediction of cardiac death or nonfatal AMI is shown in Table 3.Univariable predictors of cardiac death or nonfatal AMI were LVESV (HR 1.17, 95% CI 1.01 to 1.36, p = 0.04), LVEF (HR 0.59, 95% CI 0.38 to 0.90, p = 0.01) and elevated CDI (HR 1.20, 95% CI 1.05 to 1.36, p = 0.008). Patients with higher LVESV had an increased likelihood of death or nonfatal AMI at any time. A 10-U increase in volume resulted in the hazard of death or nonfatal AMI increasing by 17%. High values of LVEF were associated with a lower risk of death or nonfatal AMI at any time. A 0.1-U increase in CDI increased risk of death or nonfatal AMI by 20%. However, when multivariable analysis was performed, only elevated CDI emerged as a significant predictor (HR 1.26, 95% CI 1.09 to 1.44, p = 0.002) of death or nonfatal AMI.
Receiver-operator characteristic analysis
Receiver-operator characteristic curve analyses showing sensitivity and specificity of different cutoff points with CDI for the predictions of death and death or nonfatal AMI were performed. The areas under the receiver-operator characteristic curve with CDI for the predictions of death (Fig. 1A)and death or nonfatal AMI (Fig. 1B) were 0.82 and 0.80, respectively. A CDI value of 1.86 provided the optimum cutoff, with sensitivity and specificity of 87% and 84% for cardiac death and 62% and 84% for cardiac death or nonfatal AMI, respectively.
Kaplan-Meier survival curve analysis
Kaplan-Meier survival curve analysis was performed using a cutoff CDI of 1.86 for the prediction of cardiac death (Fig. 2A)and a cutoff CDI of 1.67 for the prediction of cardiac death or nonfatal AMI (Fig. 2B). The log rank test showed that patients with CDI >1.86 had a significantly (p = 0.0001) high cardiac mortality (33%), compared with only 1% in patients with CDI ≤1.86. Similarly, patients with a CDI >1.67 had a significantly higher incidence of cardiac death or nonfatal AMI (39%) compared with those with a CDI ≤1.67, who had an event rate of 5% over a 4-year follow-up period. The CDI maintained its ability to predict cardiac events irrespective of the technique used. A CDI ≤1.67 separated patients with and without cardiac events in both the Optison (4% and 38%, respectively) and the Sonovue (5% and 44%, respectively) groups.
Relationship between LVEF, infarct size, and prognosis
The relationship between LVEF, infarct size, and prognosis is shown in Figures 3 and 4.⇓⇓Of the 95 patients, 46 (48%) had LVEF >50%. Of these, the majority, i.e., 37 (80%), had a only small AMI defined as CDI ≤1.67. In this group, 5% of the patients with a small infarction versus 20% with a large infarction suffered cardiac death or nonfatal AMI (p = 0.33). However, in patients with LVEF ≤50%, although none with a small infarction had cardiac event, in patients with large infarctions, the incidence of cardiac events increased significantly to 53% (p = 0.002). This trend was noted even in patients with a lower LVEF, i.e., ≤45% (p = 0.015) and ≤40% (p = 0.08).
This is the first study showing that residual myocardial viability as shown by low-power MCE is an independent predictor of long-term hard cardiac events in patients after AMI. In this population, in which survival was 92% over a 4-year follow-up period, the presence of myocardial viability determined by MCE predicted 99% survival, compared with only 67% survival in patients with no significant myocardial viability. Whereas 39% of patients suffered either cardiac death or nonfatal AMI with no significant myocardial viability, only 5% experienced such events in the presence of significant myocardial viability in a population with a hard cardiac event rate of 16% over 4 years. This study also showed that larger infarctions predict mortality, whereas AMI is more likely to occur in patients with relatively smaller infarctions and hence viable myocardium. It is intuitive that after AMI and thrombolysis, for a given LVEF, infarct size and inducible myocardial ischemia are likely to be the ultimate determinants of outcome.
The LVEF and LVESV, which reflect the extent of infarction burden, were strong univariate predictors of mortality or mortality and AMI, but only the extent of infarction assessed by MCE remained an independent predictor of both outcomes. The LV function early after AMI tends to overestimate the extent of true infarction because of myocardial stunning, which can persist for up to 6 weeks. Thus in our study, MCE, which assesses the extent of infarction more directly than the parameters of LV function, provided the most prognostic information in patients with LV dysfunction. Myocardial contrast echocardiography did not provide significant incremental information over LV function parameters in patients with preserved LVEF, because in these patients most (80%) had only a small infarction, as already predicted by normal LV function. Several previous studies using radionuclide imaging and DSE, both of which also detect the extent of infarction performed early after AMI, have provided superior prognostication compared with LV function, as also shown in our study using MCE (4–6). In our study, the severity of CAD, which is a marker of ischemia, was also an independent predictor of cardiac death.
Mechanism of detection of myocardial viability by MCE
Myocardial contrast echocardiography detects contrast microbubbles at the capillary level within the myocardium (16). When a steady state of microbubble concentration is achieved in the myocardium during a contrast administration, the acoustic signal intensity observed provides a measure of myocardial blood volume fraction. Because 90% of myocardial blood volume fraction comprises capillary blood, a single MCE image provides an assessment of capillary density in the different myocardial regions (16). Contrast signal intensity during MCE has been shown to correlate directly with capillary density and indirectly with fibrosis in biopsied myocardium in patients with chronic heart failure (17). Contrast perfusion assessed at 15 s during destruction-replenishment imaging has been shown to correlate well with ultimate infarct size in animal models of AMI (11,12). After AMI, collateral blood flow has been shown to be generally less than normal flow in areas showing myocardial viability (18,19). Regions with normal blood flow fill within 5 s after destruction imaging. Regions with collateral flow fill later depending on the magnitude of flow. If they do not fill within 15 s, then flow to the region is markedly reduced and will result in necrosis (11). This method of assessment of myocardial viability by MCE has been shown to predict recovery of regional function and transmural extent of infarction, and also has been shown to predict recovery of both regional and global LV function (8,13,14).
Risk stratification after AMI
The most important determinants of outcome in patients who are deemed stable after AMI and thrombolysis are resting LV function, extent and degree of residual myocardial viability, and myocardial ischemia at the site of AMI or remote territory. The MCE is a unique technique because it can address all of the above outcome measures in a single examination. An MCE during stress testing has been shown to accurately detect the severity of CAD in both the stable and the acute setting (20–24). In a recent study after AMI, MCE accurately detected residual infarct related artery and multivessel disease, which also are strong prognostic markers of outcome after AMI (23). The present study confirms the prognostic value of MCE-determined myocardial viability after AMI.
Radionuclide perfusion imaging, DSE, and late gadolinium-enhanced cardiac magnetic resonance are techniques widely used to assess myocardial viability after AMI. However, MCE, compared with radionuclide perfusion imaging, is portable, can be performed at the bedside, allows rapid acquisition of data, and has no radiation burden. Compared with DSE, MCE has been shown to be more sensitive for the detection of myocardial viability after AMI, and unlike DSE can detect myocardial viability at rest (25,26). Although gadolinium-enhanced cardiac magnetic resonance has excellent spatial resolution, a recent study comparing the 2 modalities did not show any significant difference in the prediction of recovery of LV function after AMI (13). Thus, with these advantages over other competitive techniques, MCE may well become the technique of choice for the assessment of myocardial viability after AMI. In a recent study, after only anterior AMI patients who underwent primary percutaneous coronary intervention, MCE predicted outcome but not hard cardiac events. The likely explanation could be a relatively lower-risk study population, because all of them had undergone primary percutaneous coronary intervention (27). However, larger multicenter studies are required before MCE can be widely recommended in the assessment of myocardial viability after AMI and thrombolysis.
Quantitative analysis of MCE was not performed, which may have made the data more robust. Nevertheless, the objective of the study was to test whether there was any clinical value in the use of a technique that gives a rapid and accurate qualitative assessment of microvascular perfusion. The study population is not large, but nevertheless, sufficient cardiac events had occurred to allow us to conclude reasonably.
The extent of residual myocardial viability predicted by MCE is a powerful predictor of hard cardiac events in patients after AMI.
Supported by a grant from the Cardiac Research Fund, Northwick Park Institute of Medical Research, Harrow, United Kingdom.
- Abbreviations and Acronyms
- acute myocardial infarction
- coronary artery disease
- contrast defect index
- dobutamine stress echocardiography
- left ventricular
- left ventricular end-diastolic volume
- left ventricular ejection fraction
- left ventricular end-systolic volume
- myocardial contrast echocardiography
- Received January 16, 2007.
- Revision received March 1, 2007.
- Accepted March 12, 2007.
- American College of Cardiology Foundation
- Rogers W.J.,
- Canto J.G.,
- Lambrew C.T.,
- et al.
- Braunwald E.,
- Kloner R.A.
- Swinburn J.M.,
- Senior R.
- Carlos M.E.,
- Smart S.C.,
- Wynsen J.C.,
- Sagar K.B.
- Gibson R.S.,
- Watson D.D.
- Kaul S.,
- Force T.
- Ito H.,
- Maruyama A.,
- Iwakura K.,
- et al.
- Swinburn J.M.,
- Lahiri A.,
- Senior R.
- Coggins M.P.,
- Sklenar J.,
- Le D.E.,
- et al.
- Lafitte S.,
- Higashiyama A.,
- Masugata H.,
- et al.
- Senior R.,
- Sridhara B.S.,
- Basu S.,
- et al.
- Shimoni S.,
- Frangogiannis N.G.,
- Aggeli C.J.,
- et al.
- Jeetley P.,
- Hickman M.,
- Kamp O.,
- et al.
- Senior R.,
- Janardhanan R.,
- Jeetley P.,
- Burden L.
- Senior R.,
- Lepper W.,
- Pasquet A.,
- et al.
- Janardhanan R.,
- Senior R.
- Peltier M.,
- Vancraeynest D.,
- Pasquet A.,
- et al.