Background
Although retrospective studies have suggested that myocardial perfusion and wall motion analysis with real-time myocardial contrast echocardiography (RTMCE) improves the detection of coronary artery disease (CAD) during dobutamine or exercise stress echocardiography, a prospective randomized comparison with conventional stress echocardiography that did not use RTMCE has not been performed.
Methods
A total of 1,776 patients with preserved resting left ventricular wall motion undergoing dobutamine or exercise stress echocardiography for suspicion of CAD were randomized to either non-RTMCE, for which contrast was used only for the approved indication of enhancing left ventricular opacification, or RTMCE, for which contrast infusion was used in all cases to examine both wall motion and myocardial perfusion. Comparisons in test positivity, and positive predictive value in those subsequently referred for quantitative coronary angiography, were performed.
Results
Patients randomized to RTMCE had significantly higher test positivity (22% for RTMCE vs 15% with non-RTMCE, P = .0002). The increased test positivity occurred without a difference in positive predictive value in predicting >50% diameter stenoses by quantitative coronary angiography (67% for non-RTMCE, 73% for RTMCE). The mechanism for increased detection of CAD with RTMCE was mostly due to the detection of subendocardial wall thickening abnormalities that would have gone undetected when examining transmural wall thickening.
Conclusions
RTMCE improves the detection of CAD during dobutamine and exercise stress echocardiography, mainly by the detection of subendocardial ischemia.
Real-time myocardial contrast echocardiography (RTMCE) is a technique that allows the simultaneous analysis of myocardial perfusion (MP) and wall motion (WM) during stress echocardiography. Retrospective and preclinical studies have shown that MP data obtained with RTMCE can be incremental to WM analysis in detecting coronary artery disease (CAD) and can improve the predictive value of the test. Gaibazzi et al. recently showed that MP imaging, in combination with coronary flow reserve measurements during vasodilator stress echocardiography, had high sensitivity and accuracy in detecting CAD. Reant et al. demonstrated increased sensitivity and accuracy of MP imaging when combined with the assessment of two-dimensional strain during dobutamine stress echocardiography (DSE). However, no prospective comparison of RTMCE with state-of-the-art harmonic stress echocardiography has been undertaken. In this study, we prospectively compared the ability of conventional stress echocardiography that did not use real-time myocardial contrast enhancement techniques, for which contrast was used only when needed to define endocardial borders, with that of RTMCE in detecting CAD in a real-world setting. Because RTMCE detects subendocardial perfusion defects, we also sought to study the ability of this technique to detect subendocardial wall thickening abnormalities in the presence of seemingly normal transmural thickening and how this affects the positive predictive value (PPV) of RTMCE in identifying angiographically significant CAD.
Methods
Study Population
As required by our data safety and monitoring board, we performed an interim analysis of 1,964 patients with suspected coronary artery disease undergoing DSE or treadmill exercise stress echocardiography (ESE) in the Prospective Analysis and Comparison of Conventional Stress Echocardiography and Real-Time Myocardial Contrast Stress Echocardiogram study ( ClinicalTrials.gov identifier NCT00575549 ). Consecutive patients referred to the echocardiography laboratory at the University of Nebraska Medical Center between 2007 and 2010 were asked to participate in the study if they scored between 7 and 9 for appropriateness (i.e., rated as appropriate indications for stress echocardiography). Those who consented to participate in the study were randomized, using an Internet-based site ( http://www.random.org/lists ), to undergo either conventional high–frame rate harmonic imaging (non-RTMCE) or RTMCE as their imaging technique. Any patient with a resting WM abnormality involving more than one segment was excluded from this analysis. After this exclusion, a total of 1,776 patients (896 randomized to the non-RTMCE group and 880 to RTMCE) were evaluated. Of these patients, only a small number (59 and 35 patients in the RTMCE and non-RTMCE groups, respectively) had one-segment resting WM abnormalities ( Figure 1 ). All patients in the RTMCE group received contrast to examine MP and WM, while those randomized to non-RTMCE received contrast only if clinically indicated for the US Food and Drug Administration (FDA) indication of enhancing left ventricular opacification (LVO). Patients were advised to stop taking β-blockers for 24 hours before the stress test.
A total of four experienced stress echocardiography readers who had interpreted >100 non–real-time myocardial contrast echocardiographic and real-time myocardial contrast echocardiographic studies with contrast, and one experienced reader who had interpreted >1,000 contrast studies (T.P.), analyzed the studies at the time stress echocardiography was performed. In 27 studies, the interpretation was performed by a less experienced physician but under the supervision of one of the experienced physicians. All interpreting physicians had access to clinical indications and were aware of patient risk factors. All patients gave written informed consent, and the study protocol was approved by the University of Nebraska Medical Center institutional review board.
RTMCE
The contrast agent used for the study was the commercially available lipid-encapsulated microbubble Definity (Lantheus Medical Imaging, North Billerica, MA). This agent was administered as a 3% intravenous continuous infusion at 4 to 6 mL/min under resting conditions and during stress. The infusion was adjusted to optimize myocardial opacification while minimizing attenuation from left ventricular cavity contrast.
RTMCE was performed with ultrasound scanners equipped with low–mechanical index (MI) real-time pulse sequence schemes, which use either interpulse amplitude modulation (Power Modulation, Sonos 5500; Philips Medical Systems, Bothell, WA) or interpulse phase and amplitude modulation (Contrast Pulse Sequencing, Siemens Acuson Sequoia; Siemens Medical Solutions USA, Inc., Mountain View, CA). The MI was kept at ≤0.25, and the frame rate was kept at 20 to 25 Hz during rest and stress imaging, while the brief high-MI impulses to destroy myocardial microbubbles were 1.3 to 1.9 MI. Time gain compensation and two-dimensional gain settings were adjusted to suppress any nonlinear signals from tissue before contrast injection. Digitized images during the myocardial contrast replenishment (MCR) phase after high-MI impulses were obtained from apical views (four, two, and three chamber) at rest and at maximal stress after the patients had achieved a test end point.
For non-RTMCE, Definity contrast was administered only when two contiguous segments could not be visualized, as recommended by the 2008 American Society of Echocardiography guidelines. The contrast was administered as a 3% infusion using B-mode harmonic imaging at a reduced MI of <0.3.
Stress Protocol
The decision to perform dobutamine or exercise treadmill stress echocardiography was made by the referring physician. In either case, patients were instructed to discontinue β-blockers ≥24 hours before the stress test. Patients undergoing exercise stress underwent maximal symptom-limited treadmill exercise according to the Bruce protocol. Patients undergoing DSE received an intravenous dobutamine infusion at a starting dose of 5 μg/kg/min, followed by increasing doses of 10, 20, 30, 40, up to a maximal dose of 50 μg/kg/min, in 3-min to 5-min stages. Atropine (up to 2.0 mg) was injected in patients not achieving 85% of the predicted maximal heart rate (220 − age in years). A 12-lead electrocardiogram and blood pressure were recorded at baseline and at the end of each stress level. The end points of the DSE were achievement of the target heart rate (85% predicted maximal heart rate), maximal dobutamine or atropine dose, ST-segment elevation ≥2 mm at an interval of 80 msec after the J point in non-Q-wave leads, sustained arrhythmias, severe chest pain, or intolerable adverse effects considered to be due to physical exertion, dobutamine, or atropine.
Image Analysis
All studies were analyzed by the reviewer at the time of the study. For RTMCE, both perfusion and WM were analyzed simultaneously during the replenishment phase of contrast after brief high-MI impulses at baseline and at peak stress (>85% predicted maximum heart rate), while for non-RTMCE, WM was analyzed (with or without the aid of enhanced border delineation with contrast) at baseline and peak stress. In accordance with known data regarding capillary blood flow and ultrasound beam elevation width, resting MCR was considered normal when it occurred within 4 sec of the high-MI impulse, while at peak stress, a perfusion defect was defined as delay of >2 sec in transmural or subendocardial contrast replenishment after a high-MI impulse. To ensure that any delays were not due to a reduced contrast input function (e.g., too slow an infusion rate, too long a high-MI impulse), replenishment was also compared with adjacent wall segments at the same depth. Abnormal findings on non-RTMCE were defined as new wall thickening abnormality (hypokinetic, akinetic, or dyskinetic segment) appearing at peak stress, and abnormal findings on RTMCE were based on either a new wall thickening abnormality or a perfusion abnormality at peak stress. Perfusion and WM were both assessed using a standard 17-segment model. In the non-RTMCE arm, the reviewers had access to at least two clips of cardiac cycles obtained at peak stress, to compare side by side with at least one cardiac cycle of resting images. These digitized loops of the three apical windows were displayed side by side for rest and stress comparisons.
Image Analysis by a Second Reviewer
To determine the mechanism for the improved detection of WM abnormalities with RTMCE, wall thickening and contrast enhancement of abnormal studies were analyzed by a second trained reviewer who was blinded to the clinical history of the patient and the original interpretation. Perfusion and WM were analyzed at two time points after the high-MI impulse. The first analysis was within the first 2 sec after the high-MI impulse, before any MCR occurred. These were referred to as pre-MCR images. The second analysis of wall thickening was performed when replenishment of myocardium had begun, usually during a cardiac cycle 2 sec after the high-MI impulse (referred to as MCR images).
The second trained reviewer also analyzed 57 studies (30 by the experienced reviewers and all 27 performed by less experienced reviewers) to assess for interobserver agreement on study interpretation.
Coronary Angiography
Angiography was performed if clinically indicated in the judgment of the referring cardiologist. Only coronary angiograms obtained within 3 weeks of stress echocardiography in otherwise clinically stable patients were used for analysis of PPV. Studies were performed and interpreted by experienced interventional cardiologists. Digital quantitative angiography was performed on all cases, with moderate coronary stenosis defined as a luminal diameter reduction of 50% to 70% and severe stenosis as a luminal diameter reduction of >70%.
Statistical Analysis
All data are expressed as mean ± SD. Patient characteristics at baseline were compared between non-RTMCE and RTMCE using χ 2 and t tests. A multivariate logistic regression model was used to determine if either imaging technique was predictive of an abnormal test result after adjusting for any pertinent differences in baseline characteristics. All analyses were done using SAS version 9.2 (SAS Institute, Inc., Cary, NC).
Results
Patient Characteristics
Table 1 lists the demographic data of the patients included in the study, comparing those randomized to non-RTMCE and to RTMCE. Patients were similar with respect to age, gender, risk factors other than hyperlipidemia, ejection fraction, and most medications. However, clopidogrel and β-blocker use were more common in the RTMCE group, and more patients in the RTMCE group had prior percutaneous coronary intervention or remote histories of myocardial infarction. DSE was performed with higher frequency in patients randomized to RTMCE (61% vs 55% for non-RTMCE, P = .01).
| Characteristic | RTMCE ( n = 880) | Non-RTMCE ( n = 896) | P |
|---|---|---|---|
| DSE | 61% | 55% | .01 |
| Age (y) | 59 ± 12 | 59 ± 13 | .40 |
| Men | 46% | 45% | .48 |
| BMI (kg/m 2 ) | 32 ± 8 | 32 ± 9 | .95 |
| Diabetes | 27% | 24% | .17 |
| Family history of CAD | 34% | 33% | .6 |
| Hypertension | 61% | 59% | .25 |
| Smoking | 35% | 34% | .53 |
| Angiotensin-converting enzyme inhibitor use | 24% | 24% | .96 |
| Aspirin | 35% | 33% | .33 |
| Clopidogrel | 6% | 3% | .0038 |
| β-blockers | 41% | 35% | .0045 |
| Left ventricular ejection fraction (%) ∗ | 60 ± 9 | 58 ± 10 | .52 |
| Diuretics | 14% | 16% | .28 |
| Antilipid agents | 46% | 40% | .0074 |
| Obesity † | 35% | 33% | .38 |
| Prior revascularization | 17% | 10% | <.0001 |
| Conduction abnormality at rest | 4% | 4% | .77 |
| Previous myocardial infarction | 8% | 5% | .0077 |
| Previous PTCA/PCI | 12% | 7% | .0003 |
| Prior CABG | 7% | 5% | .13 |
∗ Determined using either Simpson’s biplane or single-plane formula.
DSE and ESE
A total of 412 patients (46%) in the non-RTMCE group received contrast to improve LVO. During DSE, the average increases in heart rate and systolic blood pressure were not different between the non-RTMCE and RTMCE groups ( Table 2 ). There were also no differences between groups in those who failed to attain the 85% of their target heart rates for their age. The stress test was stopped prematurely in eight patients (four non-RTMCE, four RTMCE) because of hypertensive responses, shortness of breath, and supraventricular or nonsustained ventricular tachycardia.
| DSE | ESE | Overall | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | RTMCE ( n = 535) | Non-RTMCE ( n = 494) | P | RTMCE ( n = 335) | Non-RTMCE ( n = 400) | P | RTMCE ( n = 880) | Non-RTMCE ( n = 896) | P |
| Arrhythmias (%) | 32 | 28 | .16 | 22 | 16 | .028 | 28 | 23 | .0075 |
| Chest pain | 16 | 11 | .012 | 7 | 6 | .65 | 13 | 9 | .0096 |
| Dyspnea | 1 | 1 | .89 | 11 | 11 | .84 | 5 | 6 | .66 |
| Conduction abnormality at stress | 2 | 4 | .088 | 1 | 2 | .52 | 2 | 3 | .085 |
| Peak systolic BP (mm Hg) | 152 ± 33 | 153 ± 30 | .43 | 151 ± 19 | 152 ± 19 | .55 | 152 ± 29 | 153 ± 25 | .35 |
| Peak diastolic BP (mm Hg) | 67 ± 19 | 67 ± 19 | .72 | 81 ± 8 | 82 ± 8 | .59 | 72 ± 17 | 74 ± 17 | .12 |
| Peak heart rate (beats/min) | 143 ± 12 | 143 ± 13 | .99 | 156 ± 21 | 156 ± 23 | .95 | 148 ± 18 | 149 ± 19 | .40 |
| Percentage of maximal predicted heart rate (%) | 90 ± 11 | 91 ± 10 | .15 | 95 ± 13 | 94 ± 11 | .43 | 92 ± 12 | 92 ± 10 | .40 |
| Rate-pressure product | 21,543 ± 4,816 | 21,620 ± 5,332 | .87 | 23,901 ± 5,415 | 23,769 ± 4,936 | .82 | 22,382 ± 5,156 | 22,645 ± 5,252 | .48 |
During the exercise stress test, the average increases in heart rate and systolic blood pressure, as well as rate-pressure product achieved, were not different between groups ( Table 2 ). Most patients in each group attained 85% of their target heart rates. Among patients who underwent non-RTMCE, 80 (9%) developed chest pain and 206 (23%) developed stress-induced arrhythmias. Among patients undergoing RTMCE, 114 (13%) developed chest pain and 244 (28%) had stress-induced arrhythmias ( P < .05 vs non-RTMCE). The additional arrhythmia (premature ventricular contractions) in the RTMCE group was related to the brief high-MI pulse. However, the arrhythmia induced by this was one beat in duration; none were sustained.
MP and WM Analysis
Patients randomized to RTMCE had a significantly higher positivity rate compared with non-RTMCE (22% vs 15%, P = .0002). This higher positivity rate was observed even after removing the patients with resting WM abnormalities in one segment (18% vs 13%, P = .018). Furthermore, WM was considered abnormal in a higher percentage of RTMCE cases (19% of RTMCE cases vs 15% with non-RTMCE, P = .056; Figure 2 ). The higher proportion of WM abnormalities with RTMCE was observed irrespective of whether the patient underwent DSE (24% vs 17%, P = .009) or ESE (19% vs 13%, P = .04). Using a univariate logistic model, the likelihood of a positive test was 1.6 times more likely if the patient was randomized to RTMCE rather than non-RTMCE ( P = .0002). In a multivariate logistic model adjusting for differences in risk factor profiles between the groups (including the frequency of hyperlipidemia, clopidogrel or β-blocker use, prior percutaneous coronary intervention, or prior myocardial infarction), the likelihood of an abnormal test was still 1.4 times higher if randomized to RTMCE ( P = .015). There was also no significant interaction between test type (ESE vs DSE) or test technique (RTMCE vs non-RTMCE), indicating that the higher likelihood of test positivity was true for both DSE and ESE.
Because of the larger number of prior percutaneous revascularizations and prior myocardial infarctions in the RTMCE group, the analysis of test positivity in the non-RTMCE and RTMCE groups was repeated after excluding these patients, leaving 1,550 patients for analysis. In the non-RTMCE group, the abnormal rate was 12% compared with 17% in the RTMCE group ( P = .010). With a univariate logistic model, an abnormal test result with RTMCE was 1.5 times more likely, even after removing patients with prior revascularization or infarctions. Table 3 demonstrates that a repeat multivariate logistic model adjusting for hyperlipidemia, clopidogrel use, β-blocker use, and hyperlipidemia, as well as test type (treadmill or dobutamine stress), an abnormal test result was 1.4 times more likely with RTMCE than with non-RTMCE after excluding patients with prior infarctions or prior revascularization ( P = .035). There was still no significant interaction between test type and technique, which indicates that the effect of technique in predicting echocardiographic results did not differ depending on use of DSE or ESE ( P = .81).
