Background
No large study has demonstrated that any stress test can risk-stratify future hard cardiac events (cardiac death or myocardial infarction) in patients with suspected acute coronary syndromes (ACS), nondiagnostic electrocardiographic (ECG) findings, and normal troponin levels. The aim of this study was to test the hypothesis that combined contrast wall motion and myocardial perfusion echocardiographic assessment (cMCE) during stress echocardiography can predict long-term hard cardiac events in patients with suspected ACS, nondiagnostic ECG findings, and normal troponin.
Methods
A total of 545 patients referred for contrast stress echocardiography from the emergency department for suspected ACS but nondiagnostic ECG findings and normal troponin levels at 12 hours were followed up for cardiac events. Patients underwent dipyridamole-atropine echocardiography with adjunctive myocardial perfusion imaging using a commercially available ultrasound contrast medium (SonoVue).
Results
During a median follow-up period of 12 months, 25 cardiac events (4.6%) occurred (no deaths, 12 nonfatal myocardial infarctions, 13 episodes of unstable angina). Abnormal findings on cMCE were the most significant predictor of both hard cardiac events (hazard ratio, 22.8; 95% confidence interval, 2.9–176.7) and the combined (cardiac death, myocardial infarction, or unstable angina requiring revascularization) end point (hazard ratio, 10.7; 95% confidence interval, 3.7–31.3). The inclusion of the cMCE variable significantly improved multivariate models, determining lower Akaike information criterion values and higher discrimination ability.
Conclusions
cMCE during contrast stress echocardiography provided independent information for predicting hard and combined cardiac events beyond that predicted by stress wall motion abnormalities in patients with suspected ACS, nondiagnostic ECG findings, and normal troponin levels.
The evaluation of patients presenting to the hospital with suspected acute coronary syndromes (ACS) remains a challenge. The majority of patients with suspected ACS present with normal or nondiagnostic electrocardiographic (ECG) findings and normal cardiac markers both on presentation and on serial measurements. Current guidelines recommend stress testing for this cohort of patients. Contrast stress echocardiography (SE), which assesses both myocardial wall motion (WM) and myocardial perfusion (MP) simultaneously, has been shown to be accurate for the assessment of coronary artery disease (CAD) and to provide incremental information for the prediction of cardiac events over clinical data in suspected CAD. However, in the subset of patients presenting with suspected ACS but nondiagnostic ECG findings and normal troponin levels, contrast SE has not been shown to predict hard cardiac events, probably because of small study populations. We hypothesized that contrast SE, by virtue of its excellent spatial and temporal resolution and its ability to assess function and perfusion simultaneously, is likely to predict hard cardiac events in patients presenting to the hospital with suspected ACS but nondiagnostic ECG findings and normal 12-hour troponin levels.
Methods
Patient Population
Consecutive patients presenting to the emergency department at Parma University Hospital from December 2007 to November 2008 with suspected ACS but nondiagnostic ECG findings and normal 12-hour troponin levels, who were referred for clinical SE, were approached to participate in the study if they did not meet any of the exclusion criteria. Patients who enrolled in the study underwent contrast SE between 24 and 104 hours after their episodes of chest pain. Patients were followed for a median of 361 days or until they developed an end point.
Exclusion criteria were (1) left ventricular ejection fraction < 30%, (2) severe valvular disease, (3) sustained ventricular arrhythmias, (4) chest pain within 24 hours of contrast SE, (5) unable to discontinue β-blockers in the past 24 hours, (6) known allergy to sulfonamides, and (7) pregnancy or lactation. All patients underwent full clinical assessment and Thrombolysis In Myocardial Infarction risk evaluation. ECG findings were assessed as either normal or abnormal (<0.5-mm ST-segment depression, nonspecific T-wave inversion, or complete bundle branch block). Troponin I was assessed at admission and at 12 hours. This study was approved by our institutional ethics committee. All patients gave written informed consent to the examination and follow-up study. The study complied with the Declaration of Helsinki.
Contrast SE
In the clinical protocol, contrast SE was performed first at the stress stage, and resting studies were performed at the recovery stage only when either MP or WM abnormalities were observed during stress ( Figure 1 ). All images were acquired in the apical four-chamber, two-chamber, and three-chamber views using an S5-1 broadband transducer (Philips Medical Systems, Eindhoven, The Netherlands). SonoVue (Bracco, Milan, Italy) was infused using a rotating pump (BR-NF100; Bracco, Geneva, Switzerland) 1 min after the completion of dipyridamole infusion followed after 1 min by imaging. Low-power (mechanical index = 0.10) continuous imaging was performed for WM assessment. For MP, a flash-replenishment sequence was used (destructive pulse of eight frames at a mechanical index of 1.0, followed by low-power imaging for 10 cycles) both in the continuous (40 frames/sec) and end-systolic triggered modes.
Interpretation of WM
Regional WM was evaluated using a semiquantitative WM score (normal, hypokinesia, or akinesia) using a 17-segment model of the left ventricle. Positive results were defined as the occurrence in at least one myocardial segment of either a new dyssynergy in a region with normal rest function or worsening of rest dyssynergy. In our protocol, both rest and stress WM assessments were intentionally performed without the use of contrast.
Interpretation of MP
Perfusion was defined as abnormal if the myocardium replenishment was delayed beyond 2 sec after the flash impulse at peak hyperemia. The cutoff for normal replenishment after aminophylline infusion (recovery stage) was considered 4 sec after the end of the flash impulse. 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, with the exception of a fixed perfusion defect in the specific site of a known prior acute myocardial infarction (AMI), which was considered a “normal” MP result because it is expected and compatible with the known clinical history. In the study protocol, the only clips considered for WM analysis evaluation were those acquired without contrast (at rest and after atropine); peak contrast real-time clips (acquired before atropine for MP analysis) were also available to the readers and may have retained some WM information. As a consequence, MP data could not formally be defined as stand-alone MP data in our analysis but rather as a combined contrast MP and WM echocardiographic assessment (cMCE). Standard WM and cMCE were two independent parameters available from the test, which was finally scored as (1) negative for both WM and cMCE, (2) positive for both WM and cMCE, or (3) positive for cMCE only; there were in fact no tests with positive results for WM and negative results for cMCE, as expected from pathophysiology. WM and cMCE clips were separately analyzed in a blinded manner by consensus by two observers. Disagreement was resolved by a third observer.
The decision to proceed to coronary angiography after contrast stress echocardiographic results was left to the referring physician.
Quantitative Coronary Angiography
Only coronary angiographic studies performed <60 days after SE were considered. Quantitative coronary angiography was performed by an experienced cardiologist. Any visually evident stenosis was measured using a handheld electronic caliper (Tesa S.A., Renes, Switzerland) operated with custom-developed software. Measurements were expressed as the percentage diameter narrowing using the diameter of the nearest normal-appearing region as the reference. CAD was defined as >50% luminal diameter stenosis in one or more major coronary arteries. We have previously reported on the increase in diagnostic accuracy by the addition of contrast perfusion imaging analysis (89%; 95% confidence interval [CI], 83%–91%) beyond WM assessment (73%; 95% CI, 66%–77%) ( P < .05) in a subset of patients with similar clinical characteristics compared with those enrolled in the present study.
Follow-Up
Follow-up data were obtained from review of patients’ hospital charts and electronic records as well as telephone interviews with patients. Follow-up time was calculated from the time of the test to the end point date or the date of final contact with the patient. Patients who underwent revascularization procedures after the tests were not excluded. A hard cardiac end point was defined as cardiac death or nonfatal AMI. A combined end point was defined as cardiac death, nonfatal AMI, or unstable angina requiring urgent revascularization. Cardiac death was defined as death associated with known or suspected AMI, life-threatening arrhythmia, or heart failure. Nonfatal AMI was defined on the basis of criteria of typical chest pain, elevated cardiac enzyme levels, with or without typical ECG changes; revascularization-related AMI was not considered an end point for the purpose of the study.
Statistical Analysis
Continuous variables are expressed as mean ± SD; categorical variables are expressed as numbers of cases and proportions. Chi-square analysis was used for comparison of categorical variables. In a random sample of 20 subjects, both WM and cMCE were independently evaluated by two observers, and intraobserver and interobserver agreement for each variable was described using Cohen’s κ statistics. Kaplan-Meier curves were used to evaluate time to hard cardiac events and time to combined cardiac events. Differences between survival curves were compared using log-rank tests. Univariate and multivariate Cox proportional-hazards models were used to evaluate the risk for cardiac events, on the basis of any potentially significant predictor. All variables significant at 1% in the univariate analysis were evaluated in multivariate models. Covariate selection for multivariate model entry was based on the clinical meaning of the variable and statistical approach. Stepwise methods along with clinical considerations were used to select candidate models. For the purpose of verifying the importance of considering abnormal cMCE findings in model building, the Akaike information criterion (AIC) was also used for model comparison; the model with the smallest AIC was considered the best one, and although the limited number of events also limited the number of covariates considered, the discrimination ability of models was evaluated with Harrell’s c-index. Receiver operating characteristic curves on the basis of predicted values from Cox models were also used to evaluate the sensitivity and specificity of either abnormal cMCE findings or abnormal stress WM findings with respect to combined cardiac events. The proportional-hazards assumption was checked using the Schoenfeld test. P values < .05 was considered significant. Analysis were performed using Stata release 10 (StataCorp LP, College Station, TX).
Results
Patient Characteristics
Of the 557 patients enrolled, 12 (2%) were excluded because of suboptimal acoustic windows; follow-up was available in the remaining 545 patients (58% men), with a mean age of 67 ± 11 years ( Table 1 ). Of these, 218 patients (40%) had either previous AMI or histories of revascularization or known obstructive CAD. Thrombolysis In Myocardial Infarction risk scores were low (0 or 1) in 44% and intermediate (2–4) in the remaining 56%.
| Variable | n (%) |
|---|---|
| Risk factors | |
| Age ≥ 65 y | 350 (64) |
| Male gender | 317 (58) |
| TIMI risk score | |
| 0 or 1 | 240 (44) |
| 2–4 | 305 (56) |
| Family history of premature CAD | 125 (22) |
| Hypercholesterolemia | 318 (58) |
| Hypertension | 417 (77) |
| Diabetes | 128 (24) |
| Cigarette smoking | 136 (25) |
| Obesity | 75 (14) |
| Abnormal rest ECG findings | 147 (27) |
| Cardiovascular history | |
| Previous AMI | 151 (28) |
| Previous PCI | 130 (24) |
| Previous CABG | 39 (7) |
| Medications | |
| Aspirin | 332 (61) |
| β-blockers | 307 (55) |
| Statins | 296 (54) |
| Test results | |
| Rest WM abnormalities | 149 (27) |
| Reversible WM abnormalities | 101 (19) |
| Abnormal cMCE findings | 195 (36) |
| Reversible | 172 (32) |
| Fixed | 23 (4) |
Contrast SE
The results of SE were interpreted as concordantly normal for both WM response and cMCE in 350 patients (64%) and concordantly abnormal in 101 (18%), while 94 patients (17%) had normal WM responses but abnormal cMCE findings. All patients with abnormal WM responses had abnormal cMCE findings. In the 195 patients with abnormal cMCE findings, two vials of SonoVue were infused (one vial for stress and one for recovery cMCE), while in the 350 patients with normal cMCE findings, just one vial was used, at the stress stage. The average quantity of contrast used per patient was 6.5 mL. No contrast-related side effects were recorded; mild reversible flushing or headache were related to dipyridamole infusion in 228 of 545 patients (42%). The mean heart rate at rest was 69 ± 12 beats/min, the mean stress heart rate was 105 ± 17 beats/min, the mean rest systolic blood pressure was 142 ± 19 mm Hg, and the man stress systolic blood pressure was 135 ± 22 mm Hg.
Intraobserver and Interobserver Agreement
Agreement between observers interpreting WM and cMCE findings for test positivity versus test negativity on a patient level in the 20 randomly selected studies was 95% (κ = 0.77) and 90% (κ = 0.76), respectively; intraobserver agreement for WM and cMCE was 95% (κ = 0.77) and 90% (κ = 0.79), respectively.
Coronary Arteriography
Of the 545 patients, 170 underwent coronary arteriography at the discretion of the attending physician who had knowledge of the imaging data. Significant CAD was detected in 130 patients (76%), of whom 65 had single-vessel, 44 had two-vessel, and 21 had three-vessel CAD. Left anterior descending coronary artery stenoses were observed in 114 patients (67%). Eighty-two patients had either percutaneous coronary intervention ( n = 65) or coronary artery bypass grafting ( n = 17) within 3 months of SE.
Cardiac Events
The median follow-up period for patients who did not experience events was 361 days (first quartile, 255 days; third quartile, 445 days). A total of 25 patients (4.6%) had cardiac events during follow-up.
Hard events occurred at a median of 163 days (range, 23–370 days) after the index stress test and included nonfatal AMIs in 12 patients (2.2%), which were unrelated to revascularization; unstable angina requiring revascularization was recorded in 13 patients at a median of 129 days (range, 37–362 days). No cardiac deaths were recorded. One patient died from a noncardiac cause.
Cardiac Events and Contrast SE
Table 2 summarizes the occurrence of cardiac events in relation to cMCE and WM results. Figure 2 shows the highest survival for those with normal cMCE findings for both hard cardiac events ( P < .001) and combined cardiac events ( P < .001). Figure 3 illustrates that also for patients with normal WM responses, survival was significantly higher for both hard cardiac events ( P < .01) and combined cardiac events ( P < .01). Figure 4 shows that when both WM and cMCE findings were normal, survival to hard and combined cardiac events was higher with respect to both abnormal cMCE findings only and combined abnormal cMCE and WM results. Figure 5 demonstrates that cMCE status determined significantly different survival for hard cardiac events in patients with both low and intermediate Thrombolysis In Myocardial Infarction risk scores ( P < .01).
| Cardiac death | Nonfatal AMI | Unstable angina | Hard cardiac event rate (%) | Combined cardiac event rate (%) | |
|---|---|---|---|---|---|
| All patients ( n = 545) | 0 | 12 | 13 | 2.2 | 4.6 |
| Normal cMCE findings and normal WM ( n = 350) | 0 | 0 | 3 | 0 | 0.8 |
| Abnormal cMCE findings ( n = 195) | 0 | 12 | 10 | 6.2 | 11.3 |
| Reversible cMCE abnormalities ( n = 172) | 0 | 10 | 10 | ||
| Fixed cMCE abnormalities ( n = 23) | 0 | 2 | 0 | ||
| Normal WM ( n = 444) | 0 | 6 | 8 | 1.4 | 3.2 |
| Abnormal WM ( n = 101) | 0 | 6 | 5 | 5.9 | 10.9 |
| Normal WM but abnormal cMCE findings ( n = 94) | 0 | 6 | 5 | 6.4 | 11.7 |
Univariate and Multivariable Cox Regression Analysis
Tables 3 and 4 present the univariate predictors of risk for both hard and combined cardiac events at follow-up. Multivariate analysis was not feasible for hard cardiac events because of the small number of events; for the same reason, only two variables were entered in multivariate models for combined cardiac events.
| Variable | HR | 95% CI | P |
|---|---|---|---|
| Age | 1.037 | (0.977–1.1) | .230 |
| Three or more risk factors | 0.990 | (0.314–3.12) | .986 |
| Cigarette smoking | 0.981 | (0.266–3.625) | .977 |
| Hypercholesterolemia | 0.701 | (0.226–2.175) | .539 |
| Diabetes | 2.353 | (0.747–7.416) | .144 |
| Hypertension | 1.571 | (0.344–7.172) | .560 |
| Previous AMI | 0.520 | (0.114–2.374) | .399 |
| Statins | 2.673 | (0.723–9.882) | .140 |
| Aspirin | 7.383 | (0.953–57.203) | .056 |
| β-blockers | 4.097 | (0.897–18.708) | .069 |
| Previous AMI, CAD, or revascularization | 1.115 | (0.354–3.512) | .853 |
| TIMI risk score low/intermediate | 1.530 | (0.46–5.085) | .488 |
| Abnormal rest ECG findings | 1.325 | (0.399–4.399) | .646 |
| Rest WM abnormalities | 1.374 | (0.414–4.563) | .604 |
| Abnormal vs normal WM | 4.916 | (1–583–15.268) | .006 |
| Abnormal vs normal cMCE findings | 22.778 | (2.936–176.730) | .003 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
