Author + information
- Received January 29, 1997
- Revision received June 17, 1997
- Accepted July 1, 1997
- Published online October 1, 1997.
- Michael C Kontos, MDA,* (, )
- Robert L Jesse, MD, FACCA,
- Kristin L Schmidt, BAA,
- Joseph P Ornato, MD, FACCA and
- James L Tatum, MDA
- ↵*Dr. Michael C. Kontos, Box 980128, Medical College of Virginia/Virginia Commonwealth University, Richmond, Virginia 23298-0128.
Objectives. This study sought to determine the ability of early perfusion imaging using technetium-99m sestamibi to predict adverse cardiac outcomes in patients who present to the emergency department with possible cardiac ischemia and nondiagnostic electrocardiograms (ECGs).
Background. Evaluation of patients presenting to the emergency department with possible acute coronary syndromes and nondiagnostic ECGs is problematic. Accurate risk stratification is necessary to prevent serious adverse outcomes. Initial results suggest that early perfusion imaging using technetium-99m sestamibi enables reliable risk stratification.
Methods. Patients presenting to the emergency department with a low to moderate probability of acute coronary syndromes underwent rapid sestamibi injection with gated single-photon emission computed tomographic imaging. Studies showing perfusion defects with associated wall motion abnormalities were considered positive.
Results. A total of 532 consecutive patients underwent serial myocardial marker analysis and rest perfusion imaging. Of these patients, perfusion imaging was positive in 171 (32%). Positive perfusion imaging was the only multivariate predictor of myocardial infarction (MI) (p < 0.0001, odds ratio [OR] 33, 95% confidence interval [CI] 7.7 to 141) and was the most important independent predictor of MI or revascularization (p < 0.0001, OR 14, 95% CI 7.3 to 25), followed by diabetes (p < 0.01, OR 2.8, 95% CI 1.5 to 5.1), typical angina (p = 0.01, OR 2.1, 95% CI 1.2 to 3.7) and male gender (p = 0.03, OR 1.9, 95% CI 1.1 to 3.5). The sensitivity of positive perfusion imaging for MI was 93% (95% CI 77% to 98%), and for MI or revascularization it was 81% (95% CI 71% to 88%), with negative predictive values of 99% (95% CI 98% to 100%) and 95% (95% CI 92% to 97%), respectively.
Conclusions. Positive rest perfusion imaging accurately identified patients at high risk for adverse cardiac outcomes, whereas negative perfusion imaging identified a low risk patient group. Early perfusion imaging allows for rapid and accurate risk stratification of emergency department patients with possible cardiac ischemia and nondiagnostic ECGs.
The evaluation of patients presenting to the emergency department with chest pain is often problematic. Limitations in the diagnostic accuracy of clinical, historic and electrocardiographic (ECG) data result in the hospital admission of the majority of these patients [1, 2], even though many have nonischemic etiologies for their chest pain [3, 4]. Despite this low threshold for admission, 4% to 8% of patients with myocardial infarction (MI) are inappropriately discharged from emergency departments, often with adverse consequences [5, 6].
Recent studies have used technetium-99m sestamibi (Cardiolite, Dupont Pharma) in an effort to increase diagnostic accuracy in patients presenting to the emergency department with chest pain. Sestamibi has several advantages when used in these patients, including minimal redistribution , which allows patients to be injected while they are symptomatic and to be imaged later after clinical stabilization. Initial studies have demonstrated that sestamibi has a higher sensitivity and specificity for detecting acute ischemic syndromes compared with clinical and ECG variables [8–10]. However, these studies included relatively small numbers of patients who were selectively enrolled in investigational protocols. The diagnostic accuracy of rest sestamibi perfusion imaging when used in a larger, nonselected patient group is unknown. The purpose of this study was to determine the accuracy of early perfusion imaging with sestamibi when used in patients considered to have a low to moderate risk for acute coronary syndrome.
1.1 Patient selection.
All patients who present to the Emergency Department of the Medical College of Virginia/Virginia Commonwealth University with symptoms suggestive of cardiac ischemia undergo prompt clinical evaluation, which includes a history, physical examination and ECG. Our chest pain evaluation protocol has been previously described . In general, high risk patients (those with abnormal ECGs or those with nondiagnostic ECGs, known coronary disease and typical symptoms) are admitted to the coronary care unit (CCU). Lower risk patients undergo further risk stratification according to a protocol that includes early rest myocardial perfusion imaging. This study includes all patients who were evaluated in the emergency department from June 1994 through August 1995, who underwent early perfusion imaging and serial myocardial marker analysis. These guidelines were used by emergency department residents and attending physicians for patient assessment, but clinical judgment was permitted, occasionally resulting in the imaging of patients who retrospectively did not meet these guidelines (21 patients with positive and 26 patients with negative imaging results). Exclusion of these patients did not significantly change the results, and therefore they were included in all analyses.
1.2 Imaging and interpretation.
The technique of early perfusion imaging has been previously reported . Briefly, patients were injected with ∼20 mCi of sestamibi in the emergency department, and the presence or absence of symptoms was recorded. Although patients were not required to be symptomatic at the time of tracer injection, none were injected more than 6 h after the last episode of pain. Approximately 60 to 90 min after injection, gated single-photon emission computed tomographic (SPECT) myocardial perfusion imaging was performed using a triple-headed gamma camera system. Immediately after data acquisition, predefined commercial protocols were used to generate short- and long-axis static reconstructions and multilevel gated cine films for visual interpretation.
Perfusion images were evaluated by an experienced nuclear medicine attending physician, and all data were made available to the physicians treating the patient. For purposes of this study, images were classified as either positive or negative for acute MI or ischemia. A positive study required a discrete perfusion defect with associated abnormalities in wall motion and thickening. Studies visually interpreted as normal, equivocal or consistent with cardiomyopathy were considered negative for acute coronary syndromes. Normal studies had normal perfusion and systolic function without regional wall motion or thickening abnormalities. Equivocal studies typically displayed low grade perfusion abnormalities but normal wall motion and regional thickening. Studies consistent with cardiomyopathy showed reduced systolic function on cinematic replay with either normal perfusion or perfusion defects without accompanying segmental wall motion abnormalities.
1.3 End points and definitions.
The patients were observed in the CCU and underwent serial sampling of myoglobin, creatine kinase–MB fraction (CK-MB) mass and total creatine kinase (CK). Decisions regarding further diagnostic evaluation were made by the attending cardiologist in the CCU. Stress testing was performed using either a standard Bruce protocol or pharmacologic stress with dipyridamole or dobutamine. Coronary angiography was performed using the Judkins technique with views of the coronary arteries obtained in multiple projections. Significant coronary artery disease was defined as ≥70% stenosis in a major coronary artery or its branches or ≥50% left main coronary artery stenosis.
The primary end points were cardiac death, MI and revascularization (coronary artery bypass graft surgery [CABG] or percutaneous transluminal coronary angioplasty [PTCA]) during the initial evaluation or within 5 days of admission. Myocardial infarction was defined as CK-MB mass ≥8.0 ng/dl with a relative index ([CK-MB mass/total CK] × 100) ≥4.0. For patients having both MI and revascularization, only MI was counted as an event. Anginal symptoms were considered typical if they were described as pressure, tightness, squeezing, burning, heaviness, crushing or indigestion, or were similar to previous symptoms of angina. Only the initial hospital admission was used for patients admitted more than once.
1.4 Statistical analysis.
Results are presented as mean value ± SD. Comparisons were made using the Student ttest and chi-square analysis for categoric and continuous variables, respectively. A p value <0.05 was considered significant. Sensitivity, specificity and positive and negative predictive values were calculated in the standard fashion. Stepwise logistic regression analysis using the variables in Table 1(age was considered a dichotomous variable) was used to determine the best model that predicted the specified clinical outcome. Variables were entered into the model sequentially according to the largest adjusted chi-square that had a p value ≤0.10 at the specific step and were removed if the p value was ≥0.10.
A total of 542 patients were observed in the CCU after undergoing sestamibi injection in the emergency department. Ten patients (1.9%) were excluded from the analysis (five refused to be imaged and five had nondiagnostic image quality), leaving 532 patients in whom myocardial markers and perfusion images could be compared. Perfusion imaging was positive in 171 patients (32%) and negative in 361 (68%). The negative studies consisted of 325 (90%) that were normal, 28 (7.8%) that were equivocal and 8 (2.2%) that were consistent with cardiomyopathy. Comparisons of patients with positive and negative perfusion studies are shown in Table 1. Patients with positive rest perfusion images were older and more likely to be ≥60 years old, be male, have hypertension and diabetes and have had a previous MI. They were also significantly more likely to have a new MI (odds ratio [OR] 32, 95% confidence intervals [CI] 7.5 to 137), to have an MI or undergo revascularization (OR 15, 95% CI 8.5 to 28) or to have an MI, revascularization or significant coronary artery disease (OR 17, 95% CI 10 to 28) (Fig. 1).
2.1 Myocardial infarction.
Twenty-eight patients were diagnosed with MI (5.3%), 26 of whom (93%) had positive rest perfusion images (Table 2). Only 2 (0.6%) of the 361 patients (95% CI 0% to 2.4%) with negative perfusion images had an acute MI. Both patients had non–Q wave MIs and both had peak CKs <500 U/liter and single-vessel disease.
Multivariate analysis demonstrated that positive rest perfusion imaging was the only independent variable predictive of MI (p < 0.001, OR 33, 95% CI 7.7 to 141). Early rest perfusion imaging had a sensitivity of 93% (95% CI 77% to 98%), a specificity of 71% (95% CI 66% to 74%) and a negative predictive value of 99.4% (95% CI 98% to 100%) for identifying or excluding MI. There was no significant difference in the sensitivity for diagnosing MI in men compared with women (94% vs. 92%).
2.2 Revascularization and coronary angiography.
Ninety-two patients who had positive rest perfusion imaging without a new MI underwent coronary angiography. Significant disease was found in 65 patients (71%), including 45 who then underwent revascularization (CABG in 9, PTCA in 36). Among the 20 patients who were not revascularized, 7 had left main coronary or three-vessel disease, 9 had two-vessel disease and four had one-vessel disease. Seventy patients who had negative early rest perfusion imaging also underwent diagnostic coronary angiography; significant disease was found in 21 patients (30%), 14 of whom underwent revascularization (CABG in 4, PTCA in 10). Among the seven patients who were not revascularized, six had single-vessel and one had three-vessel disease.
Patients with either MI or revascularization (n = 87) were compared with those with neither (n = 445) (Table 3). Male gender, diabetes, typical angina, previous MI and positive early perfusion imaging were significantly more common in patients with MI or revascularization. On multivariate analysis, positive early perfusion imaging was the most important predictor (p < 0.0001, OR 14), followed by diabetes (p < 0.01, OR 2.8), typical angina (p = 0.01, OR 2.1) and male gender (p = 0.03, OR 2.0) (Fig. 2). Positive early rest perfusion imaging had a sensitivity of 82% (95% CI 71% to 88%), a specificity of 76% (95% CI 72% to 80%) and a negative predictive value of 95% (95% CI 92% to 97%) for predicting MI or revascularization. There was no significant difference in the sensitivity when men were compared with women (47 of 55 [86%] vs. 24 of 32 [75%], p = 0.28).
Twenty-seven of the 92 patients with positive early perfusion images without a new MI underwent coronary angiography but did not have significant coronary lesions (Table 4). Of these, 18 patients had a history of MI. Two patients had at least one vessel with 50% to 70% stenosis. In addition to the perfusion imaging, four patients had either ECG or echocardiographic evidence consistent with an acute coronary syndrome, despite minimal disease on angiography. Thus, 24 (89%) of the 27 patients with abnormal perfusion images but without angiographically significant coronary lesions had symptoms or other diagnostic information, or both, consistent with an acute coronary syndrome or previous MI.
2.3 Presence of chest pain at time of injection.
Fourteen (61%) of the 23 patients with negative rest perfusion imaging who had an MI or significant disease were symptom free at the time of injection. The sensitivity of positive early rest perfusion imaging was not significantly different between patients injected with and those injected without pain, regardless of which end points were used: MI, revascularization or significant coronary disease (82% vs. 86%); MI or revascularization (83% vs. 78%); or revascularization or significant disease (76% vs. 73%).
2.4 Exclusion of patients with previous MI.
Specificity may be reduced by inclusion of patients with a previous MI. Therefore, the data were reanalyzed after excluding the 114 patients with historic or ECG evidence of a previous MI (70 positive and 44 negative). The only univariate predictor of MI remained positive early rest perfusion imaging (p < 0.001, OR 78, 95% CI 10 to 590). The sensitivity of positive early rest perfusion imaging for predicting MI was essentially unchanged (95%; 95% CI 75% to 99%), and specificity significantly increased from 71% to 80% (95% CI 79% to 87%, p = 0.005). The specificity for predicting MI or revascularization was also significantly increased to 83% (95% CI 79% to 87%, p = 0.01), again without a significant change in sensitivity (78%; 95% CI 65% to 87%).
We found that performing early rest myocardial perfusion imaging in lower risk patients presenting to the emergency department with possible myocardial ischemia accurately identified those at increased risk for cardiovascular complications. Using multivariate analysis, positive perfusion imaging was the only independent predictor of MI and was the most important independent predictor of the combined end point of MI or revascularization. The sensitivity for predicting MI or revascularization was high, as was the negative predictive value for excluding them. Excluding patients with a previous MI significantly improved the specificity for predicting MI alone and the combination of MI or revascularization, with no loss in sensitivity.
Each year ∼6 million patients are evaluated in emergency departments for chest pain or other symptoms consistent with cardiac ischemia , accounting for 7% of all emergency department visits . In patients with obvious ECG or clinical indicators of cardiac ischemia, diagnosis and initial triage are usually straightforward. In patients with nonischemic ECGs, accurate diagnosis is more difficult. Although chest pain characteristics, known coronary disease, age and gender have some discriminating value [1, 13–16], the overlap in initial presentations between patients with and those without cardiac ischemia results in significant diagnostic uncertainty [1, 2]. This uncertainty leads to the hospital admission of many patients whose symptoms are ultimately attributed to nonischemic causes; the cost for the evaluation of these patients is estimated to be $4 billion annually .
We found that early rest perfusion imaging was a powerful predictor of MI in this patient group for which early and accurate diagnosis has been difficult. Although the incidence of MI in these patients is low, it is not insignificant [18, 19]and is associated with a clinically important mortality rate . Similar to other studies, we found that traditional risk factors were not helpful in distinguishing patients with MI [15, 20–22].
Importantly, we found that early rest perfusion imaging could accurately identify patients with acute ischemia without MI. The importance of this identification is increasingly being recognized. The number of patients admitted with unstable angina has increased dramatically in recent years , approaching the numbers of patients admitted with MI [2, 24]. Early recognition and aggressive medical management can significantly reduce the high, short-term cardiovascular morbidity and mortality of these patients [25, 26]. Although cardiac markers may provide early diagnosis of patients with MI, myocardial necrosis is necessary before the markers are released, so elevations will not be detectable in patients experiencing ischemia only. Even after MI is excluded, provocative testing before discharge may be necessary to exclude significant myocardial ischemia [27, 28]. Positive early perfusion imaging has the advantage of providing early identification of patients with ischemia, allowing rapid initiation of appropriate treatment that may prevent the occurrence of MI.
Negative early perfusion imaging also provides important prognostic information. Patients with negative rest perfusion imaging can safely undergo early stress testing and therefore be eligible for early discharge, potentially resulting in significant cost savings. Although the initial costs of evaluating patients with chest pain may increase because of expansion of the emergency department coverage required for immediate perfusion imaging, total costs may be decreased through significant reductions in hospital resource costs [29, 30].
In the current study, two patients with negative rest perfusion imaging had MI. Although no previous study identified MI in a patient with negative rest perfusion imaging, the upper confidence limits in the current study (2.4%) were consistent with previous studies [8, 9]. Although SPECT imaging can increase the sensitivity for diagnosing small infarcts over that of planar imaging [32, 33], all imaging techniques have reduced sensitivities in patients with small or non–Q wave MIs [34–36]. The fact that some MIs may not be detected by early perfusion imaging underscores the importance of using a comprehensive multitiered diagnostic strategy, including serum markers, rather than relying on a single diagnostic tool for evaluating patients with a possible MI.
3.1 Pain versus pain-free injection.
We found that sensitivity did not significantly differ between patients who were and were not experiencing symptoms at the time of injection. Other studies have reported similar results. Varetto et al. found that sensitivity was high for diagnosing significant disease in patients injected 2 to 8 h after symptom relief. Bilodeau et al. found that perfusion defects persisted in 17 (68%) of 25 patients who had an MI excluded and were restudied after resolution of symptoms. Initial imaging studies using thallium-201 for diagnosis of acute ischemia found a high sensitivity, even though most patients were pain free at the time of injection [37, 38]. However, in the current study, two-thirds of the patients with negative rest studies who subsequently underwent revascularization were injected while pain free. This suggests that although the presence of symptoms is not always necessary for demonstrating myocardial ischemia, provocative testing may be required for optimal exclusion of unstable angina in patients who were injected while pain free and who had normal rest images.
3.2 Comparison with other studies.
Our results are in agreement with previous studies demonstrating that positive early rest perfusion imaging accurately identifies patients who have ongoing ischemia or infarction [8–10]. In contrast, one study found that early perfusion imaging had a low accuracy for identifying patients at risk for MI . The lower specificity in this study may have been due to the inclusion of a high percentage of patients with a previous MI, a lack of gating of images and use of planar imaging with a portable camera.
Our study differs from others in several important ways. We evaluated a large number of unselected consecutive patients, resulting in a more accurate estimate of the overall predictive ability of early perfusion imaging. Despite using broad inclusion criteria, we found sensitivities and specificities similar to those of previous studies [8–10]. Because all patients underwent imaging as part of a systematic clinical protocol, biases associated with preselecting patients were avoided, making the results more representative of the capabilities of early perfusion imaging in the clinical setting. Images were gated whenever possible, allowing simultaneous correlation of perfusion defects with wall motion abnormalities , thereby improving specificity by identifying artifacts or tissue attenuation .
3.3 Study limitations.
A limitation of early perfusion imaging is the inability to differentiate between acute infarction, previous infarction and acute ischemia. Elevations in cardiac markers allow early identification of the patient with an acute infarction. Discriminating among patients with ischemia and a previous infarction may be more difficult. Although perfusion defects are often seen in patients with a previous MI, considering them “false positive” results is not entirely correct; a history of MI is an important risk factor in patients undergoing evaluation for myocardial ischemia [1, 14–16, 20], and early discharge is not usually appropriate. Appropriate clinical correlation is often able to distinguish between those with and those without a previous MI. Not all patients underwent invasive testing, so the true sensitivity and specificity of early perfusion imaging for identification of coronary disease were not determined. However, the purpose of this study was to evaluate the accuracy of early sestamibi perfusion imaging when used in clinical practice, with all the inherent limitations. We used revascularization as a surrogate for unstable angina, an end point that has been used in previous studies of unstable angina [42, 43]. Because physicians had knowledge of the perfusion imaging results, a selection bias for angiographic referral is possible. Despite this potential for bias, revascularization implies the presence of significant coronary disease and may provide a more objective end point for unstable angina than the clinical diagnoses used in other studies.
We conclude that early rest perfusion imaging using sestamibi in low to moderate risk patients with nondiagnostic ECGs undergoing evaluation for possible myocardial ischemia can successfully identify high and low risk groups. Positive rest perfusion imaging predicted a high incidence of MI or revascularization, whereas a negative study identified patients with a low incidence of cardiovascular complications. Further investigation is necessary to determine if such a routine approach results in reduced costs compared with current standard practice.
☆ To discuss this article on-line, visit the ACC Home Page at www.acc.org/membersand click on the JACC Forum
This study was presented in part at the 69th Sessions of the American Heart Association, New Orleans, Louisiana, November 1996.
- coronary artery bypass graft surgery
- coronary care unit
- confidence interval
- creatine kinase
- electrocardiogram, electrocardiographic
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- odds ratio
- single-photon emission computed tomography
- Received January 29, 1997.
- Revision received June 17, 1997.
- Accepted July 1, 1997.
- The American College of Cardiology
- Aase O,
- Jonsbu J,
- Liestol K,
- Rollag A,
- Erikssen J
- Okada RD,
- Glover D,
- Gaffney T,
- Williams S
- Bilodeau L,
- Theroux P,
- Gregoire J,
- Gagnon D,
- Arsenault A
- Hilton TC,
- Thompson RC,
- Williams HJ,
- Saylors R,
- Fulmer H,
- Stowers SA
- Varetto T,
- Cantalupi D,
- Altieri A,
- Orlandi C
- Gushee DE,
- Holroyd BR
- ↵Braunwald E, Mark DB, Jones RH, et al. Unstable Angina: Diagnosis and Management. Clinical Practice Guideline Number 10 (amended) AHCPR Publication No. 94-0602. Rockville (MD): Agency for Health Care Policy and Research and the National Heart, Lung, and Blood Institute, Public Health Service, U.S. Department of Health and Human Services; May 1994.
- van Miltenburg-van Zijl AJ,
- Simoons ML,
- Veerhoek RJ,
- Bossuyt PM
- Tamaki S,
- Kambara H,
- Kadota K,
- et al.
- Boucher CA. Detection and location of myocardial infarction using technetium-99m sestamibi imaging at rest. Am J Cardiol 1990;66:32–5E.
- Wackers FJ,
- Lie KI,
- Liem KL,
- et al.
- Kjoller E,
- Nielsen SL,
- Carlsen J,
- et al.
- Mannting F,
- Morgan-Mannting MG
- DePuey EG,
- Rozanski A
- Anderson HV,
- Cannon CP,
- Stone PH,
- et al.