The goal of cardiac testing is to help stratify patients thought to be at risk for coronary artery disease. Risk stratification of chest pain patients in the emergency department (ED) also includes interpretation of the history, physical examination, ECG, and cardiac biomarkers. Cardiac testing encompasses diagnostic coronary angiography (invasive) or a variety of noninvasive tests. This article focuses on the noninvasive testing modalities. These include exercise stress testing, pharmacologic stress testing, myocardial perfusion imaging, stress echocardiography, and cardiac CT and MRI. The noninvasive tests can be performed on an outpatient basis, in a physician’s office, in a hospital, or for observation unit and admitted patients.
An understanding of these tests is important to the emergency physician (EP) for two primary reasons. First, patients frequently present that have undergone prior noninvasive testing. Knowing the value and limitations of that testing can be essential to the care of such patients. Second, with the recent expansion of observation medicine, it has become the responsibility of emergency physicians to choose and utilize the results of noninvasive cardiac testing in many hospitals. Noninvasive cardiac testing is an important adjunct to the broader scheme used to risk stratify chest pain patients. Use of cardiac biomarkers alone without additional noninvasive testing has not been shown to confer a low-enough risk to safely discharge a large proportion of emergency department chest pain patients.1, 2, 3
Ideally, physicians use a Bayesian model to interpret results of cardiac tests. They generate a pretest probability of disease for an individual patient based on history, ECG, laboratory results, and other clinical factors. Then by using the sensitivity and specificity of a given test for the population of interest, a post-test probability is calculated, which can guide their decision making. In day-to-day practice, this is performed more qualitatively than quantitatively. In addition, this process is reflected in diagnostic protocols for chest pain.
This article discusses the physiology, technique, interpretation, and utility of the most common noninvasive cardiac tests.
Exercise Tolerance Test
Test physiology and technique
Physical exercise places stress on the cardiopulmonary system. The physiologic response to exercise stress increases myocardial oxygen demand in response to increased heart rate and systolic blood pressure. The ECG response and development of angina in response to exercise closely correlates with myocardial ischemia due to coronary artery disease. Exercise capacity is reduced by myocardial ischemia but is also influenced by many other factors. The goal of ED exercise testing is typically to evaluate for coronary ischemia and not for exercise capacity per se. The typical clinical paradigm anticipates discharge to home of patients with a negative initial evaluation, negative cardiac biomarkers, and a negative exercise test.
Multiple protocols exist for exercise tolerance tests. A bicycle ergometer or treadmill is most often used. The goal is to increase workload incrementally to induce ischemia or until a predetermined workload is reached. One common protocol is to have the patient start walking on a treadmill and then to increase the treadmill speed and gradient until the patient experiences symptoms or ECG changes, the heart rate or blood pressure reaches preset limits, or the patient reaches a predetermined metabolic workload.
Multiple studies have validated the safety and efficacy of exercise testing in low-risk chest pain patients. Low risk, in this context, is defined as patients presenting with chest pain who remain pain-free during a 6- to 12-hour period of observation and have normal initial and repeat cardiac biomarker levels.4 It is also assumed that other serious diagnoses such as pulmonary embolism or aortic dissection are not present. Studies have also reported on the safety and efficacy of “immediate” exercise testing in low-risk patients who have normal initial ECG findings and biomarker levels and are not serially evaluated prior to stress testing.5
Certain patients do not benefit from exercise electrocardiography; this group includes patients with resting ECG abnormalities (left bundle-branch block, paced rhythm, preexcitation syndromes, or ST depressions at rest), inability to exercise, and others. Test interpretation may be compromised in patients taking certain medications such as digoxin, beta-blockers, certain calcium channel blockers, and other antihypertensive medications. Other tests, such as nuclear cardiac scanning, may be useful in this subgroup. In addition, clinicians should be familiar with contraindications to stress testing prior to ordering or performing the test. Contraindications include the following.
Sustained ventricular arrhythmias, SVT, high-grade heart block
Wellens syndrome (highly correlated with CAD and sudden death) (see Media file 1)
Aortic stenosis (hemodynamically significant)*
Serious coexisting illness (eg, pneumonia, DKA)
Active venous thromboembolic disease (DVT, PE)
Pericarditis, myocarditis, endocarditis
*May be candidates for pharmacologic stress testing
Exercise tolerance test (ETT) results are centered on the ST response, with ST depression greater than or equal to 1 mm signifying a positive test result. The probability and severity of coronary artery disease is related directly to the amount of depression and to the down-slope of the ST segment. Severity of coronary artery disease and prognosis is correlated with the lower workload at which ST-segment depression occurs.
ST-segment elevation in patients with no Q waves on the resting ECG is a rare finding, which signifies significant ischemia. ST-segment elevation in leads with previous Q waves appears to be related to the presence of dyskinetic areas or ventricular aneurysms, which does not signify acute ischemia.
Patients are instructed to terminate the test for significant chest pain, as chest pain consistent with angina constitutes a positive test. Chest pain becomes more predictive of coronary artery disease if it is associated with ST depression. Signs of poor perfusion, such as a drop in skin temperature or peripheral cyanosis and symptoms of lightheadedness or vertigo, may indicate inadequate cardiac output.
Exercise capacity frequently is reported in metabolic equivalents of task (METs). METs indicate units equivalent to the metabolic equivalent of resting oxygen uptake while sitting. An exercise capacity of 5 METs or less is associated with a poor prognosis in patients younger than 65 years. In patients with CAD, exercise capacity of at least 10 METs signifies a good prognosis with medical therapy, similar to that of coronary artery bypass surgery. An exercise capacity of 13 METs indicates a good prognosis even with an abnormal exercise ECG response.6
Systolic blood pressure at peak exertion is considered a clinically useful estimation of the inotropic capacity of the heart. A drop of systolic blood pressure below that at rest is associated with increased risk in patients with a prior myocardial infarction (MI) or myocardial ischemia. Heart rate response to exercise can be affected by left ventricle dysfunction, ischemia, cardioactive drugs, or autonomic dysfunction. Chronotropic incompetence, or failure to achieve 80% of the age-predicted maximum exercise heart rate, was associated with an 84% increase in all-cause mortality over 2 years in a 1996 Cleveland Clinic Study.7 The heart rate recovery pattern, or change in heart rate after the patient stops exercising, also has prognostic significance, as do changes in blood pressure, with a slower reversion to the patient’s baseline vital signs associated with higher long-term mortality.
The American College of Cardiology and the American Heart Association performed a meta-analysis of the diagnostic accuracy of exercise stress testing on 147 consecutively published reports involving 24,045 patients who underwent coronary angiography and ETT. The results indicated a mean sensitivity of 68% (range, 23-100%; standard deviation, 17%) and a mean specificity of 77% (range, 17-100%; standard deviation, 17%). When the studies that included patients with a previous MI were excluded, the meta-analysis involving 11,691 patients showed a mean sensitivity of 67% and mean specificity of 72% of exercise stress testing for diagnosing coronary artery disease.
The few studies that removed workup bias by having patients agree to undergo both procedures beforehand showed a sensitivity of 50% and a specificity of 90%.8 However, the purpose of stress testing in the context of the ED evaluation of chest pain is not to definitively rule out coronary artery disease. Rather, it is a short-term prognostic tool to aid in the safe disposition of patients. Other studies have shown excellent short-term (1-6 mo) cardiovascular prognosis for patients discharged from the ED or observation unit after a negative exercise test result.9
Myocardial Perfusion Imaging
Test physiology and technique
Myocardial perfusion imaging offers a method of visualizing blood flow to the heart by injection of a radioactive cardiac specific tracer. This improves the diagnostic accuracy of a stress test since it gives another method of detecting perfusion defects aside from measuring ST depression on the ECG. Myocardial perfusion imaging offers the additional advantage of estimating left ventricular function. The technique is also occasionally used independent of a stress test when evaluating patients with acute active chest pain.
Two agents are widely agents for myocardial perfusion imaging. Thallium-201 was the first agent used in clinical practice. It is a cation that acts similarly to potassium and is taken into viable cardiac myocytes. It distributes in cardiac tissue roughly in proportion to regional blood flow. It has a half-life of 73 hours. In practice, it is injected while the patient is at peak exercise or shortly after the pharmacologic stress agent is administered. Images are taken with a photon camera shortly after and then again in 3-4 hours. Defects on the initial image can represent regional ischemia or nonviable myocardium.
After the cardiac stress is discontinued, the thallium-201 redistributes and fills in areas that were underperfused due to ischemia (reversible defect). Regions of the heart that have been irreversibly damaged by previous myocardial infarction do not demonstrate resolution of the defect on the delayed image (fixed defect). In this way, the test can discriminate between regions of inducible ischemia at risk for future myocardial infarction and areas that have already been irreversibly damaged by prior myocardial infarction.
The second widely used agent is technetium-99 sestamibi. It acts as a calcium analog when taken up by the heart. It has a shorter (6 h) half-life. Once taken up by the cardiac myocytes, redistribution does not occur as it does for the thallium-201. Consequently, when performing a technetium-99 scan, a second injection is given at the time of the delayed image. Interpretation of the stress and delayed images is similar to that of thallium-201. Occasionally, stress and delayed images are obtained with a combination of the two agents.
Images are obtained by a gamma camera that rotates around the body obtaining a tomographic image. This is termed single-photon emission computed tomography (SPECT). Planar images are also available but less accurate. The images are interpreted qualitatively and can also be analyzed quantitatively by a variety of automated protocols.
Image quality can be improved by gating. This is a technique where image acquisition is timed to only occur while the heart is in diastole, offering an image with greater resolution.
A positive test is one that demonstrates reversible ischemia. Information on the size of the perfusion defect has additional prognostic value. A manual or automated scoring system may be used as well. These scoring systems have been validated and correlate with cardiac mortality. In addition, the results of MPI provide prognostic value independent of the treadmill ECG results. In many centers, initial interpretation of test results is performed by a cardiologist and reported back to the ordering physician.
Because myocardial perfusion imaging increases diagnostic accuracy of stress testing, the American College of Cardiology/American Heart Association (ACC/AHA) guidelines recommend it be used in several patient subsets. It should be used if there is any baseline ECG abnormality that would interfere with measurement of stress-induced ST-segment changes, such as LVH, bundle branch blocks, and digoxin use. In addition, it should always be used as an adjunct to pharmacologic stress testing. Finally, it is useful in higher risk patients such as those with diabetes.10
The ACC/AHA guidelines report that when both exercise and pharmacologic stress tests with SPECT imaging are compared with angiography, the test is 87-89% sensitive and 73-75% specific for significant (>50%) stenosis.11
In a 6-year follow-up study of 1,137 patients with normal thallium-201 perfusion study results, the annual rate of myocardial infarction or cardiac death was only 0.88%.12 In a meta-analysis of 14 trials with more than 12,000 patients, normal technetium-99 sestamibi imaging results were associated with a cardiac event rate of 0.6% per year.13
One difficulty that arises is when the electrocardiographic evidence and myocardial perfusion imaging on a stress test disagree. Soman et al studied 473 patients with chest pain, and two thirds of whom had abnormal ST segment response to exercise. In this study, normal technetium-99 sestamibi SPECT study results were associated with an annual mortality rate of 0.2%.14 When interpreting stress tests, more importance is generally placed on the myocardial perfusion results than the electrocardiographic results.
A technetium-99 sestamibi scan exposes a patient to approximately 8 millisieverts of radiation. This is roughly half the radiation exposure from a chest or abdomen CT. The thallium test exposure is approximately equal to that of a CT.
Equivocal results can result from poor image quality. Interference by breast tissue or the diaphragm can impair image quality in some patients.
Test physiology and technique
Another method of detecting coronary artery disease is to perform echocardiography while the heart is undergoing exercise or pharmacologically induced ischemia. Wall motion abnormalities can be visualized with the technique. The exercise is performed using a treadmill or a bicycle ergometer. If a treadmill is used, images are obtained prior to exercise and then within 60-90 seconds of completing exercise. Bicycle ergometry has the advantage of being able to perform the echocardiogram at different stages of exercise. Supine ergometry provides the most information since 4 cardiac views can be obtained. Dobutamine is the most common pharmacologic agent used in conjunction with echocardiography. Image quality can be enhanced by injection of echogenic microbubbles.
A positive stress echocardiogram is defined by stress-induced decrease in regional wall motion, decreased wall thickening, or regional compensatory hyperkinesis. In experienced hands this can have a diagnostic accuracy similar to that of nuclear stress testing. However, results are operator dependent.15
Advantages to stress echocardiography are that it is a faster test to perform than a nuclear stress test because delayed images are obtained much sooner. It has no associated radiation exposure. It is less costly than nuclear stress testing, and therefore performs well on cost analysis studies. The test can be more readily performed in an office setting.
In a meta-analysis that included data from 24 studies, Fleischmann et al found that exercise echocardiography had a sensitivity of 85% and a specificity of 77% when compared with coronary angiography. The results were felt to be similar to those for SPECT imaging.16
As stated above, the test is dependent on the experience of the operator. Obesity, lung disease, and tachycardia can limit image quality. Up to 10% of cases have inadequate image quality.
Calcium deposits are commonly found in atherosclerotic coronary plaques. The total amount of coronary calcium is predictive of future cardiac events. Cardiac computed tomography (CCT) can measure the density and extent of calcifications in coronary artery walls. The technique of CCT was established with electron beam scanners, but it has been refined and made more widely available with the introduction of multidetector scanners. The technique relies on ECG “gating” to compensate for cardiac motion. No contrast is used. The coronary lumen itself is not visualized. A related technique is cardiac CT angiography (CCTA). CCTA uses intravenous contrast material to provide direct visualization of the coronary lumen. Gating is also used to decrease motion artifact. CCTA has been shown to have good correlation with the criterion standard of conventional coronary angiography.
Coronary CTA techniques are under rapid development. A low and regular heart rate is necessary for optimal imaging and it may be necessary to administer beta-blockers to achieve an adequately low heart rate (approximately 65 bpm or less).
Test outcomes and interpretation
The amount of calcium seen in coronary vessels on CT is usually expressed as an “Agatston score,” which is based on the area and the density of the calcified plaques. A typical report provides an Agatston score for the major coronary arteries as well as a total Agatston score. A test result is considered to be positive if any calcification is detected within the coronary arteries. A positive test result is nearly 100% specific for atheromatous coronary plaque but not highly correlated with obstructive disease. A negative test result has a 96-100% negative predictive value for obstructive lesions. Agatston scores of less than 10, 11-99, 100-400, and above 400 have been proposed to categorize individuals into groups having minimal, moderate, increased, or extensive amounts of calcification, respectively.
Calcium scores greater than 1000 have been associated with significant increases in morbidity and mortality independent of other risk factors. Scores greater than 100 are consistent with a high risk (>2% annually) of a coronary event within 5 years. The amount of calcification can give, to some extent, an indication of the overall amount of atherosclerosis. In addition, a greater amount of calcification and a higher Agatston score increase the likelihood that coronary angiography will detect significant coronary artery stenosis. However, there is not a 1-to-1 relationship between a high score and the presence of coronary artery stenosis. In other words, a positive scan result indicates atherosclerosis but not necessarily significant stenosis.17
Individuals with Agatston scores greater than 400 have an increased occurrence of coronary procedures (bypass, stent placement, angioplasty) and events (myocardial infarction and cardiac death) within the 2-5 years after the test. Individuals with very high Agatston scores (>1000) have a 20% chance of suffering a myocardial infarction or cardiac death within a year. Even among elderly patients (>70 y), who frequently have calcification, an Agatston score greater than 400 was associated with a higher risk of death. In one study, patients with calcium scores greater than 1000 were found to have a relative risk of death at 5 years of 4.03 (95% confidence interval [CI], 2.52-6.40). However, calcium scores reflect overall risk and cannot be used to diagnose the presence of an obstructing lesion.18
Studies have investigated the use of CCT in the ED. These studies report a negative predictive value (NPV) of 97-100%. For example, in one study, CCT was performed in 192 patients presenting to the ED with chest pain, with an average follow-up interval of 50 months. The negative predictive value of the test was 99%. Patients with the absence of coronary artery calcium (CAC) had a 0.6% annual cardiovascular event rate. In another study of ED chest pain patients, a negative test result (absence of coronary calcification) was associated with a very low adverse event rate over a 7-year follow-up period. Increasing score quartiles were strongly correlated with risk (p<0.001).19
As with other noninvasive techniques, CCT cannot be used to identify or rule out the presence of an unstable plaque. A problem with the use of CCT is that calcification is present much more often than significant stenosis. Most patients with coronary calcification who go on to conventional invasive catheter angiography will therefore not have significant obstructive disease. CCTA may be a less invasive alternative in these cases, but there are limitations of the currently available data for CCTA. These include the fact that most reports have been based on single-center experiences and have been conducted with a subset of symptomatic middle-aged white men who had a high prevalence of CAD. Multicenter trials and studies with intermediate-risk populations are warranted.
The studies evaluating CCTA are relatively small. They have found good negative predictive value of CCTA compared with the criterion standard of catheter angiography. A normal CCTA study reliably rules out significant stenosis. Large outcome based studies of CCTA in acutely symptomatic patients are presently lacking. In one study of CCTA in low risk ED patients published in abstract form, CCTA result was considered negative if no vessel had more than a 50% stenosis and the calcium score was less than 100. Patients with a negative study result were discharged. Of the 407 discharged patients, 402 had 30-day follow up. None (0%) died from a cardiovascular cause, needed revascularization, or had an MI. This result has a 95% confidence interval of 0-0.9%. The authors concluded that low-risk chest pain patients with a negative CCTA result can be safely discharged.20
Future Directions in Testing
Magnetic resonance angiography
Cardiac magnetic resonance angiography (MRA) allows visualization of coronary vessels without radiation or contrast dye. With contrast and the addition of vasodilators or dobutamine, MRA can be used to assess myocardial viability as well. By synchronizing image acquisition with the patient’s cardiac cycle, new protocols allow the patient to breathe during the test. While cardiac MRI/MRA continues to evolve, it shows promise as the only imaging modality that can combine angiography with perfusion and wall motion assessments.
Carotid intima-media thickness
Carotid artery ultrasonography and measurement of the intima-media thickness is another area of investigation. Observational studies have shown that intima-media thickness is an independent marker of cardiovascular risk, but whether it is more accurate than traditional risk factors is unclear. However, it could prove valuable as a rapid, low-cost, low-risk test easily obtainable in the emergency department.
Combined CT studies for chest pain evaluation
Conceptually, a CT scan with intravenous contrast can combine imaging of the coronary arteries, ascending aorta, and pulmonary arteries. This allows assessment of coronary artery disease, pulmonary embolism, and disease of the thoracic aorta (dissection) with a single study. Technical aspects of the study differ than for CCTA with a wider field of view and a different protocol for the administration of intravenous contrast. At the present time, this type of imaging lacks validation for sensitivity and specificity compared with protocols specific for the various conditions under consideration.
Cardiac Testing in Women
Cardiovascular disease is the leading cause of death for women in the United States, but a considerable body of research has demonstrated that women have different patterns of coronary artery disease and different responses to cardiac testing than their male counterparts. Women are more likely to have nonobstructive or single-vessel disease when compared with men, which decreases the diagnostic accuracy of stress testing. For example, treadmill testing in one meta-analysis was shown to have a sensitivity and specificity of 61% and 70%, respectively, for women compared with 72% and 77%, respectively, for men.21
Calcium scoring is limited because women tend to have 3- to 5-fold greater mortality rates for a given calcium score than men, suggesting that separate guidelines for interpreting scores in women should be developed.
SPECT imaging is technically limited in women because breast tissue and relatively small left ventricle size can generate false-positive results. Technetium is less prone to attenuation artifacts than thallium and thus has higher specificity. The American Heart Association has recommended that the exercise tolerance test is still the initial test of choice for a low-risk or intermediate-risk symptomatic woman with no contraindications.22
Pharmacologic Stress Testing
Test physiology and technique
Pharmacologic stress testing differs from exercise testing in that it does not rely on the patient’s own ability to increase cardiac oxygen demand. Rather, the patient can remain at rest while the heart’s response to a drug is measured. The most widely available pharmacologic agents for stress testing are dipyridamole (Persantine), adenosine, and dobutamine.
Dipyridamole and adenosine are cardiac vasodilators and rely on a unique mechanism of action to demonstrate coronary stenosis. By dilating coronary vessels, they lead to an increased blood velocity and flow rate in normal coronary vessels. However, stenotic vessels do not respond as vigorously to these drugs. Because of the heterogeneous dilation, a “steal” of flow occurs. This can be seen as perfusion defects in cardiac nuclear scans or as ST-segment changes.10 Dobutamine is a cardiac inotrope and chronotrope and exerts an effect on the heart similar to that of exercise.
Pharmacologic stress testing is indicated for patients who would be unable to adequately perform an exercise stress test. An exercise test is considered inadequate when a patient cannot reach 85% of predicted maximum heart rate, or reach a workload of 5 metabolic equivalents of task (METs) for 3 minutes. A pharmacologic test is preferred over an exercise test in patients with aortic stenosis, left bundle branch block, a paced rhythm, recent MI, and severe hypertension even if they were able to exercise adequately.23
The pharmacologic stress test is interpreted in a similar manner to the exercise stress test (see above). Additionally, myocardial perfusion imaging is advisable in all patients undergoing pharmacologic stress testing.
Pharmacologic stress testing with nuclear imaging is equivalent to an exercise stress test with nuclear imaging at detecting coronary artery disease. Note however, that since patients undergoing pharmacologic stress testing tend to have more comorbidities, the post-test probability of disease is higher in patients who have undergone a pharmacologic test. A normal pharmacologic stress test result confers a 1-2% per year cardiac event rate, whereas a normal exercise test result with nuclear imaging has a rate less than 1% per year.24
Theophylline can reduce ischemic changes on the ECG with vasodilator stress testing. Caffeine has been reported to have a similar effect. However, a recent study demonstrated that one cup of coffee, one hour prior to stress testing did not attenuate the results of adenosine nuclear imaging.25 Calcium channel blockers, beta-blockers, and nitrates can also alter perfusion defects on pharmacologic stress tests and therefore ideally should be withheld for 24 hours prior to pharmacologic stress testing. Dipyridamole and adenosine can lead to bronchospasm; they are generally avoided in patients with severe reactive airway disease or active wheezing. Dobutamine is safe to use in these patients.
Noninvasive cardiac testing is used as part of a broader scheme of risk stratification for patients with possible acute coronary syndromes. Several tests exist, and each has unique advantages and disadvantages. Patient characteristics and local resources dictate which of the cardiac tests are chosen. Variability exists in how well noninvasive cardiac tests correlate with angiographic findings. Despite this variability, most of the tests are useful for determining short-term risk of myocardial infarction and death.
Noninvasive cardiac tests are improving as new diagnostic technologies and methods are being developed. As future studies reveal the true diagnostic characteristics and capabilities of these tests, ED physicians can better assess patients’ risk of coronary artery disease based on their previous test results and more effectively recommend further testing and interventions.
As with all diagnostic tests, none of the cardiac tests are ideal. They are useful as part of a risk stratification scheme, but, with the current state of diagnostic testing, some cases of serious coronary disease will always be missed.