CHAPTER 13: ELECTROCARDIOGRAPHIC EXERCISE TESTING

For 50 years, exercise testing has served as the backbone of the noninvasive evaluation of coronary artery disease (CAD) and as an essential component of a clinician’s tool kit. Its rationale is compelling: it is simple, widely available, and inexpensive. Underused as stand-alone testing, a sophisticated and critical evaluation of the ECG and the patient during exercise and recovery, and a true maximal exercise effort, often provides the test sensitivity required for decision making without adjunctive nuclear or echocardiographic imaging. When adjunctive imaging is used, this same discriminatory power can enhance accuracy and confidence in the combined study.

When standardized exercise testing was developed by Master and colleagues in 1929,¹ the gold standard for exercise test accuracy was the development of angina and the occurrence of cardiac events.^2,3 The advent of coronary angiography in 1958⁴ created a new gold standard, easily performed after exercise testing to study the correlation between exercise test results and coronary luminal anatomy. The following year, Prinzmetal and coworkers identified that ischemia could be caused by coronary vasospasm in the presence or absence of obstructive anatomic disease.⁵ Coronary vasospasm was thought to be infrequent, however, and the accuracy of exercise testing continued to be measured by the presence of obstructive CAD on coronary angiography. A few decades later, the development of thallium- and technetium-based single photon emission computed tomography (SPECT) imaging allowed myocardial perfusion imaging (MPI) to be considered a gold standard.

Studies assessing the physiologic effect of coronary stenoses with fractional flow reserve (FFR) call the use of coronary angiography to predict ischemia, and thus serve as a gold standard, into question. Meanwhile, recent advances in quantitative positron emission tomography (PET) imaging create the opportunity to accurately measure global myocardial blood flow, and cardiac magnetic resonance imaging (MRI) has been shown to accurately detect ischemia that is solely subendocardial. These two techniques may demonstrate ischemia associated with coronary lesions that are angiographically less than 50% stenotic. Thus, what is the proper gold standard against which exercise testing should be compared? As exercise testing detects ischemia that is predominately subendocardial, the assessment of global coronary flow by quantitative pharmacologic PET or ischemia by cardiac MRI should serve as the new gold standards. If ischemia in the distribution of a particular coronary artery needs to be evaluated, FFR can serve as the gold standard. Our previous understanding of the accuracy of exercise testing to detect ischemia must be tested against these new gold standards. The bulk of the extensive literature assessing the sensitivity and specificity of exercise stress testing used coronary angiography as the gold standard. As such, the basis of our understanding of exercise test sensitivity, specificity, and predictive values may be called into question. As new correlative studies have yet to be performed, we must rely on previous studies to best understand the accuracy of exercise testing, understanding that future research may challenge some of our assumptions.

Gibbons and fellow authors of the American College of Cardiology (ACC) and American Heart Association (AHA) 2002 guideline for exercise testing reviewed individual studies and meta-analyses evaluating the sensitivity and specificity of exercise testing.⁶ They cited a meta-analysis by Detrano and colleagues of 147 exercise testing studies that found a mean sensitivity and specificity of 68% and 77%, respectively.⁷ Sensitivity and specificity were altered by the selection of which cut point is used to discriminate between those with and without disease, particularly when measuring a continuous variable such as assessment of CAD with an exercise test. The 147 studies discussed above used 1 mm (1 mV) of horizontal or downsloping ST-segment depression as their cut point. The Bayes theorem of conditional probability states that the patient’s pretest likelihood of disease affects the post-test likelihood of disease following a test. For example, if a patient with a low (eg, 5%) pretest likelihood of disease undergoes a test with a sensitivity of 70% and a specificity of 90% and the test is positive, the post-test likelihood of disease is only 27%. However, if the pretest likelihood is 50%, a positive test yields a post-test likelihood of disease of 80%.⁸

A change in a cut point to diagnose disease alters test sensitivity and specificity. If a more lenient cut point is used to diagnosis diabetes, such as a lower fasting blood sugar, sensitivity will increase but specificity will necessarily decrease and do so inversely with the increase in sensitivity. Sensitivity and specificity theoretically do not change with the prevalence of disease in a population, whereas positive and negative predictive values do change. Performing exercise testing in a population with low disease prevalence, such as asymptomatic young men or women, will result in a low positive predictive value, making testing unhelpful in such a cohort. However, exercise testing in an older population with symptoms suggestive of CAD would be doing so in a population with an intermediate likelihood of CAD, yielding a higher and clinically acceptable positive predictive value.

Disease severity does influence sensitivity and specificity.^9,10 If those with the disease have a more severe form, the diagnostic test will be more accurate. This is relevant for exercise testing because patients now present with less advanced CAD, more often single vessel than multivessel. Ischemia is thus less likely to be demonstrated on exercise testing and its manifestations are often less obvious. In the case of exercise testing, less ST depression may be apparent. In this era of less CAD severity, it is appropriate to reconsider the traditional cut points used in the characterization of an exercise test as positive or negative.

Prior to delving further into exercise testing, a review of exercise physiology, adapted from Lipinski and Froelicher, is in order.¹¹

Exercise Physiology

Two basic principles of exercise physiology are central to exercise testing. The first is a physiologic principle: total body oxygen uptake and myocardial oxygen uptake are distinct in their determinants and in the way they are measured or estimated (Table 13–1). Total body or ventilatory oxygen uptake (volume oxygen consumption [VO₂]) is the amount of oxygen extracted from inspired air as the body performs work. The determinants of VO₂ are cardiac output and the peripheral arteriovenous oxygen difference. Maximal arteriovenous difference is physiologically limited to roughly 15 to 17 mL/dL. As maximal arteriovenous difference behaves more or less as a constant, maximal oxygen uptake serves as an indirect estimate of maximal cardiac output.

Myocardial oxygen uptake is the amount of oxygen consumed by the myocardium. The determinants of myocardial oxygen uptake include intramyocardial wall tension (left ventricular pressure and end-diastolic volume), contractility, and heart rate. Myocardial oxygen uptake can be estimated by the product of heart rate and systolic blood pressure (the double product or rate pressure product). This information is valuable clinically because exercise-induced angina often occurs at the same myocardial oxygen demand and thus double product. The higher the double product achieved, the better the myocardial perfusion and prognosis. Therefore, it is not surprising that the double product during exercise testing has long been known to be a significant independent predictor of myocardial ischemia¹² and prognosis.^13,14

The second principle is one of pathophysiology: considerable interaction takes place between the exercise test manifestations of abnormal myocardial perfusion and function. The electrocardiographic response and development of angina are closely related to myocardial ischemia (and extent of CAD), whereas exercise capacity, systolic blood pressure, and heart rate response to exercise can be determined by the presence of myocardial ischemia, myocardial dysfunction, or peripheral responses. Exercise-induced ischemia can cause cardiac dysfunction, resulting in exercise impairment and a blunted systolic blood pressure response.

Metabolic Equivalents

Because exercise testing fundamentally involves the measurement of work, there are several key concepts regarding work. The common biologic measure of total body work is oxygen uptake, which is usually expressed as a rate (making it a measure of power) in liters per minute. The metabolic equivalent of a task (MET) is a term used clinically to express the oxygen requirement of the work rate during an exercise test. One MET is equal to the resting metabolic rate (~3.5 mL of O₂/kg/min) and a MET value achieved from an exercise test is a multiple of the resting metabolic rate, either estimated from the maximal workload achieved or measured directly by oxygen uptake. It also represents a method of reporting work performed with different stress test protocols.

Acute Cardiopulmonary Responses to Exercise

Figure 13–1 summarizes the hemodynamic response to exercise. The cardiovascular system responds to acute exercise with a series of adjustments that ensure the following:

• Exercising muscles receive blood supply commensurate to their metabolic needs.

• Heat generated by the muscles is dissipated.

• Blood supply to the brain and heart is maintained.

This response requires a major redistribution of cardiac output, occurring in tandem with a number of local metabolic changes. The usual measure of the capacity of the body to deliver and use oxygen is the maximal oxygen consumption (VO₂max). Thus, the limits of the cardiopulmonary system are defined by VO₂max, which can be expressed by the Fick principle:

Cardiac output must closely match ventilation in the lung to deliver oxygen to the working muscle. VO₂max is determined by the maximal amount of ventilation (volume of expired gas [VE]) moving into and out of the lung and by the fraction of this ventilation that is extracted by the tissues:

where VE is minute ventilation, and FiO₂ and FeO₂ are the fractional concentration of oxygen in the inspired and expired air, respectively. To measure VO₂ accurately, CO₂ in the expired air (carbon dioxide elimination [VCO₂]) must also be measured; the main purpose of VCO₂ in this equation is to correct for the difference in ventilation between inspired and expired air. Because the rate of increase in VCO₂ relative to the work rate or ventilation parallels the severity of heart failure, VCO₂ is a valuable measurement clinically and a powerful prognostic marker for chronic heart failure patients.

Cardiopulmonary limits (VO₂max) are therefore defined by cardiac output, which describes the capacity of the heart to function as a pump, and the arteriovenous oxygen difference, which characterizes the capacity of the lung to oxygenate the blood delivered to it as well as the capacity of the working muscle to extract this oxygen from the blood.

Chief Determinants of Maximal Oxygen Consumption

Heart Rate

Sympathetic and parasympathetic nervous system influences underlie the increase in heart rate, the cardiovascular system’s first response to exercise. Vagal withdrawal is responsible for the initial 10 to 30 beats/min increase, whereas the remainder is thought to be largely caused by increased sympathetic outflow. Of the two major components of cardiac output, heart rate and stroke volume, heart rate is responsible for most of the increase in cardiac output during exercise, particularly at higher levels. Heart rate increases linearly with workload and oxygen uptake.

Stroke Volume

The product of stroke volume (the volume of blood ejected per heartbeat) and heart rate equals cardiac output. The stroke volume is equal to the difference between end-diastolic and end-systolic volumes. Thus, greater diastolic filling (preload) will increase stroke volume. Alternatively, factors that increase arterial blood pressure (afterload) will resist ventricular outflow and result in reduced stroke volume. During exercise, stroke volume increases up to about 50% to 60% of maximal exercise capacity, after which increases in cardiac output are a result of further increases in heart rate.

End-Systolic Volume

End-systolic volume depends on two factors: contractility and afterload. Contractility describes the forcefulness of the heart’s contraction. Increasing contractility reduces end-systolic volume, which results in a greater stroke volume and thus greater cardiac output. Contractility is commonly quantified by the ejection fraction, the percentage of blood ejected from the ventricle during systole. Afterload is a measure of the force resisting the ejection of blood by the heart. Increased afterload (aortic pressure) results in a reduced ejection fraction and increased end-diastolic and end-systolic volumes. During dynamic exercise, the force resisting ejection in the periphery (total peripheral resistance) is reduced by vasodilation, secondary to the effect of local metabolites on the skeletal muscle vasculature. Thus, despite even a five-fold increase in cardiac output among normal subjects during exercise, mean arterial pressure increases only moderately.

Arteriovenous Oxygen Difference

Oxygen extraction by the tissues during exercise reflects the difference between the oxygen content of the arteries (~18–20 mL O₂/100 mL at rest) and the oxygen content in the veins (~13–15 mL O₂/100 mL at rest), yielding a typical arteriovenous oxygen difference (A-VO₂) at rest of 4 to 6 mL O₂/100 mL, approximately 23% extraction. During exercise, this difference widens as the working tissues extract greater amounts of oxygen; venous oxygen content reaches very low levels, and A-VO₂ can be as high as 16 to 18 mL O₂/100 mL with exhaustive exercise. Some oxygenated blood always returns to the heart, however, as smaller amounts of blood continue to flow through metabolically less active tissues that do not fully extract oxygen. The A-VO₂ is generally considered to widen by a relatively fixed amount during exercise, and differences in VO₂max are predominantly explained by differences in cardiac output. Some patients with cardiovascular or pulmonary disease, however, exhibit reduced VO₂max values that can be attributed to a combination of central and peripheral factors. For example, in a patient with chronic obstructive pulmonary disease, impaired oxygen uptake will necessarily diminish VO₂max.

Determinants of Arterial Oxygen Content

Arterial oxygen content is related to the partial pressure of arterial oxygen, which is determined in the lungs by alveolar ventilation and pulmonary diffusion capacity and in the blood by hemoglobin content. In the absence of pulmonary disease, arterial oxygen content and saturation are generally normal throughout exercise. Patients with symptomatic pulmonary disease often neither ventilate the alveoli adequately nor diffuse oxygen from the lung into the bloodstream normally, resulting in a decrease in arterial oxygen saturation during exercise.

Determinants of Venous Oxygen Content

Venous oxygen content reflects the capacity to extract oxygen from the blood as it flows through the muscle. It is determined by the amount of blood directed to the muscle and by capillary density. Muscle blood flow increases in proportion to the increase in work rate and thus oxygen requirements. The increase in blood flow is brought about not only by the increase in cardiac output, but also by a preferential redistribution of the cardiac output to the exercising muscle. Locally produced vasodilatory mechanisms along with possible neurogenic dilatation resulting from higher sympathetic activity reduce local vascular resistance and mediate the greater skeletal muscle blood flow. A marked increase in the number of open capillaries reduces diffusion distances; increases capillary blood volume; and increases mean transit time, facilitating oxygen delivery to the muscle.

As with any cardiac test, patient selection is critical and patient safety of foremost concern. In their seminal 1971 study of 170,000 exercise tests, Rochmis and Blackburn observed a mortality rate of approximately 1 in 10,000 exercise tests.¹⁵ Eight subsequent studies summarized in 1993 found the mortality rate had decreased to 0 to 0.5 per 10,000 tests.¹⁶ The clinician supervising the test, be it a physician, nurse, exercise physiologist, or another professional, should be very comfortable in his or her ability to interpret a rest and exercise electrocardiogram (ECG) for ischemia and arrhythmia, recognize exercise-induced symptoms of ischemia and cardiac failure, and handle an exercise-induced emergency. Given the significant potential for harm in patients with valvular heart disease, particularly severe aortic stenosis, the supervising clinician should also be skilled in cardiac auscultation, to be performed prior to exercise testing.

Qualifications of Testing Clinicians

In the United States, exercise testing is increasingly supervised by nonphysicians. As reviewed by Myers and colleagues,¹⁷ the AHA stated in 1979 that an experienced nonphysician could conduct exercise testing, although a physician must be immediately available during testing.¹⁸ In 1990, a multispecialty task force of the ACC, AHA, and American College of Physicians recommended specific clinical competencies for physicians and nonphysicians in order to perform exercise testing.¹⁹ Subsequent statements by these organizations, as well as the American College of Sports Medicine and American Association of Cardiovascular and Pulmonary Medicine, further defined the level of supervision during which patients can be safely tested by nonphysician providers and the detailed training required. Central to these recommendations are the risk stratification of patients prior to exercise testing and the immediate availability of a physician during patient testing. Personal supervision by a physician in the room is recommended for high-risk patients as listed in the bullets below, which have been adapted from Myers and colleagues.¹⁷ These patients should be identified by careful pretest evaluation and good clinical judgment.

• Moderate to severe aortic or mitral valve stenosis with or without symptoms

• Hypertrophic cardiomyopathy: risk stratification and exercise gradient assessment

• History of malignant or exertional arrhythmias, sudden cardiac death

• History of exertional syncope or presyncope

• Intracardiac shunts

• Genetic channelopathies

• Within 7 days of myocardial infarction or acute coronary syndrome

• New York Heart Association class III heart failure

• Severe left ventricular dysfunction

• Severe pulmonary arterial hypertension

• In the broader context of potential instability resulting from noncardiovascular comorbidities, (eg, frailty, dehydration, orthopedic limitations, chronic obstructive lung disease)

All exercise laboratory staff assisting in exercise testing should be credentialed in basic life support. Supervising nonphysicians should be certified in Advanced Cardiac Life Support. If supervising physicians are not certified in Advanced Cardiac Life Support, they should be competent in identifying and handling such emergencies based on past and continued training and experience. Given a dearth of comprehensive continuing education pathways to achieve expertise in exercise testing, the training of nonphysicians, such as registered nurses, nurse practitioners, and physician assistants can be challenging. Previous cardiac experience can be leveraged with available continuing medical education courses, textbooks, and articles, as well as extensive on the job training and supervision to develop competence. Both nonphysicians and physicians should perform a minimum of 200 exercise tests during training to meet the level 1 Core Cardiology Training Symposium standard established by the ACC in 2008.²⁰ In general, a minimum of 50 tests should be performed annually to maintain and enhance competency. With substantial experience, nonphysicians may perform basic interpretation of the exercise test with the final responsibility for the report resting with the physician. Any physician supervising and ultimately responsible for the exercise test should meet ACC level 1 guidelines.²⁰ Physicians who are immediately available during testing by nonphysicians should either meet these guidelines or be skilled in emergency medicine.¹⁷

Exercise Equipment

Exercise testing can be performed with bicycle ergometry or treadmill testing. The latter is much preferred if space in the exercise laboratory allows, because higher workloads can be obtained with treadmill compared to bicycle exercise given the greater muscle mass involved.

Given the low but finite risk of prolonged myocardial ischemia, infarction, or arrhythmia, and of cardiac arrest, a “crash cart” should be immediately available, and supervising clinicians should have the knowledge to use the medications therein. A defibrillator should also be immediately available, and staff should be trained in its use. Laboratory staff should be well versed in the laboratory’s predefined plan to respond to an emergency. In an outpatient facility, this includes an understanding of who will directly attend to the patient, who will record unfolding events, and who will call for emergency medical transport personnel. When an event does occur, a postevent debriefing presents the opportunity to discuss opportunities for improvement.

Pretest Preparation

Patient preparation is key and should begin on test scheduling. The patient should be provided with digital or printed information on what to expect. They should be advised to wear clothes and shoes fit for exercising. Because caffeine is a competitive antagonist of adenosine receptors and thus blocks the effect of dipyridamole, adenosine, and regadenoson, it should be withheld for at least 12 hours if there is a possibility of the patient undergoing vasodilator stress the same day. However, we are unaware of investigative data on the impact of caffeine on stand-alone exercise testing. Because the vast majority of persons consume caffeine regularly, testing caffeinated patients would be akin to testing them in their typical physiologic state. If there is little or no chance that a patient would undergo vasodilator stress the same day as an exercise test, prohibition against caffeine use prior to exercise testing is not required. We have also found that a light meal prior to the exercise test, if desired, does not present problems.

Tobacco has traditionally been prohibited for 3 hours prior to testing. Tobacco can potentiate vasospasm, which can result in a positive exercise test in the absence of obstructive CAD. However, if the patient smokes regularly, a test without a tobacco prohibition would reflect their day-to-day natural state. The prohibition can be implemented or relaxed depending on the clinical question to be answered.

Avoiding vigorous exercise on the day of exercise testing is appropriate to enhance the patient’s ability to achieve maximum exercise on the test.

If the exercise test is being performed to determine if a patient has CAD and consequent ischemia, antianginal medications should be held to minimize their anti-ischemic impact. As best observed in studies of patients undergoing SPECT imaging, beta-blockers, nitrates, and calcium channel blockers can all turn a test from ischemic to nonischemic.^21,22 A practical approach to discontinuing medications for diagnostic testing is to hold beta-blockers for 24 hours and hold nitrates and calcium channel blockers the day of the study. Patients should be instructed to bring these medications to the test and to take them following the treadmill test. This is ideally accomplished while the patient is still in the exercise laboratory in order to facilitate compliance in restarting their medication.

Holding medications is particularly important when using exercise testing for patients being evaluated from the emergency department for chest pain. If a patient is already on antianginal therapy on arrival for exercise testing from the emergency department or elsewhere, sensitivity will be decreased. If the test is negative for ischemia, the reporting clinician may consider noting in the report either that test sensitivity may be decreased by the patient’s medical regimen or that ischemia was not seen at the level of exertion or double product achieved. If the test is being performed for prognostic reasons—that is, if CAD is known to be present and the clinician wishes to determine if the patient is having ischemia on their current medical regimen—medications need not be withheld.

On arrival, staff should review the test protocol and obtain informed consent by informing the patient of the risks and benefits. Instructions should be provided on vocalizing if chest pain or other ischemic symptoms occur prior to, during, or after exercise.

Exercise Protocols and Testing

The most commonly used multistage treadmill protocols in the United States are those developed by Bruce and Ellestad, often regarded as the founders of exercise treadmill testing in the United States. Both protocols begin with what they regarded as a warm-up stage (stage 1), 1.7 miles/hour at 10% grade. For patients with limited ability to exercise, the Bruce protocol can be modified to include a stage 0 in which the patient exercises at 1.7 miles/hour at 0% grade for 3 minutes prior to moving into stage 1 of the standard Bruce protocol. Ideally, a protocol should be used such that patients exercise for 5 to 6 minutes or more.²³ A comparison of Bruce and Ellestad treadmill protocols is shown in Fig. 13–2.²⁴ Pollock and colleagues compared the physiologic parameters of these two protocols in a study of 51 men who underwent both tests.²⁴ Peak VO₂max, METs, heart rate, and systolic and diastolic blood pressures obtained were similar. As with other multistage treadmill protocols, VO₂max correlated directly with exercise duration (Fig. 13–3). Time to achieving peak exercise was about 1 minute sooner with the Ellestad protocol, likely secondary to its 2-minute stages following stage 1 compared to the 3-minute stages of the Bruce protocol.²⁵ Other protocols are available, including those by Naughton and Balke. Exercise duration correlates with VO₂max in each and achieves similar results.

A brief cardiac examination, including auscultation of the heart and lungs, should be performed prior to testing, and the patient should be questioned to determine the reason for the test and any symptoms that may have prompted the referral. Particularly important are any symptoms the patient may have been experiencing that could be cardiac in nature. If the symptoms being evaluated clinically appear consistent with ischemia and the patient develops these during exercise, test termination is generally indicated. Likewise, if the patient describes symptoms prior to exercise testing that are consistent with new onset unstable angina, which have not resolved, invasive coronary angiography or computed tomography (CT) coronary angiography will often be a safer choice than exercise testing. A careful history and clinical judgment are a prerequisite to the safe execution of exercise testing.

ECG lead placement should ensure proper contact in order to minimize artifacts, which can obfuscate the ability to interpret ischemia. The Mason-Likar “torso” ECG used for exercise testing distorts the axis rightward, which can result in a loss of inferior Q waves masking a previous inferior myocardial infarction (MI).²³ Nevertheless, this is the ECG that will be used as the comparator to the exercise ECG and should be reviewed prior to exercise testing. Ideally, a standard 12-lead ECG with electrodes placed on the limbs should be performed prior to exercise testing. Previous records, often accessed via an electronic medical record system, should be reviewed for pertinent symptoms and history, including past exercise testing and a recent resting ECG. Electrocardiographic tracings from the most recent exercise test allow a direct comparison of the timing of onset and extent of previous ECG changes. Comparing exercise duration between the previous and current test is particularly helpful. For example, if a patient undergoes a follow-up exercise test and the clinician performing and/or interpreting the test has expended the effort to review the patient’s previous exercise duration and electrocardiographic response and finds them either unchanged or to have changed between the two tests, this information will be particularly helpful and appreciated by the referring clinician.

During this review, the patient and records should be carefully interrogated for clinical conditions in which the risk of exercise testing outweighs the benefits and is contraindicated (Table 13–2).^2,6 Relative contraindications are listed in Table 13–3.^2,26 Indications for the test should be assessed and directly matched to the most recent societal Appropriate Use Criteria.²⁷ If exercise testing does not appear to be indicated to address the patient’s clinical condition, the ordering physician should be contacted to discuss their understanding of the clinical situation and potential alternatives to exercise testing.

The patient’s usual level of activity should also be assessed to determine if the patient would be expected to achieve ≥ 85% of maximum predicted heart rate (MPHR) and if so, to determine whether to modify the planned protocol. Unfortunately, laboratories are increasingly using pharmacologic rather than exercise stress as the stressor for MPI. As American Society of Nuclear Cardiology guidelines recommend, pharmacologic stress is only appropriate if the patient is unable to achieve ≥ 85% MPHR and five METs.²⁸ Although pharmacologic MPI provides adequate perfusion images to assess ischemia, MPI performed with exercise provides key additional opportunities to assess ischemia, including exercise duration, assessment of symptoms, and electrocardiographic changes. With exercise, the expense and risks of the pharmacologic stress agent is also not incurred.

A review of the resting ECG should include an evaluation for ischemia, conduction defects, and arrhythmias, and it should be compared to prior tracings if they are available. Obtaining an ECG during hyperventilation should be avoided before testing. Subjects with and without CAD can exhibit ST-segment changes with hyperventilation; therefore, hyperventilation to identify false-positive responders is no longer considered helpful.

Once exercise is initiated, the technologist handling the treadmill and the supervising clinician should focus on the patient, the ECG tracings on the monitor, the intermittent 12-lead ECG obtained at the end of each stage, and the patient’s symptoms and appearance. If changes suggestive of ischemia develop on the continuous monitor, a 12-lead ECG should be immediately performed. Electrocardiographic changes should also elicit a question to the patient about any associated symptoms.

For a stand-alone exercise test, if a patient develops convincing evidence of ischemia by ECG, symptoms, or both, the test can be terminated, because the question of the presence of ischemia has been answered.³ If ischemia is not convincingly observed, the patient should continue until they can safely or comfortably go no further. A Borg-perceived exertion scale of 6 to 20 (Table 13–4) posted in the patient’s view allows the patient to quantitate their level of exertion.²⁹ This is particularly useful during nuclear stress testing when judging how much more time the patient can exercise in order to gauge when to inject the radioisotope. A Borg scale of ≥ 15 suggests the anaerobic threshold has been met and > 18 generally indicates maximal exercise.²³ In many patients, particularly those younger and not on medication, a zero can be added to the Borg score to estimate the patient’s exertional heart rate. This is by design.²⁹

Reassurance

The clinician should talk intermittently to patients during the test, reassuring them as to their progress and asking how they feel. Toward the end of each stage, patients should be informed of the upcoming increase in treadmill speed and incline and asked if they can continue. Conversation during the test is reassuring, particularly for those undergoing testing for the first time.

Maximum Heart Rate

Although age-adjusted MPHR has been assessed since the early years of treadmill testing, it has proved too variable among individuals to serve as a suitable end point to terminate an exercise test. If safe, patients should be encouraged to exercise as long as they are safely able. Exercise testing is meant to be symptom limited, not terminated by an arbitrary age-related heart rate of 85% or 100% of predicted maximum. Stopping exercise prematurely once 85% of an estimated maximal heart rate is achieved decreases exercise testing sensitivity and minimizes the opportunity to assess ischemia electrocardiographically and with adjunctive imaging.³⁰

Obtaining a heart rate of at least 85% of MPHR is particularly relevant for exercise testing with nuclear or echocardiographic testing. Bairey and coworkers compared patients undergoing exercise MPI with a normal perfusion scan relative to whether they achieved 85% of MPHR. The annual event rate of death, nonfatal MI, or late revascularization was 1.9% for those who reached ≥ 85% MPHR compared to 8.6% for those who achieved < 85%.³¹ In those who had not achieved 85% MPHR, the predictive value of MPI was minimized as a normal MPI study in the setting of a submaximal heart rate failed to adequately detect ischemia and thus failed to serve as an adequate risk stratifier. Iskandrian and colleagues found MPI sensitivity to be about 20% less among patients who achieved a heart rate < 85% of MPHR compared to those who achieved ≥ 85% MPHR or who stopped on reaching an ischemic end point. Given that ischemia as assessed by MPI is heart rate dependent, ischemia detection by ECG should be heart rate dependent as well.^32,33 Cumming demonstrated this concept in his comparison of submaximal to maximal bicycle exercise testing in 63 subjects who had an abnormal ECG response during maximal exercise or during recovery from maximal exercise.³⁴ Although all were ischemic by ECG at maximum exercise or in recovery, ischemic ECG changes were present at heart rates of 85% or less of MPHR in only 46% of subjects.

Noting that 40% of 100 consecutive laboratories applying for nuclear cardiology accreditation used 85% MPHR as their primary termination criteria for ending an exercise test, Jain and colleagues evaluated 232 of their own patients who had ischemic ECG changes at peak exercise and who exercised for more than 1 minute beyond the time at which they achieved ≥ 85% MPHR.³⁰ At 85% MPHR, the ECG was ischemic in only 144 (62%) patients. Mean ST-segment depression was 1.2 and 2.3 mm at 85% MPHR and peak exercise, respectively. They thus emphasized the admonition of Bruce and colleagues in the first description of Bruce’s multistage treadmill protocol in 1963, that the patient should exercise until “exhausted by fatigue.”³⁵ Achieving either maximum effort or an ischemic end point is key for exercise testing performed with or without imaging.

An anecdote of how the commonly used, easy-to-remember 220 – age formula was created to estimate MPHR is both interesting and highlights its limitations. On a plane to a medical convention in 1970, Haskell and Fox realized that in about 10 studies of young and middle-aged men undergoing exercise testing, maximum heart rate achieved could be roughly estimated by the formula 220 – age.³⁶ The paper in which they define the formula describes its use not to estimate MPHR on an exercise test but to develop a patient’s exercise prescription.^30,37 Formulas with a greater evidence base are recommended to estimate MPHR. Tanaka and colleagues, for example, found that 220 – age results in an underestimation of maximum heart rate, particularly in older patients.³⁸ In a meta-analysis of 351 studies involving 18,712 patients, the empirically derived formula of 208 – 0.7 × age best predicted the maximum heart rate. The formula was predictive in men and women and independent of habitual activity level. They then tested the formula prospectively in 514 subjects, finding excellent prediction of maximum heart rate. Evidence-based tables of age-related maximum heart rates can also be found in Sheffield³⁹ and Ellestad.⁴⁰ Of these examples, the formula of Tanaka and colleagues is preferred.

Walking on the Treadmill

Once exercise has started, patients should use the side rails of the treadmill for balance and safety but not to minimize exertional effort by a tight hold. Some patients will need to be briefly supported by the laboratory team as they start exercising to be certain they are comfortable walking on the belt of the treadmill. In patients unfamiliar with walking on a treadmill, the clinician will often need to ask that they stand erect, lengthen their stride, and move toward the front of the treadmill. This is illustrated in Fig. 13–4.

Symptoms should be carefully solicited and recorded. Chest pain that is intermittent and of only a few seconds’ duration is far less suggestive of ischemia than chest pain that increases with exercise. If premature ventricular contractions or premature atrial contractions occur during exercise or recovery, patients should be asked if they feel them. In this case, patients can be informed of the cause of the feeling and likely reassured.

The clinician and assisting staff member should work to obtain artifact-free ECGs during exercise and recovery. This is particularly important for ECGs at peak exercise and in immediate recovery. Additionally, obtaining diagnostic and artifact-free ECGs during exercise may require obtaining ECGs more frequently than at the end of each stage. Blood pressure should be measured toward the end of each stage and when patients report or acknowledge symptoms. Although automated sphygmomanometers are available, manual blood pressure measurement is the most reliable method. It is often difficult to obtain an accurate blood pressure when patients are walking or jogging at speeds of greater than 4 mph on the treadmill. If this is the case, the general appearance of patients—their skin color and the apparent strength and vigor of their walk—will provide clues to the adequacy of their hemodynamic response.

Recovery Protocols

Two common protocol options exist for the recovery period: an immediate stop after peak exercise or a short cool-down period of 1 to 2 minutes of walking. An immediate stop provides the opportunity to obtain an artifact-free immediate postexercise ECG while the patient is still standing on the treadmill. The patient is then immediately laid supine or semirecumbent on a gurney or its equivalent. The cool-down option of slowing the patient down over 1 to 2 minutes can delay⁴¹ or eliminate the appearance of ST-segment depression in recovery. The act of placing a patient supine immediately postexercise enhances ischemic electrocardiographic abnormalities in recovery, likely as a result of the supine position increasing venous return resulting higher left ventricular end-diastolic pressure and thus higher wall tension and oxygen consumption of the myocardium. A representative series of ECG tracings from a patient instructed to move from sitting to laying down at various times during recovery serves as an example of inducing worsening ST-segment depression with the supine position (Fig. 13–5).² Given the advantages of an immediate stop and a swift move to the supine position, this recovery protocol is recommended for most patients undergoing stand-alone exercise testing.

A rapid stop may occasionally induce a vagal response. This occurs particularly in young persons exercising to exhaustion and may result in lightheadedness or presyncope. The alternative recovery protocol, a 1- to 2-minute slow cool-down walk, minimizes the vagal response and may be better tolerated. In patients undergoing stress echocardiography, the immediate stop protocol facilitates the patient quickly coming to the echocardiography machine. The immediate-recovery ECG can be performed immediately on the patient laying down for post–stress echocardiography. For patients undergoing MPI, the slow walking cool-down protocol is appropriate because some radioisotope will still be circulating and can be taken up by the myocardium during a slow walking recovery.

Recovery should last 6 or more minutes and include monitoring of the ECG and symptoms throughout with a 12-lead ECG and blood pressure obtained every minute. If any of these parameters remains abnormal, monitoring should continue until a return to normal and until any symptoms resolve. A markedly abnormal exercise test merits a prompt call to the referring physician and consideration of further care in an acute care setting.

Indications for Termination of Exercise

The patient should be continuously observed during exercise testing. The patient should stop exercising when they can go no further or are close to exhaustion, or if their ECG, symptoms, heart rhythm, or hemodynamics warrant stopping because of demonstration of ischemia or concern for safety. As adapted from those outlined in Ellestad² and by Gibbons and colleagues in the 2002 ACC/AHA guidelines,⁶ exercise should be terminated when:

1. Anginal pain increases to moderate severity. This is a decision based on clinical judgment and consideration of the chronicity of symptoms and the clinical situation. In many patients, the development of mild angina can answer the question at hand and further exercise is not needed.

2. ST-segment depression has become moderate, generally ≥ 2 mm.

3. ST-segment depression is present at rest and there is a progressive increase in ST-segment depression with modest exercise or ST-segment depression increases by ≥ 2 mm.

4. ST-segment elevation is ≥ 1 mm in precordial (anterior) or inferior leads that do not have a resting Q wave.

5. Atrial tachycardia, atrial fibrillation, or atrial flutter supervenes.

6. There is onset of Mobitz type 1 or 2 second-degree or third-degree heart block.

7. Premature ventricular contractions occur in pairs with increasing frequency as exercise increases, or when at least three-beat ventricular tachycardia occurs. Termination secondary to increasing ventricular arrhythmia should be influenced “by the company it keeps.” If associated with ischemic electrocardiographic changes, termination would be prudent with increasing ventricular arrhythmia. If the test is being performed specifically to evaluate for induction of arrhythmia and is not associated with other markers of ischemia, more tolerance can be given.

8. Systolic blood pressure drops progressively in the face of continuing exercise, particularly if accompanied by another indication of ischemia.

9. The patient is unable to continue because of dyspnea, fatigue, or lightheadedness.

10. Musculoskeletal pain becomes moderate, such as that which might occur with arthritis or claudication.

11. The patient looks vasoconstricted—pale and clammy.

12. Systolic blood pressure becomes greater than 250 mm Hg or diastolic blood pressure becomes greater than 115 mm Hg.

13. The patient has reached or exceeded MPHR. As stated, if the patient is able and wishes to go further, he or she can do so in the absence of other indications for termination.

14. Equipment problems occur, such as loss of the ECG on the monitor.

Electrocardiographic Patterns and Their Significance

ST-Segment Depression

The most common manifestation of ischemia during exercise testing is ST-segment depression. By convention, the magnitude of ST segment is measured 0.08 seconds (80 milliseconds; two boxes on a standard ECG tracing) after the J point of the QRS, which represents the end of the QRS complex and the beginning of the ST segment. The deviation of the ST segment is measured against the location on the ECG in millivolts (mV) at the beginning of the QRS complex (the PQ junction in mV). It is measured here because as the heart rate increases, the segment between the T wave and the P wave often shortens to the point where it is no longer interpretable.⁴² One millivolt on the ECG corresponds to 1 mm and one box on the ECG. Two mm of ST-segment depression was the initial criterion for an abnormal exercise test and was soon changed to 1.5 mm as the criteria for being abnormal.² However, studies using outcome or coronary angiography as the gold standard demonstrated that 1.0 to 1.5 mm of ST-segment depression also predicted adverse outcomes or one or more stenoses of 50% or more on angiography.^43,44

Studies as long ago as 1941, as well as more recently, observed either adverse clinical outcomes or angiographic coronary disease in individuals exhibiting ST-segment depression of 0.5 to 1.0 mm.^45,46 Although lowering the cut point (threshold) to characterize an exercise study with 0.5 to 1.0 mm of ST-segment depression as abnormal increased sensitivity, specificity decreased to such a degree that a consensus developed to accept at least 1.0 mm of ST-segment depression as the standard cut point for ischemia. However, if patients exhibit other signs of ischemia, those with 0.5 to 1.0 mm of ST-segment depression are likely to have angiographically significant CAD.

The ST-segment depression discussed immediately above refers to horizontal or downsloping ST-segment depression. Figure 13–6 compares the normal ECG response to different categories of ST-segment depression: J point, also called rapid upsloping, slow upsloping, horizontal, and downsloping.

Downsloping ST-segment depression is most predictive of angiographic CAD whether it occurs during exercise or recovery, the latter termed “ischemic evolution” and highly specific for CAD.³ Horizontal ST-segment depression is also predictive of angiographic CAD, although less so than if downsloping. More controversial is ST-segment depression that is upsloping but remains 1 mm or more 0.08 seconds after the J point (slow upsloping ST-segment depression). ST-segment depression at the J point that slopes up rapidly such that the ST-segment depression is < 1 mm at 0.08 seconds after the J point of the QRS is commonly observed and is not indicative of ischemia.⁴⁷

Slow Upsloping ST Depression

In a 6-year outcome study by Stuart and Ellestad of 2700 patients undergoing exercise testing between 1964 and 1975, 438 had 2 mm or more of ST-segment depression when evaluated in immediate recovery.⁴⁸ (When exercise treadmill testing was initially performed, ECGs were performed during recovery rather than during exercise and recovery.) Using a composite end point of cardiac death, nonfatal MI, and the onset or progression of angina, patients whose ST-segment depression was downsloping had the worst prognosis, followed by those with ST-segment depression that was either horizontal or slow upsloping (Fig. 13–7). Those with no or < 2 mm of ST-segment depression experienced a favorable prognosis. In a separate group of patients in the same study who underwent coronary angiography and whose antecedent exercise test demonstrated at least 1 mm of ST-segment depression, those whose ST-segment depression was downsloping included numerically more subjects with two- or three-vessel CAD of at least 50% stenosis (62%) than those with horizontal (60%) or upsloping (57%) ST-segment depression. This cohort of patients from several generations ago had substantially more CAD than patients currently undergoing exercise testing in the United States today,⁴⁹ resulting in increased test sensitivity.

Contemporary with the study of Stuart and Ellestad,⁴⁸ Brody,⁵⁰ Goldschlager,³ and Kurita and colleagues⁵¹ found 1.5 mm of upsloping ST-segment depression to be highly predictive of angiographically significant CAD. Using 2 mm of upsloping ST-segment depression as their threshold for a positive test, Rijneke and associates found improved sensitivity in men and women when adding upsloping ST-segment depression to horizontal and downsloping ST-segment depression.³² Other investigators studying a more recent patient cohort found upsloping ST-segment depression to predict CAD, but with specificity progressively decreasing if ≥ 2, ≥ 1.5, or ≥ 1 mm of upsloping ST-segment depression was used as the threshold for positivity.⁵²

Recommended Criteria for an Abnormal Electrocardiographic Response to Exercise

The criteria for an abnormal ST-segment response of horizontal or downsloping ST-segment depression of 1 mm or more was recommended by the AHA in a 2001 statement adding that slow upsloping ST-segment depression of 2 mm or more was a borderline response.²³ In the ACC/AHA 2002 exercise testing guidelines, only at least 1 mm of horizontal or downsloping ST-segment depression is characterized as abnormal.⁶ Using only ST-segment depression that is horizontal or downsloping diminishes the sensitivity of exercise testing. In their 2013 AHA scientific statement on exercise standards for testing and training, Fletcher and coauthors characterize ≥ 1 mm of slow upsloping as equivocal. If exercise testing is used to “screen” a population of normal individuals, disease prevalence is low enough to exclude slow upsloping ST-segment depression as a criterion for positivity, as the positive predictive value would be too low. In patients referred for exercise testing, however, we recommend that slow upsloping ST-segment depression of ≥ 1.5 mm be included as a criterion for a positive exercise test in addition to the conventional criterion of ≥ 1.0 mm of horizontal or downsloping ST-segment depression.⁴⁰ The addition of ≥ 1.5 mm of slow upsloping ST-segment depression facilitates opportune test sensitivity in an era of diminished CAD severity and allows a better use of exercise testing in patients actually referred for exercise testing.

In order to be reliable, the ST-segment depression being measured should be present in multiple consecutive beats in at least one ECG lead. T-wave changes during exercise, such as the development of T-wave inversion with exercise or pseudonormalization, an inverted T wave at rest that becomes upright with exercise, are nonspecific findings.²

Onset of ST-segment Depression

The earlier during exercise that ST-segment depression occurs and the lower the rate pressure product of this depression, as well as the longer it lasts during recovery, the more severe the CAD as manifested by the incidence of multivessel and left main disease of coronary angiography.^3,12

Although ST-segment depression during exercise often persists into recovery, it may not manifest until exercise has been terminated. In a study of 214 subjects without known CAD who had asymptomatic ST-segment depression during exercise testing, Rywik and colleagues observed ST-segment depression only during recovery in 29% of patients. Over the ensuing 6 years, the development of angina, nonfatal MI or cardiac death was 2.5 times greater in those with ST-segment depression regardless of whether it occurred during exercise or recovery (Fig. 13–8).⁵³

In a study of 168 men with ST-segment depression during exercise testing who underwent subsequent coronary angiography, Lachterman and coworkers found that 15% demonstrated ST-segment depression only during recovery. Recovery-only ST-segment depression had the same predictive power of angiographic CAD as ST-segment depression occurring during exercise.⁵⁴ The work of Goldschlager and colleagues in a group of 330 patients undergoing exercise testing and coronary angiography emphasizes the importance of a careful review of the ECGs obtained during every minute of recovery. They found ST-segment depression to occur more frequently and be more abnormal during the first 3 minutes of recovery than during exercise itself. For example, in the 90 patients who developed ≥ 2 mm of slow upsloping ST-segment depression during exercise, 63% developed horizontal or downsloping ST-segment depression of ≥ 1 mm during recovery, indicative of ischemic evolution.

Location of ST-segment Depression

ST-segment depression during exercise testing in the precordial leads is more predictive of CAD than in the inferior leads, potentially because of atrial depolarization affecting the ST segment in the inferior leads. Location of ST-segment depression does not predict the artery/region of myocardium involved.⁴³

ST-Segment Elevation

ST-segment elevation occurring during exercise in contiguous leads without Q waves is a particularly ominous sign of ischemia and poor outcomes, often consistent with a high-grade proximal coronary lesion. It is also arrhythmogenic and is frequently accompanied by peaked T waves in the same leads, not unlike the ECG of a ST-segment elevation MI. Unlike ST-segment depression during exercise testing, ST-segment elevation during exercise in contiguous leads with an R wave localizes to the coronary artery involved.⁵⁵ Although ST-segment depression is consistent with subendocardial ischemia, ST-segment elevation is most consistent with transmural ischemia. Exercise is best terminated if ≥ 1 mm of ST-segment elevation occurs during exercise testing; continuing exercise in the face of ST-segment elevation in leads without Q waves creates substantial risk to the patient. If ST-segment elevation in leads without Q waves is associated with angina or other high-risk features during exercise, coronary angiography should be considered, often urgently.² ST-segment elevation present at rest in the setting of Q waves from a previous MI may result in an increase in ST-segment elevation with exercise that is generally not indicative of ischemia.

Lead aVR

Lead aVR is often unnecessarily neglected during exercise testing. On Einthoven’s classic triangle, aVR represents current going toward –150 degrees, which is roughly opposite to the left ventricular apex. Thus, the ST segment responds in a reciprocal fashion, becoming elevated in the setting of subendocardial apical ischemia. Thus, ST-segment depression in V2 to V6 or leads 2 and 3 is often reflected as ST-segment elevation in aVR. At times, diffuse ischemia from three vessel or left main disease results in cancelling out ST-segment deviation in other leads but ST-segment elevation in aVR often remains.^2,56 Monitoring and evaluation of aVR with exercise is now endorsed within the 2013 AHA scientific statement on exercise standards for testing.⁴²

Blood Pressure Response

Systolic blood pressure should rise with increasing treadmill workload, whereas diastolic blood pressure usually remains approximately the same or decreases (see Fig. 13–1). Although exertional hypotension has been defined in many ways, it has been shown to predict severe angiographic CAD and is associated with a poor prognosis.⁵⁷ A drop in systolic blood pressure below preexercise values is the most ominous sign. A failure of systolic blood pressure to adequately increase is particularly worrisome in patients who have a past history of MI, although in some patients this may be secondary to aggressive blood pressure control with medical therapy.

Exercise Duration

As emphasized by Bruce at the advent of treadmill testing, the longer the exercise duration, the better the prognosis. For a number of years, the chief parameter assessed during exercise testing was not a measure of ischemia per se, but exercise duration, measured in either METs or minutes of exercise on a treadmill protocol. A patient in whom exercise duration on a treadmill is prolonged, however, cannot always expect to have a good prognosis. In the Coronary Artery Surgery Study, Weiner and colleagues found prognosis to be good if the patient reached stage 4 on the standard Bruce protocol (10 METs). This held true as long as the patient had < 2 mm of ST-segment depression. If the patient had > 2 mm of ST-segment depression, prognosis was poor despite achieving stage 4 of the Bruce protocol (Fig. 13–9).⁵⁸

Influence of the Resting Electrocardiogram

The closer the ECG is to normal at rest, the more accurate the evaluation of the ECG during the heart rate increase of exercise. Among patients with ≥ 1 mm of ST-segment depression at rest, an increase of > 2 mm in horizontal ST-segment depression or > 1 mm of downsloping ST-segment depression is considered an abnormal response. Left bundle branch block, ventricular preexcitation (ie, Wolff-Parkinson-White pattern), changes associated with digoxin use in the preceding 2 weeks, and ventricular paced rhythms preclude the ability to interpret ischemic changes electrocardiographically.⁶ Left ventricular hypertrophy with repolarization abnormalities diminishes sensitivity, but specificity is maintained.⁶

Chronotropic Response

Ellestad and colleagues found that chronotropic response inversely predicted the development of the composite end point of cardiac death, nonfatal MI, and the onset or progression of angina.⁴³ From the same group, Chin and colleagues demonstrated that among patients not taking beta-blockers, a peak heart rate achieved on exercise testing that was two standard deviations less than MPHR predicted CAD.^2,59 Lauer and associates found that patients who were not taking beta-blockers and who did not reach 85% of MPHR (measured as 220 – age) experienced substantially increased all-cause mortality.⁶⁰

Heart Rate Response in Recovery

In their 1929 abstract first describing the standardized step test, Master and Oppenheimer described a rapid return of heart rate toward normal as indicative of “circulatory efficiency” or fitness.¹ The decrease in heart rate postexercise is considered to be a function of the reactivation of vagal tone.⁶¹ Lauer further characterized this heart rate recovery response with quantitation and outcome evaluation. He found that that those who fail to decrease their heart rate by 12 beats/min 1 minute postexercise during a recovery walking 1.5 miles/hour at a grade of 2.5% experienced increased all-cause mortality.^62,63,64

The Effect of Medications

Antianginal medications diminish ischemia with exercise stress and thus the opportunity to observe ischemic ECG changes. The effect of beta-blockers is particularly important, given their impact on the maximum heart rate achieved. Digoxin renders ST-segment evaluation difficult for up to 2 weeks following its last dose.

Treadmill Scores

The most widely used score to integrate parameters of exercise testing is the Duke Treadmill Score. This was developed in 1987 among a group of 2842 consecutive inpatients with angina undergoing treadmill testing within 6 weeks of coronary angiography performed between 1969 and 1980.⁶⁵ The derived score utilized the components of the exercise test that best predicted cardiovascular death or nonfatal MI over a mean 5 years of follow-up.

In the exercise angina index,

• 0 = no angina on the treadmill

• 1 = angina occurred

• 2 = angina caused termination of exercise

Note that parameters that predict improved outcome are positive numbers (eg, duration of exercise), whereas negative predictors, such as ST-segment deviation, are negative numbers. A score with an excellent prognosis would be +25. A poor prognostic score would be –25. The researchers arbitrarily grouped patients in the study into those at lowest risk (a score of +5 or greater), moderate risk (+4 to –10) and highest risk (–11 or lower). During follow-up, 24% of patients underwent coronary artery bypass grafting (CABG). They were censured at the time of surgery from further follow-up. To obtain the event-free survival outcomes shown in Fig. 13–10, 24% of patients crossed over from medical to surgical therapy.

Mark and colleagues further validated the score in a study of 613 consecutive outpatients undergoing Bruce protocol exercise testing and followed for a mean of 4 years using cardiovascular death as the end point.⁶⁶ As seen in Table 13–5, the score was a good predictor of outcome in both inpatients⁶⁵ and outpatients.⁶⁶ The crossover rate to CABG or percutaneous coronary angiography in the outpatient population was 8%. Because patients currently undergoing exercise testing have a lower burden of CAD compared to the study time period,^49,67 scores will be skewed to the lower risk category, resulting in less robust risk stratification. However, its use as a continuous variable and rather than a simple categorization into low, moderate, and high risk adds valuable predictive information to an exercise test report.

Exercise Testing with a Negative Exercise Myocardial Perfusion Imaging Study

Because diagnostic test accuracy is greater for MPI than stand-alone exercise testing, there is a tendency to ignore the results of a positive exercise test if perfusion imaging is normal. In other words, to accept the premise that imaging trumps the exercise ECG. However, applying Bayes theorem in this situation demonstrates that although a normal MPI study decreases the probability that the exercise ECG is falsely positive, it does not eliminate it. In 208 consecutive patients undergoing exercise thallium-201 MPI, Raiker and coworkers tracked cardiac death, nonfatal MI, and the development of unstable angina over 13 months. Of those with a normal or nondiagnostic exercise test and normal perfusion, the event rate was 2/175 (1%). A positive exercise test was defined as ≥ 2 mm of horizontal or downsloping ST-segment depression. Among those with a positive ECG exercise test and normal perfusion, the event rate was 3 of 33 (9%). Thus, both perfusion and exercise response add information, and clinical judgment should be applied when they are discordant.

Exercise Testing with a Negative Exercise Stress Echocardiogram

McCully and colleagues followed 1325 patients for approximately 3 years following an exercise stress echocardiogram in which wall motion of all 17 segments that were normal at rest were either normal or hyperdynamic poststress—defined as a normal stress echocardiogram.⁶⁸ The cardiac event end point of cardiac death, nonfatal MI and coronary revascularization averaged 0.9% per year over the time period. However, if the patient developed angina during exercise testing, despite the normal poststress wall motion, the relative risk of a subsequent cardiac event was 4.1 times greater than if they did not.

Rather than routinely accepting that imaging trumps the ECG and symptoms in patients undergoing exercise testing with adjunctive imaging, clinical judgment should be used in the evaluation entertaining the potential that imaging may be falsely negative. Depending on the pretest likelihood of CAD and degree of ECG changes or exercise-induced symptoms, a calcium score with noncontrast cardiac CT or CT coronary angiography can prove helpful in these situations.

Like “plain old balloon angioplasty,” stand-alone exercise testing may be less appealing to the clinician than exercise testing with adjunctive imaging. Exercise testing is, however, remarkably effective as an initial and often final test in patients able to exercise and with ECGs that do not preclude the interpretation of ischemia with exercise. The authors of the 2013 Multimodality Appropriate Use Criteria for the Detection and Risk Assessment of Stable Ischemic Heart Disease guidelines, written by representatives from leading US cardiology societies, recommend the stand-alone exercise stress test in appropriately selected patients.²⁷ The primary recommended use of adjunctive imaging is among patients who are either unable to achieve adequate exercise stress, have abnormal resting ECGs, or who require a test with greater sensitivity.²⁷

The ACC/AHA 2002 guidelines for exercise testing include comprehensive indications for exercise testing.²⁶ One of the key class I recommendations is that exercise testing serve as the initial test for adult patients in the diagnosis of obstructive CAD with an intermediate pretest probability of CAD on the basis of age, sex, and symptoms.⁶ Class III exclusions to this indication are patients in whom the ECG is uninterruptible, including those with ventricular preexcitation, paced ventricular rhythm, > 1 mm of resting ST-segment depression, or left bundle branch block. Exercise testing is also a class I recommendation in the assessment of risk and prognosis in patients with symptoms or a prior history of CAD. This includes (1) patients with suspected or previously evaluated known CAD presenting with a change in clinical status, (2) low-risk unstable angina patients 8 to 12 hours after presentation who have been free of active ischemic or heart failure symptoms, and (3) intermediate-risk unstable angina patients 2 to 3 days after presentation who have been free of active ischemic or heart failure symptoms.

Submaximal exercise testing in which exercise is stopped at a predetermined end point, such as a peak heart rate of 120 beats/min, 70% of MPHR, or a peak MET level of 5, can be used as a class I indication 4 to 7 plus days post-MI for evaluation of medical therapy, prognostic assessment, and/or development of an activity prescription.

Exercise testing also has class I indications to demonstrate ischemia prior to revascularization and in the evaluation of patients with recurrent symptoms that suggest ischemia after revascularization.

Exercise Testing in Women

Traditionally exercise testing has thought to be less accurate in women than men.²⁶ Levisman and colleagues point out, however, that many of the studies on which this premise is based excluded women older than 65 years, an age when CAD prevalence and severity are greater.⁶⁹ Levisman and colleagues thus age-stratified 111 women undergoing both exercise testing and coronary angiography. The positive predictive value of exercise testing of the overall group was 51%. For women 35 to 50 years, the positive predictive value was 36%, compared to 68% for those older than 65 years. Among those with true positive tests, three-vessel disease was present in about 55% of those older than 65 years and about 25% of those 35 to 50 years. Thus, the expected lower sensitivity and specificity traditionally observed in women may be explained by differing CAD prevalence and severity. For testing in a population of women with CAD severity as great as in men, sensitivity is likely similar between the two sexes. Another challenge in evaluating test sensitivity in women is the use of obstructive CAD on coronary angiography as the gold standard. More than men, women experience ischemia and events, including MI, with coronary stenoses of < 50%.⁶⁸

In the What is the Optimal Method for Ischemia Evaluation in Women (WOMEN) Trial, Shaw and coworkers randomized 824 symptomatic women with interpretable ECGs who were expected to achieve five METs to stand-alone exercise testing or to exercise MPI.⁷⁰ At the end of 2 years, there was no difference in the primary end point, survival free from a major cardiac adverse event (98.0% for exercise testing and 97.7% for MPI).

Kohli,⁷¹ Borque and colleagues,⁷² and Mieres and the coauthors of the AHA consensus statement on the role of noninvasive testing in women⁷³ emphasize the use of parameters other than only the use of ST-segment deviation in women. They comment on the powerful additional predictive information demonstrated in women as well as men of exercise duration, the Duke Treadmill Score, and heart rate recovery.

Consistent with the ACC/AHA 2002 guidelines for both sexes,⁶ the 2014 AHA consensus statement on the role of noninvasive testing in women recommends exercise testing as the initial test for symptomatic women with interpretable ECGs who are able to exercise.⁷³

Novel Electrocardiogram Parameters

Building on previous work,⁷⁴ Sharir and colleagues evaluated the use of an advanced signal processing ECG lead system designed to measure low-amplitude, high-frequency electrical signals during the QRS complex.⁷⁵ These signals are termed high-frequency mid-QRS (HFQRS) waveforms. The amplitude of the standard waveforms measured on the standard ECG at rest and during exercise are measured in millivolts and at frequencies generally less than or equal to 100 Hz. HFQRS waveforms are about 100 times smaller in amplitude and are thus measured in microvolts rather than millivolts and are high frequency, 150 to 250 Hz. Abboud and colleagues previously demonstrated an approximately 33% decrease in HFQRS during ischemia induced by balloon inflation in a coronary artery.⁷⁴ As reviewed by Amit and colleagues, the normal extensive branching His-Purkinje system results in a vast array of myocardial activation sites and thus a depolarization wave front with numerous HFQRS signals.⁷⁶ Ischemia, however, slows myocyte-to-myocyte conduction, resulting in a decrease in the amplitude of high-frequency waveforms arising from this area of ischemia. This is useful for exercise testing, because the development of ischemia during exercise results in decreasing amplitude of HFQRS waveforms. In the 996-patient exercise testing study of Sharir, she found that when advanced signal processing to measure HFQRS was added to standard ST-segment analysis, this resulted in a significant increase in sensitivity and specificity using ischemia by MPI as the gold standard.⁷⁵

Rosenbaum and colleagues evaluated the addition of HFQRS to ST-segment evaluation during exercise testing in 113 women. Using coronary angiography as the gold standard, the addition of HFQRS improved specificity. Also, the number of leads with decreased HFQRS amplitude correlated with the number of stenotic coronary arteries.⁷⁷ A large multicenter prospective trial evaluating the value and limitations of HFQRS analysis is strongly recommended to evaluate this promising technology.

Combining Calcium Scoring with Exercise Testing

In a retrospective study of 153 symptomatic patients who underwent exercise testing, calcium scoring, and coronary angiography, Lamont and colleagues described how the combination of calcium scoring could theoretically enhance the accuracy of exercise testing.⁷⁸ Coronary calcification was present in 81% of the cohort and nearly 99% of those with angiographic CAD. Of the 27% of patients without CAD, 66% had no coronary calcium.

In 2013 Rozanski and colleagues coined the term the calcium treadmill test.⁷⁹ They suggested that this combination of exercise testing and a coronary artery calcium (CAC) score could evaluate not only ischemia and functional capacity with exercise testing but also calcific coronary atherosclerosis with CAC. They posit the use of this combination test may be initially best used in stable, mildly symptomatic patients able to exercise.

Chang and coworkers evaluated the results of about 1000 patients who underwent exercise testing and CAC scoring as part of previously published prospective trial.⁸⁰ Increasing CAC added substantial prognostic information over the 7-year follow-up period to the exercise test parameters of ST-segment depression, Duke Treadmill Score, and exercise capacity in the prediction of cardiac death, nonfatal MI, and revascularization.

Thompson and coworkers previously evaluated how unprompted clinicians responded to a CAC score of more than 100 Agatston units performed in relation to an exercise or pharmacologic MPI that was normal. At a mean follow-up of 9 months, a CAC score of more than 100 resulted in a significant increase in aspirin use and a near doubling in statin use.⁸¹

CAC results show great promise in appropriately influencing post-test management in patients with normal or abnormal exercise tests.⁸² Prospective trials of the potential role of the calcium treadmill test are warranted.

Exercise testing is a powerful predictor of CAD diagnosis and prognosis. Its careful and thoughtful performance creates the opportunity for this inexpensive test to continue to serve as the backbone of the noninvasive evaluation of CAD.

We graciously thank Drs. Michael J. Lipinski and Victor F. Froelicher, who authored the previous version of this chapter in the 13th edition.