Reviews |

Stress Testing in Cardiac Evaluation*: Current Concepts With Emphasis on the ECG FREE TO VIEW

Morton E. Tavel, MD, FCCP
Author and Funding Information

* From the Indiana Heart Institute, Care Group, Inc, and the Department of Medicine, Indiana University School of Medicine, Indianapolis, IN.

Correspondence to: Morton E. Tavel, MD, FCCP, 8333 Naab Rd, Suite 200, Indianapolis, IN 46260; e-mail: mtavel6986@aol.com

Chest. 2001;119(3):907-925. doi:10.1378/chest.119.3.907
Text Size: A A A
Published online

The stress test combined with ECG was originally used primarily for the detection of ST changes secondary to myocardial ischemia. Modern exercise testing, however, is not limited to observation of these changes. Important information is derived from exercise capacity, BP response, development of arrhythmias, and whether or not symptoms such as chest pain develop during exercise. This allows for assessment of presence and severity of ischemia, prognosis, overall functional capacity, and efficacy of therapeutic interventions. Although stress testing is often combined with radionuclide or echocardiographic imaging, I shall focus primarily on the ECG component. The proper selection of stress tests in clinical evaluation will also be discussed.

The major indications for performing a stress test are summarized in Table 1 . Table 2 lists the generally accepted contraindications for such testing, as adapted from the guidelines provided by the American College of Cardiology/American Heart Association Task Force in 1997.1

Stress testing is a relatively safe procedure. Before 1980, an overall mortality rate of 1 in 20,000 tests was observed. In the contemporary era, however, this frequency has been found to be even lower, generally < 1 in 50,000.2 Nonfatal complications, such as myocardial infarction, occur at the rate of < 4 per 10,000 tests.2In subjects with histories of ventricular tachycardia or fibrillation, serious but nonfatal arrhythmias occurred during 2.3% of the tests.3 In the absence of such a history, the incidence of such complications is approximately 0.05%.

The most commonly performed stress test is the graded exercise test, using either the treadmill or cycle ergometer. The patient is generally subjected to increasing workloads at 2- or 3-min intervals. The test is stopped for any of the reasons listed in Table 3 . The ECG is monitored not only during exercise but also afterward, for 5 to 11% of patients with abnormal responses may not display such findings until reaching the recovery period45 (see below).

The protocol used for treadmill testing varies among different institutions. One of the most widely used is that of Bruce and Hornsten,6but the procedure may be customized to allow for 6 to 12 min of exercise.7A modified version of this protocol is detailed in Table 4 and is especially useful because it facilitates extrapolation from maximum treadmill performance to levels of work and recreational activity. It also allows for estimation of severity of cardiac decompensation (New York Heart Association classes). The estimated workload is reported in metabolic equivalents (METs), a unit that facilitates comparison of different exercise protocols as well as allowing for comparison with work or recreational effort requirements. This term actually represents the energy cost of activity in multiples of resting oxygen consumption (1 MET = 3.5 mL/kg/min). Inasmuch as oxygen consumption is determined primarily by cardiac output in the absence of pulmonary or skeletal limitations, this information allows for rough estimates of cardiac function. Although oxygen uptake is not actually measured in most clinical laboratories, one can estimate these approximate values simply by consulting published information derived from the various treadmill workloads. Increasing age and deconditioning reduce normal maximum values.8 In some clinics, the test is stopped at an arbitrary target point of 85% of the predicted maximal heart rate for the subject’s age. This maximal rate is estimated by subtracting the subject’s age from 220. This practice is no longer recommended in favor of stressing individuals to the point of exhaustion or the development of warning signs or symptoms (Table 3).,1 If the subject can continue, I usually terminate the test on reaching 100% of the expected maximum heart rate for age.

The ECG and BP are monitored throughout exercise and for several minutes thereafter. In most instances, the test may be conducted by a properly trained nonphysician,2 especially with subjects at low risk for cardiac events. In subjects at high risk, ie, with chest pain suggestive of angina pectoris or with known heart disease, a physician should be in attendance or in close proximity during the test. In all instances, a physician should be near enough to be readily available should there be an emergent need.

In response to the increased stroke volume and systolic contractile force, the systolic BP normally rises progressively with increasing workloads, reaching approximately 160 to 220 mm Hg with maximum effort. Because exercise lowers overall peripheral vascular resistance, the diastolic pressure exhibits little or no change (< 10 mm Hg). If the systolic pressure fails to rise to ≥ 130 mm Hg or falls by ≥ 10 mm Hg in response to exercise, this frequently indicates left ventricular dysfunction and often signals the presence of severe coronary artery disease (CAD) associated with extensive myocardial ischemia.910 However, possibly in a substantial number of patients, myocardial ischemia may produce an abnormal fall in BP, presumably resulting from an excessive vasodilator reflex in nonexercising vascular beds.11 In this instance, cardiac output actually may be increased during exercise.

An abnormal rise in exercise systolic pressure to a level ≥ 214 mm Hg in a subject with a normal resting pressure predicts increased risk for future sustained hypertension, estimated at approximately 10 to 26% for the next 5 to 10 years.12This is associated with a relatively high prevalence of left ventricular hypertrophy.13 Some investigators12 have found a slightly greater 5- to 10-year rate of subsequent cardiovascular events within this group, but others14 have not observed this outcome.

Normally, the systolic pressure falls rapidly after cessation of exercise, dropping by an average of ≥ 15% at 3 min after stopping. Myocardial ischemia may reduce the rate at which this level falls: a 3-min postexercise level of ≥ 90% in comparison with the peak systolic level during exercise suggests the presence of ischemia.1517 To minimize the inaccuracies of BP measurement at peak exercise, McHam et al17 suggested comparing pressures at 1 and 3 min after exercise. Abnormality existed if the value at 3 min equaled or exceeded that at 1 min. Such a retarded pressure drop is an insensitive indicator of ischemia, but it is fairly specific for this disorder, reportedly exceeding 80%.17 Although the mechanism for this abnormal pressure response is uncertain, it may result from ischemic suppression of left ventricular function during exercise combined with the subsequent recovery of contractility during recovery. This response has been found usually to signal profound and extensive ischemia, with the systolic pressure ratio increasing proportionally with the number of diseased coronary arteries.16

In general, the heart rate rises proportionately with the intensity of the workload. Excessively rapid rises in rate result primarily from reduced left ventricular stroke volume, which, in turn, may be caused by physical deconditioning or cardiac disease. Under such circumstances, the heart rate reaches its peak level relatively early, and this limits maximum exercise capacity. However, when the heart rate response to exercise is excessively attenuated (in the absence of rate-limiting drugs), this condition is termed chronotropic incompetence. Patients showing this response often have significant organic heart disease, and among patients with known or suspected coronary disease, it is independently predictive of higher all-cause mortality.18Unfortunately, however, definitions of chronotropic incompetence vary: they range from the inability to achieve 85% of expected maximum heart rate, to failure to achieve 100 beats/min at maximal exertion,19 to heart rate responses in terms of percentage or standard deviations around a mean heart rate for each stage on the treadmill.18,20Because of such limitations, analysis of cardiac chronotropic responses currently has limited practical applicability.21

The rate with which the heart rate slows in the early recovery period can also provide information about ventricular function and prognosis. A recent study by Cole et al22 demonstrated that a drop in rate by ≤ 12 beats/min at 1 min after peak exercise during the cool-down phase in early recovery (while walking 1.5 mph at 2.5% grade) signaled a poor prognosis. These subjects were found to have a subsequent 6-year mortality rate four times greater than those have with a more rapid fall in heart rate. The retarded heart rate drop during recovery probably signifies reduced vagal tone, which is often associated with decreased myocardial function and exercise capacity. An abnormal rate drop may add independent prognostic information that extends beyond such factors as effort tolerance and rate response during the exercise period.

Cardiac auscultation should be performed before and immediately after exercise. Resting abnormalities, such as murmurs and third and fourth heart sounds, should be noted and compared with postexercise findings. This facilitates recognition of important valvular abnormalities, such as aortic stenosis, that might preclude or modify testing. I have also encountered individuals with hypertrophic subaortic stenosis in whom intense systolic murmurs became manifest only after exercise. The appearance of a third heart sound or mid-diastolic gallop sound after exercise is usually indicative of reduced resting systolic function of the left ventricle, which may be associated with cardiomyopathy or prior infarction, often associated with diabetes mellitus and left bundle branch-block, although extensive stress-induced ischemia is occasionally found.23

The conventional 12-lead ECG is the most widely chosen one for the graded exercise test. Almost all the significant ST-segment changes, however, can be demonstrated in leads I and V3 through V6,24and approximately 90% of that information is contained in lead V5. Some investigators have suggested that the addition of right precordial leads, ie, leads V1R through V4R, may improve the diagnostic accuracy in detection of ischemia in general, especially in the distribution of the right and circumflex left coronary arteries.26

Equipment for computer averaging of the ECG signals is commonly available, and this provides assistance in the analysis of various changes—especially ST depression. I agree with others27 who have noted that such records often produce spuriously abnormal ST changes. Thus, the clinician should always evaluate the actual ECG recording strips.

Horizontal or Downsloping ST-Segment Depression of ≥1 mm

The PR segment is the reference point with which the ST segment is compared (Fig 1 ). The criterion of ≥ 1 mm of horizontal or downsloping depression (Fig 1, D and E) is the most generally accepted. Junctional (J point) depression with slowly upsloping ST segments (Fig 1C) also is generally considered to be an abnormal response,2830 although the definition of slowly upsloping varies. Various investigators31have reported that including upsloping ST responses that remain 1 to 2 mm below the baseline at 60 to 80 ms after the J point significantly increases test sensitivity without degrading specificity. However, Sansoy et al32noted that specificity was reduced, especially if 1-mm depression was selected at 80 ms. For this reason and until further data are available, in the case of upsloping ST segments, I agree with others33 who suggest using 1.5-mm depression at 80 ms after the J point as a reasonable criterion for a positive response.

The depth of ST depression caused by ischemia appears to be influenced by the overall amplitude of the ECG signal as displayed by the R-wave height in lead V5. Hollenberg et al34reported that increased R-wave amplitude might give rise to exaggerated ST depression during exercise. Accuracy was improved by normalizing the ST depression for the R-wave amplitude. Ellestad et al35noted that the correction of ST depression for R-wave amplitude is especially useful in subjects with a low R wave in V5. This group36 found that if the R wave decreases by ≥ 1 mm in V5 at the end of exercise, ST depression of ≥ 0.5 mm constituted a positive ischemic response and would thereby increase the sensitivity of the test to detect disease. Although of interest, these observations require further confirmation, for Froelicher et al37 were unable to find improvement in test accuracy by adjusting for R-wave amplitude.

ST-Segment Elevation

Although horizontal or downsloping ST-segment depression is the typical ischemic response in stress testing (in all leads except aVR), some patients exhibit ST-segment elevation of ≥ 1 mm (Fig 1F). Generally, in the absence of prior infarction, this finding is uncommon but implies severe transmural ischemia exceeding that associated with isolated ST depression.38 The reported incidence among patients with chest pain ranges from 0.2 to 1.7%.3842 High-grade proximal coronary stenosis is usually found, and this combination is associated with an ominous prognosis.38 The correlation between the site of the ST-segment elevation and the artery involved is generally quite good.3843 Probably representing a variant of ST elevation, isolated transient increase in height of T waves in the anterior leads (V1 through V3) strongly suggests severe narrowing of the left anterior descending coronary artery.44

Exercise-induced ST-segment elevation is seen most commonly in patients who have had previous myocardial infarctions.4041 Most studies demonstrated an incidence of 14 to 27%.4042,4550 Patients with anterior myocardial infarction are more likely to have exercise-induced ST-segment elevation than are those with inferior myocardial infarction.49,51 The ST-segment elevation almost always occurs in the leads with abnormal Q waves.50 It also is associated with a left ventricular wall motion abnormality, either dyskinetic or akinetic, in the corresponding site in > 90% of cases.41,5052 Overall poor left ventricular systolic function is usually found.45,52Although some studies suggest that such ST changes denote residual myocardial ischemia and contractile reserve within this infarct area,5354 passive segmental left ventricular wall motion abnormality (with or without aneurysm formation) is probably the underlying mechanism for the ST-segment elevation in most cases.54Findings in one study,55however, suggested that residual viability could be found only in those instances in which ST elevation was associated with reciprocal ST depression in those leads taken from the opposite side of the heart. In this context, therefore, ST depression may simply be a secondary response to remote ischemia rather than a primary marker of ischemia in itself—as is the usual interpretation. In the case of prior inferior myocardial infarction, the ST elevation found accompanying the Q waves in the inferior leads (II, III, and aVF) often gives rise to reciprocal ST depression in the high lateral leads (I and aVL).56

Ten to 30% of patients with variant angina also may have ST-segment elevation with exercise.51 The leads that show ST-segment elevation are usually the same leads that record elevation during angina at rest. All studies51,5758 regularly report the occurrence of spasm of a major coronary artery supplying the area of the myocardium corresponding to the site of the ST-segment elevation. Most patients with variant angina and exercise-induced ST-segment elevation, however, also have significant fixed coronary lesions.59

Transient, exercise-induced ST elevation has been reported to occur in conjunction with acute pericarditis60and may be mistaken for ischemic pain in the acute care setting. As noted previously, patients with known pericarditis are usually not subjected to stress testing. When the diagnosis is uncertain, however, persistence of ST elevation in response to exercise stress testing may aid in the distinction between pericarditis and early repolarization (a normal variant) inasmuch as in the latter condition, ST elevation returns to the isoelectric line.61Exercise-induced resolution of ST elevation, however, although uncommon, may occur in pericarditis as well.62

The degree of ST-segment displacement in relation to the increase in heart rate (ΔST/HR index) with exercise has been suggested by some investigators to be a more accurate indicator of the presence and severity of CAD.6367 This method requires meticulous or computerized measurement of ST displacement. Other studies6871 have not demonstrated the superiority of theΔ ST/HR index in the prediction of CAD. The usefulness of changes of the heart rate-adjusted ST-segment depression in the detection of CAD, therefore, remains controversial72and, at best, may add only limited incremental diagnostic value.73 Thus, this method has not gained wide acceptance.

Transient inversion of U waves—even in the absence of an abnormal ST-segment response—has been suggested as a marker of extensive ischemia in the anterior myocardium74or elsewhere.75Analogous to U-wave inversion, others7677 have found that exercise-induced increased magnitude of the U wave in the precordial leads (≥ 0.05 MV) was strongly suggestive of ischemia in the distribution of the left circumflex or right coronary arteries, presumably representing reciprocal changes from posterior U-wave inversion when myocardial ischemia occurred in this latter location. Detection of all U-wave changes is difficult in the presence of tachycardia, a factor that limits the usefulness of these observations.

Bonoris et al78initially reported an increase in the R-wave amplitude immediately after exercise in patients with severe multivessel coronary artery narrowing and ventricular dysfunction, presumably caused by transient ventricular dilatation. The results of later studies varied, some supporting and others not supporting the original observation.7984 Nevertheless, although the sensitivity of this finding is low, when the R-wave amplitude increases by > 2 mm at peak exercise, this has been said to strongly suggest ischemia.85 In general, however, analysis of R-wave amplitude is not used in clinical practice.

Normally, the Q-wave depth in lead V5 increases in response to exercise,8687 presumably because of septal thickening in response to inotropic stimulation. A decrease or no change in this wave has been found with stenosis of the left anterior descending coronary artery, usually in association with multivessel disease.86

The P wave shortens normally by approximately 0.02 s in response to exercise, whereas in the presence of ischemia, it may lengthen slightly or remain unchanged.88 Transient elevation of atrial and left ventricular filling pressure induced by ischemia is the assumed mechanism for such P-wave prolongation. This interesting observation, if confirmed, may provide a useful secondary means to confirm the significance of other changes, such as ST depression.

The mean frontal plane QRS axis normally shifts toward the right in response to exercise. Exercise-induced leftward axis shift8990 or absence of rightward shift89 is reported to be highly specific for narrowing of the left anterior descending coronary artery, presumably caused by ischemia of the left anterior fascicle. However, transient rightward axis shift to > 90° is a rare occurrence, but said to be highly specific for CAD,91 presumably as a consequence of septal ischemia.

Exercise induction of complete right or left bundle branch block is generally nonspecific. When found together with other evidence of ischemia, however, such as angina pectoris, the conduction abnormality is said to be strongly suggestive of myocardial ischemia, especially in the distribution of the proximal left anterior descending coronary artery.90

The QRS duration normally remains unchanged or shortens slightly in response to exercise. In the presence of ischemia, the QRS may lengthen slightly (> 3 to 5 ms) and this may allow detection of ischemia with greater sensitivity and specificity than ST segment changes alone,9295 even in patients with recent myocardial infarction.96 Michaelides et al95 found that the QRS prolongation correlated with severity of ischemia, prolonging progressively with one (9.7 ms), two (13.6 ms), and three (16.3 ms) ischemic areas as demonstrated by nuclear scintigraphy. When prolongation of S waves (10 to 12 ms) occurs in subjects with resting right bundle branch block or left anterior hemiblock, this is believed to be highly suggestive of left anterior descending coronary artery stenosis.97 Because of the small increments in duration of any of the above intervals, higher recording paper speeds or computer-aided measuring techniques such as signal averaging93,95 would be generally required to achieve sufficient accuracy.

Normally, the QT interval (corrected for heart rate) shortens with exercise. Some investigators have found that this interval fails to shorten or lengthens when ischemia is present.9899 Others100101 have suggested that abnormal exercise-induced QT dispersion, ie, the difference between shortest and longest QT intervals when multiple leads are compared, is greater in patients with ischemia. Measurement of this interval, however, is difficult, especially when tachycardia is present, and this limits its potential clinical value.

Despite its limitations, the coronary arteriogram remains the most common standard against which the diagnostic value (especially sensitivity) of stress testing for CAD is measured. As noted, the use of arteriography as a standard usually suffers from the disadvantages of referral bias, ie, a positive ECG stress test is often involved in the decision to select those referred for subsequent coronary arteriography. The effect of such selective referral is to raise falsely sensitivity estimates, and, conversely, to reduce falsely specificity values.

Noninvasive techniques (eg, stress thallium perfusion imaging or stress echocardiography) have also been used to confirm the presence of coronary disease. When such studies are used to evaluate the performance of stress ECG, referral bias can be minimized, and this usually results in lower sensitivity but relatively high specificity values.

Stress-induced ST-segment depression is probably caused largely by reduction of perfusion to the subendocardium, the zone most vulnerable to ischemia. The sensitivity of such changes in the detection of coronary disease varies widely in published reports,102104 undoubtedly reflecting various confounding factors such as referral bias, severity of disease in the selected population, and so forth. Overall, test sensitivity is most often found to lie in the 60 to 70% range. In studies in which referral bias was minimized, however, lower sensitivity values are reported, falling in the range of 45 to 60%.37,105In general, the greater the degree and extent of ischemia, the more likely will ST depression occur, and, therefore, the higher the sensitivity of the stress ECG. Most investigators have reported a sensitivity in the range of 40%, 66%, and 76% for one-, two-, and three-vessel coronary disease, respectively.106107 We have also found that with larger nuclear perfusion defects in response to stress imaging, ST depression was more often encountered.105 Notwithstanding the results of smaller studies,108 we have also found that larger areas of hypoperfusion also increased the average depth of ST depression.105

In general, the distribution of leads manifesting ST-segment depression does not appear to be helpful in localizing the obstructive coronary lesions,105106,109112 for as noted, the lead V5 is most apt to reflect this change whenever ischemia is present. However, some studies,43,113 have noted that ST depression of lead V1—either isolated or associated with other lead changes—suggests the presence of ischemia in posterior myocardial regions, ie, in those areas supplied by the left circumflex or right coronary arteries. Such depression in these leads presumably arises as a reciprocal response to ST elevation posteriorly, a location usually inaccessible to conventional lead systems. Extending these observations further, one group,26 has suggested that ST deviation in the right precordial leads (V3R and V4R) may not only enhance the recognition of posterior ischemia, but when these changes are combined with abnormalities in the conventional lead systems, the sensitivity of ECG stress testing in general could be greatly enhanced for detection of coronary disease. They claimed a sensitivity ranging from 89% in single-vessel involvement to as high as 95% in triple-vessel coronary disease. To enter the clinical mainstream, however, these latter observations require confirmation.

In contrast to the almost universal changes produced in the left precordial leads (V4 through V6), myocardial ischemia seldom produces ST depression confined to the inferior leads (II, III, and aVF), and when one encounters such a limited distribution, this usually denotes a false-positive test response,105,114 possibly often attributable to P-wave repolarization (see below).

In general, the myocardial location of ischemia also appears to play no role in the likelihood of ST depression,105,110 for we have demonstrated that the likelihood of its appearance depends primarily on the extent of ischemia rather than its location.105

The configuration, time of onset, and duration of depressed ST segment during and after treadmill exercise have important diagnostic significance.115Multivessel or left main coronary disease is present in approximately 90% of patients who have changes appearing at low workloads (Bruce stage I or II) or persisting for > 8 min after exercise.116117 Although such early and prolonged ST responses are highly specific for ischemia, some studies suggest that they correlate less well with subsequent prognosis,118and even arteriographic or scintigraphic severity may be variable.119Although T-wave changes alone are not helpful in diagnosis, the presence of deep T-wave inversion (≥ 5 mm) when combined with ST depression has been found to be highly specific for multivessel CAD with multiple severe narrowings.120

ST-segment depression occasionally begins only after cessation of exercise. The diagnostic and prognostic significance of such a delayed response is generally similar to those occurring during exercise121123; however, when the onset of such a change is delayed by > 2 to 3 min into recovery, this suggests a false-positive response.124When ST changes during exercise are equivocal, the finding of progressively greater downsloping ST depression during recovery is a fairly specific sign of severe ischemia and also signals a greater incidence of cardiac events in follow-up.125

The pattern of regression of ST changes in the recovery period may be useful in distinguishing ischemic responses from those encountered occasionally in normal subjects who are falsely positive.124,126 In true ischemia, the major ST depression tends to coincide with the termination of exercise and continues—often intensifying—for ≥ 2 to 3 min after cessation.124 The persistence of this depression in recovery usually parallels its onset, ie, when it begins early at low workloads, it is more persistent during recovery. When it reaches a maximum later in the exercise phase, it usually regresses relatively early in recovery, but it usually continues for at least 3 min after stopping. In contrast, false-positive ST depression tends to reach its maximum immediately before and at peak exertion but regresses quickly on cessation, frequently returning to normal within 1 to 3 min of recovery. In this latter instance, as the heart rate slows in recovery, the depth of ST depression is less when compared with corresponding cycle lengths during the exercise phase, whereas those subjects with true ischemia usually show equal or greater ST displacement at comparable cycle lengths.,126

The development of typical anginal chest pain during the test generally signifies fairly extensive ischemia and thus increases the likelihood of ST changes, adding significantly to test sensitivity.105 Moreover, typical chest pain during testing is almost as predictive of ischemia as is ST-segment depression.127

As already noted for test sensitivity, the specificity of ST depression in the evaluation of coronary ischemia has also been found to vary considerably, with a mean value derived from a meta-analysis reported to be 72%.102 Referral bias, however, probably reduced these values spuriously, for when this type of bias is minimized, these values generally approach or exceed 90%.37,128129

Differences in results of stress tests between men and women have been the subject of considerable controversy. In general, women are far more likely than men to manifest a false-positive response,130132 but the difference may be attributable to the lower rate of CAD in the populations of women subjected to testing. Based on Bayes theorem, the lower general prevalence of CAD among women results in a relatively low predictive value of a positive test in this group. Some investigators133 have suggested that hormonal effects, especially those of progestin, might be responsible for the production of false-positive ST responses. This effect, if present at all, would be small, for the rate of false-positive stress tests in premenopausal women is low129and only marginally greater than that of men.

Test specificity, ie, the percentage of negative responders in a population known to be free of disease, is of utmost importance in clinical evaluation, and many mistakenly believe that this value is unacceptably low in women. For proper determination of test specificity, a group must be defined that is uniformly proven to be free of disease and in which no prior ECG tests have been performed. Coronary cineangiography is often used as the definitive test to rule out CAD. Unfortunately, such a design is basically flawed because it is a rare individual who has not been subjected to prior stress testing and in whom the results of this testing were not instrumental in the decision to obtain the cineangiographic study. This process produces so-called referral bias, ie, the inclusion of an inordinately high percentage of positive test responders in a disease-free group selected in this manner, thus yielding falsely low specificity values. Such bias may produce greater distortions of test specificity in women simply because, as noted above, falsely abnormal stress responders are more plentiful in mixed populations of women (with and without disease) from which these subjects are drawn. When this type of bias is minimized, however, the difference in specificity between the sexes all but disappears, with a false-positive response rate of ≤ 10% for each.128129 Therefore, in accordance with Bayesian principles, a negative test result encountered in a subject from an unselected population of women generally carries a high negative predictive value and thus is useful in excluding CAD. These data suggest that, in general, the initial evaluation of an individual woman should be identical to that of a man, ie, accomplished with a standard ECG stress test. It need not involve costly nuclear or echocardiographic techniques unless the former test result is positive—a relatively uncommon occurrence in the absence of coronary disease.

Numerous situations appear to give rise to exercise-induced ST depression in the absence of obstruction to the major coronary arteries. Mechanical lesions that place a greater burden on left ventricular dynamics and oxygen requirements include such abnormalities as mitral or aortic valvular dysfunction,134135 pulmonary hypertension,135pericardial constriction,136 and left ventricular hypertrophy.136137 Relative coronary insufficiency is probably also the responsible mechanism in patients with left ventricular hypertrophy. The frequency of positive test results in such cases has been found to be as high as 38%.138 Patients with increased left ventricular mass, even in the absence of standard ECG voltage criteria for this diagnosis, may have false-positive ECG exercise responses.137138

A wide variety of miscellaneous situations has also been associated with falsely positive ST responses to exercise.139These include digitalis administration,140141 hypokalemia,142143 normal postprandial changes,144hyperventilation,145146 postural changes,147vasoregulatory abnormalities,148149 mitral valve prolapse,150151 pectus excavatum,152and intraventricular conduction defect including left bundle branch block and Wolff-Parkinson-White syndrome.153155 There is undoubtedly no common mechanism for ST shifts in these diverse situations; however, effects brought about by electrolyte shifts and sympathetic nervous stimulation at the cellular level may play an important role.

Digitalis is known to cause a false-positive exercise test result in both normal subjects and in patients with heart disease,140141,156 occurring as often as 25% in healthy subjects,141 showing a greater prevalence with increasing age. Hamasaki et al156 found that digitalis-induced ST-segment depression occurs gradually as the heart rate increases in response to exercise, a pattern differing from that usually seen in myocardial ischemia, in which the ST depression progresses more rapidly as peak heart rates are approached. This might aid in distinguishing between drug effect and ischemia.

Hypokalemia is often associated with abnormal exercise responses,142 and these changes can be abolished after potassium repletion. Therefore, in patients who are taking diuretics, the results of the ECG stress test should be interpreted with caution.

Food intake may induce ST-segment and T-wave changes in the resting ECG.157 Significant ST depression also may develop after glucose ingestion in subjects who otherwise had normal exercise ECGs.144 For this reason, exercise testing should be proscribed until ≥ 2 h after a meal to avoid this source of variability.

Although usually causing changes primarily confined to the T waves, hyperventilation is known to produce ST-segment changes in response to exercise that mimic those of myocardial ischemia.145146,158 If this cause is suspected in a given individual with suspicious stress-induced ST changes, that subject may be instructed to voluntarily hyperventilate for a period of 2 to 3 min during rest and with ECG monitoring. If ST changes are produced by this maneuver, they are compared with those encountered during the stress test, and if similar, this suggests a false-positive result induced by hyperventilation. This maneuver, however, should be selectively performed only after the standard stress test, for routine performance before the test is usually unnecessary and may produce dizziness and discomfort and interfere with the proper execution of the standard test.

A peculiar syndrome, sometimes labeled syndrome X, is occasionally encountered, especially in young and middle-aged women, consisting of anginal-type chest pain and abnormal exercise ECG but normal coronary arteriograms.159Exercise-induced coronary spasm and microvascular disease160 are possible causes of this syndrome. Therefore, in this context, the abnormal ECG stress test result may not be truly false positive.

Patients with the mitral valve prolapse and normal coronary arteriograms may have false-positive exercise responses.150 This phenomenon is clinically important because chest pain and vasoregulatory abnormalities are occasionally encountered in these patients.

The secondary ST-segment and T-wave changes in patients with intraventricular conduction defects such as left bundle branch block, ventricular paced rhythm, and Wolff-Parkinson-White (preexcitation) syndrome interfere with proper interpretation of the exercise response.153155,161162 Both false-positive and false-negative responses may be seen in patients with left bundle branch block. Studies in a limited number of patients, however, suggested that the exercise test might be useful even if this latter conduction abnormality is present.163164 Recently, Ibrahim et al165 found that additional exercise-induced J point depression in leads II, aVF, and V5 was suggestive of ischemia in this group. Most useful in their study was a change of ≥ 0.5 mm in lead II. This finding, if confirmed, would be of significant clinical value. False-positive changes are particularly common in patients with the Wolff-Parkinson-White syndrome,,154being observed in as many as 100% of such cases.155

In the case of right bundle branch block, the resting anterior ST-T changes secondary to this conduction abnormality interfere with interpretation of the exercise response. Thus, changes in leads V1 through V3 often falsely suggest ischemia.166The test is still reliable if the ST segment depression is recorded in leads V4 through V6. In one small study,167 however, this conduction abnormality was found to be capable of masking the usual ischemic ST depression in these latter leads, thus producing false-negative ST responses to stress.

Exaggerated atrial repolarization waves may produce spurious depression of ST segments.168170 The atrial repolarization wave is directionally opposite to the P wave and can extend well into the ST segment. Thus, with exercise-induced tachycardia, P wave and atrial repolarization wave amplitudes increase, and the PR segment shortens, thus shifting the atrial repolarization wave toward the ST segment. In practice, this phenomenon may be suspected when apparent ST depression is found together with a prominent P wave with a short, sharply downsloping PR segment, especially notable in the inferior leads.

Apart from the inherent limitations of sensitivity of the exercise test in the detection of myocardial ischemia, certain drugs are known to limit its value even further. Drugs which limit the heart rate response to exercise (β-adrenergic blocking agents, diltiazem, and verapamil) reduce the heart rate and maximum systolic arterial BP during exercise, thus decreasing the left ventricular work and myocardial oxygen requirements and reducing or eliminating the ST-segment depression.171In the process of increasing exercise capacity, nitrates may also prevent or minimize changes in exercise ECG.172 In general, if one wishes to maximize test sensitivity, treatment with any of the drugs noted above should be withdrawn for a sufficient period to allow removal from the body before testing. Depending on the information desired from the test and the need for continued medication, however, the clinician may decide to perform the test in selected cases without drug withdrawal.

A number of reports suggest that exercise testing may be of value even when the resting ECG is abnormal.173175 Most of the patients with abnormal resting ECGs show nonspecific ST-segment and T-wave changes. In general, additional ST depression of ≥ 1 mm with exercise has a diagnostic accuracy approaching that found in the absence of resting changes.173175 Such changes also have been found to have similar prognostic significance as those found in patients with a normal resting ECG.175

In the presence of resting T-wave inversion, normalization of this abnormality in response to exercise may occur in different clinical settings176178: it may result from regional myocardial ischemia or abnormal left ventricular wall motion, but may also occur in normal subjects. Thus, in general, this finding has little specificity. However, if such normalization occurs in conjunction with an infarcted dysfunctional myocardial zone, it may indicate higher coronary flow reserve, and this in turn suggests better preservation of coronary microcirculatory function and the likely presence of myocardial viability.176177 This concept, however, has been challenged by others.52

ST-segment elevation may be seen in the resting ECGs of healthy subjects because of early repolarization. In such cases, the ST segment returns to the isoelectric baseline with exercise, whereas those with significant CAD may have horizontal ST-segment depression.179 Therefore, even in the presence of ST elevation in the baseline ECG, the usual criteria for the interpretation of the exercise test are probably still applicable. As noted above, persistent or increasing ST elevation during exercise may be encountered in myocardial ischemia, prior infarction, or pericarditis.

When large groups of unselected individuals are screened with stress tests, studies consistently show that subjects manifesting abnormal ST responses are at greater risk for subsequent cardiovascular events (angina, acute myocardial infarction, or sudden death).180186 In general, after a follow-up period of 5 to 13 years, those with positive ECG responses show approximately a four- to sixfold greater incidence of such events in comparison with those responding normally to stress. Inability to exercise > 6 min on a standard Bruce protocol (approximately 6 to 7 METs) and inability to increase the heart rate to 85% of age-predicted normal maximum values also were significant indicators of increased risk of coronary events. Individuals demonstrating high exercise tolerance (≥ 10 METs) generally enjoy an excellent prognosis, regardless of the ECG response and even in the presence of known CAD.187189

The combination of various exercise test variables has improved the estimation of long-term prognosis.187,190 Variables found to be independently associated with time to cardiovascular death were weighted to create an equation that calculates a numeric score. The Duke score,187 which is the most widely used and confirmed by others,175 is based on three exercise variables: exercise tolerance in METs, the largest measured ST-segment depression during exercise, and whether angina pectoris could be induced by the test. (Duke Score = exercise time − 5 × {exercise-induced ST depression in millimeters} − 4 × treadmill angina index. Where exercise time is in minutes of Bruce protocol; ST depression is largest stress-induced downward displacement in mm, and angina index is given as the following: 0 for no angina during exercise, 1 for typical angina, and 2 for angina leading to discontinuation of exercise. The Duke score [DS] is then used to calculate annual cardiovascular mortality [CVM]: CVM = −0.00018[DS]3 − 0.0071[DS]2 − 0.143[DS] + 1.60.) A Long Beach Veteran’s Administration score,190 incorporates slightly different variables but yields similar prognostic information. The use of such indexes enhances the practical value of stress testing far beyond that obtained from simple analysis of ECG changes alone, further emphasizing the importance of using symptom-limited protocols for testing. Although patient management should be individualized, the projection of an annual mortality rate of≤ 1% would usually warrant a conservative medical approach to management in favor of aggressive invasive or surgical procedures.

The induction of typical angina pectoris during treadmill stress signifies that the extent of ischemia is greater than in subjects lacking this symptom,105 although diabetics are generally less likely to experience pain. The subsequent survival rate decreases incrementally in proportion to the reduction in exercise duration and effort tolerance.191As a rule, those who have good exercise capacity (approximately ≥10 METs) enjoy a good prognosis, experiencing a 5-year survival of ≥ 95%. However, the probability of survival is much lower in those patients who have ≥ 1 mm ST-segment depression and who are able to achieve a level of exercise equivalent to only Bruce stage 1 (5 METs) or lower. Survival at 5 years in this latter group ranges from 50 to 72%. In view of the data cited above, one may formulate the following general strategy for the use of ECG stress testing in deciding whether and which further tests are required for patients with known or suspected CAD. If an ECG stress test is negative or indicates a low risk (yearly mortality ≤ 1%), medical management with risk-factor control is generally preferred. For those demonstrating high risk (projected mortality ≥ 5% yearly), direct intervention with coronary cineangiocardiography should be strongly considered. For those determined to be at intermediate risk (1 to 5% annual mortality), further stratification with nuclear imaging should be considered before choosing further evaluation or treatment.192 Multiple or extensive nuclear perfusion abnormalities signal the need for invasive study, whereas a normal scan or one containing a small defect warrants a conservative approach. Obviously, initial evaluation with nuclear or echocardiographic procedures with pharmacologic provocation is required for those individuals unable to exercise or who possess disqualifying resting ECG abnormalities. Moreover, other clinical features, such as age and comorbidities, enter into these testing and management decisions.

To assess the presence of ischemia or left ventricular dysfunction and to gauge future risk more effectively, stress testing is often used as early as 4 to 7 days after an acute myocardial infarction193194 and, more recently, 3 days after such an event.195In the past, it was performed typically after an interval of ≥ 2 months had elapsed.196199 The predictive value of early testing, especially when symptom limited, is similar to that of testing performed 6 weeks after infarction.193,200201 If performed early, exercise on the treadmill is commonly stopped when the subject reaches an arbitrary heart rate, which is generally 120 or 130 beats/min, or 70% of predicted maximum heart rate for age. More recently, however, several investigators193195 have found that symptom-limited testing could be performed early and safely in subjects with uncomplicated myocardial infarctions (average 4 to 7 days after the event). Stress-induced ST depression after a single previous myocardial infarction usually identifies those with ischemia resulting from multivessel coronary disease.45,48 A poor BP response during exercise is suggestive of reduced left ventricular function.199 With maximal testing, the ability to achieve a high heart rate–systolic pressure product (≥ 21,700) implies good myocardial blood flow and a favorable prognosis (6 months mortality, 0.8% vs 2% in those failing to reach this product).202 Exercise-induced ST-segment elevation in leads possessing pathologic Q waves usually is associated with greater impairment of left ventricular function because of more extensive damage rather than extent of the CAD,203 but, as noted above, whether it indicates the presence of viable myocardium is controversial. Combining various features of stress testing enhances assessment of prognosis: exercise-induced angina, ST-segment displacement, falling BP or failure to increase BP to ≥ 110 mm Hg, repetitive ventricular arrhythmias, poor exercise tolerance (< 6 METs), and inability to reach exercise heart rate of 120 beats/min (in the absence of β-blockers) are predictive of higher risk for future cardiac events such as unstable angina, recurrent myocardial infarction, and cardiac death.193194,196,204207 The estimated 1-year mortality ranges from approximately 1% if none of these features was present to 17% if three or more were present. Similar findings have been observed both in patients with Q-wave and in those with non–Q-wave infarctions.193194,208 Those receiving thrombolytic treatment apparently enjoy a better prognosis even when these test results are positive.207

With the increasing popularity of various stress-imaging modalities, physicians must consider the proper strategy of patient management after a myocardial infarction before discharge from the hospital. In a meta-analysis,207 stress ECG was compared with myocardial nuclear perfusion and ventricular function imaging. Positive results of all forms of testing yielded relatively low positive predictive values for subsequent events, ie, < 20% chance of death or reinfarction in 1 year. In general, tests to detect reduced left ventricular function (peak stress-induced ejection fraction ≤ 40%) were more predictive of an adverse outcome, averaging from 30 to 40% chance of death or reinfarction in 1 year. Those tests that disclosed myocardial perfusion defects, such as ST depression or reversible nuclear perfusion or wall motion abnormalities, were less predictive of subsequent events. However, negative results of all tests were comparably and independently predictive of lower (< 10%) event rates at 1 year.

From these data above, one might formulate the following strategy for noninvasive testing after a myocardial infarction. In patients undergoing immediate angiography or angioplasty, a predischarge stress test is generally superfluous, and this test may be deferred until 4 to 6 weeks after hospital discharge.195 In the remainder who have sustained an uncomplicated infarction (absence of congestive heart failure, reduced resting ejection fraction, or other adverse markers such as persistent chest pain, hypotension, and so forth), one would proceed first with an ECG stress test, preferably symptom-limited in type. If such patients exercise to > 6 METs without ECG or hemodynamic abnormalities, they are at low risk of a recurrent cardiac event during the ensuing year. Through this means, no further tests are required, costs can be minimized, and the patient can be reassured. If results of this test are abnormal or if resting ECG abnormalities preclude assessment of stress-induced changes, I suggest stress imaging with nuclear or echocardiographic techniques. Exercise is the preferable form of stress unless limitations of patient performance necessitate pharmacologic stimulation with agents such as dobutamine or dipyridamole. In those patients with complicated infarctions or who exhibit evidence of reduced resting left ventricular function, one would proceed directly to stress imaging or invasive study.

Stress testing may be useful in risk assessment before any type of elective surgical procedure. The indications for testing are basically the same as noted in Table 1, especially categories 1 and 2. The features indicative of high risk for perioperative cardiac events are, as expected, poor functional capacity, marked exercise-induced ST-segment shift or angina at low workloads, and a drop or an inability to increase the BP with progressive exercise.209 Those shown to be at low long-term risk generally may proceed directly to surgery without further testing. Patients requiring surgical procedures, however, especially for peripheral vascular disease, are often unable to exercise, and, therefore, evaluation must necessarily begin with nuclear or echocardiographic imaging with pharmacologic provocation. Selection for additional testing, including invasive procedures, before elective surgery is generally the same as that described above for the general population. This subject has been reviewed previously by Chaitman and Miller.209

Premature ventricular beats are often induced by exercise. In studies of middle-aged or older subjects without overt heart disease, approximately 3% or more have premature ventricular beats at rest, increasing by more than twofold (to as high as 50%) at peak exercise, which often includes the new appearance of repetitive ventricular ectopic beats.181,210212 The incidence of such arrhythmias appears to increase with age.213 In general, more arrhythmias are seen on recovery than during exercise. Although ventricular ectopy is more easily evoked in patients with CAD than in normal subjects, the considerable overlap between those with and without ischemia prevents this finding from possessing diagnostic value. The long-term prognosis of asymptomatic subjects manifesting exercise-induced ventricular arrhythmias, including nonsustained ventricular tachycardia, appears to be benign.181,214

In patients with CAD, the reported incidence of exercise-induced ventricular arrhythmias ranges from 38 to 65%.211212,215 In general, the survival rate of patients with CAD, including those with recent myocardial infarction, is decreased if they have exercise-induced complex ventricular arrhythmias.205,216218 Some reports dispute such an association, however,219at least in low-risk patients with demonstrable stable coronary disease.220221 Significant multivessel disease is likely to be present in patients with angina and exercise-induced ventricular arrhythmias,215,222especially if ST-segment changes consistent with myocardial ischemia also are present.223

Thus, the appearance of ventricular arrhythmias in response to exercise in asymptomatic subjects has no diagnostic or prognostic value. When they are found in association with known CAD or with other markers of ischemia, such as ST depression or anginal-type chest discomfort, they generally predict an inordinately high incidence of subsequent cardiac events.

Authorities disagree about whether any form of exercise testing should be performed in asymptomatic individuals.1 In this category, I believe it justifiable to test individuals with multiple risk factors for atherosclerosis (including associated diseases such as chronic renal failure), sedentary individuals who plan to start vigorous exercise, and those who are involved in occupations in which impairment might impact public safety. The use of ultrafast (electron beam) CT is being used with increasing frequency in asymptomatic individuals for the detection of coronary artery calcification.224The presence and amount of calcium detected by this means appear to correlate with the amount of atherosclerotic plaque, but they do not necessarily signal the presence of coronary arterial flow restricting disease. It may offer an earlier means to assess risk and observe responses to therapy. The proper role of this technique, however, is currently controversial and is under review.225

In patients presenting to emergency facilities with acute chest pain syndromes, stress testing with electrocardiography may provide an important means for triage and management with the attendant reduction of medical costs.226227 In those subjects with suspected cardiac pain demonstrating a low or moderate risk for immediate adverse outcomes, management may proceed swiftly to include the following. First, serum enzymes are assessed (total serum creatine kinase and its MB isoenzyme and troponin levels), and if these are normal, a treadmill test is performed either immediately225or after a 6-h observation period with repeat enzyme determination.226 Individuals undergoing this strategy would include those with ECGs that are normal or with only nonspecific repolarization abnormalities, stable BP, no clinical evidence of cardiac decompensation, and absence of prolonged chest pain associated with dynamic ST changes. A normal result of a stress test allows for immediate hospital discharge, whereas an equivocal or positive test result is followed by hospitalization with additional study.

When given a choice among ECG and imaging methods in evaluating patients with recurrent chest pain with intermediate or high likelihood of a cardiac origin, the physician should opt for the most direct and cost-effective means of initial evaluation—the ECG stress test. Men and women should be approached in the same fashion. Even though stress nuclear and echocardiographic imaging are more sensitive in the detection of ischemia, there is insufficient evidence to assume that information so obtained can be more cost-effective or used to improve outcomes beyond that obtained through prognostic assessment used in conjunction with stress ECG alone. If the ECG changes suggest ischemia and clinical circumstances warrant it, an imaging study may then be used. When resting ECG abnormalities preclude interpretation, then initial evaluation should combine stress testing with imaging modalities. When patients are unable to exercise for any reason or if an unsatisfactory rate response to exercise is anticipated (as exemplified by the inability to withdraw treatment with rate-limiting drugs), then imaging studies with pharmacologic provocation should be considered.

Modern exercise testing with ECG monitoring remains a cornerstone of cardiovascular evaluation, providing a valuable source for several types of information. (1) Changes in the ECG pattern—especially depression of the ST segment—indicate the presence and often the severity of myocardial ischemia. (2) When combined with ECG changes, symptoms of dyspnea and chest pain, limitation of maximum performance, and BP response provide valuable markers of the presence and severity of disease and its prognosis. Combined with the history and physical examination, a prognostic index indicative of low risk for future cardiovascular events may allow the clinician to avoid aggressive and costly procedures such as cardiac catheterization.228 (3) Maximum effort tolerance is useful in gauging work and recreational limitations and for use in monitoring efficacy of treatment and in prescribing a safe exercise program. (4) Diagnostic and prognostic data derived in this way can supplement those obtained through simultaneously performed imaging techniques, ie, stress echocardiography or nuclear perfusion study.

Abbreviations: CAD = coronary artery disease; MET = metabolic equivalent

Table Graphic Jump Location
Table 1. Reasons for Stress Testing
Table Graphic Jump Location
Table 2. Contraindications to Stress Testing
Table Graphic Jump Location
Table 3. Indications for Terminating Exercise Testing
Table Graphic Jump Location
Table 4. Relationship Between Treadmill Workloads, General Activities, and Cardiac Classification (New York Heart Association)
Figure Jump LinkFigure 1. Schematic representation of various ST-segment patterns potentially produced by exercise. A: normal; B: junctional depression that returns to baseline (level of PR segment) within 0.08 s (arrow); C: junctional depression that remains below baseline at 0.08 s; D: horizontal ST depression; E: downsloping ST depression;. F: ST elevation. See text for explanation.Grahic Jump Location
Gibbons, RJ, Balady, GJ, Beasley, JW, et al (1997) ACC/AHA guidelines for exercise testing: a report of the ACC/AHA task force on practice guidelines.J Am Coll Cardiol30,260-311. [CrossRef] [PubMed]
Franklin, B, Timmis, GC, O’Neill, WW Is direct physician supervision of exercise stress testing routinely necessary?Chest1997;111,262-265. [CrossRef] [PubMed]
Young, DZ, Lampert, S, Graboys, TB, et al Safety of maximal exercise testing in patients at high risk for ventricular arrhythmia.Circulation1984;70,184-191. [CrossRef] [PubMed]
Martin, CM, McConahay, DR Maximal treadmill exercise electrocardiography: correlations with coronary arteriography and cardiac hemodynamics.Circulation1972;46,956-962. [CrossRef] [PubMed]
Mason, RE, Likar, I, Biern, RO, et al Multiple-lead exercise electrocardiography: experience in 107 normal subjects and 67 patients with angina pectoris, and comparison with coronary cinearteriography in 84 patients.Circulation1967;36,517-525. [CrossRef] [PubMed]
Bruce, RA, Hornsten, TR Exercise stress testing in evaluation of patients with ischemic heart disease.Prog Cardiovasc Dis1969;11,371-390. [CrossRef] [PubMed]
Myers, J, Froelicher, VF Optimizing the exercise test for pharmacological investigations.Circulation1990;82,1839-1846. [CrossRef] [PubMed]
Fletcher, GG, Froelicher, VF, Hartley, LH, et al Exercise standards: a statement for health professionals from the American Heart Association.Circulation1990;82,2286-2322. [CrossRef] [PubMed]
Sanmarco, ME, Pontius, S, Selvester, RH Abnormal blood pressure response and marked ischemic ST-segment depression as predictors of severe coronary artery disease.Circulation1980;61,572-578. [CrossRef] [PubMed]
Weiner, DA, McCabe, CH, Cutler, SS, et al Decrease in systolic blood pressure during exercise testing: reproducibility, response to coronary bypass surgery and prognostic significance.Am J Cardiol1982;49,1627-1631. [CrossRef] [PubMed]
Lele, SS, Scalia, G, Thomson, H, et al Mechanism of exercise hypotension in patients with ischemic heart disease.Circulation1994;90,2701-2709. [CrossRef] [PubMed]
Allison, TG, Cordeiro, MA, Miller, TD, et al Prognostic significance of exercise-induced systemic hypertension in healthy subjects.Am J Cardiol1999;83,371-375. [CrossRef] [PubMed]
Molina, L, Elosua, R, Marrugat, J, et al Relation of maximum blood pressure during exercise and regular physical activity in normotensive men with left ventricular mass and hypertrophy.Am J Cardiol1999;84,890-893. [CrossRef] [PubMed]
Campbell, L, Marwick, HT, Pashkow, JF, et al Usefulness of an exaggerated systolic blood pressure response to exercise in predicting myocardial perfusion defects in known or suspected coronary artery disease.Am J Cardiol1999;84,1304-1310. [CrossRef] [PubMed]
Taylor, AJ, Beller, GA Postexercise systolic blood pressure response: association with the presence and extent of perfusion abnormalities on thallium-201 scintigraphy.Am Heart J1995;129,227-234. [CrossRef] [PubMed]
Tsuda, M, Hatano, K, Hayashi, H, et al Diagnostic value of post exercise systolic blood pressure response for detecting coronary artery disease in patients with or without hypertension.Am Heart J1993;125,718-725. [CrossRef] [PubMed]
McHam, SA, Marwick, TH, Pashkow, FJ, et al Delayed systolic blood pressure recovery after graded exercise: an independent correlate of angiographic coronary disease.J Am Coll Cardiol1999;34,754-759. [CrossRef] [PubMed]
Lauer, MS, Francis, GS, Okin, PM, et al Impaired chronotropic response to exercise stress testing as a predictor of mortality.JAMA1999;281,524-529. [CrossRef] [PubMed]
Dreifus, LS, Fisch, C, Griffin, JC, et al Guidelines for implantation of cardiac pacemakers and antiarrhythmic devices: a report of the ACC/AHA task force on assessment of diagnostic and therapeutic cardiovascular procedures.Circulation1991;84,455-467. [CrossRef] [PubMed]
Ellestad, MH, Wan, M Predictive implications of stress testing: follow-up of 2700 subjects after maximum treadmill stress testing.Circulation1975;51,363-369. [CrossRef] [PubMed]
Wilkoff, BL, Miller, RE Exercise testing for chronotropic assessment.Cardiol Clin1992;10,705-717. [PubMed]
Cole, CR, Blackstone, EH, Pashkow, MJ, et al Heart-rate recovery immediately after exercise as a predictor of mortality. N Engl J Med. 1999;;341 ,.:1351. [CrossRef] [PubMed]
Tavel, ME The appearance of gallop rhythm after exercise stress testing.Clin Cardiol1996;19,887-891. [CrossRef] [PubMed]
Viik, J, Lehtinen, R, Turjanmaa, V, et al Correct utilization of exercise electrocardiographic leads in differentiation of men with coronary artery disease from patients with a low likelihood of coronary artery disease using peak exercise ST-segment depression.Am J Cardiol1998;81,964-969. [CrossRef] [PubMed]
Braat, SH, Kingma, JH, Brugada, P, et al Value of lead V4R in exercise testing to predict proximal stenosis of the right coronary artery.J Am Coll Cardiol1985;5,1308-1311. [CrossRef] [PubMed]
Michaelides, AP, Psomadaki, ZD, Dilaveris, PE, et al Improved detection of coronary artery disease by exercise electrocardiography with the use of right precordial leads.N Engl J Med1990;340,340-345
Miliken, JA, Abdollah, H, Burggraf, GW False-positive treadmill exercise tests due to computer averaging.Am J Cardiol1990;65,946-948. [CrossRef] [PubMed]
McHenry, PL, Fisch, C Clinical applications of the treadmill exercise test.Mod Concepts Cardiovasc Dis1977;46,21-25. [PubMed]
Rijneke, AD, Ascoop, CA, Talmon, JL Clinical significance of upsloping ST segments in exercise electrocardiography.Circulation1980;61,671-678. [CrossRef] [PubMed]
Greenberg, PS, Friscia, DA, Ellestad, MH Predictive accuracy of Q-X/Q-T ratio, Q-Tc interval, S-T depression and R wave amplitude during stress testing.Am J Cardiol1979;44,18-23. [CrossRef] [PubMed]
Sheffield, LT Upsloping ST segments: easy to measure, hard to agree upon.Circulation1991;84,426-428. [CrossRef] [PubMed]
Sansoy, V, Watson, DD, Beller, GA Significance of slow upsloping ST-segment depression on exercise stress testing.Am J Cardiol1997;79,709-712. [CrossRef] [PubMed]
Stuart, RJ, Ellestad, MH Upsloping ST segments in exercise testing.Am J Cardiol1976;37,19-22. [CrossRef] [PubMed]
Hollenberg, M, Go, M, Jr, Massie, BM, et al Influence of R-wave amplitude on exercise-induced ST depression: need for a “gain factor” correction when interpreting stress electrocardiograms.Am J Cardiol1985;56,13-17. [CrossRef] [PubMed]
Ellestad, MH, Crump, R, Surber, M The significance of lead strength on ST change during treadmill stress test.J Electrocardiol1992;25,31-34. [CrossRef] [PubMed]
Cheng, S, Ellestad, MH, Selvester, RH Significance of ST-segment depression with R-wave amplitude decrease on exercise testing.Am J Cardiol1999;83,955-959. [CrossRef] [PubMed]
Froelicher, VF, Lehmann, KG, Thomas, R, et al The electrocardiographic exercise test in a population with reduced workup bias: diagnostic performance, computerized interpretation, and multivariable prediction.Ann Intern Med1998;128,965-974. [PubMed]
Galik, DM, Mahmarian, JJ, Verani, MS Therapeutic significance of exercise-induced ST-segment elevation in patients without previous myocardial infarction.Am J Cardiol1993;72,1-7. [CrossRef] [PubMed]
Longhurst, JC, Kraus, WL Exercise-induced ST elevation in patients without myocardial infarction.Circulation1979;60,616-629. [CrossRef] [PubMed]
Stiles, GL, Rosati, RA, Wallace, AG Clinical relevance of exercise-induced S-T segment elevation.Am J Cardiol1980;46,931-936. [CrossRef] [PubMed]
Waters, DD, Chaitman, BR, Bourassa, MG, et al Clinical and angiographic correlates of exercise-induced ST-segment elevation: increased detection with multiple ECG leads.Circulation1980;61,286-296. [CrossRef] [PubMed]
Lahiri, A, Balasubramanian, V, Millar, CMW, et al Exercise-induced ST segment elevation: electrocardiographic, angiographic, and scintigraphic evaluation.Br Heart J1980;43,582-588. [CrossRef] [PubMed]
Halon, DA, Mevorach, D, Rodeanu, M, et al Improved criteria for localization of coronary artery disease from the exercise electrocardiogram.Noninvasive Cardiol1994;84,331-338
Lee, JH, Crump, R, Ellestad, MH Significance of precordial T-wave increase during treadmill stress testing.Am J Cardiol1995;76,1297-1299. [CrossRef]