10 Exercise Testing

169

From: Essential Cardiology: Principles and Practice, 2nd Ed.

Edited by: C. Rosendorff © Humana Press Inc., Totowa, NJ

Gregory Engel, MD and Victor Froelicher, MD

INTRODUCTION

Exercise can be considered the true test of the heart because it is the most common everyday stress that humans undertake. The exercise test is the most practical and useful procedure in the clinical evaluation of cardiovascular status.

Despite the many recent advances in technology related to the diagnosis and treatment of cardiovascular disease, the exercise test remains an important diagnostic modality. Its many appli- cations, widespread availability, and high yield of clinically useful information continue to make it an important gatekeeper for more expensive and invasive procedures. The numerous approaches to the exercise test however, have been a drawback to its proper application. Excellent guidelines have been developed based on a multitude of research studies over the last 20 yr, and have led to greater uniformity in methods.

ADVANTAGES AND DISADVANTAGES OF EXERCISE TESTING

The standard exercise test surprisingly has characteristics not dissimilar from newer, more expen- sive tests. Table 1 lists its disadvantages and advantages.

INDICATIONS

The common clinical applications of exercise testing to be discussed in this chapter are diagnosis and prognosis. The other applications listed in Table 2 are discussed elsewhere (1). The ACC/AHA Guidelines will be followed in regard to diagnosis and prognosis (2).

METHODS

Safety Precautions and Risks

The safety precautions outlined by the American Heart Association are very explicit in regard to

the requirements for exercise testing. Everything necessary for cardiopulmonary resuscitation must

be available, and regular drills should be performed to ascertain that both personnel and equipment

are ready for a cardiac emergency. The classic survey of clinical exercise facilities by Rochmis

and Blackburn (3) showed exercise testing to be a safe procedure, with approximately one death

and five nonfatal complications per 10,000 tests. Perhaps due to an expanded knowledge concern-

ing indications, contraindications, and endpoints, maximal exercise testing appears safer today than

20 yr ago. Gibbons et al. (4) reviewed 71,914 tests conducted over 16 yr and reported a compli-

cation rate of only 0.8 per 10,000 tests. The risk of exercise testing in coronary artery disease patients

cannot be disregarded, however, even with its excellent safety record. Cobb and Weaver (5) esti-

mated that the risk of arrhythmic events may be 100 times higher in the recovery period.

Most problems can be prevented by having an experienced physician, nurse, or exercise physio- logist standing next to the patient, measuring blood pressure, and assessing patient appearance during the test. The exercise technician should operate the recorder and treadmill, take the appro- priate tracings, enter data on a form, and alert the physician to any abnormalities that may appear on the monitor. If the patient’s appearance is worrisome, if systolic blood pressure drops or plateaus, if there are alarming electrocardiographic abnormalities, if chest pain occurs and becomes worse than the patient’s usual pain, or if a patient wants to stop the test for any reason, the test should be stopped, even at a submaximal level. In most instances, a symptom-limited maximal test is preferred, but is usually advisable to stop if 2 mm of additional ST segment elevation occurs, or if 2 mm of flat or downsloping ST depression occurs. In some patients estimated to be at high risk because of their clinical history, it may be appropriate to stop at a submaximal level since it is not unusual for severe ST segment depression, dysrhythmias, or both to occur only after exercise. If the mea- surement of maximal exercise capacity or other information is needed, it is better to repeat the test later, once the patient has demonstrated a safe performance of a submaximal workload.

Exercise testing should be an extension of the history and physical examination. A physician obtains the most information by being present to talk with, observe, and examine the patient in con- junction with the test. A physical examination should always be performed to rule out significant obstructive aortic valvular disease. In this way, patient safety and an optimal yield of information are assured. In some instances, such as when asymptomatic, apparently healthy subjects are being screened, or a repeat treadmill test is being done on a patient whose condition is stable, a physician need not be present, but should be in close proximity and prepared to respond promptly. The phys- ician’s reaction to signs or symptoms should be moderated by the information the patient gives regarding his or her usual activity. If abnormal findings occur at levels of exercise that the patient

Table 1

Advantages and Disadvantages of the Standard Exercise Test Advantages

Low cost Convenience

Availability of equipment

Ready availability of trained personnel Exercise capacity determined

Patient acceptability

Takes less than an hour to accomplish Disadvantages

Relatively limited sensitivity and specificity No localization of ischemia or coronary lesions No estimate of LV function

Not suitable for certain groups

Requires cooperation and the ability to walk strenuously

Table 2

Additional Applications of the Exercise Test Treatment evaluation

Exercise capacity determination After myocardial infarction Screening

Cardiac rehabilitation

Exercise prescription

Arrhythmia evaluation

Intermittent claudication

Preoperative evaluation

usually performs, then it may not be necessary to stop the test for them. Also, the patient’s activity history should help determine appropriate work rates for testing.

Contraindications

Table 3 lists the absolute and relative contraindications to performing an exercise test. Good clinical judgment should be foremost in deciding the indications and contraindications for exer- cise testing. In selected cases with relative contraindications, testing can provide valuable informa- tion even if performed submaximally.

Patient Preparation Preparations for exercise testing include the following:

1. The patient should be instructed not to eat or smoke at least 2 to 3 h prior to the test, and to come dressed for exercise.

2. A history and physical examination (particularly for systolic murmurs) should be accomplished to rule out any contraindications to testing.

3. Specific questioning should determine which drugs are being taken, and potential electrolyte abnormalities should be considered.

4. If the reason for the exercise test is not apparent, the referring physician should be contacted.

5. A 12-lead electrocardiogram should be obtained in both the supine and standing positions. A base- line abnormality may prohibit testing.

6. There should be careful explanations of why the test is being performed, of the testing procedure including its risks and possible complications, and of how to perform the test. The patient should be told that he or she can hold on initially, but later on should use the rails only for balance.

bbbbb-Blockers

With patients subgrouped according to b-blocker administration as initiated by their referring physician, no differences in exercise score test performance were found in a consecutive group of males being evaluated for possible coronary artery disease; however, if only ST segment elevation/

depression criteria are to be used then 0.5 mm is needed to maintain sensitivity (6). Though perhaps optimal, for routine exercise testing it is unnecessary for physicians to accept the risk of stopping

Symptomatic severe aortic stenosis Uncontrolled symptomatic heart failure

Acute pulmonary embolus or pulmonary infarction Acute myocarditis or pericarditis

Relative

Left main coronary stenosis or its equivalent Moderate stenotic valvular heart disease Electrolyte abnormalities

Severe arterial hypertension

Tachyarrhythmias or bradyarrhythmias

Hypertrophic cardiomyopathy and other forms of outflow tract obstruction Mental or physical impairment leading to inability to exercise adequately High-degree atrioventricular block

a

Relative contraindications can be superseded if benefits outweigh risks of exercise

b

In the absence of definitive evidence, a systolic blood pressure of 200 mmHg and a diastolic

blood pressure of 110 mmHg seem reasonable criteria.

b-blockers before testing. Because of the life-threatening rebound phenomena associated with dis- continuing b-blockers, if they are going to be stopped, it should be done gradually with careful super- vision of the tapering process by a physician or nurse.

Protocols

The many different exercise protocols in use have led to some confusion regarding how phys- icians compare tests between patients and serial tests in the same patient. The most common pro- tocols, their stages, and the predicted oxygen cost of each stage are illustrated in Fig. 1. When tread- mill and cycle ergometer testing were first introduced into clinical practice, practitioners adopted protocols used by major researchers, i.e., Balke (7), Astrand (8), Bruce (9), and Ellestad (10) and their co-workers. In 1980, Stuart and Ellestad surveyed 1375 exercise laboratories in North America and reported that of those performing treadmill testing, 65.5% use the Bruce protocol for routine clinical testing (11). This protocol uses relatively large and unequal 2 to 3 metabolic equivalent (MET) increments in work every 3 min. Large and uneven work increments such as these have been shown to result in a tendency to overestimate exercise capacity (12). Investigators have since recommended protocols with smaller and more equal increments (13,14).

R

T

An approach to exercise testing that has gained interest is the ramp protocol, in which work increases constantly and continuously (Fig. 2). The recent call for “optimizing” exercise testing would appear to be facilitated by the ramp approach, as work increments are small, and, because it allows for increases in work to be individualized, a given test duration can be targeted.

To investigate this, our laboratory compared ramp treadmill and bicycle tests to protocols more commonly used clinically (15). Ten patients with chronic heart failure, 10 with coronary artery dis- ease who were limited by angina during exercise, 10 with coronary artery disease who were asymp- tomatic during exercise, and 10 age-matched normal subjects performed three bicycle tests (25 W/

2 min stage, 50 W/2 min stage, and ramp) and three treadmill tests (Bruce, Balke, and ramp) in ran- domized order on different days. For the ramp tests on the bicycle and treadmill, ramp rates were individualized to yield a test duration of approx 10 min for each subject. Maximal oxygen uptake was significantly higher (18%) on the treadmill protocols versus the bicycle protocols collectively, confirming previous observations. Only minor differences in maximal oxygen uptake, however, were observed between the treadmill protocols themselves or between the cycle ergometer proto- cols themselves.

Fig. 1. The most common protocols, their stages, and the predicted oxygen cost of each stage are illustrated.

Because this approach appears to offer several advantages, we currently perform all our clinical and research testing using the ramp. This approach is empirical and more data from other labora- tories are needed to confirm its utility. A number of equipment manufacturers have developed tread- mills that can perform ramping but simple individual manual stepping up of the grade and speed is possible with any device.

H

R

Age-predicted maximal heart rate targets are relatively useless for clinical purposes. A consis- tent finding in population studies has been a relatively poor relationship of maximal heart rate to age. Correlation coefficients of -0.4 are usually found with a standard error of the estimate of 10 to 25 beats/min. Since prediction of maximal heart rate is an inaccurate science, exercise should be symptom-limited and not targeted on achieving a certain heart rate.

Exertional hypotension, best defined as a drop in systolic blood pressure below standing rest or a drop of 20 mmHg after a rise, is very predictive of severe angiographic coronary artery disease (CAD) and a poor prognosis. A failure of systolic blood pressure to rise is particularly worrisome after an MI. Until automated devices are adequately validated, we strongly recommend that blood pressure be taken manually with a cuff and stethoscope.

Exercise Capacity

The MET is a unit of basal oxygen consumption equal to approx 3.5 mL O

per kilogram of body weight per minute and is the approximate amount of oxygen required to sustain life in the resting state. An individual’s maximal oxygen uptake is normally estimated from the workload reached using a formula based on speed and grade. Maximal oxygen uptake is most precisely determined by direct measurement using ventilatory gas-exchange techniques. In certain circumstances where precision is important, such as in athletics, research studies, and patients considered for cardiac trans- plantation, a direct measurement is essential.

Fig. 2. An approach to exercise testing that has gained interest is the ramp protocol, in which work increases

constantly and continuously.

Exercise capacity should always be reported in METs and not minutes of exercise. In this way, the results from different protocols and exercise modalities can be compared directly. Achieved workload in METs has been shown to be a major prognostic variable.

ST Analysis

ST segment depression is a representation of global subendocardial ischemia, with a direction determined largely by the placement of the heart in the chest. ST depression does not localize coro- nary artery lesions. V

is the lead that most frequently demonstrates significant ST depression. ST depression in the inferior leads (II, AVF) is most often due to the atrial repolarization wave that begins in the PR segment and can extend to the beginning of the ST segment. When ST depression is isolated to these leads and there are no diagnostic Q waves, it is usually a false-positive. ST seg- ment depression limited to the recovery period does not generally represent a “false-positive”

response. Inclusion of analysis during this time period increases the diagnostic yield of the exer- cise test.

When the resting ECG shows Q waves of an old MI, ST elevation can be due to wall motion abnormalities, whereas accompanying ST depression can be due to a second area of ischemia or reciprocal changes. When the resting ECG is normal, ST elevation is due to severe ischemia (spasm or a critical lesion). Such ST elevation is uncommon and very arrhythmogenic. Exercise-induced ST elevation (not over diagnostic Q waves) and ST depression both represent ischemia, but they are quite distinctive: elevation is due to transmural ischemia, is arrhythmogenic, has a 0.1% prev- alence, and localizes the artery where there is spasm or a tight lesion, while depression is due to subendocardial ischemia, is not arrhythmogenic, has a 5 to 50% prevalence, is rarely due to spasm, and does not localize. Figure 3 illustrates the various patterns. The standard criterion for abnor- mality is 1 mm of horizontal or downsloping ST depression below the PR isoelectric line or 1 mm further depression if there is baseline depression.

The most important times to look for ST depression are during maximal exercise in lead V

and 3 min into recovery (16). Patients should be placed supine as soon as possible after exercise, avoid- ing a cool-down walk, to maximize sensitivity for ST abnormalities. ECG recordings should con- tinue for 5 min in recovery or until any new changes from baseline normalize.

Nonsustained ventricular tachycardia is uncommon during routine clinical treadmill testing (prevalence less than 2%), is well tolerated, and its prognosis is determined by any accompanying ischemia or LV damage (17).

DIAGNOSTIC USE OF EXERCISE TEST The ACC/AHA Guidelines

The task force to establish guidelines for the use of exercise testing met and produced guide- lines in 1986, 1997, and 2002 (2). Over the years some dramatic changes have occurred, including the recommendation that the standard exercise test be the first diagnostic procedure in women and in most patients with resting ECG abnormalities rather than performing imaging studies. The following is a synopsis of the recommendations from these evidence-based guidelines for the use of exercise testing to diagnose obstructive coronary artery disease.

Class I. Conditions for which there is evidence and/or general agreement that the standard exercise test is useful and effective.

• Adult male or female patients (including those with complete right bundle branch block or with less than 1 mm of resting ST depression) with an intermediate pretest probability* of coronary artery disease (specific exceptions are discussed in Class II and III sections).

*Pretest probability was determined from the Diamond-Forrester estimates by age, symptoms, and

gender (see Table 4).

Fig. 3. The various patterns of exercise ST shifts possible with ischemia. ST amplitude measurement depends on the ST level at baseline.

Table 4

Pretest Probability of Coronary Disease by Symptoms, Gender, and Age

Typical/definite Atypical/probable Nonanginal

Age

Gender angina pectoris angina pectoris chest pain Asymptomatic

30–39 Males Intermediate Intermediate Low Very low

Females Intermediate Very low Very low Very low

40–49 Males High Intermediate Intermediate Low

Females Intermediate Low Very low Very low

50–59 Males High Intermediate Intermediate Low

Females Intermediate Intermediate Low Very low

60–69 Males High Intermediate Intermediate Low

Females High Intermediate Intermediate Low

a

High = >90%, intermediate = 10–90%, low = <10%, and very low = <5%.

b

There are no data for patients younger than 30 or older than 69 but it can be assumed that coronary artery disease prevalence increases with age.

Class IIa. Conditions for which there is conflicting evidence and/or a divergence of opinion that the standard exercise test is useful and efficacious. Weight of evidence/opinion is in favor of usefulness/efficacy.

• Patients with vasospastic angina

Class IIb. Conditions for which there is conflicting evidence and/or a divergence of opinion that the

standard exercise test is useful and efficacious. Usefulness/efficacy is less well established

by evidence/opinion.

• Patients with a high pretest probability of coronary artery disease.*

• Patients with a low pretest probability of coronary artery disease.*

• Patients taking digoxin with less than 1 mm of baseline ST depression.

• Patients with ECG criteria for left ventricular hypertrophy with less than 1 mm of baseline ST depression.

Class III. Conditions for which there is evidence and/or general agreement that the standard exercise test is not useful or efficacious and in some cases may be harmful.

• Patients who demonstrate the following baseline ECG abnormalities:

• Preexcitation (Wolff-Parkinson-White) syndrome

• Electronically paced ventricular rhythm

• More than 1 mm of ST depression

• Complete left bundle branch block

• Patients who have had a well-documented myocardial infarction or significant disease demonstrated on coronary angiography.

Test Performance Definitions

Sensitivity and specificity are the terms used to define how reliably a test distinguishes diseased from nondiseased individuals. They are parameters of the accuracy of a diagnostic test. Sensitivity is the percentage of times that a test gives an abnormal (“positive”) result when those with the disease are tested. Specificity is the percentage of times that a test gives a normal (“negative”) result when those without the disease are tested. This is quite different from the colloquial use of the word “specific.” The methods of calculating sensitivity, specificity, and other test character- istics are shown in Table 5. Table 6 presents an example in which test performance characteristics are calculated and the effects of differences in the prevalence of disease are demonstrated.

Standards for Studies of Diagnostic Test Performance

Reid, Lachs, and colleagues updated the seven “methodological standards” for diagnostic tests in 1995 (18). The standards are: (1) specify spectrum of evaluated patients, (2) report test indexes for clinical subgroups, (3) avoid workup bias, (4) avoid review bias, (5) provide numerical pre-

Table 5

Definitions and Calculation of the Terms Used to Quantify Test Diagnostic Accuracy True-positive (TP) = number of patients with the disease and a positive result

False-negative (FN) = number of patients with the disease but with a negative result True-negative (TN) = number of patients without the disease and a negative result False-positive (FP) = number of patients without the disease but with a positive result Total population = TP + TN + FP + FN

Sensitivity = percentage of those with the disease who test positive: TP / (TP + FN) ´ 100 Specificity = percentage of those without the disease who test negative: TN / (TN + FP) ´ 100 Positive predictive value (PV+) = percentage of those with a positive test result who have the disease:

TP / (TP + FP) ´ 100

Negative predictive value (PV -) = percentage of those with a negative test who do not have the disease:

TN / (TN + FN) ´ 100

Predictive accuracy (PA) = percentage of correct classifications, both positive and negative:

(TP + TN) / Total population ´ 100

Range of characteristics (ROC) curve = plot of sensitivity vs specificity for the range of measurement cutpoints

Risk ratio (RR) = The ratio of the disease rate in those with a positive result compared to those with a

negative result : TP/(TP + FP) / FN/(FN + TN) = PV+ / FN/(FN + TN)

cision for test indexes, (6) report frequency and management of indeterminate results when cal- culating test indexes, and (7) specify test reproducibility. In their evaluation of 112 studies, there was limited application of the standards with only one standard being fulfilled by more than half the studies, demonstrating inadequate evaluation of most diagnostic tests. The purpose of refining these standards was to improve patient care, reduce health care costs, improve the quality of diag- nostic test information, and to eliminate useless tests or testing methodologies.

Some of the logical and easily appreciated ways to conform to diagnostic test standards in eval- uating exercise testing are by blinding to test interpretation, exclusion of patients with prior MIs, and classifying chest pain. Two subtle standards that are least understood but affect test perfor- mance drastically and are most commonly not fulfilled are limited challenge and workup bias.

Limited challenge could be justified as the first step of looking at a new measurement or test. An investigator may choose both healthy and sick persons, test them using a new measurement, and see if they are different. If no difference were noted then further investigation would not be indi- cated. This approach favors the measurement and a better test would be performed in consecutive patients presenting for evaluation. A measurement or test may function well to separate the extremes but fail in a clinical situation. Workup bias means that the decision of who goes to catheterization is made by the physician using the test results and his/her clinical acumen. So the patients in the study are different from patients presenting for evaluation before this selection process occurs.

This can only be avoided by having patients agree to both procedures (an exercise test and a cardiac catheterization) prior to performing any testing.

Populations chosen for test evaluation that fail to avoid limited challenge will result in predictive accuracy and range of characteristics (ROC) curves greater than truly associated with the test mea- surement. Workup bias can affect the calibration of the measurement cutpoints. That is, a score or ST measurement can have a different sensitivity and specificity for a particular cutpoint when work- up bias is present.

The two studies that have removed workup bias by protocol have included 2000 patients and have considerably different test characteristics (19).

5% 500 with CAD 50% sensitive 250 (TP) 250 (FN) 250/250 + 950

9500 w/o CAD 90% specific 950 (FP) 8550 (TN) = 21%

50% 5000 with CAD 50% sensitive 2500 (TP) 2500 (FN) 2500/3000

5000 w/o CAD 90% specific 500 (FP) 4500 (TN) = 83%

Disease prevalence

sensitivity/specificity Predictive value of an abnormal test Risk Ratio

5% 50% 5% 50%

70/90% 27% 88% 27 ´ 3 ´

90/70% 14% 75% 14 ´ 5 ´

90/90% 32% 90% 64 ´ 9 ´

66/84% 18% 80% 9 ´ 3 ´

Calculation of the predictive value of an abnormal test (positive predictive value) using a test with a sensitivity of

50% and a specificity of 90% in two populations of 10,000 patients: one with a coronary artery disease prevalence of

5% and the other with 50% prevalence. This demonstrates the important influence that prevalence has on the positive

predictive value.

Meta-Analysis of Exercise Testing Studies

Gianrossi et al. investigated the variability of the reported diagnostic accuracy of the exercise electrocardiogram by performing a meta-analysis (20). One hundred forty-seven consecutively pub- lished reports, involving 24,074 patients who underwent both coronary angiography and exercise testing, were summarized. Details regarding population characteristics and methods were evalu- ated including number of ECG leads, exercise protocol, pre-exercise hyperventilation, definition of an abnormal ST response, exclusion of certain subgroups, and blinding of test interpretation. Wide variability in sensitivity and specificity was found (the mean sensitivity was 68% with a range of 23 to 100% and a standard deviation of 16%; the mean specificity was 77% with a range of 17 to 100% and a standard deviation of 17%). The median predictive accuracy (percentage of total true calls) is approx 73%.

To more accurately portray the performance of the exercise test, only the results in 41 studies out of the original 147 were considered (21). These 41 studies removed patients with a prior MI from this meta-analysis, fulfilling one of the criteria for evaluating a diagnostic test, and provided all the numbers for calculating test performance. These 41 studies, including nearly 10,000 patients, demonstrated a lower mean sensitivity of 68% and a lower mean specificity of 74%; this means that there also is a lower predictive accuracy of 71%. In several studies in which workup bias has been lessened, fulfilling another major criterion, the sensitivity is approximately 50% and the spe- cificity 90%; the predictive accuracy is 70% (22). This demonstrates that the key feature of the standard exercise test is high specificity and that low sensitivity is a problem.

Effects of Digoxin, LVH, and Resting ST Depression

LVH, resting ST depression, and digoxin were evaluated in the meta-analysis done as part of the guidelines (2). Only those studies that provided sensitivity, specificity, and total patient numbers and included more than 100 patients were considered. The conclusion from this analysis was that only digoxin had a major effect on test performance.

Gender

There has been controversy regarding the use of the standard exercise ECG test in women. In fact, some experts have recommended that only imaging techniques be used for testing women because of the impression that the standard exercise ECG did not perform as well as it did in men. The recent ACC/AHA guidelines reviewed this subject in detail and came to another conclusion. This posi- tion was based on evidence from meta-analysis as well as 15 studies that considered only women.

Exercise testing for the diagnosis of significant obstruction coronary disease in adult patients including women, with symptoms or other clinical findings suggestive of intermediate probability of coronary artery disease is a Class I indication (i.e., definitely indicated). Women with intermedi- ate pretest probability are those age 30 to 59 with typical or definite angina pectoris, 50 to 69 with atypical or probable pectoris, and 60 to 69 with nonanginal chest pain (see Table 4).

Exercise Test Scores

Improved exercise test characteristics can be obtained by considering additional information

in addition to the ST response. Studies have confirmed that this approach is effective (23). The

Duke score, originally developed for prognostic use, has been extended to diagnosis (24). Simpli-

fied scores derived from multivariable equations have been developed to determine the probability

of disease and prognosis. All variables are coded with the same number of intervals so that the coef-

ficients will be proportional. For instance, if 5 is the chosen interval, dichotomous variables are 0

if not present and 5 if present. Continuous variables such as age and maximum heart rate are coded

into groups associated with increasing prevalence of disease. The relative importance of the selected

variables is obvious and the health care provider merely compiles the variables in the score, mul-

tiples by the appropriate number and then adds up the products. Calculation of the “simple” exer-

cise test score can be done using Fig. 4 (25) for men and Fig. 5 (26) for women.

COMPARISON WITH OTHER DIAGNOSTIC TESTS Nuclear Perfusion and Echocardiography

The performances of exercise echocardiography and exercise nuclear perfusion scanning in the diagnosis of coronary artery disease were compared in a meta-analysis of 44 studies published between 1990 and 1997 (27). Articles were included if they discussed exercise echocardiography and/or exercise nuclear imaging with thallium or sestamibi for detection and/or evaluation of coro- nary artery disease, if data on coronary angiography were presented as the reference test, and if the absolute numbers of true-positive, false-negative, true-negative, and false-positive observations were available or derivable from the data presented. Studies performed exclusively in patients

Fig. 4. Calculation of the simple exercise test score for men.

Fig. 5. Calculation of the simple exercise test score for women.

after myocardial infarction, after percutaneous transluminal coronary angioplasty, after coronary artery bypass grafting, or with recent unstable coronary syndromes were excluded. When the discrim- inatory abilities of exercise echo and exercise nuclear were compared to exercise testing without imaging, both echo and nuclear performed significantly better than the exercise ECG.

Predictive Accuracy

Some test results are dichotomous (normal vs abnormal, positive vs negative) rather than contin- uous like a score. Predictive accuracy (true positive plus true negatives divided by the total popula- tion studied) can be used to compare dichotomous test results. Any score can also be dealt with as a dichotomous variable by choosing a cutpoint. An advantage of predictive accuracy is that it pro- vides an estimate of the number of patients correctly classified by the test out of 100 tested. How- ever, when predictive accuracy is used to compare tests, populations with roughly the same preva- lence of disease should be considered. Table 7, based on published meta-analyses, summarizes the sensitivity, specificity, and predictive accuracy of the major diagnostic tests for coronary artery disease currently available (28). While the nonexercise stress tests are very useful, the results shown are probably better than their actual performance because of patient selection. For studies of diag- nostic characteristics, patients with a prior MI should be excluded, as diagnosis of coronary dis- ease is not an issue in them.

PROGNOSTIC USE OF EXERCISE TEST ACC/AHA Guidelines (2)

Indications for exercise testing to assess risk and prognosis in patients with symptoms or a prior history of coronary artery disease:

Class I. Conditions for which there is evidence and/or general agreement that the standard exercise test is useful and effective.

• Patients undergoing initial evaluation with suspected or known CAD. Exceptions are noted below in Class IIb.

• Patients with suspected or known CAD previously evaluated with significant change in clinical status.

• Low-risk unstable angina patients 8–12 h after presentation who have been free of active ischemic or heart failure symptoms.

• Intermediate-risk unstable angina patients 2–3 d after presentation who have been free of active ische- mic or heart failure symptoms.

Table 7

Comparison of Exercise Testing Subgroups and Different Test Modalities

Patients Sensitivity Specificity Predictive

Grouping Studies (n) (%) (%) accuracy

Meta-analysis of standard ETT 147 24,047 68% 77% 73%

Meta-analysis w/o MI 58 11,691 67% 72% 69%

Meta-analysis of treadmill scores 24 11,788 80%

Consensus treadmill score 1 2000 85% 92% 88%

Electron beam computed tomography 4 1631 90% 45% 68%

Thallium scintigraphy 59 6038 85% 85% 85%

SPECT w/o MI 27 2136 86% 62% 74%

Persantine thallium 11 85% 91% 87%

Exercise ECHO 58 5000 84% 75% 80%

Exercise ECHO w/o MI 24 2109 87% 84% 85%

Dobutamine ECHO 5 88% 84% 86%

Class IIb. Conditions for which there is conflicting evidence and/or a divergence of opinion that the standard exercise test is useful and efficacious. Usefulness/efficacy is less well established by evidence/opinion.

• Patients who demonstrate the following ECG abnormalities:

• Pre-excitation (Wolff-Parkinson-White) syndrome

• Electronically paced ventricular rhythm

• More than 1 mm of resting ST depression

• QRS duration greater than 120 ms

• Periodic monitoring to guide management of patients with a stable clinical course.

Class III. Conditions for which there is evidence and/or general agreement that the standard exercise test is not useful or efficacious and in some cases may be harmful.

• Patients with severe comorbidity likely to limit life expectancy and/or candidacy for revascularization.

• High-risk unstable angina patients.

Prognostic Scores

The DUKE and VA predictive equations represent the “state of the art” in prognostication.

D

T

S

N

Mark et al. studied 2842 consecutive patients who underwent cardiac catheterization and exer- cise testing and whose data were entered into the Duke computerized medical information system (29). The median follow-up for the study population was 5 yr and 98% complete. All patients under- went a Bruce protocol exercise test and had standard ECG measurements recorded. A treadmill angina index was assigned a value of 0 if angina was absent, 1 if typical angina occurred during exercise, and 2 if angina was the reason the patient stopped exercising. Before the test, 54% of the patients had taken propranolol and 11% had taken digoxin. ST measurements considered were sum of the largest net ST depression and elevation, sum of the ST displacements in all 12 leads, the number of leads showing ST displacement of 0.1 mV or more, and the product of the number of leads showing ST displacement and the largest single ST displacement in any lead. This nomogram and an example are shown in Fig. 6.

VA P

E

On the basis of clinical and exercise test data, patients with signs and symptoms of coronary heart disease can be classified into low- and high-risk categories. The latter clearly should be considered for cardiac catheterization, while the former should not, unless their symptoms dictate otherwise.

The problem lies in justifying intervention to improve survival for patients whose symptoms are satisfactorily managed medically. Cardiac catheterization is not needed to do so in the majority of such patients.

The VA predictive rules demonstrate that simple noninvasive clinical indicators can stratify

these patients with stable coronary artery disease into high- or low-risk groups. In our VA popu-

lation, a simple score based on one item of clinical information (history of congestive heart failure

or digoxin use) and three exercise test responses (ST depression, exercise capacity, and change

in systolic blood pressure) can identify a group of patients at high risk for cardiovascular death

(30). Clinical judgment must be applied to decide whether intervention is likely to improve survi-

val in our high-risk patients.

Recovery Heart Rate

Heart rate usually falls rapidly at the end of a bout of progressive exercise. While the rate of the drop in heart rate is related to fitness, more recently it has been shown to be inversely related to survival (31). In general, a decline in heart rate of less than 20 bpm by the first or second minute of recovery is associated with an increased risk of death (32).

SUMMARY

The exercise test is relatively inexpensive and readily available. It can be performed in the doc- tor’s office and does not require injections or exposure to radiation. It can provide diagnostic and prognostic information and can also help determine functional capacity and degree of disability.

Many physicians have learned from experience that the exercise test complements the medical history and the physical exam as no other test can, so it remains the second most commonly per- formed cardiology procedure next to the routine ECG. The renewed efforts to control costs will undoubtedly win more over to this group of enlightened physicians. Convincing evidence that tread- mill scores enhance the diagnostic and prognostic power of the exercise test will add this movement.

The ACC/AHA guidelines for exercise testing clearly indicate the correct uses of exercise test- ing. Since the last guidelines, exercise testing has been extended as the first diagnostic test in women and in individuals with right bundle branch block and resting ST segment depression. When diag- nostic scores and prognostic scores are used with exercise testing, test characteristics approach that of the nuclear and echocardiographic add-ons to the exercise test.

The following rules are important to follow for getting the most information from the standard exercise test:

• The exercise protocol should be progressive, with even increments in speed and grade whenever possible; consider using a manual or automated ramp protocol.

• The treadmill protocol should be adjusted to the patient; one protocol is not appropriate for all patients.

• Report exercise capacity in METs, not minutes of exercise.

• ST-segment measurements should be made at ST

(J-junction) and ST segment depression should be considered abnormal only if horizontal or downsloping.

• Raw ECG waveforms should be considered first and then supplemented by computer-enhanced (filtered and averaged) waveforms when the raw data are acceptable.

Fig. 6. The Duke Treadmill prognostic nomogram and an example.

• A treadmill score should be calculated for every patient; use of multiple scores or a computerized consensus score should be considered as part of the treadmill report.

To ensure the safety of exercise testing, the following list of the most dangerous circumstances in the exercise testing lab should be considered:

• Testing patients with aortic valvular disease should be done with great care because these patients can have a cardiovascular collapse.

• ST segment elevation without diagnostic Q waves is due to transmural ischemia and can be asso- ciated with dangerous arrhythmias and infarction.

• Exertional hypotension accompanied by ischemia (angina or ST depression) can be an ominous sign.

• A cool-down walk is advisable in any high-risk patient where ST segment changes are not critical to diagnosis or the evaluation purpose.

REFERENCES

1. Froelicher VF, Myers J. Exercise and the Heart, 4th ed. Saunders/Mosby Year Book Medical Publishers, 1999.

2. Gibbons RJ, Balady GJ, Bricker JT, et al. ACC/AHA 2002 guideline update for exercise testing: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Exercise Testing). Circulation 2002;106:1883–1892.

3. Rochmis P, Blackburn H. Exercise tests: a survey of procedures, safety, and litigation experience in approximately 170,000 tests. JAMA 1971;217:1061–1066.

4. Gibbons L, Blair SN, Kohl HW, Cooper K. The safety of maximal exercise testing. Circulation 1989;80:846–852.

5. Cobb LA, Weaver WD. Exercise: a risk for sudden death in patients with coronary heart disease. J Am Coll Cardiol 1986;7:215–219.

6. Gauri AJ, Raxwal VK, Roux L, et al. Effects of chronotropic incompetence and beta-blocker use on the exercise treadmill test in men. Am Heart J 2001;142:136–141.

7. Balke B, Ware R. An experimental study of physical fitness of air force personnel. US Armed Forces Med J 1959;

10:675–688.

8. Astrand PO, Rodahl K. Textbook of Work Physiology. McGraw-Hill, New York, 1986, pp. 331–365.

9. Bruce RA. Exercise testing of patients with coronary heart disease. Ann Clin Res 1971;3:323–330.

10. Ellestad MH, Allen W, Wan MCK, Kemp G. Maximal treadmill stress testing for cardiovascular evaluation. Circu- lation 1969;39:517–522.

11. Stuart RJ, Ellestad MH. National survey of exercise stress testing facilities. Chest 1980;77:94–97.

12. Sullivan M, McKirnan MD. Errors in predicting functional capacity for postmyocardial infarction patients using a modified Bruce protocol. Am Heart J 1984;107:486–491.

13. Webster MWI, Sharpe DN. Exercise testing in angina pectoris: the importance of protocol design in clinical trials.

Am Heart J 1989;117:505–508.

14. Panza JA, Quyyumi AA, Diodati JG, et al. Prediction of the frequency and duration of ambulatory myocardial ische- mia in patients with stable coronary artery disease by determination of the ischemic threshold from exercise testing:

importance of the exercise protocol. J Am Coll Cardiol 1991;17:657–663.

15. Myers J, Buchanan N, Walsh D, et al. A comparison of the ramp versus standard exercise protocols. J Am Coll Cardiol 1991;17:1334–1342.

16. Lachterman B, Lehmann KG, Abrahamson D, Froelicher VF. “Recovery only” ST-segment depression and the predictive accuracy of the exercise test. Ann Intern Med 1990;112:11–16.

17. Yang JC, Wesley RC, Froelicher VF. Ventricular tachycardia during routine treadmill testing. Risk and prognosis.

Arch Internal Med 1991;151:349–353.

18. Reid M, Lachs M, Feinstein A. Use of methodological standards in diagnostic test research. JAMA 1995;274:645–

651.

19. 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. Veterans Affairs Cooperative Study in Health Services b016 (QUEXTA) Study Group. Quantitative Exercise Testing and Angiogra- phy. Ann Intern Med 1998;128:965–974.

20. Gianrossi R, Detrano R, Mulvihill D, et al. Exercise-induced ST depression in the diagnosis of coronary artery dis-

ease: A meta-analysis. Circulation 1989;80:87–98.

21. Marcus R, Lowe R, Froelicher VF, Do D. The exercise test as gatekeeper. Limiting access or appropriately directing resources? Chest 1995;107:1442–1446.

22. Morise A, Diamond GA. Comparison of the sensitivity and specificity of exercise electrocardiography in biased and unbiased populations of men and women. Am Heart J 1995;130:741–747.

23. Yamada H, Do D, Morise A, Froelicher V. Review of studies utilizing multivariable analysis of clinical and exercise test data to predict angiographic coronary artery disease. Progress in CV Disease 1997;39:457–481.

24. Shaw LJ, Peterson ED, Shaw LK, et al. Use of a prognostic treadmill score in identifying diagnostic coronary disease subgroups. Circulation 1998;98:1622–1630.

25. Raxwal V, Shetler K, Do D, Froelicher V. A simple treadmill score. Chest 2000;113:1933–1940.

26. Morise AP, Lauer MS, Froelicher VF. Development and validation of a simple exercise test score for use in women with symptoms of suspected coronary artery disease. Am Heart J 2002;144:818–825.

27. Fleischmann KE, Hunink MG, Kuntz KM, Douglas PS. Exercise echocardiography or exercise SPECT imaging?

A meta-analysis of diagnostic test performance. JAMA 1998;280:913–920.

28. O’Rourke RA, Brundage BH, Froelicher VF, et al. American College of Cardiology/American Heart Association Expert Consensus Document on electron-beam computed tomography for the diagnosis and prognosis of coronary artery disease. J Am Coll Cardiol 2000;36:326–340.

29. Mark DB, Hlatky MA, Harrell FE, et al. Exercise treadmill score for predicting prognosis in coronary artery disease.

Ann Int Med 1987;106:793–800.

30. Morrow K, Morris CK, Froelicher VF, Hideg A. Prediction of cardiovascular death in men undergoing noninvasive evaluation for CAD. Ann Int Med 1993;118:689–695.

31. Cole CR, Blackstone EH, Pashkow FJ, et al. Heart-rate recovery immediately after exercise as a predictor of mortal- ity. N Engl J Med 1999;341:1351–1357.

32. Shetler K, Marcus R, Froelicher VF, et al. Heart rate recovery: validation and methodologic issues. J Am Coll Cardiol 2001;38:1980–1987.

RECOMMENDED READING

Ashley EA, Froelicher V. The post myocardial infarction exercise test: still worthy after all of these years. Eur Heart J 2001;22:273–276.

Ashley EA, Myers J, Froelicher V. Exercise testing in clinical medicine. Lancet 2000;356:1592–1597.

Atwood JE, Do D, Froelicher VF, et al. Can computerization of the exercise test replace the cardiologist? Am Heart J 1998;136:543–552.

Do D, Marcus R, Froelicher VF, et al. Predicting severe angiographic coronary artery disease using computerization of clinical and exercise test data. Chest 1998;114:1437–1445.

Do D, West JA, Morise A, Froelicher VF. A consensus approach to diagnosing coronary artery disease based on clinical and exercise test data. Chest 1997;111:1742–1749.

Fletcher GF, Froelicher VF, Hartley LH, et al. Exercise standards. A statement for health professionals from the American Heart Association. Circulation 1990;82:2286–2321. Revised, Circulation 1995;91:580–632.

Froelicher VF. Manual of Exercise Testing, 2nd ed. Mosby Year Book Medical Publishers, 1995.

Froelicher VF, Myers J. Research as part of clinical practice: use of Windows-based relational data bases. Veterans Health System Journal March 1998:53–57.

Froelicher VF, Quaglietti S. Handbook of Exercise Testing, Little, Brown, Boston, 1995.

Morrow K, Morris CK, Froelicher VF, Hideg A. Prediction of cardiovascular death in men undergoing noninvasive eval- uation for CAD. Ann Int Med 1993;118:689–695.

www.cardiology.org