Treadmill exercise testing, part II

Note: This is Part II of a two-partpresentation. The November issue was Part I. INTERPRETATION OBJECTIVE

OF EXERCISE

TEST

RESULTS

RESULTS

FUNCTIONAL CAPACITY. -There are few data available to establish the diagnostic value of maximal cardiac output or maximal oxygen consumption. These are the best measurements of the functional capacity of the cardiovascular system but require techniques and equipment which make their measurement impractical in most clinical situations. Also, there is a wide biologic scatter of these parameters in healthy individuals even when age, sex and activity status are considered. Since both maximal cardiac output and maximal oxygen consumption decline with age, the effects of age and disease are difficult to separate. McDonough et al. measured maximal cardiac output in cardiac patients and found its reduction to be the major hemodynamic consequence of symptomatic coronary artery disease that resulted in functional impairment. l7 Acute reduction in left ventricular performance manifested by decreasing stroke volume and increasing pulmonary artery pressure appeared to be the mechanism limiting cardiac output. Maximal oxygen consumption was linearly related to maximal cardiac output. Patterson et al. studied 43 patients with cardiac disease and compared their functional classification by maximal oxygen consumption and by clinical assessmentB When a discrepancy occurred, the hemodynamic data from cardiac catheterization usually indicated that maximal oxygen consumption more accurately reflected the degree of impairment. Patients who had begun to experience limiting symptoms were found to have a maximal oxygen consumption of less than 22 cc O,/kg/min, and those who considered themselves severely limited had a maximal oxygen consumption of 16 cc O,lkg/min or less. Bruce has found that the inability to complete stage 1 of his treadmill protocol is one of the most important risk markers in the Seattle heart watch study. 77 Margolis et al. found that patients with coronary artery disease achieving stage 4 of the Bruce protocol have an excellent prognosis regardless of ST segment response or subsequent therapy, whereas patients stopping in stage 1 have a poor prognosis that may improve with surgery.62 Other investigators have reported a worse prognosis or greater severity of angiographic coronary artery disease in patients with ST segment depression at relatively low work loads69 61 Functional aerobic impairment should prove to have a 5

prognostic and diagnostic value, although it awaits the Seattle heart watch. As with any test result, ceptions; for example, an asymptomatic man with ally high maximal oxygen consumption has been triple-vessel coronary artery disease.78

definition by there are exan exceptionreported with

F SAMUEL M. FOX, III: Conversely, we have seen middle-aged men suffer infarctions after producing no ST segment displacement or vector

loop change on a recent exercise test. In one case, the patient’s aerobic capacity was found to be distinctly below what would be expected from his activity program. This 37-year-old vigorous bicycle commuter (with a type II hyperlipidemia under good control) had a heart rate of 196 and brief trigeminy while completing only 10 ME%. Within 13 months, he developed a major anterior infarction. We must learn how to sort out

these responses to protect patients better. As technological advances make measuring maximal oxygen consumption and cardiac output convenient and practical, it may be possible to develop limits or discriminant values for these measurements dependent upon age, activity status and sex. Maximal oxygen consumption and maximal cardiac output should have a prognostic implication similar to ejection fraction and left ventricular end diastolic pressure, while ST segment changes should parallel the degree of angiographic coronary artery disease. Until these techniques are available, maximal oxygen consumption estimated from the work load achieved during treadmill testing will remain clinically valuable. F ROBERT A. O'RQURKE: The maximum oxygen consumption and cardiac output reflect the left ventricular performance, whereas the ST segment changes may indicate the extent and severity of myocardial ischemia due to occlusive disease of the coronary arterial circulation.

HEART RATE AND BLOODPRESSURERESPONSE.-The factors that affect the heart rate response to exercise and the wide physiologic variation in maximal heart rate have been discussed previously. Relatively elevated heart rates during submaximal exercise, maximal exercise or recovery could be due to vasoregulatory asthenia, to any condition decreasing vascular volume or peripheral resistance, prolonged bed rest, anemia or metabolic disorders. Relatively low heart rates at any point during submaximal exercise could be due to physical conditioning, enhanced stroke volume or drugs. The common use of beta blockers in patients with coronary artery disease complicates the interpretation of their heart rate response to exercise. Winkle et al. studied 301 healthy middle-aged men with 6 hr of ambulatory ECG monitoring and followed them for 7 years.7g The 34 men with “relative bradycardia” experienced a greater proportion of acute cardiac deaths than the other men. The 26 6

men who died from coronary artery disease had heart rates below the population mean. Ellestad and Wan have defined “chronotropic incompetence” as a heart rate response to the Ellestad treadmill protocol below the 95% confidence limits for age and sex determined for patients referred to their exercise laboratory.81 59 Conditioned individuals were excluded. In a follow-up study, patients with chronotropic incompetence had the same incidence of coronary artery disease as patients with ST segment depression. The mechanism of chronotropic incompetence is poorly understood, but many of these patients have poor left ventricular function and most have multivessel disease. Bruce et al. define the term “heart rate impairment” as the percentage deviation in measured maximal heart rate from predicted heart rates0 They have reported heart rate impairment in patients with coronary artery disease, but its prognostic value has not been demonstrated. Systolic blood pressure can rise above the limits illustrated in Figure 4 with no ominous implications or complications reported. Inadequate systolic blood pressure rises can be due to aortic outflow obstruction or left ventricular dysfunction. Thomson and Kelemen have reported that serious coronary artery disease was found in all of their patients who developed hypotension along with angina during exercise testing.81 Seventeen patients had a systolic blood pressure below resting levels when they developed chest pain and ST segment depression. Two of these patients died suddenly 6 weeks after testing. The remaining 15 were studied with coronary angiography and all had significant disease of the left anterior descending coronary artery and five also had left main lesions. Six of six patients had normal blood pressure responses to exercise testing without angina or ST depression following coronary bypass surgery. Bruce defines the term “left ventricular impairment” as the percentage deviation in the pressure-rate product at maximal exercise from the predicted value. Patients with coronary artery disease have had abnormal values for left ventricular impairment but its prognostic value has not been demonstrated. Morris and McHenry have reported the prevalence of exercise-induced hypotension in 450 patients with definite or suspected coronary artery disease.so Exercise-induced hypotension was defined as a sustained decrease of 10 mm Hg or more in systolic blood pressure. A lesion > 75% in one or more coronary arteries was seen in 272 patients, of whom 25 had exercise-induced hypotension. None of the 88 patients with single vessel coronary artery disease manifested this, while 6 of 96 with two-vessel disease and 15 of 68 with three-vessel disease did so. A decrease in systolic blood pressure occurred in three patients (with obstructive cardiomyopathies) out of 22 patients with noncoronary heart disease. Only one of 156 patients with a normal cardiac cathe7

terization had exercise-induced hypotension. Over 500 healthy men had no decrease in systolic blood pressure during their treadmill tests. Therefore, 15 of 27 patients with exercise-induced hypotension had triple-vessel disease. Exercise-induced hypotension was not found to be a reliable sign in women, however. b SAMUEL M. Fox, III: The further the patient pushes to meet or exceed

his/her maximum heart rate, the more likely it is that there will be a systolic “roll over” with a decline that may exceed 10 mm Hg. This phenomenon is most significant, based on our experience, when it occurs before this state of exhaustion and is heralded by a softening of the upper Korotkoff tones in pressure taken prior to the occurrence of the arithmetic “roll over.” Its occurrence in the admittedly sedentary and obviously deconditioned male appears to be less ominous than the equivalent early and inappropriate systolic decline in an individual who has made a reasonable effort to maintain physical endurance capacity but finds light-headedness and fatigue closing in on rapidly diminishing exertional intensities. The emphasis on total clinical appraisal is long overdue.

The emphasis placed on the exercise ECG has tended to deemphasize the exercise measurements discussed in this section. These measurements may improve the diagnostic value of exercise testing, and may be useful indices for identifying the presence and/or severity of coronary artery disease. The value of any measurement in providing diagnostic information from exercise testing depends upon: (1) the accuracy and completeness with which the measurement has been made in healthy individuals (the reference values), and (2) the effectiveness of which certain limits of the measurement (discriminant values) separate healthy individuals from subgroups with disease. Hopefully, the complete set of reference values presented in Figure 4 will stimulate investigators to determine discriminant values for separating patient groups. Using these discriminant values, sensitivity and specificity can be determined similar to that done for an abnormal ST segment response. Most of these measurements are non-Guassian, requiring that nonparametric statistical tests be used and that discriminant values be determined as percentiles rather than standard deviations or confidence limits. ST SEGMENT RESPONSE. -Early investigators reported changes in the ECG during exercise in normal subjects. The changes included decreased R wave amplitude, right axis deviation, junctional depression of the ST segment and decreased T wave amplitude. The depression of the J junction during exercise in normal subjects was demonstrated to be directly related to increasing heart rate. An exercise-induced increase in P wave amplitude was also demonstrated. Quantitative vectorcardiographic WCG) analyses utilizing 8

modern computer techniques have confirmed these findings.82 The 12-lead ECG is still preferred by most cardiologists to the VCG, but the VCG lead systems make practical the simultaneous recording of the heart’s three-dimensional electrical activity. The resulting time coherence makes possible the measurement and separation of changes in the heart’s electrical activity, both in magnitude and in direction. During exercise, the interval between the spatial maximum of the P wave and the onset of the QRS complex decreases. The magnitude of the P wave increases and the direction of the P vector does not change, consistent with right atria1 overload. There is no significant change in QRS duration. As expected, the QT interval shortens with increasing heart rate. In lateral leads like V, or X, R wave duration decreases while the S wave increases in duration and depth. The magnitude and spatial orientation of the maximum QRS vector remain constant. The terminal QRS vector and the ST vector gradually shift toward the right and superiorly. This is reflected by a deepening S wave and a depression of the J junction with an ascending ST segment in lateral leads (V, and X) and inferior leads (II-aV, and Y). T wave amplitude in lateral and inferior leads is lower than at rest. During the first minute of recovery, the P and T magnitudes markedly increase and gradually all of the measurements normalize. The application of computerized quantitative VCG to exercise testing has made possible some important discoveries regarding ST segment changes. Three factors have been found to influence ST segment changes during exercise: (1) physiologic factors, including heart rate, sex and intracardiac blood volume; (2) changes in myocardial action potentials due to myocardial ischemia; and (3) dyskinetic myocardial areas causing QRS abnormalities in the resting ECG.= Since the VCG leads approximate certain of the standard 12 leads, much of this information can be applied to them. Similar changes can be seen in the following lead groups: Y-II, aV, (inferior/superior), X-I, V,, V, (lateral) and Z-V,, V, (anterior/posterior). The normal ST segment vector response to tachycardia and to exercise is a shift rightward and upward. The degree of this shift appears to have a fair amount of biologic variation. Most normals will have mild early repolarization at rest, which will shift to the isoelectric PR segment line in inferior, lateral and anterior leads with exercise. As previously mentioned, this shift can be influenced further by ischemia and by myocardial scars. When later portions of the ST segment are affected, flattening or downward depression can be recorded. Both local effects and the direction of the spatial changes during repolarization cause the ST segment to have a different appearance in the many surface sites that can be monitored. Though not well demonstrated, it 9

would seem that the more leads with “ischemic” shifts, the more severe the ischemia. Experiments using epicardial mapping, mainly recording action potentials under the epicardial electrodes, have shown that ST segment elevation occurs over severely ischemic areas while areas of lesser ischemia show ST segment depression. In the surface ECG, which is a summation of the many myocardial action potentials, the ST segment response to exercise-induced ischemia can be (1) ST segment elevation, (2) normalization or no change or (3) ST segment depression. Exercise-induced ST segment elevation. - Fortuin and Friesinger have reported the angiographic and clinical findings and the 2-year follow-up of 12 patients with 0.1 mv or more ST segment elevation during or after exercise. 83These patients were selected from about 400 patients who had both coronary angiography and exercise testing. Seven of the subset had previous myocardial infarctions. Nine of the 10 who had angina by history developed it with the exercise test. One patient with atypical chest pain had normal coronaries and improved over the follow-up period. Seven of the eight with exercise-induced ST segment elevation in V, through V, had left anterior descending disease. All four with inferior elevation had right coronary disease. None of the subset had ST segment elevation at rest, but many had Q waves and/or T wave inversion. The authors stated that ventricular aneurysms did not explain the findings. Within 2 years of their exercise, four of the patients had died suddenly; one had a documented myocardial infarction and two had increasing symptoms. Bobba et al. presented four similar patients with exerciseinduced ST segment elevation .84 An increase in R wave amplitude was noted in the leads with ST segment elevation, and ST depression was often noted in other leads. Coronary lesions were again demonstrated in arteries supplying the areas of ST segment elevation. Hegge et al. found 11% of the patients (18/X8) they studied with maximal treadmill testing and coronary angiography to have exercise-induced ST segment elevation in the postexercise 12-lead ECG.85 This relatively high prevalence probably is explained by the additional use of Viv2, leads not monitored in other studies. The elevation was present in the precordial leads only in 12 patients, in the inferior leads only in 5 patients and in both in one patient. Seventeen patients had severe coronary artery disease in the coronary artery supplying the appropriate area and the remaining patient had a normal angiogram. Chahine et al. reviewed the prevalence of exercise-induced ST segment elevation in 840 consecutive patients.@ Twenty-nine, or 3.5% of these patients developed ST segment elevation, similar to the 3% (12/400) reported by Fortuin and Friesinger. In Chahine’s study, only 8 (28%) developed angina during exercise test10

ing, while 25 (86%) had an anterior myocardial infarction on their resting ECG. Twenty-one of the patients had coronary angiography; 90% of these showed critical left anterior descending lesions, and 86% had an abnormal left ventriculogram. CM, and CM, were the only leads monitored, so that only lateral wall ST segment elevation was detected. In the three previous studies, changes occasionally were noted in V, -V, only, and a third of the changes occurred only in II and aV,. It appears that all of Chahine’s patients with ST segment elevation had significant coronary artery disease (no false positives). Of all those who had an old anterior myocardial infarction or a critical left anterior descending lesion among the.840 patients, only about 20% of each group showed ST segment elevation. Sixty-four percent of the patients with left ventricular apical aneurysm displayed ST segment elevation. Manvi and Ellestad presented their findings in 29 patients with coronary artery disease who had abnormal left ventriculogram~.~~ ST segment elevation in CF, was considered. Fortyeight percent of the 29 patients demonstrated ST segment elevation, 33% developed ST segment depression and the remaining 1% had no changes. Elevation occurred in 26 of 2,000 (1.3%) consecutive treadmill tests they performed. Simoons et al. investigated the spatial orientation of the exercise-induced ST changes in relation to the presence of dyskinetic areas as documented by left ventriculography.33 In patients with an anterior infarct, the ST vectors were widely scattered, but were most often directed to the left, anterior and superior. Patients with an inferior myocardial infarction had ST vectors directed rightward and anteriorly and also inferiorly if inferior dyskinesis was present. Anteriorly oriented ST changes were associated with anterior or apical dyssynergy in patients with anterior infarctions. The ST vector shifts with dyskinesis resulted in ST elevation over the dyskinetic area. In patients with dyskinetic areas, the direction of the ST changes varied so widely that only the magnitude of the changes can be used as a criterion of exercise-induced ischemia. An important goal of future investigations would be to determine whether the myocardial areas associated with ST segment elevation have any particular characteristics-that is, are they scars or are they areas which could be improved by revascularization? The angiographic diagnosis of dyskinesis has been subjective and ST segment elevation has not been related to true left ventricular aneurysms. It is more accurate to say that exercise-induced ST segment elevation reflects severe ischemia in underlying myocardium since it usually occurs in the area supplied by vessels with angiographic lesions. This is in contrast to exercise-induced ST segment depression which does not appear to accurately predict the site of coronary lesions. 11

Normal repolurization due to cancellation. -The frequency with which cancellation forces result in normalization of ST-T changes or no ST changes in the surface ECG during exerciseinduced ischemia is unknown. Its recognition is somewhat dependent upon the number of leads monitored, i.e., ST elevation in inferior leads cancelling depression in lateral or anterior leads or vice versa. Patients with three-vessel coronary artery disease would be most likely to have cancellation occur, but they have the highest prevalence of abnormal tests. Manvi and Ellestad reported that 20% of their patients with dyskinesis and coronary artery disease had normal tests and Chahine et al. reported that about 25% of their patients with dyskinesis and coronary artery disease normalized or minimally elevated their ST segments. 86,87 Noble et al. have reported normalization of both resting inverted T waves and depressed ST segments in 11 patients during exercise-induced angina.“” When exercise testing fails to produce ST segment depression or elevation in a patient with known coronary artery disease, this could be due to two or more severely ischemic myocardial segments causing cancelling ST vectors. Sweet and Sheffield have reported a patient with minor ST depression and T wave inversion in V, who normalized or “improved” his EGG during treadmill testing only to have an acute infarction 10 min after the testF9 In summary, the prevalence of the cancelling of surface ST changes by multiple ischemic vectors is unknown. Chronotropic incompetence and the inability of patients to give an adequate effort are more likely explanations of the largest portion of false negative exercise tests in patients with multivessel coronary artery disease. The decreased sensitivity of exercise testing in single vessel disease is most likely due to insufficient myocardial ischemia. Exercise-induced ST segment depression. -The classic criterion for abnormal ST segment depression has been 0.1 mv or more of horizontal or downward sloping ST segment depression for 80 msec. The probability and severity of coronary artery disease are directly related to the amount of J junctional depression and are inversely related to the slope of the ST segment. Because of these related factors, computer indices such as ST index and ST integral, which consider both slope and depression, should prove to be superior to classic criteria.6> I1 Downward sloping ST depression is more serious than horizontal depression and both are more serious than upsloping depression.s1 However, patients with upsloping ST segment depression, especially when the slope is less than 1 mv/sec, probably are at increased risk.@’ If a slowly ascending slope is utilized as a criteria for abnormality, the specificity of exercise testing will decrease (more false positives) though the test becomes more sensitive. One electrode can show upsloping ST segment depression while an adjacent 12

electrode shows horizontal or downward sloping depression. If a borderline appearing ST segment with an inadequate slope is recorded in a single precordial lead in a patient highly suspected of having coronary artery disease, multiple precordial leads should be scanned before the exercise test is called normal. An upsloping depressed ST segment may be the precursor of abnormal ST depression that will follow in the recovery period or with a higher heart rate during greater work loads. It is preferable to label tests with a slowly upsloping ST segment and ST segment depression “borderline” responses, and put added emphasis on other parameters. A number of factors can modify both the amount and slope of the ST segment. One is the negative reference for the monitored precordial leads. When both CM, and V, are simultaneously recorded, there is usually more ST segment depression and a more positive slope in CM, as compared to V?.90 The position that the patient assumes in the postexercise perrod also is influential. In accordance with the Laplace relationship, increasing heart volume causes an increase in myocardial oxygen consumption, and so the assumption of the supine position after exercise enhances ST segment depression compared to sitting. Having patients perform a cool-down walk after the treadmill test can delay and alter the appearance of postexercise ST segment depression. 91 Also, maximal testing can increase the prevalence of abnormal responses and worsen the amount and time of the depression compared to submaximal testing.53 In addition to these considerations, it is uncertain that the classic criteria have the same significance in all leads. The optimal criteria for abnormality in the inferior leads, various bipolar leads and in the Frank X lead most likely differ from those for V,, b FRANK I. MARCUS: It is important to realize that the degree of ST segment depression may differ depending upon the lead system used. EXERCISE-INDUCED VENTRICULAR DYSRHYTHMIAS.Since exercise increases myocardial oxygen consumption and may produce myocardial ischemia in the presence of coronary artery disease, it is logical to assume that exercise may also provoke ectopic electrical activity. Exercise-induced supraventricular dysrhythmias are unusual and have not been related to coronary artery disease. Of prime concern, however, are exercise-induced premature ventricular complexes. Do they have any diagnostic or prognostic value, and when should they be treated? Table 3 summarizes the major clinical and population studies regarding exercise-induced premature ventricular contractions (PVCs). A follow-up study of men screened with maximal treadmill testing has demonstrated the diagnostic value of exercise-induced PVCs in an asymptomatic populationq2. In this study, 13

NA

NA

11

26

5

NA

1

214

8

31-37

84-9.5

NA

NA

NA

462

Maximal

Healthy policemen

THE

47-49

18.5-20

NA

NA

NA

81

37 treadmill

Possible ccxO”ar” disease-

ET AL.!” -

EVALUATING DYSRHYTHMIAS

--FARIS

1 1

150

183

3I 35

196

1,390

40-59 Submaximal to heart rate 150

37 Mammal treadmill

No of subJects Mean maximal heart rate W with ~2 PVc&ni” t with -‘2 PVWmin % with multiform PVC8 % with WuDlets or big&“y % with 23 consecutive PVCe Total ‘% with PVCs

BLACKBU~N ET a.*5

High risk, hut without overt disease

Asymptomatic ai rerewnlen

3. -SIX STUDIES VENTRICULAR

FROELTCHERET AL.=

TABLE

-

IN

0 44

NA

6

10

35

175

144

44

Healthy

Cl 16

NA


2

14

150

Chest pain: “0lTlUd CQlV”iWY alteries 44 Maximal or 85% submaximal treadmill 141

3 29

NA

4

10

20

128

197

48

ChO”~Y disease

7 56

13

5

12

NA

136

Maxunal or nearmaximal

16 88

24

23

4

NA

57 24-hr ambulatory monitor 100

patients

RYAN ET AI..~~ Coronary

OF EXERCISE-INDUCED POPULATIONS

MCHENFX ET AL.~~--

PREVALENCE DIFFERENT

ATKINS ET AL.~

15 38

NA

NA

31

NA

86

Isometrics 45

55

2 22

NA

NA

20

NA

122

MiXl”LXl bicycle

Heart disease patients and one normal subject

TABLE 4.-INITIAL TIME OF APPEARANCE OF OTHER THAN SINGLE OR OCCASIONAL PREMATURE VENTRICULAR CONTRACTIONS TlME OF APPEARANCEOF PVCS Late exercise Throughout exercise Early in exercise Midexercise Recovery only

72 OF MEN 40 30 15 11 4

1,390 men were followed for a mean of 6.3 years and the endpoints for coronary artery disease were angina, sudden death or acute myocardial infarction. PVCs were considered “frequent” when they occurred at a rate of 20% or more in any series of 50 beats. PVCs were considered “ominous” when their rate became “frequent” during exercise, when two consecutive PVCs occurred with other ectopy that increased with exercise or when three or more PVCs occurred consecutively. More than a third of the men developed exercise-induced PVCs; 12% developed only a single PVC and 14% developed occasional PVCs. Table 4 lists the initial time of appearance of PVCs other than single or occasional beats. PVCs were usually first seen during exercise. However, in patients who developed frequent PVCs, the PVCs often were most frequent during recovery. As shown in Table 5, the occurrence rate of PVCs was directly related to age. In Table 6 are listed the patterns that occurred in the 2.1% of the men who developed “ominous” PVCs. As shown in Table 7, the diagnostic value of exercise-induced “ominous” PVCs is limited both by a low sensitivity and a low predictive value for endpoints of coronary artery disease. Blackburn et al. studied 196 asymptomatic men and found that 21% had some kind of exercise-induced PVCs, while 7.7% had couplets, bigeminy or ventricular tachycardia.93 Of the patients with PVCs, 30% developed them only at the highest exerTABLE

5. -PERCENTAGE OF MEN EACH AGE GROUP WITH OTHER THAN SINGLE OR OCCASIONAL PREMATURE VENTRICULAR CONTRACTIONS IN

AGE IYB) 20-29 30-39 40-53

c~OMINOUs” ONLY I%)

ALL OTHE&S t R )

0.8 1.0 3.5

6.8 7.6 13.1 15

TABLE 6.-“OMINOUS” PATTERNS OF EXERCISE-INDUCED PREMATURE VENTRICULAR CONTRACTIONS DETECTED IN 29 OF 1,390 ASYMPTOMATIC MEN %MINOUS”

PVCs that increase 3 consecutive PVCs 2 consecutive PVCs exercise 3 consecutive PVCs 4 or more consecutive

P&lTERNS

NO.

to frequent

with

exercise

plus

that

increase

others

plus others PVCs plus

4 5 with 9 10 1

others

cise workload. The PVC prevalence was less than 3% when the heart rate was 73- 102 beatslmin but increased to 50% when the heart rate was greater than 173 beats/min. After 18 months of exercise training, PVC prevalence tended to decrease. Faris et al. studied the ventricular dysrhythmias induced by maximal treadmill testing of 543 apparently healthy policemen.g4 Dysrhythmia prevalence for the group as a whole was not significantly different between two treadmill tests performed 3 years apart. The age group 25- 34 years had a PVC prevalence of 30% in the first test and 36% in the second test. The age group 45- 54 years had a PVC prevalence of 36% in the first test and 42% in the second test. The reproducibility in individual patients was only slightly greater than by chance and without relationship to type or complexity. Fifteen percent of the men with possible coronary artery disease had exercise-induced PVCs that were reproducible. The poor reproducibility of exercise-induced PVCs limits the ability of a single exercise test to predict future coronary artery disease. However, those subjects with reproducible PVCs may be at high risk. McHenry et al. compared exercise-induced PVCs in normal subjects and in patients with documented coronary artery disease.g5 Three subgroups were studied: (1) apparently healthy subjects, (2) patients with chest pain and normal coronary angiograms and (3) patients with documented coronary disease. The prevalence of PVCs was 44%, 16%, and 29%, respectively. Complex PVCs were not frequent in any group, although venTABLE 7. -DIAGNOSTIC VALUE OF EXERCISE-INDUCED “OMINOUS” VENTRICULAR ECTOPY FOUND IN 29 OF THE 1,390 ASYMPTOMATIC MEN IN THE FOLLOW-UP STUDY sENSITI”ITY 6.7% 16

SPECIFICITY 97%

PREDICTIVE VALUE 10%

REX 3.4x

tricular tachycardia was most often seen in patients with coronary artery disease. There was a higher prevalence of PVCs in the normal group because they exercised to a higher workload and heart rate. No correlation was noted between ST segment depression and PVCs. Also, no relationship was seen between PVCs and angina; most of the patients had PVCs before angina. In a similar study, Jelinek and Lown found that a quarter of their angina-limited patients had PVCs during exercise compared to 42.7% of their coronary patients who were able to exercise to fatigue.% In the study by McHenry et al. the importance of exercise heart rate was also demonstrated. There was a 27% prevalence of PVCs with heart rates up to 130 beats/min in the coronary patients, compared to a prevalence of 7% in healthy subjects. Malignant forms of PVCs were also more common at submaximal heart rates in the patients with coronary artery disease. The association of three vessel disease and abnormal left ventricular wall motion with an increased prevalence of exercise-induced PVCs was in agreement with the study of Goldschlager et aZ.% PVCs were suppressed by continued exercise in approximately one third of the patients with coronary artery disease. w FRANK I. MARCUS:This complete review of the subject should dispel the popular, yet erroneous. belief that extrasystoles disappear with increasing exercise, are “benign” and are not often associated with ischemit heart disease as compared with those that increase in frequency with exercise.

Ryan et al. studied 100 patients with coronary artery disease with both ambulatory ECG monitoring and treadmill testing.98 They found an 88% prevalence of PVCs in the patients during ambulatory monitoring compared to 56% prevalence during treadmill testing. Multiformity, couplets, bigeminy and ventricular tachycardia were also more common during ambulatory monitoring. Hinkle et al., using ambulatory monitoring in healthy, middle-aged men, found 62% to have PVCs with 33% showing multiform beats, 21% couplets and bigeminy and even 3% showing ventricular tachycardia.7Y Ambulatory monitoring reveals a higher prevalence of PVCs and is more effective in detecting serious dysrhythmias, including ventricular tachycardia, than treadmill testing. Atkins et al. compared the induction of dysrhythmias using isometric exercise and then using bicycle ergometry in 45 patients.99 There was almost twice the prevalence of PVCs during isometric exercise as compared to bicycle exercise and ventricular tachycardia was also more frequent. This difference occurred in spite of a much lower heart rate during isometric exercise and may be due to a relatively increased sympathetic drive. A major problem still exists in the diagnosis and management 17

of patients considered to be normal but who have frequent PVCs. Helfant et al. studied 60 patients with PVCs.““’ Thirtyeight patients developed or increased PVCs with exercise. Of these 38 patients, 16 or 42% did not have coronary artery disease. The remaining 22 patients decreased their ectopic activity with exercise. Of these 22 patients, 6 or 27% had coronary artery disease, supporting the previous finding that patients with coronary disease could have dysrhythmias suppressed by exercise. In the 14 normals, the PVC frequency decreased in 8 and increased in 6, making an individual diagnosis impossible. Kennedy and Underhill reported 25 extensively studied healthy subjects with PVCs who had been followed for an average of 6 years .lol PVCs were frequent with all patients having bigeminy, trigeminy or both, and 57% having ventricular couplets. Maximal treadmill testing completely suppressed PVCs in 92% of these patients. In summary, PVCs occur in approximately one third of the asymptomatic men who perform a maximal treadmill test and the PVC prevalence is directly related to age. PVCs occur most frequently at maximal exercise and often are not reproducible on repeat testing. A subgroup of apparently healthy men (about 2%) will have serious exercise-induced PVCs. This subgroup will have three times the normal risk of developing coronary artery disease, but only about 10% of them will do so. Furthermore, only 7% of those who develop disease will have had “ominous” PVCs. Patients with coronary artery disease usually have a higher prevalence of serious PVCs, and their PVCs occur at lower heart rates than in healthy subjects. Dysrhythmias suppressed by exercise do not rule out the presence of coronary artery disease. Ambulatory monitoring and isometric exercise can demonstrate PVCs in more people than can dynamic exercise testing. The demonstration of the prognostic significance of exercise-induced PVCs in coronary patients and the value of medical suppression will require follow-up studies. However, studies have shown that serious PVCs detected by ambulatory monitoring performed 3 weeks after a myocardial infarction will identify patients at high risk. lo2, lo3 PVCs induced by exercise testing may be as predictive but more specific in these patients. In addition, the total information gleaned from an exercise test may be more helpful in patient management and more cost-effective than ambulatory monitoring. b SAMUEL M. Fox, III: Although not frequent, the “R-on-T,” or ‘rprompt,” PVC is considered ominous, occurring in or near the vulnerable period remaining from the preceding ventricular activation. When seen we believe the test usually should be stopped, serum electrolytes drawn immediately and a subsequent evaluation undertaken with a reenforced team with an intravenous drip or “heparin-lock” in place. We 18

have seen two cases with volume depletion, hypokalemia and ST segment displacement but without significant coronary artery disease. Since this text was prepared the Stanford group has published their careful evaluation of early postinfarct exercise testing demonstrating its usefulness relative to both ventricular ectopy and ST displacement.140 ) ROBERT A. O'ROURKE: At the present time, continuous ambulatory ECG monitoring is an accurate and sensitive means of detecting serious ventricular premature depolarizations and ventricular tachycardia in patients with ischemic heart disease including both those with and without a prior myocardial infarction. As mentioned above, the poor prognostic significance of serious ventricular arrhythmias detected by this method in patients with ischemic heart disease has been demonstrated. The significance of premature ventricular depolarizations in these patients during treadmill exercise, particularly at peak exercise and during early recovery, remains unknown (Kotter et al.: Circulation 47:959, 1973 and Crawford et al.: Circulation 50:301, 1974).

SUBJECTIVE

RESULTS

APPEARAN~E,~KINTEMPERATURE,FINDINGS INATION. - Careful observation of the

ONPHYSICALEXAM-

patient’s appearance is necessary for the safe performance of an exercise test and is helpful in the clinical assessment of a patient. It is usually easy to identify patients who exaggerate their limitations or symptoms and those unwilling to cooperate. A drop in skin temperature during exercise can indicate an inadequate cardiac output with secondary vasoccmstriction and can be an indication for not encouraging a patient to a higher workload. Neurologic manifestations such as lightheadedness or vertigo can also be an indication of an inadequate cardiac output. ) SAMUEL M. Fox, III: We have found skin temperature useful but often hard to evaluate under a profuse perspiration. For those patients who develop a bright erythema on the upper back we have found the postpressure release refilling rate and color pattern of value. If the red color promptly returns and is even accentuated where the erythema has been pressed out by finger pressure, we find the patient will maintain good cardiac output relative to demand a bit longer. When the color returns slowly, unevenly and in purplish blotches we find the Korotkoff tones getting soft at the top, a systolic “roll over” starting and the patient’s verbal responses less acute.

Findings on physical examination can be helpful, but their sensitivity and specificity have not been demonstrated. Gallop sounds, a mitral regurgitant murmur or a precordial bulge could be due to left ventricular dysfunction. The findings of congestive heart failure, including rales and neck vein distention, should rarely be encountered in patients referred for exercise testing. However, some exercise testing laboratories use the sitting posi19

tion for the recovery period in order to avoid problems patient who develops orthopnea.

with the

) SAMUEL M. Fox, III: The S, and the much less frequent S., found before exercise in the well-trained athlete are almost always absent or late in returning after exertion. This is probably due to the rapid relaxation of the healthy, compliant myocardium. The occurrence or accentuation of an apical S, and S, after exercise is highly suggestive of a dysfunctional left ventricle. Some of these patients, however, respond well and promptly to a physical conditioning program. Although adding to suspicion, a postexertional gallop is not very specific for coronary artery disease. Occasionally we find a postexertional S, and El, at the xiphoid and not at the apex, especially with pectus excavatum deformity, which we believe is due to right ventricular distention. It usually disappears after 2 - 3 minutes and we have not found any ominous connotation as yet. When not previously present, the appearance of a mitral regurgitant murmur after exercise has the ominous connotation of papillary muscle dysfunction secondary to the ischemic effects of coronary artery narrowing. Although seen in cardiomyopathies as a result of other causes, the largest number are associated with severe coronary disease. In contrast, the murmur of mitral valve prolapse frequently decreases in intensity and/or duration. This may be secondary to left ventricular dilation pulling the chordae tendineae down so that the leaflets prolapse less deeply into the atrium.

Attempts to make the findings of the physical examination less subjective include the use of phonocardiography, apexcardiography and displacement cardiography. Signal averaging techniques make it possible to record gallop sounds during exercise.‘” Apexcardiography and displacement cardiography (a new technique using electromagnetic energy) enable the detection of dyskinetic areas in the anterior left ventricular wall during the recovery period after exercise if the patient can hold his breath. The finding of a dyskinetic area along with ST segment depression may improve the specificity of the exercise test for diagnosing coronary artery disease, but this combination could also occur in a cardiomyopathy. Normal anterior wall movement possibly could identify false positive ECG responses; however, ischemia does not always cause anterior wall dyskinesis. ) FRANK I. MAFXUS:The use of systolic time intervals has been shown to enhance the sensitivity of the stress exercise testing (Broder and Anbe: Clin. Res. 20365, 1972; also Lewis et al.: Am. J. Cardiol. 37: 787,1976). In patients with ischemic heart disease, the left ventricular ejection time increased by more than 30 m/set 4 minutes after exercise, as compared to the pre-exercise tracing. This was not observed in subjects with normal coronary arteries. This finding was seen in some patients with significant coronary disease who did not have a positive test by other criteria. CHESTPAIN 20

INDUCED

BYEXERCISE

TESTING. -Weineret

al. have

reported 281 consecutive patients studied with treadmill testing and coronary angiography with the following responses: (1) 76 patients with an abnormal ST response and treadmill test-induced chest pain; (2) 85 patients with an abnormal ST response and no chest pain; (3) 40 patients with treadmill test-induced chest pain, who had no ST changes; and (4) 80 patients with neither chest pain nor ST changes.lo5 They found that 91% of the first group, 65% of the second group, 72% of the third group and only 35% of the fourth group had significant angiographic coronary artery disease. Cole and Ellestad followed 95 patients with an abnormal maximal treadmill test.lM At 5 years of follow-up, the incidence of coronary disease was 73% in those with both chest pain and an abnormal ST segment response as compared with 43% who only had an abnormal ST segment response. Mortality was also twice as high in those with both ST segment changes and chest pain induced by the treadmill test. The results of these studies suggest that ischemic chest pain induced by the exercise test predicts the presence of coronary artery disease as well as ST segment depression and when they occur together, they are even more predictive of disease than either alone. THE DIAGNOSTIC

VALUE

OF EXERCISE

TESTING

In order to make a probability statement regarding the presence or absence of coronary artery disease in a patient after an exercise test, it is essential to know the sensitivity and specificity of the test. Sensitivity is the percentage of patients with significant angiographic disease who have abnormal exercise tests; specificity is the percentage of patients with normal angiograms who have normal exercise tests. Though there are a number of objective and subjective results that must be considered in deciding whether an exercise test is normal or abnormal, this section will deal only with the diagnostic value of ST segment depression. Coronary angiography is an objective indicator for coronary artery disease and can be used to validate the exercise test, but it has limitations.1u7 The ECG response to the exercise test identifies functional myocardial ischemia, whereas the angiogram provides anatomic evidence of coronary obstruction. For example, many patients who have had a previous myocardial infarction no longer have ischemic myocardium and will have a normal ST response, yet have coronary artery disease. b SAMUEL M. FOX, III: Although angiography can be an objective indicator of coronary artery disease, we are impressed with the frequent difficulty in assessing the luminal obstruction that exists. This comment is not made to “knock” angiography, for it is a technique of tremendous

value when performed well and interpreted with care. As with many other techniques, however, it has its share of interobserver and from time to time intraobserver variability. Precise angiographic criteria which establish or rule out the presence of &hernia do not exist. Among the reports reviewed here, criteria for abnormality vary from 50% to 75% diameter reduction of a major coronary artery. However, comparing the sensitivity and specificity of exercise testing according to the angiographic criterion used reveals no difference between studies employing 50% as a criterion versus those using 75% diameter reduction. The studies which carefully evaluated sensitivity in terms of angiographic severity of coronary disease all found increasing sensitivity of the test as more vessels were involved (Table 8). Most false negatives were found among patients with singlevessel disease. Sensitivity is decreased for the milder degrees of disease, but it is also likely that some patients with single vessel disease do not have myocardial ischemia. Most reported studies use the criterion of 0.1 mv horizontal or downward sloping ST segment depression to indicate an abnormal exercise test. A number of reports have used 0.05 mv depression, and all found the decrease in specificity unacceptable despite the increase in sensitivity. Conversely, the criterion of 0.2 mv depression is lacking in sensitivity despite its greater specificity.” Few studies have indicated the occurence of ST segment elevation or used it as a criterion for abnormality. Although many of the available studies do not comment on the resting ECG as a significant parameter in the evaluation of the results of an exercise test, several studies were limited to patients with a normal resting ST segment in the monitored leads. This improves the results since specificity is reduced when ST segment depression is present at rest. Unfortunately, restricting the exercise test to patients with normal resting ST segments would exclude many of the patients referred for the test. A major advantage of monitored progressive treadmill tests over step tests has been the safe performance of an increased work load. Maximal and submaximal have been used to describe the amount of exercise performed. Confusion has arisen because these terms have been applied both to the degree of the patient’s effort and to a specified target heart rate. A patient may have performed a maximal exercise test, stopping well before achieving anywhere near a “maximal” heart rate because of fatigue. In most reported studies, patients have exercised to a target heart rate (85- 100% of age-adjusted maximum) unless unable to continue exercise because of symptoms or unless marked ST segment depression or serious arrhythmias occur. In the event of an 22

8. -STUDIES PATIENTS

Kassebaum et ~1.‘“” Martin and McConahaylw McHenry et aZ.11u Helfant et al.“* Bartel et al.112 Goldschlager et ~1.~’ Mason et ~1.36 Roitman et al.‘l3 Average

STUDY

TABLE

8% 95%

83% 94% 93% 89%

82%(

63 609

410 84 100 90%

97%

68

3Y39)

(V37) pY.50) (95/30) ('7Ylsl) (Vl41) (V35) (V31) (V.51)

43%

-

40% Ws7)

39%

60% (g/m)

61%

25% (Y8) 35% p/23)

1 YESSEL

THE DIAGNOSTIC STUDIED WITH

(3%4)

BEEN

100 166

NO.

EVALUATING WHO HAVE

67%

-

83% 62% 63%

91%

38% 67%

W8e)

(“1/m)

(‘o/~z)

(?43)

(‘2/m)

(T13)

2 “ES8Els

VALUE OF CORONARY

69%

86%

64%

7% 63%

81%

53% 62%

78% 73%

W116)

('o/11) (9%27)

W12)

P/22)

(V13)

IN

-

79%

73%

91%

100%

85% 86%

3 YE9aEI-S

EXERCISE TESTING ANGIOGRAPHY

(%32)

( '73/2e9) t38/49)

(‘Vza)

(V33)

(“‘/se)

( 'E/34) (39/e.3)

abnormal test: a lower than normal maximal heart rate does not affect sensitivity. However, some false negatives may be due to poor cooperation resulting in an inadequate heart rate response. It can be difficult to separate patients with chronotropic incompetence who should not be urged beyond their physiologic limits from patients who are uncooperative or deconditioned. Beta blocking agents decrease maximal heart rate and reduce the sensitivity of the test. )

SAMUELM. Fox, III: In assessing whether the patient is at the point of significant or “near maximal” stress, circumoral pallor and a poor rate of postcompression refill of erythematous areas on the trunk are both of assistance. These signs usually do not occur until a patient is up to 85% of maximum. The systolic Korotkoff tcnes can also give useful information when a dark complexion makes pallor difficult to evaluate:

Rejecting the test results of patients not achieving a target heart rate in the absence of ST depression as uninterpretable increases the sensitivity of the test by 15-20%. However, this would render the test useless for evaluating 20-30% of the referred patients. Realization of the importance for achieving an adequate heart rate should decrease this problem. Many patients who stop during an exercise test because of anxiety or fear can be encouraged to perform a greater effort safely by an interested clinician. Patients may also be safely exercised beyond the onset of mild angina and this may increase the yield of abnormal responders. ) SAMUELM. Fox, III: As with many types of clinical examinations, it may be worthwhile to have the patient repeat the test on another day. Although this may be economically unfeasible, there are some circumstances that can only be defined by a subsequent test. Examples include patients with early dysrhythmia at low exertional levels, those on beta blockers and those who have a rate-dependent left bundle branch block. A repeat test can be very important to these patients and may actually save money and further examinations in the long run. Thallium nuclide myocardial “perfusion” scanning has been of assistance in some cases, but its place in the diagnostic approach has not been adequately established.

Sensitivity and specificity are important for determining the value of a diagnostic test, but the information most important to the clinician is the likelihood of the patient having disease once his test result is known. Bayes’s theorem stresses that such a likelihood cannot be accurately estimated from the test result and the diagnostic characteristics of the test alone, but also requires a knowledge of the prior probability of the patient having the disease before the test was administered. Bayes’s theorem says that the odds of a patient having the disease after the test 24

will be the product of the odds before the test and the odds that the test result was a true one. The odds of a test result being true is known as the likelihood ratio (8). It is the ratio of true results to false results. In the case of an abnormal test result: et=

% with disease with abnormal test % without disease with abnormal test

Or

sensitivity 1 - specificity

In the case of a normal test result: 8- = % without disease with normal test % with disease with normal test Bayes’s theorem

specificity Or 1 - sensitivity

may be expressed in the following

fashion:

post-test odds _ pre-test odds X likelihood ratio of disease of disease of test result The clinician often makes such calculations intuitively as when he suspects as a false result the abnormal exercise test of a 30-year-old woman with stabbing chest pain (low prior odds). The same abnormal response would be accepted as a true result in a 60-year-old man with angina and a previous myocardial infarction (high prior odds). The likelihood ratio is an indicator of the diagnosticity of a test; the higher it is, the greater the diagnostic impact of the test. Using conventional techniques of ST segment analysis, the multistage exercise test has a sensitivity of 60- 80% and a specificity of about 90% (see Table 8). Using 70% as an average figure for sensitivity, the likelihood ratio for an abnormal test result equals: (e+) = For a normal

0.70 = 7.0 1 - 0.90

test result, the likelihood (e-) =

ratio equals:

0.90 = 3.0 1 - 0.70

In most clinical situations, effective use of Bayes’s theorem is impaired by insufficient data for the estimation of prior probability. However, in the case of coronary artery disease, good data exist. A number of studies have investigated the prevalence of significant angiographic disease with different chest pain syndromes (Table 9). Since chest pain is the presenting complaint in the majority of patients on whom a diagnostic exercise test is performed, the nature of the pain would seem a practical basis by which to estimate prior probability. 25

TABLE 9. -PERCENTAGE OF MIDDLE-AGED SIGNIFICANT ANGIOGRAPHIC CORONARY ACCORDING TO CHEST PAIN

Proudfit Friesinger McConahay Campeau Average

et a1.114

790

8% P%so)

and

251 100 131

92% (99/S?) 96% W59) 8% (44/55) 9%

Smith”5

et al.“~ et al.“’

PATIENTS ARTERY HISTORY.

58% 66% 42% 42% 52%

P/MI) (Vr1) P/19)

P?‘zd

WITH DISEASE

10% 16% 12% 4% 10%

wz6s) (‘8/m) p/25) (%o)

Chest pain syndromes can be classified into three categories by the following definitions. 1. ‘Typical angina poctoris is a chest pain or discomfort: (a) usually involving part of the sternum; (b) usually precipitated by physical exertion or other stresses; (c) relieved within lo- 15 minutes after cessation of stress; and (d) relieved within several minutes by nitroglycerin. 2. Atypical angina pectoris is a pain meeting at least two of the above criteria but having some of the following atypical features: (a) unusual location; (b) prolonged duration; (c) inconstant precipitating factors; and (d) failure to respond to nitroglycerin. 3. ~onanginal chest pain is a chest pain having none of the above characteristics. Table 9 demonstrates a remarkable agreement among different studies which have investigated the coronary angiographic anatomy in middle-aged patients with these clinical SYndromes. The patient with typical angina has apProxima~lY a 96% chance of having significant coronary artery disease, but a 10% chance of being free of it. The patient with atypical angma has about a 59% chance of having disease and the patient with chest pain thought unlikely to be ischemic has about a 10% chance of havin sigrnficant disease. In comparison, Gensini and Kelly found a f ~5% prevalence of significant angiographic disease among 278 patients undergoing cardiac catheterization but not clinically suspected of cmwnary artery disease.“* THE

PATIENT

WITH TYPICAL

ANGINA

PECTORIS

The patient with typical angina has a 96% chance or a 9: 1 chance of having significant coronary artery disease. An abnormal test result with a likelihood ratio (8 +) of 7 will increase the odds from 9: I to approximately 63 : 1. Such an impressive change in odds represents a relatively small increase in the probability of disease from 96% to $X3%, a gain of questionable practical value. (The probability is calculated from the odds ratio; i.e., 63 divided by 63 + 1 equals 98%.) The major purpose for 26

s

performing an exercise test on patients with typical angina from a diagnostic point of view would be to identify the 10% falsely labeled as having ischemic heart disease. Under optimal conditions (normal resting ST segment, adequate maximal heart rate) with a likelihood ratio (8 -) of 3, the odds of disease would decrease from 9: 1 to 9: 3, but the probability of disease after a negative test result would still be 75%, (i.e., 9 divided by 9 + 3.) Thus, a coronary angiogram would be required to definitely rule out coronary artery disease. If the purpose for performing the exercise test is to achieve a more accurate diagnosis in the patient with typical angina pectoris, only certain patients should be tested. The greatest diagnostic impact will be on patients with normal resting ST segments and those likely to cooperate in reaching an adequate maximal heart rate. The patient should be one in whom it would be reasonable to perform a coronary angiogram in the event of a normal test to confirm that no coronary disease is present. However, the finding of chronotropic incompetence, hypotension, dysrhythmias, poor functional capacity or marked mation for patient

ST segment management.

changes

may be valuable

infor-

b SAMUEL M. Fox, III: The figures presented suggest that the diagnostic value of an abnormal exercise test in a patient with typical angina is questionable. However, this refers only to a tight mathematical matrix without the clinically important information of when the abnormality occurs, and with what relation to the patient’s activity needs. We are firmly convinced that all typical angina1 patients should have a progressive multistage exercise test to help determine the pattern and duration characteristics of any associated ischemia as well as the hemodynamic, dysrhythmic and heart sound findings. Exercise tolerance testing provides far more information than a simplistic “yes” or “no” relative to more than 0.10 mv of additional ST displacement. Angiograms, exerciseinduced electrocardiographic, hemodynamic and rhythm responses and the recently applied nuclide myocardial perfusion scanning and ventriculographic techniques are more complementary than competitive in providing clinically useful information. When a “normal” exercise response occurs after previous “typical” angina pectoris, another reasonable approach is to encourage physical reconditioning of the patient and retest the person 3 - 6 months later. This life-style improvement, combined with the probability of higher performance in a retest situation, often produces results that can be a source of encouragement for all concerned. This approach should cost less and have an acceptably low hazard as well. This is not presented in opposition to the use of angiograms to confirm the absence of significant coronary disease. It is very important that some individuals have definite clearance, including angios, even if it involves an additional low order of hazard and added expense. w FRANK I. MARCUS:One of the most vexing problems in the practice of cardiology is to determine which patients with known or suspected ischemit heart disease should undergo cardiac catheterization and coronary arteriography. The report of Margolis (Circulation 51 and 52 27

[Suppl. 111:109,1975), if verified, is important, since it indicates that exercise testing should help in the selection of patients with coronary artery disease who should have coronary arteriography and possible coronary bypass. They observed that patients who achieved Bruce stage 4 had an excellent prognosis as compared to those who could not complete stage 1, indicating a poor prognosis. Since the patient with ischemic heart disease who achieves stage 4 of the Bruce protocol has an excellent prognosis, why perform coronary arteriography to define the extent of the coronary artery disease? A similar observation was made by Ellestad and Wan (Circulation 51:363, 1975). They found that patients with known or suspected coronary disease who developed &hernia at an exercise level of 4 METS had an annual coronary heart disease event rate of 15%. When &hernia appeared at about 8 METS, the annual incidence was 4%. Cheitlin et al. (Am. Heart J. 89305, 1975) reported that the great majority of patients with triple-vessel or left main disease may be identified by the amount of ST segment depression with exercise. They stated that the chance of missing significant triple-vessel or left main coronary artery disease was less than 6% in the presence of an adequate treadmill test that showed less than 1 mm of ST segment depression. As mentioned in this review, patients with ischemic heart disease who develop ST segment depression early in the course of exercise testing or have this finding to a depth of > 2 mm, and those who have persistent ST segment depression after exercise for longer than 8 minutes during recovery, have a high prevalence of triple-vessel left main disease and should be considered for arteriography if clinically indicated. It is not known whether there is an increased risk of exercising patients with “unstable angina” defined as patients who have new onset of angina or who have a crescendo pattern, both associated with ST segment depression at rest in the presence of pain. These patients should not be exercised if they continue to have ischemic pain at rest, especially if the pain persists after bedrest and treatment with nitrites and propranolol. However, the patient whose pain promptly subsides after rest and medical therapy can probably be exercised safely after several days, providing an acute myocardial infarction has been excluded. Since the cooperative multicenter trials have not documented that bypass grafting results in a decrease in morbidity or mortality in unstable angina, exercise testing may help determine which patients should undergo coronary arteriography to identify those who have left main or triple-vessel disease. I should again emphasize that I do not recommend exercise stress testing in patients with unstable angina who continue to have pain at rest. THE

PATIENT

WITH

ATYPICAL

ANGINA

PECTORIS

The patient with atypical angina has approximately chance of having coronary artery disease. An abnormal sult (likelihood ratio [0 +I of 7) would increase the odds to 7: 1, the probability of disease to 88% and for practical es establish the diagnosis. The importance of a normal sult would be dependent upon the heart rate achieved resting ST segments, but with a likelihood ratio (0 -) 28

a 50% test refrom 1: 1 purpostest reand the of 3, the

odds would be reduced cal angina by exercise

decrease to 1: 3 and the probability of disease would to approximately 25%. Thus, the patient with atypiwould be an ideal candidate for diagnostic evaluation testing.

b FRANK I. MARCUS:The use of 201thallium, injected at peak exercise, is still being assessed to determine its place as an adjunct to evaluation of the electrocardiogram in the patient with suspected coronary disease. Ritchie et al. (Circulation 56:66, 1977) have reported an increase in the sensitivity of detecting significant coronary disease from 65% to 91% by the use of this isotopic technique. The experience at the University of Arizona confirms that there is an increase in sensitivity with the use of 20’thallium, but not to the same extent as reported by Ritchie. Disadvantages of the use of 201thallium are its high cost (about $250 for the complete isotopic procedure) and the relative unavailability and short halflife of the isotope. False negatives have been observed in patients with single-vessel disease and in those with triple-vessel disease. In the latter case, it is assumed that global ischemia prevents selective decrease in mvocardial nerfusion.

THE

PATIENT

WITH

NONANGINAL

CHEST

PAIN

The middle-aged male patient with nonanginal chest pain has a 10% chance of having coronary artery disease. This is slightly higher than an asymptomatic individual. An abnormal test will occur in about 10-E% of such patients, but when it occurs, it will increase the odds from 1: 9 to 7: 9. The probability of disease will rise from 10% to 44%, thereby identifying patients in whom coronary angiography might be indicated. An alternate approach would be to offer these patients risk factor modification and medical follow-up. A negative test would reduce the probability from 10% to 4%, which is not a very important change. Nevertheless, the experience of exercising and being told that the exercise test was normal is very reassuring and may be quite valuable in convincing the patient with nonanginal chest pain that his symptoms are not due to cardiac disease. THE PATIENT CHEST PAIN

WITH A PRIOR MYOCARDIAL

INFARCTION

AND

The patient with a myocardial infarction has an extremely high prior probability of having coronary artery disease. The clinical problem in such patients often is whether or not their chest pain is cardiac in origin. The exercise test has a special value in such patients, in contrast to coronary angiography, which will be abnormal regardless of the cause of the pain. As in the case of the patient without a history of myocardial infarction, the prior probability of the pain being ischemic should be 29

STUDY

IO. -RESULTS

11.3 16 10 12.8 12.5 -

-

3.3 4.7

6.3 3

2.3

9

INCIDENCE OF CORoNAnY DIBEAsB (%I

5 5

56WITHABNORMAL mmw-UP EXERCISE TEST PEQIOD (YE)

10 13

13.6 13.6 14.3

60 67 61 58 62

90 91

92

91 92

46 20 25 26

test.

13.6

PREDICTIVE VALUEcall

ASYMPTOMATIC

SPECIFIcmY (8)

SCREENING

BENBIT*YITY :%I

STUDIES DISEASE

RELATlVE RlSKcQ*

OF EXERCISE TESTING IN FOUR PROSPECTIVE MEN FOR LATENT CORONARY ARTERY

“Relative risk = relative chance of developing coronary disease with an abnormal test compared to those with a normal tPredictive value = percentage of abnormal responders who developed coronary artery disease over the follow-up period.

Bruce and McDonough”” Aronow and Cassidy73 Froelicher et al?* Cumming et ~1.‘~~ Average

TABLE

established by the chest pain history. The exercise test result will be an additional indicator of the probability that the chest pain is ischemic. Observation for chest pain during the exercise test may be of special value. EXERCISE TESTING A~YMPWMATIC PERSONS

Exercise testing has been used as an epidemiologic screening tool as well as a detector of latent coronary artery disease. From available data, it appears that the probability of disease in an asymptomatic middle-aged person is 5- 8%. According to Bayes’s theorem, a normal exercise test would only reduce the probability to l-2%, whereas an abnormal test would increase the odds from 1: 14 to 7: 14 and the probability of disease to approximately 33%. These calculations are in remarkable agreement with the reported predictive value of an abnormal exercise test among asymptomatic men (Table 10). The predictive value is equal to the probability of an abnormal responder having coronary artery disease. Thus, exercise testing among asymptomatic men identifies a subgroup of approximately 12% with an increased probability and risk for disease. Risk factor modification and medical follow-up would be recommended for these patients. Coronary angiography would usually be reserved for those in whom sudden incapacitation could pose a public safety problem. Chronotropic incompetence, relative hypotension, ST segment elevation or marked or prolonged depression, or chest pain precipitated by the exercise test are very unusual in a healthy population. Such findings, when encountered, should be considered with individualized clinical judgment. SAMUEL M. FOX, III: Table 10 presents a number of clinically important points. The “relative risk,” or risk ratio, is a strong arrow in the quiver of facts to present to the patient in support of life-style change. The “predictive value” is directly time-related and all these studies were of relatively short duration. However, a 26% chance of developing a manifestation of coronary artery disease within 3 - 6 years should be a strong motivator for life-style change with or without coronary surgery. b

Table 11 presents the probabilities for coronary artery disease given the various clinical presentations and responses to exercise testing as discussed in this section. The probability for disease after an abnormal test response is equivalent to the predictive value of an abnormal test response. Another way of stating Bayes’s theorem is that the predictive value of an abnormal response is directly related to the prevalence of disease in the population tested. 31

TABLE 11. -APPLICATION OF BAYES’S THEOREM IN ORDER TO ASSESS LIKELIHOOD OF CORONARY ARTERY DISEASE ACCORDING TO CHEST PAIN SYNDROMES AND RESPONSE TO EXERCISE TESTING APmoxIY*TE

PROB*BILITY ARTERY

FREsmwlNG 8YMpToMS

Typical angina Atypical angina Nonanginal chest Asymptomatic

SPECIFIC

pain

B&FOB& TEST

AmERAmmEMN. EXErlcI(IB

90 50 10 5

98 88 44 27

P%)

OP

HAVINO

coRc%**RY

DI8EAm

TEST

APTEB No%MAL EXERCISE TEST

75 25 4 2

PROBLEMS

WOMEN

The diagnostic value of exercise testing is not as well-defined in women as it is in men. Three types of studies are reviewed (1) four studies screening asymptomatic, healthy women, (2) two studies of women with chest pain and normal coronary angiography and (3) one study using the same exercise testing and angiographic techniques to evaluate both men and women presenting with chest pain. Table 12 lists four studies which screened a total of 761 healthy, asymptomatic women with exercise testing. The total percentage of abnormal responders was approximately 30%. The frequency of abnormal tests was agedependent, as in men, but the frequency was 2-3 times higher than when screening asymptomatic men of comparable ages in spite of a much higher disease prevalence in men. Follow-up data are not available and so the relative risk of an abnormal response in women is not known. However, the specificity of exercise testing asymptomatic women must obviously be lower than in men. A group of patients, mainly younger women who have angina but normal coronary angiograms, has been described. In two series of such patients approximately half of them had exercise TABLE 12. -RESULTS OF FOUR STUDIES SCREENING ASYMPTOMATIC WOMEN FOR LATENT CORONARY ARTERY DISEASE WITH EXERCISE TESTING

Lepeshkin Astrand’= Profant Cumming 32

and

Surawia’*’

et a1.lz3 et a1.1z4

143 117 144 357

41 50 45 44

24 35 33 26

TABLE 13.-RESULTS OF SKETCH ET AL.. EVALUATING MEN AND WOMEN PRESENTING WITH CHEST PAIN USING CORONARY ANGIOGRAPHY AND EXERCISE TESTING12’

Women (No.= 56) Men (No. = 195)

50

20

27

46

78

33

48

57

32

51

94

92

tests performed and 30-6m of those tested had abnormal responses.125* 126Both of these studies demonstrated a good prognosis for these patients. The results of Sketch et al. evaluating men and women presenting with chest pain using angiography and treadmill testing are presented in Table 13. lz7 The finding that the women presenting with chest pain had a lower prevalence of significant angiographic disease than the men is not explained by age difference. The women had nearly the same prevalence of abnormal treadmill tests in spite of a lower prevalence of coronary artery disease. Sensitivity was relatively low in both sexes, possibly because a submaximal test was used and only the postexercise ECG was analyzed. In the men, the specificity was comparable to other studies but specificity was lower in the women. The predictive value of an abnormal test (percentage of abnormal responders with disease) was much lower in the women; conversely, the false positive rate was higher. This probably was due both to the lower prevalence of disease and to the lower specificity of exercise testing in women. In summary, exercise testing in women has a lower specificity than it does in men. However, much less exercise testing has been reported in women and often the reported studies are complicated by concomitant drug treatment, including estrogen, or abnormal resting ECGs. The explanation for the increased number of false positives in women is not known. Kusumi et al. have presented data suggesting that it may be due to their blood pressure response during exercise. 12* It is likely that the physiologic response to exercise in women is different from that in men and that sex-specific ST segment criteria will be necessary in order to obtain valid information when exercise testing women. w ROBERT A. O'ROURKE: A large number of women undergoing treadmill exercise testing do so because of atypical chest pain that frequently is associated with anxiety. Many of these patients are receiving drugs which affect the central nervous system and may be associated with ST-T wave changes on the resting or exercise ECG. Many of these patients are

receiving

diaxepam

(Valium)

or phenothiazine

derivatives.

Another

sub33

set of these women undergoing treadmill testing are patients who are shown to have mitral valve prolapse by subsequent echocardiographic or left ventricular angiographic studies. These patients may also have ST-T wave changes at rest and frequently develop ST segment depression during exercise despite no angiographic evidence of large vessel coronary artery disease. When women receiving medications affecting the central nervous system and those with mitral valve prolapse are excluded, the sensitivity of ST segment depression during treadmill exercise testing is more similar to that occurring in men (Linhart et aL: Circulation 50: 1173,1974). EXERCISE TESTING WITH RESTING ST SEGMENT DEPREMON Although patients with ST segment depression at rest are at increased risk for coronary artery disease, the exercise ECG can still give additional useful information. However, the interpretation of the exercise ECG in these patients is difficult, since criteria for exercise-induced ST changes have not been established. Should the classic O.l-mv depression be added to the resting ST segment depression, to the PR isoelectric line or should more depression be required to classify the response as abnormal?

Kansal et al. have reported 37 patients with chest pain and at least 0.05 mv ST depression at rest who were studied with treadmill testing and coronary angiography.12g None of their patients had ECG left ventricular

hypertrophy,

intraventricular

conduc-

tion defects, electrolyte abnormalities or were on digitalis. Using 0.1 mv additional ST segment depression as an abnormal response to the exercise test, they demonstrated a 92% sensitivity and a 75% specificity for angiographic coronary artery disease. The degree of ST depression did not help improve differentiation

or relate

to the severity

of disease. This sensitivity

is

higher and the specificity lower than when using exercise testing in those with normal resting EC&. This is in agreement with the high prior probability that these patients have coronary artery disease and demonstrates that exercise testing yields more false positives in patients with resting ST segment depression. Simoons et al. have presented a spatial analysis of the ECG changes during exercise in patients with abnormal resting ECGs.33 The spatial orientation of the exercise-induced ST changes was correlated with left ventricular dyskinetic areas as documented by left ventriculography. These patients were compared with a reference group of normal subjects and with patients with angiographic disease but with normal resting ECGs. In the patients with a normal ECG at rest, but with coronary artery disease, horizontal ST segment depression occurred in leads X and Y and some posterior displacement was found in the ST segment in lead Z. Previous studies have disagreed whether 34

the normal exercise-induced ST segment vector shift is anterior or posterior. A posterior shift may be very specific for ischemia when the resting ECG is normal and an anterior ST shift may be very specific for ischemia when the resting ECG shows loss of anterior forces. The direction of the ST changes in patients with coronary artery disease with a normal ECG at rest was the same as in healthy subjects. This might indicate that subendocardial ischemia is present during exercise in normal subjects and that it directs itself along the long axis of the heart. The ST segment changes during exercise in patients with an abnormal ECG at rest are partially dependent on the amount of myocardial scar tissue present and partially on the location of the scar tissue. Since the direction of the ST segment changes varied widely in patients with an abnormal ECG at rest, the magnitude, but not the direction, of these changes should be used as an indication of exercise-induced ischemia. Until spatial magnitude criteria are derived, the exercise test will have the specificity limitations demonstrated by the study of Kansal et al. in patients with resting ST segment depression.lzg In screening asymptomatic men, an abnormal exercise test (using 0.1 mv or more horizontal or downsloping ST segment depression from the PR segment) was found to have a higher predictive value for angiographic coronary artery disease in those with minor ST segment depression at rest than in those with normal ST segments.4o In a study using computerized ST segment analysis, it was found that those with abnormal tests had nearly isoelectric ST segments at rest while those with normal tests had mild early repolarization.90 These findings suggest that consideration of even subtle measurements of the resting ST segment will have importance in analyzing the exercise test. FALSE POSITIVES

Table 14 lists some of the conditions that can possibly result in exercise-induced ST segment depression not due to coronary artery disease. Simonson has suggested that in populations with a high prevalence of another type of heart disease, an abnormal exercise test would be as diagnostic for that disease as it would be for coronary artery disease in populations with a high prevalence of coronary artery disease.’ An abnormal response in patients with the first four cardiac disorders in Table 13 could be a false positive. Digitalis and other drugs can cause exercise-induced repolarization abnormalities in normal individuals. In a study of 23 healthy young men, 33% developed abnormal exercise-induced ST segment depression after the administration of digoxin. 130 Patients with anemia, electrolyte abnormalities or 35

TABLE 14.-CONDITIONS INDUCED ABNORMAL PRESENCE (FALSE Valvular Congenital

ST

THAT CAN SEGMENT

RESULT RESPONSE

IN

AN EXERCISEWITHOUT THE

OF CORONARY ARTERY DISEASE POSITIVE EXERCISE TEST) Left ventricular hypertrophy Wolff-Parkinson-White syndrome Iown-Ganong-Levine syndrome Pre-excitation variants Mitral valve prolapse syndrome

heart disease heart disease

Cardiomyopathies Pericardial disorders JWP Electrolyte abnormalities Nonfasting state Anemia Sudden excessive exercise Inadequate recording equipme Bundle branch block Improper interpretation

Vasoregulatory abnormality Hyperventilation abnormalities

nt

asthenia or repoiarization

Hypertension Excessive double product Improper lead systems Incorrect criteria

who are on medications who have abnormal responses should be retested when these conditions are eliminated. Since meals and glucose ingestion can alter the ST segment and T wave in the resting and exercise ECG, failure to fast prior to an exercise test can cause a false positive response. To avoid this problem, all ECG studies should be performed after at least a 4-hr fast. b SAMUELM. FOX, III: We have modified

our pre-exercise

restriction

on

eating and encouraged our subjects to have unsweetened orange juice, cereal with minimal sugar and low fat milk or a clear soup and salad type lunch prior to coming to the laboratory. By the time of the test 1Y2 - 2 hours have usually gone by and we feel there is more gamed than lost by providing some “fuel in the furnace.” Large, heavily sugared meals should be avoided as well as tobacco, fats, stimulating beverages and alcohol. It has been assumed that ST segment changes induced by exercise could be used to diagnose coronary artery disease when the baseline ECG showed bundle branch block. Whinnery et al. reported 31 asymptomatic men with acquired left bundle branch block who were studied with both treadmill testing and coronary angiography.131 They demonstrated that there can be a marked amount of exercise-induced ST segment depression in addition to that present at rest in healthy men with left bundle branch block. No difference was found between the ST segment response to exercise in those with and in those without significant coronary artery disease. The ST segment response to exercise testing cannot be used to make diagnostic decisions about patients with left bundle branch block. In a second study, Whinnery et ~1. reported the response to maximal treadmill testing of 40 asymptomatic men with acquired right bundle branch block.‘“2 There 36

was no exercise-induced ST segment depression in an inferior or a lateral lead in either those with or those without significant disease. Subsequently, we have seen two patients with symptomatic coronary artery disease and right bundle branch block who developed exercise-induced ST segment depression in inferior and lateral leads. Friedman and Marcus reported exerciseinduced ST segment depression in anterior precordial leads in patients with right bundle branch block.133 This is most apparent in V leads with an rSR’ or a notched R wave which show a downward sloping ST segment at rest. This finding probably is not indicative of myocardial ischemia, in contrast to the depression in the inferior or lateral leads. Individuals with left ventricular hypertrophy and strain pattern on the resting ECG are at high risk for coronary artery disease. Harris et al. have reported that left ventricular hypertrophy in and of itself may cause an abnormal exercise test.134 Healthy individuals with the Wolff-Parkinson-White pattern can have exercise-induced ST segment depression. Some individuals with pre-excitation with a short PR interval and a normal QRS complex may have a false positive exercise test.ls5 Patients with the prolapsing mitral valve syndrome have been reported to have abnormal exercise tests but normal coronary angiograms.136 Most aircrewmen with this syndrome in the United States Air Force are asymptomatic and have normal exercise tests. ) SAMUELM. Fox, III: For patients with bundle branch block, hypertrophy and strain, pre-excitation syndromes (both Wolff-Parkinson-White and Lawn-Ganong-Levine syndromes) and those with vasoregulatory asthenia, radionuclide studies have great potential use. This is particularly true in the areas of gated blood pool analyses of wall motion and rate of ejection and relaxation. While patients with “click-murmur” mitral prolapse also may be evaluated with benefit, the degree of mitral regurgitation mav make wall motion studies hard to evaluate.

Individuals with vasoregulatory asthenia and orthostatic or vasoregulatory abnormalities can have abnormal exercise-induced ST segment changes without coronary artery disease.“” The same can be said for those with hyperventilation-induced repolarization abnormalities. 56 However, these conditions can occur in the presence of coronary artery disease. It is a good procedure to screen all patients for orthostatic and hyperventilation repolarization changes prior to treadmill testing or at least to use these maneuvers only on patients who have an abnormal response. Such changes are unusual in our experience and have rarely been responsible for false positive exercise tests. Hyperventilation may also induce chest pain and/or PVCs in patients with coronary artery disease, so it may be appropriate to have a physician

in attendance

during this maneuver. 37

Hypothetically, individuals with hypertension or excessive blood pressure-heart rate products during exercise could have a physiologic imbalance between myocardial oxygen surjply and demand. Barnard et al. have demonstrated that sudden high work loads of treadmill exercise can yield ST segment depression in healthy individuals on this basis.13’ A recorder with an inadequate frequency response can either induce ST segment’ depression in normals or show upsloping depression when horizontal depression is actually present. Use of the proper equipment should avoid this type of distortion. In conclusion, the conditions discussed above can be avoided and should not be the major causes of false positive responses in a good exercise testing laboratory. The most common cause of a false positive test should be the normal variant who has a physiologic ST segment vector that is similar to that produced by ischemia. EXERCISE TESTING AFTER AN ACUTE MY~CARDIAL INFARCTION

Exercise testing can be valuable in evaluating patients after an acute myocardial infarction by identifying high risk patients who may benefit from treatment and by determining what activities a patient can safely perform. It is advisable to exercise a patient under monitored conditions rather than to allow him to go home to activities without any knowledge of how his cardiovascular system will respond. One of the authors (G. K.). has studied 60 patients with treadmill testing and ambulatory monitoring at 3 and 8 weeks after their myocardial infarction. These patients were New York Heart Association Functional Class I or II and were not on digitalis or any antiarrhythmic agents. At the 3d week after their infarct, the treadmill protocol consisted of three progressive 3min stages. The speed was constant at about 2 mph with grades of 0, 5 and 10%. The approximate oxygen cost of each of these stages is 9, 12 and 16 cc O,/kg/min. All but 10 patients completed this 9-min test without difficulty. No serious dysrhythmias were encountered, two patients were stopped by angina and five by fatigue. At the 8th week postinfarction, a maximal treadmill test was performed. The majority of patients stopped either because of leg fatigue or shortness of breath. Five patients were stopped by angina and two by ST segment depression. No serious complications were encountered. The approximate mean maximal pressure rate product during exercise was 17 X 103 at 3 weeks and 27 X 103 at 8 weeks. A shown in Table 15, the mean maximal heart rate at 3 weeks was 116 beats/min and at 8 weeks it was 157 beatslmin. Angina with or without ST segment depression was an infrequent response in both the 3d (18%) and 8th week (28%). Exercise-induced ST segment depression developed in 13% of the patients during the 3d-week test 38

and in 32% during the &h-week test. The exercise test during the 3d week allowed for assessment prior to hospital discharge and the test during the 8th week assessed the patient’s ability to return to work and helped formulate an exercise prescription. Because exercise testing measures a patient’s exercise capacity, evaluates psychologic reaction to exercise and gives an objective assessment of dysrhythmias and ischemia, it is particularly helpful in the management of postinfarction patients. w FRANK 1. MARCUS:Modified exercise stress testing at an average of 18 days after infarction has been performed rather routinely at the University of Arizona Medical Center since 1972, and our experience was recently reported (Smith et al.: Circulation 54 [Suppl. IIl:lO, 1976). We concluded that the test was safe and was helpful in the management of postinfarction patients. A recent report from Sweden (Granath et cd.: Br. Heart J. 39758, 1977) provides data that this test, done within 3 weeks after an acute myocardial infarction, also provides prognostic information. Patients with angina pectoris and those who had substantially increased heart rates during exercise or major ventricular arrhythmias with exercise, had a significantly increased risk of death during the subsequent 2 - 5 years. Patients with ventricular arrhythmias who responded to antiarrhythmic therapy appear to have a lesser risk than those who did not. These data still need to be confirmed utilizing a larger number of subjects.

Table 15 summarizes four studies performed at 3 weeks after an acute myocardial infarction enabling a comparison of exercise testing and ambulatory monitoring for detecting dysrhythmias. Antiarrhythmic agents and digitalis often were administered to the patients in the other series. No relationship was noted to either angina or ST segment depression and PVC prevalence. Only 30-40% of the postinfarction patients were found to have PVCs during their exercise tests while ambulatory monitoring detected PVCs in approximately 75% of them. A much higher prevalence of ominous PVCs were also noted with ambulatory monitoring. In the one study with the procedures repeated at 8 weeks after infarction, the dysrhythmias were the same in type and prevalence as at 3 weeks and were usually reproducible. These data support the conclusion that exercise testing will underestimate the prevalence of PVCs, particularly ominous PVCs, in patients 3 weeks postinfarction. The finding of ominous PVCs at this time has been a risk marker for morbidity and mortality in these patients, particularly in the first few months where the highest mortality occur~.‘~~~lo3 Treadmill testing after an acute myocardial infarction should be used to evaluate the functional capacity and possible presence of ischemia while ambulatory ECG monitoring is more sensitive for dysrhythmia detection. 39

11 32 10

20

NA

17 NA NA

42

8

7

33

0.5

49

24

3 wk after AMI 60 Maximal bicycle 209 129

No. of subjects Mean maximal heart rate % with ischemic response 00.1 mv) % with exerciseinduced angina % with (2 PVCs/min or <30 PVC&r % with 22 PVCs/min or 230 PVCs/hr 8 with multiform PVCS % with couplets or bigeminy % with 13 consecutive PVCS Total % with PVCs

to 120

AM1

62 Submaximal treadmill beatslmin 100 118

after

EVALUATING PATIENTS WITH EXERCISE TESTING

Mean age (yr) Testing procedure

tAMI)

3 wk

15.-STUDIES

Population

TABLE

29

0

5

0

2

21

18

13

116

Submaximal treadmill to 1.7 mph, 10% grade

3 wk after

AM1

79

7

14

34

9

48

60

54 24-hr AECG

8 wk

25

0

4

2

5

14

28

32

157

after

MYOCARDIAL (AECG)

Maximal treadmill

WITHIN 3-8 WEEKS AFTER ACUTE AND/OR AMBULATORY MONITORING

81

4

25

33

9

44

24-hr AECG

AMI

9

72

4

18

20

NA

100

3 wk after AMI 57 6-hr AECG

INFARCTION

w SAMUEL M. FOX, III: We thoroughly endorse the usefulness of predischarge exercise evaluation of the stable and relatively uncomplicated postinfarct patient. We have not usually exceeded a heart rate of 12O/min or 4 METS (14 cc 0, min/kg body weight) using 2.0 mph and 7% grade as our peak exertional level. Approximately a third of the patients present evidence justifying a lower than standard level of postdischarge activities. The majority, however, show no adverse reactions and can be permitted slightly greater freedom. Unless there is unusual persuasion to explore higher levels, we do not care to go above a 140/minute heart rate earlier than 3 months postinfar& We have found no data refuting the 1939 information of Kenneth Mallory and Paul D. White indicating that the scar may not be healed fully until 3 months postinfarct. Unless unusual circumstances provide rare justification, we do not wish to counter the contraction of the scar by imposing even brief high tension stresses beyond those that help restore personal confidence and avoid deconditionine: comnlications

SUMMARY Exercise testing is supported by the available data as a valuable tool in the modern practice of medicine. Although the major application of exercise testing in clinical practice is the evaluation of the patient with chest pain, other important applications have also been reviewed. The section on exercise physiology and the methodology of exercise testing should facilitate the safe and proper use of the test. Proper interpretation of the exercise test requires that each of its results be considered separately, i.e., functional capacity, ST segment changes, dysrhythmias, heart rate, blood pressure, physical findings and symptomatology. A basic understanding of statistics, likelihood ratios and the determination of the probability of disease are helpful in understanding the diagnostic implications of the results of exercise testing. An understanding of Bayesian statistics and the predictive model explain why exercise testing yields different results in different populations. While the sensitivity and the specificity of ST segment depression have been demonstrated by numerous studies and are approximately 70 and 90%, respectively, further studies are needed to clarify the significance of the other exercise test responses. ST segment elevation is a rare but highly specific finding for coronary artery disease that has not been adequately appreciated. There are numerous problem areas in exercise testing, but only those with new information available were reviewed. We feel that the quality and quantity of the studies evaluating exercise testing have established it as one of the safest and most valuable procedures for evaluating the cardiovascular system. As studies demonstrate how to optimize and standardize the methodology and interpretation, the value of the exercise 41

test will increase even further. Hopefully, this review both will help the practicing clinician apply the test clinically in an up-todate manner as well as direct the clinical researcher to areas for future investigation. ACKNOWLEDGMENT This study could not have Lou Richardson, Marilyn always, Julie. T. J. Reeves and catalysts, and the USAF

been done without the help of Mary Stewart, Rosa Rodriguez, and, as and L. T. Sheffield were teachers provided the milieu.

REFERENCES”’ 77. Bruce, R. A.: Exercise testing for evaluation of ventricular function, N. Engl. J. Med. 296:671,1977. 78. Thompson, A. J., and Froelicher, V. F.: Kugel’s artery as a major collateral channel in severe coronary disease, Aerosp. Med. 45:1276, 1974. 79. Hinkle, L. E., Carver, S. T., and Stevens, M.: The frequency of asymptomatic disturbances of cardiac rhythm and conduction in middleaged men, Am. J. Cardiol. 24:629,1969. 80. Bruce, R. A., Gey, G. O., Cooper, M. N., Fisher, L. D., and Peterson, D. R.: Seattle Heart Watch: Initial clinical, circulatory, and electrocardiographic responses to maximal exercise, Am. J. Cardiol. 33:459,1974. 81. Thomson, P. D., and Kelemen, M. H.: Hypotension accompanying the onset of exertional angina, Circulation 52:28,1975. 82. Sirnoons, M. L., and Hugenholtz, P. G.: Gradual changes of ECG waveform during and after exercise in normal subjects, Circulation 52:570,1975. 83. Fortuin, N. J., and Friesinger, G. C.: Exercise-induced ST segment elevation, Am. J. Med. 49459.1970. 84. Bobba, PI, Vecchio, C., Di Guglielmo, L., Salerno, J., Kasari, A., and Montemartine. C.: Exercise induced RS-T elevation. CardioloavVI 57162. 1972. 85. Hegge, F. N., Tuna, N., and Burchill, H. B.: Coronary arteriographic findings in patients with axis shifts or ST segment elevation on exercise stress testing, Am. Heart J. 86:603,1973. 86. Chahine, R. A., Raziner, A. E., and Ishimori, T.: The clinical significance of exercise induced ST segment elevation, Circulation 54:209,1976. 87. Manvi, K. N., and Ellestad, M. H.: Elevated ST segments with exercise in ventricular aneurysm, J. Electrocardiol. 5:317, 1972. 88. Noble. R. J.. Rothbaum. D. A.. Knoebel. S. B.. McHenrv. P. L.. and Anderson, G.’ J.: Normalization of abnormal T waves in i&hernia, ‘Arch. Intern. Med. 136:391,1976. 89. Stuart, R. J., and Ellestad, M. H.: Upsloping S-T segments in exercise stress testing, Am. J. Cardiol. 37:19, 1976. 90. Froelicher, V. F., Wolthuis, R., and Keiser, N.: A comparison of two bipolar exercise electrocardiographic leads to lead V,, Chest 70:530,1976. 91. Gutman, R. A., Alexander, E. R., Li, Y. B., Chiang, B. N., Watten, R. H., Ting, N., and Bruce, R. A.: Delay of ST depression after maximal exercise by walking for 2 minutes, Circulation 42:229,1970. 92. Froelicher, V. F., Thomas, M., Pillow, C., et al.: An epidemiological study of asymptomatic men screened with exercise testing for latent coronary heart disease, Am. J. Cardiol. 34:770,1974. 93. Blackburn, H., Taylor, H. L., Hamrell, B., Buskirk, E., Nicholas, W. C., and ‘:‘References 1 - 76 appeared in Part I (November 1977) 42

94. 95. 96. 97. 98. 99. 100. 101. 102. 103.

104. 105. 106. 107. 108.

109. 110. 111. 112. 113.

Thorsen, R. D.: Premature ventricular complexes induced by stress testing, Am. J. Cardiol. 31:441,1973. Faris, J. V., McHenry, P. L., and Jordan, J. W.: The prevalence and reproducibility of exercise-induced PVCs during maximal exercise in normal men., Am. J. Cardiol. 37:617, 1976. McHenry, P. L., Fisch, C., and Jordan, J. W.: Cardiac arrhythmias observed during maximal treadmill testing in clinically normal men, Am. J. Cardiol. 29331,1972. Jelinek, M. V., and Lown, B.: Exercise stress testing for exposure of cardiac arrhythmias, Prog. Cardiovasc. Dis. l&497,1974. Goldachlager, N., Cake, D., and Cohn, K.: Exercise-induced ventricular arrhythmias in patients with coronary artery disease, Am. J. Cardiol. 31: 434,1973. Ryan, M., I-own, B., and Horn, H.: Comparison of ventricular ectopic activity during 24 hour monitoring and exercise testing in patients with coronary artery disease, N. Engl. J. Med. 292:224,1975. Atkins, J. M., Matthews, 0. A., Blomquist, C. G., and Mullins, C. B.: Comparison of dysrhythmias with bicycle ergometry and isometric exercise, Br. Heart. J. 38:465,1976. Helfant, R. H., Pine, R., Kabde, V., and Banka, V.: Exercise-related PVCs in coronary heart disease, Ann. Intern. Med. 80:589,1974. Kennedy, H. L., and Underhill, S. J.: Frequent or complex ventricular ectopy in apparently healthy subjects, Am. J. Cardiol. 38:141,1976. Schulze, R. A., Strauss, H. W., and Pitt, B.: Sudden death in the year following myocardial infarction, Am. J. Med. 62:192, 1977. Moss, A. J., DeCamilla, J., Mietlowski, W., Greene, W. A., Goldstein, S., and Locksley, R.: Prognostic grading and significance of ventricular premature beats after recovery from myocardial infarction, Circulation 51 & 52 (Suppl. III): 204, 1975. Siegel, W.: A computer-assisted system for clinical use in maximal exercise testing, Cleve. Clin. Q. 43:143, 1976. Weiner, D. A., McCabe, C., Hueter, D., Hood, W. B., and Ryan, T.: The predictive value of chest pain as an indicator of coronary disease during exercise testing (Abstract), Circulation 53 & 54 (Suppl II):lO, 1976. Cole, J. P., and Elk&ad, M. H.: Correlation of chest pain during treadmill exercise electrocardiography and coronary events (Abstract), Circulation 53 & 54 (SupplII):206,1976. Froelicher, V. F.: Screening asymptomatic men for coronary disease, Ann. Rev. Med. 28:1,1977. Kassebaum, D. G., Sutherland, K. I., and Judkins, M. P.: A comparison of hypoxemia and exercise electrocardiography in coronary artery disease. Diagnostic precision of the methods correlated with coronary angiography, Am. Heart. J. 75:759, 1968. Martin, C. M., and McConahay, D. R.: Maximal treadmill exercise electrocardiography. Correlations with coronary arteriography and cardiac hemodynamica, Circulation 46:956,1972. McHenry, P. L., Phillips, J. F., and Knoebel, S. B.: Correlation of computerquantitated treadmill exercise electrocardiogram with arteriographic location of coronary arterv disease. Am. J. Cardiol. 36747.1972. Helfant, R. H., Sanka, V.“S., De Villa, M. A., et al.: Use of bicycle ergometry and sustained handgrip exercise in the diagnosis of presence and extent of coronary artery disease., Br. Heart J. 351321, 1973. Bartel, A. G., Behar, V. S., Peter, R. H., et al,: Graded exercise stress tests in angiographically documented coronary artery disease, Circulation 49: 348,1974. Roitman, D., Jones, W. B., and Sheffield, L. T.: Comparison of submaximal exercise ECG test with coronary cineangiography, Ann. Intern. Med. 72:641, 1970. 43

114. Proud&, W. L., Shirey, E. K., and Sones, F. M.: Selective tine coronary angiography: Correlation with clinical findings in 1000 patients, Circulation 33901, 1966. 115. F&singer, G. C., and Smith, R. F.: Correlation of electrocardiographic studies and arteriographic findings with angina pectoris, Circulation 46: 1173,1972. 116. McConahay, D. R., McCallister, B. D., and Smith, R. E.: Post-exercise electrocardiography: Correlation with coronary arteriography and left ventricular hemodynamics, Am. J. Cardiol. 28:1, 1972. 117. Campeau, L., Bourassa, M. G., Bois, M. A., et al.: Clinical significance of selective coronary cinearteriography, Can. Med. Assoc. J. 99:1063,1968. 118. Gensini, G. G., and Kelly, A. E.: Incidence and progression of coronary artery disease. An angiographic correlation in 1,263 patients, Arch. Intern. Med. 129814, 1972. 119. Bruce, R. A., and McDonough, J. R.: Stress testing in screening for cardiovascular disease, Bull. N. Y. Acad. Med. 45:1288,1969. 120. Cumming, G. R., Samm, J., Borysyk, L., et al.: Electrocardiographic changes during exercise in asymptomatic men: 3-year follow-up, Can. Med. Assoc. J. 112:578,1975. 121. Lepeshkin, E., and Surawicz, B.: Characteristics of true positive and false positive results of electrocardiographic Master-two-step exercise test, N. Engl. J. Med. 258:511,1968. 122. Astrand, I.: Exercise electrocardiograms recorded twice with an 8-year interval in a group of 204 women and men 48-63 years oid, Acta Med. Scand. 178:27,1966. 123. Profant. G. R.. Earlv. R. G., Nilson, K. L., et aZ.: Responses to maximal exercise in healthy middle-aged women, J. Appl. Physiol. 33595, 1972. 124. Cummine. G. R.. Dufresne. C.. and Samm. J.: Exercise ECG chances in normal women, dan. Med. Ass&. J. 109:108; 1973. 125. Waxler, E. B., Kimlins, D., and Dreifus, L. S.: The fate of women with normal coronary arteriograms and chest pain resembling angina pectoris, Am. J. Cardiol. 28251971. 126. Kemp, H. G., Jr., Vokonas, P. S., Cohn, P. F., and Gorlin, R.: The angina1 syndrome associated with normal coronary arteriograms, Am. J. Med. 54: 735,1973. 127. Sketch, M. H., Mohinddin, S. M., Lynch, J. D., Zencka, A. E., and Runco, V.: Significant sex differences in the correlation of electrocardiographic exercise testing and coronary arteriograms, Am. J. Cardiol. 36:169,1975. 128. Kusumi, F., Bruce, R. A., Ross, M. A., Trimble, S., and Voigt, A. E.: Elevated arterial pressure and post-exertional ST segment depression in middle-aged women, Am. Heart J. 92:576,1976. 129. Kansal, S., Roitman, D., and Sheffield, L. T.: Stress testing with STsegment depression at rest: An angiographic correlation, Circulation 54: 636.1976. 130. Jaeschke, M., and Belx, G.: Influence of digoxin on the exercise electrocardiogram of healthy men, Med. Klin. 71:2059, 1976. 131. Whinnery, J. E., Froelicher, V. F., Stewart, A., et ul.: The electrocardiographic response to maximal treadmill exercise of asymptomatic men with left bundle branch block, Am. Heart J. 94:316,1977. 132. Whinnery, J. E., Froelicher, V. F., Stewart, A., et al.: The electrocardiographic response to maximal treadmill exercise of asymptomatic men with right bundle branch block, Chest 71:335,1977. 133. Friedman, M., and Marcus, F. I.: Exercise testing in patients with right bundle branch block, Am. J. Cardiol., in press. 134. Harris, C., Aronow, W., Parker, D., et aE.: Treadmill stress test in left ventricular hypertrophy, Chest 63:353,1973. 135. Astrand, I., Blomquist, G., and Orinius, E.: ST changes at exercise in patients with short P-R interval, Acta Med. Stand. 185:205, 1969. 44

136.

137. 138.

139.

140.

Lobstein, H. P., Horwitz, L. D., Curry, G. C., et al.: Electrocardiographic abnormalities and coronary arteriograms in the mitral click-murmur syndrome, N. Engl. J. Med. 289:127,1973. Barnard, R., MacAlpin, R., Kattus, A., et al.: Ischemic response to sudden strenuous exercise in healthy men, Circulation 48936, 1973. Ericsson. M.. Granath. A.. Ohldn. P.. Siidermark. T.. and Volne. U.: Arrhythmias’ and symptoms during treadmill testing three weeks ‘after myocardial infarction in 100 patients, Br. Heart J. 35:787,1973. Ibsen, H., Kjeller, E., Styperek, J., and Pederson, A.: Routine exercise ECG three weeks after acute myocardial infarction, Acta Med. Stand. 198463, 1975. Markiewia, W., Houston, N., and DeBusk, R. F.: Exercise testing soon after myocardial infarction, Circulation 5626, 1977.

SELF-ASSESSMENT 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

ANSWERS

a, b and c a 22 and 16 a, b and c a andd b a b False a, b, c and d c b a=X%,b=9O%,c=lO% a, b, c and d True aandb

45