THE WESTERN Journal of Medicine
Refer to: Abbott JA, Tedeschi MA, Cheitlin MD: Graded treadmill stress testing-Patterns of physician use and abuse. West J Med 126:173-178, Mar 1977
Graded Treadmill Stress Testing Patterns of Physician Use and Abuse JOSEPH A. ABBOTT, MD; MARY ANN TEDESCHI, BS, and MELVIN D. CHEITLIN, MD San Francisco
Treadmill stress testing is used in assessing the condition of patients with known or suspected heart disease. We did a prospective study to clarify physician ordering and integration of the test. Ordering criteria were always complied with, although most tests were ordered for evaluation of atypical chest pain and only a few for high risk patients with known cardiac dysfunction, indicating a misplaced emphasis on the diagnostic capabilities of the test. Tests in patients with atypical chest pain and stress-induced ischemic changes were always integrated, but in 30 percent of patients with atypical pain and no stress-induced electrocardiographic changes, the tests were not used in patient management. This was often due to the misconception that negative findings on a stress test excluded coronary disease. Physicians should be alerted to this misplaced emphasis and misconception.
As EMPHASIZED in a host of publications, the major uses of graded treadmill exercise stress testing are assisting physicians in substantiating a diagnosis of myocardial ischemia, establishing an anginal threshold level, provoking exercise-induced arrhythmias, establishing a level of activity after recovery from myocardial infarction, providing an objective index of effectiveness of therapy and reassuring an anxious patient, among others.'-5 However, few authors examine the actual utilization of treadmill tests. The prospective study reported here evaluated the pattern of physician use of graded treadmill exercise stress tests. From the Medical Service, San Francisco General Hospital; and the Department of Medicine, University of California, San Francisco. Presented in part at the annual meeting of the Western Section, American Federation for Clinical Research, February 5th and 6th, 1976, Carmel, California. Submitted, revised, June 21, 1976. Reprint requests to: Joseph A. Abbott, MD, Room SG I, San Francisco General Hospital, 1001 Potrero Avenue, San Francisco, CA 94110.
Materials and Methods The study had two purposes: (1) to determine if the relatively expensive graded exercise tests are ordered for appropriate reasons at San Francisco General Hospital and (2) to find if the information obtained from the tests affect, in any way, patients' subsequent management. Therefore, we designed the study using a prospective format in which acceptable criteria for ordering the test and acting upon the results were determined before initiation of the test.67 In order to make the experiment both relevant and educational, the predetermined criteria were established by both house staff and attending physicians before the inception of the study; they are outlined in Tables 1 and 2. We attempted to further ensure general acceptance of the ordering and action criteria via a staff conference on graded treadmill testing, which ensured wide promulgation to the rHE WESTERN JOURNAL OF MEDICINE
173
TREADMILL STRESS TESTING
house staff and attending physicians. Because most of the tests were ordered by the physicians in the outpatient division, collection of data was facilitated because outpatient care for each patient is the continuous responsibility of one house officer under staff supervision, and the problem-oriented medical record is used throughout the hospital. All graded exercise tests done in the cardiopulmonary unit at San Francisco General Hospital between July 1, 1974 and June 30, 1975 were examined. At this teaching medical facility, all exercise tests are ordered by interns, residents and staff physicians. A physician telephones the cardiology secretary and asks her to schedule an exercise appointment. Subsequently, he must also complete an official request and mail it to the staff cardiologist who carries out the test. This request form includes a succinct outline of the type of patient who is a subject for testing. The form strongly implies that the test should be used in patients with known cardiac disease. It is not required that any test request be approved by the cardiology staff physician. Each patient referred for testing was interviewed, examined and stress tested by one of the authors (J.A.). Before each test, the reason for the test was ascertained, the patient signed an informed consent and a resting 12-lead electrocardiogram was done immediately before testing. Testing was done on a Quinton treadmill, model No. 18-54. The exercise stages were adapted from the recommendations of Sheffield,8 and consisted of three minutes of exercise, at each sequential stage, starting first at 1.5 miles per hour (mph) and a 10 percent grade, then at staged increments TABLE 1.-Acceptable Criteria for Ordering Graded Treadmill Tests
Attempting confirmation of diagnosis of angina that is uncertain from the clinical history. Evaluating whether arrhythmias improve or worsen with effort. Ascertaining appropriate postmyocardial infarction exercise prescription. Screening high risk patients-namely those with stable angina pectoris who are at high risk of having severe three-vessel coronary disease or its equivalent and symptomatic patients at high risk of having coronary disease, such as firemen, policemen, persons with diabetes or hypertension, and persons with strong family history of coronary disease. Evaluating the results of antianginal therapy, including drugs and surgical procedures. Evaluating for industrial or disability purposes. Evaluating selected patients with aortic stenosis, subaortic stenosis or mitral valve disease. Reassuring anxious patients without heart disease.
174
MARCH 1977 * 126 * 3
of 1 mph to a maximum of 4.5 mph or the rate of speed that is just below that when the patient must jog. Thereafter, exercise stages were achieved by increasing the incline of the treadmill from a 10 percent grade at Stages I through IV to more steep grades at increments of 2.5 percent per exercise level. Electrocardiographic monitoring was accomplished by a single bipolar precordial lead system of the type recommended by Ellestad and co-workers9 and by a tachometer, oscilloscope, and continuous strip chart and magnetic tape recordings.9 Blood pressure was recorded by sphygmomanometer before, at each exercise stage during and immediately after exercise, and every two minutes thereafter for 11 minutes. Exercise was continued until the patient achieved 90 percent of the maximum predicted heart rate or developed untoward symptoms, including an anginal syndrome, frequent ventricular irritability, arterial hypotension, extreme fatigue, breathlessness, or lightheadedness. Patient safety was emphasized. The patterns of the electrocardiographic ST segments-including those at rest, at the third minute of each exercise stage, at the peak exercise achieved, and at each minute after exercise for 10 minutes-were reviewed independently by two of the authors (J.A. 'and M.C.) .* Using the criteria of Ellestad, the exercise test was considered positive for ischemia when ST-segment depression of *The two cardiologists reviewing the stress tests for positivity or negativity of ST-segment depression in the atypical angina subgroup of patients agreed in 41 instances (78 percent) and disagreed in 11 (21 percent). These findings are much better than those reported by Blackburn and co-workers10 for noncomputerized methods of stress testing interpretation and internally invalidated the testing procedure.
TABLE 2.-Acceptable Criteria for Integrating Graded Treadmill Test Results If test result is positive: Medical or surgical therapy is initiated or continued. Physician restricts or liberalizes patient activities. Coronary angiography is advised and carried out. Medical strategy is changed. If test result is negative, borderline or inconclusive: The clinical diagnosis of heart disease is still certain and the physician proceeds as in above. The clinical diagnosis of heart disease is excluded and the physician seeks another cause for the patient's complaints and discontinues therapy for heart disease.
The clinical diagnosis of heart disease is unresolved and the physician orders further diagnostic tests which may include coronary angiography. Further drug therapy for arteriosclerotic heart disease is legitimate only if the reason is explicitly stated in medical records and the patient responds to this drug therapy.
TREADMILL STRESS TESTING
at least 0.1 mV in depth lasting for at least 0.08 seconds was present on any of the exercise or recovery tracings.1' Arrthythmias were tabulated, where appropriate, and observations regarding development of congestive heart failure and state of physical condition were made and reported. When the cardiologists disagreed with the interpretation of electrocardiographic changes, the opinion of the testing cardiologist was used in the final report. All exercise test reports were telephoned to the ordering physician without any judgmental statement regarding the necessity for the test or suggestions for future patient management, and a typed report was placed in the patient's medical record to ensure that the results reached the attending physician. This report emphasized the duration of exercise for symptoms or ischemia to develop (a reflection of the patient's cardiac disability) or if the patient achieved 90 percent of maximum heart rate, the duration of exercise was related to physical fitness. If other adverse or untoward effects developed, such as an inordinant rise or fall in blood pressure, excessive fatigue or exhaustion, cerebral symptoms or heart failure, then these aspects of the test were also duly reported in the typed report. After a minimum of one month, the charts of all subjects were analyzed by two of the authors (J.A. and M.T.) for acceptable integration of the test results into patient management using the predetermined appropriate action criteria. The action criteria were liberally interpreted so that a physician's response was deemed appropriate if it could be inferred from the actual course that transpired or from the physician's ordering of medications even if it was not explicitly stated in the medical record. These included continuing or discontinuing therapy, carrying out coronary angiography and limiting or liberalizing exercise (Table 2).
Results In all, 90 graded treadmill tests were done during the study period and reviewed; 86 of the patients were referred from the outpatient division. Of the tests, 40 were ordered by members of the house staff; the remainder by cardiac or medical full-time staff physicians. The reasons for ordering the tests are summarized in Table 3. Most of the tests were ordered to assist in resolving the problem of chest pain (60 percent), but relatively few were ordered for evaluating high risk patients (18 percent) including patients with known angina pectoris, diseases that predispose to coronary
TABLE 3.-All Graded Exercise Treadmill Tests Done at San Francisco General Hospital From July 1, 1974 to June 30, 1975 Physicians' Reasons for Ordering Test
No. of Patients
54 Chest pain-atypical angina pectoris Arrhythmias improve or worsen ....... 11 5 Postmyocardial infarction ............ 1 Classic angina pectoris .............. Evaluate medicine and/or surgical 5 coronary therapy ....... ........... 6 Functional evaluation ................ 5 Valvular and congenital heart disease 3 Reassurance of anxious patient ........ TOTAL ............................ 90 Acceptable reasons for ordering test .. 90
Percent
60 12 5 1 6 7 6 3 100
100
artery disease such as hypertension and diabetes, postmyocardial infarction and arrhythmias. However, at least one of the predetermined acceptable criteria for ordering the treadmill exercise tests was met in each instance; no test was ordered that was inappropriate. Because the group of patients with chest pain of uncertain cause, possibly atypical angina, was the largest studied, each of these patients' medical records was reviewed for integration of the test results into patient management. In this review the predetermined acceptable criteria of physician action were used. Only 52 charts for the 54 patients in this subgroup were available for review. The numbers of patients in each of the remaining seven categories were too few to yield significant data. Although all 52 patients were referred for evaluation of chest pain atypical of angina pectoris, the interview before the test often showed a syndrome completely compatible with angina-that is, exercise or mental stress resulted in a precordial sensation of chest heaviness or tightness that was self-limited, lasted less than 20 minutes or was relieved with the sublingual use of nitroglycerin. Twelve such patients were identified; the variability of the physician-patient interview in obtaining historical documentation of the anginal syndrome has been discussed by others.'2 Atypical angina was presumed present if some of the patient's complaints were compatible with, but unusual for, the anginal syndrome, such as sharp rather than heavy or constricting precordial pains, symptoms present at rest but not exercise, or pain limited to a localized area over the apex of the heart. Because no significant differences were noted when the results of physician integration of the test into patient management were tabuTHE WESTERN JOURNAL OF MEDICINE
175
TREADMILL STRESS TESTING
lated for those patients who had clear-cut angina or for those who had atypical angina, the results were grouped for the sake of clarity. The results showed that when the graded treadmill test was positive for ischemic ST-segment depression in six of 52 instances, acceptable criteria for integration of the test results into patient management were always met. However, when the ST-segment depression was negative (39 of 52) or borderline (seven of 52), an inordinate number of physicians did not integrate the exercise data into patient management (Table 4). A total of 39 graded treadmill stress tests were negative for ischemic ST-segment depression or development of a typical anginal syndrome in patients referred for atypical chest pains. In 26 (67 percent), physician integration of the results was appropriate. The reason for the appropriate physician action did not have to be stated explicitly in the hospital record but at a minimum could be inferred by the reviewers from subsequent physician actions, whether diagnostic or therapeutic, including prescribing behavior. In 13 instances (33 percent), inappropriate physician action was judged present. In 11, the test results were not integrated into patient management in any explicit or implicit fashion even though the symptoms of chest pain prompted the patients to seek medical attention and persisted in all; neither cause of the complaints was determined nor was symptomatic therapy to relieve them instituted. Subsequent personal interviews with the attending physicians caring for these patients showed that the physicians believed a negative graded exercise test excluded any possibility of coronary artery disease and that consequently their patients' chest complaints could not be due to cardiac disease; no further attempt was made to find out the origin of the patients' symptoms. In two patients, the test results explicitly mentioned the exclusion of coronary artery disease, but in spite of this, antianginal therapy was continued. The rationale for doing this and the results of such treatment were unclear. The staff caring for these 13 patients included ten interns, one resident, one cardiac fellow and one medical attending staff physician. In six subjects, the graded treadmill exercise tests were positive for ischemic ST-segment depression, and in each subject appropriate physician action (as determined by the ordering physicians before the initiation of the study) followed the reporting of the patient's exercise test (Table
176
MARCH 1977 * 126 * 3
TABLE 4.-Integration of Exercise Test Results into Management of 54 Patients with Atypical Angina at San Francisco General Hospital From July 1, 1974 to June 30, 1975 Appropriate Integration
Test Results
Positive ........ Negative ....... Borderline or Inconclusive
No. of No. of Patients Patients Percent
Inadequate Integration No. of Patients Percent
6 39
6 26
100 67
0 13
0 33
7
5
72
2
28
52 TOTAL .. Charts not located 2
37
15
2). In three patients with ischemic response, typical angina developed at the time of stress testing. In one the test resulted in restriction of activities, in the second propranolol was added to nitroglycerine therapy and in the third the magnitude of the ST depression played a role in the decision to carry out coronary cineangiograms during which significant three-vessel disease was found. Three of the six patients with ischemic responses reproduced their atypical anginal symptoms with effort. In one with 1-mm exercise-induced, ST-segment depression, coronary cineangiography was done because of persistence of chest pain unresponsive to medical management and a low grade stenosis of the right coronary artery proximal to the crux was shown. In the second patient, a mitral valve prolapse syndrome was diagnosed clinically and verified by echocardiography. Although it was never determined whether or not the positive ischemic response to exercise was falsely positive in the third, medical therapy with nitroglycerin successfully relieved the symptoms.
Discussion These results show that the physicians ordering the graded treadmill exercise test were totally in compliance with explicit and predetermined ordering criteria. This is in harmony with findings of other medical care audits regarding ordering patterns from this hospital.7 In addition, compliance was in no way affected by preselection, because the cardiology division did not screen for approval the tests scheduled by the ordering physicians. When medical use of one specific laboratory test (versus panels of laboratory tests) is audited, malutilization of the test is more a subtle phenomenon than heretofore emphasized.
TREADMILL STRESS TESTING
We have previously noted that malutilization of specific laboratory tests is often due to lack of integration of the test results into patient management.7 Regarding graded treadmill exercise testing, failure to exercise high risk and known coronary patients and failure to always properly integrate the results into patient management appeared to be caused by a dual misconception on the part of the patients' physicians. The first misconception, which was derived from physician interview of the 11 patients in whom there were negative stress test results but in whom symptoms persisted, is the idea that exercise stress testing is the definitive noninvasive tool to verify the existence, or exclude the presence, of ischemic coronary artery disease. Although we concur that stress testing should be used to evaluate atypical chest pain, our results show that at our hospital inordinate emphasis is placed on treadmill tests to resolve these questions concerning coronary disease. In six patients the exercise test gave positive findings, indicating to the referring physician that coronary disease was responsible for the patient's complaints. Although the number of patients was small, 100 percent compliance with the explicit predetermined integration criteria was always present and management goals were clear to the physician. In those patients with possible but atypical angina, a negative finding on a graded treadmill test too often resulted in less clearly defined diagnostic and subsequent therapeutic maneuvers. In 30 percent, this led to physician reassutrance that the patient's symptoms were not due to a cardiac cause; even though the complaints persisted, the exercise test was the only diagnostic test ordered. No further therapy, even symptomatic treatment, was prescribed. In these patients, the physicians concluded that a negative finding on a treadmill test definitely excluded the presence of coronary artery disease. Patently, this attitude is fallacious, because it has been amply shown that ST-segment depression during exercise testing correlates to a high degree with the presence of significant and severe coronary artery disease in two or three vessels, but the absence of such depression does not exclude it.13 Borer and co-workers'4 recently reemphasized this point when they found negative exercise tests in 39 percent of patients with angiographically documented severe coronary disease and typical angina. When there is less clear-cut historical information suggesting angina, there is also a lower
probability that stress testing will resolve the issue of the presence or absence of coronary disease.15 Because there is ample evidence that coronary angiography has a significant morbidity and mortality as well as expense, the physician's attitude toward subjecting all patients with possible atypical angina to coronary angiography is natural. Possibly, stress testing combined with radioisotope scanning of the heart will increase the sensitivity of noninvasive diagnosis in such patients. The second misconception we encountered in our hospital is that there is little practical benefit from exercise testing subjects with known coronary artery disease, especially those with stable angina pectoris or recovered myocardial infarction. It must be emphasized that this is not a result of our failure to see large numbers of patients with coronary artery disease because 200 patients with acute myocardial infarction were admitted, treated and discharged from the hospital during the study period, nor is it a result of our explicit teaching. We believe it reflects our failure to communicate the proper emphasis of the ordering criteria. Only one patient with healed myocardial infarction was referred for exercise testing during the study year. There is ample evidence that the patients most in need of graded treadmill stress testing are those with known coronary artery disease. For example, we believe that any patient, even one with stable angina, in whom pronounced ST-segment depression is noted during graded treadmill testing may be a serious candidate for coronary angiography if only because pronounced ST-segment depression has been shown to correlate well with left main stem or three vessel coronary disease, lesions which may be surgically remediable.6- In addition, we believe that exercise testing is clearly called for in a patient who has recovered from an acute myocardial infarction in order to assist the attending physician in prescribing a rational and safe rehabilitation program. Although we concur that graded treadmill stress tests certainly should be used to evaluate patients with atypical chest pain, we conclude that stress testing is used inordinately as the definitive test in the differential diagnostic workup of these patients. On the other hand, stress testing is underused in those patients with known and stable coronary disease especially those with angina or healed myocardial infarction. Too often a negative stress test result is ignored and used only by the ordering physician to reassure himself that THE WESTERN JOURNAL OF MEDICINE
177
TREADMILL STRESS TESTING
a patient does not have severe coronary disease and it is not used in guiding management of the persistent symptoms after the test is completed. REFERENCES 1. Rochmis P, Blackburn H: Exercise tests-A survey of procedtires, safety, and litigation experience in approximately 170,000 tests. JAMA 217:1061-1066, 1971 2. Bruce RA, Kusumi F, Hosmer D: Maximal oxygen intake and nomographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 85:546-562, 1973 3. Kemp GL, Ellestad MH: The current application of maximal treadmill stress testing. Calif Med 107:406-412, Nov 1967 4. Blomqvist CG: Use of exercise testing for diagnostic and functional evaluation of patients with arteriosclerotic heart disease. Circulation 44:1120-1136, 1971 5. DeBusk R: The value of exercise stress testing. JAMA 232:956-958, 1975 6. Brown CR Jr, Uhl HSM: Mandatory continuing educationSense or nonsense? JAMA 213:1660-1668, 1970 7. Goldberg GA, Abbott JA: Letter to the Editor: Explicit criteria for use of laboratory tests. Ann Intern Med 81:857-858, 1974 8. Sheffield LT: Graded exercise test (GXT) for ischemic heart disease-A submaximal test to a target heart rate, In Kattus AA,
Brock, LL, Bruce RA, et al (Eds): Exercise Testing and Training of Apparently Healthy Individuals: A Handbook for Physicians. New York, American Heart Association, 1972, pp 35-38 9. Ellestad MH, Allen W, Wan MCK, et al: Maximal treadmill stress testing for cardiovascular evaluation. Circulation 39:517222, 1969 10. Blackburn H, Blomqvist G, Freiman A, et al: The exercise electrocardiogram: Differences in interpretation-Report of a technical group on exercise electrocardiography. Am J Cardiol 21:871-880, 1968 11. Ellestad MH: Stress Testing-Principles and Practice. Philadephia, F.A. Davis Company, 1975, pp 95-107 12. Rose G: Variability of angina-Some implications for epidemiology. Br J Prev Soc Med 22:12-15, 1968 13. Martin CM, McConahay DR: Maximal treadmill exercise electrocardiography-Correlations with coronary arteriography and cardiac hemodynamics. Circulation 46:956-962, 1972 14. Borer JS, Brensike JF, Redwood DR, et al: Limitations of the electrocardiographic response to exercise in predicting coronary-artery disease. N Engl J Med 293:367-371, 1975 15. Redwood DR, Epstein SE: Uses and limitations of stress testing in the evaluation of ischemic heart disease. Circulation 46:1115-1131, 1972 16. Cheitlin MD, Davia JE, deCastro CM, et al: Correlation of "critical" left coronary artery lesions with positive submaximal exercise tests in patients with chest pain. Am Heart J 89:305-310, 1975
A Reliable Method of Determining Intravascular Volume We find the most reliable method for following a patient's intravascular volume is the measurement of the blood pressure in two positions, mainly looking for postural hypotension. If I had to have one clinical method, I would take that. You can't use the dryness of the mouth to determine dehydration because some patients are mouth-breathers and they all have dry mouths, or many of them do. You can't use axillary perspiration... But the most reliable thing clinically is to measure their blood pressure lying in bed and sitting up with their feet dangling. Now, it isn't enough just to have them sit up in bed. They must dangle their feet. I must agree that occasionally you may have a problem with an old person who has been lying in bed for many months and suddenly you rush in and have the patient sit up. That may be the last time you examine him. Obviously, you must pick your case. But if you get greater than a 20 mm drop of mercury mean pressure, you would then have a pretty good indication that the patient has hypovolemia. Now, there are certain problems here... Again if the patient is an old person and has been lying in bed for a very long time, you may get a postural drop and he may not be hypovolemic. If the patient has autonomic dysfunction (diabetes, let's say), you may get a postural drop and he may not be hypovolemic. The clue here is, of course, that the pulse rate will not go up. You may also be fooled by a patient who is taking an antihypertensive drug, such as methyldopa (Aldometg), which produces postural hypotension. But again, the best clinical guide we have is this postural change. -NORMAN LASKER, MD, Philadelphia Extracted from Audio-Digest Internal Medicine, Vol. 23, No. 14, in the Audio-Digest Foundation's subscription series of taperecorded programs. For subscription information: 1930 Wilshire Blvd., Suite 700, Los Angeles, CA 90057
178
MARCH 1977 * 126 * 3
Refer to: Abbott JA, Tedeschi MA, Cheitlin MD: Graded treadmill stress testing-Patterns of physician use and abuse. West J Med 126:173-178, Mar 1977
Graded Treadmill Stress Testing Patterns of Physician Use and Abuse JOSEPH A. ABBOTT, MD; MARY ANN TEDESCHI, BS, and MELVIN D. CHEITLIN, MD San Francisco
Treadmill stress testing is used in assessing the condition of patients with known or suspected heart disease. We did a prospective study to clarify physician ordering and integration of the test. Ordering criteria were always complied with, although most tests were ordered for evaluation of atypical chest pain and only a few for high risk patients with known cardiac dysfunction, indicating a misplaced emphasis on the diagnostic capabilities of the test. Tests in patients with atypical chest pain and stress-induced ischemic changes were always integrated, but in 30 percent of patients with atypical pain and no stress-induced electrocardiographic changes, the tests were not used in patient management. This was often due to the misconception that negative findings on a stress test excluded coronary disease. Physicians should be alerted to this misplaced emphasis and misconception.
As EMPHASIZED in a host of publications, the major uses of graded treadmill exercise stress testing are assisting physicians in substantiating a diagnosis of myocardial ischemia, establishing an anginal threshold level, provoking exercise-induced arrhythmias, establishing a level of activity after recovery from myocardial infarction, providing an objective index of effectiveness of therapy and reassuring an anxious patient, among others.'-5 However, few authors examine the actual utilization of treadmill tests. The prospective study reported here evaluated the pattern of physician use of graded treadmill exercise stress tests. From the Medical Service, San Francisco General Hospital; and the Department of Medicine, University of California, San Francisco. Presented in part at the annual meeting of the Western Section, American Federation for Clinical Research, February 5th and 6th, 1976, Carmel, California. Submitted, revised, June 21, 1976. Reprint requests to: Joseph A. Abbott, MD, Room SG I, San Francisco General Hospital, 1001 Potrero Avenue, San Francisco, CA 94110.
Materials and Methods The study had two purposes: (1) to determine if the relatively expensive graded exercise tests are ordered for appropriate reasons at San Francisco General Hospital and (2) to find if the information obtained from the tests affect, in any way, patients' subsequent management. Therefore, we designed the study using a prospective format in which acceptable criteria for ordering the test and acting upon the results were determined before initiation of the test.67 In order to make the experiment both relevant and educational, the predetermined criteria were established by both house staff and attending physicians before the inception of the study; they are outlined in Tables 1 and 2. We attempted to further ensure general acceptance of the ordering and action criteria via a staff conference on graded treadmill testing, which ensured wide promulgation to the rHE WESTERN JOURNAL OF MEDICINE
173
TREADMILL STRESS TESTING
house staff and attending physicians. Because most of the tests were ordered by the physicians in the outpatient division, collection of data was facilitated because outpatient care for each patient is the continuous responsibility of one house officer under staff supervision, and the problem-oriented medical record is used throughout the hospital. All graded exercise tests done in the cardiopulmonary unit at San Francisco General Hospital between July 1, 1974 and June 30, 1975 were examined. At this teaching medical facility, all exercise tests are ordered by interns, residents and staff physicians. A physician telephones the cardiology secretary and asks her to schedule an exercise appointment. Subsequently, he must also complete an official request and mail it to the staff cardiologist who carries out the test. This request form includes a succinct outline of the type of patient who is a subject for testing. The form strongly implies that the test should be used in patients with known cardiac disease. It is not required that any test request be approved by the cardiology staff physician. Each patient referred for testing was interviewed, examined and stress tested by one of the authors (J.A.). Before each test, the reason for the test was ascertained, the patient signed an informed consent and a resting 12-lead electrocardiogram was done immediately before testing. Testing was done on a Quinton treadmill, model No. 18-54. The exercise stages were adapted from the recommendations of Sheffield,8 and consisted of three minutes of exercise, at each sequential stage, starting first at 1.5 miles per hour (mph) and a 10 percent grade, then at staged increments TABLE 1.-Acceptable Criteria for Ordering Graded Treadmill Tests
Attempting confirmation of diagnosis of angina that is uncertain from the clinical history. Evaluating whether arrhythmias improve or worsen with effort. Ascertaining appropriate postmyocardial infarction exercise prescription. Screening high risk patients-namely those with stable angina pectoris who are at high risk of having severe three-vessel coronary disease or its equivalent and symptomatic patients at high risk of having coronary disease, such as firemen, policemen, persons with diabetes or hypertension, and persons with strong family history of coronary disease. Evaluating the results of antianginal therapy, including drugs and surgical procedures. Evaluating for industrial or disability purposes. Evaluating selected patients with aortic stenosis, subaortic stenosis or mitral valve disease. Reassuring anxious patients without heart disease.
174
MARCH 1977 * 126 * 3
of 1 mph to a maximum of 4.5 mph or the rate of speed that is just below that when the patient must jog. Thereafter, exercise stages were achieved by increasing the incline of the treadmill from a 10 percent grade at Stages I through IV to more steep grades at increments of 2.5 percent per exercise level. Electrocardiographic monitoring was accomplished by a single bipolar precordial lead system of the type recommended by Ellestad and co-workers9 and by a tachometer, oscilloscope, and continuous strip chart and magnetic tape recordings.9 Blood pressure was recorded by sphygmomanometer before, at each exercise stage during and immediately after exercise, and every two minutes thereafter for 11 minutes. Exercise was continued until the patient achieved 90 percent of the maximum predicted heart rate or developed untoward symptoms, including an anginal syndrome, frequent ventricular irritability, arterial hypotension, extreme fatigue, breathlessness, or lightheadedness. Patient safety was emphasized. The patterns of the electrocardiographic ST segments-including those at rest, at the third minute of each exercise stage, at the peak exercise achieved, and at each minute after exercise for 10 minutes-were reviewed independently by two of the authors (J.A. 'and M.C.) .* Using the criteria of Ellestad, the exercise test was considered positive for ischemia when ST-segment depression of *The two cardiologists reviewing the stress tests for positivity or negativity of ST-segment depression in the atypical angina subgroup of patients agreed in 41 instances (78 percent) and disagreed in 11 (21 percent). These findings are much better than those reported by Blackburn and co-workers10 for noncomputerized methods of stress testing interpretation and internally invalidated the testing procedure.
TABLE 2.-Acceptable Criteria for Integrating Graded Treadmill Test Results If test result is positive: Medical or surgical therapy is initiated or continued. Physician restricts or liberalizes patient activities. Coronary angiography is advised and carried out. Medical strategy is changed. If test result is negative, borderline or inconclusive: The clinical diagnosis of heart disease is still certain and the physician proceeds as in above. The clinical diagnosis of heart disease is excluded and the physician seeks another cause for the patient's complaints and discontinues therapy for heart disease.
The clinical diagnosis of heart disease is unresolved and the physician orders further diagnostic tests which may include coronary angiography. Further drug therapy for arteriosclerotic heart disease is legitimate only if the reason is explicitly stated in medical records and the patient responds to this drug therapy.
TREADMILL STRESS TESTING
at least 0.1 mV in depth lasting for at least 0.08 seconds was present on any of the exercise or recovery tracings.1' Arrthythmias were tabulated, where appropriate, and observations regarding development of congestive heart failure and state of physical condition were made and reported. When the cardiologists disagreed with the interpretation of electrocardiographic changes, the opinion of the testing cardiologist was used in the final report. All exercise test reports were telephoned to the ordering physician without any judgmental statement regarding the necessity for the test or suggestions for future patient management, and a typed report was placed in the patient's medical record to ensure that the results reached the attending physician. This report emphasized the duration of exercise for symptoms or ischemia to develop (a reflection of the patient's cardiac disability) or if the patient achieved 90 percent of maximum heart rate, the duration of exercise was related to physical fitness. If other adverse or untoward effects developed, such as an inordinant rise or fall in blood pressure, excessive fatigue or exhaustion, cerebral symptoms or heart failure, then these aspects of the test were also duly reported in the typed report. After a minimum of one month, the charts of all subjects were analyzed by two of the authors (J.A. and M.T.) for acceptable integration of the test results into patient management using the predetermined appropriate action criteria. The action criteria were liberally interpreted so that a physician's response was deemed appropriate if it could be inferred from the actual course that transpired or from the physician's ordering of medications even if it was not explicitly stated in the medical record. These included continuing or discontinuing therapy, carrying out coronary angiography and limiting or liberalizing exercise (Table 2).
Results In all, 90 graded treadmill tests were done during the study period and reviewed; 86 of the patients were referred from the outpatient division. Of the tests, 40 were ordered by members of the house staff; the remainder by cardiac or medical full-time staff physicians. The reasons for ordering the tests are summarized in Table 3. Most of the tests were ordered to assist in resolving the problem of chest pain (60 percent), but relatively few were ordered for evaluating high risk patients (18 percent) including patients with known angina pectoris, diseases that predispose to coronary
TABLE 3.-All Graded Exercise Treadmill Tests Done at San Francisco General Hospital From July 1, 1974 to June 30, 1975 Physicians' Reasons for Ordering Test
No. of Patients
54 Chest pain-atypical angina pectoris Arrhythmias improve or worsen ....... 11 5 Postmyocardial infarction ............ 1 Classic angina pectoris .............. Evaluate medicine and/or surgical 5 coronary therapy ....... ........... 6 Functional evaluation ................ 5 Valvular and congenital heart disease 3 Reassurance of anxious patient ........ TOTAL ............................ 90 Acceptable reasons for ordering test .. 90
Percent
60 12 5 1 6 7 6 3 100
100
artery disease such as hypertension and diabetes, postmyocardial infarction and arrhythmias. However, at least one of the predetermined acceptable criteria for ordering the treadmill exercise tests was met in each instance; no test was ordered that was inappropriate. Because the group of patients with chest pain of uncertain cause, possibly atypical angina, was the largest studied, each of these patients' medical records was reviewed for integration of the test results into patient management. In this review the predetermined acceptable criteria of physician action were used. Only 52 charts for the 54 patients in this subgroup were available for review. The numbers of patients in each of the remaining seven categories were too few to yield significant data. Although all 52 patients were referred for evaluation of chest pain atypical of angina pectoris, the interview before the test often showed a syndrome completely compatible with angina-that is, exercise or mental stress resulted in a precordial sensation of chest heaviness or tightness that was self-limited, lasted less than 20 minutes or was relieved with the sublingual use of nitroglycerin. Twelve such patients were identified; the variability of the physician-patient interview in obtaining historical documentation of the anginal syndrome has been discussed by others.'2 Atypical angina was presumed present if some of the patient's complaints were compatible with, but unusual for, the anginal syndrome, such as sharp rather than heavy or constricting precordial pains, symptoms present at rest but not exercise, or pain limited to a localized area over the apex of the heart. Because no significant differences were noted when the results of physician integration of the test into patient management were tabuTHE WESTERN JOURNAL OF MEDICINE
175
TREADMILL STRESS TESTING
lated for those patients who had clear-cut angina or for those who had atypical angina, the results were grouped for the sake of clarity. The results showed that when the graded treadmill test was positive for ischemic ST-segment depression in six of 52 instances, acceptable criteria for integration of the test results into patient management were always met. However, when the ST-segment depression was negative (39 of 52) or borderline (seven of 52), an inordinate number of physicians did not integrate the exercise data into patient management (Table 4). A total of 39 graded treadmill stress tests were negative for ischemic ST-segment depression or development of a typical anginal syndrome in patients referred for atypical chest pains. In 26 (67 percent), physician integration of the results was appropriate. The reason for the appropriate physician action did not have to be stated explicitly in the hospital record but at a minimum could be inferred by the reviewers from subsequent physician actions, whether diagnostic or therapeutic, including prescribing behavior. In 13 instances (33 percent), inappropriate physician action was judged present. In 11, the test results were not integrated into patient management in any explicit or implicit fashion even though the symptoms of chest pain prompted the patients to seek medical attention and persisted in all; neither cause of the complaints was determined nor was symptomatic therapy to relieve them instituted. Subsequent personal interviews with the attending physicians caring for these patients showed that the physicians believed a negative graded exercise test excluded any possibility of coronary artery disease and that consequently their patients' chest complaints could not be due to cardiac disease; no further attempt was made to find out the origin of the patients' symptoms. In two patients, the test results explicitly mentioned the exclusion of coronary artery disease, but in spite of this, antianginal therapy was continued. The rationale for doing this and the results of such treatment were unclear. The staff caring for these 13 patients included ten interns, one resident, one cardiac fellow and one medical attending staff physician. In six subjects, the graded treadmill exercise tests were positive for ischemic ST-segment depression, and in each subject appropriate physician action (as determined by the ordering physicians before the initiation of the study) followed the reporting of the patient's exercise test (Table
176
MARCH 1977 * 126 * 3
TABLE 4.-Integration of Exercise Test Results into Management of 54 Patients with Atypical Angina at San Francisco General Hospital From July 1, 1974 to June 30, 1975 Appropriate Integration
Test Results
Positive ........ Negative ....... Borderline or Inconclusive
No. of No. of Patients Patients Percent
Inadequate Integration No. of Patients Percent
6 39
6 26
100 67
0 13
0 33
7
5
72
2
28
52 TOTAL .. Charts not located 2
37
15
2). In three patients with ischemic response, typical angina developed at the time of stress testing. In one the test resulted in restriction of activities, in the second propranolol was added to nitroglycerine therapy and in the third the magnitude of the ST depression played a role in the decision to carry out coronary cineangiograms during which significant three-vessel disease was found. Three of the six patients with ischemic responses reproduced their atypical anginal symptoms with effort. In one with 1-mm exercise-induced, ST-segment depression, coronary cineangiography was done because of persistence of chest pain unresponsive to medical management and a low grade stenosis of the right coronary artery proximal to the crux was shown. In the second patient, a mitral valve prolapse syndrome was diagnosed clinically and verified by echocardiography. Although it was never determined whether or not the positive ischemic response to exercise was falsely positive in the third, medical therapy with nitroglycerin successfully relieved the symptoms.
Discussion These results show that the physicians ordering the graded treadmill exercise test were totally in compliance with explicit and predetermined ordering criteria. This is in harmony with findings of other medical care audits regarding ordering patterns from this hospital.7 In addition, compliance was in no way affected by preselection, because the cardiology division did not screen for approval the tests scheduled by the ordering physicians. When medical use of one specific laboratory test (versus panels of laboratory tests) is audited, malutilization of the test is more a subtle phenomenon than heretofore emphasized.
TREADMILL STRESS TESTING
We have previously noted that malutilization of specific laboratory tests is often due to lack of integration of the test results into patient management.7 Regarding graded treadmill exercise testing, failure to exercise high risk and known coronary patients and failure to always properly integrate the results into patient management appeared to be caused by a dual misconception on the part of the patients' physicians. The first misconception, which was derived from physician interview of the 11 patients in whom there were negative stress test results but in whom symptoms persisted, is the idea that exercise stress testing is the definitive noninvasive tool to verify the existence, or exclude the presence, of ischemic coronary artery disease. Although we concur that stress testing should be used to evaluate atypical chest pain, our results show that at our hospital inordinate emphasis is placed on treadmill tests to resolve these questions concerning coronary disease. In six patients the exercise test gave positive findings, indicating to the referring physician that coronary disease was responsible for the patient's complaints. Although the number of patients was small, 100 percent compliance with the explicit predetermined integration criteria was always present and management goals were clear to the physician. In those patients with possible but atypical angina, a negative finding on a graded treadmill test too often resulted in less clearly defined diagnostic and subsequent therapeutic maneuvers. In 30 percent, this led to physician reassutrance that the patient's symptoms were not due to a cardiac cause; even though the complaints persisted, the exercise test was the only diagnostic test ordered. No further therapy, even symptomatic treatment, was prescribed. In these patients, the physicians concluded that a negative finding on a treadmill test definitely excluded the presence of coronary artery disease. Patently, this attitude is fallacious, because it has been amply shown that ST-segment depression during exercise testing correlates to a high degree with the presence of significant and severe coronary artery disease in two or three vessels, but the absence of such depression does not exclude it.13 Borer and co-workers'4 recently reemphasized this point when they found negative exercise tests in 39 percent of patients with angiographically documented severe coronary disease and typical angina. When there is less clear-cut historical information suggesting angina, there is also a lower
probability that stress testing will resolve the issue of the presence or absence of coronary disease.15 Because there is ample evidence that coronary angiography has a significant morbidity and mortality as well as expense, the physician's attitude toward subjecting all patients with possible atypical angina to coronary angiography is natural. Possibly, stress testing combined with radioisotope scanning of the heart will increase the sensitivity of noninvasive diagnosis in such patients. The second misconception we encountered in our hospital is that there is little practical benefit from exercise testing subjects with known coronary artery disease, especially those with stable angina pectoris or recovered myocardial infarction. It must be emphasized that this is not a result of our failure to see large numbers of patients with coronary artery disease because 200 patients with acute myocardial infarction were admitted, treated and discharged from the hospital during the study period, nor is it a result of our explicit teaching. We believe it reflects our failure to communicate the proper emphasis of the ordering criteria. Only one patient with healed myocardial infarction was referred for exercise testing during the study year. There is ample evidence that the patients most in need of graded treadmill stress testing are those with known coronary artery disease. For example, we believe that any patient, even one with stable angina, in whom pronounced ST-segment depression is noted during graded treadmill testing may be a serious candidate for coronary angiography if only because pronounced ST-segment depression has been shown to correlate well with left main stem or three vessel coronary disease, lesions which may be surgically remediable.6- In addition, we believe that exercise testing is clearly called for in a patient who has recovered from an acute myocardial infarction in order to assist the attending physician in prescribing a rational and safe rehabilitation program. Although we concur that graded treadmill stress tests certainly should be used to evaluate patients with atypical chest pain, we conclude that stress testing is used inordinately as the definitive test in the differential diagnostic workup of these patients. On the other hand, stress testing is underused in those patients with known and stable coronary disease especially those with angina or healed myocardial infarction. Too often a negative stress test result is ignored and used only by the ordering physician to reassure himself that THE WESTERN JOURNAL OF MEDICINE
177
TREADMILL STRESS TESTING
a patient does not have severe coronary disease and it is not used in guiding management of the persistent symptoms after the test is completed. REFERENCES 1. Rochmis P, Blackburn H: Exercise tests-A survey of procedtires, safety, and litigation experience in approximately 170,000 tests. JAMA 217:1061-1066, 1971 2. Bruce RA, Kusumi F, Hosmer D: Maximal oxygen intake and nomographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 85:546-562, 1973 3. Kemp GL, Ellestad MH: The current application of maximal treadmill stress testing. Calif Med 107:406-412, Nov 1967 4. Blomqvist CG: Use of exercise testing for diagnostic and functional evaluation of patients with arteriosclerotic heart disease. Circulation 44:1120-1136, 1971 5. DeBusk R: The value of exercise stress testing. JAMA 232:956-958, 1975 6. Brown CR Jr, Uhl HSM: Mandatory continuing educationSense or nonsense? JAMA 213:1660-1668, 1970 7. Goldberg GA, Abbott JA: Letter to the Editor: Explicit criteria for use of laboratory tests. Ann Intern Med 81:857-858, 1974 8. Sheffield LT: Graded exercise test (GXT) for ischemic heart disease-A submaximal test to a target heart rate, In Kattus AA,
Brock, LL, Bruce RA, et al (Eds): Exercise Testing and Training of Apparently Healthy Individuals: A Handbook for Physicians. New York, American Heart Association, 1972, pp 35-38 9. Ellestad MH, Allen W, Wan MCK, et al: Maximal treadmill stress testing for cardiovascular evaluation. Circulation 39:517222, 1969 10. Blackburn H, Blomqvist G, Freiman A, et al: The exercise electrocardiogram: Differences in interpretation-Report of a technical group on exercise electrocardiography. Am J Cardiol 21:871-880, 1968 11. Ellestad MH: Stress Testing-Principles and Practice. Philadephia, F.A. Davis Company, 1975, pp 95-107 12. Rose G: Variability of angina-Some implications for epidemiology. Br J Prev Soc Med 22:12-15, 1968 13. Martin CM, McConahay DR: Maximal treadmill exercise electrocardiography-Correlations with coronary arteriography and cardiac hemodynamics. Circulation 46:956-962, 1972 14. Borer JS, Brensike JF, Redwood DR, et al: Limitations of the electrocardiographic response to exercise in predicting coronary-artery disease. N Engl J Med 293:367-371, 1975 15. Redwood DR, Epstein SE: Uses and limitations of stress testing in the evaluation of ischemic heart disease. Circulation 46:1115-1131, 1972 16. Cheitlin MD, Davia JE, deCastro CM, et al: Correlation of "critical" left coronary artery lesions with positive submaximal exercise tests in patients with chest pain. Am Heart J 89:305-310, 1975
A Reliable Method of Determining Intravascular Volume We find the most reliable method for following a patient's intravascular volume is the measurement of the blood pressure in two positions, mainly looking for postural hypotension. If I had to have one clinical method, I would take that. You can't use the dryness of the mouth to determine dehydration because some patients are mouth-breathers and they all have dry mouths, or many of them do. You can't use axillary perspiration... But the most reliable thing clinically is to measure their blood pressure lying in bed and sitting up with their feet dangling. Now, it isn't enough just to have them sit up in bed. They must dangle their feet. I must agree that occasionally you may have a problem with an old person who has been lying in bed for many months and suddenly you rush in and have the patient sit up. That may be the last time you examine him. Obviously, you must pick your case. But if you get greater than a 20 mm drop of mercury mean pressure, you would then have a pretty good indication that the patient has hypovolemia. Now, there are certain problems here... Again if the patient is an old person and has been lying in bed for a very long time, you may get a postural drop and he may not be hypovolemic. If the patient has autonomic dysfunction (diabetes, let's say), you may get a postural drop and he may not be hypovolemic. The clue here is, of course, that the pulse rate will not go up. You may also be fooled by a patient who is taking an antihypertensive drug, such as methyldopa (Aldometg), which produces postural hypotension. But again, the best clinical guide we have is this postural change. -NORMAN LASKER, MD, Philadelphia Extracted from Audio-Digest Internal Medicine, Vol. 23, No. 14, in the Audio-Digest Foundation's subscription series of taperecorded programs. For subscription information: 1930 Wilshire Blvd., Suite 700, Los Angeles, CA 90057
178
MARCH 1977 * 126 * 3
