Author + information
- Received June 15, 1998
- Revision received January 25, 1999
- Accepted February 15, 1999
- Published online June 1, 1999.
- William R Lewis, MD, FACCa,* (, )
- Ezra A Amsterdam, MD, FACCa,
- Samuel Turnipseed, MDa,b and
- J.Douglas Kirk, MDa,b
- ↵*Reprint requests and correspondence: Dr. William R. Lewis, 4150 V St., Suite 3500, Sacramento, California 95817
The purpose of this study was to demonstrate the safety and utility of immediate exercise treadmill testing (IETT) of low risk patients presenting to the emergency department with known coronary artery disease (CAD).
More than 70% of the two million patients admitted to U.S. hospitals annually for suspected acute myocardial infarction (AMI) are found not to have had a cardiac event. We have previously demonstrated the safety and efficacy of IETT of selected low risk patients without known CAD presenting to the emergency department with chest pain. This study extends this approach to selected patients with a history of CAD.
One hundred patients evaluated by the chest pain emergency room to rule out AMI underwent IETT using a modified Bruce protocol upon admission to the hospital (median time <1 h).
Twenty-three patients (23%) had positive exercise electrocardiograms (ExECGs); an uncomplicated non–Q wave AMI was diagnosed in two patients. Thirty-eight patients (38%) had negative ExECGs and 39 patients (39%) had nondiagnostic ExECGs. Of these 100 patients, 64 were discharged immediately after IETT, 19 were discharged in less than 24 h after negative serial cardiac enzymes and stable electrocardiograms and 17 were discharged after further evaluation and treatment. There were no complications from exercise testing and no late deaths or AMI during six-month follow-up.
Immediate exercise treadmill testing of low risk patients with chest pain and known CAD is effective in further stratifying this group into patients who can be safely discharged and those who require hospital admission.
“Patients should be admitted to the coronary care unit solely on suspicion of having an acute myocardial infarction” has been, until recently, the standard of care for patients presenting to the emergency room with chest pain suggestive of myocardial origin (1). This practice has resulted in more than 2 million patients hospitalized every year for suspected acute myocardial infarction (AMI). Ultimately, as few as one third of these admissions are found to have had a coronary event (2), at a cost of $10 billion (3). Because of the relatively low coronary event rate and high cost of this approach, alternate management strategies have been proposed (4–13). These approaches have shown that low risk patients can be identified on the basis of symptoms, coronary risk factors, physical examination and the electrocardiogram (ECG) (14,15). The coronary event rate in selected patients without known coronary artery disease (CAD) is as low as 5%, and in patients with CAD as high as 26% (13). We have previously demonstrated the safety and efficacy of immediate exercise treadmill testing (IETT) in low risk patients without known coronary disease admitted to the hospital after presenting to the emergency department with chest pain (16,17). This study extends this management strategy to selected low risk patients with known coronary disease admitted to the chest pain emergency room (CPER) for suspected myocardial ischemia.
The CPER was established in 1993 to manage low risk cardiac patients with expected hospital stay of less than 24 h (17). All patients with known CAD evaluated by the CPER with chest pain of suspected cardiac etiology were eligible for this study. Evaluation in the emergency department included history, physical examination, bilateral arm blood pressure measurements, chest roentgenogram, 12-lead ECG and, in most cases, an initial total serum creatine kinase and its MB isoenzyme. From this population, low risk patients were selected on the basis of ECG, clinical findings and the ability to exercise. The resulting population was about 20% of patients screened, with inability to exercise being the main reason for exclusion. Chest pain description varied from typical to atypical for cardiac ischemia, and ECGs were normal or had only nonspecific repolarization abnormalities. Patients with ECGs diagnostic of AMI or ischemia were excluded as were patients with ST-segment and T-wave changes that would preclude accurate interpretation of the exercise electrocardiogram (ExECG). Changes of old myocardial infarction as well as the presence of persistent chest pain did not preclude IETT. Patients with chest roetgenograms demonstrating acute pulmonary or vascular disease were excluded, as were patients with clinical findings of heart failure (S3 gallop, rales, elevated jugular venous pressure) or pulse deficits suspicious for aortic dissection. Patients with chest pain believed to be from noncardiac etiologies, such as gastrointestinal bleeding or pulmonary embolism, were also not included.
Immediate exercise treadmill testing was performed using continuous 12-lead ECG monitoring and a modified Bruce protocol (two 1.7-mph stages at 0% and 5% incline preceding the standard Bruce protocol). Immediate exercise treadmill testing was performed as soon as possible after the CPER physician completed the evaluation. Exercise was terminated for ECG changes diagnostic of ischemia (≥1 mm horizontal or down-sloping ST-segment depression or elevation 80 ms after the J point), significant arrhythmia, systolic blood pressure decrease of ≥10 mm Hg or significant symptoms. A nondiagnostic test was defined as a negative exercise ECG with a peak heart rate <85% of age maximal predicted heart rate (MPHR).
Myocardial infarction was diagnosed based on the presence of chest pain and either serial ECG changes of AMI or cardiac isoenzymes that were abnormal based on the established norms for our institution. Unstable angina was defined as symptoms consistent with myocardial ischemia plus either an early positive IETT (<80% MPHR) or, when the IETT was nondiagnostic, a positive stress nuclear scintigram or stress echocardiogram.
Study patients were not consecutive, and study referral and subsequent disposition after IETT was by the attending physician to whom the patient was referred. A minority of potentially eligible patients were referred for the study due to restrictions on availability of immediate treadmill testing and reluctance of many physicians to refer patients. Follow-up was attempted for a minimum of six months. This protocol was approved by the Human Studies Committee of the University of California, Davis.
Statistical analyses were by two-tailed Students ttest for continuous data and chi-square for categorical data. Differences were considered significant if p < 0.05.
One hundred patients underwent IETT between September 1993 and February 1998 (Table 1). There were no significant differences between men and women except for hypertension, which was more common in women. Twenty-nine patients had prior coronary artery bypass grafting, 25 patients had prior percutaneous transluminal coronary angioplasty, 15 had significant disease documented on coronary arteriography (>50% stenosis in a coronary artery or major branch vessel), 17 patients had a history of myocardial infarction and one had an abnormal nuclear scintigram. In 14 patients, records could not be obtained documenting coronary disease, but the patients gave a history consistent with coronary events believed to be valid by the referring physician and the investigators.
Results of IETT
Median time from hospital admission to IETT was <1 h. The minority of patients had positive (n = 23) ExECGs, and an equal number had negative (n = 38) and nondiagnostic (n = 39) tests.
Complications and coronary events
There were no complications of IETT. Ten patients (10%) were diagnosed with a coronary event. Of this group, 7 patients had early positive IETT (<80% MPHR); 5 patients were diagnosed with unstable angina, and a final diagnosis of non–Q wave AMI was made in 2 patients. The remaining three patients had nondiagnostic IETT and were diagnosed with unstable angina based on findings of imaging studies.
Further acute evaluation
Of the 7 patients (7%) with early positive ExECGs, 1 patient refused further evaluation, 1 patient with previously angiographically documented two-vessel disease had a negative myocardial perfusion scintigram, 3 patients had cardiac catheterization and 2 patients had non–Q wave AMI. Of the 3 patients undergoing cardiac catheterization, 2 underwent angioplasty and 1 patient had inoperable disease. Of the 16 patients (16%) with late positive ExECGs (>80% MPHR), 5 patients underwent further testing, with one low risk positive stress nuclear scintigram, two negative stress echocardiograms (1 patient with a prior angioplasty and one with a prior coronary artery bypass graft) and two cardiac catheterizations demonstrating significant but noncritical disease.
The two patients with non–Q wave infarctions were 59- and 77-year-old men. The patients had stress-induced ST-segment depression at 59% and 77% MPHR which resolved spontaneously during postexercise recovery. One patient had chest pain during stress and one did not. After the early positive stress tests, blood that had been held in the emergency department was sent for creatine kinase assay. Initial enzyme elevation was consistent with onset of AMI before presentation and did not appear to be temporally related to IETT. Serial ECGs revealed non–Q wave myocardial infarctions with only nonspecific ST-segment and T-wave changes (unchanged from prior tracings in the case of one patient). Both patients had a history of coronary artery bypass grafting, and both patients underwent subsequent cardiac catheterization, which revealed significant disease with a single culprit vessel identified and treated with angioplasty and stent placement. Hospital course was unremarkable. One patient returned within six months with recurrent chest pain, had a positive IETT and was found to have progression of disease in an unrelated vessel that was treated with angioplasty and stent placement.
Nondiagnostic exercise tests
Thirty-nine patients (39%) had nondiagnostic tests. Twenty-two were discharged immediately after IETT without further inpatient evaluation (mean MPHR 74%). Eleven were discharged after an 8-h observation period in the telemetry unit with serial ECGs and three sets of fractionated creatine kinase enzymes (mean MPHR 75%). Six patients (mean MPHR 66%) underwent further inpatient evaluation with two negative and two positive stress nuclear scintigrams (a small inferior reversible defect in one patient and a small anteroseptal reversible defect in the other patient), one coronary arteriogram without critical disease and one positive stress echocardiogram with subsequent coronary arteriography and angioplasty. No AMIs were diagnosed.
Negative exercise tests
Of the 38 patients (38%) with negative IETT, 33 were discharged directly after their tests, and 5 completed an 8-h observation in the telemetry unit with serial creatine kinase enzyme assays. No AMIs were found.
Six-month follow-up data were available in 80 patients (80%), and there were no adverse events in this group. The outcome is summarized in Table 2and is correlated with hospital outcome by patient subgroup. One patient, a 62-year-old woman with a negative IETT, underwent elective coronary arteriography, angioplasty and subsequently coronary artery bypass graft for recurrent chest pain. No follow-up was attainable in 20 patients, but review of county coroner records revealed no mortalities in this locale for this group. Review of the social security death registry located one patient who died three years later in another state. These 20 patients were not significantly different (p > 0.05) from the overall patient population with respect to age, gender, percentage of age MPHR attained during IETT or cardiac risk factors. Of these 20 patients, 6 had positive tests, 7 had negative tests and 7 had nondiagnostic tests.
Our study demonstrates that selected patients with coronary artery disease presenting to the emergency room with chest pain can be managed successfully with IETT, resulting in risk stratification of this population into a group that can be discharged and a group that may benefit from admission. All of our patients had been referred from the emergency department physician to the CPER service for evaluation of a suspected coronary event. The patients studied were a selected group in whom low risk was defined by an ECG demonstrating no acute ischemia or injury and no clinically apparent disease that would preclude exercise treadmill testing as outlined in the Methods section.
Our approach is based on evidence that low risk patients with a suspected coronary event can be identified clinically by gender, age, symptoms, past history of CAD and ECG. Lee et al. (13)reported an event rate of 5% in patients with normal ECGs on presentation to the emergency department for acute chest pain. Brush et al. (14)demonstrated that a normal initial ECG identified low risk patients with a life-threatening complication rate of only 0.6%, whereas life-threatening events were 23 times higher in patients with abnormal initial ECGs.
Based on published algorithms for identifying low risk patients (6,13), we predicted that 6% of our patients would have had a coronary event. Our finding that 10 patients (10%) had a coronary event confirms the clinical efficacy of these risk stratification approaches. It is important to note that a diagnosis of unstable angina unassociated with evidence of ischemia in the resting ECG defines a low risk group as opposed to high risk unstable angina patients with ECG evidence of ischemia (13,15,18,19).
Evolution of immediate treadmill testing
Innovative approaches to the management of low risk patients with chest pain have been evolving for more than a decade (20). A number of new noninvasive methods are undergoing investigation for their potential role in this problem. At this time, each has demonstrated advantages and limitations. Immediate exercise treadmill testing of selected low risk patients presenting to the emergency room with chest pain is a management strategy that we have utilized to stratify patient risk and thereby identify a group requiring hospital admission and a group that can be discharged (16,17,19).
In our initial study (16), we excluded patients with known CAD. This study suggests that IETT in selected coronary disease patients utilizing a modified Bruce protocol is safe even though there is the possibility of exercising patients with a non–Q wave AMI. In our two patients with non–Q wave AMI, creatine kinase had not been measured before IETT. Our protocol has been amended and currently calls for measurement of a single serum creatine kinase or troponin I before IETT to further reduce the possibility of exercising patients with non–Q wave AMI.
Disposition of patients
In our study, 55 patients (55%) were discharged immediately after IETT, 27 (27%) after an 8-h observation period and 18 (18%) after further evaluation. Clearly, patients with early positive IETT require admission and patients with negative IETT do not. Further investigation will be required to determine how best to manage patients with late positive tests and nondiagnostic tests. In this regard, we have found in prior studies of patients without CAD that a nondiagnostic IETT associated with a heart rate >75% of MPHR identifies low risk (17). The current study suggests that a heart rate >80% MPHR identifies low risk in patients with known CAD.
Our study has several limitations, and most important of which is the relatively small number of patients evaluated. We do not believe that nonconsecutive patient enrollment is a major limitation because, by the nature of the study, the patients were selected. Furthermore, the incidence of events was as predicted by prior data (13). This suggests that the patients studied represent a low risk group with coronary disease that is typical of those admitted during daily practice. In these other studies (6,13), unstable angina was diagnosed solely on clinical suspicion. Our study added the requirement of noninvasive testing to support the clinical diagnosis. Not all of our patients had CAD confirmed by medical records, but this problem is also common in practice and is not a significant limitation of this study as the coronary event rate was higher than predicted. Further limitations include the fact that not all patients completed an 8-h observation period with three sets of creatine kinase subfractions, and it is possible that non–Q wave AMIs were missed. However, follow-up data revealed no early untoward events, further confirming low risk. Finally, follow-up at six months was not possible in all patients, but, as noted, this subgroup was low risk in profile, did not differ significantly from the entire group and included no reported mortalities in our locale or in the social security death registry.
The authors gratefully acknowledge the assistance of Alisha Willis and Elani Streja in the preparation of the manuscript.
- acute myocardial infarction
- coronary artery disease
- chest pain emergency room
- exercise electrocardiogram
- immediate exercise treadmill testing
- maximal predicted heart rate
- Received June 15, 1998.
- Revision received January 25, 1999.
- Accepted February 15, 1999.
- American College of Cardiology
- Roberts R,
- Kleiman N.S
- Horowitz R.S,
- Morganroth J,
- Parrotto C,
- Chen C.C,
- Soffer J,
- Pauletto F.J
- Sabia P,
- Afrookteh A,
- Touchstone D.A,
- Keller M.W,
- Esquivel L,
- Kaul S
- Henneman P.L,
- Mena I.G,
- Rothstein R.J,
- Garrett K.B,
- Pleyto A.S,
- French W.J
- Gazes P.C,
- Mobley E.M Jr.,
- Favis H.M Jr.,
- Duncan R.C,
- Humphries G.B
- Braunwald E,
- Jones R.H,
- Mark D.B,
- et al.