Methods

Patients were identified by review of the University of California (Davis) Medical Center CPU database. The study group comprised consecutive low-risk women aged 30 to 50 years admitted to the CPU from February 1995 to November 2006 with symptoms suggestive of ACS. Patients were excluded if the had (1) history of cardiovascular disease (ACS, stroke, ischemic noninvasive stress test, obstructive coronary artery disease on angiography, and heart failure), (2) serious co-morbidity, (3) diabetes mellitus, (4) history of smoking, or (5) cocaine or amphetamine use within 6 months. Low risk for ACS was based on clinical stability (no arrhythmias or hemodynamic dysfunction), normal results (<0.5 mm ST depression and no T-wave inversion) on electrocardiogram (ECG), and negative cardiac injury markers (troponin-I or creatine kinase-MB). The latter marker was used early in the study in only a small number of patients. Patients fulfilling low-risk criteria were transferred from the ED to the CPU for further evaluation, which could include exercise treadmill test, stress echocardiography, myocardial stress scintigraphy (MSS), or no predischarge test. Length of stay (LOS) from admission to the CPU to time of hospital discharge was determined in all patients. This study received approval from our institutional review board.

After initial negative evaluation in the CPU, patients were discharged either with or without a predischarge cardiac test at the discretion of the attending physician. No patients were scheduled for early (24 to 72 hours) cardiac testing after discharge. Treadmill testing was performed according to the Bruce or modified Bruce protocols. Exercise end points included symptoms (e.g., chest pain, undue dyspnea, and dizziness), ECG evidence of myocardial ischemia (1.0-mm horizontal ST shift 80 ms after the J point), 10 mm Hg decrease in systolic blood pressure, ≥3 consecutive ventricular extrasystoles, or a sustained supraventricular tachyarrhythmia. The criterion for a test to be positive for ischemia was the previously noted exercise-induced ST alteration; a nondiagnostic test was defined by absence of ECG ischemia at a heart rate <85% of age-predicted maximum.

In patients who could not exercise or who had baseline ECG changes precluding interpretation of the exercise ECG, stress echocardiography or MSS was used. MSS was performed by single-photon emission computed tomography. Positive MSS comprised a new stress-induced myocardial perfusion defect. For patients undergoing stress echocardiography (exercise or dobutamine), the criterion for a test to be positive was a new stress-induced left ventricular wall motion abnormality. Coronary angiography was performed in multiple projections according to standard techniques. Significant coronary artery disease was defined as >50% stenosis in ≥1 major coronary arteries. We reviewed all patients’ medical records to determine clinical status after discharge. Follow-up also included telephone interviews of patients or family members. Clinical end points were all-cause mortality and cardiovascular events (myocardial infarction, stroke, and revascularization).

Results

The study group included 403 patients ( Table 1 ). Two thirds were aged 40 to 49 years, the largest ethnic group was white, and >80% had 0 to 1 cardiac risk factors. Cardiac testing was performed in 48% (n = 192) of patients: treadmill ( Table 2 ) 84% (n = 162); stress imaging 14% (n = 26); coronary angiography 2% (n = 4). Predischarge testing was not performed in 52% (n = 211) of patients. There was no difference in age or ethnicity between patients who underwent predischarge testing and those who did not (42.2 vs 42.0 years). However, those who were tested had statistically significantly greater number of risk factors (1.0 vs 0.6, p <0.0001), number of serial cardiac injury markers measured (2.1 vs 1.9, p <0.04), number of 12-lead ECGs (1.9 vs 1.7, p <0.04), and LOS (10 vs 6 hours, p <0.0001).

Treadmill test results were negative in a large majority of patients, <20% were positive, and the remainder was nondiagnostic ( Table 2 ). All but 2 patients with negative treadmill tests were discharged directly from the CPU. These 2 patients were discharged after further evaluation by negative cardiac stress imaging. Of the patients with positive treadmill tests (n = 28, 17%), 11 had further evaluation (stress imaging 10, coronary angiography 2 [1 patient had both imaging and angiography]), which was negative in all, and they were discharged directly from the CPU. In the remaining 17 patients, the treadmill test results were considered false positive and/or low risk, and these patients were also directly discharged. Of the 19 nondiagnostic treadmill tests, 18 were considered low risk, and these patients were directly discharged from the CPU without further testing. The nineteenth patient in this group was discharged after negative MSS. Approximately half of the patients who underwent treadmill testing had a functional capacity of 8 to 14 METs and almost 90% had no exercise-induced chest pain.

Stress imaging as the initial cardiac test was performed by stress echocardiography (n = 19) or MSS (n = 7). All except 2 of these tests were negative, resulting in discharge. Two tests were positive for reversible perfusion defects, 1 of which was a false positive based on normal coronary angiography and 1 was considered low risk because the perfusion defect was small. Coronary angiography result was normal in 7 patients, and they were directly discharged. In 4 patients, this was the only test performed based on attending physician discretion. It was applied in an additional 3 patients for further evaluation of positive noninvasive tests (2 positive treadmill tests, 1 positive MSS).

LOS varied from <2 to >24 hours. Mean LOS in patients without cardiac stress testing was significantly shorter than in those with testing (6 vs 10 hours, p <0.0001). In patients who underwent cardiac testing, LOS was <2 hours in 32% and ≤12 hours in 66%. In patients who did not undergo cardiac stress testing, LOS was <2 hours in 50% and ≤12 hours in 84%.

Follow-up status was obtained in 87% of patients (n = 349) at 30 days after discharge and in 80% of patients (n = 321) at 6 months. There were no major adverse outcomes in any of the patients. Thirteen percentage of patients (n = 54) were lost to follow-up, and in 20%, follow-up was <6 months. Although age was greater in the followed group (p <0.05), this difference was not clinically important (followed 42.3 vs no followed 41.3 years). Average number of cardiac risk factors in the followed and no followed groups was <1 in each.