American College of Cardiology

GIBBONS ET AL., MANAGEMENT OF PATIENTS WITH CHRONIC STABLE ANGINA UPDATE
http://www.acc.org/clinical/guidelines/stable/update_index.htm

ACC/AHA 2002 Guideline Update for the Management of Patients With Chronic Stable Angina

A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for the Management of Patients With Chronic Stable Angina)

This is a Guideline Update of the 1999 Chronic Stable Angina Guidelines. To highlight the changes, deleted text is indicated by strikeout, and revised text is presented in brown. A clean version of the document, with changes fully incorporated, is available for download and print.

II. Initial Evaluation and Management

A. Clinical Assessment

Patients with suspected ACS must be evaluated rapidly. Decisions made based on the initial evaluation have substantial clinical and economic consequences⁽¹¹⁾. When the patient first makes contact with the medical care system, a critical decision must be made about where the evaluation will take place. The physician then must place the evaluation in the context of 2 critical questions: Are the symptoms a manifestation of an ACS? If so, what is the prognosis? The answers to these 2 questions lead logically to a series of decisions about where the patient will be managed, what medications will be prescribed, and whether an angiographic evaluation will be required.

Given the large number of patients with symptoms compatible with ACS, the heterogeneity of the population, and the clustering of events shortly after the onset of symptoms (Figure 3), a strategy for the initial evaluation and management is essential. Healthcare providers may be informed about signs and symptoms of ACS over the telephone or in person (and perhaps in the future over the Internet). The objectives of the initial evaluation are first to identify signs of immediate life-threatening instability and then to ensure that the patient is moved rapidly to the most appropriate environment for the level of care needed based on diagnostic criteria and an estimation of the underlying risk of specific negative outcomes.

Recommendation for Telephone Triage
Class I

Patients with symptoms that suggest possible ACS should not be evaluated solely over the telephone but should be referred to a facility that allows evaluation by a physician and the recording of a 12-lead ECG. (Level of Evidence: C)

Health practitioners frequently receive telephone calls from patients who are concerned that their symptoms may reflect heart disease. Most such calls regarding chest discomfort of possible cardiac origin in patients without known CAD do not represent an emergency; rather these patients usually seek reassurance that they do not have heart disease or that there is little risk due to their symptoms. Despite the frequent inclination to dismiss such symptoms over the telephone, physicians should advise patients with possible accelerating angina or angina at rest that such an evaluation cannot be carried out solely via the telephone. This advice is essential because of the need for a physical examination and an ECG and the potential importance of blood tests to measure cardiac markers.

Patients with known CAD—including those with chronic stable angina or recent MI or who have had coronary artery bypass graft surgery (CABG) or a PCI—who contact a physician because of worsening or recurrence of symptoms should be urged to go directly to an ED equipped to perform prompt reperfusion therapy. Alternatively, they may enter the emergency medical services system directly by calling 9-1-1. Patients who have recently been evaluated and who are calling for advice regarding modification of medication as part of an ongoing treatment plan represent exceptions.

Even in the most urgent subgroup of patients who present with acute-onset chest pain, there usually is adequate time for transport to an environment in which they can be evaluated and treated ⁽¹²⁾. In a large study of consecutive patients with chest pain suspected to be of cardiac etiology who were transported to the ED via ambulance, one third had a final diagnosis of AMI, one third had a final diagnosis of UA, and one third had a final diagnosis of a noncardiac cause. Only 1.5% of these patients developed cardiopulmonary arrest before arrival at the hospital or in the ED ⁽¹³⁾. These findings suggest that patients with acute chest pain might be better served by transport to an adequately staffed and equipped ED than by sending them to a less well staffed and equipped facility, thereby compromising the quality of the care environment in an attempt to shorten the initial transport time.

Patients must retain the ultimate responsibility for deciding whether to seek medical attention and, if so, in what environment. The physician cannot be expected to assume responsibility for a patient with a potentially serious acute cardiac disorder who does not present in person for urgent evaluation and declines after being advised to do so. Physicians should be cautious not to inappropriately reassure patients who are inclined not to seek further medical attention.

1. ED or Outpatient Facility Presentation
Recommendation
Class I

Patients with a suspected ACS with chest discomfort at rest for greater than 20 min, hemodynamic instability, or recent syncope or presyncope should be strongly considered for immediate referral to an ED or a specialized chest pain unit. Other patients with a suspected ACS may be seen initially in an ED, a chest pain unit, or an outpatient facility. (Level of Evidence: C)

Although no data are available that compare outcome as a function of the location of the initial assessment, this recommendation is based on evidence that symptoms and signs of an ACS may lead to a clinical decision that requires a sophisticated level of intervention. When symptoms have been unremitting for greater than 20 min, the possibility of STEMI must be considered. Given the strong evidence for a relationship between delay in treatment and death ^(14-16), an immediate assessment that includes a 12-lead ECG is essential. Patients who present with hemodynamic instability require an environment in which therapeutic interventions can be provided, and for those with presyncope or syncope, the major concern is the risk of sudden death. Such patients should be encouraged to seek emergency transportation when it is available. Transport as a passenger in a private vehicle is an acceptable alternative only if the wait for an emergency vehicle would impose a delay of greater than 20 to 30 min.

Patients without any of these high-risk features may be seen initially in an outpatient facility.

2. Questions to be Addressed at the Initial Evaluation
The initial evaluation should be used to provide information about the diagnosis and prognosis. The attempt should be made to simultaneously answer 2 questions:

• What is the likelihood that the signs and symptoms represent ACS secondary to obstructive CAD (Table 5)?
• What is the likelihood of an adverse clinical outcome (Table 6)? Outcomes of concern include death, MI (or recurrent MI), stroke, heart failure, recurrent symptomatic ischemia, and serious arrhythmia.

For the most part, the answers to these questions form a sequence of contingent probabilities. Thus, the likelihood that the signs and symptoms represent ACS is contingent on the likelihood that the patient has underlying CAD. Similarly, the prognosis is contingent on the likelihood that the symptoms represent acute ischemia.

B. Early Risk Stratification

Recommendations for Early Risk Stratification
Class I

1. A determination should be made in all patients with chest discomfort of the likelihood of acute ischemia caused by CAD as high, intermediate, or low. (Level of Evidence: C)
2. Patients who present with chest discomfort should undergo early risk stratification that focuses on anginal symptoms, physical findings, ECG findings, and biomarkers of cardiac injury. (Level of Evidence: B)
3. A 12-lead ECG should be obtained immediately (within 10 min) in patients with ongoing chest discomfort and as rapidly as possible in patients who have a history of chest discomfort consistent with ACS but whose discomfort has resolved by the time of evaluation. (Level of Evidence: C)
4. Biomarkers of cardiac injury should be measured in all patients who present with chest discomfort consistent with ACS. A cardiac-specific troponin is the preferred marker, and if available, it should be measured in all patients. CK-MB by mass assay is also acceptable. In patients with negative cardiac markers within 6 h of the onset of pain, another sample should be drawn in the 6- to 12-h time frame (e.g., at 9 h after the onset of symptoms). (Level of Evidence: C)

Class IIa

For patients who present within 6 h of the onset of symptoms, an early marker of cardiac injury (e.g., myoglobin or CK-MB subforms) should be considered in addition to a cardiac troponin. (Level of Evidence: C)

Class IIb

C-reactive protein (CRP) and other markers of inflammation should be measured. (Level of Evidence: B)

Class III

Total CK (without MB), aspartate aminotransferase (AST, SGOT), beta-hydroxybutyric dehydrogenase, and/or lactate dehydrogenase should be the markers for the detection of myocardial injury in patients with chest discomfort suggestive of ACS. (Level of Evidence: C)

1. Estimation of the Level of Risk
The medical history, physical examination, ECG, and biochemical cardiac marker measurements in patients with symptoms suggestive of ACS at the time of the initial presentation can be integrated into an estimation of the risk of death and nonfatal cardiac ischemic events. The latter include new or recurrent MI, recurrent UA, disabling angina that requires hospitalization, and/or urgent coronary revascularization. Estimation of the level of risk is a multivariable problem that cannot be accurately quantified with a simple table; therefore, Tables 5 and 6 are meant to be illustrative of the general relationships between clinical and ECG findings and the categorization of patients into those at a low, an intermediate, or a high risk of events.

2. Rationale for Risk Stratification
Because patients with ischemic discomfort at rest as a group are at an increased risk of cardiac death and nonfatal ischemic events, an assessment of the prognosis often sets the pace of the initial evaluation and treatment. An estimation of risk is useful in 1) selection of the site of care (coronary care unit, monitored step-down unit, or outpatient setting) and 2) selection of therapy, especially platelet glycoprotein (GP) IIb/IIIa inhibitors (see Section III. B) and coronary revascularization (see Section IV). For all modes of presentation of an ACS, a strong relationship exists between indicators of the likelihood of ischemia due to CAD and prognosis (Tables 5 and 6). Patients with a high likelihood of ischemia due to CAD are at a greater risk of an untoward cardiac event than are patients with a lower likelihood of CAD. Therefore, an assessment of the likelihood of CAD is the starting point for the determination of prognosis in patients who present with symptoms suggestive of an ACS. Other important elements for prognostic assessment are the tempo of the patient's clinical course, which relates to the short-term risk of future cardiac events, principally AMI, and the patient's likelihood of survival should an AMI occur.

Patients may present with ischemic discomfort but without ST-segment elevation on the 12-lead ECG in a variety of clinical scenarios, including no known prior history of CAD, a prior history of stable CAD, soon after MI, and after myocardial revascularization with CABG or PCI ^(7,17,18). As a clinical syndrome, ischemic discomfort without ST-segment elevation (UA and NSTEMI) shares ill-defined borders with severe chronic stable angina, a condition associated with lower risk, and with STEMI, a presentation with a higher risk of early death and cardiac ischemic events. This fact is illustrated by data from the Duke Cardiovascular Databank that describe the rate of cardiac death in 21,761 patients treated for CAD without interventional procedures at Duke University Medical Center between 1985 and 1992 and that were published in the AHCPR-NHLBI guidelines ⁽¹⁾, now supplemented with data from large clinical trials in ACS ⁽¹⁰⁾ (Figure 3). The highest risk of cardiac death was at the time of presentation, and the risk declined so that by 2 months, mortality rates for patients with ACS were at the same level as those for patients with chronic stable angina. Data from randomized controlled trials of patients with UA/NSTEMI have also shown that the rate of nonfatal cardiac ischemic events such as recurrent MI and recurrent angina is highest during the initial hospitalization and declines thereafter ^(4,10,19-21).

Two large clinical trials, Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) ⁽¹⁰⁾ and Efficacy and Safety of Subcutaneous Enoxaparin in Non-Q wave Coronary Events (ESSENCE) ⁽²²⁾, have evaluated the clinical and ECG characteristics associated with an increased risk of death and nonfatal MI in 24,774 patients with UA/NSTEMI. The critical clinical features associated with an increased risk of death were age (greater than 65 years), presence of positive markers for myocardial necrosis on admission, lighter weight, more severe (CCS class III or IV) chronic angina before the acute admission, rales on physical examination, and ST-segment depression on the admission ECG. In the PURSUIT trial, either tachycardia or bradycardia and lower blood pressure were associated with a higher risk of death or MI. These findings allow the stratification of patients with UA/NSTEMI into those at higher risk and those at lower risk.

3. The History
Patients with suspected UA/NSTEMI may be divided into those with and those without a history of documented CAD. Particularly when the patient does not have a known history of CAD, the physician must determine whether the patient's presentation, with its constellation of specific symptoms and signs, is most consistent with chronic ischemia, with acute ischemia, or with an alternative disease process. The 5 most important factors derived from the initial history that relate to the likelihood of ischemia due to CAD, ranked in the order of importance, are 1) the nature of the anginal symptoms, 2) prior history of CAD, 3) sex, 4) age, and 5) the number of traditional risk factors present ^(23-25).

a. Anginal Symptoms
The characteristics of angina are described in the ACC/AHA/ACP-ASIM Guidelines for the Management of Patients With Chronic Stable Angina ⁽²⁶⁾. Angina is characterized as a deep, poorly localized chest or arm discomfort that is reproducibly associated with physical exertion or emotional stress and is relieved promptly (i.e., less than 5 min) with rest and/or the use of sublingual nitroglycerin (NTG) (Table 5). Patients with UA may have discomfort that has all of the qualities of typical angina except that the episodes are more severe and prolonged, may occur at rest, or may be precipitated by less exertion than previously. Some patients may have no chest discomfort but present solely with jaw, neck, ear, arm, or epigastric discomfort. If these symptoms have a clear relationship to exertion or stress or are relieved promptly with NTG, they should be considered equivalent to angina. Occasionally, such "anginal equivalents" that occur at rest are the mode of presentation of a patient with UA, but without the exertional history, it may be difficult to recognize the cardiac origin. Other difficult presentations of the patient with UA include those without any chest (or equivalent) discomfort. Isolated unexplained new-onset or worsened exertional dyspnea is the most common anginal equivalent symptom, especially in older patients; others include nausea and vomiting, diaphoresis, and unexplained fatigue. Elderly patients, especially women with ACS, often present with atypical angina.

Features that are not characteristic of myocardial ischemia include the following:

• Pleuritic pain (i.e., sharp or knife-like pain brought on by respiratory movements or cough)
• Primary or sole location of discomfort in the middle or lower abdominal region
• Pain that may be localized at the tip of 1 finger, particularly over the left ventricular (LV) apex
• Pain reproduced with movement or palpation of the chest wall or arms
• Constant pain that lasts for many hours
• Very brief episodes of pain that last a few seconds or less
• Pain that radiates into the lower extremities

Documentation of the evaluation of a patient with suspected UA/NSTEMI should include the physician's opinion of whether the discomfort is in 1 of 3 categories: high, intermediate, or low likelihood of acute ischemia caused by CAD (Table 5).

Although typical characteristics substantially raise the probability of CAD, features not characteristic of chest pain, such as sharp stabbing pain or reproduction of pain on palpation, do not exclude the possibility of ACS. In the Multicenter Chest Pain Study, acute ischemia was diagnosed in 22% of patients who presented to the ED with sharp or stabbing pain and in 13% of patients with pain with pleuritic qualities. Furthermore, 7% of patients whose pain was fully reproduced with palpation were ultimately recognized to have ACS ⁽²⁷⁾. The Acute Cardiac Ischemia Time-Insensitive Predictive Instrument (ACI-TIPI) project ^(28,29) found that older age, male sex, the presence of chest or left arm pain, and the identification of chest pain or pressure as the most important presenting symptom all increased the likelihood that the patient was experiencing acute ischemia.

b. Demographics and History in Diagnosis and Risk Stratification
In most studies of ACS, a prior history of MI has been associated not only with a high risk of obstructive CAD ⁽³⁰⁾ but also with an increased risk of multivessel CAD.

There are differences in the presentations of men and women with ACS (see Section VI. A). A smaller percentage of women than men present with STEMI, and of the patients who present without ST-segment elevation, fewer women than men have MIs ⁽³¹⁾. Women with suspected ACS are less likely to have CAD than are men with a similar clinical presentation, and when CAD is present in women, it tends to be less severe. If STEMI is present, the outcome in women tends to be worse even when adjustment is made for the older age and greater comorbidity of women. However, the outcome for women with UA is significantly better than the outcome for men, and the outcomes are similar for men and women with NSTEMI ^(32,33).

Older patients (see Section VI. D) have increased risks of both underlying CAD ^(34,35) and multivessel CAD; furthermore, they are at higher risk for an adverse outcome than are younger patients. The slope of the increased risk is steepest beyond age 70. This increased risk is related in part to the greater extent and severity of underlying CAD and the more severe LV dysfunction in older patients, but age itself appears to exert an independent prognostic risk as well, perhaps because of comorbidities. Elderly patients are also more likely to have atypical symptoms on presentation.

In patients with symptoms of possible ACS, some of the traditional risk factors for CAD (e.g., hypertension, hypercholesterolemia, cigarette smoking) are only weakly predictive of the likelihood of acute ischemia ^(29,36) and are far less important than are symptoms, ECG findings, and cardiac markers. Therefore, the presence or absence of these traditional risk factors ordinarily should not be used to determine whether an individual patient should be admitted or treated for ACS. However, the presence of these risk factors does appear to relate to poor outcomes in patients with established ACS. Although a family history of premature CAD raises interesting issues of the genetic contribution to the development of this syndrome, it has not been a useful indicator of diagnosis or prognosis in patients evaluated for possible symptoms of ACS. However, several of these risk factors have important prognostic and therapeutic implications. Diabetes and the presence of extracardiac (peripheral or carotid) arterial disease are major risk factors for poor outcome in patients with ACS (see Section VI. B). For both ST-segment elevation ⁽³⁷⁾ and non-ST-segment elevation ACS ⁽¹⁰⁾, patients with these conditions have a significantly higher mortality rate and risk of acute heart failure. For the most part, this increase in risk is due to a greater extent of underlying CAD and LV dysfunction, but in many studies, diabetes carries prognostic significance over and above these findings. Similarly, a history of hypertension is associated with an increased risk of poor outcome.

Surprisingly, current smoking is associated with a lower risk of death in the setting of ACS ^(38-40), predominantly because of the less severe underlying CAD. This "smokers' paradox" seems to represent a tendency for smokers to develop thrombi on less severe plaques and at an earlier age than nonsmokers.

Cocaine use has been implicated as a cause of ACS, presumably due to the ability of this drug to cause coronary vasospasm and thrombosis in addition to its direct effects on heart rate and arterial pressure and its myocardial toxic properties (see Section VI. E). It is important to inquire about the use of cocaine in patients with suspected ACS, especially younger patients (less than 40 years).

c. The Estimation of Early Risk at Presentation
Risk has been assessed using multivariable regression techniques in patients presenting with UA/NSTEMI in several large clinical trials. These have not yet been validated in large registries of such patients. Boersma et al. analyzed the relation between baseline characteristics and the incidence of death as well as the composite of death or myocardial (re)infarction at 30 days ⁽⁵¹⁶⁾. The most important baseline features associated with death were age, heart rate, systolic blood pressure, ST-segment depression, signs of heart failure, and elevation of cardiac markers. From this analysis, a simple risk estimation score that should be useful in clinical decisionmaking was developed ⁽⁵¹⁶⁾.

Antman et al. developed a 7-point risk score (age greater than 65 years, more than 3 coronary risk factors, prior angiographic coronary obstruction, ST-segment deviation, more than 2 angina events within 24 hours, use of aspirin within 7 days, and elevated cardiac markers) ⁽⁵¹⁷⁾. The risk of developing an adverse outcome—death, (re)infarction, or recurrent severe ischemia requiring revascularization—ranged from 5% to 41% with the "TIMI (Thrombolysis In Myocardial Infarction) risk score" defined as the sum of the individual prognostic variables. The score was derived from data in the TIMI 11B trial ⁽¹⁷⁰⁾ and has been validated in 3 additional trials—ESSENCE ⁽¹⁶⁹⁾, TACTICS-TIMI 18 ⁽⁵¹⁸⁾, and PRISM-PLUS ⁽²¹⁾. Among patients with UA/NSTEMI, there is a progressively greater benefit from newer therapies such as low-molecular-weight heparin ^(169,170), platelet GP IIb/IIIa inhibition ⁽⁵¹⁹⁾, and an invasive strategy ⁽⁵¹⁸⁾ with increasing risk score.

4. Noncardiac Causes of Exacerbation of Symptoms Secondary to Myocardial Ischemia
Recommendation
Class I

The initial evaluation of the patient with suspected ACS should include a search for noncoronary causes that could explain the development of symptoms. (Level of Evidence: C)

Information from the initial history, physical examination, and ECG (Table 5) will enable the physician to recognize and exclude from further assessment patients classified as "not having ischemic discomfort." This includes patients with noncardiac pain (e.g., musculoskeletal discomfort, esophageal discomfort) or cardiac pain not caused by myocardial ischemia (e.g., acute pericarditis). The remaining patients should undergo a more complete evaluation of secondary causes of UA that might alter management. This evaluation should include a physical examination for evidence of other cardiac disease, an ECG to screen for arrhythmias, measurement of body temperature and blood pressure, and determination of hemoglobin or hematocrit. Cardiac disorders other than CAD that may cause myocardial ischemia include aortic stenosis and hypertrophic cardiomyopathy. In secondary angina, factors that increase myocardial oxygen demand or decrease oxygen delivery to the heart may provoke or exacerbate ischemia in the presence of significant underlying CAD. Previously unrecognized gastrointestinal bleeding is a common secondary cause of worsened CAD and the development of ACS symptoms due to anemia. Acute worsening of chronic obstructive pulmonary disease (COPD) (with or without superimposed infection) may lower oxygen saturation levels sufficiently to intensify ischemic symptoms in patients with CAD. Evidence of increased cardiac oxygen demand can be judged from the presence of fever, signs of hyperthyroidism, sustained tachyarrhythmias, or markedly elevated blood pressure. Another cause of increased myocardial oxygen demand is arteriovenous (AV) fistula in patients receiving dialysis.

The majority of patients seen in the ED with symptoms of possible ACS will be judged after their workup to not have a cardiac problem. A recent clinical trial of a predictive instrument evaluated 10,689 patients with suspected ACS ⁽¹¹⁾. To participate, patients were required to be greater than 30 years old with a chief symptom of chest, left arm, jaw, or epigastric pain or discomfort; shortness of breath; dizziness; palpitations; or other symptoms suggestive of acute ischemia. After the evaluation, 7,996 patients (75%) were deemed not to have acute ischemia.

5. Assessment of Risk of Death in Patients with UA/NSTEMI
In patients who meet the diagnostic criteria for UA/NSTEMI, the recent tempo of ischemic symptoms is the strongest predictor of risk of death. The AHCPR guidelines "Unstable Angina: Diagnosis and Management" identified low-risk patients as those without rest or nocturnal angina and with a normal or an unchanged ECG ⁽¹⁾. High-risk patients were identified as those with pulmonary edema; ongoing rest pain greater than 20 min in duration; angina with S₃ gallop, rales, or new or worsening mitral regurgitation (MR) murmur; hypotension; or dynamic ST-segment change greater than or equal to 1 mm. Patients without low- or high-risk features were termed to be at "intermediate risk."

These simple clinical criteria were prospectively tested in a consecutive sample of patients who presented with symptoms suggestive of ACS ⁽⁴¹⁾. After prescreening was conducted to exclude patients with AMI or cardiac arrest, patients receiving thrombolytic therapy, and patients diagnosed as having noncardiac conditions, only 6% of the remaining patients diagnosed with UA were categorized as being at low risk. This low-risk population experienced no death or MI in the 30 days after the initial presentation to the ED. In contrast, the 30-day mortality rate was 1.2% for patients at intermediate risk and 1.7% for patients deemed at high risk. These observations confirmed the management recommendations made in the earlier guidelines. Patients with low-risk UA can be managed expeditiously as outpatients. Patients with high-risk UA deserve rapid clinical stabilization in an acute-care environment in the hospital. Patients at intermediate risk require individualization of management based on clinical judgment. These patients should usually be admitted to the hospital and require monitoring but do not ordinarily require an intensive care unit.

The tempo of angina is characterized by an assessment of changes in the duration of episodes, their frequency, and the anginal threshold. In particular, it is useful to determine whether the amount of physical or emotional stress that provokes symptoms has declined, whether symptoms are occurring at rest, and whether they awaken the patient from sleep. The integration of these factors into a score can improve the predictions of outcome ^(42,43). Although new-onset angina itself is associated with greater risk than is continued stable angina, most of its contribution to an adverse prognosis is determined by its severity, frequency, and tempo ^(42,44).

Multiple studies have demonstrated that prior MI is a major risk factor for poor outcome in both STEMI and UA/NSTEMI ⁽¹⁰⁾. Patients with symptoms of acute and/or chronic heart failure are also at a substantially higher risk.

a. Physical Examination
The major objectives of the physical examination are to identify potential precipitating causes of myocardial ischemia such as uncontrolled hypertension or thyrotoxicosis and comorbid conditions such as pulmonary disease and to assess the hemodynamic impact of the ischemic event. Every patient with suspected ACS should have his or her vital signs measured (blood pressure in both arms, heart rate, temperature) and undergo a thorough cardiovascular and chest examination. Patients with evidence of LV dysfunction on examination (rales, S₃ gallop) or with acute MR have a higher likelihood of severe underlying CAD and are at a high risk of a poor outcome. Just as the history of extracardiac vascular disease is important, the physical examination of the peripheral vessels can also provide important prognostic information. The presence of bruits or pulse deficits that suggest extracardiac vascular disease (carotid, aortic, peripheral) identifies patients with a higher likelihood of significant CAD.

Elements of the physical examination can be critical in making an important alternative diagnosis in patients with chest pain. In particular, several disorders carry a significant threat to life and function if not diagnosed acutely. Aortic dissection is suggested by pain in the back, unequal pulses, or a murmur of aortic regurgitation. Acute pericarditis is suggested by a pericardial friction rub, and cardiac tamponade may be evidenced by pulsus paradoxus. Pneumothorax is suspected when acute dyspnea, pleuritic chest pain, and differential breath sounds are present.

Recently, the importance of cardiogenic shock in patients with NSTEMI was emphasized. Although most literature on cardiogenic shock has focused on STEMI, the SHould we emergently revascularize Occluded Coronaries for cardiogenic shocK (SHOCK) ⁽⁴⁵⁾, Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO)-II ^(45a), and PURSUIT ⁽¹⁰⁾ trials have found that cardiogenic shock occurs in up to 5% of patients with NSTEMI and that mortality rates are greater than 60%. Thus, hypotension and evidence of organ hypoperfusion constitute a medical emergency in NSTEMI.

6. Tools for Risk Stratification
a. Electrocardiogram
The ECG is critical not only to add support to the clinical suspicion of CAD but also to provide prognostic information that is based on the pattern and magnitude of the abnormalities ^(46-49). A recording made during an episode of the presenting symptoms is particularly valuable. Importantly, transient ST-segment changes (greater than or equal to 0.05 mV) that develop during a symptomatic episode at rest and that resolve when the patient becomes asymptomatic strongly suggest acute ischemia and a very high likelihood of underlying severe CAD. Patients whose current ECG suggests acute CAD can be assessed with greater diagnostic accuracy if a prior ECG is available for comparison (Table 5) ^(50,51).

Although it is imperfect, the 12-lead ECG lies at the center of the decision pathway for the evaluation and management of patients with acute ischemic discomfort (Figure 1, Table 5). The diagnosis of AMI is confirmed with serial cardiac markers in more than 90% of patients who present with ST-segment elevation of greater than or equal to 0.1 mV in greater than or equal to 2 contiguous leads, and such patients should be considered potential candidates for acute reperfusion therapy. Patients who present with ST-segment depression are initially considered to have either UA or NSTEMI; the distinction between the 2 diagnoses is based ultimately on the detection in the blood of markers of myocardial necrosis ^(6,18,52).

Patients with UA and reversible ST-segment depression have an increase in thrombin activity reflected in elevated levels of circulating fibrinopeptides and complex lesions that suggest thrombosis on coronary angiography ⁽⁵³⁾. Up to 25% of patients with NSTEMI and elevated CK-MB go on to develop Q-wave MI, whereas the remaining 75% have non-Q-wave MI. Acute reperfusion therapy is contraindicated for ACS patients without ST-segment elevation, except for those with isolated acute posterior infarction manifested as ST-segment depressions in leads V1 to V3 and/or isolated ST-segment elevation in posterior chest leads ⁽⁵⁴⁾. Inverted T waves may also indicate ischemia or non-Q-wave infarction. In patients suspected on clinical grounds to have ACS, marked (greater than or equal to 0.2 mV) symmetrical precordial T-wave inversion strongly suggests acute ischemia, particularly that due to a critical stenosis of the left anterior descending coronary artery (LAD) ⁽⁵⁵⁾. Patients with this ECG finding often exhibit hypokinesis of the anterior wall and are at high risk with medical treatment ⁽⁵⁶⁾. Revascularization will often reverse both the T-wave inversion and wall motion disorder ⁽⁵⁷⁾. Nonspecific ST-segment and T-wave changes, usually defined as ST-segment deviation of less than 0.05 mV or T-wave inversion of less than or equal to 0.2 mV, are less helpful than the foregoing findings. Established Q waves greater than or equal to 0.04 s are also less helpful in the diagnosis of UA, although by suggesting prior MI, they do indicate a high likelihood of significant CAD. Isolated Q waves in lead III may be a normal finding, especially in the absence of repolarization abnormalities in any of the inferior leads. A completely normal ECG in a patient with chest pain does not exclude the possibility of ACS, because 1% to 6% of such patients eventually are proved to have had an AMI (by definition, an NSTEMI), and greater than or equal to 4% will be found to have UA ^(47,58,59).

The common alternative causes of ST-segment and T-wave changes must be considered. In patients with ST-segment elevation, the diagnoses of LV aneurysm, pericarditis, Prinzmetal's angina, early repolarization, and Wolff-Parkinson-White syndrome should be considered. Central nervous system events and drug therapy with tricyclic antidepressants or phenothiazines can cause deep T-wave inversion.

Several investigators have shown that a gradient of risk of death and cardiac ischemic events can be established based on the nature of the ECG abnormality ^(48,60,61). Patients with ACS and confounding ECG patterns such as bundle-branch block, paced rhythm, or LV hypertrophy are at the highest risk for death, followed by patients with ST-segment deviation (ST-segment elevation or depression); at the lowest risk are patients with isolated T-wave inversion or normal ECG patterns. Importantly, the prognostic information contained within the ECG pattern remains an independent predictor of death even after adjustment for clinical findings and cardiac marker measurements ^(60-63).

In addition to the presence or absence of ST-segment deviation or T-wave inversion patterns as noted earlier, there is evidence that the magnitude of the ECG abnormality provides important prognostic information. Thus, Lloyd-Jones et al. ⁽⁶⁴⁾ reported that the diagnosis of acute non-Q-wave MI was 3 to 4 times more likely in patients with ischemic discomfort who had greater than or equal to 3 ECG leads that showed ST-segment depression and/or ST-segment depression of greater than or equal to 0.2 mV. Investigators from the TIMI III registry ⁽⁶⁰⁾ reported that the 1-year incidence of death or new MI in patients with greater than or equal to 0.05-mV ST-segment deviation was 16.3% compared with 6.8% for patients with isolated T-wave changes and 8.2% for patients with no ECG changes.

Because a single 12-lead ECG recording provides only a snapshot view of a dynamic process, the usefulness of obtaining serial ECG tracings or performing continuous ST-segment monitoring was studied ⁽⁴⁶⁾. Although serial ECGs increase the ability to diagnose AMI ^(65-67), the yield is higher with serial cardiac marker measurements ⁽⁶⁸⁾. Continuous 12-lead ECG monitoring to detect ST-segment shifts, both symptomatic and asymptomatic, can be performed with microprocessor-controlled, programmable devices. Clinical experience suggests that continuous ECG monitoring identifies episodes of ischemia that are missed with standard 12-lead ECGs obtained on presentation and that such episodes of transient ischemia provide independent prognostic information that indicates an increased risk of death, nonfatal MI, and the need for urgent revascularization ^(69,70). However, the ultimate clinical usefulness of continuous 12-lead ECG monitoring requires additional clarification.

7. Decision Aids that Combine Clinical Features and ECG Findings
ECG findings have been incorporated into mathematics-based decision aids for the triage of patients who present with chest pain ⁽⁴⁶⁾. The goals of these decision aids include the identification of patients at low risk of cardiac events, those with cardiac ischemia or AMI, and the estimation of prognosis ^{(28,58,71-76)}.

8. Biochemical Cardiac Markers
Biochemical cardiac markers are useful for both the diagnosis of myocardial necrosis and the estimation of prognosis. The loss of membrane integrity of myocytes that undergo necrosis allows intracellular macromolecules to diffuse into the cardiac interstitium and then into the lymphatics and cardiac microvasculature ⁽⁷⁷⁾. Eventually, these macromolecules, which are collectively referred to as biochemical cardiac markers, are detectable in the peripheral circulation. For optimum diagnostic usefulness, a marker of myocardial damage in the bloodstream should be present in a high concentration in the myocardium and absent from nonmyocardial tissue ^(52,77,78). It should be rapidly released into the blood after myocardial injury with a direct proportional relationship between the extent of myocardial injury and the measured level of the marker. Finally, the marker should persist in blood for a sufficient length of time to provide a convenient diagnostic time window with an easy, inexpensive, and rapid assay technique. Although no biochemical cardiac marker available at the present satisfies all of these requirements, as discussed later, the cardiac-specific troponins have a number of attractive features and are gaining acceptance as the biochemical markers of choice in the evaluation of patients with ACS ⁽⁶⁾.

For patients who present without ST-segment elevation, in whom the diagnosis may be unclear, biochemical cardiac markers are useful to confirm the diagnosis of MI. In addition, they provide valuable prognostic information, because there is a quantitative relationship between the magnitude of elevation of marker levels and the risk of an adverse outcome ⁽⁷⁹⁾.

a. Creatine Kinase
CK-MB has until recently been the principal serum cardiac marker used in the evaluation of ACS. Despite its common use, CK-MB has several limitations. Low levels of CK-MB in the blood of healthy persons limit its specificity for myocardial necrosis. CK-MB may also be elevated with severe damage of skeletal muscle ^(52,80,81). CK-MB isoforms exist in only 1 form in myocardial tissue (CK-MB2) but in different isoforms (or subforms) in plasma (CK-MB1). The use of an absolute level of CK-MB2 of greater than 1 U/L and a ratio of CK-MB2 to CK-MB1 of greater than 1.5 has improved sensitivity for the diagnosis of MI within the first 6 h compared with conventional assays for CK-MB, but this test has the same lack of absolute cardiac specificity as that of CK-MB itself ⁽⁸²⁾. Moreover, the assay is not widely available.

b. Cardiac Troponins
The troponin complex consists of 3 subunits: TnT, TnI, and troponin C (TnC) ⁽⁸¹⁾. Monoclonal antibody-based immunoassays have been developed to detect cardiac-specific TnT (cTnT) and cardiac-specific TnI (cTnI), because the amino acid sequences of the skeletal and cardiac isoforms of both TnT and TnI have sufficient dissimilarity. Because cardiac and smooth muscle share isoforms for TnC, no immunoassays of TnC have been developed for clinical purposes. Therefore, in these guidelines, the term "cardiac troponins" refers to either cTnT or cTnI or to both.

Because cTnT and cTnI are not generally detected in the blood of healthy persons, the cutoff value for elevated cTnT and cTnI levels may be set to slightly above the upper limit of the assay for a normal healthy population, leading some investigators to use the term "minor myocardial damage" or "microinfarction" for patients with detectable troponin but no CK-MB in the blood ⁽⁸³⁾. Case reports exist that confirm histological evidence of focal myocyte necrosis (e.g., microinfarction) in patients with elevated cardiac troponin levels and normal CK-MB values ^(6,84,85), indicating that myocardial necrosis can be recognized with increased sensitivity. It is estimated that approximately 30% of patients who present with rest pain without ST-segment elevation and who would otherwise be diagnosed as having UA because of a lack of CK-MB elevation actually have NSTEMI when assessed with cardiac-specific troponin assays. Although troponins are accurate in identifying myocardial necrosis ⁽⁵²⁰⁾, such necrosis is not necessarily secondary to atherosclerotic CAD. Therefore, in making the diagnosis of NSTEMI, cardiac troponins should be used in conjunction with appropriate symptoms or signs and/or ECG changes.

Elevated levels of cTnT or cTnI convey prognostic information beyond that supplied by the clinical characteristics of the patient, the ECG at presentation, and a predischarge exercise test ^{(61,62,86-88)}. Furthermore, among patients without ST-segment elevation and normal CK-MB levels, elevated cTnI or cTnT concentrations identify those at an increased risk of death ^(61,62). Finally, there is a quantitative relationship between the quantity of cTnI or cTnT that is measured and the risk of death in patients who present with an ACS (Figure 4). The incremental risk of death or MI in troponin-positive vs. troponin-negative patients is summarized in Tables 7 and 8. However, troponins should not be relied on as the sole markers for risk, because patients without troponin elevations may still exhibit a substantial risk of an adverse outcome. Neither marker is totally sensitive and specific in this regard. With currently available assays, cTnI and cTnT are of equal sensitivity and specificity in the detection of cardiac damage ⁽⁹⁰⁾. The choice should be made on the basis of cost and the availability of instrumentation at the institution.

Patients who present without ST-segment elevation who have elevated cardiac-specific troponin levels may receive a greater treatment benefit from platelet GP IIb/IIIa inhibitors and low-molecular-weight heparin (LMWH). For example, in the c7E3 Fab Antiplatelet Therapy in Unstable Refractory Angina (CAPTURE) trial, UA patients with an elevated cTnT level at presentation had a rate of death or nonfatal MI of 23.9% when treated with placebo vs. 9.5% when treated with abciximab (p = 0.002) ⁽⁹¹⁾, whereas among patients with a normal cTnT level, the rate of death or MI was 7.5% in the placebo group vs. 9.4% in the abciximab group (p = NS). Similar results have been reported for cTnI and cTnT with use of the GP IIb/IIIa inhibitor tirofiban ⁽⁹²⁾, and similar results were found in the Fragmin during Instability in Coronary Artery Disease (FRISC) trial of UA patients randomized to dalteparin or placebo. In the placebo group, the rate of death or nonfatal MI through 40 days increased progressively across the cTnT strata from 5.7% in the lowest tertile to 12.6% and 15.7% in the second and third tertiles, respectively ⁽⁹³⁾. In the dalteparin groups, the rates were 4.7%, 5.7%, and 8.9% across the tertiles of cTnT levels, corresponding to a 17.5% reduction in events in the lowest tertile but 43% and 55% reductions, respectively, in events in the upper 2 tertiles of cTnT levels.

c. Myoglobin
Although myoglobin, a low-molecular-weight heme protein found in both cardiac and skeletal muscle, is not cardiac specific, it is released more rapidly from infarcted myocardium than is CK-MB or the troponins and may be detected as early as 2 h after the onset of myocardial necrosis. However, the clinical value of serial determinations of myoglobin for the diagnosis of MI is limited by the brief duration of its elevation (less than 24 h) and by its lack of cardiac specificity. Thus, an isolated elevated concentration of myoglobin within the first 4 to 8 h after the onset of chest discomfort in patients with a nondiagnostic ECG should not be relied on to make the diagnosis of AMI but should be supplemented by a more cardiac-specific marker, such as CK-MB, cTnI, or cTnT ^(106,107). However, because of its high sensitivity, a negative test for myoglobin when blood is sampled within the first 4 to 8 h after onset is useful in ruling out myocardial necrosis.

d. Comparison of Cardiac Markers
The Diagnostic Marker Cooperative Study was a large, prospective, multicenter, double-blind study of patients who presented to the ED with chest pain in whom the diagnostic sensitivity and specificity for MI for total CK-MB (activity and mass), CK-MB subforms, myoglobin, and cTnI and cTnT were compared ⁽¹⁰⁸⁾. CK-MB subforms and myoglobin were most efficient for the early diagnosis (within 6 h) of MI, whereas cTnI and cTnT were highly cardiac specific and were particularly efficient for the late diagnosis of MI.

Table 9 compares the advantages and disadvantages of various cardiac markers for the evaluation and management of patients with suspected ACS but without ST-segment elevation on the 12-lead ECG. The troponins offer greater diagnostic sensitivity due to their ability to identify patients with lesser amounts of myocardial damage, which has been referred to as "minor myocardial damage." Nonetheless, these lesser amounts of damage confer a high risk in patients with ACS, because they are thought to represent microinfarctions that result from microemboli from an unstable plaque; the instability of the plaque rather than the actual amount of myocardial necrosis may be what places the patient at an increased risk. In addition, analyses from clinical trials suggest that the measurement of cardiac troponin concentrations provides prognostic information above and beyond that contained in simple demographic data such as the patient's age, findings on the 12-lead ECG, and measurement of CK-MB ^(61,62). Thus, measurement of cardiac troponin concentrations provides an efficient method for simultaneously diagnosing MI and providing prognostic information. Although not quite as sensitive or specific as the troponins, CK-MB by mass assay remains a very useful marker for the detection of more than minor myocardial damage. A normal CK-MB, however, does not exclude the minor myocardial damage and its attendant risk of adverse outcomes detectable by cardiac-specific troponins. As noted earlier, the measurement of CK-MB isoforms is useful for the extremely early diagnosis (less than 4 h) of MI. However, to date, experience with the measurement of CK-MB isoforms has been limited predominantly to dedicated research centers, and its "field performance" in widespread clinical use remains to be established. Because of its poor cardiac specificity in the setting of skeletal muscle injury and its rapid clearance from the bloodstream, myoglobin should not be used as the only diagnostic marker for the identification of patients with NSTEMI, but its early appearance makes it quite useful for ruling out myocardial necrosis.

Cardiac-specific troponins are gaining acceptance as the primary biochemical cardiac marker in ACS. Commercially available assays are undergoing refinement, with several versions of assays in clinical use with different diagnostic cutoffs, underscoring the need for careful review of the cardiac troponin results reported in local hospital laboratories ^(6,109). As with any new testing procedure, there may be a period of adjustment as the laboratory introduces the troponin assays and the clinician becomes familiar with their use. Clinicians are encouraged to work closely with their colleagues in laboratory medicine to minimize the transition phase in making troponin measurements available in their institutions. The continued measurement of CK-MB mass is advisable during this transition. It should be emphasized that troponin levels may not rise for 6 h after the onset of symptoms, and in the case of a negative troponin level at less than 6 h, the measurement should be repeated 8 to 12 h after the onset of pain.

9. Integration of Clinical History with Serum Marker Measurements
Given the overlapping time frame of the release pattern of biochemical cardiac markers, it is important that clinicians incorporate the time from the onset of the patient's symptoms into their assessment of the results of biochemical marker measurements ^{(6,110,111,111a)} (Figure 5). The earliest marker of myocardial necrosis, myoglobin, is a sensitive test but lacks cardiac specificity. Later appearing markers, such as TnT and TnI, are more specific but have a lower sensitivity for the very early detection of myocardial necrosis (e.g., less than 6 h) after the onset of symptoms, and if an early (less than 6 h) troponin test is negative, a measurement should be repeated 8 to 12 h after the onset of symptoms. Although the release kinetics of the troponins provide a wider diagnostic window for the diagnosis of MI at a time when CK-MB elevations have returned to normal, the more protracted period of elevation of troponin levels after an MI must be recognized. One possible disadvantage of the use of cardiac-specific troponins is their long (up to 10 to 14 days) persistence in the serum after release. Thus, if a patient who had an MI several days earlier presents with recurrent ischemic chest discomfort, a single, slightly elevated cardiac-specific troponin measurement may represent either old or new myocardial damage. Serum myoglobin, although less cardiac specific than the troponins, may be helpful in this situation. A negative value suggests that the elevated troponin is related to recent (less than 10 to 14 days) but not acute myocardial damage.

A promising method to both identify and exclude AMI within 6 h of symptoms is to rely on changes ( values) in concentrations. Because assays are becoming ever more sensitive and precise, this method permits the identification of significantly increasing values while still in the normal range of assay. Thus, by relying on values, patients without ST-segment elevation can be selected for therapy with GP IIb/IIIa inhibitors, and those with negative values can be considered for early stress testing ^(112-114).

a. Bedside Testing for Cardiac Markers
Cardiac markers can be measured in the central chemistry laboratory or with point-of-care instruments in the ED with desktop devices or hand-held bedside rapid qualitative assays ⁽⁸³⁾. When a central laboratory is used, results should be available within 60 min, preferably within 30 min. Point-of-care systems, if implemented at the bedside, have the advantage of reducing delays due to transportation and processing in a central laboratory and can eliminate delays due to the lack of availability of central laboratory assays at all hours. These advantages of point-of-care systems must be weighed against the need for stringent quality control and appropriate training of ED personnel in assay performance and the higher costs of point-of-care testing devices relative to determinations in the central laboratory. In addition, these point-of-care assays at present are qualitative or, at best, semiquantitative. The evolution of technology that will provide quantitative assays of multiple markers that are simple to use will improve the diagnosis and management of patients with suspected ACS in the ED. Portable devices are becoming available that allow the simultaneous rapid measurement of myoglobin, CK-MB, and TnI at the point of care ⁽¹¹²⁾, and they are likely to be useful in the assessment of patients with ACS.

10. Other Markers
Other biochemical markers for the detection of myocardial necrosis are less well studied than those mentioned earlier. Although the available evidence does not support their routine use, these other markers are of scientific interest, and if measured in a patient with chest pain, they may provide useful supportive diagnostic information that can be incorporated into the overall assessment of the likelihood of CAD and the level of risk of the patient for death and cardiac ischemic events.

Markers of activity of the coagulation cascade, including elevated plasma levels of fibrinopeptide ⁽¹¹⁵⁾ and fibrinogen ⁽¹¹⁶⁾, appear to indicate an increased risk in ACS patients.

Given the increasing interest in the hypothesis that destabilization of atherosclerotic plaques may result from inflammatory processes, several groups have evaluated markers of the acute phase of inflammation, such as CRP, serum amyloid A ⁽¹¹⁷⁾, and interleukin-6 in patients with UA. Patients without biochemical evidence of myocardial necrosis but who have an elevated CRP level are at an increased risk of an adverse outcome, especially those whose CRP levels are markedly elevated (e.g., highest quintile in population studies) ^(118-121). Elevated levels of interleukin-6, the major determinant of acute phase reactant proteins in the liver, and serum amyloid A, another acute phase reactant protein, have been shown to have a similar predictive value of an adverse outcome as CRP ^(119,121). Increased levels of circulating soluble adhesion molecules, such as intercellular adhesion molecule-1, vascular cell adhesion molecule-1, and E-selectin, in patients with UA are under investigation as markers of increased risk ⁽¹²²⁾.

C. Immediate Management

Recommendations
Class I

1. The history, physical examination, 12-lead ECG, and initial cardiac marker tests should be integrated to assign patients with chest pain into 1 of 4 categories: a noncardiac diagnosis, chronic stable angina, possible ACS, and definite ACS. (Level of Evidence: C)
2. Patients with definite or possible ACS, but whose initial 12-lead ECG and cardiac marker levels are normal, should be observed in a facility with cardiac monitoring (e.g., chest pain unit), and a repeat ECG and cardiac marker measurement should be obtained 6 to 12 h after the onset of symptoms. (Level of Evidence: B)
3. In patients in whom ischemic heart disease is present or suspected, if the follow-up 12-lead ECG and cardiac marker measurements are normal, a stress test (exercise or pharmacological) to provoke ischemia may be performed in the ED, in a chest pain unit, or on an outpatient basis shortly after discharge. Low-risk patients with a negative stress test can be managed as outpatients. (Level of Evidence: C)
4. Patients with definite ACS and ongoing pain, positive cardiac markers, new ST-segment deviations, new deep T-wave inversions, hemodynamic abnormalities, or a positive stress test should be admitted to the hospital for further management. (Level of Evidence: C)
5. Patients with possible ACS and negative cardiac markers who are unable to exercise or who have an abnormal resting ECG should undergo a pharmacological stress test. (Level of Evidence: B)
6. Patients with definite ACS and ST-segment elevation should be evaluated for immediate reperfusion therapy. (Level of Evidence: A)

By integrating information from the history, physical examination, 12-lead ECG, and initial cardiac marker tests, clinicians can assign patients into 1 of 4 categories: noncardiac diagnosis, chronic stable angina, possible ACS, and definite ACS (Figure 6).

Patients who arrive at a medical facility in a pain-free state, have unchanged or normal ECGs, are hemodynamically stable, and do not have elevated cardiac markers represent more of a diagnostic than an urgent therapeutic challenge. Evaluation begins in these patients by obtaining information from the history, physical examination, and ECG (see Tables 5 and 6) to be used to confirm or reject the diagnosis of UA/NSTEMI.

Patients with a low likelihood of CAD should be evaluated for other causes of the presentation, including musculoskeletal pain; gastrointestinal disorders such as esophageal spasm, gastritis, peptic ulcer disease, or cholecystitis; intrathoracic disease, such as pneumonia, pleurisy, pneumothorax, or pericarditis; and neuropsychiatric disease, such as hyperventilation or panic disorder (Figure 6, B1). Patients who are found to have evidence of one of these alternative diagnoses should be excluded from management with these guidelines and referred for appropriate follow-up care (Figure 6, C1). Reassurance should be balanced with instructions to return for further evaluation if symptoms worsen or if the patient fails to respond to symptomatic treatment.

Chronic stable angina may also be diagnosed in this setting (Figure 6, B2), and patients with this diagnosis should be managed according to the ACC/AHA/ACP-ASIM Guidelines for the Management of Patients With Chronic Stable Angina ⁽²⁶⁾.

Patients with possible ACS (Figure 6, B3 and D1) are candidates for additional observation in a specialized facility (e.g., chest pain unit) (Figure 6, E1). Patients with definite ACS (Figure 6, B4) are triaged based on the pattern of the 12-lead ECG. Patients with ST-segment elevation Figure 6, C3) are evaluated for immediate reperfusion therapy (Figure 6, D3) and managed according to the ACC/AHA Guidelines for the Management of Patients With Acute Myocardial Infarction ⁽⁵⁾, whereas those without ST-segment elevation (Figure 6, C2) are either managed by additional observation (Figure 6, E1) or admitted to the hospital (Figure 6, H3). Patients with low-risk ACS (Table 5) without transient ST-segment depressions of greater than or equal to 0.05 mV and/or T-wave inversions of greater than or equal to 0.2 mV, without positive cardiac markers, and without a positive stress test (Figure 6, H1) may be discharged and treated as outpatients (Figure 6, I1).

1. Chest Pain Units
To facilitate a more definitive evaluation while avoiding the unnecessary hospital admission of patients with possible ACS (Figure 6, B3) and low-risk ACS (Figure 6, F1) and the inappropriate discharge of patients with active myocardial ischemia without ST-segment elevation (Figure 6, C2), special units have been devised that are variously referred to as "chest pain units" and "short-stay ED coronary care units." Personnel in these units use critical pathways or protocols designed to arrive at a decision about the presence or absence of myocardial ischemia and, if present, to characterize it further as UA or NSTEMI and to define the optimal next step in the care of the patient (e.g., admission, acute intervention) ⁽¹²³⁾. The goal is to arrive at such a decision after a finite amount of time, which usually is between 6 and 12 h but may extend up to 24 h depending on the policies in individual hospitals. Although chest pain units are useful, other appropriate observation areas in which patients with chest pain can be evaluated may be used as well.

The physical location of the chest pain unit or site where patients with chest pain are observed is variable, ranging from a specifically designated area of the ED to a separate unit with the appropriate equipment ⁽¹²⁴⁾. Similarly, the chest pain unit may be administratively a part of the ED and staffed by emergency physicians or may be administered and staffed separately. Suggestions for the design of chest pain units have been presented by several authoritative bodies and generally include provisions for continuous monitoring of the patient's ECG, ready availability of cardiac resuscitation equipment and medications, and appropriate staffing with nurses and physicians. Given the evolving nature of the field and the recent introduction of chest pain units into clinical medicine, ACEP has published guidelines that recommend a program for the continuous monitoring of outcomes of patients evaluated in such units as well as the impact on hospital resources ⁽¹²⁵⁾. A Consensus Panel statement from ACEP emphasized that chest pain units should be considered 1 part of a multifaceted program that also includes efforts to minimize patient delays in seeking medical care and delays in the ED itself ⁽¹²⁵⁾.

Several groups have studied the impact of chest pain units on the care of patients with chest pain who present to the ED. It has been reported, both from studies with historical controls and from randomized trials, that the use of chest pain units is cost saving compared with an in-hospital evaluation to "rule-out MI" ^(126,127).

A common clinical practice is to minimize the chance of "missing" an MI in a patient with chest discomfort by admitting to the hospital all patients with suspected ACS and by obtaining serial 12-lead ECGs and biochemical cardiac marker measurements to either exclude or confirm the diagnosis of MI. Such a practice typically results in a low percentage of admitted patients actually being confirmed to have an MI. Given the inverse relationship between the percentage of patients with a "rule-out MI evaluation" and the "MI miss rate," the potential cost savings of a chest pain unit varies depending on the practice pattern for the disposition of chest pain patients at individual hospitals ⁽¹²⁶⁾. Hospitals with a high admission rate of low-risk patients to "rule-out MI" (70% to 80%) will experience the largest cost savings by implementing a chest pain unit approach but will have the smallest impact on the number of missed MI patients. In contrast, hospitals with relatively low admission rates of such patients (30% to 40%) will experience greater improvements in the quality of care because fewer MI patients will be missed but will have a smaller impact on costs because of the low baseline admission rate.

a. Potential Expansion of the Use of Chest Pain Units for Intermediate-Risk Patients
Farkouh et al. ⁽¹²⁸⁾ extended the use of a chest pain unit in a separate portion of the ED to include patients at an intermediate risk of adverse clinical outcome based on the previously published AHCPR guidelines for the management of UA ⁽¹⁾ (Table 6). They reported a 46% reduction in the ultimate need for hospital admission in intermediate-risk patients after a median stay of 9.2 h in the chest pain unit. Extension of the use of chest pain units to intermediate-risk patients in an effort to reduce inpatient costs is facilitated by making available diagnostic testing modalities such as treadmill testing and stress imaging (echocardiographic or nuclear) 7 days a week ⁽¹²⁹⁾.

b. Triage of Patients
Patients with chest discomfort for whom a specific diagnosis cannot be made after a review of the history, physical examination, initial 12-lead ECG, and biochemical cardiac marker data should undergo a more definitive evaluation. Several categories of patients should be considered according to the algorithm shown in Figure 6:

• Patients with possible ACS (Figure 6, B3) are those who had a recent episode of chest discomfort at rest not entirely typical of ischemia but are pain free when initially evaluated, have a normal or unchanged ECG, and have no elevations of cardiac markers.
• Patients with a recent episode of typical ischemic discomfort that either is of new onset or severe or exhibits an accelerating pattern of previous stable angina (especially if it has occurred at rest or is within 2 weeks of a previously documented MI) should initially be considered to have a "definite ACS" (Figure 6, B4). However, such patients may be at a low risk if their ECG obtained at presentation has no diagnostic abnormalities and the initial serum cardiac markers (especially cardiac-specific troponins) are normal (Figure 6, C2 and D1). As indicated in the algorithm, patients with either "possible ACS" (Figure 6, B3) or "definite ACS" (Figure 6, B4) but with nondiagnostic ECG and normal initial cardiac markers (Figure 6, D1) are candidates for additional observation in the ED or in a specialized area such as a chest pain unit (E1). In contrast, patients who present without ST-segment elevation but have features indicative of active ischemia (ongoing pain, ST-segment and/or T-wave changes, positive cardiac markers, or hemodynamic instability) (Figure 6, D2) should be admitted to the hospital (H3).

2. Discharge from ED or Chest Pain Unit
The initial assessment of whether a patient has UA/NSTEMI and which triage option is most suitable generally should be made immediately on the patient's arrival at a medical facility. Rapid assessment of a patient's candidacy for additional observation can be accomplished based on the status of the symptoms, ECG findings, and serum cardiac marker measurements.

Patients who experience recurrent ischemic discomfort, evolve abnormalities on a follow-up 12-lead ECG or cardiac marker measurements, or develop hemodynamic abnormalities such as new or worsening congestive heart failure (CHF) (Figure 6, D2) should be admitted to the hospital (Figure 6, H3) and managed as described in Section III.

Patients who are pain free, have either a normal or nondiagnostic ECG or one that is unchanged from previous tracings, and have a normal set of initial cardiac marker measurements are candidates for further evaluation to screen for nonischemic discomfort (Figure 6, B1) vs. a low-risk ACS (Figure 6, D1). If the patient is low risk (Table 6) and does not experience any further ischemic discomfort and a follow-up 12-lead ECG and cardiac marker measurements after 6 to 8 h of observation are normal (Figure 6, F1), the patient may be considered for an early stress test to provoke ischemia (Figure 6, G1). This test can be performed before the discharge and should be supervised by an experienced physician. Alternatively, the patient may be discharged and return for a stress test as an outpatient within 72 h. The exact nature of the stress test may vary depending on the patient's ability to exercise on either a treadmill or bicycle and the local expertise in a given hospital setting (e.g., availability of different testing modalities at different times of the day or different days of the week) ⁽¹³⁰⁾. Patients who are capable of exercise and are free of confounding features on the baseline ECG, such as bundle-branch block, LV hypertrophy, or paced rhythms, can be evaluated with routine symptom-limited conventional exercise stress testing. Patients who are incapable of exercise or who have an uninterpretable baseline ECG should be considered for pharmacological stress testing with either nuclear perfusion imaging or two-dimensional echocardiography ^(46,131). Because LV function is so integrally related to prognosis and heavily affects therapeutic options, strong consideration should be given to the assessment of LV function with echocardiography or radionuclide ventriculography in patients with documented ischemia. In sites at which stress tests are not available, low-risk patients may be discharged and the test scheduled to be carried out within 72 h.

Patients who develop recurrent pain during observation or in whom the follow-up studies (12-lead ECG, cardiac markers) show new abnormalities (Figure 6, F2) should be admitted to the hospital (Figure 6, H3).

Because continuity of care is important in the overall management of patients with a chest pain syndrome, the patient's primary physician (if not involved in the care of the patient during the initial episode) should be notified of the results of the evaluation and should receive a copy of the relevant test results. Patients with a noncardiac diagnosis and those with low risk or possible ACS with a negative stress test should be counseled to make an appointment with their primary care physician as outpatients for further investigation into the cause of their symptoms (Figure 6, I1). They should be seen by a physician within 72 h of discharge from the ED or chest pain unit.

Patients with possible ACS (Figure 6, B3) and those with a definite ACS but a nondiagnostic ECG and normal biochemical cardiac markers when they are initially seen (Figure 6, D1) at institutions without a chest pain unit (or equivalent facility) should be admitted to an inpatient unit. The inpatient unit to which such patients are to be admitted should have the same provisions for continuous ECG monitoring, availability of resuscitation equipment, and staffing arrangements as described earlier for the design of chest pain units.