ACC/AHA GUIDELINES
ACC/AHA/ACP-ASIM guidelines for the management of patients with chronic stable angina1
A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Management of Patients With Chronic Stable Angina)
Raymond J. Gibbons, MD, FACC, Committee Member, Chair,
Kanu Chatterjee, MB, FACC, Committee Member,
Jennifer Daley, MD, FACP, Committee Member,
John S. Douglas, MD, FACC, Committee Member,
Stephan D. Fihn, MD, MPH, FACP, Committee Member,
Julius M. Gardin, MD, FACC, Committee Member,
Mark A. Grunwald, MD, FAAFP, Committee Member,
Daniel Levy, MD, FACC, Committee Member,
Bruce W. Lytle, MD, FACC, Committee Member,
Robert A. O’Rourke, MD, FACC, Committee Member,
William P. Schafer, MD, FACC, Committee Member,
Sankey V. Williams, MD, FACP, Committee Member,
James L. Ritchie, MD, FACC, Task Force Member, Chair,
Raymond J. Gibbons, MD, FACC, Task Force Member, Vice Chair,
Melvin D. Cheitlin, MD, FACC, Task Force Member,
Kim A. Eagle, MD, FACC, Task Force Member,
Timothy J. Gardner, MD, FACC, Task Force Member,
Arthur Garson, Jr, MD, MPH, FACC, Task Force Member,
Richard O. Russell, MD, FACC, Task Force Member,
Thomas J. Ryan, MD, FACC, Task Force Member and
Sidney C. Smith, Jr, MD, FACC, Task Force Member
 |
Committee members
|
|---|
Table of contents.
- Introduction and Overview......2093
- Organization of Committee and Evidence Review......2093
- Scope of the Guidelines......2094
- Overlap With Other Guidelines......2094
- Magnitude of the Problem......2095
- Organization of the Guidelines......2097
- Diagnosis......2098
- History and Physical......2098
- Associated Conditions......2105
- Noninvasive Testing......2106
- ECG/Chest X-Ray......2106
- Exercise ECG for Diagnosis......2107
- Echocardiography (Resting)......2111
- Stress Imaging Studies—Echo and Nuclear......2112
- Invasive Testing: Value of Coronary Angiography......2119
- Risk Stratification......2121
- Clinical Assessment......2121
- ECG/Chest X-Ray......2123
- Noninvasive Testing......2123
- Resting LV Function (Echo/Radionuclide Imaging)......2123
- Exercise Testing for Risk Stratification and Prognosis......2124
- Stress Imaging Studies (Radionuclide and Echocardiography)......2127
- Coronary Angiography and Left Ventriculography......2133
- Treatment......2135
- Pharmacologic Therapy......2135
- Definition of Successful Treatment and Initiation of Treatment......2145
- Education of Patients with Chronic Stable Angina2147
- Coronary Disease Risk Factors and Evidence That Treatment Can Reduce the Risk for Coronary Disease Events......2149
- Revascularization for Chronic Stable Angina......2161
- Patient Follow-up: Monitoring of Symptoms and Antianginal Therapy......2167
- Index......2191
|
Preamble
|
|---|
It is important that the medical profession play a significant role in critically evaluating the use of diagnostic procedures and therapies in the management or prevention of disease states. Rigorous and expert analysis of the available data documenting relative benefits and risks of those procedures and therapies can produce helpful guidelines that improve the effectiveness of care, optimize patient outcomes, and have a favorable impact on the overall cost of care by focusing resources on the most effective strategies.
The American College of Cardiology (ACC) and the American Heart Association (AHA) have jointly engaged in the production of such guidelines in the area of cardiovascular disease since 1980. This effort is directed by the ACC/AHA Task Force on Practice Guidelines, whose charge is to develop and revise practice guidelines for important cardiovascular diseases and procedures. Experts in the subject under consideration are selected from both organizations to examine subject-specific data and write guidelines. The process includes additional representatives from other medical practitioner and specialty groups where appropriate. Writing groups are specifically charged to perform a formal literature review, weigh the strength of evidence for or against a particular treatment or procedure, and include estimates of expected health outcomes where data exist. Patient-specific modifiers, comorbidities and issues of patient preference that might influence the choice of particular tests or therapies are considered as well as frequency of follow-up and cost-effectiveness.
The ACC/AHA Task Force on Practice Guidelines makes every effort to avoid any actual or potential conflicts of interest that might arise as a result of an outside relationship or personal interest of a member of the writing panel. Specifically, all members of the writing panel are asked to provide disclosure statements of all such relationships that might be perceived as real or potential conflicts of interest. These statements are reviewed by the parent task force, reported orally to all members of the writing panel at the first meeting, and updated yearly and as changes occur.
These practice guidelines are intended to assist physicians in clinical decision making by describing a range of generally acceptable approaches for the diagnosis, management, and prevention of specific diseases or conditions. These guidelines attempt to define practices that meet the needs of most patients in most circumstances. The ultimate judgment regarding care of a particular patient must be made by the physician and patient in light of all of the circumstances presented by that patient.
The executive summary and recommendations are published in the June 1, 1999 issue of Circulation. The full text is published in the June 1999 issue of the Journal of the American College of Cardiology. Reprints of the full text and the executive summary are available from both organizations.
James L. Ritchie, MD, FACC
Chair, ACC/AHA Task Force on Practice Guidelines
|
I. Introduction and overview
|
|---|
A. Organization of committee and evidence review.
The ACC/AHA Task Force on Practice Guidelines was formed to make recommendations regarding the diagnosis and treatment of patients with known or suspected cardiovascular disease. Ischemic heart disease is the single leading cause of death in the U.S. The most common manifestation of this disease is chronic stable angina. Recognizing the importance of the management of this common entity and the absence of national clinical practice guidelines in this area, the task force formed the current committee to develop guidelines for the management of patients with stable angina. Because this problem is frequently encountered in the practice of internal medicine, the task force invited the American College of Physicians-American Society of Internal Medicine (ACP-ASIM) to serve as a partner in this effort by naming four general internists to serve on the committee.
The committee reviewed and compiled published reports (excluding abstracts) through a series of computerized literature searches of the English language research literature since 1975 and a manual search of selected final articles. Details of the specific searches conducted for particular sections are provided as appropriate. Detailed evidence tables were developed whenever necessary on the basis of specific criteria outlined in the individual sections. The recommendations were based primarily on these published data. The weight of the evidence was ranked high (A) if the data were derived from multiple randomized clinical trials with large numbers of patients and intermediate (B) if the data were derived from a limited number of randomized trials with small numbers of patients, careful analyses of nonrandomized studies or observational registries. A low rank (C) was given when expert consensus was the primary basis for the recommendation.
The customary ACC/AHA classifications I, II and III are used in tables that summarize both the evidence and expert opinion and provide final recommendations for both patient evaluation and therapy:
Class I
Conditions for which there is evidence or general agreement that a given procedure or treatment is useful and effective.
Class II
Conditions for which there is conflicting evidence or a divergence of opinion about the usefulness/efficacy of a procedure or treatment.
Class IIa
Weight of evidence/opinion is in favor of usefulness/efficacy.
Class IIb
Usefulness/efficacy is less well established by evidence/opinion.
Class III
Conditions for which there is evidence and/or general agreement that the procedure/treatment is not useful/effective and in some cases may be harmful.
A complete list of many publications on various aspects of this subject is beyond the scope of these guidelines; only selected references are included. The committee consisted of acknowledged experts in general internal medicine from the ACP-ASIM, family medicine from the American Academy of Family Physicians (AAFP), and general cardiology as well as persons with recognized expertise in more specialized areas, including noninvasive testing, preventive cardiology, coronary intervention, and cardiovascular surgery. Both the academic and private practice sectors were represented. This document was reviewed by three outside reviewers nominated by the ACC, three outside reviewers nominated by the AHA, three outside reviewers nominated by the ACP-ASIM, and two outside reviewers nominated by the AAFP. This document was approved for publication by the governing bodies of the ACC, AHA, and ACP-ASIM. The task force will review these guidelines one year after publication and yearly thereafter to determine whether revisions are needed. These guidelines will be considered current unless the task force revises or withdraws them from distribution.
B. scope of the guidelines.
These guidelines are intended to apply to adult patients with stable chest pain syndromes and known or suspected ischemic heart disease. Patients who have "ischemic equivalents," such as dyspnea or arm pain with exertion, are included in these guidelines. Some patients with ischemic heart disease may become asymptomatic with appropriate therapy. As a result, the follow-up sections of the guidelines may apply to patients who were previously symptomatic. However, the diagnosis, risk stratification and treatment sections of the guidelines are intended to apply to symptomatic patients. Asymptomatic patients with "silent ischemia" or known coronary artery disease (CAD) that has been detected in the absence of symptoms are beyond the scope of these guidelines. Pediatric patients are also beyond the scope of these guidelines because ischemic heart disease is very unusual in such patients and is primarily related to the presence of coronary artery anomalies. Patients with chest pain syndromes following cardiac transplantation are also not included in these guidelines.
Patients with nonanginal chest pain are generally at lower risk for ischemic heart disease. Often their chest pain syndromes have identifiable noncardiac causes. Such patients are included in these guidelines if there is sufficient suspicion of heart disease to warrant cardiac evaluation. If the evaluation demonstrates that ischemic heart disease is unlikely and noncardiac causes are the primary focus of evaluation, such patients are beyond the scope of these guidelines. If the initial cardiac evaluation demonstrates that ischemic heart disease is possible, subsequent management of such patients does fall within these guidelines.
Acute ischemic syndromes are not included in these guidelines. For patients with acute myocardial infarction (MI), the reader is referred to the "ACC/AHA Guidelines for the Management of Patients With Acute Myocardial Infarction" (1). For patients with unstable angina, the reader is referred to the Agency for Health Care Policy and Research (AHCPR) clinical practice guideline on unstable angina (2), which was endorsed by the ACC and the AHA. This guideline for unstable angina did describe some low-risk patients who should not be hospitalized but instead evaluated as outpatients. Such patients are indistinguishable from many patients with stable chest pain syndromes and are therefore within the scope of the present guidelines. Patients whose recent unstable angina was satisfactorily treated by medical therapy and who then present with a recurrence of symptoms with a stable pattern fall within the scope of the present guidelines. Similarly, patients with MI who subsequently present with stable chest pain symptoms >30 days after the initial event are within the scope of the present guidelines.
The present guidelines do not apply to patients with chest pain symptoms early after revascularization by either percutaneous techniques or coronary artery bypass grafting. Although the division between "early" and "late" symptoms is arbitrary, the committee believed that these guidelines should not be applied to patients who develop recurrent symptoms within six months of revascularization.
C. Overlap with other guidelines.
These guidelines will overlap with a large number of recently published (or soon to be published) clinical practice guidelines developed by the ACC/AHA Task Force on Practice Guidelines; the National Heart, Lung, and Blood Institute (NHLBI); and the ACP-ASIM (Table 1).
This report includes text and recommendations from many of these guidelines, which are clearly indicated. Additions and revisions have been made where appropriate to reflect more recently available evidence. This report specifically indicates rare situations in which it deviates from previous guidelines and presents the rationale. In some cases, this report attempts to combine previous sets of similar and dissimilar recommendations into one set of final recommendations. Although this report includes a significant amount of material from the previous guidelines, by necessity the material was often condensed into a succinct summary. These guidelines are not intended to provide a comprehensive understanding of the imaging modalities, therapeutic modalities, and clinical problems detailed in other guidelines. For such an understanding, the reader is referred to the original guidelines listed in the references.
D. Magnitude of the problem.
There is no question that ischemic heart disease remains a major public health problem. Chronic stable angina is the initial manifestation of ischemic heart disease in approximately one half of patients (3,4). It is difficult to estimate the number of patients with chronic chest pain syndromes in the U.S. who fall within these guidelines, but clearly it is measured in the millions. The reported annual incidence of angina is 213/100,000 population >30 years old (3). When the Framingham Heart Study (4) is considered, an additional 350,000 Americans each year are covered by these guidelines. The AHA has estimated that 6,200,000 Americans have chest pain (5); however, this may be a conservative estimate.
The prevalence of angina can also be estimated by extrapolating from the number of MIs in the U.S. (1). About one half of patients presenting at the hospital with MI have preceding angina (6). The best current estimate is that there are 1,100,000 patients with MI each year in the U.S. (5); about one half of these (550,000) survive until hospitalization. Two population-based studies (from Olmsted County, Minnesota, and Framingham, Massachusetts) examined the annual rates of MI in patients with symptoms of angina and reported similar rates of 3% to 3.5% per year (4,7). On this basis, it can be estimated that there are 30 patients with stable angina for every patient with infarction who is hospitalized. As a result, the number of patients with stable angina can be estimated as 30 x 550,000, or 16,500,000. This estimate does not include patients who do not seek medical attention for their chest pain or whose chest pain has a noncardiac cause. Thus, it is likely that the present guidelines cover at least six million Americans and conceivably more than twice that number.
Ischemic heart disease is important not only because of its prevalence but also because of its associated morbidity and mortality. Despite the well-documented recent decline in cardiovascular mortality (8), ischemic heart disease remains the leading single cause of death in the U.S. (Table 2) and is responsible for 1 of every 4.8 deaths (9). The morbidity associated with this disease is also considerable: each year >1,000,000 patients have an MI. Many more are hospitalized for unstable angina and evaluation and treatment of stable chest pain syndromes. Beyond the need for hospitalization, many patients with chronic chest pain syndromes are temporarily unable to perform normal activities for hours or days, thereby experiencing a reduced quality of life. According to the recently published data from the Bypass Angioplasty Revascularization Investigation (10), about 30% of patients never return to work following coronary revascularization, and 15% to 20% of patients rated their own health fair or poor despite revascularization. These data confirm the widespread clinical impression that ischemic heart disease continues to be associated with considerable patient morbidity despite the decline in cardiovascular mortality.
The economic costs of chronic ischemic heart disease are enormous. Some insight into the potential cost can be obtained by examining Medicare data for inpatient diagnosis-related groups (DRGs) and diagnostic tests. Table 3 shows the number of patients hospitalized under various DRGs during 1995 and associated direct payments by Medicare. These DRGs represent only hospitalization of patients covered by Medicare. The table includes estimates for the proportion of inpatient admissions for unstable angina, MI, and revascularization for patients with a history of stable angina. Direct costs associated with non-Medicare patients hospitalized for the same diagnoses are probably about the same as the covered charges under Medicare. Thus, the direct costs of hospitalization are >$15 billion.
Table 4 shows the Medicare fees and volumes of commonly used diagnostic procedures in ischemic heart disease. Although some of these procedures may have been performed for other diagnoses and some of the cost of the technical procedure relative value units (RVUs) may have been for inpatients listed in Table 3, the magnitude of the direct costs is considerable. When the 1998 Medicare reimbursement of $36.6873 per RVU is used, the direct cost to Medicare of these 61.2 million RVUs can be estimated at $2.25 billion. Again, assuming that the non-Medicare patient costs are at least as great, the estimated cost of these diagnostic procedures alone would be about $4.5 billion.
These estimates of the direct costs associated with chronic stable angina obviously do not take into account the indirect costs of workdays lost, reduced productivity, long-term medication, and associated other effects. The indirect costs have been estimated to be almost as great as direct costs (4). The magnitude of the problem can be succinctly summarized: chronic stable angina affects many millions of Americans, with associated annual costs that are measured in tens of billions of dollars.
Given the magnitude of this problem, the need for practice guidelines is self-evident. This need is further reinforced by the available information, which suggests considerable regional differences in the management of ischemic heart disease. Figure 1 shows published information from the Medicare database for rates of coronary angiography in different counties of the country (11). Three- and four-fold differences in adjusted rates for this procedure in different counties within the same state are not uncommon, suggesting that the clinical management of such patients is highly variable. The reasons for such variation in management are unknown.
View larger version (53K):
[in this window]
[in a new window]
|
Figure 1 Map depicting coronary angiography rates in the U.S. HRR = hospital referral region. From Wennberg et al. (11) with permission.
|
|
E. Organization of the guidelines.
These guidelines are arbitrarily divided into four sections: diagnosis, risk stratification, treatment and patient follow-up. Experienced clinicians will quickly recognize that the distinctions between these sections may be arbitrary and unrealistic in individual patients. However, for most clinical decision making, these divisions are helpful and facilitate presentation and analysis of the available evidence.
The three flow diagrams that follow summarize the management of stable angina in three algorithms: clinical assessment (Fig. 2), stress testing/angiography (Fig. 3), and treatment (Fig. 4). The treatment mnemonic (Fig. 5) is intended to highlight the 10 treatment elements that the committee considered most important.
View larger version (43K):
[in this window]
[in a new window]
|
Figure 2 Clinical assessment. AHCPR = Agency for Health Care Policy and Research; MI = myocardial infarction; PTCA = percutaneous transluminal coronary angioplasty; CABG = coronary artery bypass graft; ACC = American College of Cardiology; AHA = American Heart Association; LV = left ventricular; and ECG = electrocardiogram.
|
|
View larger version (58K):
[in this window]
[in a new window]
|
Figure 4 Treatment. CAD = coronary artery disease; NTG = nitroglycerin; MI = myocardial infarction; NCEP = National Cholesterol Education Program; JNC = Joint National Committee. *Vasodilators, excessive thyroid replacement, vasoconstrictors, profound anemia, uncontrolled hypertension, hyperthyroidism, hypoxemia, tachyarrhythmias, bradyarrhythmias, valvular heart disease (especially aortic stenosis) and hypertrophic cardiomyopathy. **On the basis of coronary anatomy, severity of anginal symptoms, and patient preferences, it is reasonable to consider evaluation for coronary revascularization. Unless a patient has documented left main, three-vessel, or two-vessel CAD with significant stenosis of the proximal left anterior descending coronary artery, there is no demonstrated survival advantage associated with revascularization in low-risk patients with chronic stable angina. Thus, medical therapy should be attempted in most patients before considering percutaneous transluminal coronary angioplasty or coronary artery bypass graft.
|
|
Although the evaluation of many patients will require all three algorithms, this is not always true. Some patients may require only clinical assessment to determine that they do not belong within these guidelines. Others may require only clinical assessment and treatment if the probability of CAD is high and patient preferences and comorbidities preclude revascularization (and therefore the need for risk stratification). The stress testing/angiography algorithm may be required either for diagnosis (and risk stratification) in patients with a moderate probability of CAD or for risk stratification only in patients with a high probability of CAD.
|
II. Diagnosis
|
|---|
A. History and physical.
Recommendations
Class I
In patients presenting with chest pain, a detailed symptom history, focused physical examination, and directed risk-factor assessment should be performed. With this information, the clinician should estimate the probability of significant CAD (i.e., low, intermediate, high). (Level of Evidence: B)
|
Definition of angina
|
|---|
Angina is a clinical syndrome characterized by discomfort in the chest, jaw, shoulder, back or arm. It is typically aggravated by exertion or emotional stress and relieved by nitroglycerin. Angina usually occurs in patients with CAD involving  1 large epicardial artery. However, angina can also occur in persons with valvular heart disease, hypertrophic cardiomyopathy and uncontrolled hypertension. It can be present in patients with normal coronary arteries and myocardial ischemia related to spasm or endothelial dysfunction. Angina is also a symptom in patients with noncardiac conditions of the esophagus, chest wall or lungs. Once cardiac causes have been excluded, the management of patients with these noncardiac conditions is outside the scope of these guidelines.
|
Clinical evaluation of patients with chest pain
|
|---|
History.
The clinical examination is the most important step in the evaluation of the patient with chest pain, allowing the clinician to estimate the likelihood of clinically significant CAD with a high degree of accuracy (29). Significant CAD is defined angiographically as CAD with 70% diameter stenosis of 1 major epicardial artery segment or 50% diameter stenosis of the left main coronary artery. Although lesions of less stenosis can cause angina, they have much less prognostic significance (30).
The first step, a detailed description of the symptom complex, enables the clinician to characterize the chest pain (31). Five components are typically considered: quality, location, duration of pain, factors that provoke the pain and factors that relieve the pain. Various adjectives have been used by patients to describe the quality of the anginal pain: "squeezing," "griplike," "pressurelike," "suffocating" and "heavy" are common. Not infrequently, patients insist that their symptom is a "discomfort" but not "pain." Angina is almost never sharp or stabbing, and it usually does not change with position or respiration.
The anginal episode is typically minutes in duration. Fleeting discomfort or a dull ache lasting for hours is rarely angina. The location of angina is usually substernal, but radiation to the neck, jaw, epigastrium, or arms is not uncommon. Pain above the mandible, below the epigastrium, or localized to a small area over the left lateral chest wall is rarely anginal. Angina is generally precipitated by exertion or emotional stress and commonly relieved by rest. Sublingual nitroglycerin also relieves angina, usually within 30 s to several minutes.
After the history of the pain is obtained, the physician makes a global assessment of the symptom complex. One classification scheme for chest pain in many studies uses three groups: typical angina, atypical angina or noncardiac chest pain (32) (Table 5).
Angina is further classified as stable or unstable (2). Unstable angina is important in that its presence predicts a much higher short-term risk of an acute coronary event. Unstable angina is operationally defined as angina that presents in one of three principal ways: rest angina, severe new-onset angina, or increasing angina (Tables 6 and 7 ). Most important, unstable angina patients can be subdivided by their short-term risk (Table 8). Patients at high or moderate risk often have coronary artery plaques that have recently ruptured. Their risk of death is intermediate, between that of patients with acute MI and patients with stable angina. The initial evaluation of high- or moderate-risk patients with unstable angina is best carried out in the inpatient setting. However, low-risk patients with unstable angina have a short-term risk not substantially different from those with stable angina. Their evaluation can be accomplished safely and expeditiously in an outpatient setting. The recommendations made in these guidelines do not apply to high- and moderate-risk unstable angina but are applicable to the low-risk unstable angina group.
After a detailed chest pain history is taken, the presence of risk factors for CAD (23) should be determined. Cigarette smoking, hyperlipidemia, diabetes, hypertension and a family history of premature CAD are all important. Past history of cerebrovascular or peripheral vascular disease increases the likelihood that CAD will be present.
Physical.
The physical examination is often normal in patients with stable angina (33). However, an exam made during an episode of pain can be beneficial. An S4 or S3 sound or gallop, mitral regurgitant murmur, a paradoxically split S2 or bibasilar rales or chest wall heave that disappears when the pain subsides are all predictive of CAD (34). Even though the physical is generally not helpful for confirming CAD, a careful cardiovascular exam may reveal other conditions associated with angina, such as valvular heart disease or hypertrophic cardiomyopathy. Evidence of noncoronary atherosclerotic disease—a carotid bruit, diminished pedal pulse or abdominal aneurysm—increases the likelihood of CAD. Elevated blood pressure, xanthomas and retinal exudates point to the presence of CAD risk factors. Palpation of the chest wall often reveals tender areas in patients whose chest pain is caused by musculoskeletal chest wall syndromes (35). However, pain produced by pressure on the chest wall may be present even if the patient has angina due to ischemic heart disease. The presence of a rub will point to pericardial or pleural disease.
|
Developing the probability estimate
|
|---|
When the initial history and physical are complete, the physician and patient find themselves asking the same question: "Is it the heart?" In certain instances, the physician can confidently assure the patient that it is not. Patients with noncardiac chest pain are generally at lower risk for ischemic heart disease. As indicated on the flow diagram, the history and appropriate diagnostic tests will usually focus on noncardiac causes of chest pain. Appropriate treatment and follow-up for the noncardiac condition can be prescribed, and the patient can be educated about CAD and risk factors, especially if he or she rarely sees a physician.
When there is sufficient suspicion of heart disease to warrant cardiac evaluation, the clinician should make a probability estimate of the likelihood of CAD. The importance of doing so is obvious when considering how this estimate affects the utility of a commonly used diagnostic test: the standard exercise test. Consider how interpretation of the standard exercise test would be affected by varying the pretest probability of disease from 5% to 50% to 90% (36). In this example, the exercise test is considered positive if 1-mm ST-segment depression is observed. The test sensitivity is 50% and specificity 90% (14).
In patients with a low probability of CAD (5%), the positive predictive value of an abnormal test result is only 21%. If 1,000 low-probability patients are tested, 120 will test positive. Of these, 95 will not have significant CAD. Before testing such a group, the clinician must weigh the value of correctly diagnosing CAD in 25 patients against the cost of a stress test for all 1,000 patients plus the cost of misdiagnosis—undue anxiety, further invasive testing, unnecessary medications or higher insurance premiums—for the 95 patients with a false-positive test result. In patients with a high probability of CAD (90%), a positive test result raises the probability of disease to 98% and a negative test result lowers probability to 83%. Although exercise testing has prognostic value in these patients (see Section III, C-2) (37), a negative test result obviously does not allow the clinician to discard the diagnosis of CAD. In patients with a 50% probability of CAD, a positive test result increases the likelihood of disease to 83% and a negative test result decreases the likelihood to 36%. The test separates this group of patients into two distinct subgroups: one in whom CAD almost certainly exists and the other for whom the diagnosis, although far from being excluded, is doubtful. An accurate estimate of the likelihood of CAD is necessary for interpretation of further test results and good clinical decision making about therapy.
Although it may seem premature to predict the probability of CAD after the history and physical, the clinicopathological study performed by Diamond and Forrester (38) demonstrated that it is possible. By combining data from a series of angiography studies performed in the 1960s and the 1970s, they showed that the simple clinical observations of pain type, age, and gender were powerful predictors of the likelihood of CAD. For instance, a 64-year-old man with typical angina has a 94% likelihood of having significant CAD. A 32-year-old woman with nonanginal chest pain has a 1% chance of CAD (14).
The value of the Diamond and Forrester approach was subsequently confirmed in prospective studies at Duke and Stanford. In these studies, both men and women were referred to cardiology specialty clinics for cardiac catheterization (39,40) or cardiac stress testing (41), and the initial clinical exam characteristics most helpful in predicting CAD were determined. With these characteristics, predictive models (logistic regression equations) were developed. When prospectively applied to another group of patients referred to the same specialty clinic, the models worked well. As in Diamond and Forrester’s original work, age, gender and pain type were the most powerful predictors. Other characteristics that strengthened the predictive abilities of the models were smoking (defined as a history of smoking half a pack or more of cigarettes per day within five years of the study or at least 25 pack-years), Q wave or ST-T-wave changes, hyperlipidemia (defined as a cholesterol level >250 mg/dL) and diabetes (glucose >140). Of these risk factors, diabetes had the greatest influence on increasing risk. Other significant risk factors, such as family history and hypertension, were not as strongly predictive and did not improve the power of equations.
Although these models worked well prospectively in the settings in which they were developed, clinicians must assess how reliable they will be when used in their own practices. The Diamond and Forrester probabilities were compared with those published in the Coronary Artery Surgery Study (CASS) (42), a large 15-center study that compared clinical and angiographic findings in >20,000 patients. In both studies, probability tables were presented in which patients were categorized by age, gender, and pain type. Tables with 24 patient groupings were published. With the exception of adults <50 years old with atypical angina for whom the CASS data estimated a probability of disease 17% higher than the Diamond-Forrester data, the agreement between studies was very close: the difference averaged 5%. Because the results were so similar, the committee combined the probabilities from both studies in one evidence table (Table 9).
It is more difficult to compare the Duke data directly with the CASS and Diamond-Forrester tables because within each age, gender, and pain type grouping, the patient’s predicted probability of disease varies, depending on the presence or absence of electrocardiogram (ECG) findings (Q waves or ST-T changes) or risk factors (smoking, diabetes, hyperlipidemia). Table 10 presents the Duke data for mid-decade patients (35, 45, 55, and 65 years old). Two probabilities are given. The first is for a low-risk patient with no risk factors and a normal ECG. The second is for a high-risk patient who smokes and has diabetes and hyperlipidemia but has a normal ECG. The presence of ECG changes would increase the probability of coronary disease even more. When Tables 9 and 10 are compared, the correlation between studies is quite strong. Apparent in the Duke data is the importance of risk factors in modifying the likelihood of disease. This becomes more important the younger the patient and the more atypical the pain. For example, the likelihood of disease for women <55 years old with atypical angina and no risk factors is <10%, but if diabetes, smoking and hyperlipidemia are present, the likelihood jumps to 40%.
View this table:
[in this window]
[in a new window]
|
Table 10 Comparing Pretest Likelihoods of CAD in Low-Risk Symptomatic Patients With High-Risk Symptomatic Patients—Duke Database (41)
|
|
|
Applicability of models to primary-care practices
|
|---|
All the studies mentioned above were university-based. The patients used to develop the models were largely referred. The only study that directly looked at applicability of the university-derived model to primary-care practices was the Stanford study (40). The university-derived equation was used and the likelihood of CAD was predicted for patients presenting to two urban primary-care clinics. The equation worked well for typical angina patients but substantially overpredicted CAD for patients at less risk.
Referral (or ascertainment) bias in these studies likely explains these differences (43,44) because the clinical decision-making process before the patient was referred is unknown. Primary-care providers do not unselectively refer all chest pain patients for cardiac evaluation. The disease probabilities for high-risk patients will vary little from the study because few primary-care physicians will fail to recommend cardiac evaluation for typical angina patients. However, younger patients with less classic pain stories will often be referred only after therapeutic trials, time or noncardiac diagnostic studies fail to eliminate CAD as a possibility. Correcting for referral bias is required before these models can be applied to primary-care practices. The Stanford study showed that it was possible to correct the model predictions by using the overall prevalence of CAD in the primary care population (40). Unfortunately, while Bayesian analysis might help a primary care provider improve the models, there are no studies examining how accurately providers calculate the prevalence of CAD among their chest pain patients, or how the prevalence of CAD varies among primary care settings. Primary-care physicians must therefore exercise caution when using these predictive equations, tables, or nomograms with patients presenting for the first time with chest pain. Whether the difference between the model estimates and actual likelihood of CAD is great enough to lead to a different diagnostic and therapeutic strategy is not known.
Ideally, the strategy a clinician uses to evaluate a patient with chest pain will also take into account the patient’s preferences. Two patients with the same pretest probability of CAD may prefer different approaches because of variations in personal beliefs, economic situation or stage of life. Patient preference studies that inform physicians about what is an acceptable balance between the underdiagnosis and overdiagnosis of CAD have not been done.
|
B. Associated conditions
|
|---|
Recommendations for initial laboratory tests for diagnosis.
Class I
- Hemoglobin. (Level of Evidence: C)
- Fasting glucose. (Level of Evidence: C)
- Fasting lipid panel, including total cholesterol, HDL cholesterol, triglycerides, and calculated LDL cholesterol. (Level of Evidence: C)
Using information gathered from the history and physical examination, the clinician should consider possibilities other than CAD in the differential diagnosis, because a number of other conditions can both cause and contribute to angina. In those patients with risk factors for CAD but an otherwise low probability history for angina, alternative diagnoses should be considered (Table 11).
In all patients, particularly those with typical angina, comorbid conditions that may precipitate "functional" angina (i.e., myocardial ischemia in the absence of significant anatomic coronary obstruction) should be considered. Generally, these are pathological entities that cause myocardial ischemia either by placing increased myocardial oxygen demands on the heart or by decreasing the myocardial oxygen supply (Table 12).
Increased oxygen demand can be produced by such entities as hyperthermia, hyperthyroidism, and cocaine abuse. Hyperthermia, particularly if accompanied by volume contraction due to diaphoresis or other fluid losses, can precipitate angina in the absence of significant CAD (47).
Hyperthyroidism, with its associated tachycardia and increased metabolic rate, increases oxygen demand and, perhaps because of inceased platelet aggregation, may also decrease supply. These effects can readily lead to angina. In addition, elderly patients may not present with a typical clinical picture of thyrotoxicosis. Therefore, this possibility should be considered in the setting of minimal risk factors accompanied by a history of typical angina, particularly in older patients.
Sympathomimetic toxicity, of which cocaine is the prototype, not only increases myocardial oxygen demand but, through coronary vasospasm, simultaneously decreases supply, sometimes leading to infarction in young patients. Long-term cocaine use may also lead to development of angina by causing premature development of CAD (48).
Angina may occur in patients with severe uncontrolled hypertension due to increased wall tension, which increases myocardial oxygen demand, and increased left ventricular (LV) end-diastolic pressure, which decreases subendocardial perfusion. These same mechanisms contribute to angina in hypertrophic cardiomyopathy and aortic stenosis; however, in these conditions, wall tension may be even greater due to an outflow tract gradient, and end-diastolic pressure may be even higher due to severe LV hypertrophy.
Sustained tachycardia, either ventricular or supraventricular, may also increase myocardial oxygen demand. Paroxysmal tachycardias are more frequent conditions that contribute to angina. Unfortunately, they are often more difficult to diagnose.
Conditions that reduce myocardial oxygen supply must also be considered in the differential diagnosis of patients with angina.
Anemia reduces the oxygen-carrying capacity of the blood and also increases the cardiac workload. An increased cardiac output is associated with <9 g/dL of hemoglobin, and ST-T wave changes (depression or inversion) may be seen when hemoglobin drops below 7 g/dL.
Hypoxemia resulting from pulmonary disease (e.g., pneumonia, asthma, chronic obstructive pulmonary disease, pulmonary hypertension, interstitial fibrosis, obstructive sleep apnea) may also precipitate angina. Obstructive sleep apnea should be seriously considered in patients with only nocturnal symptoms.
Conditions that are associated with increased blood viscosity can increase coronary resistance and thereby decrease coronary artery blood flow, precipitating angina in patients without severe coronary stenoses. Increased viscosity is seen with polycythemia, leukemia, thrombocytosis and hypergammaglobulinemia.
|
C. Noninvasive testing
|
|---|
1. ECG/chest X-Ray.
Recommendations for Electrocardiography, Chest X-Ray, or Electron Beam Computed Tomography in the Diagnosis of Chronic Stable Angina
Class I
- Rest ECG in patients without an obvious non-cardiac cause of chest pain. (Level of Evidence: B)
- Rest ECG during an episode of chest pain. (Level of Evidence: B)
- Chest X-ray in patients with signs or symptoms of congestive heart failure, valvular heart disease, pericardial disease, or aortic dissection/aneurysm. (Level of Evidence: B)
Class IIa
Chest X-ray in patients with signs or symptoms of pulmonary disease. (Level of Evidence: B)
Class IIb
- Chest X-ray in other patients. (Level of Evidence: C)
- Electron beam computed tomography (EBCT). (Level of Evidence: B)
A rest 12-lead ECG should be recorded in all patients with symptoms suggestive of angina pectoris; however, it will be normal in 50% of patients with chronic stable angina (49). A normal rest ECG does not exclude severe CAD. ECG evidence of LV hypertrophy or ST-T wave changes consistent with myocardial ischemia favor the diagnosis of angina pectoris (50). Evidence of prior Q-wave MI on the ECG makes CAD very likely. However, certain Q-wave patterns are equivocal, such as an isolated Q in lead III or a QS pattern in leads V1 and V2.
The presence of arrhythmias such as atrial fibrillation or ventricular tachyarrhythmia on the ECG in patients with chest pain also increases the probability of underlying CAD; however, these arrhythmias are frequently caused by other types of cardiac disease. Various degrees of AV block can be present in patients with chronic CAD but have many other causes and a very low specificity for the diagnosis. Left anterior fascicular block, right bundle-branch block and left bundle-branch block often occur in patients with CAD and frequently indicate the presence of multivessel CAD. However, these findings also lack specificity in the diagnosis of chronic stable angina.
An ECG obtained during chest pain is abnormal in  50% of patients with angina who have a normal rest ECG. Sinus tachycardia occurs commonly; bradyarrhythmia is less common. The ST-segment elevation or depression establishes a high likelihood of angina and indicates ischemia at a low workload, portending an unfavorable prognosis. Many high-risk patients need no further noninvasive testing. Coronary arteriography usually defines the severity of coronary artery stenoses and the necessity and feasibility of myocardial revascularization. In patients with ST-T wave depression or inversion on the rest ECG, "pseudonormalization" of these abnormalities during pain is another indicator that CAD is likely (51). The occurrence of tachyarrhythmias, atrioventricular block, left anterior fascicular block or bundle-branch block with chest pain also increases the probability of coronary heart disease (CHD) and often leads to coronary arteriography.
The chest roentgenogram is often normal in patients with stable angina pectoris. Its usefulness as a routine test is not well established. It is more likely to be abnormal in patients with previous or acute MI, those with a noncoronary artery cause of chest pain and those with noncardiac chest discomfort. Cardiac enlargement may be attributable to previous MI, acute LV failure, pericardial effusion or chronic volume overload of the left ventricle such as occurs with aortic or mitral regurgitation. Abnormal physical findings, associated chest X-ray findings (e.g., pulmonary venous congestion), and abnormalities detected by noninvasive testing (echocardiography) may indicate the correct etiology.
Enlargement of the upper mediastinum often results from an ascending aortic aneurysm with or without dissection. Pruning or cutoffs of the pulmonary arteries or areas of segmental oligemia may indicate pulmonary infarction/embolism or other causes of pulmonary hypertension.
Coronary artery calcification increases the likelihood of symptomatic CAD. Fluoroscopically detectable coronary calcification is correlated with major-vessel occlusion in 94% of patients with chest pain (52); however, the sensitivity of the test is only 40%.
Ultrafast computed tomography.
Ultrafast (electron beam) computed tomography is being used with increased frequency for the detection and quantification of coronary artery calcification (25). In seven studies including 50 to 710 patients, calcium of the coronary arteries detected by EBCT was an important indicator of angiographic coronary stenoses. In these studies of selected patients, the sensitivity of a positive EBCT detection of calcium for the presence of CAD varied from 85% to 100%; specificity ranged from only 41% to 76%; the positive predictive value varied considerably from 55% to 84% and negative predictive value from 84% to 100% (25). The presence and amount of calcium detected in coronary arteries by EBCT in two studies appeared to correlate with the presence and associated amount of atherosclerotic plaque (53,54). However, several studies (55–57) have shown a marked variability in repeated measures of coronary calcium by EBCT. Therefore, the use of serial EBCT scans in individual patients for identification and serial assessment of the progression or regression of calcium remains problematic. The proper role of EBCT is controversial and will be the subject of future ACC/AHA statements.
|
2. Exercise ECG for diagnosis
|
|---|
Recommendations for Diagnosis of Obstructive CAD With Exercise ECG Testing Without an Imaging Modality
Class I.
Patients with an intermediate pretest probability of CAD based on age, gender and symptoms, including those with complete right bundle-branch block or <1 mm of ST depression at rest (exceptions are listed below in classes II and III). (Level of Evidence: B)
Class IIa.
Patients with suspected vasospastic angina. (Level of Evidence: C)
Class IIb.
- Patients with a high pretest probability of CAD by age, gender and symptoms. (Level of Evidence: B)
- Patients with a low pretest probability of CAD by age, gender and symptoms. (Level of Evidence: B)
- Patients taking digoxin whose ECG has <1 mm of baseline ST-segment depression. (Level of Evidence: B)
- Patients with ECG criteria for LV hypertrophy and <1 mm of baseline ST-segment depression. (Level of Evidence: B)
Class III.
- Patients with the following baseline ECG abnormalities.
- Pre-excitation (Wolff-Parkinson-White) syndrome. (Level of Evidence: B)
- Electronically paced ventricular rhythm. (Level of Evidence: B)
- More than 1 mm of ST depression at rest. (Level of Evidence: B)
- Complete left bundle-branch block. (Level of Evidence: B)
- Patients with an established diagnosis of CAD due to prior MI or coronary angiography; however, testing can assess functional capacity and prognosis, as discussed in section III. (Level of Evidence: B)
Description of the exercise testing procedure.
Exercise testing is a well-established procedure that has been in widespread clinical use for many decades. Detailed descriptions of exercise testing are available in other publications (58–60). This section provides a brief overview based on the "ACC/AHA Guidelines for Exercise Testing" (14).
Although exercise testing is generally a safe procedure, both MI and death occur at a rate of  1/2500 tests (61). The absolute contraindications to exercise testing include acute MI within two days, cardiac arrhythmias causing symptoms or hemodynamic compromise, symptomatic and severe aortic stenosis, symptomatic heart failure, acute pulmonary embolus or pulmonary infarction, acute myocarditis or pericarditis and acute aortic dissection (14,59). Additional factors are relative contraindications: left main coronary stenosis, moderate aortic stenosis, electrolyte abnormalities, systolic hypertension >200 mm Hg, diastolic blood pressure >110 mm Hg, tachyarrhythmias or bradyarrhythmias, hypertrophic cardiomyopathy and other forms of outflow tract obstruction, mental or physical impairment leading to an inability to exercise adequately and high-degree atrioventricular block (14,59). In the past, unstable angina was a contraindication to exercise testing. However, new information suggests that exercise treadmill (62–64) and pharmacologic (65–68) testing are safe in low-risk outpatients with unstable angina and in low- or intermediate-risk patients hospitalized with unstable angina in whom an MI has been ruled out and who are free of angina and congestive heart failure.
Both treadmill and cycle ergometer devices are used for exercise testing. Although cycle ergometers have important advantages, fatigue in the quadricep muscles in patients who are not experienced cyclists usually makes them stop before reaching their maximum oxygen uptake. As a result, treadmills are more commonly used in the U.S.
There are clear advantages in customizing the protocol to the individual patient to allow exercise lasting 6 to 12 min (69). Exercise capacity should be reported in estimated METs of exercise. (One MET is the standard basal oxygen uptake of 3.5 mL/kg per min.) If exercise capacity is also reported in minutes, the protocol should be described clearly.
Exercise testing should be supervised by an appropriately trained physician (70), although personal supervision (as defined by the Health Care Financing Administration [HCFA]) is not always required. The ECG, heart rate and blood pressure should be carefully monitored and recorded during each stage of exercise as well as during ST-segment abnormalities and chest pain. The patient should be monitored continuously for transient rhythm disturbances, ST-segment changes and other ECG manifestations of myocardial ischemia. Although exercise testing is commonly terminated when subjects reach a standard percentage (often 85%) of age-predicted maximum heart rate, there is great variability in maximum heart rates among individuals, so predicted values may be supramaximal for some patients and submaximal for others. Therefore, it is important to monitor the patient closely for other indications for stopping the test. Absolute indications for stopping include a drop in systolic blood pressure of >10 mm Hg from baseline blood pressure despite an increase in workload when accompanied by other evidence of ischemia; moderate to severe angina; increasing ataxia, dizziness or near syncope; signs of poor perfusion such as cyanosis or pallor; technical difficulties monitoring the ECG or systolic blood pressure; the subject’s desire to stop; sustained ventricular tachycardia; or ST elevation 1 mm in leads without diagnostic Q waves (other than V1 or aVR). Relative indications for stopping include a drop in systolic blood pressure of >10 mm Hg from baseline blood pressure despite an increase in workload in the absence of other evidence of ischemia; >2 mm of horizontal or downsloping ST-segment depression; marked axis deviation; arrhythmias such as multifocal premature ventricular complexes (PVCs), triplets of PVCs, supraventricular tachycardia, heart block or bradyarrhythmias; symptoms such as fatigue, shortness of breath, wheezing, leg cramps or claudication; bundle-branch block or IVCD that cannot be distinguished from ventricular tachycardia; increasing chest pain; systolic blood pressure >250 mm Hg; or diastolic blood pressure >115 mm Hg (59). Rating the level of perceived exertion with the Borg scale (71) helps measure patient fatigue, and fatigue-limited testing is especially important when assessing functional capacity.
Interpretation of the exercise test.
Interpretation of the exercise test should include symptomatic response, exercise capacity, hemodynamic response, and ECG response. The occurrence of ischemic chest pain consistent with angina is important, particularly if it forces termination of the test. Abnormalities in exercise capacity, systolic blood pressure response to exercise and heart rate response to exercise are important findings. The most important ECG findings are ST depression and ST elevation. The most commonly used definition for a positive exercise test is 1 mm of horizontal or downsloping ST-segment depression or elevation for 60 to 80 ms after the end of the QRS complex, either during or after exercise (14).
Cost and availability.
The exercise ECG is the least costly diagnostic test, with the cost of stress echocardiography 2-fold higher, stress SPECT myocardial imaging at least 5-fold higher, and coronary angiography 20-fold higher. A lower cost of the treadmill exercise test alone does not necessarily result in a lower overall cost of patient care, however, as the cost of additional testing and intervention may be higher because the exercise test is less accurate.
Treadmill exercise tests are performed frequently but somewhat less often than the most frequent imaging procedure, which is stress SPECT myocardial perfusion imaging. An estimated two thirds of the treadmill exercise tests charged to Medicare in 1994 were performed as office procedures, and 33% of these charges were submitted by noncardiologists (14).
|
Rationale
|
|---|
Diagnostic characteristics of exercise tests.
The sensitivity of the exercise test measures the probability that a patient with obstructive CAD will have a positive test result, whereas the specificity measures the probability that a patient without obstructive CAD will have a negative test result. Sensitivity and specificity are used to summarize the characteristics of diagnostic tests because they provide standard measures that can be used to compare different tests. Sensitivity and specificity alone, however, do not provide the information needed to interpret the results of exercise testing. That information can be calculated and expressed as predictive values. These calculations require the sensitivity and specificity of the exercise test along with the pretest probability that the patient has obstructive CAD.
The numerator refers to positive test results that are true-positive, and the denominator refers to all positive test results, true-positive and false-positive. The positive predictive value is the probability that the patient has obstructive CAD when the exercise test result is positive.
The numerator refers to negative test results that are true-negative, and the denominator refers to all negative test results, both true-negative and false-negative. The negative predictive value is the probability that the patient does not have obstructive CAD when the exercise test result is negative. Therefore, knowing the sensitivity and specificity of the exercise test and the patient’s pretest probability of obstructive CAD is especially important when interpreting the results of exercise testing.
When interpreting estimates of the sensitivity and specificity of exercise testing, it is important to recognize a type of bias called workup verification, or posttest referral bias. This type of bias occurs when the results of exercise testing are used to decide which patients have the diagnosis of CAD verified or ruled out with a gold-standard procedure. This bias also occurs when patients with positive results on exercise testing are referred for coronary angiography and patients with negative results are not. Such a selection process curtails the number of true-negative results. The result of this type of bias is to raise the measured sensitivity and lower the measured specificity in relation to their true values.
Sensitivity and specificity of the exercise test.
A meta-analysis of 147 published reports describing 24,074 patients who underwent both coronary angiography and exercise testing found wide variation in sensitivity and specificity (14). Mean sensitivity was 68% with a standard deviation of 16%; mean specificity was 77% with a standard deviation of 17%. When the analysis considered only results from the 58 studies that focused on diagnostic tests by excluding patients with a prior MI, mean sensitivity was 67% and mean specificity 72%. When the analysis was restricted to the few studies that avoided workup bias by including only patients who agreed before any testing to have both exercise testing and coronary angiography, sensitivity was 50% and specificity 90% (73,74). In a more recent study of 814 men that was carefully designed to minimize workup bias, sensitivity was 45% and specificity 85% (75). Therefore, the true diagnostic value of the exercise ECG lies in its relatively high specificity. The modest sensitivity of the exercise ECG is generally lower than the sensitivity of imaging procedures (12,13).
Although the sensitivity and specificity of a diagnostic test are usually thought to be characteristics of the tests themselves and not affected by patient differences, this is not always the case. For instance, the exercise test has a higher sensitivity in the elderly and persons with three-vessel disease than in younger persons and those with one-vessel disease. The test has a lower specificity in those with valvular heart disease, LV hypertrophy and rest ST depression and those taking digoxin (14).
Physicians are often urged to consider more than just the ST segment when interpreting the exercise test, and some studies that use complex formulas to incorporate additional test information have found diagnoses made with this approach to be more accurate than those based only on the ST response (76,77). However, the diagnostic interpretation of the exercise test still centers around the ST response because different studies produce different formulas, and the formulas provide similar results when compared with the judgment of experienced clinical cardiologists (75,78,79).
Pretest probability.
Diagnostic testing is most valuable when the pretest probability of obstructive CAD is intermediate: for example, when a 50-year-old man has atypical angina and the probability of CAD is 50% (see Table 9). In these conditions, the test result has the largest effect on the posttest probability of disease and thus on clinical decisions.
The exact definition of the upper and lower boundaries of intermediate probability (e.g., 10% and 90%, 20% and 80%, 30% and 70%) is a matter of physician judgment in an individual situation. Among the factors relevant to the choice of these boundaries are the degree of uncertainty that is acceptable to physician and patient; the likelihood of an alternative diagnosis; the reliability, cost, and potential risks of further testing; and the benefits and risks of treatment in the absence of additional testing. Pauker and Kassirer (80) have described the application of decision analysis to this important issue. As indicated earlier, it should be recognized that the initial evaluation of patients with noncardiac pain will focus on noncardiac conditions. Clinical judgment in such patients may indicate that they are at low probability and do not require cardiac evaluation.
For the diagnosis of CAD, one possible arbitrary definition of intermediate probability that appears in published research is between 10% and 90%. This definition was first advocated 20 years ago (81), and has been used in several studies (82,83) and the "ACC/AHA Guidelines for Exercise Testing" (14). Although this range may seem very broad, many sizable patient groups (e.g., older men with typical angina and younger women with nonanginal pain) fall outside the intermediate probability range. When the probability of obstructive CAD is high, a positive test result only confirms the high probability of disease, and a negative test result may not decrease the probability of disease enough to make a clinical difference. Although the exercise test is less useful for the diagnosis of CAD when pretest probability is high, it can provide information about the patient’s risk status and prognosis (see Section III). When the probability of obstructive CAD is very low, a negative test result only confirms the low probability of disease, and a positive test result may not increase the probability of disease enough to make a clinical difference.
Influence of other factors on test performance.
Digoxin
Digoxin produces abnormal exercise-induced ST depression in 25% to 40% of apparently healthy normal subjects (84,85). The prevalence of abnormal responses is directly related to age.
Beta-adrenergic blocking agent therapy
Whenever possible, it is recommended that beta-blockers (and other anti-ischemic drugs) be withheld for four to five half-lives (usually about 48 h) before exercise stress testing for the diagnosis and initial risk stratification of patients with suspected CAD. Ideally, these drugs should be withdrawn gradually to avoid a withdrawal phenomenon that may precipitate events (86,87). When beta-blockers cannot be stopped, stress testing may detect myocardial ischemia less reliably, but it usually will still be positive in patients at the highest risk.
Other drugs
Antihypertensive agents and vasodilators can affect test performance by altering the hemodynamic response of blood pressure. Short-term administration of nitrates can attenuate the angina and ST depression associated with myocardial ischemia. Flecainide has been associated with exercise-induced ventricular tachycardia (88,89).
Left bundle-branch block
Exercise-induced ST depression usually occurs with left bundle-branch block and is not associated with ischemia (90).
Right bundle-branch block
Exercise-induced ST depression usually occurs with right bundle-branch block in the anterior chest leads (V1–3) and has no association with ischemia (91). However, when it occurs in the left chest leads (V5,6) or inferior leads (II, aVF), it has the same significance as it does when the rest ECG is normal.
LV hypertrophy with repolarization abnormality
on the rest ECG is associated with more false-positive test results due to decreased specificity.
Rest ST-segment depression
is a marker for adverse cardiac events in patients with and without known CAD (92–99). Additional exercise-induced ST-segment depression in the patient with 1 mm rest ST-segment depression is a reasonably sensitive indicator of CAD.
|
ST-segment interpretation issues
|
|---|
Lead selection.
Twelve-lead ECGs provide the greatest sensitivity. The V5 alone consistently outperforms the inferior leads and the combination of V5 with lead II. In patients without prior MI and with a normal rest ECG, the precordial leads alone are a reliable marker for CAD. In patients with a normal rest ECG, exercise-induced ST-segment depression confined to the inferior leads is of little value (100).
Upsloping ST depression.
Patients with ST-segment depression that slopes upward at <1 mV/s probably have an increased probability of coronary disease (101,102). However, the ACC/AHA/ACP-ASIM Committee to Develop Guidelines for the Management of Chronic Stable Angina favors the use of the more common definition for a positive test, which is 1 mm of horizontal or downsloping ST depression or elevation for 60 to 80 milliseconds after the end of the QRS complex (72), because most of the published literature is based on this definition.
Atrial repolarization.
Atrial repolarization waves are opposite in direction to P waves and may extend into the ST segment and T wave. Exaggerated atrial repolarization waves during exercise can cause downsloping ST depression in the absence of ischemia (103,104). Patients with false-positive exercise tests have a high peak exercise heart rate, absence of exercise-induced chest pain, and markedly downsloping PR segments in the inferior leads. This issue of atrial repolarization waves was not addressed in the "ACC/AHA Guidelines for Exercise Testing" (14).
ST elevation.
When the rest ECG is normal, ST elevation (other than in aVR or V1) is very rare, represents transmural ischemia caused by spasm or a critical lesion, greatly increases the likelihood of arrhythmias, and localizes the ischemia. When the rest ECG shows Q waves from an old MI, the significance of ST elevation is controversial. Some studies have suggested that it is due to wall motion abnormalities (105,106); other studies (107–109) have found it to be a marker of residual viability in the infarcted area.
R-wave changes.
A multitude of factors affect the R-wave response to exercise (110), and the response does not have diagnostic significance (111,112).
Heart rate adjustment.
Several methods of heart rate adjustment have been proposed to increase the diagnostic accuracy of the exercise ECG (113–116), but there is no convincing evidence of benefit (115–119). It is more important to consider exercise capacity than heart rate.
Computer processing.
Although computer processing of the exercise ECG can be helpful, it can also result in false-positive ST depression (120). To avoid this problem, the interpreting physician should always compare the unprocessed ECG with any computer-generated averages.
|
Special groups
|
|---|
Women.
The use of exercise testing in women presents difficulties that are not experienced in men. These difficulties reflect the differences between men and women regarding the prevalence of CAD and the sensitivity and specificity of exercise testing.
Although obstructive CAD is one of the principal causes of death in women, the prevalence (and thus the pretest probability) of this disease is lower in women than it is in men of comparable age, especially in premenopausal women. When compared with men, the lower pretest probability of disease in women means that more test results are false-positive. For example, almost half the women with anginal symptoms in the CASS study, many of whom had positive exercise test results, had normal coronary arteriograms (121).
Exercise testing is less sensitive in women than it is in men, and some studies have found it also to be less specific (14,73,83,122–131). Among the proposed reasons for these differences are the use of different criteria for defining coronary disease, differences in the prevalence of multivessel disease and prior MI, differences in the criteria for ST-segment positivity (132,133), differences in type of exercise, the inability of many women to exercise to maximum aerobic capacity (134,135), the greater prevalence of mitral valve prolapse and syndrome X in women, differences in microvascular function (leading perhaps to coronary spasm) and possibly, hormonal differences. To compensate for the limitations of the test in women, some investigators have developed predictive models that incorporate more information from the test than simply the amount and type of ST-segment change (130,131). Although this approach is attractive, its clinical application remains limited.
The difficulties of using exercise testing for diagnosing obstructive CAD in women have led to speculation that stress imaging may be preferred over standard stress testing (129). Although the optimal strategy for diagnosing obstructive CAD in women remains to be defined, the ACC/AHA/ACP-ASIM Committee to Develop Guidelines for the Management of Chronic Stable Angina believes there are currently insufficient data to justify replacing standard exercise testing with stress imaging when evaluating women for CAD. In many women with a low pretest likelihood of disease, a negative exercise test result will be sufficient, and imaging procedures will not be required (83).
The elderly.
Few data have been published about the use of exercise testing in people 70 years old. The 1989 National Health Interview Survey (136) found that the diagnosis of CAD was reported by 1.8% in men and 1.5% in women >75 years old. Silent ischemia is estimated to be present in 15% of 80-year-olds (137).
The performance of exercise testing poses additional problems in the elderly. Functional capacity often is compromised from muscle weakness and deconditioning, making the decision about an exercise test versus a pharmacologic stress test more important. More attention must be given to the mechanical hazards of exercise, and less challenging protocols should be used (138). Elderly patients are more likely to hold the hand rails tightly, thus reducing the validity of treadmill time for estimating METs. Arrhythmias occur more frequently with increasing age, especially at higher workloads (138). In some patients with problems of gait and coordination, a bicycle exercise test may be more attractive (139), but bicycle exercise is unfamiliar to most elderly patients.
The interpretation of exercise test results in the elderly differs from that in the young. The greater severity of coronary disease in this group increases the sensitivity of exercise testing (84%), but it also decreases the specificity (70%). The high prevalence of disease means that more test results are false-negative (140). False-positive test results may reflect the coexistence of LV hypertrophy from valvular disease and hypertension, as well as conduction disturbances. Other rest ECG abnormalities that complicate interpretation, including prior MI, also are more frequent.
Exercise testing in the elderly is more difficult both to do and to interpret, and the follow-up risks of coronary angiography and revascularization are greater. Despite these differences, exercise testing remains important in the elderly because the alternative to revascularization is medical therapy, which also has greater risks in this group.
|
3. Echocardiography (resting)
|
|---|
Recommendations for Echocardiography for Diagnosis of Cause of Chest Pain in Patients With Suspected Chronic Stable Angina Pectoris
Class I.
- Patients with systolic murmur suggestive of aortic stenosis or hypertrophic cardiomyopathy. (Level of Evidence: C)
- Evaluation of extent (severity) of ischemia (e.g., LV segmental wall motion abnormality) when the echocardiogram can be obtained during pain or within 30 min after its abatement. (Level of Evidence: C)
Class IIb.
Patients with a click or murmur to diagnose mitral valve prolapse (15). (Level of Evidence: C)
Class III.
Patients with a normal ECG, no history of MI and no signs or symptoms suggestive of heart failure, valvular heart disease, or hypertrophic cardiomyopathy. (Level of Evidence: C)
Echocardiography can be a useful tool for assisting in establishing a diagnosis of CAD. Echocardiography can also assist in defining the consequences of coronary disease in selected patients with chronic chest pain presumed to be chronic stable angina. However, most patients undergoing a diagnostic evaluation for angina do not need an echocardiogram.
Cause of chest pain unclear: confounding or concurrent cardiac diagnoses.
Transthoracic echocardiographic imaging and Doppler recording are useful when there is a murmur or other evidence for conditions such as aortic stenosis or hypertrophic cardiomyopathy coexisting with CAD. Echo-Doppler techniques usually provide accurate quantitative information regarding the presence and severity of a coexisting lesion, such as 1) whether there is concentric hypertrophy or asymmetric hypertrophy of the ventricular septum, LV apex, or free wall; 2) the severity of any aortic valvular or subvalvular gradient; and 3) the status of LV function (13).
Echocardiography is useful for establishing or excluding the diagnosis of mitral valve prolapse and establishing the need for infective endocarditis prophylaxis (15).
Global LV systolic function.
Chronic ischemic heart disease, whether associated with angina pectoris or not, can result in impaired systolic LV function. The extent and severity of regional and global abnormalities are important considerations in choosing appropriate medical or surgical therapy. Routine estimation of parameters of global LV function, such as LV ejection fraction, is unnecessary for the diagnosis of chronic angina pectoris. For example, in patients with suspected angina and a normal ECG, no history of MI, and no signs or symptoms of heart failure, echocardiography and radionuclide imaging are not indicated (141,142).
Segmental LV wall motion abnormalities.
Echocardiographic findings that may help establish the diagnosis of chronic ischemic heart disease include regional systolic wall motion abnormalities, e.g., hypokinesis (reduced wall motion), akinesis (absence of wall motion), dyskinesis (paradoxical wall motion) and failure of a wall segment to thicken normally during systole (13). Care must be taken to distinguish chronic CAD as a cause of ventricular septal wall motion abnormalities from other conditions, such as left bundle-branch block, presence of an intraventricular pacemaker, right ventricular volume overload, or prior cardiac surgery (13).
The extent of regional (segmental) LV dysfunction can be described by scoring LV wall segments individually as to degree of wall motion abnormality (e.g., hypokinesis, dyskinesis, akinesis) or by using a scoring system that describes the summated wall motion score reflecting the normality or abnormality of each segment (143–146). Segmental wall motion abnormalities are often detected in patients with a prior history of MI or significant Q waves on their ECGs. Their locations correlate well with the distribution of CAD and pathologic evidence of infarction (143,144,147–154). Regional wall motion abnormalities can also be seen in patients with transient myocardial ischemia, chronic ischemia (hibernating myocardium) and myocardial scar and in some patients with myocarditis or other conditions not associated with coronary occlusion (13). In patients in whom the LV endocardium is suboptimally imaged by standard transthoracic echocardiography, tissue harmonic imaging (155,156) with newer transducers and contrast echocardiography with intravenous injections of encapsulated gaseous microbubbles represent promising new solutions (157–159).
In patients with chronic stable angina pectoris without previous MI, LV wall motion is typically normal on the rest echocardiogram in the absence of ischemia. However, in the uncommon situation in which an echocardiogram can be recorded during ischemia or, in some cases (e.g., with stunned myocardium), up to 30 min after ischemia, the presence of regional systolic wall motion abnormalities (in a patient without known CAD) is a moderately accurate indicator of an increased likelihood of clinically significant CAD. According to pooled data, the positive predictive accuracy of this finding for acute ischemia or infarction has been reported to be 50% (13). Conversely, the absence of regional wall motion abnormalities identifies a subset of patients at low risk for an acute infarction (147,160), with a pooled negative predictive accuracy of about 95%.
Ischemic mitral regurgitation.
Other structural and functional alterations can complicate chronic ischemic heart disease associated with stable angina pectoris. Mitral regurgitation may result from global LV systolic dysfunction (161), regional papillary muscle dysfunction (162), scarring and shortening of the submitral chords (163), papillary muscle rupture (164), or other causes. The presence, severity and mechanism of mitral regurgitation can be reliably detected with transthoracic imaging and Doppler echocardiographic techniques. Potential surgical approaches to mitral valve repair or replacement can also be defined echocardiographically (15).
|
4. Stress imaging studies: echocardiographic and nuclear
|
|---|
Recommendations for Cardiac Stress Imaging as the Initial Test for Diagnosis in Patients With Chronic Stable Angina Who Are Able to Exercise
Class I.
- Exercise myocardial perfusion imaging or exercise echocardiography in patients with an intermediate pretest probability of CAD who have one of the following baseline ECG abnormalities:
- Pre-excitation (Wolff-Parkinson-White) syndrome. (Level of Evidence: B)
- More than 1 mm of ST depression at rest. (Level of Evidence: B)
- Exercise myocardial perfusion imaging or exercise echocardiography in patients with prior revascularization (either PTCA or CABG). (Level of Evidence: B)
- Adenosine or dipyridamole myocardial perfusion imaging in patients with an intermediate pretest probability of CAD and one of the following baseline ECG abnormalities:
- Electronically paced ventricular rhythm. (Level of Evidence: C)
- Left bundle-branch block. (Level of Evidence: B)
Class IIb.
- Exercise myocardial perfusion imaging and exercise echocardiography in patients with a low or high probability of CAD who have one of the following baseline ECG abnormalities:
- Pre-excitation (Wolff-Parkinson-White) syndrome. (Level of Evidence: B)
- More than 1 mm of ST depression. (Level of Evidence: B)
- Adenosine or dipyridamole myocardial perfusion imaging in patients with a low or high probability of CAD and one of the following baseline ECG abnormalities:
- Electronically paced ventricular rhythm. (Level of Evidence: C)
- Left bundle-branch block. (Level of Evidence: B)
- Exercise myocardial perfusion imaging or exercise echocardiography in patients with an intermediate probability of CAD who have one of the following:
- Digoxin use with <1 mm ST depression on the baseline ECG. (Level of Evidence: B)
- LVH with <1 mm ST depression on the baseline ECG. (Level of Evidence: B)
- Exercise myocardial perfusion imaging, exercise echocardiography, adenosine or dipyridamole myocardial perfusion imaging or dobutamine echocardiography as the initial stress test in a patient with a normal rest ECG who is not taking digoxin. (Level of Evidence: B)
- Exercise or dobutamine echocardiography in patients with left bundle-branch block. (Level of Evidence: C)
Recommendations for Cardiac Stress Imaging as the Initial Test for Diagnosis in Patients With Chronic Stable Angina Who Are Unable to Exercise
Class I.
- Adenosine or dipyridamole myocardial perfusion imaging or dobutamine echocardiography in patients with an intermediate pretest probability of CAD. (Level of Evidence: B)
- Adenosine or dipyridamole stress myocardial perfusion imaging or dobutamine echocardiography in patients with prior revascularization (either PTCA or CABG). (Level of Evidence: B)
Class IIb.
- Adenosine or dipyridamole stress myocardial perfusion imaging or dobutamine echocardiography in patients with a low or high probability of CAD in the absence of electronically paced ventricular rhythm or left bundle-branch block. (Level of Evidence: B)
- Adenosine or dipyridamole myocardial perfusion imaging in patients with a low or a high probability of CAD and one of the following baseline ECG abnormalities
- Electronically paced ventricular rhythm. (Level of Evidence: C)
- Left bundle-branch block. (Level of Evidence: B)
- Dobutamine echocardiography in patients with left bundle-branch block. (Level of Evidence: C)
When to do stress imaging.
Patients who are good candidates for cardiac stress testing with imaging, as opposed to exercise ECG, include those in the following categories (see also Section II.C.3) (14): 1) complete left bundle-branch block, electronically paced ventricular rhythm, preexcitation (Wolff-Parkinson-White) syndrome and other similar ECG conduction abnormalities; 2) patients who have >1 mm of ST-segment depression at rest, including those with LV hypertrophy or taking drugs such as digitalis; 3) patients who are unable to exercise to a level high enough to give meaningful results on routine stress ECG who should be considered for pharmacologic stress imaging tests; and 4) patients with angina who have undergone prior revascularization, in whom localization of ischemia, establishing the functional significance of lesions and demonstrating myocardial viability are important considerations.
Exercise and pharmacologic modalities used in stress imaging.
A variety of methods can be used to induce stress: 1) exercise (treadmill or upright or supine bicycle [see Section II.C.3]), and 2) pharmacologic techniques (either dobutamine or vasodilators). When the patient can exercise to develop an appropriate level of cardiovascular stress (e.g., 6 to 12 min), exercise stress testing (generally with a treadmill) is preferable to pharmacologic stress testing (see Section II.C.3). However, when the patient cannot exercise to the necessary level or in other specified circumstances (e.g., when stress echocardiography is being used in the assessment of myocardial viability), pharmacologic stress testing may be preferable. Three drugs are commonly used as substitutes for exercise stress testing: dipyridamole, adenosine and dobutamine. Dipyridamole and adenosine are vasodilators that are commonly used in conjunction with myocardial perfusion scintigraphy, whereas dobutamine is a positive inotropic (and chronotropic) agent commonly used with echocardiography.
Dipyridamole indirectly causes coronary vasodilation by inhibiting cellular uptake of adenosine, thereby increasing the blood and tissue levels of adenosine, which is a potent, direct coronary vasodilator and markedly increases coronary blood flow. The flow increase with adenosine or dipyridamole is of a lesser magnitude through stenotic arteries, creating heterogeneous myocardial perfusion, which can be observed with a perfusion tracer. Although this mechanism can exist independent of myocardial ischemia, in some patients, true myocardial ischemia can occur with either dipyridamole or adenosine because of a coronary steal phenomenon.
Both dipyridamole and adenosine are safe and well tolerated despite frequent mild side effects, which occur in 50% (165) and 80% (166,167) of patients, respectively. With dipyridamole infusion, the most common side effect was angina (18% to 42%), with arrhythmia occurring in <2%. Noncardiac side effects have included headache (5% to 23%), dizziness (5% to 21%), nausea (8% to 12%), and flushing (3%) (165). With adenosine infusion, chest pain has been reported in 57%, headache in 35%, flushing in 25%, shortness of breath in 15%, and first-degree AV block in 18%. Severe side effects are rare, but both dipyridamole and adenosine may cause severe bronchospasm in patients with asthma or chronic obstructive lung disease; therefore, they should be used with extreme caution—if at all—in these patients. Dipyridamole and adenosine side effects are antagonized by theophylline, although this drug is ordinarily not needed after adenosine because of the latter’s ultrashort half-life (<10 s).
Dobutamine in high doses (20 to 40·µg kg–1·min–1) increases the three main determinants of myocardial oxygen demand, namely, heart rate, systolic blood pressure and myocardial contractility, thereby eliciting a secondary increase in myocardial blood flow and provoking ischemia. The flow increase (2-fold to 3-fold baseline values) is less than that elicited by adenosine or dipyridamole but is sufficient to demonstrate heterogeneous perfusion by radionuclide imaging. Although side effects are frequent during dobutamine infusion, the test appears to be relatively safe, even in the elderly (168–173). The most frequently reported noncardiac side effects (total 26%) in a study of 1118 patients included nausea (8%), anxiety (6%), headache (4%) and tremor (4%) (172). Common arrhythmias included premature ventricular beats (15%), premature atrial beats (8%), and supraventricular tachycardia and nonsustained ventricular tachycardia (3% to 4%). Atypical chest pain was reported in 8% and angina pectoris in approximately 20%.
Factors affecting accuracy of noninvasive testing.
As already described for exercise ECG, apparent test performance can be altered by the pretest probability of CAD (38,174,175). The positive predictive value of a test declines as the disease prevalence decreases in the population under study, whereas the negative predictive accuracy increases (176). Stress imaging should generally not be used for routine diagnostic purposes in patients with a low or high pretest probability of disease. However, although stress imaging is less useful for diagnosis when the pretest probability of CAD is high, it can provide information about the patient’s risk status and prognosis (see Section III.C.3).
As it is for exercise electrocardiography, the phenomenon of workup, verification, or posttest referral bias is an important factor influencing the sensitivity, specificity and predictive value of myocardial perfusion imaging and stress echocardiography (see Section II.C.3). The effects of posttest referral bias have been similar for myocardial perfusion/imaging (177,178) and exercise echocardiography (179). Correction for posttest referral bias results in strikingly lower sensitivity and higher specificity for both techniques (Tables 13 through 17). As a result of these changes in sensitivity and specificity, in a patient with an intermediate pretest probability of disease, correction for verification bias actually improves the diagnostic value of a positive test result while the value of a negative test result decreases (175).
Diagnostic accuracy of stress imaging techniques.
Radionuclide imaging
An excellent review of the use of radionuclide imaging in the diagnosis and localization of CAD was included in the "ACC/AHA Guidelines for Clinical Use of Cardiac Radionuclide Imaging," which was published in 1995 (12). This discussion, which focuses on myocardial perfusion imaging, borrows from this previous document but has been updated to reflect more recent publications. In patients with suspected or known chronic stable angina, the largest accumulated experience in myocardial perfusion imaging has been with the tracer 201Tl, but the available evidence suggests that the newer tracers 99mTc sestamibi and 99mTc tetrofosmin yield similar diagnostic accuracy (180–190). Thus, for the most part, 201Tl, 99mTc sestamibi or 99mTc tetrofosmin can be used interchangeably with similar diagnostic accuracy in CAD.
Myocardial perfusion imaging may use either planar or single-photon emission computed tomographic (SPECT) techniques and visual analyses (191–194) or quantitative techniques (195–202). Quantification (e.g., using horizontal [195] or circumferential [196–198] profiles) may improve the sensitivity of the test, especially in patients with one-vessel disease (194,198–202). For 201Tl planar scintigraphy, average reported values of sensitivity and specificity (not corrected for posttest referral bias) have been in the range of 83% and 88%, respectively, by visual analysis (191–194), and 90% and 80%, respectively, for quantitative analyses (194–203). The 201Tl SPECT is generally more sensitive than planar imaging for diagnosing CAD, localizing hypoperfused vascular territories, identifying left anterior descending and left circumflex coronary artery stenoses (204), and correctly predicting the presence of multivessel CAD (205). The average (uncorrected for referral bias) sensitivity and specificity of exercise 201Tl SPECT imaging are in the range of 89% and 76%, respectively, for qualitative analyses (201,202,206) and 90% and 70%, respectively, for quantitative analyses (206).
The less-than-perfect sensitivity and specificity may in part be explained by the fact that visually estimated angiographic severity of coronary stenoses does not closely correlate with functional severity as assessed by coronary flow reserve after maximal pharmacologic coronary vasodilation (203). Furthermore, the lower-than-expected specificity in the more recent series, which has generally involved SPECT rather than planar imaging, may well be related to posttest referral bias (see above). Although patient selection undoubtedly plays a role in decreasing the specificity observed with SPECT compared with planar imaging, other factors, such as photon attenuation and artifacts created by the tomographic reconstruction process, are also likely important.
Since the introduction of dipyridamole-induced coronary vasodilation as an adjunct to 201Tl myocardial perfusion imaging (207–209), pharmacologic interventions have become an important tool in noninvasive diagnosis of CAD (165,166,168–171,208–217). Dipyridamole planar scintigraphy has a high sensitivity (90% average, uncorrected) and acceptable specificity (70% average, uncorrected) for detection of CAD (165). Dipyridamole SPECT imaging with 201Tl or 99mTc sestamibi appears to be at least as accurate as planar imaging (218–220). Results of myocardial perfusion imaging during adenosine infusion are similar to those obtained with dipyridamole and exercise imaging (212–214,216). Dobutamine perfusion imaging has significant limitations compared with vasodilator (dipyridamole or adenosine) perfusion imaging because it does not provoke as great an increase in coronary flow (221,222). Its use should therefore be restricted to patients with contraindications to dipyridamole and adenosine, although dobutamine perfusion imaging has reasonable diagnostic accuracy (223). Because it should be used far less commonly than dipyridamole and adenosine, dobutamine perfusion imaging is not included in the recommendations.
Exercise and dobutamine radionuclide angiography are now performed very infrequently and are therefore also not included in the recommendations.
Stress echocardiography
Stress echocardiography relies on imaging LV segmental wall motion and thickening during stress compared with baseline. Echocardiographic findings suggestive of myocardial ischemia include 1) a decrease in wall motion in 1 LV segment with stress, 2) a decrease in wall thickening in 1 LV segment with stress, and 3) compensatory hyperkinesis in complementary (nonischemic) wall segments. The advent of digital acquisition and storage, as well as side-by-side (or quad screen) display of cineloops of LV images acquired at different levels of rest or stress, has facilitated efficiency and accuracy in interpretation of stress echocardiograms (13).
Stress echocardiography has been reported to have sensitivity and specificity for detecting CAD approximately in the range reported for stress myocardial imaging. In 36 studies reviewed that included 3,210 patients, the range of reported overall sensitivities, uncorrected for posttest referral bias, ranged from 70% to 97%; an average figure was approximately 85% for overall sensitivity for exercise echocardiography and 82% for dobutamine stress echocardiography (13). As expected, the reported sensitivity of exercise echocardiography for multivessel disease was higher (73% to 100%, average approximately 90%) than the sensitivity for one-vessel disease (63% to 93%, average approximately 79%) (13). In this series of studies, specificity ranged from 72% to 100%, with an average of approximately 86% for exercise echocardiography and 85% for dobutamine echocardiography.
Pharmacologic stress echocardiography is best accomplished with the use of dobutamine because it enhances myocardial contractile performance and wall motion, which can be evaluated directly by echocardiography. Dobutamine stress echocardiography has substantially higher sensitivity than vasodilator (dipyridamole or adenosine) stress echocardiography for detecting coronary stenoses (170,224,225). In a recent review of 36 studies, average sensitivity and specificity (uncorrected for referral bias) of dobutamine stress echocardiography in the detection of CAD were in the range of 82% (86% for multivessel disease) and 85%, respectively (13). Although dipyridamole echocardiography is performed abroad, it is used far less commonly in the U.S. and is not included in the recommendations.
|
Special issues related to stress cardiac imaging
|
|---|
Concomitant use of drugs.
The sensitivity of the exercise imaging study for diagnosis of CAD appears to be lower in patients taking beta-blockers (226–230). As recommended earlier for the exercise ECG (Section II.C.2), whenever possible, it is recommended that beta-blockers (and other anti-ischemic drugs) be withheld for four to five half-lives (usually about 48 h) before exercise imaging studies for the diagnosis and initial risk stratification of patients with suspected CAD. Nonetheless, in patients who exercise to a submaximal level because of the effect of drugs, perfusion or echocardiographic imaging still affords higher sensitivity than the exercise ECG alone (231).
Bundle-branch blocks.
Several studies have observed an increased prevalence of myocardial perfusion defects during exercise imaging, in the absence of angiographic coronary disease, in patients with left bundle-branch block (232–234). These defects often involve the interventricular septum, may be reversible or fixed and are often absent during pharmacologic stress. Their exact mechanism is uncertain. Multiple studies (involving >200 patients) have found that perfusion imaging with pharmacologic vasodilation is more accurate for identifying CAD in patients with left bundle-branch block (235–243). In contrast, only one small study of 24 patients has reported on the diagnostic usefulness of stress echocardiography in the presence of left bundle-branch block (244). The committee therefore believed that adenosine or dipyridamole myocardial perfusion imaging is preferred in these patients. Right bundle-branch block and left anterior hemiblock are not ordinarily associated with such perfusion defects.
Cardiac Stress Imaging in Selected Patient Subsets (Female, Elderly or Obese Patients, and Patients With Special Occupations).The treadmill ECG test is less accurate for diagnosis in women, who have a generally lower pretest likelihood of CAD than men (14). Myocardial perfusion imaging or echocardiography could be a logical addition to treadmill testing in this circumstance. However, the sensitivity of thallium perfusion scans may be lower in women than men (203,245). Artifacts due to breast attenuation, usually manifest in the anterior wall, can be an important caveat in the interpretation of women’s perfusion scans, especially when 201Tl is used as a tracer. More recently, the use of gated 99mTc sestamibi SPECT imaging has been associated with an apparent reduction in breast artifacts (246,247). In a recent prospective study of 115 women with either suspected or a low pretest likelihood of CAD, both 201Tl SPECT and 99mTc sestamibi had a similar sensitivity for detection of CAD in women (84.3% and 80.4% for 70% stenosis) (248). However, 99mTc sestamibi SPECT imaging had a better specificity (84.4% vs. 67.2%) and was further enhanced to 92.2% with ECG gating. Similarly, exercise or pharmacologic stress echocardiography may help avoid artifacts specifically due to breast attenuation. However, echocardiographic imaging in obese persons tends both to be technically more difficult and to produce images of less quality. As indicated earlier (Section II.C.2), the committee believes that there currently are insufficient data to justify replacing standard exercise testing with stress imaging in the initial evaluation of women.
Although some elderly patients can perform an adequate exercise test, many are unable to do so because of physical impairment. Pharmacologic stress imaging is an appropriate option in such patients. Very obese patients constitute a special problem because most imaging tables used for SPECT have weight-bearing limits (often 300 lb [135 kg]) that preclude imaging very heavy subjects. These subjects can still be imaged by planar scintigraphy. Obese patients often have suboptimal perfusion images, especially with 201Tl, owing to the marked photon attenuation by soft tissue. In these patients, 99mTc sestamibi is probably most appropriate and should yield images of better quality than 201Tl. Positron emission tomographic imaging is also likely to be superior to conventional myocardial perfusion imaging in these subjects. As noted above, exercise or pharmacologic stress echocardiography may yield suboptimal images in significantly obese subjects. Suboptimal images are also not uncommon in patients with chest deformities and significant lung disease.
Persons whose occupation may affect public safety (e.g., airline pilots, truckers, bus drivers, railroad engineers, firefighters and law enforcement officers) or who are professional or high-profile athletes not uncommonly undergo periodic exercise testing for assessment of exercise capacity and prognostic evaluation of possible CAD (14). Although there are insufficient data to justify this approach, these evaluations are done for statutory reasons in some cases (27). For patients in these groups with chronic chest pain who are in the intermediate-to-high likelihood range for CAD, the threshold for adding imaging to standard exercise electrocardiography may properly be lower than in the average patient. Specifically, one might recommend that for persons in this risk category, in whom stress testing is being contemplated, stress imaging (with echocardiography or radionuclide perfusion imaging) should be the initial stress test.
|
Comparison of myocardial perfusion imaging and echocardiography
|
|---|
Sensitivity and specificity.
In an analysis of 11 studies involving 808 patients who had contemporaneous treadmill (or pharmacologic) stress echocardiography and perfusion scintigraphy, the overall (uncorrected for referral bias) sensitivity was 83% for stress perfusion imaging versus 78% for stress echocardiography (p = NS). On the other hand, overall specificity (uncorrected for referral bias) tended to favor stress echocardiography (86% vs. 77%; p = NS) (249).
More recently, Fleischmann et al. (250) performed a meta-analysis on 44 articles (published between 1990 and 1997), which examined the diagnostic accuracy of exercise tomographic myocardial perfusion imaging or exercise echocardiography. The overall sensitivity and specificity, respectively, were 85% and 77% for exercise echocardiography, 87% and 64% for exercise myocardial perfusion imaging, and 52% and 71% for exercise ECG. These estimates were not adjusted for referral bias. On the basis of receiver operator characteristic curves, which were also not adjusted for referral bias, exercise echocardiography had significantly better discriminatory power than exercise myocardial perfusion imaging.
Localization of disease to individual coronary arteries.
Use of SPECT has provided diagnostic improvement over planar imaging for more precise localization of the vascular territories involved, particularly the identification of left circumflex coronary artery stenoses and prediction of multivessel CAD (201,202,206). O’Keefe et al. (141) reviewed data on 770 patients (1,328 diseased coronary arteries) in multiple studies who had exercise perfusion imaging versus 200 (704 diseased coronary arteries) who had undergone exercise echocardiography. In these data derived from 10 published studies, there was a nonsignificant trend toward improvement in localization of coronary disease by the radionuclide technique (79% vs. 65% uncorrected sensitivity; p = NS). For localization of disease to the circumflex coronary artery, however, the radionuclide method conferred a significant advantage in sensitivity (72% vs. 33%, uncorrected p < 0.001).
Importance of local expertise and facilities.
Echocardiographic and radionuclide stress imaging have complementary roles, and both add value to routine stress electrocardiography under the circumstances outlined above. The choice of which test to perform depends on issues of local expertise, available facilities and considerations of cost-effectiveness (see following text). Because of its lower cost and generally greater portability, stress echocardiography is more likely to be performed in the physician’s office than stress radionuclide imaging; the availability of stress imaging in the office setting has both advantages and disadvantages for the patient (176).
Cost-effectiveness considerations.
In this era of managed care, cost-effectiveness considerations have come into sharper focus in medical decision making. Commonly used measures of cost-effectiveness include the change in quality-adjusted life-years (QALY) per dollar of cost. The cost/ QALY ratio is importantly affected by the pretest likelihood of CAD, test accuracy, and the cost and complication rates of the test (176,251,252). Patterson and Eisner (251) used an assumption for detection of significant CAD of 75% sensitivity and 90% specificity for stress echocardiography and 84% sensitivity and 87% specificity for SPECT perfusion imaging. They found that the cost/ QALY ratio was 8% to 12% higher for stress echocardiography than for SPECT thallium imaging (251). However, Marwick (252) has argued that if Medicare reimbursement rates had been substituted for costs quoted by the authors and sensitivity/specificity data adjusted to 80% and 85%, respectively, for stress echocardiography, and 70% and 90%, respectively, for SPECT thallium imaging, that the cost/ QALY ratios would have decreased for both tests. Marwick also argued that the cost/ QALY ratio would have been slightly lower for stress echocardiography (compared with stress perfusion imaging) at coronary disease probability rates of 20% to 30% and slightly higher for stress echocardiography at probability rates of 40% to 80%.
A subsequent decision and cost-effectiveness analysis (253) used published data (uncorrected for referral bias) to compare exercise electrocardiography, exercise thallium perfusion imaging, exercise echocardiography and coronary angiography for the diagnosis of suspected CAD in a 55-year-old woman. Coronary angiography was most cost-effective in a woman of this age with definite angina, whereas exercise echocardiography was most cost-effective in the presence of atypical angina or nonanginal chest pain.
A summary of comparative advantages of stress myocardial perfusion imaging and stress echocardiography is provided in Table 18.
View this table:
[in this window]
[in a new window]
|
Table 18 Comparative Advantages of Stress Echocardiography and Stress Radionuclide Perfusion Imaging in Diagnosis of CAD
|
|
Recent technical developments.
The published comparisons between stress echocardiography and stress myocardial perfusion imaging do not fully reflect the ongoing developments in both techniques.
For stress echocardiography, recent developments include tissue harmonic imaging and intravenous contrast agents, which can improve detection of endocardial borders (254,255).
For stress myocardial perfusion imaging, newer-generation gamma cameras and scatter correction improve resolution (256), and gating permits assessment of global and regional function (257) as well as more accurate characterization of equivocal findings (247). Attenuation correction is under development (258).
These recent advances in both stress echocardiography and stress myocardial perfusion imaging should improve diagnostic accuracy.
|
D. invasive testing: value of coronary angiography
|
|---|
Recommendations for Coronary Angiography to Establish a Diagnosis in Patients With Suspected Angina, Including Those With Known CAD Who Have a Significant Change in Anginal Symptoms
Class I.
- Patients with known or possible angina pectoris who have survived sudden cardiac death. (Level of Evidence: B)
Class IIa.
- Patients with an uncertain diagnosis after noninvasive testing in whom the benefit of a more certain diagnosis outweighs the risk and cost of coronary angiography. (Level of Evidence: C)
- Patients who cannot undergo noninvasive testing due to disability, illness or morbid obesity. (Level of Evidence: C)
- Patients with an occupational requirement for a definitive diagnosis. (Level of Evidence: C)
- Patients who by virtue of young age at onset of symptoms, noninvasive imaging, or other clinical parameters are suspected of having a nonatherosclerotic cause for myocardial ischemia (coronary artery anomaly, Kawasaki disease, primary coronary artery dissection, radiation-induced vasculoplasty). (Level of Evidence: C)
- Patients in whom coronary artery spasm is suspected and provocative testing may be necessary. (Level of Evidence: C)
- Patients with a high pretest probability of left main or three-vessel CAD. (Level of Evidence: C)
Class IIb.
- Patients with recurrent hospitalization for chest pain in whom a definite diagnosis is judged necessary. (Level of Evidence: C)
- Patients with an overriding desire for a definitive diagnosis and a greater-than-low probability of CAD. (Level of Evidence: C)
Class III.
- Patients with significant comorbidity in whom the risk of coronary arteriography outweighs the benefit of the procedure. (Level of Evidence: C)
- Patients with an overriding personal desire for a definitive diagnosis and a low probability of CAD. (Level of Evidence: C)
This invasive technique for imaging the coronary artery lumen remains the most accurate for the diagnosis of clinically important obstructive coronary atherosclerosis and less common nonatherosclerotic causes of possible chronic stable angina pectoris such as coronary artery spasm, coronary anomaly, Kawasaki disease, primary coronary artery dissection and radiation-induced coronary vasculopathy (322–331). Early case studies correlating symptoms with the findings at coronary angiography reported that between 26% and 65% of patients with chest discomfort that was suggestive of but was not classical angina (i.e., atypical symptoms) had significant coronary stenoses due to atherosclerosis (38,332–334). In many patients with symptoms suggestive but not typical of chronic stable angina pectoris (i.e., pretest probability 50%), the incremental value of noninvasive testing, when considered with other clinical data, may permit a sufficiently accurate diagnosis on which to base clinical management strategies (12–14) (see Section II.C). Incumbent on the physician is the responsibility for estimating the probability that the patient’s symptoms are due to myocardial ischemia and matching the intensity of the evaluation to this estimation. All decisions regarding testing for possible CAD must be modulated by patient preference and comorbidity. It is important to reemphasize the value of a history of typical effort angina in middle-aged or elderly men in whom 90% have significant coronary disease (38,332–334) and many have multivessel disease (see Fig. 6). In women, only about one half with classic angina pectoris have significant obstructive coronary disease (see following section).
View larger version (42K):
[in this window]
[in a new window]
|
Figure 6 Coronary angiography findings in patients with chronic effort-induced angina pectoris. Top: Percentage of men with one-vessel, two-vessel, three-vessel, left main or no coronary artery disease on coronary angioraphy. Bottom: Percentage of women with one-vessel, two-vessel, three-vessel, left main, or no coronary artery disease on coronary angiography. N = normal or <50% stenosis; 1 = one-vessel disease; two = 2-vessel disease; three = 3-vessel disease; LM = left main disease. Data from Douglas and Hurst (333).
|
|
|
Indications for coronary angiography
|
|---|
Direct referral for diagnostic coronary angiography may be indicated in patients with chest pain possibly attributable to myocardial ischemia when noninvasive testing is contraindicated or unlikely to be adequate due to illness, disability or physical characteristics. For example, a patient with chest pain suggestive of chronic stable angina and coexisting chronic obstructive pulmonary disease who is not a candidate for exercise testing because of dyspnea, perfusion imaging with dipyridamole or adenosine because of bronchospasm and theophylline therapy or stress-echocardiography because of poor images may undergo coronary angiography with minimal risk.
Patients in whom noninvasive testing is abnormal but not clearly diagnostic may warrant clarification of an uncertain diagnosis by coronary angiography or in some cases by a second noninvasive test (imaging modality), which may be recommended for a low-likelihood patient with an intermediate-risk treadmill result (335). Coronary angiography may be most appropriate for a patient with a high-risk treadmill outcome.
In patients with symptoms suggestive but not characteristic of stable angina, direct referral to coronary angiography may be indicated when the patient’s occupation or activity could constitute a risk to themselves or others (pilots, firefighters, police, professional athletes or serious runners) (27). In certain patients with typical or atypical symptoms suggestive of stable angina and a high clinical probability of severe CAD, direct referral to coronary angiography may be indicated and prove cost-effective (335). The diagnosis of chronic stable angina in diabetic persons can be particularly difficult because of the paucity of symptomatic expressions of myocardial ischemia due to autonomic and sensory neuropathy, and a lowered threshold for coronary angiography is appropriate (336). The use of coronary angiography in patients with a high pretest probability of disease is in some patients as important in risk assessment (see Section III.A) as in diagnosis.
Women.
The evaluation of chest pain in women has been scrutinized recently, and available data suggest that gender differences in presentation and disease manifestations should be considered (337). Atypical chest pain is more common in women, perhaps in relation to an increased prevalence of vasospasm as well as mitral valve prolapse and noncoronary chest pain syndromes. ECG treadmill exercise testing has a higher false-positive rate in women (38% to 67%) than men (7% to 44%) (338), largely because of the lower pretest likelihood of disease (339), but a low false-negative rate, which indicates that routine testing reliably excludes the presence of coronary disease when the results of noninvasive tests are negative. Despite the limitations of routine exercise ECG testing in women, it has been shown to reduce procedures without loss of diagnostic accuracy. Only 30% of women (in whom a reasonably certain diagnosis of CAD could not be reached or excluded) need be referred for further testing (83). Recent studies examining the outcome of patients undergoing diagnostic testing indicate that women with positive stress ECGs or stress thallium examinations were less frequently referred for additional noninvasive testing (4% vs. 20% for men) or coronary angiography (34% vs. 45%) (340). Although these findings suggest that a gender-based difference in clinical practice exists in this country, 2 reports indicate that the reduced referral rate of women was clinically appropriate (341,342). As mentioned in Section II.C, a recent cost-effectiveness analysis concluded that coronary angiography was the preferred initial diagnostic test in a 55-year-old woman with typical angina (253).
The elderly.
The evaluation of chest pain syndromes in the elderly can be difficult because complaints of chest discomfort, weakness and dyspnea are common, and comorbid conditions that mimic angina pectoris are frequently present. Reduced activity levels and blunted appreciation of ischemic symptoms become the norm with advancing age (343). In large community studies of men and women 65 years old, those with atypical symptoms and typical angina were shown to have similar three-year cardiac mortality rates (344). An increased frequency of abnormal ECGs at rest and inability to exercise complicate noninvasive diagnostic testing, as does the increased prevalence of disease, which reduces the value of a negative noninvasive test. Diagnostic coronary angiography has very little increased risk (compared with younger patients) in older patients undergoing elective evaluation and is commonly used; in many centers, most patients who undergo this study are >65-years-old (345).
Coronary spasm.
Coronary artery spasm is a well-recognized cause of chest pain at rest (346) and may also lead to variable threshold effort angina (323,324), but in a 10-year study of 3,447 patients who underwent provocative testing with ergonovine maleate, coronary spasm was most often associated with an atypical chest pain syndrome (322) and cigarette smoking. The lack of a classic presentation and requirement for provocative testing during coronary angiography may hinder making this diagnosis. Although some investigators have advocated noninvasive, provocative testing for coronary spasm (347), there is some risk of irreversible coronary spasm (348); for this reason, most recommend that provocative testing for coronary spasm be done in the cardiac catheterization environment, where administration of intracoronary nitroglycerin and other vasodilators is feasible and other support systems are available (349).
Coronary anomaly.
The anomalous origin and course of coronary arteries is an uncommon cause of chronic stable angina usually recognized unexpectedly at coronary angiography, but this diagnosis may be suspected in younger patients with signs or symptoms of myocardial ischemia (325–327) and recognized by noninvasive imaging modalities such as transesophageal echocardiography, computerized tomography, or magnetic resonance imaging. The presence of a continuous murmur can point to a diagnosis of anomalous origin of the left anterior descending or circumflex artery from the pulmonary artery or coronary arteriovenous fistula that should be confirmed by coronary angiography.
Resuscitation from ventricular fibrillation or sustained ventricular tachycardia.
Most patients experiencing sudden cardiac arrest or malignant arrhythmia have severe CAD (350). Therefore, coronary arteriography is warranted to establish as precise a diagnosis as possible as well as to establish revascularization options (see Section IV).
|
III. Risk stratification
|
|---|
A. Clinical assessment.
Prognosis of CAD for death or nonfatal MI: general considerations
Coronary artery disease is a chronic disorder with a natural history that spans multiple decades. In each affected person, the disease typically cycles in and out of clinically defined phases: asymptomatic, stable angina, progressive angina and unstable angina or acute MI. Although the specific approach to risk stratification of the coronary disease patient can vary according to the phase of the disease in which the patient presents, some general concepts apply across the spectrum of disease.
The patient’s risk is usually a function of four types of patient characteristic. The strongest predictor of long-term survival with CAD is the functioning of the LV. Ejection fraction is the most commonly used measure of the extent of LV dysfunction. A second patient characteristic is the anatomic extent and severity of atherosclerotic involvement of the coronary tree. The number of diseased vessels is the most common measure of this characteristic. A third characteristic provides evidence of a recent coronary plaque rupture, indicating a substantially increased short-term risk for cardiac death or nonfatal MI. Worsening clinical symptoms with unstable features is the major clinical marker of a plaque event. The fourth patient characteristic is general health and noncoronary comorbidity.
The probability that a given patient will progress to a higher or lower risk disease state depends primarily on factors related to the aggressiveness of the underlying atherosclerotic process. Patients with major cardiac risk factors, including smoking, hypercholesterolemia, diabetes mellitus and hypertension, are most likely to have progressive atherosclerosis with repeated coronary plaque events. Patients presenting at a younger age also may have more aggressive disease.
A growing body of pathologic, angiographic, angioscopic and intravascular ultrasonographic data support a pathophysiologic model in which most major cardiac events are initiated by microscopic ulcerations of vulnerable atherosclerotic plaques. Several lines of evidence have shown that the majority of vulnerable plaques appear "angiographically insignificant" before their rupture (<75% diameter stenosis). In contrast, most of the "significant" plaques (>75% stenosis) visualized at angiography are at low risk for plaque rupture. Thus, the ability of stress testing of any type to detect vulnerable atherosclerotic lesions may be limited by the smaller size and lesser effect on coronary blood flow of these plaques and may explain the occasional acute coronary event that occurs shortly after a negative treadmill test result.
|
Risk stratification with clinical parameters
|
|---|
Rigorous evidence for predictors of severe CAD (three-vessel and left main disease) derived solely from the history and physical examination in patients with chest pain is surprisingly limited. Presumably, this is because physicians routinely incorporate additional information (e.g., an ECG) into risk stratification.
Nevertheless, very useful information relevant to prognosis can be obtained from the history. This includes demographics such as age and gender as well as a medical history focusing on hypertension, diabetes, hypercholesterolemia, smoking, peripheral vascular or arterial disease and previous MI. As previously discussed, the description of the patient’s chest discomfort can usually be easily assigned to one of three categories: typical angina, atypical angina and nonanginal chest pain (38).
The physical examination may also aid in risk stratification by determining the presence or absence of signs and symptoms that might alter the probability of severe CAD. Useful findings include those suggesting vascular disease (abnormal fundi, decreased peripheral pulses, bruits), long-standing hypertension (blood pressure, abnormal fundi), aortic valve stenosis or idiopathic hypertrophic subaortic stenosis (systolic murmur, abnormal carotid pulse, abnormal apical pulse), left-heart failure (third heart sound, displaced apical impulse, bibasilar rales), and right-heart failure (jugular venous distension, hepatomegaly, ascites, pedal edema).
Several studies have examined the value of clinical parameters for identifying the presence of severe (three-vessel or left main) CAD. Pryor et al. (134) identified 11 clinical characteristics that are important in estimating the likelihood of severe CAD: typical angina, previous MI, age, gender, duration of chest pain symptoms, risk factors (hypertension, diabetes, hyperlipidemia, smoking), carotid bruit and chest pain frequency. In a subsequent study, Pryor et al. (41) provided detailed equations for the prediction of both severe CAD and survival based on clinical parameters.
Hubbard et al. (351) identified five clinical parameters that were independently predictive of severe (three-vessel or left main) CAD: age, typical angina, diabetes, gender and prior MI (history or ECG). Hubbard then developed a five-point cardiac risk score. A composite graph (Fig. 7) estimates the probability of severe CAD. Each curve shows the probability of severe CAD as a function of age for a given cardiac risk score. As shown on this graph, some patients have a high likelihood (>1 chance in 2) of severe disease on the basis of clinical parameters alone. Such patients should be considered for direct referral to angiography, as noninvasive testing is highly unlikely to be normal and, if it is, may conceivably be false-negative. An example would be a 50-year-old male patient with diabetes, taking insulin, with typical angina and history and ECG evidence of previous MI. His estimated likelihood of severe disease is 60%; such a patient should be considered for angiography without further testing.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 7 Nomogram showing the probability of severe (three-vessel or left main) coronary disease based on a five-point score. One point is awarded for each of the following variables: male gender, typical angina, history and electrocardiographic evidence of myocardial infarction, diabetes and use of insulin. Each curve shows the probability of severe coronary disease as a function of age. From Hubbard et al. (135), with permission.
|
|
Descriptive information about the chest pain is very important in assessment of patient prognosis as well as risk of severe CAD. However, because the extent and location of angiographically demonstrated occlusion, together with the degree of LV dysfunction, appear to have substantially greater prognostic power than symptom severity (96,352), many clinicians have come to rely almost exclusively on these "objective" measurements of disease and very little on the patient’s history in choosing among the alternative management strategies for their patients. However, clinical parameters should not be ignored for risk stratification (41,353,354).
Califf et al. (95) have provided evidence that the aggregation of certain historical and ECG variables in an "angina score" offers prognostic information that is independent of and incremental to that detected by catheterization. The angina score was composed of three differentially weighted variables: the "anginal course," anginal frequency, and rest ECG ST-T wave abnormalities. Two features of the prognostic power of the angina score seem intuitively correct: 1) it had a greater impact on short-term prognosis than long-term prognosis, presumably reflecting the importance of a plaque rupture; and 2) it had greater prognostic value when the LV was normal than when it was abnormal, presumably because so much of the overall prognosis was determined by LV function when it was abnormal.
Peripheral vascular disease is another clinical parameter that is useful in stratifying risk. The presence of a carotid bruit, like male gender and previous MI, nearly doubles the risk for severe CAD (134). In addition to peripheral vascular disease, signs and symptoms related to congestive heart failure (CHF), which reflect LV function, convey an adverse prognosis.
All the studies evaluating clinical characteristics as predictors of severe CAD use only patients referred for further evaluation of chest pain and cardiac catheterization. Although it does not undercut internal validity, this bias in the assembly of a cohort severely limits the generalizability (external validity) of study findings to all patients with CAD. However, it is likely that the overall "risk" of an unselected population is lower, so that patients described as "low risk" by these findings are still likely to be low risk.
Risk stratification of patients with stable angina using clinical parameters may facilitate the development of clearer indications of referral for exercise testing and cardiac catheterization. Long-term follow-up data from the CASS Registry (352) showed that 72% of the deaths occurred in the 38% of the population that had either LV dysfunction or severe coronary disease. The prognosis of patients with a normal ECG (which implies normal LV function at rest) and a low clinical risk for severe CAD is therefore excellent. Pryor et al. (41) showed that 37% of outpatients referred for noninvasive testing met the criteria for low risk. Fewer than 1% of these patients had left main artery disease or died within three years. The value of additional testing for risk stratification in such patients is modest. Lower-cost options such as treadmill testing should therefore be used whenever possible, and only the most abnormal results (described in Section III.2) should be referred to angiography.
B. ECG/chest X-ray.
Patients with chronic stable angina who have rest ECG abnormalities are at greater risk than those with normal ECGs (355). Evidence of 1 prior MI on ECG indicates an increased risk for cardiac events. In fact, the presence of Q waves in multiple ECG leads, often accompanied by an R wave in lead V1 (posterior infarction), is frequently associated with a markedly reduced LV ejection fraction, an important determinant of the natural history of patients with suspected atherosclerotic CHD (356). A "QRS score" has been used to indicate the extent of old or new MI (357), with the higher scores being associated with lower LV ejection fractions and a poorer long-term prognosis. The presence of persistent ST-T wave inversions, particularly in leads V1 to V3 of the rest ECG, is associated with an increased likelihood of future acute coronary events and a poor prognosis (358–361). A decreased prognosis for patients with angina pectoris is also likely when the ECG shows left bundle-branch block, bifascicular block (often left anterior fascicular block plus right bundle-branch block), second- or third-degree atrioventricular block, atrial fibrillation or ventricular tachyarrhythmias (362). The presence of LV hypertrophy by ECG criteria in a patient with angina pectoris is also associated with increased morbidity and mortality (361,363).
On the chest roentgenogram, the presence of cardiomegaly, an LV aneurysm or pulmonary venous congestion is associated with a poorer long-term prognosis than that which occurs in patients with a normal chest X-ray result. The presence of left atrial enlargement, indicating a higher likelihood of pulmonary venous congestion or mitral regurgitation, is also a negative prognostic factor.
As indicated previously, the presence of calcium in the coronary arteries on chest X-ray or fluoroscopy in patients with symptomatic CAD suggests an increased risk of cardiac events (364). The presence and amount of coronary artery calcification by EBCT also correlates to some extent with the severity of CAD, but there is considerable patient variation.
C. Noninvasive testing.
- Resting LV Function (Echocardiographic/Radionuclide Imaging)
Recommendations for Measurement of Rest LV Function by Echocardiography or Radionuclide Angiography in Patients With Chronic Stable Angina
Class I
- Echocardiography or radionuclide angiography (RNA) in patients with a history of prior MI, pathologic Q waves or symptoms or signs suggestive of heart failure to assess LV function. (Level of Evidence: B)
- Echocardiography in patients with a systolic murmur suggesting mitral regurgitation to assess its severity and etiology. (Level of Evidence: C)
- Echocardiography or RNA in patients with complex ventricular arrhythmias to assess LV function. (Level of Evidence: B)
Class III
- Routine periodic reassessment of stable patients for whom no new change in therapy is contemplated. (Level of Evidence: C)
- Patients with a normal ECG, no history of MI and no symptoms or signs suggestive of CHF. (Level of Evidence: B)
Importance of assessing LV function.
Most patients undergoing a diagnostic evaluation for angina do not need an echocardiogram. However, in the chronic stable angina patient who has a history of documented MI or Q waves on ECG, measurement of global LV systolic function (e.g., ejection fraction) may be important in choosing appropriate medical or surgical therapy and making recommendations about activity level, rehabilitation and work status (13,365). Similarly, in patients who, in addition to chronic stable angina, have clinical signs or symptoms of heart failure, cardiac imaging may be helpful in establishing pathophysiologic mechanisms and guiding therapy. For example, a patient with heart failure might have predominantly systolic LV dysfunction, predominantly diastolic dysfunction, mitral or aortic valve disease, some combination of these abnormalities or a noncardiac cause for symptoms. The best treatment of the patient can be planned more rationally knowing the status of LV systolic and diastolic function (by echocardiography or radionuclide imaging), valvular function, and pulmonary artery pressure (by transthoracic echo-Doppler techniques).
Assessment of global LV function.
Left ventricular global systolic function and volumes have been well documented to be important predictors of prognosis in patients with cardiac disease. In patients with chronic ischemic heart disease, LV ejection fraction measured at rest by either echocardiography (352) or RNA (96,352,365) is predictive of long-term prognosis; as LV ejection fraction declines, mortality increases (352). A rest ejection fraction of <35% is associated with an annual mortality rate >3% per year.
Current echocardiographic techniques permit a comprehensive assessment of LV size and function (366,367). Two-dimensional echocardiographic LV ejection fraction may be measured quantitatively or reported qualitatively (by visual estimation) as increased; normal; or mildly, moderately, or severely reduced. When performed by skilled observers, visual estimation has been reported to yield ejection fractions that correspond closely to those obtained by angiography (368) or gated blood pool scanning (369). In addition to measures of LV systolic function, echo-Doppler characteristics of the pulsed Doppler transmitral velocity pattern can help assess diastolic function (370), although its independent prognostic value has not been established.
Left ventricular mass and wall thickness-to-chamber radius ratio, as measured from echocardiographic images, have both been shown to be independent of cardiovascular morbidity and mortality (371–373). The LV mass can be measured from two-dimensional or two-dimensionally directed M-mode echocardiographic images.
Radionuclide ejection fraction may be measured at rest using a gamma camera, a 99mTc tracer, and first-pass or gated equilibrium blood pool angiography (13) or gated SPECT perfusion imaging (257). Diastolic function can also be assessed by radionuclide ventriculography (374,375). It should be noted that LV ejection fraction and other indexes of myocardial contractile performance are limited by their dependence on loading conditions and heart rate (146,376).
Although magnetic resonance imaging is less widely disseminated, it may also be used to assess LV performance, including ejection fraction (377).
LV segmental wall motion abnormalities.
In patients with chronic stable angina and a history of previous MI, segmental wall motion abnormalities can be seen not only in the zone(s) of prior infarction but also in areas with ischemic "stunning" or "hibernation" of myocardium that is nonfunctional but still viable (143,148,151,378–380). The sum of these segmental abnormalities reflects total ventricular functional impairment, which may overestimate true anatomic infarct size or radionuclide perfusion defect (380). Thus, echocardiographically derived infarct size (143) correlates only modestly with 201Tl perfusion defects (151), peak creatine kinase levels (148,381), hemodynamic changes (143) and pathologic findings (379). However, it does predict the development of early (382) and late (383) complications and mortality (143,384).
As mentioned previously (Sections II.C.3 and II.C.4), recent developments in both echocardiography (tissue harmonic imaging and intravenous contrast agents to assess the endocardium) and myocardial perfusion imaging (gated SPECT imaging to assess global and regional function) should improve the ability of both techniques to assess LV function.
Ischemic mitral regurgitation, LV aneurysm, and LV thrombosis.
In patients with chronic ischemic heart disease, mitral regurgitation may result from global LV systolic dysfunction (161), regional papillary muscle dysfunction (162), scarring and shortening of the submitral chords (163), papillary muscle rupture (164), or other causes. The presence, severity and mechanism of mitral regurgitation can be reliably detected by transthoracic imaging and Doppler echocardiographic techniques (13). Potential surgical approaches also can be defined. In addition, chronic stable angina patients who have ischemic mitral regurgitation have a worse prognosis than those without regurgitation.
In patients with chronic angina and concomitant heart failure or significant ventricular arrhythmias, the presence or absence of ventricular aneurysm can generally be established by transthoracic echocardiography (385,386). When an aneurysm is demonstrated, the function of the nonaneurysmal portion of the left ventricle is an important consideration in choosing medical or surgical therapy (387).
Echocardiography is the definitive test for detecting intracardiac thrombi (388–394). The LV thrombi are most common in stable angina pectoris patients who have significant LV wall motion abnormalities.
In patients with anterior and apical infarctions (388,392–394), the presence of LV thrombi denotes an increased risk of both embolism (389) and death (391). In addition, the structural appearance of a thrombus, which can be defined by transthoracic (or transesophageal) echocardiography, has some prognostic significance. Sessile, laminar thrombi represent less of a potential embolic risk than do pedunculated and mobile thrombi (13).
|
2. Exercise testing for risk stratification and prognosis
|
|---|
Recommendations for Risk Assessment and Prognosis in Patients With an Intermediate or High Probability of CAD
Class I.
- Patients undergoing initial evaluation. (Exceptions are listed below in classes IIb and III.) (Level of Evidence: B)
- Patients after a significant change in cardiac symptoms. (Level of Evidence: C)
Class IIb.
- Patients with the following ECG abnormalities:
- Preexcitation (Wolff-Parkinson-White) syndrome. (Level of Evidence: B)
- Electronically paced ventricular rhythm. (Level of Evidence: B)
- More than 1 mm of ST depression at rest. (Level of Evidence: B)
- Complete left bundle-branch block. (Level of Evidence: B)
- Patients who have undergone cardiac catheterization to identify ischemia in the distribution of a coronary lesion of borderline severity. (Level of Evidence: C)
- Postrevascularization patients who have a significant change in anginal pattern suggestive of ischemia. (Level of Evidence: C)
Class III.
- Patients with severe comorbidity likely to limit life expectancy or prevent revascularization. (Level of Evidence: C)
Risk stratification for death or MI: general considerations.
Risk stratification with the exercise test does not take place in isolation but as part of a process that includes other data from the clinical examination and other laboratory tests. Thus, the value of exercise testing for risk stratification must be considered in light of what is added to what is already known about the patient’s risk status. Most research on exercise testing has concentrated on its relationship with future survival and, to a lesser extent, freedom from MI. The summary presented here is based on the "ACC/AHA Guidelines for Exercise Testing" (14).
Risk stratification with the exercise test.
The risk of exercise testing in appropriately selected candidates is extremely low, and thus the main argument for not performing an exercise test is that the extra information provided would not be worth the extra cost of obtaining that information or the test might provide misinformation that could lead to inappropriate testing or therapy.
Unless cardiac catheterization is indicated, symptomatic patients with suspected or known CAD should usually undergo exercise testing to assess the risk of future cardiac events unless they have confounding features on the rest ECG. Furthermore, documentation of exercise-induced ischemia is desirable for most patients who are being evaluated for revascularization (72,395).
The choice of initial stress test should be based on the patient’s rest ECG, physical ability to perform exercise, local expertise and available technologies. Patients with a normal rest ECG constitute a large and important subgroup. Most patients who present with angina for the first time have a normal rest ECG (49). Such patients are very likely (92% to 96%) to have normal LV function (141,142,396) and therefore an excellent prognosis (49). The exercise ECG has a higher specificity in the absence of rest ST-T changes, LV hypertrophy and digoxin.
Several studies have examined the incremental value of exercise imaging procedures compared with the exercise ECG in patients with a normal rest ECG who are not taking digoxin (Table 19). In analyses (397,398) that included clinical and exercise ECG parameters for the prediction of left main or three-vessel disease, the modest benefit of imaging does not appear to justify its cost, which has been estimated at $20,550 per additional patient correctly classified (397). For the prediction of subsequent cardiac events, four separate analyses have failed to demonstrate incremental value. Mattera et al. (399) did find some incremental value, but only for the prediction of hard and soft events (including unstable angina) and only if the exercise ECG was abnormal. They still favored a stepwise strategy that used the exercise ECG as the initial test, like that proposed by others (83,400).
View this table:
[in this window]
[in a new window]
|
Table 19 Studies Examining the Incremental Value of Exercise Imaging Studies for the Prediction of Severe CAD and Subsequent Cardiac Events in Patients With a Normal Resting ECG*
|
|
For these reasons, the committee favored a stepwise strategy in which the exercise ECG, and not stress imaging procedures, is performed as the initial test in patients who are not taking digoxin, have a normal rest ECG, and are able to exercise. In contrast, a stress-imaging technique should be used for patients with widespread rest ST depression (>1 mm), complete left bundle-branch block, ventricular paced rhythm or preexcitation. Although exercise capacity can be assessed in such patients, exercise-induced ischemia cannot. Patients unable to exercise due to physical limitations such as reduced exercise capacity, arthritis, amputations, severe peripheral vascular disease or severe chronic obstructive pulmonary disease should undergo pharmacologic stress testing in combination with imaging.
The primary evidence that exercise testing can be used to estimate prognosis and assist in management decisions consists of seven observational studies (354,355,401–405). One of the strongest and most consistent prognostic markers is maximum exercise capacity. This measure is at least partly influenced by the extent of rest LV dysfunction and the amount of further LV dysfunction induced by exercise. However, the relationship between exercise capacity and LV function is complex, because exercise capacity is also affected by age, general physical conditioning, comorbidities and psychological state, especially depression (406). Exercise capacity is measured by maximum exercise duration, maximum MET level achieved, maximum workload achieved, maximum heart rate and double product. The specific variable used to measure exercise capacity is less important than including exercise capacity in the assessment. The translation of exercise duration or workload into METs provides a standard measure of performance regardless of the type of exercise test or protocol used.
A second group of prognostic markers is related to exercise-induced ischemia. The ST-segment depression and elevation (in leads without pathological Q waves and not in aVR) best summarize the prognostic information related to ischemia (401). Other variables are less powerful, including angina, the number of leads with ST-segment depression, the configuration of the ST depression (downsloping, horizontal or upsloping), and the duration of ST deviation into the recovery phase.
The Duke treadmill score combines this information and provides a way to calculate risk (37,401). The Duke treadmill score equals the exercise time in minutes minus (5x the ST-segment deviation, during or after exercise, in millimeters) minus (4x the angina index, which has a value of "0" if there is no angina, "1" if angina occurs, and "2" if angina is the reason for stopping the test). Among outpatients with suspected CAD, the two thirds of patients with scores indicating low risk had a four-year survival rate of 99% (average annual mortality rate 0.25%), and the 4% who had scores indicating high risk had a four-year survival rate of 79% (average annual mortality rate 5%) (see Table 20). The score works well for both inpatients and outpatients, and preliminary data suggest that the score works equally well for men and women (37,409,410). Only a small number of elderly patients have been studied, however. Comparable scores have been developed by others (402).
Because of its simplicity, lower cost and widespread familiarity with its performance and interpretation, the standard exercise test is the most reasonable one to select for men with a normal rest ECG who are able to exercise. The optimal testing strategy remains less well defined in women. Until adequate data are available to resolve this issue, it is reasonable to use exercise testing for risk stratification in women.
Use of exercise test results in patient management.
The results of exercise testing may be used to titrate medical therapy to the desired level of effectiveness. For example, a normal heart rate response to exercise suggests that the dose of beta-blocker should be increased. Testing for this purpose should generally be performed with the patient on medication. The other major management step addressed by the exercise test is whether to proceed with additional testing, which might lead to revascularization.
Proceeding with additional testing usually involves imaging. Although both stress echocardiography and stress SPECT perfusion imaging have been used after exercise testing, only SPECT perfusion imaging has been studied in patients divided into risk groups based on the Duke treadmill score (410). In patients with an intermediate-risk treadmill score, imaging appears to be useful for further risk stratification. In patients with a high-risk treadmill score, imaging may identify enough low-risk patients who can avoid cardiac catheterization to justify the cost of routine imaging, but further study is required. Few patients (<5%) who have a low-risk treadmill score will be identified as high risk after imaging, and thus the cost of identifying these patients argues against routine imaging (410).
Patients with a predicted average annual cardiac mortality rate of 1% per year (low-risk score) can be managed medically without the need for cardiac catheterization. Patients with a predicted average annual cardiac mortality rate 3% per year (high-risk score) should be referred for cardiac catheterization. Patients with a predicted average annual cardiac mortality rate of 1% to 3% per year (intermediate-risk score) should have either cardiac catheterization or an exercise imaging study. Those with known LV dysfunction should have cardiac catheterization.
Recommendation for exercise testing in patients with chest pain 6 months after revascularization.
Class IIb
Patients with a significant change in anginal pattern suggestive of ischemia. (Level of Evidence: B)
|
Rationale
|
|---|
There are two postrevascularization phases. In the early phase, the goal of exercise testing is to determine the immediate result of revascularization. In the late phase, which begins six months after revascularization and is the focus of this discussion, the goal is to assist in the evaluation and management of patients with chronic established CAD. Exercise testing also may be helpful in guiding a cardiac rehabilitation program and return-to-work decisions.
Exercise testing after CABG.
Exercise testing distinguishes cardiac from noncardiac causes of chest pain, which is often atypical after surgery. After CABG, the exercise ECG has a number of limitations. Rest ECG abnormalities are frequent, and if an imaging test is not incorporated into the study, more attention must be paid to symptom status, hemodynamic response, and exercise capacity. Because of these considerations and the need to document the site of ischemia, stress imaging tests are preferred for evaluating patients in this group.
Exercise testing after PTCA.
Similar considerations apply to angioplasty patients. Restenosis is more frequent, however. Although most restenosis occurs <6 months after angioplasty, when these recommendations do not apply, restenosis does occur later. The exercise ECG is an insensitive predictor of restenosis, with sensitivities ranging from 40% to 55%, which are significantly less than those with SPECT (12,411) or exercise echocardiography (13,412). Because of these considerations and the need to document the site of ischemia, stress imaging tests are preferred for evaluating symptomatic patients in this group.
Some authorities advocate routine testing for all patients in the late phase after PTCA with either exercise ECGs or stress imaging, as restenosis commonly induces silent ischemia. The rationale for this approach is that ischemia, whether painful or silent, worsens prognosis (413,414). This approach seems particularly attractive for high-risk patients, for example, those with decreased LV function, multivessel CAD, proximal left anterior descending artery disease, previous sudden death, diabetes mellitus, hazardous occupations and suboptimal PTCA results. If routine testing is done, there are insufficient data to justify a particular frequency of testing after angioplasty. The alternative approach, which the committee labeled class IIb because the prognostic benefit of controlling silent ischemia needs to be proved, is to selectively evaluate only patients with a significant change in anginal pattern.
|
3. Stress imaging studies (radionuclide and echocardiography)
|
|---|
Recommendations for Cardiac Stress Imaging as the Initial Test for Risk Stratification of Patients With Chronic Stable Angina Who Are Able to Exercise
Class I.
- Exercise myocardial perfusion imaging or exercise echocardiography to identify the extent, severity and location of ischemia in patients who do not have left bundle-branch block or an electronically paced ventricular rhythm and have either an abnormal rest ECG or are using digoxin. (Level of Evidence: B)
- Dipyridamole or adenosine myocardial perfusion imaging in patients with left bundle-branch block or electronically paced ventricular rhythm. (Level of Evidence: B)
- Exercise myocardial perfusion imaging or exercise echocardiography to assess the functional significance of coronary lesions (if not already known) in planning PTCA. (Level of Evidence: B)
Class IIb.
- Exercise or dobutamine echocardiography in patients with left bundle-branch block. (Level of Evidence: C)
- Exercise, dipyridamole, adenosine myocardial perfusion imaging or exercise or dobutamine echocardiography as the initial test in patients who have a normal rest ECG and who are not taking digoxin. (Level of Evidence: B)
Class III.
- Exercise myocardial perfusion imaging in patients with left bundle-branch block. (Level of Evidence: C)
- Exercise, dipyridamole, adenosine myocardial perfusion imaging or exercise or dobutamine echocardiography in patients with severe comorbidity likely to limit life expectation or prevent revascularization. (Level of Evidence: C)
Recommendations for Cardiac Stress Imaging as the Initial Test for Risk Stratification of Patients With Chronic Stable Angina Who Are Unable to Exercise
Class I.
- Dipyridamole or adenosine myocardial perfusion imaging or dobutamine echocardiography to identify the extent, severity and location of ischemia in patients who do not have left bundle-branch block or electronically paced ventricular rhythm. (Level of Evidence: B)
- Dipyridamole or adenosine myocardial perfusion imaging in patients with left bundle-branch block or electronically paced ventricular rhythm. (Level of Evidence: B)
- Dipyridamole or adenosine myocardial perfusion imaging or dobutamine echocardiography to assess the functional significance of coronary lesions (if not already known) in planning PTCA. (Level of Evidence: B)
Class IIb.
Dobutamine echocardiography in patients with left bundle-branch block. (Level of Evidence: C)
Class III.
Dipyridamole or adenosine myocardial perfusion imaging or dobutamine echocardiography in patients with severe comorbidity likely to limit life expectation or prevent revascularization. (Level of Evidence: C)
Available stress imaging approaches.
Stress imaging studies in which radionuclide myocardial perfusion imaging techniques or two-dimensional echocardiography at rest and during stress are useful for risk stratification and determination of the most beneficial management strategy for patients with chronic stable angina (415–417). Whenever possible, treadmill or bicycle exercise should be used as the most appropriate form of stress because it provides the most information concerning patient symptoms, cardiovascular function and hemodynamic response during usual forms of activity (14). In fact, the inability to perform a bicycle or exercise treadmill test is in itself a negative prognostic factor for patients with chronic CAD.
In patients who cannot perform an adequate amount of bicycle or treadmill exercise, various types of pharmacologic stress are useful for risk stratification (12,13,217,418). The selection of the type of pharmacologic stress will depend on specific patient factors such as the patient’s heart rate and blood pressure, the presence or absence of bronchospastic disease, the presence of left bundle-branch block or a pacemaker and the likelihood of ventricular arrhythmias.
Pharmacologic agents are often used to increase cardiac workload as a substitute for treadmill or bicycle exercise or to cause an increase in overall coronary blood flow (224,225). For the former effect, adrenergic-stimulating drugs (such as dobutamine or arbutamine) are usually used, and for the latter effect, vasodilating agents (such as dipyridamole or adenosine) are generally used (12,13,217,224,225,418) (see Section II.C.4).
Radionuclide imaging has played a major role in risk stratification of patients with CAD. Either planar (three conventional views) or SPECT (multiple tomographic slices in three planes) imaging with 201Tl or 99mTc perfusion tracers with images obtained at stress and during rest provide important information about the severity of functionally significant CAD (180–188,191,192,199,204, 205,419).
More recently, stress echocardiography has been used for assessing patients with chronic stable angina; thus, the amount of prognostic data obtained with this approach is somewhat limited. Nevertheless, the presence or absence of inducible myocardial wall motion abnormalities has useful predictive value in patients undergoing exercise or pharmacologic stress echocardiography. A negative stress echocardiography study denotes a low cardiovascular event rate during follow-up (420–428).
|
Important findings on stress perfusion studies for risk stratification
|
|---|
Normal poststress thallium scan results are highly predictive of a benign prognosis even in patients with known coronary disease (12). A collation of 16 studies involving 3,594 patients followed up for a mean of 29 months indicated a rate of cardiac death and MI of 0.9% per year (429), nearly as low as that of the general population (430). In a recent prospective study of 5,183 consecutive patients who underwent myocardial perfusion studies during stress and later at rest, patients with normal scans were at low risk (<0.5% per year) for cardiac death and MI during 642 ± 226 days of mean follow-up, and rates of both outcomes increased significantly with worsening scan abnormalities (431). The presence of a normal stress myocardial perfusion scan indicates such a low likelihood of significant CAD that coronary arteriography is usually not indicated as a subsequent test. Although the published data are limited, the single exception would appear to be patients with high-risk treadmill scores and normal images (431).
The number, extent, and site of abnormalities on stress myocardial perfusion scintigrams reflect the location and severity of functionally significant coronary artery stenoses. Lung uptake of 201Tl on postexercise or pharmacologic stress images is an indicator of stress-induced global LV dysfunction and is associated with pulmonary venous hypertension in the presence of multivessel CAD (432–435). Transient poststress ischemic LV dilation also correlates with severe two- or three-vessel CAD (436–439). Several studies have suggested that SPECT may be more accurate than planar imaging for determining the size of defects, detecting coronary and particularly left circumflex CAD and localizing abnormalities in the distribution of individual coronary arteries (180,204,419). However, more false-positive results are likely to result from photon attenuation during SPECT imaging (12).
The number, size, and location of perfusion abnormalities, the amount of lung uptake of 201Tl on poststress images, and the presence or absence of poststress ischemic LV dilation can be combined to maximize the recognition of high-risk patients, including those with multivessel disease, left main CAD and disease of the proximal portion of the left anterior descending coronary artery (LAD). Incremental prognostic information from the results of stress myocardial perfusion imaging can determine the likelihood of subsequent important cardiac events. The number of transient perfusion defects, whether provoked by exercise or pharmacologic stress, is a reliable predictor of subsequent cardiac death or nonfatal MI (180,419,440–447). The number of stenotic coronary arteries may be less predictive than the number of reversible perfusion defects (440–450). The magnitude of the perfusion abnormality was the single most prognostic indicator in a study that demonstrated independent and incremental prognostic information from SPECT 201Tl scintigraphy compared with that obtained from clinical, exercise treadmill and catheterization data (451). As indicated previously, increased lung uptake of thallium induced by exercise or pharmacologic stress is associated with a high risk for cardiac events (12,452).
Information concerning both myocardial perfusion and ventricular function at rest may be helpful in determining the extent and severity of coronary disease (181,183,453). This combined information can be obtained by performing two separate exercise tests (e.g., stress perfusion scintigraphy and stress RNA) or combining the studies after one exercise test (e.g., first-pass RNA with 99mTc-based agents followed by perfusion imaging or perfusion imaging using gating). However, an additional benefit of the greater information provided by combined myocardial perfusion and ventricular function exercise testing has not been shown in clinical outcome or prognostic studies (12). Thus, one determination of LV function at rest and one measure of exercise/pharmacologic stress-induced myocardial perfusion or exercise ventricular function, but not both, are appropriate (12). The prognostic value of stress myocardial perfusion imaging in chronic stable angina is summarized in Table 21 (studies with >100 patients, who did not have recent MI, and that included both positive and negative perfusion images).
Application of myocardial perfusion imaging to specific patient subsets.
Patients with a normal rest ECG
Myocardial perfusion imaging has little advantage over the less expensive treadmill exercise test in this subset of patients. Three separate studies (402,404,405) have demonstrated little (if any) incremental value of myocardial perfusion imaging in the initial evaluation of such patients. As mentioned previously (Section III.2), many such patients will have low-risk treadmill scores and will not require further evaluation.
Concomitant use of drugs
As mentioned previously (Sections II.2 and II.4), beta-blockers (and other anti-ischemic drugs) should be withheld for four to five half-lives before testing. However, even if these drugs are continued, most high-risk patients will usually still be identified (14). Nitrates may also decrease the extent of perfusion defects or even convert abnormal exercise scan results to normal results (462).
Women, the elderly, or obese patients
The treadmill ECG test is less accurate for the diagnosis of CHD in women who have a lower pretest likelihood than men (194). However, the sensitivity of thallium perfusion scans may be lower in women than men (194,245). Artifacts due to breast attenuation, usually manifest in the anterior wall, can be an important consideration in the interpretation of women’s scans, especially when 201Tl is used as a tracer (12). As mentioned previously, 99mTc sestamibi may be preferable to 201Tl scintigraphy for determining prognosis as well as diagnosing CAD in women with large breasts or breast implants (248).
Although many elderly patients can perform an exercise test, some are unable to do so because of physical impairment. Pharmacologic stress imaging is an appropriate option for risk stratification in such patients. Very obese patients constitute a specific problem because most imaging tables used for SPECT have weight-bearing limits (usually 300 to 450 lb) that preclude imaging very heavy subjects. These subjects can still be imaged by planar scintigraphy (12). Obese patients often have suboptimal perfusion images, especially with thallium-201 because of the marked photon attenuation by soft tissue. In these patients, technetium sestamibi is probably the most appropriate and should provide images of better quality than 201Tl.
Left Bundle-Branch Block. As mentioned previously (Section II.4), pharmacologic stress perfusion imaging is preferable to exercise perfusion imaging in patients with left bundle-branch block. Recently, 245 patients with left bundle-branch block underwent SPECT imaging with 201Tl (n = 173) or 99mTc sestamibi (n = 72) during dipyridamole (n = 153) or adenosine (n = 92) stress testing (463). Patients with a large, severe fixed defect, a large reversible defect or cardiac enlargement and either increased pulmonary uptake (thallium) or decreased ejection fraction (sestamibi) (n = 20) were classified as high-risk patients. The rest were classified as low risk. The three-year overall survival rate was 57% in the high-risk group compared with 87% in the low-risk group (p = 0.001). Patients with a low-risk scan had an overall survival rate that was not significantly different from that of the U.S.-matched population (p = 0.86). The value of pharmacologic perfusion imaging for prognostication was confirmed in three other studies (464–466), which included >300 patients followed for a mean of nearly three years. Normal dipyridamole or adenosine scans were associated with a low cardiac event rate; large defects and increased pulmonary uptake were associated with a high cardiac event rate.
After Coronary Angiography. Myocardial perfusion imaging is useful in planning revascularization procedures because it demonstrates whether a specific coronary stenosis is associated with the stress-induced perfusion abnormality (12). Myocardial perfusion imaging is particularly helpful in determining the functional importance of single or multiple stenoses when PTCA is targeted to the "culprit lesion," that is, the ischemia-provoking stenosis (12,463,467–469).
After Myocardial Revascularization. Myocardial perfusion imaging can be useful in several situations after coronary bypass surgery. In patients with ST-T wave abnormalities at rest, recurrent myocardial ischemia during stress can be better evaluated by exercise scintigraphy than ECG treadmill testing. In addition, approximately 30% have an abnormal ECG response on treadmill exercise testing early after bypass surgery (470); these patients can be assessed for potential and incomplete revascularization and the extent of myocardium affected. Patients with initial negative postoperative treadmill test results that later become positive usually have progressive ischemia due to either graft closure or progression of disease in the native circulation (471). Myocardial perfusion scintigraphy can be useful in determining the location, extent and severity of such ischemia (12). Its prognostic value has been demonstrated both early (472) and late (473–475) after CABG.
After Exercise Testing. In patients who perform a treadmill exercise test that is not associated with an adequate exercise effort necessary to risk-stratify the patient appropriately, a repeat exercise test with thallium scintigraphy or a myocardial perfusion imaging test with pharmacologic stress may give a better indication of the presence or absence of high-risk coronary disease (14).
|
Important findings on stress echocardiography for risk stratification
|
|---|
Stress echocardiography is both sensitive and specific for detecting inducible myocardial ischemia in patients with chronic stable angina (13) (see Section II.C.4). Compared with standard exercise treadmill testing, stress echocardiography provides an additional clinical value for detecting and localizing myocardial ischemia. The results of stress echocardiography may provide important prognostic value. Several studies indicate that patients at low, intermediate and high risk for cardiac events can be stratified on the presence or absence of inducible wall motion abnormalities on stress echocardiography testing. A positive stress echocardiographic study can be useful in determining the location and severity of inducible ischemia, even in a patient with a high pretest likelihood that disease is present. A negative stress echocardiographic evaluation predicts a low risk for future cardiovascular events (420–428).
However, the value of a negative study compared with a negative thallium study must be further documented because there are less follow-up data compared with radionuclide imaging. Recently, McCully et al. (476) assessed the outcomes of 1,325 patients who had normal exercise echocardiograms with overall and cardiac event-free survival as end points. Cardiac events included cardiac death, nonfatal MI, and coronary revascularization. The event-free survival rates were 99.2% at one year, 97.8% at two years, and 97.4% at three years. Table 22 summarizes the prognostic value of stress echocardiography from the literature (studies with >100 patients who did not have recent MI and that included both positive and negative echocardiograms). The presence of ischemia on the exercise echocardiogram is independent and incremental to clinical and exercise data in predicting cardiac events in both men and women (477,478).
View this table:
[in this window]
[in a new window]
|
Table 22 Prognostic Value of Stress Echocardiography in Definite or Suspected Coronary Heart Disease (Studies With n > 100, Not Recent MI, Both Positive/Negative Echocardiograms)
|
|
The prognosis is not benign in patients with a positive stress echocardiographic study. In this subset, morbid or fatal cardiovascular events are more likely, but the overall event rates are rather variable. Hence, the cost-effectiveness of using routine stress echocardiographic testing to establish prognosis is uncertain.
In general, patients with a positive ECG response to treadmill stress testing but no inducible wall motion abnormality on stress echocardiography have a very low rate of adverse cardiovascular events during follow-up (13,420,421), albeit higher than in patients with negative ECG results as well. However, the number of patients followed up after both stress ECG and stress echocardiography is relatively small, and there has been no breakdown into groups with various METs achieved during ECG treadmill testing and with different risks according to the treadmill score (see Section II.C.2).
In patients with a significant clinical suspicion of CAD, stress echocardiography is appropriate for risk stratification when standard exercise testing is likely to be suboptimal (14). A variety of methods can be used to induce stress. Treadmill stress echocardiography may have lowered sensitivity if there is a significant delay from the end of exercise to the acquisition of postexercise images. Dobutamine stress echocardiography has substantially higher sensitivity than vasodilator stress echocardiography for detecting coronary stenoses (13,224,225,479). Sensitivity can also be diminished if all myocardial segments are not adequately visualized.
|
Application of stress echocardiography to specific patient subsets
|
|---|
Women, the Elderly, and Obese Patients. There are some recent data concerning the usefulness of stress echocardiography in women compared with men. Two studies by Marwick and associates (129,479) define the predictive value of exercise echocardiography as an independent predictor of cardiac events in women with known or suspected CAD. Symptom-limited exercise echocardiography was performed in 508 consecutive women (55 ± 10 years) between 1989 and 1993 (129), with a follow-up of 41 ± 10 months. Cardiac events occurred in 7% of women, and exercise echocardiography provided key prognostic information incremental to clinical and exercise testing data with a Cox proportional hazard model. In another group of women, the specificity of exercise echocardiography for indicating CAD and potential risk exceeded that of exercise electrocardiography (80 ± 3% vs. 64 + 3%, p = 0.05) and was a more cost-effective approach (129). Although these data are promising, the committee thought that in most women, ECG treadmill testing is still the first choice for detecting high-risk inducible myocardial ischemia.
The echocardiographic window and the number of myocardial segments detected during exercise or dobutamine echocardiography are often suboptimal in very obese patients and many elderly patients who have chronic obstructive lung disease and a suboptimal echocardiographic window. As mentioned previously (Section II.C.3), tissue harmonic imaging and contrast echocardiography should improve detection of the endocardium.
Left Bundle-Branch Block. Like exercise myocardial perfusion imaging studies, the significance of stress-induced echocardiography wall motion abnormalities in patients with left bundle-branch block is unreliable (13). During either exercise or dobutamine stimulation, abnormal contraction of the intraventricular septum has been a frequent occurrence in patients with left bundle-branch block who do not have underlying disease of the LAD.
After Coronary Angiography. Echocardiographic studies may help in planning revascularization procedures by demonstrating the functional significance of a given coronary stenosis. This may be of particular value in determining the need for PTCA, especially when the degree of angiographic stenosis is of uncertain physiologic significance or when multiple lesions are present (13).
After Revascularization. When symptoms persist or recur 6 months after CABG, echocardiographic testing can be useful. Abnormal baseline ECG findings after cardiac surgery are common, and postbypass patients frequently have abnormal ECG responses on standard treadmill testing. When symptoms of ischemia suggest incomplete revascularization, stress echocardiography studies may demonstrate the location and severity of residual ischemia. When symptoms recur after initial relief and stress echocardiogram demonstrates inducible ischemia, either graft closure or the development of new coronary artery obstructive lesions is likely (482).
After Treadmill Exercise Testing. As with stress myocardial perfusion imaging, stress echocardiography may provide additional information in patients unable to perform appropriate exercise on the treadmill and in those who have an intermediate risk determined by ECG criteria during exercise testing (13).
|
D. Coronary angiography and left ventriculography
|
|---|
|
Top
Committee members
Preamble
