CLINICAL GUIDELINE: Diagnosing Syncope: Part 1: Value of History, Physical Examination, and Electrocardiography

Methods

We used the MEDLINE database to identify articles related to syncope and diagnostic testing. References that evaluated the diagnostic test in near syncope and dizziness were included if they also used the test in patients with syncope. When a medical subject heading did not identify a sufficient number of references about a particular diagnostic test (such as neurovascular testing or carotid Doppler ultrasonography), keyword searches (using such terms as transcranial Doppler) were done. To be included in the review, articles had to be published in the English language between 1980 and 1995. The studies had to be randomized trials, observational studies, cohort studies, or case series of more than 10 patients (review articles and case reports were excluded); had to focus on or include patients with syncope; and had to examine only patients 18 years of age or older (except for tilt-table studies, which often included adult and pediatric cases in the same articles).

Articles that were candidates for review were evaluated in detail by one of the authors. Articles that met the selection criteria were used to prepare summary tables or paragraphs. Comparisons between groups (for example, the proportion of patients with and without heart disease who had tachyarrhythmias diagnosed by electrophysiologic testing) were made using the Fisher exact test.

Selected national experts in cardiology and neurology were asked to review the findings in their area of expertise. The opinions of these experts were incorporated into the recommendations.

Limitations of the Literature on Syncope

In syncope, there is no diagnostic gold standard against which other diagnostic tests may be measured; thus, sensitivity and specificity may not be easily calculated. Moreover, the presence of a disease, such as coronary disease, in a patient who has fainted does not prove that the disease caused the syncope. Syncope is, at its core, a symptom and not a disease. Therefore, this review is not organized around a technology or a disease entity but focuses on the physiologic states that lead to a sudden, transient loss of consciousness.

The literature that discusses syncope predominantly comprises case series or cohort studies based on referrals to tertiary care centers. We classified studies into three types: population-based studies (including unselected patients from the general population who were hospitalized or seen in emergency departments and other outpatient settings), referral-based studies (including patients referred to specialized centers for syncope workups), and small case series. To our knowledge, no randomized trials of the diagnostic workup or management strategies for patients with syncope have been done. A summary of the types of studies conducted in patients with syncope (Table 1) shows that most have been referral studies or case series.

Definitions

Organic Heart Disease

Whenever possible, our definition of organic heart disease included coronary artery disease, congestive heart failure, valvular heart disease, cardiomyopathy, and congenital heart disease. Because conduction system disease is a separate predictor of the need for special diagnostic testing, it was kept apart except where indicated. Patients who had a history and physical examination that were negative for cardiovascular symptoms or signs and a normal electrocardiogram were considered to have normal hearts; however, we recognize that some investigators think that echocardiography should be done before patients are declared free of organic heart disease.

Diagnostic Yield

For most tests, the diagnostic yield reflects the number of patients with positive diagnostic test results divided by the number of tested patients. For Holter and loop monitoring, the numerator includes the sum of the true-positive test results (arrhythmias during fainting) plus the true-negative test results (normal rhythm during symptoms). This expanded definition reflects the prognostic importance of a negative result on electrocardiography during syncope. For certain tests, the absolute value of the diagnostic yield may not be as important as the ability of the test to exclude a serious diagnosis (for example, intracardiac electrophysiologic studies may be of considerable benefit when they exclude ventricular tachycardia in a patient in whom that diagnosis was strongly considered).

Data Synthesis

Differential Diagnosis

The first category of syncope is neurally mediated syncope, which results from reflex mechanisms that are associated with inappropriate vasodilatation, bradycardia, or both (Table 2). This category includes vasovagal, vasodepressor, situational, and carotid sinus syncope. Neurocardiogenic mechanisms are also implicated in syncope associated with ventricular outflow obstruction (such as with aortic stenosis and pulmonary embolism) as well as supraventricular tachyarrhythmias [5-9]. The second category is orthostatic hypotension, which may result from age-related physiologic changes, volume depletion, medication, and autonomic insufficiency [10, 11]. Psychiatric disorders related to syncope (such as anxiety, depression, and conversion disorders) form a third category. The fourth category includes neurologic disorders, although these rarely cause syncope unless patients with seizures are included. Neurologic causes of syncope include transient ischemia (almost exclusively involving the vertebrobasilar territory), migraines (basilar artery), and seizures (atonic seizures, temporal lobe epilepsy, and unwitnessed grand mal seizures) [12].

Cardiac causes of syncope include coronary disease, congenital and valvular heart disease, cardiomyopathy, arrhythmias, and conduction system disorders. Coronary disease, congestive heart failure, ventricular hypertrophy, and myocarditis may set the stage for arrhythmia and syncope. Exertional syncope results from heart disease characterized by a fixed cardiac output that does not increase with exercise. Exertional syncope may also reflect arrhythmic or neurocardiogenic disorders or an anomalous coronary artery. Syncope may be the presenting symptom in elderly patients with acute myocardial infarction [13]; it rarely occurs with coronary artery spasm and aortic dissection.

We used five population-based studies of unselected patients to estimate the prevalence of various causes of syncope [14-18]; the summary of these studies is necessarily limited by the variability in diagnostic criteria. The most common causes of syncope were vasovagal episode, heart disease and arrhythmias, orthostatic hypotension, and seizures. The cause of syncope could not be determined in approximately 34% of patients. All of these studies were done several years ago, and the proportion of patients with unexplained syncope is probably lower now, given wider use of event monitoring, tilt testing, electrophysiologic studies, attention to psychiatric illnesses, and recognition that the cause of syncope in elderly patients may be multifactorial.

Approach to Syncope

The algorithm depicted in Figure 1 provides a diagnostic approach to syncope. It is intended to provide a framework for clinical judgment, not to replace it. Key points in the algorithm that will be discussed in the text include the following.

1. History, physical examination, and electrocardiography are the core of the workup for patients with syncope.

2. Carotid sinus massage may be useful in elderly patients but should not be done by the generalist if bruits are present, if the patient has a history of ventricular tachycardia, or in the setting of a recent stroke or myocardial infarction. A false-positive test result should be suspected if carotid massage is positive but the history does not suggest carotid hypersensitivity.

3. Special issues for elderly patients include the multifactorial nature of syncope, polypharmacy, use of carotid sinus massage, and cardiac testing (exercise stress test and echocardiography) to exclude cardiac disease.

4. Nondiagnostic arrhythmias found on Holter monitor readings should not usually be treated.

5. Intracardiac electrophysiologic studies are most useful in patients who have organic heart disease and otherwise unexplained syncope.

6. In a patient with exertional syncope, echocardiography should precede exercise stress testing.

7. The assessment of patients with a normal heart who have frequent episodes of syncope should include a loop recorder and psychiatric evaluation.

8. The workup of patients with a normal heart who have infrequent episodes of syncope should include a tilt test and psychiatric evaluation.

9. Neurologic testing, including electroencephalography, computed tomography, and carotid and transcranial Doppler ultrasonography, should be reserved for patients who have neurologic signs or symptoms or carotid bruits.

History and Physical Examination

Table 3, which includes data from six population-based studies, shows that the history and physical examination identify a potential cause of syncope in 45% of patients whose primary disorder can be diagnosed. Furthermore, organic cardiac diseases that cause syncope (such as aortic stenosis, idiopathic hypertrophic subaortic stenosis, or pulmonary embolism) and neurologic diseases (such as the subclavian steal syndrome) are frequently suspected on the basis of the history and physical examination. One study reported that suggestive findings on the history and physical examination were helpful in assigning a cause by directed testing in 8% of additional patients. History taking should focus on postural symptoms (orthostatic or vasovagal syncope), exertional symptoms or a positive family history (cardiac syncope, such as prolonged QT syndromes), palpitations (arrhythmia), postictal symptoms (neurologic syncope), situational symptoms (such as defecation and urination), use of medication, and history of organic heart disease (predisposing to arrhythmias or ischemia). A seizure without typical postictal symptoms may suggest an alternative cause, such as hypotension caused by arrhythmia or vasovagal syncope. A history taken from a family member or witness can be helpful.

Physical findings that are useful in diagnosing syncope include orthostatic hypotension, cardiovascular signs, and neurologic signs. Orthostatic hypotension is implicated in 8% of patients with syncope (range, 4% to 12%) [4, 14-19]. One study [20] found that 31% of patients with syncope had orthostasis (defined as a decline of 20 mm Hg in systolic blood pressure after standing). In 90% of patients, this was apparent within 2 minutes of standing up-right. Other important cardiovascular findings include differences in blood pressure in each arm or signs of aortic stenosis, idiopathic hypertrophic subaortic stenosis, pulmonary hypertension, myxomas, and aortic dissection.

Some patients with syncope have a history of concomitant dizziness (lightheadedness). In one study, a psychiatric cause for symptoms was implicated in many of these patients, especially those who had a history of vertigo (24% of patients with syncope and dizziness compared with 5% of patients with syncope alone [P < 0.01]) [21]. However, syncope and dizziness can also be a sign of cardiac arrhythmias. A thorough history and physical examination are thus mandatory [22] and should focus on cardiac, neurologic, and medication-related issues. In younger patients who have normal hearts, Holter monitors, loop monitors, or head-up tilt-table testing can help determine whether symptoms are caused by a cardiac or vasovagal abnormality. Older patients may have more serious cardiac arrhythmias, orthostatic hypotension, or neurologic causes.

Medications frequently cause syncope, especially in elderly patients who are receiving several medications [23]. In a referral study of adverse drug reactions and syncope [23], antihypertensive and antidepressant agents were most commonly implicated. Other medications that are often associated with syncope include antianginal agents, analgesics, and central nervous system depressants. Blood levels of medication may be useful for diagnosis, but the most important ways to confirm medication-induced syncope are to document side effects of medication (such as bradycardia or orthostatic hypotension) that can lead to syncope or to discontinue the medication and follow the patient for remission of syncope. Concerns about medications that might predispose patients to malignant arrhythmias (for example, concern about quinidine producing torsades de pointes) would mandate hospitalization. Ambulatory monitoring of blood pressure may document episodes of medication-induced orthostasis.

Although syncope may be relatively common in pregnancy, remarkably few researchers have attempted to assess its cause, natural history, and workup in this setting. Although we found more than 52 000 papers on pregnancy in a MEDLINE search of the literature published since 1980, only 7 articles focused on syncope. All of these were small case series involving seven or fewer patients. Aortocaval compression by an enlarged uterus, especially in the supine position, may lead to syncope in the third trimester [24]. Pregnant patients with known heart disease or arrhythmias, a pathologic murmur, exertional syncope, or palpitations with syncope clearly require further evaluation [25]. Further research is needed to help clinicians assess risk and the need for diagnostic evaluation in other pregnant women.

Electrocardiography at Baseline

An abnormal electrocardiogram is found in many patients with syncope. Common findings include bundle-branch block, previous myocardial infarction, and left ventricular hypertrophy [16]. It should be noted that most patients with these findings do not have an identifiable cardiac cause for syncope. Indeed, as shown in Table 3, causes of syncope were determined in only 5% of patients by electrocardiography, by rhythm strip done by paramedics, or in the emergency department [4, 14-18], primarily because of the transient nature of arrhythmias. The most common diagnoses included ventricular tachycardia; bradyarrhythmias; and, less commonly, acute myocardial infarction. Findings of first-degree heart block, bundle-branch block, and sinus bradycardia may predict a cause for syncope attributable to bradycardia, whereas previous myocardial infarction or pronounced left ventricular hypertrophy in hypertrophic cardiomyopathy may be associated with ventricular tachycardia.

Although the yield of electrocardiography is low (5%), the test is risk free and relatively inexpensive. Moreover, finding such abnormalities as bundle-branch block, previous myocardial infarction, and nonsustained ventricular tachycardia will guide further evaluation that may detect life-threatening disorders. Electrocardiography is therefore recommended in almost all patients with syncope.

Basic Laboratory Testing

Routine blood tests (blood count and tests for electrolytes, blood urea nitrogen concentration, creatinine concentration, and glucose level) rarely yield diagnostically useful information. In studies that included patients with seizures, 2% to 3% of patients had hypoglycemia, hyponatremia, hypocalcemia, or renal failure [4, 14-18]. Routine blood tests usually confirmed a clinical suspicion; in one study [14], only one unexpected finding was discovered (hyponatremia with seizures). Bleeding as a cause of syncope was usually diagnosed clinically.

Routine use of basic laboratory tests is not recommended; these tests should be done only if they are specifically suggested by the results of the history or physical examination. Pregnancy testing should be considered in women of child-bearing age, especially those for whom tilt-table or electrophysiologic testing is being considered.

Patients with a Suggestive History

Patients with exertional syncope (in whom detection of serious cardiac disease requires echocardiography and stress testing), valvular heart disease, a history that suggests pulmonary emboli or pulmonary hypertension, neurologic signs or symptoms of syncope, or a positive family history of syncope or sudden death (prolonged QT syndromes) are included in the broad category of patients with a suggestive history. This category contains patients in whom the clinician strongly suspects a diagnosis after history, physical examination, and electrocardiography. Because many of the cardiac testing indications are discussed in part II of this paper, this section focuses on indications for neurologic testing.

Neurologic Testing

Neurologic tests used for patients with syncope include electroencephalography, brain imaging (computed tomography or magnetic resonance imaging), and neurovascular studies (carotid and transcranial oppler ultrasonographic studies). To determine which patients may benefit from neurologic testing, physicians should take a particularly careful neurologic history (for example, patients should be asked about a history of seizure activity, prolonged loss of consciousness, diplopia, headache, and postictal symptoms) and perform a thorough, focused physical examination (including a search for bruits or focal neurologic signs).

Electroencephalography

In the early 1980s, electroencephalography was one of the cornerstones of the workup for patients with syncope [26]. However, several studies [4, 15, 26-29] conclusively showed that electroencephalographic monitoring was of little use in unselected patients with syncope. In the absence of a history of seizure activity, electroencephalography has provided few diagnoses in more than 500 patients reported in the literature (Table 4). Eight of 534 patients were diagnosed using electroencephalography; 2 of these 8 patients had clinical data provided, and both had a history of seizures. Thus, electroencephalography is not recommended for patients with routine syncope and may only be beneficial in patients with a history of seizures.

Computed Tomography and Magnetic Resonance Imaging

No identifiable studies have specifically evaluated the use of brain imaging for patients with syncope. Early case series of such patients [4, 12, 14, 15, 27] (Table 4) found that computed tomography produced new information only in patients with focal neurologic signs. Of 195 patients who were studied, the average yield of computed tomography was 4%; all patients who had positive scans had a focal neurologic examination or a witnessed seizure. The diagnostic utility of magnetic resonance imaging in syncope has not been studied. Thus, computed tomography and magnetic resonance imaging should be avoided unless physical or historical features of central nervous system focality are present.

Neurovascular Studies

No single study has focused on the usefulness of transcranial Doppler ultrasonography for patients with syncope. The available studies [30-32] are insufficient to evaluate the usefulness of this test, perhaps because transient ischemic attacks involving the vertebral and basilar arteries rarely result in isolated syncope. Drop attacks (that is, sudden losses of postural tone without a clear-cut loss of consciousness) [33] can be vertebrobasilar in origin, but it is unclear whether transcranial Doppler ultrasonography can identify the cause of these events.

Anterior cerebral circulatory events rarely cause syncope. To create optimal conditions for an anterior circulatory event that could result in syncope, complete occlusion of one carotid artery and nearly complete occlusion of the other would have to occur. Few studies have evaluated carotid Doppler ultrasonography in certain neurologic conditions, including syncope, and no study has examined the usefulness of this test in syncope. One referral study found occlusive plaques in the carotid artery of 3 of 46 patients who had syncope after pacemaker implantation [32], but it is uncertain whether these plaques would have caused syncope. We know of no other studies that suggest that carotid Doppler ultrasonography is beneficial for patients with syncope, unless signs of cerebrovascular disease (such as previous strokes or bruits) are present.

Neurologic testing in syncope should be guided by the history and physical findings. Specifically, if evidence of seizure activity is present, electroencephalography may be useful. Focal neurologic signs mandate brain imaging, usually with computed tomography. Carotid or transcranial Doppler ultrasonography may be performed in the presence of bruits or when the history suggests vertebrobasilar insufficiency (for example, prolonged loss of consciousness, diplopia, nausea, or hemiparesis). Patients who have seizure activity, normal results on electroencephalography, and no postictal symptoms and patients with seizures who do not respond to anticonvulsant medications should be evaluated for possible cardiac syncope [34].

Appendix

The following are members of the Clinical Efficacy Assessment Subcommittee of the Health and Public Policy Committee of the American College of Physicians: George E. Thibault, MD, Chair, John R. Feussner, MD, Co-Chair, Anne-Marie J. Audet, MD; Gottlieb C. Friesinger Jr., MD; Daniel L. Kent, MD; Keith I. Marton, MD; Valerie Anne Palda, MD; John J. Whyte, MD; and Preston L. Winters, MD.

From University of Wisconsin School of Medicine, Madison, Wisconsin; New England Medical Center, Boston, Massachusetts; and University of Pittsburgh, Pittsburgh, Pennsylvania.

Mr. Yang: University of Wisconsin School of Medicine, 1300 University Avenue, Madison, WI 53705.

Drs. Estes and Wang: New England Medical Center, Division of Cardiology, 750 Washington Street, Boston, MA 02111.

Dr. Vorperian: University of Wisconsin School of Medicine, Department of Medicine, H6/375 Clinical Science Center, 600 Highland Avenue, Madison, WI 53792.

Dr. Kapoor: University of Pittsburgh, Lothrop Hall, Room 100, 190 Lothrop Street, Pittsburgh, PA 15261.