Address correspondence to Peter O’Callaghan, Department of Cardiology, University Hospital of Wales, Cardiff CF14 4XW, U.K. E-mail: firstname.lastname@example.org
Clinicians who diagnose and manage epilepsy frequently encounter diagnoses of a nonneurological nature, particularly when assessing patients with transient loss of consciousness (T-LOC). Among these, and perhaps the most important, is cardiac syncope. As a group, patients with cardiac syncope have the highest likelihood of subsequent sudden death, and yet, unlike sudden unexpected death in epilepsy (SUDEP) for example, it is the norm for these tragic occurrences to be both easily predictable and preventable. In the 12 months following initial presentation with cardiac syncope, sudden death has been found to be 6 times more common than in those with noncardiac syncope (N Engl J Med 309, 1983, 197). In short, for every patient seen with T-LOC, two fundamental aims of the consultation are to assess the likelihood of cardiac syncope as the cause, and to estimate the risk of future sudden death for the individual. This article aims to outline for the noncardiologist how to recognize cardiac syncope, how to tell it apart from more benign cardiovascular forms of syncope as well as from seizures and epilepsy, and what can be done to predict and prevent sudden death in these patients. This is achieved through the assessment triad of a clinical history and examination, risk stratification, and 12-lead electrocardiography (ECG).
A clear understanding of fundamental concepts relating to transient loss of consciousness (T-LOC) forms the foundation for understanding cardiac syncope. T-LOC is defined as an abrupt, self-limiting complete loss of consciousness, not caused by trauma (Moya et al., 2009). Patients presenting with a single T-LOC, particularly if accompanied by other symptoms such as chest pain or shortness of breath, are more likely to have an acute symptomatic cause such as myocardial infarction, aortic dissection, or pulmonary embolism and may need to be investigated as acute medical emergencies. However, most patients who have experienced recurrent T-LOC have a narrower differential diagnosis, essentially comprising syncope, seizures, and nonepileptic attacks. Other conditions that might result in recurrent T-LOC, for example, basilar migraine, are much rarer.
Syncope is defined similarly to T-LOC but with the addition of an inferred mechanism, that is, a transient complete loss of consciousness accompanied by loss of muscular tone, resulting from global cerebral hypoperfusion. Where there is no evidence that this is the mechanism for T-LOC, the term syncope should not be used. Syncope occurs though one of two broad mechanisms: benign blood pressure regulatory causes such as reflex (neurocardiogenic) syncope (approximately 70%) and orthostatic hypotension (approximately 10%), and cardiac syncope (approximately 15%) which is potentially life-threatening. Syncope remains unexplained in a small number of patients (Del Rosso et al., 2008). Understanding the physiologic underpinnings of these varieties of syncope makes their clinical detection easier. In reflex syncope, cardiovascular reflexes are inappropriately activated, leading to either a vasodepressor (falling blood pressure), cardioinhibitory (falling heart rate), or mixed response that results in loss of cerebral perfusion. This reflex mechanism follows an identifiable trigger in most patients, and these triggers are wide ranging. In some patients, specific actions such as coughing or micturition will provoke syncope, and in others situational factors are more important, such as emotional shock or prolonged standing. The type of trigger is not as important as its presence or absence, as will be described later. In the cardioinhibitory form of reflex syncope, intense vagal activation results in prodromal nausea, which reassures the clinician that the mechanism is unlikely to be a primary cardiac/arrhythmic disorder. In orthostatic hypotension, loss of cerebral perfusion is mediated by a more fixed reduction in peripheral vascular resistance or venous return, leading to reduced cerebral perfusion pressure. Patients frequently report postural symptoms and may have a significant difference in blood pressure between supine and erect measurements. In contrast to reflex syncope, orthostatic syncope is usually secondary to some other condition, including autonomic nervous system dysfunction, vasodilating drugs (usually antihypertensives), or volume depletion/impaired venous return (e.g., dehydration). Finally, cardiac syncope is due to a disorder of the structure or function of the heart itself. The mechanism may be arrhythmia (bradycardia or tachycardia) or may be secondary to mechanical obstruction.
As with seizures and epilepsy, a witness account is invaluable in diagnosing syncope, and should be sought if at all possible. Clinical assessment should focus on the patient’s symptoms and appearance before, during, and after T-LOC, from the perspectives of both the patient and the witness. Finally, all this information must be interpreted in the context of the situation that the patient was in at the time, including any provocative events immediately before the T-LOC. Numerous attempts have been made to convert clinical data from prospective syncope cohorts into scoring systems aimed at detecting cardiac syncope, with a best estimated sensitivity of 98%, and specificity of 56% (Martin et al., 1997; Alboni et al., 2001; Brignole et al., 2006; Quinn et al., 2006; Sheldon et al., 2006; Del Rosso et al., 2008; van Dijk et al., 2008). In all varieties of syncope the features of the loss of consciousness are rather similar, and attempts to differentiate them based on these features alone will lead to diagnostic errors. Such features include the presence or absence of pallor, shallow breathing, no palpable pulse, and incontinence. These features might not even reliably differentiate syncope from seizures.
The context in which syncope occurs is useful in differentiating the cause. It is well known that reflex syncope frequently occurs in response to noxious stimuli, emotional shock, venipuncture, prolonged standing, micturition, or cough, but perhaps less recognized is its association with a postexercise state. This is important, as syncope that occurs during exercise has a strong association with cardiac syncope (resulting from some channelopathies, or cardiac outflow obstruction, for example), and so clarification of this point in individual patients is of considerable clinical value. Reflex syncope is also associated most strongly with a standing (or less commonly sitting) posture, whereas cardiac syncope can occur in any posture. As such, in recurrent T-LOC, clarifying if any events have ever occurred while lying down may raise the possibility of cardiac syncope. This feature is even more useful in orthostatic syncope, where in most cases there is a reliable association between presyncopal symptoms and standing, and syncope almost always occurs soon after standing. Caution is required here, however, as in many patients, the symptoms are significantly delayed following standing (Gibbons & Freeman, 2006). This should be borne in mind when analyzing symptoms, but also when using lying and standing blood pressure measurements to guide diagnosis (American Academy of Neurology, 1996).
Warning (prodromal) symptoms leading up to syncope are a useful indicator of the cause, and are generally grouped under the term presyncope (presyncope can occur without progression to syncope). The duration of such symptoms is variable: in reflex and orthostatic syncope the symptoms may range from a few seconds to several minutes. In both, the patient experiences a light-headed sensation and nausea most commonly. Other warning symptoms that suggest a benign syncopal mechanism include tinnitus, a feeling of distance from one’s surroundings, generalized weakness/fatigue, and sweating. A prodrome that is brief suggests cardiac syncope; indeed, most patients with cardiac syncope report no warning symptoms, and although this can also occur in reflex and orthostatic syncope it should be regarded as a feature worthy of more urgent investigation. In those patients with cardiac syncope who do have warning symptoms, these are typically more characteristic of reduced cerebral perfusion (brief period of light-headedness, tunnel vision) with a noticeable absence of autonomic symptoms (nausea and sweating).
A major contributor to erroneous epilepsy diagnoses is the presence of myoclonic jerks during syncope, a scenario known as convulsive syncope. It is important to note that these jerks, although superficially appearing epileptiform to the inexperienced observer, are arrhythmic and often asymmetric/multifocal. Electroencephalography evidence suggests these are not ictal (Aminoff et al., 1988). They are more likely to occur in prolonged syncope, which may have a cardiac cause, and syncope, where the patient falls in a way that does not afford them the usual horizontal position that assists recovery. Alternatively, a helpful bystander may misguidedly sit the patient up soon after syncope, unwittingly making matters worse. When jerks occur in this scenario, it seems it is the natural inclination of the witness to assume a seizure has occurred, but with a careful and detailed history this should not confuse clinicians. Rather infamously, such myoclonic jerks, and other epileptiform features, occurred in 90% of syncopal episodes under controlled conditions, further stressing the importance that clinicians are not misled by their presence (Lempert et al., 1994).
The recovery that a patient experiences (and that a witness observes) is crucial in determining the cause of T-LOC and syncope. Vagal parasympathetic hyperactivity in the setting of reflex syncope gives rise to the familiar postsyncopal symptoms of nausea, vomiting, fatigue, and generalized weakness. Patients may appear drowsy, but these symptoms gradually remit, and if pressed, the patient and witness usually recall that orientation and memory were resumed soon after the T-LOC was over, typically before paramedics arrive, for example. This is in stark contrast to the aftermath of cardiac syncope, where patients experience a rapid and falsely reassuring sense of wellbeing on regaining consciousness, perhaps due to endogenous sympathetic activity arising in response to the underlying condition. As such, the absence of any “hang-over” effects following syncope should be regarded as a worrying feature. One caveat in this regard is where cardiac syncope progresses to prolonged cardiac arrest and cerebral hypoxia. Such patients may have a hypoxic seizure and postevent amnesia that may last days (if they spontaneously regain cardiac output and survive), that is, a mild hypoxic-ischemic brain injury. If the event was witnessed, the delay between syncope (collapse) and the onset of the seizure, which is likely to be several minutes, provides a clue to the underlying mechanism. Readers will be familiar with the differences in the case of generalized seizures, where convulsive movements develop from the onset of T-LOC, and postictal confusion and amnesia may last a few hours, but very rarely more.
To complete the clinical assessment for cardiac syncope, the clinician should perform a cardiovascular examination searching for evidence of structural heart disease. This not only includes auscultation, but also a search for manifestations such as peripheral edema, raised jugular venous pressure, and even stigmata that might imply coronary artery disease, such as peripheral vascular disease, xanthelasma, and so on. Symptoms elicited from the history such as reduced exercise tolerance and breathlessness when lying flat may also imply the presence of structural heart disease. Together with the ECG and previous medical and family history, this forms the basis of the second fundamental part of the assessment, risk stratification, even in the absence of any physical signs.
Syncope is common, with a large prospective cohort of adults showing an incidence of 10.5% over 24 years (822 of 7,814, aged 20–96 years, mean 51.1) (Soteriades et al., 2002). The incidence of syncope increases markedly in the elderly as coronary artery disease and orthostatic intolerance (often from prescribed drugs) move toward ubiquity (Lipsitz et al., 1985). However, for the most common variety—reflex syncope—onset is usually in adolescence, with 1 in 2 women and 1 in 4 men having experienced syncope by aged 21 years (Ganzeboom et al., 2003). The task of risk stratification should be viewed as separate to the assessment of the syncopal event in question (and in many ways it is easier to perform). This is because even if the clinician is confident that the event represented reflex syncope, for example, the presence of structural heart disease still implies an increased mortality for that patient. Survival of the syncopal episode is an opportunity to prevent future sudden death. Cardiac mortality rates for those with and without (all forms of) syncope over a 1-year period are not significantly different (3% vs. 6%, p = 0.08), but if the patient is male, aged >55 years, or has congestive heart failure, this seems independently to predict a higher chance of sudden cardiac death (Kapoor & Hanusa, 1996). When the cause of syncope is thought to be cardiac, the mortality from sudden death in the 12 months following initial presentation is 24%, compared with 4% in those with noncardiac syncope (Kapoor et al., 1983). These poor outcomes continue, with 5-year sudden death rates of 33.1% for cardiac syncope, 4.9% for noncardiac syncope, and 8.5% for syncope of undetermined cause (Kapoor, 1990). This indicates that when the cause of syncope is unknown, it is safest to presume it is cardiac until proven otherwise. All-cause mortality rates are similarly increased, not just those for sudden death.
So, the presence of syncope that is judged to be cardiac, and also the presence of structural heart disease in a patient with syncope (regardless of the cause of syncope, and especially if it is “unexplained”) are both associated with significantly increased risks of future sudden death. Similarly, an abnormal 12-lead ECG confers varying degrees of increased mortality risk, but these will be covered under their respective conditions in the next section. Finally, the clinician should seek a family history of sudden death in all cases. Enquiries should extend to two or three generations, and in sensitive situations it may be best to ask whether any unexpected or unexplained tragedies have occurred in the family. For the purposes of risk stratification, national and international guidelines vary in their definition of a significant upper age limit for sudden cardiac death in a relative, with most recent United Kingdom guidance saying <40 years of age (Westby et al., 2010) and European guidance <30 years (Moya et al., 2009). The authors use a cutoff of a first-degree relative dying a sudden cardiac death at an age younger than 35 years, although decisions are never made solely on this basis. Most sudden deaths in those older than 40 years of age are due to acquired rather than genetic cardiac diseases.
This information can be distilled in to a risk stratification schema as suggested in Figure 1. In essence, a patient with a typical history of reflex or orthostatic syncope, including an identifiable trigger and prodromal symptoms with no family history of sudden death and a normal ECG, probably has no increased risk of future cardiac morbidity or mortality and can be reassured accordingly, without necessarily needing further investigation. As patients accumulate features to the contrary, their need for urgent assessment by specialist cardiologic services increases, and in select cases the clinician should consider immediate hospital admission. Another more pragmatic aspect of risk stratification is in the setting of frequent reflex syncope with only brief warning symptoms. These patients are often falling frequently and risk injuring themselves, and may need to be assessed urgently and treated for safety reasons, for example to reduce their risk of head injury.
The combination of a clinical assessment, risk stratification, and a 12-lead ECG is often sufficient to suggest cardiac syncope, but it will not always clarify the cause. Further cardiac investigations include echocardiography, noninvasive ECG monitoring, implantable loop recorders, and diagnostic electrophysiologic studies. The causes of cardiac syncope are best classified as structural heart disease (most common), arrhythmia in structurally normal hearts, and the inherited cardiac channelopathies (least common).
Structural heart disease
Cardiac syncope can arise from arrhythmia (more commonly) or inflow/outflow obstruction in patients with structural heart disease. Acquired causes are more common in older patients; the most common cause of cardiac syncope/arrest in Western societies is ventricular tachycardia in “scarred” hearts (scar-related VT). This is usually in the context of a remote myocardial infarction, but nonischemic cardiomyopathies (alcohol-related, idiopathic dilated cardiomyopathy) are also important. Twelve-lead ECG can suggest both conditions when they may have passed unrecognized by the patient. Pathologic Q waves are a common ECG marker of previous myocardial infarction and are associated with an increased risk of sudden cardiac death. These are defined as an initial downward deflection of ≥0.04 s duration (one small square in width) and ≥0.2 mV amplitude (2 small squares in depth) in two contiguous inferior, lateral, or anterior ECG leads (Fig. 2). The finding of Q waves in the context of syncope should prompt specialist cardiologic assessment.
In middle- and older-aged patients, syncope due to inflow/outflow obstruction usually has an acquired cause, most commonly valvular disease (e.g., aortic stenosis), cardiac tumors, and pericardial disease. However, in all of these conditions, patients typically experience other symptoms such as dyspnea, reduced exercise tolerance, and peripheral edema, and have clinical signs that make a cardiac cause readily apparent when syncope occurs. Therefore, they are less of a “heartache” for clinicians who usually deal with epilepsy.
In younger patients, inherited structural heart disease is more common and hypertrophic cardiomyopathy, familial dilated cardiomyopathy, and arrhythmogenic right ventricular cardiomyopathy are the major causes. Inheritance is usually autosomal dominant, and the presence of recurrent syncope, heart failure, and/or a history of familial sudden death are all indications for insertion of an implantable cardioverter defibrillator.
Structurally normal hearts
Patients whose cardiac syncope is an isolated symptom, and where clinical examination and echocardiography are normal, are far more likely to be misdiagnosed with epilepsy.
Disease affecting the sinus node or the specialized (His–Purkinje) conducting system of the heart can cause bradyarrhythmias; syncope develops rapidly in this scenario. There may be obvious ECG findings of second degree or complete heart block, but it is important to recognize that conducting system disease can intermittently progress to complete heart block. Therefore, syncope in a patient with left bundle-branch block or trifascicular block (right bundle-branch block, left-axis deviation and first-degree heart block) suggests bradycardia as the cause and requires permanent pacemaker insertion. Drug-induced bradycardia (sinus bradycardia) may require temporary intervention, but once the cause is addressed, this should not confer an increased mortality risk.
Atrial fibrillation, atrial flutter, and supraventricular tachycardia are common, but most patients experience palpitations without syncope. However, in some conditions, such as atrial flutter with a 1:1 ventricular conduction, and a stiff ventricle (that cannot sustain cardiac output at increased heart rates), syncope can occur. Moreover, the onset of supraventricular tachycardia can act as a trigger for reflex syncope. In the Wolff-Parkinson-White syndrome, the accessory conducting pathway manifests as a preexcited delta wave on a standard 12-lead ECG (Fig. 3), a feature that needs to be recognized, as patients can progress to preexcited atrial fibrillation and cardiac arrest. For most tachyarrhythmias, pharmacologic treatment is an option, but ablation is now considered first-line therapy for the management of Wolff-Parkinson-White syndrome, atrioventricular node-dependent tachycardias, and atrial flutter, especially if these tachyarrhythmias have caused syncope.
Inherited cardiac channelopathies
These conditions, although rare, represent the greatest diagnostic challenge to the clinician dealing with epilepsy; patients are young, often with no symptoms other than syncope (without warning) and may have prominent myoclonic jerks during the T-LOC. The central dilemma is that when they present mimicking a “first fit,” the second symptom may be sudden death (Chadwick et al., 2010). Ion channel mutations underlie the clinical manifestations, and most can be detected on a standard 12-lead ECG (Marban, 2002; Kaufman, 2009). It is important to note that once recognized, subsequent sudden cardiac death is usually preventable with appropriate intervention (Garratt et al., 2010). Congenital long QT syndrome (LQTS) is the most common channelopathy, characterized by prolonged myocardial depolarization—corrected QT (QTc) intervals are often visibly prolonged on 12-lead ECG but a manual calculation is advised. This is most easily done using Bazett’s formula, but this calculation is unreliable at higher ventricular rates, and both clinicians and automated machines make frequent errors (Viskin et al., 2005). T-wave morphology may also be grossly abnormal, which can both suggest the condition and make automated QTc measurement unreliable (Fig. 4). There are many subtypes, resulting from sodium or potassium channel mutations. Inheritance is usually autosomal dominant (Romano-Ward syndrome) but there is an autosomal recessive variant with associated deafness known as Jervell and Lange-Nielsen syndrome (Schwartz et al., 2006; Webster & Berul, 2008). Syncope and sudden death in congenital long QT syndrome often occurs during sleep, or during exercise, including swimming. Sudden emotional shock or surprise is a recognized trigger in some subtypes. Once recognized, β-blockade can control symptoms and reduce the risk of sudden death, but if symptoms persist, or there are other high-risk markers (e.g., previous cardiac arrest, QTc > 500 msec), implantable cardioverter defibrillator insertion is recommended.
Mutations in the SCN5A gene encoding voltage-gated sodium channels account for most cases of Brugada syndrome, an autosomal dominant cardiac channelopathy that presents with syncope or aborted cardiac arrest, typically at rest or during sleep (Brugada & Brugada, 1992; Shimizu et al., 2005). The condition is more common in male individuals and South-East Asian populations. Most patients have the characteristic ECG appearance of “pseudo” right-bundle branch block and “coved” ST elevation in the right precordial leads (V1–V3), exacerbated by fever, electrolyte disturbance, and sodium channel blockers. Where the resting ECG is normal, the ECG abnormality can be brought out by sodium channel blockers (e.g., ajmaline) to assist diagnosis (Fig. 5). Syncope/cardiac arrest is usually an indication for an implantable cardioverter defibrillator, particularly if there is a resting ECG abnormality. A family history of sudden death does not seem to confer a greater risk in this condition.
Catecholaminergic polymorphic ventricular tachycardia is an uncommon cardiac channelopathy, characterized by the reproducible association of physiologic or emotional stress with the onset of potentially fatal ventricular tachycardia or fibrillation (Leenhardt et al., 1995). Patients are usually young (childhood/adolescent onset) and may have a family history of sudden death (30%). There are both autosomal dominant and recessive forms. Mutations in the human cardiac ryanodine receptor gene (hRyR2) cause most of the autosomal dominant forms (Priori et al., 2001), and the calsequestrin-2 (CASQ2) gene causes most of the recessive types (di Barletta et al., 2006). Unfortunately, mortality risk is particularly high and all investigations (including ECG) are characteristically normal, and so recognizing the characteristic clinical phenotype is crucial. Asymptomatic carriers, identified through symptomatic family members, may be treated with β-blockers (penetrance and risk to family members is high), but all symptomatic patients should undergo implantable cardioverter defibrillator insertion. β-Blockers do not offer the same level of protection as in LQTS.
A more comprehensive review of how to use the 12-lead ECG as a neurologist is detailed elsewhere (Marsh et al., 2008). As outlined earlier, there are some easily recognized 12-lead ECG abnormalities associated with risk of sudden death with which all clinicians should be familiar (Brugada & Geelen, 1997). It should be standard practice for all patients with T-LOC to undergo a 12-lead ECG at the earliest opportunity; it is inexpensive, entirely safe, and readily available. For clinicians who regularly deal with epilepsy and T-LOC it is recommended that a pathway is in place to have these readily reviewed by an appropriate specialist (e.g., a cardiologist) where any doubt arises (Westby et al., 2010). Limitations of ECG use that clinicians should particularly be aware of are the difficulty of detecting prolonged QT intervals at increased ventricular rates, and dynamic changes as seen in LQTS and Brugada syndrome, that may mean abnormalities are missed when only the outpatient ECG is reviewed. Tracking down an ECG performed around the time of symptoms may reduce unnecessary work and risk. Twenty-four or 48-h ambulatory ECG monitoring is frequently used when patients are admitted to hospital with T-LOC, but is only likely to yield useful results when the attacks are frequent (e.g., daily). Long-term ambulatory ECG monitoring in the form of an implantable loop recorder is far more appropriate for patients with recurrent T-LOC. The role of the noncardiologist when seeing patients with syncope is to use the opportunity provided by their recovery to perform a risk assessment and direct high-risk patients to specialist cardiology services in an appropriate time frame. Many patients will go on to have echocardiography, exercise testing, or more invasive electrophysiologic studies, but these are out side of the scope of this article.
Misdiagnosis of epilepsy is common, and there are many differential diagnoses. When cardiac syncope is misdiagnosed as epilepsy, the potential consequences could hardly be more devastating; the patient remaining at risk of a sudden death that is both predictable and preventable. The major contributors to misdiagnosis are the unpredictable and sudden nature of T-LOC in both conditions, as well as the frequent occurrence of myoclonic jerks in syncope that are mistaken for the clonic jerks of generalized tonic–clonic seizures. The clinical features of T-LOC can also be used to differentiate cardiac syncope from more benign dysfunction of blood pressure regulation (e.g., reflex syncope).
All patients with T-LOC should undergo a detailed assessment of their personal and family medical history, a 12-lead ECG and a clinical examination of their cardiovascular system (looking for evidence of structural heart disease). We have described how this information can be easily combined to stratify a patient’s risk of sudden death and in doing so provide a framework for the noncardiologist to make efficient and judicious use of specialist cardiology services. The cause of cardiac syncope can be more accurately determined through specialist assessment, and thus decisions are made on the mode and degree of intervention required to prevent subsequent sudden death.
The authors have no conflicts of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.