Author + information
- Received September 19, 2000
- Revision received January 24, 2001
- Accepted February 12, 2001
- Published online June 1, 2001.
- Paolo Alboni, MD, FACC∗,* (, )
- Michele Brignole, MD†,
- Carlo Menozzi, MD‡,
- Antonio Raviele, MD§,
- Attilio Del Rosso, MD∥,
- Maurizio Dinelli, MD∗,
- Alberto Solano, MD† and
- Nicola Bottoni, MD‡
- ↵*Reprint requests and correspondence: Dr. Paolo Alboni, Division of Cardiology, Ospedale Civile, 44042 Cento (Fe), Italy
We sought to establish what historical findings are predictive of the cause of syncope.
The clinical features of the various types of syncope have not been systematically investigated.
Three hundred forty-one patients with syncope were prospectively evaluated. Each patient was interviewed using a standard questionnaire. A cause of syncope was assigned using standardized diagnostic criteria.
A cardiac cause of syncope was established in 23% of the patients, a neurally mediated cause in 58% and a neurologic or psychiatric cause in 1%, and in the remaining 18%, the cause of syncope remained unexplained. In a preliminary analysis including age, gender and the presence of suspected or certain heart disease after the initial evaluation, only heart disease was an independent predictor of a cardiac cause of syncope (odds ratio 16, p = 0.00001), with a sensitivity of 95% and a specificity of 45%. In contrast, the absence of heart disease allowed us to exclude a cardiac cause of syncope in 97% of the patients. In patients with certain or suspected heart disease, the most specific predictors of a cardiac cause were syncope in the supine position or during effort, blurred vision and convulsive syncope. Significant and specific predictors of a neurally mediated cause were time between the first and last syncopal episode >4 years, abdominal discomfort before the loss of consciousness and nausea and diaphoresis during the recovery phase. In the patients without heart disease, palpitation was the only significant predictor of a cardiac cause.
The presence of suspected or certain heart disease after the initial evaluation is a strong predictor of a cardiac cause of syncope. A few historical findings are useful to predict cardiac and neurally mediated syncope in patients with and without heart disease.
It is commonly accepted that history is an essential part of the work-up of patients with syncope. The historical findings, per se, may be diagnostic of the cause of syncope or may suggest a strategy of evaluation. Several clinical features regarding the modalities of presentation, the factors that might predispose to the syncopal attack and the sequelae of syncope have been utilized. However, their diagnostic value is generally accepted on the basis of common sense, personal experience, pathophysiologic knowledge and anecdotal reports. This is the case of symptoms such as nausea, diaphoresis, pallor, abdominal discomfort and yawning, which are commonly considered characteristic of neurally mediated syncope.
The aim of the present study was to establish what historical findings are predictive of the cause of syncope.
Patients were recruited if they had a syncopal episode in the previous two months (defined as a brief, self-limited loss of consciousness with the inability to maintain postural tone) and were ≥18 years. Three hundred fifty-six consecutive patients with the aforementioned characteristics were referred to the “Syncope Unit” of the Cardiology Division of three hospitals (Cento, Lavagna and Reggio Emilia, Italy) from the emergency room, in-patient service and out-patient clinic between January and July 1999. All patients underwent a standardized initial evaluation consisting of a history, physical and neurologic examinations performed by one of the investigators, evaluation of blood pressure in the supine and standing positions and standard electrocardiogram (ECG). Conventional laboratory tests were performed when a hemorrhagic or metabolic cause was suspected. When the results of the initial evaluation allowed us to define the cause of syncope (i.e., diagnostic [see subsequent text]), the procedures were interrupted. When the results of the initial evaluation were inconclusive, further investigations were performed according to the guidelines of the Italian “Associazione Nazionale Medici Cardiologi Ospedalieri” (1,2)(Fig. 1), with the addition of the adenosine triphosphate (ATP) test when the cause of syncope remained unexplained at the end of the diagnostic work-up. For example, patients with certain or suspected heart disease first underwent noninvasive cardiac examinations, as appropriate, and, in the event of negative results, an electrophysiologic study. Patients without signs of heart disease first underwent carotid sinus massage and the tilt test. In the event of focal neurologic symptoms before or after the loss of consciousness, witnessed prolonged loss of consciousness with tonic-clonic movements or a post-ictal state, migraine or signs of psychiatric disorders, a consultant neurologist or psychiatrist was called in. At his or her discretion, neurologic examinations were performed.
We classified the causes of syncopes, as reported in Table 1. Diagnostic criteria for the causes of syncope were developed before enrollment, on the basis of an extensive review of the pertinent published data, and assignment of a cause was based on strict adherence to these criteria. The diagnosis was assigned by one of the investigators at the end of the diagnostic procedures. Presumed diagnoses made by the house staff or attending physician were not accepted. Syncope was defined as orthostaticin the presence of a decrease in systolic blood pressure ≥20 mm Hg during 5 min after the patient assumed the upright position, associated with dizziness or syncope (3,4); typical vasovagalif a precipitating event, such as fear, severe pain, strong emotion or instrumentation, could be identified in the absence of another competing diagnosis for the cause of syncope (3); situational syncopewhen the loss of consciousness occurred during or immediately after urination, defecation, coughing or swallowing (3); tilt-induced syncopewhen the loss of consciousness was induced during the tilt test, as previously described (5); carotid sinus syncopewhen carotid sinus massage, performed both in the supine and upright positions, induced syncope in the presence of bradycardia or hypotension, or both (6); drug-induced syncopewhen a clear temporal relationship between drug assumption and syncope could be proven; adenosine-sensitive syncopeif a bolus of 20 mg of ATP induced a cardiac pause ≥6 s (7); and mechanical cardiac syncopein the presence of severe valvular stenosis or other flow obstruction. Arrhythmic syncopewas diagnosed on the basis of standard ECG or 24-h Holter recordings or during the electrophysiologic study. The following findings of standard ECG were considered diagnostic: sinus bradycardia <40 beats/min or repetitive sinoatrial blocks or sinus pauses >3 s; Mobitz II or advanced (2:1, 3:1, etc.) second-degree atrioventricular (AV) block or third-degree AV block; atrial fibrillation with a slow ventricular response (mean heart rate <50 beats/min); and sustained supraventricular or ventricular tachyarrhythmias. The following findings of 24-h Holter recording were considered diagnostic: any arrhythmia associated with syncope or presyncope, or, in the absence of neurologic symptoms, Mobitz II or advanced degree or third-degree AV block; and sustained ventricular tachyarrhythmias. The following electrophysiologic findings were considered diagnostic: sinus node recovery time >3 s; baseline HV interval ≥100 ms or appearance of infra-hisian second- or third-degree AV block during atrial pacing or after ajmaline administration; and induction of syncopal or hypotensive supraventricular or ventricular tachyarrhythmias. Metabolic syncope due to hypoglycemiawas diagnosed when the loss of consciousness was preceded by tremors, confusion, salivation, hunger and a hyperadrenergic state, and glycemia was <40 mg/dl.
The clinical findings of syncope were investigated, taking into account the last syncopal episode, by one of the authors, who used a standard questionnaire. The questionnaire was designed to identify 46 findings of the index syncopal episode. Data were collected on precipitating and predisposing factors; prodromal and recovery symptoms; the duration of the prodromal symptoms, loss of consciousness and recovery phase; and physical signs. The duration of the loss of consciousness and physical signs was investigated only if they were observed by witnesses, who were interviewed. The variables investigated are reported in Table 2.
After the initial evaluation, the presence of heart disease was suspected or diagnosed on the basis of history, physical examination or ECG abnormalities, or a combination of these. Supraventricular or ventricular tachycardia was defined as sustained (lasting >30 s) if present on the standard ECG recorded in the emergency room or 24-h Holter recording. The ECG was considered abnormal, besides the alterations reported earlier (see Methods), in the presence of signs of atrial or ventricular enlargement, intraventricular conduction disturbances, ventricular pre-excitation, old myocardial infarction, isolated sinoatrial block and in the presence of first- or second-degree Wenckebach AV block, ST segment and T-waves abnormalities and rhythm abnormalities, such as atrial fibrillation or flutter, atrial tachycardia, junctional or paced rhythm, and frequent or repetitive premature ventricular beats.
Comparisons between groups were performed by means of the Fisher exact test, ttest or rank-sum test, as appropriate. The logistic regression test was used to find those variables that were independent predictors of the cause of syncope. The data were analyzed with the StatSoft software (Version 5.1, 1997, Tulsa, Oklahoma).
During the recruitment period, 356 patients were enrolled (Fig. 2). Of these, 15 (4%) were excluded due to an incomplete evaluation or protocol violation. Therefore, 341 patients were analyzed. The mean age of the patients was 61 ± 20 years; 184 were men.
A cardiac cause of syncope was established in 78 patients (23%), a neurally mediated cause in 199 (58%) and a neurologic or psychiatric cause in 4 (1%), and in the remaining 60 (18%) the cause of syncope remained unexplained (Table 1). The diagnostic tests that established the cause of syncope are reported in Table 3. Because only four patients had a neurologic or psychiatric cause, this group was not included in the statistical analysis. Therefore, 337 patients were evaluated.
Of 51 patients with bradyarrhythmia, AV block was present in 40, sinus node dysfunction in 6 and atrial fibrillation with a slow ventricular response in 5. Of 17 patients with tachyarrhythmia, supraventricular tachycardia was present in 4 and ventricular tachycardia in 13. Of the 78 patients with a cardiac cause of syncope, the diagnosis was made during the initial evaluation in 36 (46%); in all of these, the diagnostic test was the standard ECG. Of the 199 patients with a neurally mediated cause of syncope, the diagnosis was made during the initial evaluation in 51 (26%): typical vasovagal or situational syncope in 45 and orthostatic hypotension in 6.
Comparison of clinical features during syncope
The clinical features of the patients with cardiac, neurally mediated and unexplained causes of syncope are reported in Table 2. There were some significant differences between cardiac and neurally mediated syncope. In contrast, the clinical features of unexplained syncope were very similar to those of neurally mediated syncope; indeed, among the variables examined, only male gender, time between the first and last syncopal episode and tremors showed significant differences between the two types of syncope.
We first performed a statistical analysis of the variables that were not part of the presentation of syncope: age, gender and the presence of suspected or certain heart disease after the initial evaluation. On univariate analysis, these were found to differ. Among these three variables, on multivariate analysis, the presence of suspected or certain heart disease was the only independent predictor of a cardiac cause of syncope (odds ratio 16, 95% confidence interval 5 to 48, p = 0.00001). Consequently, the patients were classified into two subgroups: one comprising the 191 patients with heart disease and the other comprising the 146 patients without it. We analyzed the historical variables of syncope separately in these two subgroups. The results are reported in Tables 4 and 5. ⇓⇓Suspected or certain heart disease after the initial evaluation was present in 74 (95%) of 78 patients with a final diagnosis of cardiac syncope, but also in 117 (45%) of 259 of patients who did not have a cardiac cause of syncope (Fig. 2). In contrast, the absence of suspected or certain heart disease allowed us to exclude a cardiac cause of syncope in 142 (97%) of these 146 patients (Fig. 2).
Patients with suspected or certain heart disease
In the 191 patients with suspected or certain heart disease after the initial evaluation, the significant univariate predictors of a cardiac cause of syncope were time between the first and last syncopal episode ≤4 years, ≤2 syncopal episodes, syncope during effort or in the supine position, blurred vision before the loss of consciousness and convulsive syncope. On multivariate analysis, duration of symptoms ≤4 years, syncope in the supine position and blurred vision remained independent predictors (Table 4).
Significant univariate predictors of a neurally mediated cause of syncope were time between the first and last syncopal episode >4 years, history of presyncope, abdominal discomfort and pallor before the loss of consciousness and nausea and diaphoresis during the recovery phase. On multivariate analysis, duration of symptoms >4 years, history of presyncope and nausea remained independent predictors (Table 4).
Patients without suspected or diagnosed heart disease
In the 146 patients without suspected or diagnosed heart disease after the initial evaluation, the only significant univariate predictor of a cardiac cause of syncope was palpitations before the loss of consciousness. The significant univariate predictors of a neurally mediated cause of syncope were ≥3 syncopal episodes and duration of prodromes >10 s. On multivariate analysis, duration of prodromes >10 s remained an independent predictor (Table 5).
The presence of suspected or certain heart disease after the initial evaluation is a strong predictor of a cardiac cause of syncope, but its specificity is low. In a minority of patients with suspected or certain heart disease, there are historical findings predictive of a cardiac cause or a neurally mediated cause of syncope. The absence of suspected or diagnosed heart disease allows us to exclude a cardiac cause of syncope if the patient does not have palpitations before the loss of consciousness. The clinical features of unexplained syncope and neurally mediated syncope are very similar. Our results do not support the usefulness of the differential diagnosis of some clinical variables commonly used in clinical practice and previously reported.
Comparison with previous studies
Two previous prospective studies have evaluated the clinical features during syncope (9,10). Martin et al. (9)obtained the clinical histories of 170 patients presenting with syncope on arrival at the emergency room. These authors reported that the duration of prodromal symptoms was longer for neurally mediated syncope than for cardiac syncope. We could not confirm this finding. However, in this study, a presumed diagnosis was established only at the time of the emergency room evaluation. Calkins et al. (10)evaluated the clinical histories in three selected groups of patients with neurally mediated syncope, syncope due to ventricular tachycardia and syncope due to AV block. The clinical histories of patients with syncope due to ventricular tachycardia or AV block were similar. In contrast, the clinical histories of patients with neurally mediated syncope differed from those of patients with cardiac syncope. We confirmed some features suggestive of neurally mediated syncope, such as diaphoresis before and after the loss of consciousness and nausea after the loss of consciousness. Other features suggestive of neurally mediated syncope, such as longer duration of prodromes, palpitations, nausea and lightheadedness before syncope and fatigue after the loss of consciousness, were not confirmed. In the study of Calkins et al. (10), patients were asked to give a description of their two most recent syncopal episodes. For this reason, the prevalence of symptoms was very likely overestimated, and the description could not be precise, as the time between two syncopal episodes is often long. Moreover, it is not clear how the duration of prodromes was evaluated. It must also be considered that the results will be different according to whether a study is carried out in a selected group of patients or in an unselected group (consecutive patients).
Results of the present study
On the basis of specific diagnostic criteria, the cause of syncope was established in 281 patients (82%) and remained unexplained in 60 (18%); the prevalence in the latter group was lower than that reported in five previous studies performed in the 1980s (mean 34%) (9,11–14), probably because of a more extensive utilization of carotid sinus massage, tilt testing and electrophysiology.
The clinical features of unexplained syncope and neurally mediated syncope were very similar (Table 2). This suggests that unexplained syncope is, at least in the vast majority of patients, a neurally mediated syncope that is difficult to diagnose with currently utilized tools.
The absence of prodromal symptoms was similar between cardiac and neurally mediated syncope (41% and 33%, respectively); therefore, this factor is not useful for diagnosing the cause of syncope. The duration of the loss of consciousness and the duration of the recovery phase were also similar between these two types of syncope.
Symptoms and signs, such as nausea, vomiting before the loss of consciousness, pallor, weakness, lightheadedness and yawning, which have been used up to now in clinical practice and considered very suggestive of neurally mediated syncope, even in the absence of a clinical investigation, show a similar prevalence in cardiac and neurally mediated syncope (Table 2)and, therefore, are not indicative of the type of syncope. These results suggest that activation of the autonomic system occur in all types of syncope. In addition, the prevalence of incontinence and injury was similar in all types of syncope.
In syncope, there is no diagnostic “gold” standard against which other diagnostic tests may be measured; therefore, no set of criteria and no algorithm for diagnosing syncope may be considered ideal. The diagnostic criteria we employed in the present study are those most commonly used; they were also officially accepted by the Italian Cardiological Association (1,2).
The prevalence of neurologic or psychiatric loss of consciousness was low (1%). It is possible that this result does not express the true prevalence of this type of loss of consciousness, because it is presumable that patients with seizure disorders or with loss of consciousness associated with focal neurologic symptoms went directly, or through the emergency room, to the neurology department. A similar low prevalence was reported in previous studies (9,12,15,16). The prevalence of typical vasovagal syncope may also be underestimated, as a population study showed that only 23% of patients with this type of syncope are referred to the hospital (17). For these reasons, the results of the present study seem to reflect the distribution of patients who are referred to the syncope unit, rather than the distribution of patients with syncope in the general population.
Abnormal neural reflexes have been shown to play an important role in the genesis of syncope in patients with sinus node dysfunction or with paroxysmal supraventricular tachyarrhythmias (18–21). Therefore, the diagnosis of arrhythmic syncope may constitute a simplification.
In clinical practice, patients with syncope undergo several examinations, very often with negative results. The results of the present study suggest a practical strategy for evaluation of patients with syncope. We have demonstrated that the presence of suspected or certain heart disease after the initial evaluation is a predictor of a cardiac cause of syncope, as already reported by Martin et al. (22), although its specificity is low.
Thus, the presence or absence of heart disease should be investigated first. In the presence of heart disease, some variables (e.g., blurred vision, syncope occurring in the supine position or during effort or convulsive syncope) suggest a cardiac cause of syncope, with a high specificity. The presence of these variables increases the pre-test probability of a cardiac cause of syncope. For example, in a patient with suspected or certain heart disease, the diagnosis of a cardiac cause of syncope is increased from a pre-test probability of 39% (Fig. 2)to a post-test probability of 79% when syncope occurs in the supine position (Table 6). Therefore, these patients should probably be admitted to the hospital, and cardiac examinations should be given priority. In contrast, a long duration (>4 years) of symptoms and syncope preceded by abdominal discomfort or followed by nausea and diaphoresis suggest a neurally mediated cause of syncope, with a high specificity and an increased post-test probability. In these cases, autonomic tests should be given priority, and, if they are positive, a neurally mediated mechanism is likely, and the diagnostic work-up should be stopped. However, it must be emphasized that the sensitivity of the most specific predictors is quite low (Table 6), and, therefore, they can be useful only in a minority of patients.
When heart disease is absent, a cardiac cause is unlikely, unless palpitations precede syncope. In the latter case, the possibility that a tachyarrhythmia is the cause of syncope should be evaluated. In the other cases, autonomic tests should be performed to evaluate the possible neurally mediated nature of syncope. Patients usually do not need to be admitted to the hospital, given the benign outcome of this type of syncope.
The results of the present study can optimize the work-up of syncope, avoiding useless tests and reducing the high diagnostic costs in patients with syncope (23).
- adenosine triphosphate
- electrocardiogram or electrocardiographic
- Received September 19, 2000.
- Revision received January 24, 2001.
- Accepted February 12, 2001.
- American College of Cardiology
- Alboni P.,
- Raviele A.,
- Vecchio C.,
- et al.
- Brignole M.,
- Gaggioli G.,
- Menozzi C.,
- et al.
- Sackett D.L.,
- Richardson W.S.,
- Rosenberg W.,
- Hayness R.B.
- Kapoor W.N.
- Kapoor W.N.,
- Snustad D.,
- Peterson J.,
- Wleand H.S.,
- Cha R.,
- Karpf M.
- Cosin Aquilar J.,
- Solez Minguez J.,
- Garcia Rivera I.,
- Ruiz Granel R.
- Leitch J.W.,
- Klein G.J.,
- Yee R.,
- Leather R.A.,
- Kim Y.H.
- Alboni P.,
- Menozzi C.,
- Brignole M.,
- et al.
- Brignole M.,
- Gianfranchi L.,
- Menozzi C.,
- et al.
- Nyman J.A.,
- Krahn A.D.,
- Bland P.C.,
- Griffiths S.,
- Monda V.