Owing to the gravitational effects on blood, reflex syncope, particularly VVS, occurs most often in the standing position. Cardiac syncope is primarily caused by arrhythmia and structural cardiac disease, with arrhythmia being the most common cause in children [16]. Arrhythmia generally occurs regardless of posture [17]. Furthermore, syncope can occur irrespective of posture if the circulation stops [17]. Therefore, syncope in the supine position is a strong warning sign of cardiac syncope, while increased orthostatic stress is predictive of reflex syncope, namely VVS [18].

However, orthostatic stress increases while standing up and sitting down. Sitting intolerance occurred in approximately 10% of children with sitting-induced hemodynamic changes [19]. A study of 111 adult patients with syncope, excluding those of cardiac origin, reported that syncope developed in standing (n=67), sitting (n=16), standing and sitting (n=23), sitting and supine (n=1), and all three positions (n=4) [20]. That is, 40% and 4.5% of patients reported syncope in the sitting and supine positions, respectively [20]. Prodromal symptoms and circumstances were similar irrespective of position; however, emotional triggers were commonly observed in patients with syncope in the sitting and supine positions [20]. This study showed that reflex syncope, although rare, may occur in sitting and supine positions. If syncope is suspected in body positions other than standing, a specific situation and activity (swallowing, coughing, or urination) or emotional trigger (pain or fear) should be carefully investigated. However, this study emphasized that no patients had syncope in the supine position alone [20]. Furthermore, no VVS occurred in the supine position in 268 syncopal children [21]. Reflex syncope is possible in the supine position; however, cardiac syncope should be considered initially.

Activity immediately before an attack is important to the differential diagnosis of syncope. However, the association between exercise and syncope remains unclear. In most cases, exerciserelated syncope is not an indicator of underlying severe cardiac disease [22]. However, midexertional syncope (i.e., before the child can stop exercising) suggests a cardiac etiology [22]. An epidemiological study of 7,568 young athletes a mean age of 16 years of age showed that exercise-related syncope occurs infrequently [22]. Among athletes diagnosed with syncope (n=474), postexertional syncope was observed in 57 athletes (12%), while exertional syncope was present in only 6 (1.3%) [22]. All episodes of postexertional syncope had the typical features of VVS, whereas 2 midexertional syncope episodes had cardiac etiologies [22]. A study of 60 midexertional syncopes in children demonstrated that 32 (53%) had a cardiac origin [23]. Neither exertional activity type nor all symptoms occurring during syncopal events (before, immediately preceding, and after the syncopal event) could be differentiated based on the presence or absence of a cardiac diagnosis [23]. These studies suggest that midexertional syncope is a rare but crucial historical aspect of cardiac syncope, even at a young age.

Postexercise syncopal events typically occur when an individual suddenly stops exercising and stands motionless for the first 5–10 minutes after exercise [24]. Postexercise syncope shares the same mechanism as VVS. Venous return is dramatically decreased due to reduced muscle pumping of the lower extremities, while cardiac contractility remains high, and blood vessel dilation occurs during syncope [24]. That is, the Bezold-Jarisch reflex may involve postexercise syncope.

On the other hand, electrical myopathy and mechanical outflow obstruction are the primary underlying etiologies of cardiac syncope in children [23]. In both cases, a decrease in cardiac output was responsible for the syncope. A brief review of the physiology of syncope during exercise in patients diagnosed with long QT syndrome (LQTS) is warranted, as it is the most common cause of cardiac syncope in children [23]. Decreased function of the delayed rectifier potassium current, either rapid (IKr) or slow (IKs), normally controls cardiac repolarization and causes LQTS [25]. Adrenergic stimulation causes quantitatively similar enhancement of both L type calcium current, an inward current tending to a prolonged cardiac action potential, and IKs, shortening the action potential [25]. The balance between them allows for cardiac contractility enhancement during exercise without excessive action potential prolongation [25]. When IKs is dysfunctional, as in LQTS 1, sustained increases in adrenergic tone develop intense action potential prolongation and subsequently predispose patients to lethal ventricular arrhythmia [25].

Prodromal symptoms are caused by 2 main mechanisms. The first is symptoms related to “autonomic activation” of sympathetic and parasympathetic activity, while the second consists of the consequences of retinal and cerebral hypoperfusion [26]. Symptoms of autonomic activation include sweating, facial pallor, nausea, pupillary dilatation, and palpitations [26]. The symptoms of hypoperfusion include lightheadedness, unclear thinking, blurred or darkened vision, and tinnitus [26]. A prodromal sensation of vision is caused by retinal hypoperfusion [27]. Unlike the brain and brainstem, the eye is not protected by the pressure-equalizing effects of cerebrospinal fluid [27]. Because retinal hypoperfusion precedes brain hypoperfusion, visual prodromes (i.e., blackening out) can occur before a TLOC [27].

Prodrome is an essential symptom of syncope, providing a clue in differentiating it from other TLOC, particularly seizures [28]. However, clinicians must recognize that prodromal symptoms can be reported only when the fall in BP is detectable and the patient’s sensations are listed for later recall [26]. Therefore, prodromal symptoms and signs are minimal or absent in abruptonset syncope caused by provoked (i.e., fainting lark) or cardiac standstill [26]. Consequently, sudden onset and lack of prodromal symptoms provide clinical clues for cardiac causes, typically like a stroke-Adams attack (i.e., complete atrioventricular block) [26]. Meanwhile, the clinical features of syncope in ventricular tachycardia differ from those in a typical stroke-Adams attack, possibly depending on differences in the BP fall rate. Loss of consciousness (LOC) usually occurs later after the onset of ventricular tachyarrhythmias (20–120 seconds)29) than after asystole (4–15 seconds) [30].

Prodrome is more commonly reported in children with VVS than in those with cardiac syncope [31]. Zhang et al. [31] reported that 48 of 55 children (87.3%) with VVS had overall prodromal symptoms versus 16 of 31 (51.6%) with cardiac syncope. These symptoms include dizziness, headaches, chest discomfort, palpitations, sweating, paleness, nausea/vomiting, blurred vision, and fatigue [31]. Among these, dizziness, headache, and chest discomfort were significantly more common in the VVS group. Meanwhile, other symptoms, including palpitations and blurred vision, did not differ between groups [31]. On multivariate analysis, a history of abnormal electrocardiogram findings and midexertional syncope remained independent predictors of cardiac syncope but not the presence of prodromal symptoms [31]. Hurst et al. [32] analyzed 3,445 pediatric patients who visited the emergency department with syncope and reported that only 3 had a previously undiagnosed cardiac cause. They determined the sensitivity and specificity of historical features to identify cardiac syncope [32]. Among the three children with cardiac syncope, 1 did not have prodromal symptoms [32]. Thus, the sensitivity and specificity of syncope without prodrome in identifying cardiac syncope were 67% and 70%, respectively [31]. Unlike Zhang et al. [31], Hurst et al. [32] analyzed palpitations as an independent clinical feature rather than a prodrome. They reported that the sensitivity and specificity of palpitations were 100% and 98%, respectively [32]. These inconsistent results suggest that the presence or absence of prodromal symptoms alone should not be used to discriminate between cardiac and noncardiac syncope. Importantly, approximately 50% of children diagnosed with cardiac syncope also report prodromal symptoms [31].

Symptoms of VVS have the following temporal sequence: the first symptoms are feeling uncomfortable, epigastric discomfort, nausea, and abdominal cramps [26]. If patients ignore these warning signs, signs and symptoms of cerebral and retinal dysfunction develop [26]. The first sign of impending vasovagal fainting is often facial pallor, which results from reduced skin blood flow due to sympathetic and vasopressin-induced vasoconstriction [26]. During this phase, palpitations are a common complaint due to an adrenergic surge preceding the vagal response. Symptoms of palpitations often present as cardiac syncope and VVS [32]. Meanwhile, Hurst et al. [32] reported that the presence of at least 2 of the following features yielded 100% sensitivity and specificity for cardiac syncope in children: absence of prodromal symptoms, a midexertional event, and chest pain or palpitations.

Interviewing individuals with witnessed syncope may be helpful. Details including fall type, LOC duration, movements during attack, and tongue biting should be asked whenever possible.

Fall type provides valuable information for the differential diagnosis of epilepsy and syncope [33]. As a rule, during an epileptic attack, a fall to the ground is tonic, whereas during syncope, it is generally flaccid [34]. Flaccidity during LOC contradicts seizures. The only exception is “atonic seizure,” [34] which occur rarely, primarily in children diagnosed with epileptic encephalopathy [35]. Therefore, if a child without neurological impairment collapses with flaccidity (kneeling over), it is reasonable to diagnose syncope rather than seizures. However, due to this common knowledge, syncope is mistaken for seizure. Stiff falls during syncope have been commonly described in laboratory-observed syncope [36]. Lempert et al. [36] reported that only half of healthy volunteers collapsed flaccidly during fainting, while the others fell with their knees and hips extended. Clinicians must understand that a stiff fall does not necessarily herald a generalized tonic-clonic seizure [37]. A stiff fall is part of the continuum of tonic posturing and shares the same mechanism as tonic posturing, which is described in the next section.

Convulsive movements, including tonic or myoclonic muscle activity, eye deviations, automatisms, and vocalizations, often accompany syncope. The reported frequencies of convulsive movements in syncope vary widely (5%–90%) depending on study design [36,38]. A video study of adolescents documented myoclonus in 91%, facial grimacing in 78%, vocalization in 65%, and head deviation in 34% of the patients during tilting [39]. Owing to these convulsive features, syncope is misdiagnosed as epilepsy. Approximately 20% of patients referred to epilepsy clinics suffer from recurrent syncope [40]. Furthermore, 35% of patients diagnosed with reflex syncope are treated with antiepileptic drugs [41]. In recent studies of children who were referred to new-onset seizure clinics, nonepileptic events were reported in 24%–45% of cases [42,43], and syncope was the most or second most common cause. Convulsions are an integral component of the cerebral response to hypoxia. Therefore, the presence of convulsive movements depends on brain hypoperfusion degree [44]. Specific features should be based on distinguishing convulsive syncope from seizures, not by the mere presence or absence of motor phenomena [44].

The most common movement in syncope is myoclonic activity [36]. Syncopal myoclonus is often multifocal with asynchronous muscle jerks and occasionally generalized with bilateral synchronous muscle activation [36]. However, unlike epileptic myoclonus, syncopal myoclonus has arrhythmic features and is sustained for several seconds [36,45]. A study that analyzed video electroencephalography (EEG) records of 65 cases of syncope and 50 cases of convulsive seizures reported myoclonic jerks in 33 (51%) and 50 cases (100%), respectively [45]. This study clearly showed different intergroup features of myoclonic jerks: the synchrony rate and jerk frequency were lower in the syncope versus seizure group [45]. More than 20 jerks were noted in all convulsive group patients [45]. Meanwhile, all but one patient in the syncope group exhibited fewer than 10 jerks [45]. These findings suggest that fewer than 10 jerks indicates syncope and more than 20 jerks indicates convulsive seizures [45].

Lateral head turns with an ipsilateral gaze have been reported; however, they never reached the forced tonic quality of epileptic head turns [36]. Syncope may also produce complex motor activity resembling epileptic automatisms, including lip-smacking, chewing, fumbling, reaching the head, turning or raising the head, and sitting up [36,39,46]. In contrast to epileptic automatisms, these movements are primarily brief and solitary rather than repetitive [44]. This convulsive activity in syncope occurs after rather than before the loss of postural tone, which may provide an essential clue for differentiating syncope from seizures [36]. As a rule, eyes are open during syncope, a feature shared by seizures but not psychogenic TLOC [47]. The most consistent ocular sign is upward turning of the eyes early in syncope, after which lateral deviation may follow [47]. In contrast to sustained or forceful epileptic eye deviation, syncopal eye turns are typically brief [45].

Studies measuring EEG changes during syncope have demonstrated the nature of convulsive movements in syncope. Slow background activity appears first [48], followed by high-amplitude delta activity associated with LOC, eye-opening, myoclonic jerks, and general stiffening [49,50]. If hypoperfusion persists, EEG flattening is accompanied by tonic posturing, vocalization, slow horizontal eye movement, and stertorous breathing [50]. Tonic posturing is “probably universal” when syncope lasts long enough to produce a flat EEG [46]. A breath-holding spell is a good example, often accompanied by intense opisthotonic stiffening caused by profound but brief cerebral hypoxia [46,51]. The “slow-flat-slow” pattern sequence is independent of syncopal mechanisms. However, mild forms of syncope may present as isolated background slowing without EEG flattening. Myoclonic jerks predominantly occur during the slow EEG phase, whereas tonic posturing occurs during the flat phase [45]. These EEG correlates suggest that jerks in syncope are likely of cortical origin, whereas tonic posturing results from brainstem disinhibition due to complete jerks [45]. Therefore, the degree of cerebral hypoperfusion during syncope determines whether convulsive activity will occur [37].

Prospective clinical observations and laboratory-induced syncope studies reported that LOC usually lasts for approximately <30 seconds [36,52]. Although eyewitnesses often report long periods of unconsciousness, the reliability of such estimates of syncope duration is debatable [26]. Body position may play a role in the duration of LOC in syncope. If a person collapses and falls to the floor during syncope, the LOC period is brief because cerebral perfusion causes immediate recovery. By contrast, cerebral perfusion may remain inadequate for longer durations in those who remain upright or slump during syncope [26].

A meta-analysis showed that tongue biting has excellent value in the differential diagnosis of syncope and seizures [53]. Tongue biting increases the probability of epileptic seizures [53]. In a prospective study of an epilepsy monitoring unit, 8 of 34 patients with epileptic seizures displayed tongue biting on the lateral side [54]. Moreover, only 1 of the 45 patients with syncope had tongue lacerations at the tip [54]. This study suggested an extremely high (99%) specificity for tongue biting in general and 100% specificity when the injury was due to seizures versus syncope [54].

Urinary incontinence may occur during both syncope and epileptic seizures, and meta-analysis demonstrated that urinary incontinence has no value in the differential diagnosis between them [55]. Incontinence rate variations have been reported to be up to 25% of syncopal events [52]. Urinary incontinence was twice as common in patients with syncope versus seizures [56]; thus, it is more likely associated with syncope than seizures [57]. However, these findings were derived from adult studies. A pediatric study revealed that urinary incontinence occurred in approximately 56% of cardiac syncope cases versus 2.6% of noncardiac syncope cases [58]. The urinary sphincter releases tone upon asystole lasting longer than 20 seconds [30]. Unlike adults, cardiac syncope is rare in children. Therefore, further studies are needed to estimate the value of incontinence in pediatric age groups to differentiate syncope from seizures.

Reorientation is usually immediate in syncope and does not exceed 30 seconds even after an extended attack [29]. Any postictal disorientation lasting longer than 30 seconds suggests an epileptic seizure [28,37]. The interval from the LOC start back to the oriented state is only approximately 20–30 seconds in abruptonset laboratory-induced syncope [36,46]. Immediate complete reorientation is also observed in cardiac syncope if the cardiac standstill is brief [30]. Additional importance in the clinical picture of VVS is the postsyncopal symptoms characterized by persistent pallor, nausea, weakness, and sweating. Overactive autonomic nerve function takes time to return to baseline after a severe VVS attack. Thus, the syncope recurs if an individual reassumes an upright posture too soon [59]. Table 2 summarizes the pearls and pitfalls in history taking for the differential diagnosis of syncope.