Timothy C. Hain,
MD Page last modified:
November 19, 2016
See also dysautonomia orthostatic hypotension Tilt table testing Autoimmune autonomic neuropathy
Causes Evaluation Prognosis
Syncope is often discussed by cardiologists or epilepsy specialists. These groups tend to look at syncope from their special perspective. This page is written from a more general perspective -- that of a "dizzy doctor". As a preview of the "take home" messages, the very definition of "syncope" and variants is poorly thought through, the syncope literature is wildly variable, there is an overemphasis on heart conditions, and "sorting" is the huge problem with syncope.
|Incidence of syncope per 1000 person years by age from Sotieriades et al (2002).|
Syncope (faint) is often considered as a subset of the symptom of transient loss of consciousness (or in other words, passing out). Using this definition syncope is the most common -- accounting for about 80% -- of the 3 conditions that cause transient loss of consciousness. The other two, accounting for another 10%, include epilepsy and pseudoseizure. Obviously then, epilepsy/pseudoseizure are minor contributors.
The most common definition of syncope is a sudden fall of blood pressure resulting in loss of consciousness. There are three syncope subvariants that differ mainly in their prognosis and need for follow-up. As this is not generally known when the syncope occurs, ths categorization process doesn't have much pragmatic use, at least urgently.
- "vasovagal" syncope, also known as "simple faint", or "neural" syncope, is a loss of consciousness, associated with lower blood pressure, triggered by an emotional event.
- "vasodepressor syncope" is syncope mediated by vasodilation.
- "cardioinhibitory syncope" is syncope mediated by slowing of the heart, or reduced contraction of the heart.
A narrower definition of syncope is that of Jones and Gibbons (2014) who stated that syncope is a "transient loss of consciousness, associated with cerebral hypoperfusion, no lasting neurologic deficits and complete recovery". This definition does not require a fall in systemic blood pressure -- one might envision instead syncope due to choking and pressure on both carotids. It might also allow for "syncope" due to occlusion of a single carotid, which we think is highly improbable. It excludes entities like stroke. Practically, this definition is more difficult to use "in the field", as rarely does one know how well the brain is perfused. We think it is better to stick with broader definitions.
We and some others (Sotieriades et al, 2002) think that syncope is better defined as simply a transient loss of consciousness. This does not greatly change the statistics, as most transient loss of consciousness (80%) is due to lowered blood pressure, but it avoids the confounding of the very large "unknown" category between individuals thought to have lowered blood pressure or not. In the following, we will qualify our statistics according to the definition.
Presyncope consists of unsteadiness, dizziness, weakness, or cognitive symptoms without loss of consciousness. Dizziness is frequently reported during presyncope, such as from a racing heart (Walfridsson and Walfridsson, 2005) Presyncope is often a symptom of orthostatic hypotension. We think that "presyncope" is a unfortunate term in its usual usage where the "syncope" is missing. How can one say something is "pre", when the event never occurs ? This sloppy use of language leads to problems with sorting patients into diagnostic categories.
Pseudosyncope is an apparent loss of consciousness unaccompanied by circulatory changes such as reduced blood pressure. The term pseudosyncope is also an unfortunate one, as it is common to also use the prefix "pseudo" in front of another potential cause of loss of consciousness, such as a seizure, that is not "pseudo" at all. Thus a seizure (to a neurologist) would meet the definition of "pseudosyncope" to an internist.
Things that are not syncope:
Disorders that might look like syncope but which are not because there is no loss of consciousness include:
Some authors also choose to excludefrom syncope loss of consciousness due to hypoxia or hypoglycemia (e.g. Moya et al, 2009) . We think this is inappropriate splitting.
- Cataplexy (oddly, this schizophrenic behavior has lost popularity in recent years)
- Drop attacks
- Falls -- i.e someone who trips over a dog is not "syncope".
- Functional (pseudosyncope or pseudoseizures) -- see commentary above about the unfortunate use of these terms.
The incidence of syncope is estimated to be about 18.1-39.7 events/1000 patient years (Ganzeboom et al, 2006). In other words, 2-4% of the population every year. Somewhere between 3%-35% (presumably the larger #) of the population experience syncope at some point in life (Savage, 1985; Ganzeboom et al, 2006).
Of these 2-4% of the population, only about 1% end up seeing their general practitioner, and only about 0.1% end up in the emergency department (Moya, 2009). It seems reasonable to assume that syncope in the ED has more of the "dangerous" faints -- who would go to the ER if they fainted at the sight of their own blood after all ? Thus the statistics about syncope from the ED, likely overestimate the danger of syncope in general.
That being said, syncope also accounts for roughly 1-3% of all emergency room visits (Kapoor, 1990; Soteriades et al, 2002; Moya et al, 2009) and 6% of all hospital visits. Between 71 and 125 children and adolescents/100,000 population experience syncope each year. The number of visits to Emergency Departments for syncope rises steadily with age (Sotieriades et al, 2002).
Hospitalization accounts for more than 75% of the health care costs of syncope (Moya et al, 2009). In the US, the mean cost was $5400/hospitalization for syncope in 2009.
Risk stratification is used for persons with syncope of unknown origin. Most risk stratification algorithms depend largely on an abnormal ECG as an effort is made to detect lethal arrhythmias or individuals with structural heart disease.
Baron-Esquivias et al (2010) reported the proportion of patients with various types of syncope in Emergency Departments in Spain. They defined syncope as as loss of consciousness with lowered blood pressure. This is shown in the figure on the right. There were 1217 patients with syncope as defined, as well as another 202 patients with "Non Syncope Transient-LOC". Putting this together, about 80% of patients with transient LOC were categorized as "Syncope", and about 20% as just "transient LOC". As both groups included very large "unknown" categories, we ourselves think it would have been better to simply consider them as together as others have done. (Sotieriades et al, 2002). Nevertheless, Baron-Esquivias did provide enough details that one can analyze them either way.
The biggest single cause of syncope in the ED as reported by Baron-Esquivez et al (2010) is "neural syncope. Other studies suggest that about 35-48% of syncope in the ED is "neural" (Moya, 2009).
Again, remember that in the study providing this data, other neurological disorders such as stroke or brain cancer (for example) were excluded from the "neural" group, so this group are actually "vasovagal" type faints. "Neural" is an unfortunate term as "neural" might also be interpreted as the loss of consciousness due to a brain illness (which were excluded). Also known as "faint", or "vasovagal syncope", the "neural" category is defined as a loss of consciousness due to an emotional event -- for example -- seeing one's own blood. About 44% of all syncope from this large study was "neural". Other studies have similar findings.
Clinical features suggestive of "neural" syncope (Brignole et al, 2004) Absence of cardiac disease Long history of syncope Prolonged standing or crowded, hot places Nassau, vomiting associated with syncope After sudden unexpected unpleasant sight, sound, smell or pain During the meal or just after meal With head rotation or pressure on carotid sinus After exertion
In syncope units (dedicated syncope clinics), even more "neural" or "reflex" type syncope are diagnosed -- 56-73% (Moya, 2009).
Closely related to the "neural" or vasovagal faints, are "situational" syncope, again syncope triggered by a bodily event.
- situational syncope (occurring after urination, defecation, swallow or cough)
- carotid sinus hypersensitivity -- syncope after pressure on the carotid sinus in the neck.
"Unknown" is the second biggest category - -accounting for 28% of all syncope in the ED ( Baron-Esquivias et al, 2010). In the general population, "unknown" is even bigger accounting for 37% (Sotierdes et al, 2002).
In the other 202 patients categorized in the ED study of Baron-Esquivias et al (2010) as "NST-LOC" -- non syncope transient loss of consciousness, "unknown" made up even more -- 43%. One might think -- how does an unknown in the "syncope" group differ from an "unknown" in the NST-LOC group ? Isn't there quite a bit of ambiguity here ? If we moved some of the "unknowns" in the "syncope" box into the "NST-LOC" box, wouldn't the proportions be very different. Why not just put them all together ... Another consideration is that what is "unknown" in the emergency department, could certainly move into some other category after more evaluation. Like arrhythmia for example. Well, anyway.
Orthostasis -- drop in blood pressure on standing is the next biggest category -- about 9.4% of all syncope (defined as transient LOC). We have another web page that discusses orthostatic hypotension here. It is surprising that there were only 9.4% given that such a large proportion of the population has orthostatic hypotension. Presumably this means that few people actually lose consciousness -- presumably they "feel it coming on" and sit down.
Cardiac events are also a relatively rare cause of syncope. In the Framingham study, a "cardiac" etiology was found in 13.2% of men and 6.7% of women (Sotieriades et al, 2002). Roughly the same as orthostasis. Spitting these out further we end up with ( Baron-Esquivias et al, 2010) subcategories of:
Arrhythmia -- The most common cardiac cause of syncope is defined as abnormal slowing or quickness or irregularity of the heart beating. Arrhythmia accounts for only 7.3% of syncope in the ED defined as transient LOC. Examples of arrhythmias are (Moya, 2009):
- bradycardia (slowness of heart beat, such as heart rate < 50)
- atrial fibrillation (irregular heart beat)
- sick-sinus (variable heart rate between slow and fast)
- AV block (normal pacemaker fails and heart is driven by other sites).
- supraventricular or ventricular tachycardia (abnormally fast)
According to Walfridsson and Walfridsson, "dizziness" was reported in 47% of patients with SVT (supraventricular tachycardia). There are numerous drugs that can cause bradycardia, especially in persons affected by the "long-QT" syndrome.
An even smaller fraction of syncope in the ED was due to "cardiopulmonary" events (4.2%) --
- aortic stenosis and other valvular disease
- pulmonary embolism or hypertension
- myocardial infarction (heart attack)
- Basically all the rest of heart disease.
There are a few miscellaneous entities not included above.
- Vasospasm induced syncope
- Basilar artery migraine
- Raynauds Disease (Guiloff et al, 2012) -- this is difficult to comprehend.
In the differential diagnosis of syncope one must also consider numerous non-cardiac causes of transient loss of consciousness. Again from Baron-Esquivias et al (2010), "Unknown" is very common in the ED (43%). One would think that in the "unknown in the ED" group there might be some rather serious conditions, motiving one to have followup for the "unknown". Remaining include
- seizure (18%)
- stroke or TIA (11.4%)
- metabolic encephalopathy and toxins (such as hypoglycemia or ethanol)
- psychiatric conditions (pseudosyncope)
- Dizziness (rare).
Common drugs that are associated with an excess risk of syncope in the elderly include:
- fluoxetine (Prozac)
- aceprometazine (Not available in the USA)
- haloperidol (Haldol)
- L-dopa (Sinemet)
- There are also many drugs that prolong the QT interval and that may precipitate heart block and syncope.
(Cherin et al, 1997). These associations seem very odd to us and we are dubious that they should be used to make any decisions about medications.
Syncope, defined as transient loss of consciousness, is dangerous. There appears to be a tendency in the literature to overestimate the risk however. As an overview, it seems to be just a little more dangerous than getting older.
Solbiati et al (2015), in a meta-analysis including 25 studies and 11158 patients in total reported that the one year mortality of syncope was 8.4%. Of course, syncope mainly occurs in older people and this statistic might be less dramatic if one subtracted the expected mortality -- i.e computed "excess mortality". From the figure on the right, taken from the Framingham study (Soteriades et al, 2002), it does not appear to us that there is an 8% mortality at 1 year -- rather the "worst case" mortality is in the 10% of syncope that are cardiac which may indeed be about 10%. For the other 90% mortality appears to be somewhere between 1 % (which is about normal), and 3% which is significant but hardly 8.4%. So there are some major differences in studies attempting to quantify expectation of death after syncope. We think the Framingham data is more reasonable and think the meta-analysis of Solbiati et al is overly gloomy.
According to Dipaola et al, the 48 hours following syncope are the highest risk times for a major adverse event. (Dipaola et al, 2014). In reality, these studies are reporting mortality in persons who have had syncope, and the underlying cause of syncope is generally the cause of death (Moya, 2009).
Soteriades et al (2002, figure to right) reported survival after syncope (defined as transient loss of consciousness), broken out by cause. This was a very large "population" type study done in Framingham, Conn.
Importantly, there was no effect on survival for individuals with "vasovagal" syncope (also called "neural syncope above"). However, all other major categories - -including "unknown", neurological causes (such as stroke), and individuals with cardiac causes of syncope did not survive as long as persons without syncope.
About 30% of people with one syncopal episode have recurrence in 3 years (Moya, 2009). Logically, one would expect that recurrent syncope would be far more dangerous than a "one time" syncope.
Entem et al (2009) reported that arrhythmia was found in 33 of 140 patients with syncope of unknown origin, using an implantable loop recorder, during a mean followup period of roughly a year. Given that one would be unlikely to convince most people to have a device implanted unless they had some very good reason, we think this statistic is far too high because of sampling bias.
History and Physical examination
According to an extensive meta-analysis by Linzer and associates (1997) for the American College of Physicians, "a careful history, physical examination combined with an ECG will yield a diagnosis in 50% of cases". Or looking at the doughnut hole, a careful history, physical exam and ECG yields a diagnosis in only 50%. As only about 10-20% of syncope is "cardiac", we presume that Linzer is saying that 20% of the diagnoses are made by the ECG, and the rest are through the general physical examination and history. This seems very reasonable. Especially the history, as talking to the patient is often the most important part of an evaluation.
History should establish the context of the syncope -- i.e. were they getting a flu shot ? (i.e. vasovagal) The warning signs and rapidity, what happened after the loss of consciousness (i.e. perhaps a convulsion ?), any available data about blood pressure -- were they pale ? No pulse ? How long did it take for them to recover ?
Reuber et al (2016) used a factor analysis to find questions that were best at discriminating between seizures, pseudoseizures and syncope -- obviously here their definition of syncope was not transient loss of consciousness as recomended above. Recall that only a small proportion of transient loss of consciousness in the ER are from seizures or "pseudo"-LOC. Questions reflecting a "catastrophic experience" -- i.e. a dramatic experience -- were endorsed to a greater extent by pseudoseizure patients, and were least endorsed by seizure patients. Questions reflecting "Mind/body/world disconnection" had a similar pattern. So in essence, dramatic answers to questions about what happened are more common in pseudoseizure patients. No big surprise, as pseudoseizures are often just mini-dramas intended to accomplish some purpose.
Moya (2009) suggested (paraphrased here) that one should ask about circumstances (position, activity, predisposing factors), Onset of attack (especially palpitations), eyewitness inquiry (fall, color, duration, breathing, convulsions, tongue biting), End of attack (e.g. confusion, chest pain, palpitations, incontinence), and background (cardiac, neurological, diabetes, medication, previous spells).
Bare minimum physical examination should include.
- blood pressure and pulse, standing and supine (i.e. orthostatic blood pressure)
- cardiac exam
Of course, a general physical and neurological examination is important as well.
In the emergency setting, blood tests including D-dimer may be useful for identifying pulmonary embolism or aortic dissection, and Troponin may be useful when a cardiac ischemic event is suspect. BNP (brain natriuretic peptide) > 300 pg/ml is a major adverse probability predictor, increasing odds as much as 7.3 fold (Reed et al, 2007). BNP is a peptide that suggests acute heart failure.
Interestingly, the orthostatic blood pressure measurement is often left out in emergency department evaluations. As about 10% of syncope in the ED is from orthostasis (Baron-Esquivias et al. 2006), this is a place where improvement is possible.
If the syncope can be triggered by a particular maneuver, such as stretching or straining, it may be helpful to have the individual attempt to trigger the syncope in a controlled setting, such as the medical office. Stretch syncope is a benign cause of fainting in young persons. (Mazzuca et al, 2007; Pelekanos et al, 1990).
Carotid sinus testing during the physical examination may also be helpful in very rare instances.
Not everyone agrees as to what should be the routine laboratory testing for syncope. This is usually left up to the discretion of whatever physician, nurse practitioner or physician's assistant that ends up caring for the patient after they are discharged from the emergency department. It is interesting to observe that many women use their OB-Gyne physician as the primary care physician. We have encountered patients who consider chiropractors to be their primary care provider. Recall that syncope has some excess mortality at 1 year (see above).
A full ECG is advised by nearly all. This is very sensible as it detects important conditions such as a heart attack or an irregular heart rhythm. A full ECG would seem unneeded in someone with a clear vasovagal faint however.
In a second meta-analysis by the Linzer (1997), the value of diagnostic testing for syncope was considered, with recommendations for the following studies in the other 50% of patients who are undiagnosed after history and physical -- i.e. "unknown" patients" is to get 1) electrophysiologic studies in patients with organic heart disease, 2) Holter monitoring or telemetry in patients known to have or suspected of having heart disease, 3) loop monitoring in patients with frequent events and normal hearts, 4) psychiatric evaluation in patients with frequent events and no injury, and 5) tilt-table testing in patients who have infrequent events or in whom vasovagal syncope is suspected. " These appear to be safe suggestions, that seem heavily weighted towards cardiac care.
The diagnostic yields of echo, stress testing, Holter monitoring and electrophysiological studies alone or in combination varies widely (5-35%). Entem et al (2009) found arrhythmia's in about a third of their "unknown" patients with bradycardia in about half and tachycardia in 11%, using implanted loop recorders. To us, this seems a little more expensive and invasive than warranted, especially considering the new availability of ambulatory monitors (see below)
Electrocardiograms and related testing for syncope.
As most syncope is not caused by electrical problems in the heart, but rather is from "vasovagal syncope", or from orthostatic hypotension, it should be apparent that electrocardiograms as well as related tests in all syncope patients will frequently be normal. One would expect from the prevalence of arrhythmia and other cardiac problems in the ED, the positive rate should be roughly 20%. Presumably the prevalence is higher in groups selected to be of higher risk -- i.e. those with recurrent syncope of unknown origin.
That being said, cardiac causes of syncope can be lethal and nearly everyone agrees that a routine ECG is often helpful in identifying abnormalities of rhythm, conduction or morphology of the heart electrical activity that give a clue as to the underlying etiology of the syncope. Here the cost of the test (Medicare pays about $60 for an EKG) is low, and the potential benefit of detection of a potentially lethal condition is high.
Recording a subset of the ECG during the spell can be achieved by using 24 hour ambulatory recording (also known as Holter monitoring), or an event recorder, or a memory loop recorder (called "loop" monitoring). These devices are constantly recording the ECG, allowing one to "go back in time" once an event is triggered. There are also some new "home devices" for personal use - -see below. The diagnostic yield of ambulatory event monitoring in syncope is 25-35% (Linzer et al, 1997). This seems a bit high to us, again perhaps due to selection bias. As event monitoring is meant to detect arrhythmia, and large studies suggest that only about 10% of syncope (defined as transient LOC) in the emergency department is diagnosed as arrhythmia (Sotiades et al, 2002) , the logical inference that many arrhythmia's are missed in the ED. This might be because the arrhythmias are intermittent, or perhaps because vasovagal patients are excluded from ambulatory event monitoring.
Twenty-four hour heart monitoring, often called "Holter monitoring", is potentially useful in persons who have frequent enough spells that can be expected to have an event during the 24 hours that they are monitored. Such individuals need to have a non-life threatening spell to make this modality safe. The yield of Holter monitoring is low, about 1-2%, and it is thus expensive in terms of cost per diagnosis (Moya et al, 2009).
The event recorder and loop memory recorder are useful in persons whose events occur less frequently than every 24 or 48 hours. These devices require the patient or an accompanying person to activate the monitor at the time of symptoms. (Hammill, 1997).
Recently implantable cardiac monitors have been made available for persons who need chronic monitoring of heart function (i.e. Reveal device, made by Medtronics). According to Burkowitz et al (2016), these devices are 4 times more productive of arrhythmia's than external event monitors. Of course, they are very invasive and very expensive.
Kardia device Kardia output on smart phone
Additionally, there are now smartphone devices (e.g. Kardia, AliveCor), that can be used to monitor the EKG on demand, using a smartphone app These devices are far less expensive (and time consuming) than undergoing ambulatory event monitoring. They do require one to be awake, or be in the company of someone who can operate the monitor, and they do not go "back in time" as do the Holter or similar devices, but their low cost (about $100) makes them very accessible. One wonders why a similar device could not be used to stream to an "app" on one's cellphone. I.e. why couldn't one make a $100 ambulatory event monitor ?
Moya et al (2009) stated that external event recorders have "no role in the evaluation of syncope". We are a little dubious about this, as we think that external event recorders as the device shown above could potentially be used in a far greater population of individuals than devices supplied by cardiology labs.
Other tests for syncope.
- Autonomic testing, including tilt-table testing, for orthostatic hypotension. It is not needed for all syncope patients -- more discussion is below.
- Carotid Doppler (very low yield). It is almost never needed.
- EEG (low yield due to combination of relative infrequency of seizures and insensitivity of EEG).
- Electrophysiological testing (catheterization, rarely done -- looking for illnesses such as WPW)
- Psychiatric evaluation (for pseudo-LOC). Rarely done.
The tilt table test procedure uses equipment to record blood pressure and pulse after a 70 degree tilt using a motorized table. It is looking for the 10% of patients who have syncope associated with orthostatic hypotension.
The "yield" on TTT varies according to author. According to Kapoor (2003), about 65% of patients with syncope but without structural heart disease have positive results on tilt-testing. As the huge majority of syncope patients are otherwise normal persons who have experienced an emotional disturbance (e.g. vasovagal), this figure seems very high (Baron-Esquivias, G., et al. 2010). We would rather expect about 10%, the prevalence of orthostatic hypotension. Grossi et al (1987) reported positive TTT in 26.69% -- again, seeming a bit high. Perhaps there is a selection bias here.
Three types of responses to the TTT are considered abnormal:
- "Neurally mediated response", characterized by the sudden onset of hypotension, bradycardia, or both when the patient is kept in the upright position. This is the most common response. The low blood pressure is also called "orthostatic hypotension".
- Positional orthostatic tachycardia
- Dysautonomia -- a gradual decrease in blood pressure with little or no change in heart rate during the procedure.
The TTT can be anxiety provoking by itself, as it involves being strapped onto a contraption, potentially with intravenous medication. Given that many people feel faint even with simply having their blood drawn or a flu shot, the "test" itself may be provoking syncope.
The sensitivity of TTT depends on the methodology. It is about 90% when used with a low-dose isoproterenol protocol or nitroglycerin, but lower when used with a high-dose isoproterenol protocol.
We ourselves do not think that TTT using IV medication is sensible in syncope as we are concerned about the risk of provoking a cardiac event, and are unconvinced that the greater sensitivity warrants the risk. That being said, Moya et al (2009) stated that there was essentially no risk of the TTT. We are doubtful.
The operators of tilt tables (usually cardiologists) are usually unsophisticated about positional vertigo due to inner ear conditions, and can easily miss the single most common cause of positional vertigo -- BPPV. We have encountered patients who pass out after their BPPV is provoked.
Syncope -- defined as transient loss of consciousness, is a major diagnosis problem, but once the diagnosis is established, treatment is generally managed by the specialties that treat the diagnosis. Somewhat unique to syncope is that there are an immense number of "unknown" patients, some of which have lethal underlying disorders.
Vacek (1991) suggested that "The initial evaluation and management of most patients with syncope can be performed by the primary care physician ... Consultation with a cardiologist is necessary only when the results of the initial evaluation are unrevealing or if findings suggest the need for a more intensive secondary evaluation. "
The following is our opinion about who should treat transient loss of consciousness.
Providers for Syncope Cause Specialty Vasovagal No treatment or evaluation is needed, unless the patient is actually "unknown" Unknown Followup with a medicine physician. It should be a specialty of physician familiar with cardiac disorders. Non-blood pressure LOC Neurology -- for example, seizures might be the cause. Cardiac syncope Cardiology Orthostatic (including "reflex") Medicine or Cardiology, possibly followed by Neurology if syncope is not due to a heart condition, medication, metabolic or dehydration as a cause.
More information about treatment of syncope associated with low blood pressure on standing up, can be found on the orthostatic hypotension page.
There a reasonably large literature concerning syncope, driving and working. This is discussed separately.