Recent Updates and Technological Developments in Evaluating Cardiac Syncope in the Emergency Department

Abstract

Syncope is a commonly encountered problem in the emergency department (ED), accounting for approximately 3% of presenting complaints. Clinical assessment of syncope can be challenging due to the diverse range of conditions that can precipitate the symptom. Annual mortality for patients presenting with syncope ranges from 0-12%, and if the syncope is secondary to a cardiac cause, then this figure rises to 18-33%. In ED, it is paramount to accurately identify those presenting with syncope, especially patients with an underlying cardiac aetiology, initiate appropriate management, and refer them for further investigations. In 2018, the European Society of Cardiology (ESC) updated its guidelines with regard to diagnosing and managing patients with syncope. We highlight recent developments and considerations in various components of the work-up, such as history, physical examination, investigations, risk stratification, and novel biomarkers, since the establishment of the 2018 ESC guidelines. We further discuss the emerging role of artificial intelligence in diagnosing cardiac syncope and postulate how wearable technology may transform evaluating cardiac syncope in ED.

1. INTRODUCTION

Syncope is a brief loss of consciousness due to a sudden fall in cerebral perfusion. In the emergency department (ED) and clinics, it is still a prevalent presenting complaint. Differential diagnosis of syncope is diverse and management depends on the specific aetiology. Physicians who are presented with a case of syncope must work swiftly and systematically by obtaining an accurate history, conducting physical examinations and appropriate investigations to identify the underlying cause. Physicians must then stratify patients into low risk and high risk groups and decide if the patient is fit for discharge or warrants further investigations and treatment as an inpatient. According to the European Society of Cardiology, syncope is defined as a ‘transient loss of consciousness (TLOC) with an inability to maintain postural tone due to cerebral hypoperfusion, characterised by a rapid onset, short duration, and spontaneous complete recovery.’ [1]. TLOC in itself is further defined as a ‘state of real or apparent LOC with loss of awareness, characterised by amnesia for the period of unconsciousness, abnormal motor control, loss of responsiveness and a short duration.’ TLOC can be broadly categorized into “non-traumatic TLOC” and “TLOC due to head trauma” (Fig. 1) [1].

Fig. (1).

Classifications of transient loss of consciousness (TLOC). TLOC can firstly be subdivided into either “Traumatic” or “Non-Traumatic” causes. Non-traumatic differentials can then be further split into four categories; syncope, epileptic seizures, psychogenic and rare causes. The differential diagnoses within each of these sections are outlined. Adapted from the 2018 European Society of Cardiology for the diagnosis and management of syncope.

The annual number of episodes of syncope is estimated to be between 18.1 and 39.7 per 1000 patients [2]. Syncope accounts for 3-5% of presenting complaints in ED, with a hospitalisation rate in 40% of cases [3]. Economic analysis suggests that syncopal episodes, which lead to hospitalisation, have an annual cost of $2.4 billion (£1.9 billion) [4]. Its frequency in the general population varies from 15% (in those under the age of 18 years of age) [5] to 23% (among the elderly) [6]. Cardiac syncope, in particular, remains the 3^rd most common (2^nd most common in patients above 60 years) cause of syncope, accounting for almost 25% of ED syncope presentations [7]. Cardiac causes of syncope remain independent predictor of mortality [8].

Syncope is caused by a partial or complete loss of blood flow to the brain. Cerebral perfusion is dependent on systemic arterial pressure (SAP). SAP is largely a product of cardiac output (CO) and systemic vascular resistance (SVR). If cerebral perfusion is disturbed for 6-8 seconds, the systolic blood pressure drops to 60mmHg, or oxygen delivery falls by 20%, syncope can occur. There are several causes of syncope and for physicians to accurately evaluate patients with a possible cardiac syncope, an overall understanding of the mechanisms and prevalence is fundamental Table 1.

Table 1.

Syncope may be classified into five main groups; reflex syncope, orthostatic syncope, neurological syncope, cardiac syncope, and other causes. The underlying aetiology of reflex syncope may be due to a vasovagal response, situational dependence, or increased sensitivity of the carotid sinus. Orthostatic syncope may occur due to sympathetic dysfunction. Cardiac syncope may result from underlying structural or channel conduct defects leading to arrhythmias. Neurological syncope may result from cerebrovascular disease, subclavian steal syndrome, or seizures. Other infrequent causes of syncope may be due to hormone imbalances and psychiatric disturbances. The mechanisms of each of these causes and their relative frequencies are described.

Classification	Causes	Mechanism	Mean Prevalence (Range) %
Reflex Syncope	Vasovagal	Nucleus tractus solitarii is activated directly or indirectly by a triggering stimulus that enhances parasympathetic stimulation and inhibits sympathetic stimulation resulting in decreased heart rate and fall in BP. Common triggers include emotional stress, pain, and sight of blood.	18 (8-37)8
	Situational	Similar to vasovagal syncope; common triggers include coughing, straining, after or during urination	5 (1-8)8
	Carotid sinus hypersensitivity	Baroreceptors on the carotid sinus are hypersensitive such that external pressure triggers the afferent glossopharyngeal and vagus nerves, which in turn activates parasympathetic response resulting in bradycardia and hypotension	1 (0-4)8
Orthostatic Syncope	Orthostatic hypotension	Sympathetic dysfunction with insufficient peripheral vasoconstriction and inadequate increase in heart rate resulting in lack of cerebral perfusion	8 (4-10)8
Cardiac Syncope	Structural heart disease	Results in haemodynamic instability and decreased cardiac output, which in turn impairs cerebral blood flow	4 (1-8)8
	Brady-/Tachyarrhythmias	Results in reduced cardiac output and impaired cerebral blood flow	14 (4-38)8
Neurological	Cerebrovascular disease	Obstruction to the flow of blood towards the brain results in poor cerebral perfusion	10 (3-32)8
	Subclavian Steal Syndrome	Stenosis of the subclavian artery results in retrograde flow from the vertebrobasilar artery, which in turn reduces blood flow to the brain
	Seizures	Short alteration of neurological function due to the abnormal electrical activity, which can mimic syncope
Other	Psychiatric causes	These include psychogenic pseudosyncope. The mechanism is not entirely understood and rather diagnosed when other causes are ruled out. There may be a history of stress.	2 (1-7)8
	Endocrinological causes	Several mechanisms may lead to syncope, including hypoglycaemia and low steroid levels (Addison’s disease), resulting in low BP	34 (13-41)8
	Unknown

Despite updates in ESC guidelines in 2018 [9], evaluating and identifying cardiac syncope in ED remains a challenge. Here, we discuss considerations of each cardiac syncope work-up component that expands beyond current guidelines. We also explore how innovative biomarkers from artificial intelligence (AI) and wearable technology could change the landscape of cardiac syncope diagnosis and therapy in the future.

2. RECENT CONSIDERATIONS IN EVALUATING PATIENTS WITH POSSIBLE CARDIAC SYNCOPE

2.1. History

A detailed history of events from the patient is essential for the accurate diagnosis of a syncopal episode. Possible efforts must be made to gain a collateral history from witnesses. Recent cohort study by Van Wijnen and colleagues [10] stressed that history taking remains the most important component for the evaluation of patients with syncope. In order to develop a robust model for symptoms associated with cardiac syncope, Berecki-Gisolf and colleagues [10, 11] evaluated four independent data sets consisting of a total sample size of 5739 patients with syncope, of whom 647 had cardiac syncope. The group reported that variables significantly associated with cardiac syncope included; age > 60 years, male gender, structural heart disease, less than three spells, supine syncope, and exertional syncope. Factors associated with a non-cardiac cause included; age < 40 years, nausea, diaphoresis, blurred vision, and long prodrome. Interestingly, the group did not find a statistically significant association with symptoms between palpitation and cardiac syncope vs. non cardiac syncope. However, symptoms of palpitations, especially in younger patients, cannot be ignored. In a study by Colman et al. [12], it was found that palpitations in patients < 40 years of age with Long QT syndrome (LQTS) were more often the only presenting complaint. Indeed, the ESC has also outlined in its guidance that palpitations prior to a syncopal episode should raise alarms of a possible cardiac aetiology. This study also highlighted the importance of family history in younger patients, as family history of syncope, sudden cardiac death, and cardiovascular disease were present more often in patients with LQTS. It is also significant to note any medications the patient might be taking; several medications have been associated with torsade de pointes and subsequent prolongation of the QT interval resulting in cardiac syncope [13]. Blood pressure lowering agents should also be accounted for in order to differentiate cardiac syncope from orthostatic hypotension. Moreover, social history, including alcohol and recreational drug use, can precipitate syncope or seizures and must be accounted for. Finally, asking about any recent stressors, the absence of which may help the physician in ruling out a psychiatric element to the presentation.

3. PHYSICAL EXAMINATION

Despite innovations in medical technology and quick access to investigation results, Lekic et al. [14] argued that the cardiovascular physical examination remains a fundamental and cost-effective method enabling physicians to arrive at a diagnosis quickly and accurately. A recent systematic review [15] studying 4317 patients with cardiac syncope concluded that the clinical exam in conjunction with the electrocardiogram (ECG) can accurately identify patients with and without cardiac syncope. Physical work-up should include examination of the cardiovascular and neurological systems. Lying and standing BP remains a pertinent difference between cardiac syncope and orthostatic hypotension. Despite this knowledge, Heldeweg et al. [16] reported that orthostatic blood pressure monitoring (OBPM) is only conducted in 16% of patients presenting to ED with syncope. This discordance between clinical practice and the ESC syncope guidelines therefore suggests more effort is required to increase the use of OBPM in the evaluation of syncope in ED. Saad Shaukat and colleagues [17] aimed to investigate the impact of adherence to guidelines based on syncope protocol in ED and hospitalization in patients presenting with syncope. The group developed a guideline based algorithm for use in ED, which incorporated OBPM in its initial work up. The authors further highlighted salient physical examination findings that should trigger suspicion of cardiac syncope, including systolic BP <90mmHg at presentation, heart rate < 40 beats per minute, and a new murmur. The study showed that after the introduction of a syncope protocol, there was an improvement in the proportion of high risk patients being admitted (68.7% to 82.1%). Taken together, this study shows that utilizing a guideline based syncope protocol in ED can aid physicians in conducting the appropriate physical examinations and identifying patients with cardiac syncope.

4. ELECTROCARDIOGRAM

Arrhythmias are a common cause of cardiac syncope. Given that cardiac syncope is most prevalent in older adults, ECG becomes a vital tool in stratifying these patients by risk. Specific ECG predictors of arrhythmia and adverse events have been variably described. In order to account for specific ECG predictors, Nishjima et al. [18] conducted a prospective study across 11 EDs in 3416 patients above the age of 60 years. Shortened PR intervals, first degree AV block, non sinus rhythm, complete left bundle branch long, dynamic Q/T wave and ST segment changes, and multiple premature ventricular beats were all shown to be associated with an increased risk of a 30-day severe cardiac arrhythmia in the study.However, detecting arrhythmias remains a challenge, especially if the arrhythmia is episodic and not captured by the ECG. To investigate the optimal duration of ECG monitoring following a syncopal episode, Thiruganasambandamoorthy et al. [19] recently conducted a prospective cohort study on 5581 patients. The group followed patients presenting with syncope to ED for 30 days categorising them by risk using the Canadian syncope risk score (CSRS). The authors found that serious underlying arrhythmias were mostly captured within the first 2 hours of arrival to ED for low-risk patients and within 6 hours for medium and high risk patients. Similar trends were outlined by Solbiati and colleagues [20] in a multicentre observational study. The authors assessed the sensitivity and specificity of the duration of ECG monitoring in identifying 7- and 30-day adverse and arrhythmic events. The study showed that for ECG monitoring of more than 12 hours, the sensitivity and specificity in identifying 7-day adverse events were 0.89 (95% CI = 0.65 to 0.99) and 0.78 (95% CI = 0.67 to 0.87), respectively. Similar results were also observed for 30-day adverse events These studies imply that ECG monitoring in individuals with potential cardiac syncope should be done for extended periods of time or repeated after several hours to guarantee that any underlying arrhythmia is detected. Whilst longer ECG monitoring could be beneficial in identifying patients with cardiac syncope, its wider impact on the healthcare system must be considered. In the United Kingdom (UK), for example, the National Health Service (NHS) has pledged that 95% of patients presenting to ED must either be admitted, transferred to another provider, or discharged within a four hour window [21]. If the department fails to achieve these standards, it may be fined.Therefore, longer ECG monitoring could increase financial pressures and make ED more vulnerable to not meeting these targets.

5. BIOMARKERS

Studies have investigated different biomarkers to enable a more accurate diagnosis of cardiac syncope. Du Fay de Lavallaz et al. [22] showed that B-type natriuretic peptide (BNP), N-terminal pro b-type natriuretic peptide (NT-proBNP), high sensitivity cardiac troponin-T (hs-cTnT), and high sensitivity cardiac troponin-I (hs-cTnI) were significantly higher in patients with cardiac syncope compared to other causes. These biomarkers individually provided a diagnostic accuracy for cardiac syncope of 77-78%; however, combining the 4 biomarkers improved diagnostic accuracy to 81% [23].

Whilst BNP and troponin related biomarkers may be familiar tests for clinicians, more novel biomarkers are under evaluation. Badertscher et al. [23] evaluated the diagnostic value for cardiac syncope of 4 novel prohormones: MRproANP, C-terminal proendothelin 1, copeptin, and mid regional-proadrenomedullin. They found that these novel biomarkers were higher in those with cardiac syncope than other causes. MRproANP had the best diagnostic value among them, with a diagnostic accuracy of 80% alone and 90% when combined with the “probable diagnosis of cardiac syncope” from the ED physician. A 99% negative predictive value was obtained when MRproANP was less than 77 pmol/L and the probability of cardiac syncope was less than 20% (as assessed by an ED physician).None of these novel biomarkers are routinely advised by the ESC, and therefore, further research is required with regards to their utility and validity for evaluating patients with syncope. Overall, there is evidence to show that these novel prohormones, particularly MRproANP, are additional helpful biomarkers and may be considered in the future to improve the early diagnosis of cardiac syncope.

6. ECHOCARDIOGRAPHY

Sarasin et al. [24] looked at the role of echocardiography in the stepwise workup of both suspected cardiac syncope and unexplained syncope. Echocardiography confirmed severe aortic stenosis in 40% of those for whom it was highly suspected after the examination. In comparison, it showed normal findings in 100% of those with unexplained syncope. The authors concluded that echocardiography was more helpful to use for patients with a history of cardiac disease or an abnormal ECG. This modality could also play a role in risk stratification; the study also showed that echocardiography detected moderate left ventricular systolic dysfunction in 27% and arrhythmia in 38%, thereby demonstrating the ability of echocardiography to assess the severity of the underlying cardiac disease.

More recently, in 2019, Probst et al. [25] repeated a similar analysis on a larger sample group. They found that echocardiography established significant findings in only 22% of those for which it was used. Subsequently, they proposed the ROMEO (Risk of Major Echocardiography findings in Older adults with syncope) score (Table 2), which determines the likelihood of significant findings on echocardiography using the following predictors: history of cardiac failure or coronary artery disease; abnormal ECG, elevated NT-proBNP, and elevated hs-CTnT. A score of 0 had 99.5% sensitivity and 15.4% specificity for excluding significant findings on echocardiography. ROMEO score may therefore be considered by a physician to identify which patients are unlikely to have cardiac aetiology, thereby preventing unnecessary use of further resources [25].

Table 2.

Since the ESC 2018 syncope guidelines, several risk stratification tools have been proposed for predicting serious adverse events following syncope. These include; ROMEO score, FAINT score, and Bozorgi and colleagues’ syncope risk score. The components used in each of these scoring methods and their measured outcome are outlined above.

Scoring Tool	Year of Origin	Predictors	Outcome Definition
ROMEO	2018	cardiac failure; coronary artery disease; abnormal ECG; elevated NT-proBNP; elevated hs-cTnT	Significant findings on echocardiography
Bozorgi et al.	2018	symptoms; history of cardiovascular disease; EF < 50%, abnormal	Serious adverse events
FAINT	2020	cardiac failure; arrhythmias; abnormal ECG; elevated NT-proBNP; elevated hs-cTnT	Serious cardiac events in 30 days

7. UPDATES IN RISK STRATIFICATION

Over the last two decades, several scoring systems have been developed to stratify the risk of adverse events following syncope. These scoring systems primarily identify cardiac related predictors as risk factors for serious adverse events, further reiterating the fatal prognosis associated with cardiac syncope. However, these scoring tools had been limited by their complexity of use and insufficiency of data showing improvement in patient outcomes. As a result, they had not been widely adopted [26].

The large majority of these scoring tools were derived before the ESC updated its guidelines for syncope in 2018. Since then, Borzogi et al. [27] proposed a new and more feasible scoring model that scored points for the following: presenting symptoms; history of cardiovascular disease; ejection fraction (EF) < 50%; abnormal ECG (Table 2). The authors tested this unnamed tool against its predecessors and reported it was better at predicting serious events. The FAINT score, derived by Probst et al. [26] in 2020, is the most recent of the scoring tools for syncope. FAINT score has 96% sensitivity and 22.2% specificity for serious cardiac events within 30 days (Table 2). The authors also showed that the FAINT score was more accurate than the unstructured judgement from a physician, with an AUC of 0.704 vs. 0.630, respectively. Therefore, these more recent scoring tools may be considered when advising patients of the risk of short-term and long-term adverse events following syncope.

Following risk stratification, appropriate advice should be given to patients regarding the impact of syncope on their daily lives. The Driver and Vehicle Licensing Agency (DVLA) is the British government authority that publishes rules on driving safety in the UK. The main cardiac indication for informing DVLA includes patients who develop TLOC whilst sitting and standing, especially with an abnormal ECG or structural heart disease [28].

8. EMERGING TECHNOLOGIES AND EXPANDING POSSIBILITIES: THE FUTURE OF CARDIAC SYNCOPE EVALUATION

AI is an evolving phenomenon whereby computer systems demonstrate the ability to perform tasks often considered to require human intelligence. Machine learning (ML) is an element of AI that involves computers learning tasks and predictions automatically through data driven algorithms. In recent years, AI research within medicine has expanded and is widely considered a promising tool to address the current gaps in clinical practice in a wide range of specialties, including; radiology, [29] gastroenterology, [30] oncology, and pathology [31].

With regards to syncope, studies have shown that physicians using ESC guidelines have an accuracy of 80% in correctly identifying the underlying aetiology (10), whereas triage nurses have a sensitivity of 21% in identifying high-risk patients with syncope [32]. Despite updates in the risk stratification models, patients with syncope in high-risk categories may still be missed. There are several pitfalls to risk scores. These scores are limited in their generalizability as adverse outcomes following a syncopal episode may be contributed to by the presence of other health conditions. Moreover, these scoring tools take several years to construct and therefore lack the ability to quickly update and integrate new research as it becomes available. ML could address these shortcomings by quickly assimilating large quantities of data through readily available electronic health records and providing a more personalised prognosis for patients with syncope, thereby shifting the paradigm from one score fits all to one score fits one.

Studies show that compared to early warning scores, AI-based technology is superior at predicting cardiac arrests in ED [32-34]. Jiang et al. [35] assessed the performance of four common ML models for triaging patients with suspected cardiovascular disease in ED. The area under the receiver-operating curve (AUC) for this model ranged from 0.90 to 0.937. Wardrope and colleagues [36] further reported that ML based technology is able to appropriately differentiate between syncope and other common causes of TLOC. Recently, in a combined retrospective and prospective cohort study involving more than 2 million patients, Zhang et al. [37] reported that machine based algorithm may be feasibly applied to predict 1-year incident cardiac dysrhythmia in the general population. Together, these emerging trials show significant promise in using AI to evaluate cardiac syncope in ED.

In addition to AI, advances in the development and adoption of wearable technologies, including smartwatches and phones, will continue to provide new tools and opportunities to evaluate potential causes of syncope. These technologies utilise a variety of features, including ECGs, accelerometers, and Photoplethysmography (PPG), a tool that measures the rate of volumetric change of the heart by measuring light reflection from subcutaneous capillaries [38]. PPG has already been shown to be capable of characterising both heart rate and blood pressure changes [39]. The integration of these technologies could allow haemodynamic changes to be characterised in syncope patients and act as predictors of significant events.

The current literature has focused on the use of this technology in the detection of cardiac arrhythmias and, more specifically, Atrial Fibrillation (AF). The Apple Heart Study explored the ability of the PPG-based tachograms in the Apple Watch to detect AF and showed a positive predictive value of 71% compared to ambulatory ECG patches in patients with known AF [40]. A subsequent multicenter randomised control trial by Koh et al. found that smartphone-based ECG recordings detected AF after a cerebral ischaemic event better than 24-hour Holter monitoring (9.5% vs. 2.0%; absolute difference 7.5%; P = 0.024), leading to a higher proportion of patients on DOAC therapy at 3 months (9.5% vs. 0%, P = 0.002) [41]. The IPED (Investigation of Palpitations in the ED) trial compared the symptomatic rhythm detection rate of a smartphone-based event recorder alongside standard care versus standard care alone for 242 patients presenting to the ED with palpitations and pre-syncope with no obvious cause. The results showed the number of patients in whom an ECG was captured during symptoms increased over five-fold to more than 55% at 90 days [42]. These studies demonstrate the feasibility and promise that wearable technologies have to improve clinical care and patient experience for those suffering from syncope.

Future emergency department practitioners are expected to use smart technology in clinical decision-making in a variety of situations. It may be that patients who routinely use smart devices are alerted to potentially ‘pre-syncopal’ arrhythmias and are informed to visit their clinician (either in the community or urgently). These events can be recorded and can either be analysed automatically or interpreted during the consultation. The Apple Watch study showed that 57% of the study population sought medical help based on the watch findings, of which 36% underwent additional testing, with 33% being referred to a specialist and 28% being started on a new medication [40]. Furthermore, combining wearable technology with advances in machine learning will also permit smart devices to more accurately detect and eventually predict clinically significant events, such as premature ventricular contractions, supraventricular tachycardia, or sinus node dysfunctions using PPG based algorithms. Studies are also currently underway to evaluate the echocardiographic detection properties of PPG [43], which could see structural analyses of the heart integrated into clinical data. Secondly, there is potential for patients who present with syncopal symptoms to be signposted to use digital/wearable technologies to record events, with data being collected for evaluation at follow up [44]. This could increase the accuracy of diagnosis whilst decreasing the invasiveness of investigations.

There are currently several barriers to the integration of clinical data derived from smart devices and wearable technology to routine decision making in the ED. Firstly, effectiveness would be dependent on uptake by clinicians and patients. Clinicians may feel uncomfortable adopting technology that they cannot decipher or communicate the significance of results to patients. Additionally, current PPG technology shows a high rate of false positives [45], which could cause undue anxiety to patients and increase low value healthcare utilisation [46]. Conversely, devices that rely on patient activation will potentially under-detect subclinical or minimally symptomatic arrhythmias, leading to a potential false negative and thus false reassurances [47]. This is why current ESC guidelines have labelled the role of mobile phone based real-time recordings as “limited” in the evaluation of syncope (1). Furthermore, the amount and quality of data recorded, as well as predictive capabilities, would be influenced by the device's battery, storage, and computing power, which would subsequently depend upon the brand and model.Patient uptake would also be a limitation. Smartphone/watch ownership decreases with age, while the prevalence of syncope increases [48], meaning that the inherent pretest probability of a clinically relevant result in the population likely to use smart devices will be low. Moreover, using wearable health systems will incur a cost that may discourage adoption. However, it is likely that as technology becomes more ubiquitous, these limitations will be overcome, and cost-benefit analyses vs. the current standards of care will likely demonstrate more benefit as larger scale production will drive down costs.

CONCLUSION

Despite updates in guidelines, syncope remains a challenging presentation to evaluate in ED. While decades of effort have gone into building risk stratification methods to enable for more accurate identification of cardiac syncope in the emergency department, these models are restricted by the time it takes to construct them and their generalizability.The emergence of novel biomarkers and the use of AI and wearable technology is promising and may address the current pitfalls in identifying patients with cardiac syncope.

AUTHORS’ CONTRIBUTION

All authors contributed to data analysis, drafting or revising the article, gave final approval of the version to be published, and agreed to be accountable for all aspects of the work.

CONSENT FOR PUBLICATION

Not applicable.

FUNDING

None.

CONFLICT OF INTEREST

The authors declare no conflict of interest, financial or otherwise.