Prospective evaluation of diagnostic work-up in syncope patients: results of the PL-US registry

Syncope is a common problem. Demographic and clinical characteristics of patients admitted to different types of centres may vary, physician’s adherence to the guidelines has been examined only in a few studies, and the requirements for implantable loop recorders (ILR) have not been well deﬁned. The aim of this study was to (i) compare demographic and clinical characteristics of patients with syncope diagnosed and treated in tertiary electrophysiology cardiac centres and those attending syncope units or general hospitals, (ii) assess how physicians adhere to the published guidelines, and (iii) calculate the requirement for ILR insertion. the syncope) ECG suggesting arrhythmic syncope. Syncope of (53%), followed by reﬂex syncope (33%). Adherence to the guidelines was less than satisfactory—measurement of blood pressure in an upright position, carotid sinus massage, exercise testing, and electrophysiological study were underused, whereas prolonged ECG monitoring and neurological consultations were overused. Unexplained syncope had 58 (9%) patients, and 42 (72%) of them had indication for ILR which accounts for 6% of the whole study population. The calculated need for ILR was 222 implants/million inhabitants/year.


Introduction
Syncope is a common symptom in patients presenting to the emergency units as well as cardiac or internal medicine departments and outpatient clinics. The most frequent cause of syncope in this unselected population is reflex syncope (from 30 to 66%), whereas cardiac or arrhythmic syncope is less common (from 8 to 21%). 1 -6 Although demographic and clinical characteristics of this group of patients have been extensively described in literature, 1 -7 data regarding patients admitted for syncope evaluation and treatment to the tertiary cardiologic electrophysiological and pacing centres are lacking.
Identification of a cause of syncope is often difficult and requires several diagnostic tests. In recent years, the guidelines of the European Society of Cardiology 8,9 as well as the scientific statement of the AHA/ACCF 10 have been published which should facilitate and make more uniform the diagnostic approach to a syncope patient worldwide. However, only a few data have been reported in literature on the adherence to these guidelines in 'real world '. Brignole et al. 2 showed that after dissemination of the guidelines and instruction of a personnel, a strict adherence to the published guidelines may be achieved even in a 'real world '. In their study, a diagnostic work-up was consistent with the guidelines in 86% of the patients, and they achieved an excellent rate of establishing final diagnosis, 94%. In another study, Bartoletti et al. 11 evaluated the appropriateness of hospital admissions and discharges in patients with syncope. They showed that decisions concerning hospital admission were in line with the guidelines in almost 70% of the patients, whereas concerning discharge in 92% of the patients (P , 0.001). It is noticeable that in both studies, a great effort was undertaken to disseminate the current guidelines among physicians and to control the ongoing management in order to adhere as closely as possible to the guidelines.
In Poland, the ESC guidelines were translated and published in 2002 12 with several articles updating the guidelines in 2004 and 2005. 13 In addition, numerous meetings in several parts of the country, attended by cardiologists and general practitioners, were dedicated to the dissemination of the guidelines. However, no data are available as to how physicians in Poland adhere to these guidelines and recommendations in 'real life'.
The introduction of implantable loop recorders (ILR) has been shown to significantly improve diagnostic accuracy in patients with syncope. 14 -21 However, there is still a debate at which diagnostic step this device should be used and what are the requirements for ILR. For example, a study from Italy showed that 34 devices/one million inhabitants/year should be implanted. 17 However, this figure may differ depending on the stage of diagnostic work-up at which the device is implanted, and may also depend on the economical and health issues in different countries.
Accordingly, there were three main aims of our project. First, to obtain detailed demographic and clinical characteristics of the population of syncope patients diagnosed and treated in tertiary electrophysiology cardiac centres, and to compare these data with characteristics of populations described in other studies or registries. Secondly, to assess how physicians adhere to the ESC recommendations as far as the syncope work-up is concerned. Finally, what is the minimal requirement for ILR insertion, using the most conservative indication which is ILR implantation after all other tests occurred inconclusive.

Study design
This was a prospective, multicentre study called the PL-US (Polish patients with Unexplained Syncope) registry. A special database was constructed and physicians participating in this registry were instructed to enter all consecutive patients diagnosed due to syncope in her/his department. The institutions that were asked to participate in the registry were the tertiary centres fully equipped with diagnostic facilities, enabling complete evaluation of a patient with a history of syncope, including tilt table testing, Holter ECG monitoring, in-hospital ECG telemetry, external ECG event recorders (loop recorders), catheterization laboratory, electrophysiology (EP) laboratory for invasive testing, and pacemaker clinic performing follow-up of patients with pacemakers and implantable cardioverter-defibrillators (ICD). The protocol of the study was approved by the central Ethics Committee.
Consecutive patients admitted to these centres for evaluation due to syncope entered the registry. The inclusion criteria were (i) at least two syncopal episodes in the past or single episode of syncope, however, associated with injury, (ii) obtained written informed consent to participate in the registry, and (iii) age 18 years. Syncope was defined as a transient, loss of consciousness and postural tone due to cerebral hypoperfusion followed by rapid and complete spontaneous recovery.
Besides instruction given via Internet and mail, all physicians from the participating centres attended a 1-day meeting during which the protocol of the study was discussed, the ESC guidelines for syncope diagnosis were summarized, and the technical issues connected with the entering data into the Internet database were discussed. Also the ILR implantation and programming training was performed.
The database itself was very detailed and consisted of numerous demographic and clinical parameters to be entered. They included not only patient's characteristics and results of tests performed in the centres participating in the registry but also data on previously performed investigations in primary or secondary centres.
In order to facilitate and make it more practical, a diagnostic algorithm, based on the ESC guidelines, was drawn up (Figure 1). The algorithm stressed the importance of initial evaluation, defined as in the ESC guidelines, and consisted of two diagnostic arms, depending on the presence or absence of structural heart disease. Physicians participating in the registry were asked to follow this diagram; however, they were free to make an exemption if, in their judgement, the clinical situation required to do so. For example, EP study was recommended in patients with organic heart disease after other non-invasive tests were non-diagnostic; however, performing EP study in a patient with palpitations preceding syncope, normal other non-invasive tests, and no cardiac structural abnormality was regarded as a proper use of the test. Also, the order of performing the tests was suggested but was not obligatory (except initial evaluation and echocardiography for confirming or ruling out organic heart disease).

Adherence to the guidelines
Data on each individual patient was analysed with regard to the adherence to the ESC guidelines and diagnostic work-up scheme suggested in the protocol. Both the appropriateness of performing and not performing a test were analysed. All tests performed in an individual patient, both before and during index hospitalization, were taken into account.
The following definitions were used when assessing the appropriateness of using a given diagnostic test.
(i) Initial evaluation: it was assumed that all patients should undergo initial evaluation. Also, all individual components of initial evaluation were taken into account, including history taking, physical examination, blood pressure (BP) measurement in supine and erect positions, as well as 12-lead ECG recording. Not taking BP in an upright position in patients with syncope associated with acute coronary syndrome (ACS), advanced atrio-ventricular Prospective evaluation of diagnostic work-up in syncope patients (AV) block, acute stroke, or other acute conditions affecting circulatory system was regarded as correct management. (ii) Echocardiography was regarded as the mandatory examination when cardiac disease was suspected. Performing echocardiography as a second step after initial evaluation, which occurred completely normal or reflex syncope was suspected, was not regarded as an inappropriate diagnostic step since echocardiography is very frequently used to objectively and definitely rule out cardiac disease and physicians were not discouraged from performing echocardiography at this early diagnostic step. The only scenario when the use of echocardiography was deemed inappropriate was in case of initial evaluation, absolutely diagnostic for reflex syncope which was not associated with injury and was infrequent. (iii) Carotid sinus massage (CSM) was indicated for all patients aged .40 years in whom initial evaluation occurred not diagnostic. Thus, not performing CSM in this subset of patients was regarded as inappropriate unless contraindications to perform CSM were present (audible bruit over carotid arteries, known carotid artery stenosis, or a history of stroke/transient ischaemia attack). Performing CSM in patients ,40 years old was regarded as inappropriate unless circumstances of syncopal episodes were suggestive of carotid sinus hypersensitivity. (iv) Tilt testing (TT): appropriate use of TT was defined as performing TT in all patients without organic heart disease in whom initial evaluation occurred non-diagnostic or in those patients with organic heart disease in whom other tests excluded or did not confirm cardiac causes of syncope, or in patients in whom initial diagnosis was diagnostic for reflex syncope but who had very frequent symptoms, syncope associated with injury or jerky movements during syncopal attacks. (v) Prolonged ECG monitoring: 24 h Holter ECG was deemed appropriate when a patient had (i) frequent (at least one/ week) syncopal episodes (in patients with more than one syncopal episode, frequency was calculated by dividing the number of syncopal episodes in the past by the time from the first syncope to entering the registry), (ii) when syncope occurred within 1 week from index hospitalization, or (iii) when a patient had a significant structural heart disease and a high risk of life-threatening arrhythmias. Otherwise, the use of Holter ECG was regarded as inappropriate. In-hospital 24 h ECG telemetry was judged appropriate when a patient was hospitalized due to syncope which was likely to be arrhythmic (history of arrhythmia, organic heart disease, and abnormal baseline ECG), whereas inappropriate when there was a low probability of arrhythmic syncope and ECG and/or echocardiography were normal. (vi) External loop recorder was used for 1 month or until syncope recurrence and its use was regarded appropriate when syncopal spells were moderately frequent (at least once/4 weeks) 9 (vii) Exercise test (ET) was regarded appropriate when syncope was associated with exercise (during or after) or was a part of a diagnostic work-up towards possible coronary artery disease (CAD) and non-appropriate when syncope was not associated with exercise or there was a low probability of CAD. (viii) Electrophysiological study was regarded as appropriately performed when a patient had organic heart disease and other noninvasive tests were inconclusive. When EP study was not performed in a patient with organic heart disease and unexplained syncope after non-invasive tests, this was defined as a lack of adherence to the guidelines. Performing EP in a patient without an organic heart disease was left to the discretion of an attending physician and when there was a suspicion of arrhythmic cause of syncope, the use of EP was justified and classified as appropriate. (ix) Neurological consultation was deemed appropriate when there was a suspicion of neurological cause of syncope, based on initial evaluation or other tests.

Indication for implantable loop recorder insertion
The need for ILR implantation was defined as (i) all other tests inconclusive and (ii) syncope associated with injury or presence of organic heart disease or past medical history and ECG suggesting arrhythmic syncope.

Statistical analysis
Results are presented as mean + standard deviation or numbers and percentages. The need for ILR implantations in Poland was calculated based on the detailed data from the emergency department of the centre #1. The number of patients admitted to that hospital due to syncope or faint (ICD code R55) during the whole year 2008 and the population of the district of this hospital were used for this computation.

Demographic and clinical characteristics
In total, 669 patients admitted to 18 cardiac tertiary centres over a mean of 3 months (range 1-10 months) entered the registry. The mode of hospital admission was elective in 475 (71%) patients and urgent (due to recent syncope) in the remaining 194 (29%) patients. In 73 (11%) patients, this was the first medical contact due to syncope. Detailed demographic and clinical characteristics are presented in Table 1. Female gender dominated, three-fourths of the patients were aged .40 years, and more than a quarter had organic heart disease. Syncope burden was high-the mean number of past syncopal episodes was 9, with almost a quarter of patients having an injury associated with syncope. Also concomitant non-cardiac conditions were relatively frequent, with hypertension being the most common. Main demographic and clinical characteristics of our cohort and of patients included in some other syncope studies or registries are compared in Table 2. Although demographic parameters were quite similar, differences were visible in the final diagnosis with the syncope of cardiac/arrhythmic origin being by far more frequent in our patients. Also the proportion of patients with recurrent syncope was higher in our registry than in other studies.
The number and proportion of patients undergoing diagnostic tests in our registry and in some other studies are presented in Table 3. There were significant differences in the utilization of several tests. Prolonged ECG monitoring, echocardiography, TT,  (4) Any psychiatric disease 13 (2) Prospective evaluation of diagnostic work-up in syncope patients EP study, coronary angiography, and neurological examination were performed more frequently in our registry, whereas CSM was performed less frequently than in other reports. The number and proportion of tests performed in the past (prior to current index hospitalization) and at the tertiary centre (index hospitalization) are also outlined in Table 3.
Adherence to the guidelines Initial evaluation All patients underwent initial evaluation; however, not all components of this diagnostic step were performed. Medical history, physical examination, and standard 12-lead ECG were done in all 669 patients, whereas BP in the supine position was measured in 660 (99%) patients ( Figure 2). Blood pressure while standing was recorded in 185 (28%) patients. Of the remaining patients, acute contraindications to BP measurement in an upright position were present in 194 subjects (ACS, 20; advanced AV block, 117; sustained ventricular tachycardia, 32 and acute pulmonary embolism, 2). Thus, 465 patients should have undergone standing BP measurement, which gives an underutilization rate of 60% (280 of 465 did not have this test performed).

Echocardiography
Echocardiography was performed in 576 (86%) patients. None of the patients with a history of cardiac disease or with suspected cardiac disease based on the results of other tests was denied echocardiography. On the other hand, 27 (12%) of 221 patients with no signs of organic heart disease and typical reflex syncope underwent echocardiographic study.

Carotid sinus massage
There were 504 (75%) patients aged .40 years. The CSM was contraindicated in 27 patients with a history of stroke and in the other 3 patients with carotid artery stenosis. Initial evaluation occurred diagnostic in 183 patients from the age-group .40 years. In addition, in the majority of these patients, TT was performed prior to scheduled CSM and occurred abnormal in 64 of them. Out of this subgroup, 7 patients underwent subsequent CSM, whereas 57 did not. Thus, the remaining 234 patients had indication for CSM; however, it was done in only 37 (16%) patients.
Carotid sinus massage was performed in 11% of the patients aged ,40 years. None of these patients had a history suggestive of carotid sinus hypersensitivity.

Tilt testing
The test was performed in 239 patients. Of these, 35 (15%) had no indication for TT. This subgroup included 9 patients with initial evaluation fully diagnostic for reflex syncope not associated with injury, 8 subjects with initial evaluation diagnostic for other causes of syncope, and 18 patients who had diagnosis already established using other tests but underwent further evaluation using TT for unknown reasons.
Out of 58 patients who had unexplained syncope after the completion of the diagnostic work-up, 14 (24%) had no TT.

Prolonged electrocardiographic monitoring
Twenty-four-hour Holter ECG or in-hospital ECG telemetry was recorded in 526 (77%) patients. Of these patients, 114 (22%) patients had at least one syncope within 7 days from the test and 63 (12%) subjects had syncope within 7 days and likely to be due to cardiac arrhythmia.
There was no patient who had no prolonged ECG monitoring but had indications to perform it.

External loop recorder
This method was used in three (0.5%) patients, of whom two had syncope recurrence during monitoring which gives a diagnostic  yield of 66%. All these three patients had at least one syncope/ month which justified the use of this type of monitoring.

Exercise test
Syncope associated with exercise was present in 67 (10%) patients. Medical history and physical evaluation were diagnostic in 3 patients, whereas in 16 patients, baseline ECG was strongly suggestive of arrhythmic or cardiac cause of syncope so the following diagnostic work-up was continued without ET performing and the investigators proceeded directly to EP study or prolonged ECG monitoring. This leaves 48 patients in whom ET was indicated because initial evaluation was not diagnostic. Of this subgroup, ET was performed in 19 (40%) subjects. On the other hand, ET was performed in 7 (26%) of the 27 patients with a diagnosis of vaso-vagal syncope, based on typical history and normal initial evaluation, and no history of exercise-associated syncope.

Electrophysiological study
After completing non-invasive evaluation, 23 patients were left with a diagnosis of unexplained syncope and the presence of organic heart disease. Of those, eight (35%) underwent EP study.

Neurological consultation
A neurologist was asked for consultation in 237 (35%) patients. Of them, 57 (24%) (27 with a history of stroke, 3 with CAS, and 27 with syncope and jerky movements) had clear indication for neurological advice. Thus, the overutilization of neurological consultation was 76% (180 out of 237 had no indication for this consultation). Of 237 patients, in 4 (1.7%), neurological cause of syncope was found.  performed in 33 (5%) patients and confirmed psychiatric abnormality in 7 (21%) out of 33 patients.

Diagnostic efficacy
The number of patients diagnosed by an individual test is shown in Table 4. Excluding initial evaluation, the mean number of tests required to establish diagnosis was 4.6 + 2.6 (range from 0 to 16) when taking into account diagnostic procedures performed both in the past and during index hospitalization in a tertiary centre, or 2.3 + 1.6 (range from 0 to 8) when considering only tests performed in a tertiary centre. The final diagnoses are presented in Table 5. The most frequent type of syncope was cardiac/arrhythmic (53%), followed by reflex syncope (33%) and by syncope due to other causes (12%).
The total number of diagnoses was 714 which exceeds the number of studied patients because in some patients more than one cause of syncope was identified.
Unexplained syncope was present in 58 (9%) patients. The demographic and clinical characteristics of this group are presented in Table 6. According to the definition used in the present study, 42 (72%) patients with unexplained syncope had indication for ILR insertion which accounts for 6% of the whole study population.
On the basis of the files from the centre #1, the need for ILR implantation was assessed. During the year 2008, there were 10 927 hospitalizations in that centre, of them 740 (7%) were due to syncope or faint. Since the overall population of the district of this hospital is 200 000 inhabitants and, according to our definitions and results of the registry, 6% of all syncope patients required an ILR insertion, we estimate that the need for ILR in such a syncope population as studied in the present report is 222 implants/million inhabitants/year.

Demographic and clinical characteristics
The patients who were entered in the present registry were slightly younger and had a higher burden of syncope than those in other published studies. 1 -7 As far as the final diagnosis is concerned, the most striking difference between our and other studies 1 -7 was a high proportion of cardiac/arrhythmic syncope and relatively low rate of reflex syncope in our registry.
The most probable explanation for the above-mentioned discrepancies in clinical characteristics between our and other registries is the fact that our study was conducted in highly specialized tertiary cardiological centres, and that a significant number of these centres were focused on arrhythmic patients such as those with  pacemakers or ICD. Also, patients with syncope associated with ACS were included in the centres with a 24 h stand-by cath lab duty. Thus, the proportion of patients with the diagnosis of cardiac/arrhythmic syncope was higher than in other registries which included patients presenting to emergency departments, general hospitals, or specialized syncope units 1 -7 or excluded patients with organic heart disease. 18 Another difference between our patients and populations reported in literature is a low number of patients (n ¼ 73, 11%) for whom admission to the tertiary centre was the first medical contact due to syncope and who had no diagnostic tests performed in the past.

Adherence to the guidelines
One of the main goals of this study was to examine how physicians in tertiary centres in Poland conduct diagnostic work-up of syncope in light of the published guidelines. The adherence to the guidelines occurred rather poor. Although some tests were heavily overused (prolonged ECG monitoring or neurological consultation), others were severely underused, including BP measurement in an upright position, CSM, ET, or EP study. The reasons for these findings are probably multifactorial.
Initial evaluation was performed in all patients; however, not all its components were properly applied. All patients had standard ECG performed which was not the case in some studied populations. 22 The proportion of patients undergoing BP measurement while standing was unexpectedly low, which may be partly explained by the fact that patients with other obvious conditions causing syncope or with contraindications for orthostatic stress were included (e.g. pacemaker failure, sustained supraventricular or ventricular tachycardia, ACS, and others).
Tilt testing was overused probably mainly because some physicians are inclined to confirm the diagnosis of reflex syncope in spite of the fact that initial evaluation occurred diagnostic and no further tests are required according to the published guidelines. 9 This happened in 15% of the patients with a clear diagnosis of vasovagal syndrome in whom TT was not necessary to confirm the diagnosis.
Prolonged ECG monitoring has been reported to be overused in syncope patients by several authors. 9,10 However, in the presence of cardiac abnormality and increased likelihood of arrhythmic syncope, physicians are inclined to perform prolonged ECG monitoring, regardless of the frequency of syncopal attacks. The diagnostic yield of Holter ECG reported in literature is variable. It may vary from 22 23 to 12% 24 when patients with a high likelihood of arrhythmic syncope are included. However, in unselected patients with syncope, diagnostic yield of 24 h ECG monitoring may be as low as 1-2%. 9 The ESC guidelines 9 recommend performing Holter ECG when syncope occurs at least once/week (this recommendation was used in our registry), whereas the US position paper recommendations are even stricter and advocate the use of Holter ECG when spontaneous syncope occurs daily. 10 It seems that physicians are quite reluctant to follow these recommendations probably because they hope that even if Holter ECG will not be diagnostic (i.e. no syncope during recording), it may give some clues as to the possible mechanisms of syncope and direct further diagnostic tests. We found a surprisingly low rate of CSM performing in the studied patients-only 16% of patients with indications for CSM actually underwent this test. The explanation may be that some physicians are afraid of performing CSM because of the fear of complications. Another explanation is that in many centres, CSM is performed directly after completing TT, using the same noninvasive equipment and tilt table. In case of a positive result of TT, CSM is deemed unnecessary because diagnosis has been already established. Thus, there may be a need for clarification in the guidelines whether CSM should also be performed in those in whom TT was positive, especially when syncope was associated with injury because there is a high likelihood that syncope was due to long pause/asystole and a pacemaker may be needed. Also the exact place of CSM in the syncope work-up scheme should be established more clearly-should it be always performed directly after non-diagnostic initial evaluation or at later stages (i.e. after TT)? Lastly, in many patients, CSM such as those with a strong suspicion of arrhythmic syncope (pacemaker dysfunction, sustained tachyarrhythmias, drug-induced proarrhythmia, and others), CSM was deemed unnecessary; however, according to the guidelines, it should had been performed.
The underutilization of ET in patients with exercise-associated syncope is difficult to explain. Clearly, all these patients should undergo ET. Also the overuse of neurological consultation is striking and may be due, at least in part, to the willingness for a formal exclusion of neurological nature of syncope by another specialist.
Another reason for overutilization of some tests (TT and prolonged ECG monitoring) is the way we defined the appropriateness of performing a given test. We took into account all tests performed in an individual patient both in the past and during index hospitalization. Table 3 showed that a significant proportion of these inappropriately performed tests had been done in a local centre, before index hospitalization in a tertiary centre. Thus, the overutilization rate of such tests as TT or prolonged ECG monitoring in the tertiary centres is lower than the overall result reported in this manuscript.

Diagnostic efficacy
The final diagnosis was established in 91% of the patients which is only a fraction lower than a 94% rate of correct diagnosis reported by Brignole et al. 2 Of note, a 91% diagnostic efficacy was achieved without the use of ILR. However, in our study, the mean number of tests required for obtaining diagnosis was much higher (4.6 + 2.6 when taking into account all diagnostic procedures or 2.3 + 1.6 when considering only those performed in a tertiary centre) than in the Brignole et al. 2 study (1.9 + 1.1). This result corresponds to the over usage of some tests in our registry and shows that we were less efficient in educating physicians participating in our study than Italian investigators.
Using the most conservative indication, ILR insertion at the end of a non-diagnostic work-up and severe or possibly arrhythmic syncope, 6% of all syncope patients were amenable for the device. It has been well established that ILR has a high diagnostic accuracy, usually oscillating around 50%. 17 -20 The place of ILR in the diagnostic tree is still under debate. In the majority of the studies, it was used as a final diagnostic resort in patients with unexplained syncope, 19,20 whereas the results of the ISSUE-2 study showed that it may be applied at the earlier stages of the diagnostic work-up. 16 We used the most conservative criterion in order to assess the minimal requirements for the ILR insertion in Poland.
It has been reported that 34 implantations/million population/ year are required in Italy. 17 This computation was based on the finding that 28% of the patients with unexplained syncope finally received an ILR. In our study, 42 (72%) patients with unexplained syncope had symptoms severe enough to proceed with a diagnostic work-up and ILR implantation. This gives the number of 222 ILR implantations/million inhabitants/year. Presumably, these numbers would become even higher if ILR was indicated at the earlier stages of the diagnostic work-up as suggested by some studies. 16 The much higher requirement for ILR suggested in our registry than that showed by Solano et al. 17 may be due to the differences in the studied populations. As previously mentioned, in our study the proportion of patients with cardiac syncope and presumed arrhythmic cause was markedly greater than in other studies, which may explain the 72% rate of patients with unexplained syncope requiring ILR compared with a 28% rate reported by others. 17 It has also to be mentioned that computations based on the diagnoses established in the emergency room and taking into consideration the population of the district of a given hospital may be subject to significant errors. First, initial diagnosis made in the emergency department is not always correct (syncope or faint may be one of several symptoms, not always the most important). Secondly, not all patients admitted to the tertiary electrophysiological cardiac centres are inhabitants of the district which is formally covered by the hospital. Thus, the number of 222 implantations/year/million inhabitants as the requirement for ILR in Poland may be overestimated. Thirdly, the definition of the need for ILR, used in the present study, was arbitrary, and if we used other definitions, the results could be different.

Limitations of the study
Although the physicians participating in the registry were instructed to include only patients with syncope, we cannot exclude the possibility that a few patients had other forms of transient loss of consciousness. Differentiation between true syncope and other forms of transient loss of consciousness is sometimes difficult in real life.
Investigators were asked to include all consecutive patients with syncope; however, no formal log-book of all patients admitted to the departments participating in the registry was used. Therefore, we cannot be sure that the registry really covers all consecutive patients with syncope.

Conclusions
Patients with syncope admitted to the tertiary electrophysiology/ haemodynamic cardiac centres are a highly selected group of patients with syncope and differ in their characteristics as well as underlying diseases than those managed at general hospitals, outpatient clinics, or special syncope units. In Poland, the adherence to the ESC guidelines on syncope is far from satisfactory, although the rate of obtaining final diagnosis is high. At least 6% of all consecutive patients with syncope are candidates for ILR insertion.