Role of extended external loop recorders for the diagnosis of unexplained syncope, pre-syncope, and sustained palpitations

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Keywords


Introduction
The diagnosis of syncope and sustained palpitations of suspected arrhythmic origin remains a difficult task in clinical cardiology, often leading to multiple admittances to the emergency room and repetition of different diagnostic tests. The clinical presentation is often not univocal, as the history of pre-syncope, syncope, and sustained palpitations coexist in the same patient.
All the recent guidelines for the management of patients with syncope, palpitations, ventricular arrhythmias and sudden death, supraventricular arrhythmias, and atrial fibrillation recommend the use of prolonged electrocardiogram (ECG) monitoring techniques to allow a better correlation between the symptoms and the arrhythmias, and to detect asymptomatic significant rhythm disturbances. 1 -6 In the diagnostic work-up of the syncope, external or implantable ECG recorders are specifically recommended in patients with a high pre-test probability of arrhythmic syncope. 1,2 The choice of monitoring technique mainly depends on the predicted recurrence rate. Routine standard 24 h Holter recorders have a low diagnostic yield as they are unlikely to capture the reoccurrence of syncope, unless in extremely frequent episodes, while it may contribute to identifying ECG markers suggestive of arrhythmic origin. Implantable loop recorders (ILR) have a higher diagnostic yield for infrequent syncope; therefore, a more extensive early usage in the initial phase of the diagnostic work-up may be recommended. 1,7 -9 However, ILR are expensive and mildly invasive, therefore unfit as first level diagnostic tools in general clinical practice. 10 -12 The role of external loop recorders (ELRs) in the diagnostic workup of the syncope is more debated, as the available studies showed conflicting results, with diagnostic yield spanning from 10 to 50%, mainly due to the duration limited to about 1 month. 2,4,13 -15 Thus, current indications for ELR are limited to patients with a high probability of recurrent events. 2,4 However, technological advances such as auto-trigger capability and mobile cardiac outpatient telemetry are now showing promising results. 15 -17 In sustained palpitations, the indication for ELR is more clearly established, since several studies showed that 50-75% of sustained palpitations recurred within 1 month. 5,18 However, most available ELR have technical limitations, due to relatively low-memory capacity and lack of auto-trigger function.
The main goal of this study was to evaluate the diagnostic yield of a new generation of ELR with high-memory capacity, extended backward and forward memory, and auto-trigger functions in consecutive patients with syncopal and/or pre-syncope or palpitations.

Study population
Since January 2005, we have established a registry that prospectively included all consecutive patients with a history of syncope (fainting with total loss of consciousness, TLOC), pre-syncope (fainting without TLOC), or sustained palpitations referred to the Non-Invasive Electrophysiological Unit of our Department, who underwent ELR monitoring by SpiderFlash recorders. The registry included a total of 307 consecutive patients receiving ELR between January 2005 and September 2012. All the patients had previously completed a routine workup for syncope or palpitation, as recommended by the ongoing guidelines at the time of monitoring (including tilt test, sinus carotid massage, and electrophysiological test whenever appropriate), in all cases not conclusive for diagnosis. All the patients had received one or more 24 h Holter monitoring (median 3 recordings each, range 1 -18 recordings), while only one patient with a history of syncope had previously had ILR before ELR.
As most patients had a history of multiple and heterogeneous symptoms (syncope and/or pre-syncope and/or palpitations), we identified two study groups. The first group consisted of patients with a history of at least one confirmed unexplained syncope (92 patients, 30%). The second group consisted of patients with a history of palpitations and/or pre-syncope (215 patients) ( Figure 1).

Recording techniques
We utilized SpiderFlash-A w (SFA w , Sorin CRM) and SpiderFlash-T w (SFT w , Sorin CRM) digital ELRs, storing one-or two-lead ECG tracings on a high-capacity removable secure digital card (which can memorize a virtually unlimited number of events), with looping memory capabilities, intended for long-term ECG monitoring up to 30 days.
SpiderFlash recorders are programmed by Hook-up2 w (v2.00) software, allowing the choice of several recording and pre-analysis parameters (such as pre-and post-event recording duration, time-table for pre-defined recordings, and type and characteristics of arrhythmias for auto-trigger functions, if available).
The first generation SFA w recorders (utilized in the period 2005 -09) had two recording modalities: (i) patient-activated and (ii) programmable time-table. In patient-activated modality, continuous one-or two-lead ECG tracings were recorded by utilizing loop-recording technique, with backward and forward memory, which can be extended up to 30 min (15 min before and after the trigger). In programmable time-table modality, it is possible to memorize a maximum of 20 ECG tracings per day of maximum 900 s duration.
The second generation SFT w recorders, available since 2009, besides (i) patient-activated and (ii) programmable time-table recordings, also have (iii) auto-trigger capability which automatically detects and records pre-defined and programmable rhythm disturbances such as pauses, bradycardia, or supraventricular and ventricular tachyarrhythmias. SFT w recorders were programmed to identify rhythm disturbances according to the parameters listed in Table 1.
SFA w and SFT w recorders (sized 75 × 50 × 19 mm) were carried by the patient in a disposable bag hung around the neck. The recorders were connected to the thorax by lead wires and disposable adhesive electrodes, which the patients were trained to change daily for personal cleaning. The patients had to press a button on the recorder to store ECG tracings in case of symptoms ('patient-activated' events); otherwise no other manipulation was required. The patients were asked to annotate on a special diary all the symptoms (palpitations, syncope, or pre-syncope) occurring during the monitoring period. The patients were instructed to return the recorder when the recording ended (shown by a status light-emitting diode on the device) or when a significant symptom occurred. If the recorder stopped before the expected minimum of 21 days, and no symptoms had occurred, the recorder was restarted and the monitoring

Analysis technique and quality control of the recordings
Analyses were performed by EventScope2 w (v2.00), software dedicated for analysis of SFA w and SFT w recorders. The auto-trigger detection algorithm was validated in a dedicated study. 19 As the first step, the recordings were screened for quality of the tracing and real use by the patient by means of the analysis of the lead impedance. No monitoring had to be excluded due to insufficient quality, since in most tracings at least one of the two leads was adequate for arrhythmia analysis, even if in each monitoring some tracings had to be excluded due to motion artefacts. As the second step, all patient-activated tracings were checked and matched to the symptoms reported in the diary to verify the correspondence between the symptoms and the tracings and between the symptoms and the arrhythmias. As the third step, in SFT w only, all auto-triggered tracings were examined and whenever possible matched against the symptoms annotated in the diary. The number of available tracings for each recording was rather variable according to the patient characteristics and the presence or the absence of symptoms, with a median of about 1000 tracings memorized for each recording (about 50 tracings per day).

Statistical analysis
All the data with normal distribution were expressed as mean + SD, and were compared by standard t-test wherever applicable. Data without normal distribution were tested by using non-parametric tests.
The reference standard for calculating the ELR diagnostic yield was 24 h Holter monitoring, which was negative or non-conclusive in all study patients. Wherever appropriate, the diagnostic yield of SFT w (with auto-trigger capability) was compared with SFA w (without autotrigger capability). In all comparisons, P value ,0.05 was considered significant.

Results
The baseline patient characteristics are illustrated in Table 2. The mean age was 58 years (range 8-94 years) and female gender was prevalent (61%). The mean duration of the recordings was 24.1 + 8.9 days (SFA w 25.2 + 10.2 days-SFT w 22.2 + 5.8 days, NS, Table 2). In 85% of the cases, the monitoring lasted 3-5 weeks, while in 47 cases (15%) the monitoring lasted 2 weeks or less ( Figure 2). Among those cases, ELR monitoring was interrupted prematurely due to the occurrence of significant symptoms (38 of 47 patients, 81%), while in the remaining 9 cases the recording was stopped due to skin reaction to the electrodes or due to patient lack of compliance. Events occurred almost evenly in each monitoring week, both for syncope and pre-syncope or palpitations.
As most patients had a history of multiple and heterogeneous symptoms (syncope and/or pre-syncope and/or palpitations), we separately analysed patients with a clear-cut history of at least one confirmed unexplained syncope from patients with more heterogeneous symptoms of palpitations and/or pre-syncope.

Patients with a history of syncope
Among the 92 patients with syncope (54 studied by SFA w and 38 studied by SFT w ), about 30% had a history of ischaemic or nonischaemic heart diseases (valvular or congenital heart disease), about 16% had a history of supraventricular arrhythmias, about 16% had conduction disturbances [first and second degree atrioventricular block (AVB), Mobitz 1 type, or left bundle branch block (LBBB) or right bundle branch block]. No statistically significant differences in the clinical baseline characteristics were found between patients studied by SFA w and SFT w .
Among them, a typical syncope reoccurred during monitoring in 17 patients. Electrocardiogram recording during syncope was available in 16 patients, while just 1 patient (male gender, 78 years) with SFA w did not activate the manual recording after recovering from syncope ( Table 3).
Among 38 patients studied by SFT w with auto-trigger capability, asymptomatic suggestive arrhythmias [pauses, bradycardia, AVB, and sustained PAF/PSVT/non-sustained ventricular tachycardia (NSVT)] were de novo detected in 11 patients (28.9%), indicating a possible arrhythmic cause for syncope ( Table 5), leading to a major medical treatment [pacemaker implantation in 3 cases and radiofrequency (RF) ablation in 1 case] while in 2 cases ILR was later implanted for further ECG monitoring.
No differences in history of cardiac disease or arrhythmias, or cardiac therapy at the time of recording were observed in patients with or without recurrence of syncope during the monitoring period.  Patients with a history of pre-syncope and/or palpitations Among 215 patients with a history of pre-syncope and/or palpitation, about 18% had ischaemic or non-ischaemic heart diseases, about 28% had a history of supraventricular arrhythmias, and about 9% had conduction disturbances. No statistically significant differences were found between patients studied by SFA w and SFT w , although patients with PAF were slightly more frequent than in most recent patients. Among them, typical palpitation or pre-syncope reoccurred during ELR monitoring in 184 patients (85.6%) ( Table 3). Many patients had multiple episodes of pre-syncope or palpitations during recording (median 3 episodes per patient). Rhythm disturbances observed during the symptoms are listed in Table 6 ( Figure 4). As many patients had multiple episodes with multiple rhythm disturbances, the patients were classified according to the most severe arrhythmia observed during monitoring [e.g. if a patient had one palpitation with recording of sinus tachycardia and one with PAF, the final classification was PAF].
In SFA w recordings, sinus rhythm or sinus tachycardia was recorded in about one-third of the cases, and about one-third had sustained supraventricular tachycardia or PAF at the time of the symptoms. In the SFT w recordings, supraventricular tachycardia or PAF was recorded in about 46% of the cases, while bradycardia or pauses was recorded in about 13% of the cases.
Those differences in ECG findings during the symptoms observed by SFA w and SFT w ( Table 6) can be partly explained by the slight prevalence of pre-syncope among patients studied by SFT w (28 vs. 21%, NS), accounting for a slightly higher incidence of bradycardia and pauses. However, when patients with a history of pre-syncope or with palpitation (51 vs. 164 patients) were analysed separately, no significant differences in the recurrence of symptoms or in the arrhythmia recorded were observed.
Among patients with palpitation studied by SFT w , asymptomatic arrhythmias were de novo detected by the auto-trigger function in seven cases (pauses in one case, PSVT in five cases, and NSVT . 5 beats in one case) ( Table 3), suggestive of possible arrhythmic origin of palpitations/pre-syncope, although the clinical relevance of such asymptomatic arrhythmias remains to be determined.  No statistically significant differences were found between patients studied by SFA w and SFT w . b These categories are not mutually exclusive and some patients may have more than one condition.

Discussion
In this prospective, single-centre, observational study of patients with a history of syncope and pre-syncope or palpitations remaining of unknown origin despite the recommended routine follow-up, the diagnostic yield of new ELRs with extended backward memory and auto-trigger function was 86% in patients with a history of palpitations or pre-syncope and 30% in patients with a history of syncope, when considering both symptomatic and asymptomatic significant arrhythmias. The diagnostic yield of the ELR was higher than standard 24 h Holter monitoring, and similar to diagnostic yield of the ILR considering the same timeframe. 7 -9 Patients with syncope The diagnostic yield for the syncope of ELR monitoring was 17.4%, when compared with standard Holter monitoring, mainly due to a longer monitoring period (median 25 days vs. median 24 h). The diagnostic yield of both SFA w and SFT w was consistently higher than the ELR of the previous generations, 4,13 -15 thanks to the extended backward memory period (up to 15 min), which allowed the activation of recording after recovering from syncope, even without auto-trigger activation. The diagnostic yield for syncope was slightly higher with SFT w than in SFA w (21.1 vs. 14.8%, NS, Table 3). It is possible that such difference reflects some changes in the characteristics of patients referred to the registry, thanks to the establishment in 2009 of a Syncope Unit in our Hospital, leading to a stricter implementation of the ECG criteria for possible arrhythmic origin of syncope, following the Guidelines for Management of Syncope. 1 The recurrence rate for syncope was similar to that recently reported in the PICTURE registry including patients with unexplained syncope studied by ILR (19% at 3 months). 9 Thus, extended ELR and ILR (both with auto-trigger capability) had similar diagnostic yields, when considering the same time interval.
An arrhythmic origin for the syncope was found in 7of 16 patients (44%), leading to symptomatic treatment with pacemaker in 3 cases with significant pauses, and ablation or change in medication in 4 patients with fast supraventricular arrhythmias ( Table 4). In the remaining cases, an arrhythmic origin of syncope could be excluded,

Figure 3
Auto-trigger activation due to pauses during syncope. Auto-trigger activation due to pauses by SFT w during traumatic syncope (occurring at Day 5 of monitoring), with recording of 60 s asystole during atrial fibrillation in a female patient, age 70 years, with a history of hypertension and chronic atrial fibrillation LBBB and previous traumatic syncope. Full disclosure of the event (30 s by line). The patient had a pacemaker implant (January 2011) and had been asymptomatic so far.
Diagnosis of unexplained syncope and palpitations by extended ELRs and the patients were addressed to further non-cardiological investigations. As shown in Table 5, asymptomatic suggestive arrhythmias (pauses, bradycardia, AVB, and sustained PAF/PSVT/NSVT) were de novo detected by SFT w in 11 out of 38 of the patients (28.9%). Asymptomatic arrhythmias were not taken as equivalent of syncope recurrence, even if they were an important finding in diagnostic workflow, leading to major medical treatment (pacemaker implant or RF ablation) or to ILR implant for further ECG monitoring.
When considering altogether the conclusive and the suggestive cases, the diagnostic yield for the syncope of ELR with extended backward memory and auto-trigger capability was comparable with ILRs, even if considering the shorter monitoring period.

Patients with palpitations or pre-syncope
In most patients, palpitations or pre-syncope recurred during monthly monitoring time. The recurrence rate of palpitations or presyncope was much higher than the recurrence rate of syncope (86 vs. 17%, P , 0.0001). When considering both symptomatic and asymptomatic arrhythmias, the diagnostic yield was 90% when compared with standard Holter monitoring. Such a result was higher than the 66 -75% diagnostic yield for palpitations reported in previous generation of ELRs, with shorter recording periods, lower memory capacities, and without auto-trigger capability. 4,13 -15 Most ECG recordings during symptoms were of good quality and generally onset and offset of arrhythmias were recorded. SpiderFlash-T w (SFT w ) recorders were more likely than SpiderFlash-A w (SFA w ) to detect bradycardia and pauses, brief PSVT, and NSVT, suggesting that ELR with auto-trigger capability had a superior capability of capturing transient rhythm disturbances (Table 6).
Following such findings, the extended ELR with auto-trigger capabilities should become the first-choice tool in the diagnostic flowchart of palpitations and pre-syncope. Only those patients whose palpitations were still unexplained after 1 month monitoring should undergo further investigation with more invasive and more expensive ILRs.

Recording of asymptomatic arrhythmias
The main clinical objective of ELR monitoring was to correlate the symptoms and the arrhythmias. However, both significant asymptomatic arrhythmias were de novo detected by the auto-trigger function in patients with syncope and palpitations and pre-syncope. De novo detection of asymptomatic arrhythmias should never be taken as equivalent of the symptoms. 1

Early use of external loop recorder after an event
Several studies have reported that syncope and palpitations tend to occur in clusters, with higher recurrence in the early phase after the first event, and the early initiation of the diagnostic work-up may increase the likelihood of recording recurrent episodes. 1,8,12 However, the early use of ILR is difficult to implement in real clinical practice, due to the high cost and minimally invasive implanting procedures. 11,12 Non-invasive and easy-to-use ELRs can be more simply provided to patients even in the early phase of the diagnostic workup. 4,5,13,23,24 To test this hypothesis, since 2011, an international prospective study (SYNARR, Monitoring of SYNcopes and/or sustained palpitations of suspected ARRhythmic origin with External Loop-Recorder SpiderFLASH) was initiated with the aim to evaluate the diagnostic yield of SpiderFlash-T enroling 372 patients from 10 centres from 5 European countries.

Future technological advances
New generation ELRs, with extended recording memory, autotrigger capability, and high-quality ECG tracings represent a major step towards the gold standard possibility to record and analyse fulldisclosure very long-term continuous ECG tracings, allowing not only a precise correlation between the symptoms and the arrhythmias, but also a quantitative evaluation of the arrhythmic burden, both for symptomatic and asymptomatic arrhythmias. This will be possible thanks to high-power long-lasting batteries and due to larger and low-consuming storage circuits and to remote or cloud storage systems.  Diagnosis of unexplained syncope and palpitations by extended ELRs Future technological advances need to overcome the main limitation of the current ELR, still relying on lead wires and adhesive electrodes, which assure a standard ECG quality but are poorly tolerated and self-limit the recording duration to 3 or 4 weeks. Possible solutions, which have already been suggested, are belt or patch electrodes or vest or wireless non-contact ECG electrodes. 25

Conclusions
In patients with a history of syncope and pre-syncope or palpitations remaining of unknown origin despite recommended routine followup, the diagnostic yield of new ELRs with extended backward memory and auto-trigger function was 86% in patients with a history of palpitations or pre-syncope and 30% in patients with a history of syncope, when considering both symptomatic and asymptomatic significant arrhythmias. In those patients candidate to longterm ECG monitoring, these results suggest that ELR rather than standard Holter monitoring or ILR should be used in the early diagnostic workflow. 26 -28 The use of ELR, sharing the favourable characteristics of standard Holter monitoring (non-invasive, inexpensive, easy-to-use in all age groups, reusable, and providing high-quality ECG tracings), could be viewed as a first-choice tool to filter patients with early recurrence of syncope or palpitations. In a step-wise approach, the more expensive and minimally invasive ILR could be reserved for those cases where a clinical diagnosis could not be obtained after 1-month ELR monitoring.