In patients with syncope the presence of structural heart disease (SHD) is considered the main predictive factor for an arrhythmia with a sensitivity of 95% and a specificity of 45%. Its absence makes an arrhythmic cause of syncope highly unlikely.1 2 However, this conclusion is true only in patients with a diagnosis derived from conventional examinations. In patients with syncope and a negative clinical investigation it is still unclear if the presence of SHD predicts an early recurrence of syncope. In these patients, the implantable loop recorder (ILR) is a useful diagnostic tool, although data comparing patients with and without SHD are limited.3 4 It is reasonable to assume that apart from SHD other risk factors trigger early recurrence of syncope. Furthermore, the mechanism of syncope recorded with the ILR is extremely heterogeneous and a wide variety of rhythm disturbances have been recorded at the time of syncope. This has been recognised in the ISSUE (International Study on Syncope of Uncertain Etiology) classification, which proposed a mechanism for the recognised rhythm detected with the ILR during syncope.5 We hypothesised that the presence of SHD is not a significant predictor for early recurrence of syncope and other risk factors are more important. We also assumed that the ISSUE classification may markedly help for precise diagnosis and treatment. Therefore, to define an acceptable standard in clinical practice, further evaluation of the ISSUE classification is mandatory.

PATIENTS AND METHODS

Study design and study group

This was a single-centre, prospective study performed in patients referred for unexplained syncope. All patients had at least two syncopal episodes before ILR implantation. The diagnosis of unexplained syncope was established using current guidelines.1 According to the ISSUE II study, structural heart disease was defined as overt heart disease at risk of arrhythmia. Our study included patients with previous myocardial infarction or cardiomyopathy with depressed left ventricular ejection fraction or non-sustained ventricular tachycardia in whom an electrophysiological study was not suspicious of a conduction problem and did not induce sustained monomorphic ventricular tachycardia, or was not performed owing to non-specificity. None of the patients had overt valvular heart disease. All subjects gave written informed consent.

Diagnostic investigation

The initial clinical assessment included a careful history, physical examination, neurological investigation, ECG, carotid sinus massage (carotid sonography to exclude significant cerebral artery disease), postural blood pressure testing, echocardiography and 24-hour ambulatory ECG. In all patients with SHD a coronary angiography was performed. Diagnostic investigation was performed in accordance with ESC guidelines on management and treatment of syncope.1 Psychiatric evaluation and diagnosis were obtained according to the international classification of disease (10th revision) (ICD-10) criteria for somatisation, panic attacks, anxiety and major depressive disorder.6

ILR specification, implantation and follow-up

In all patients a Reveal Plus ILR (Medtronic, Inc, Minneapolis, MN, USA) was implanted under local anaesthesia subcutaneously in the left parasternal region. Preimplantation mapping was performed in the left upper parasternal region to optimise device sensing of R waves and minimise T-wave amplitude. The parameters for automatic detection were set to store pauses of ⩾3 seconds and a heart rate of ⩽40 or ⩾160 beats/min. Patients and relatives were instructed to activate the device after every episode of syncope or presyncope with palpitations. Patients were seen at the outpatient clinic at least every 3 months and were followed up until a diagnosis or the end of life of the ILR was reached. Thereafter, the device was explanted.

End points, classification

The primary end point of the study was the electrocardiographic diagnosis made by the analysis of the electrocardiographic tracing obtained during a syncopal episode that was correctly recorded by the device. Syncope and presyncope with palpitations were considered specific. In contrast, presyncope alone was considered non-specific.7 8 In particular, based on the results of the ISSUE study the mechanism of syncope was classified as uncertain mechanism, neurally mediated and primary cardiac arrhythmia.5 9^–11 If battery depletion occurred before documentation of a syncopal relapse, patients were not asked to undergo a second ILR implantation.

Statistical analysis

All continuous variables are given as mean(SD) and were compared using the Student t-test or Wilcoxon non-parametric test, as appropriate. Nominal variables were compared using contingency tables in combination with a χ² test. The Kruskal–Wallis test was used for comparison of three dependent or independent variables. The Bonferroni correction for multiple-comparison correction was used when several dependent or independent tests were performed. Comparisons between Kaplan–Meier survival curves were made using the log-rank test. Univariate analysis and stepwise multivariate analysis were performed to identify statistically significant predictors for early recurrence with the following variables: body mass index, sex, age, left ventricular ejection fraction, syncope with trauma, number of ILR implantations before syncope, β-blockers, antiarrhythmic agents, ACE inhibitors, calcium antagonists, diuretics, selective serotonin reuptake inhibitors, hypertonia, major depressive disorder, diabetes, epilepsy and stroke. Statistical significance was assumed if the null hypothesis could be rejected at the 0.05 probability level. Software: SAS version 9.3.1, Cary, NC, USA.

RESULTS

Main characteristics

Seventy patients (36 women and 34 men, aged 55 (17) years, range 25–79) with recurrent syncope were enrolled and divided into two groups: 33 patients (47%) with structural heart disease with a mean (SD) number of episodes of 2.4 (1.1) composed the SHD group (ischaemic cardiomyopathy in16, dilated cardiomyopathy in 9 and hypertrophic cardiomyopathy in 8). Thirty-seven patients (53%) with a mean (SD) number of episodes of 5.2 (2.6) had no underlying heart disease and composed the no-SHD group. Patients differed significantly in left-ventricular ejection fraction (46 (3.6)% vs 61.4 (6.9)%, p<0.001). Medical history, physical and standard examination were not diagnostic of a predisposing condition.

Table 1 illustrates the type and number of diagnostic procedures performed at the discretion of the referring doctor in each patient to exclude a cardiac or neurological cause of syncopal events. Differences in type of comorbidities showed a higher rate of treated hypertension in the SHD group. Following this, there were differences in baseline drug treatment (β-blockers and calcium antagonists) between the two groups. No differences where found in baseline ECG.

Follow-up results based on ISSUE classification

The mean (SD) ILR follow-up time was 16.9 (7.7) months (SHD group) and 15.4 (8.6) months (no-SHD group), respectively. Syncopal and presyncopal recurrences during follow-up were almost equally distributed in patients with and without SHD: syncope 82% versus 73%, and presyncope with palpitations 9% versus 8%. An event–rhythm correlation was possible in 91% of cases in the SHD group and in 81% of cases in the no-SHD group (table 1). Issue classification type 1 arrhythmia was more frequent in the no-SHD group, but without statistically significant difference. ISSUE classification types 2 and 4 were equally distributed in both groups. No or slight rhythm variations (ISSUE classification type 3) were more common in the SHD group (p = 0.04). No or slight rhythm variations were most commonly found, followed by asystole, tachycardia and bradycardia. (table 2) Overall, the mechanism of syncope was similar in patients with syncope and in those with presyncope with palpitations (six patients: four, one and one patient with ISSUE classification types 3, 1 and 4, respectively).

ISSUE classification and clinical consequences

Based on ISSUE classification of ECG-documented spontaneous syncope “uncertain mechanism” was most commonly diagnosed (⩾40% in both groups) followed by neurally mediated syncope (approximately one-third in both groups) and syncope due to primary cardiac arrhythmia. Thus, in total a rhythm–symptom correlation was made in 60 patients (86%). An ILR-based treatment was prescribed in 34 patients (49%): pacemaker implantation (23%), β-blocker (19%) and amiodarone/catheter ablation (7%). There was no significant difference between the two groups (table 2).

Cox-hazard regression and survival analysis

Univariate analysis identified major depressive disorder and diuretics as significant predictors for recurrence of symptoms. On stepwise multivariate analysis major depressive disorder was the only significant predictor for recurrence of symptoms during ILR follow-up (p = 0.01, hazard ratio 3.35 (95% CI 1.1 to 7.1)) (table 3).

Based on ISSUE classification patients were stratified according to major depressive disorder (table 4). “Uncertain mechanism” was more common in patients with major depressive disorder (p = 0.001) than in those without depression. In contrast, “neurally mediated syncope” was significantly more common in patients without major depressive disorder (p = 0.003). Significantly more patients with major depressive disorder received “no treatment”, whereas “pacemaker treatment” was significantly more common in patients without major depressive disorder (p = 0.002 and p<0.001, respectively; table 4).

During follow-up (16.1 (8.3) months, interquartile range 2–23), an ECG-documented recurrence occurred in 60 patients with an actuarial rate of 30% at 3 months, 65% at 12 months and 91% at 24 months in the SHD group and with an actuarial rate of 35% at 3 months, 68% at 12 months and 87% at 24 months in the no-SHD group.

The Kaplan–Meier recurrence curves (fig 1) showed no difference between the two groups (log-rank test, p = 0.803). Irrespective of an underlying heart disease patients with major depressive disorder had their syncopal recurrence significantly earlier than patients without major depressive disorder (log-rank test, p = 0.028; fig 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 After 3, 12 and 24 months of follow-up 30%, 65% and 91% of patients with structural heart disease (SHD) and 35%, 68% and 87% of patients with no-SHD, respectively, had a syncopal recurrence (p = NS).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Syncopal recurrence stratified according to an underlying major depressive disorder. After 3 months of follow-up, 78% of patients with major depressive disorder but only 26% without major depressive disorder had at least one recurrence (p = 0.028).

A Kruskal–Wallis test for comparison of recurrence between patients with and without major depressive disorder was used. Asystole, bradycardia and tachycardia (ISSUE classification types 1, 2 and 4), sinus rhythm (ISSUE classification type 3) and patients with no events were compared. The two ISSUE classification groups differed significantly for asystole, bradycardia and tachycardia (p<0.001), sinus rhythm (p = 0.01) but not regarding patients with no events (p = 0.053; fig 3). Fifty-seven per cent of all patients with major depressive disorder had sinus rhythm during recurrence compared with 31% of patients without (p = 0.01). Conversely, no patient with major depressive disorder had asystole compared with 33% without (p<0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Comparison of recurrence between patients with and without major depressive disorder: asystole, bradycardia and tachycardia (ISSUE classification types 1, 2 and 4), sinus rhythm (ISSUE classification type 3) and patients with no events. The two groups differed significantly for asystole, bradycardia and tachycardia (p<0.001), sinus rhythm (p = 0.01) but not regarding patients with no events (p = 0.053).

Adverse events

During follow-up no patient died. Three syncope-related traumatic episodes occurred in two patients. No patient was permanently handicapped after syncope-related trauma.

DISCUSSION

The major finding of our study is that in unexplained syncope ILR-documented asystole, bradycardia and tachycardia events are equally distributed in patients with and without SHD. An underlying major depressive disorder was found to be a predictor for early syncopal recurrence without significant arrhythmias in most cases.

In this prospective, single-centre study our study group represents a well characterised sample of consecutive patients with unexplained syncope and implanted ILRs. We report a high rate of syncopal recurrences in patients with and without SHD: 77% of all patients had at least one real syncope during a mean ILR follow-up time of 16 months. When considering also presyncope with palpitations, an event–rhythm correlation was possible in the majority (86%) of patients. This recurrence rate is higher than reported in some previous studies with more heterogeneous patients in which an ILR diagnosis ranged from 17% to 58%.5^–7 9 10 Our recurrence rates are in line with smaller studies by Menozzi et al11 and Mason et al.3 We are in line with Nierop et al,12 who concluded that the ILR made a symptom–rhythm correlation possible in 83% of patients with recurrent syncope. Farewell et al demonstrated that the ILR significantly increased the rate of diagnosis in an unselected Western population with recurrent syncope.13 14 Finally, Krahn et al concluded that the aetiology of recurrent syncope is diverse and cannot be predicted by baseline clinical variables.15 Patients with structural heart disease were less likely to experience recurrent symptoms during monitoring. Only advanced age was associated with earlier recurrence of symptoms. This is in contrast to our findings where age was not associated with earlier recurrence.

Syncope itself does not necessarily herald a bad prognosis. It has been demonstrated that in patients with SHD who were inducible on electrophysiological study, syncope and sustained ventricular tachycardia do not necessarily correlate during follow-up.16 17 Electrophysiological testing has its role in patients with SHD. Solano et al demonstrated that more than 50% of patients without electrophysiological testing had a primary cardiac arrhythmia during their syncopal recurrence.4 In our study the ILR documented a low incidence of tachyarrhythmias in 13.5% of the patients with SHD and in only 3% of patients without SHD. The low incidence of cardiac arrhythmias in patients with SHD confirms the results of previous studies which evaluated patients with SHD and a negative electrophysiological study.9 11

The main objective of the diagnostic investigation in unexplained syncope is to identify the underlying cause. Therefore, an electrocardiographic classification of spontaneous syncope documented by an ILR was proposed in 2005.5 This classification is based on the results of the International Study on Syncope of Uncertain Etiology (ISSUE). In the present study we used this classification in a series of consecutive patients drawn from a single university cardiac centre. We demonstrated that asystole, bradycardia and tachycardia events are equally distributed in patients with and without SHD.

Thirty per cent of our study patients had a major depressive disorder. This is in line with findings of Kapoor et al,18 who showed that the prevalence of psychiatric illness markedly differs between patients with unexplained (25%), non-cardiac (19%) and cardiac (10%) syncope. Moreover, Kouakam et al reported a prevalence of anxiety and panic in patients with syncope of 30% and 20%, respectively.19 The reported syncopal recurrence rate was higher in patients who fulfilled criteria for affective disorders. This relationship between unexplained syncope and psychiatric illness was also described by Linzer et al.20 In our study major depressive disorder was found to be the only predictive factor for early syncopal recurrence. This is in line with Giada et al who studied patients with syncope and positive tilt testing and healthy controls. Psychiatric disorders were more common among patients than controls. Psychiatric disorders also predicted the risk of recurrence.21

Anxiety and major depressive disorders exert a well-known negative influence on cardiovascular disease and are often unrecognised and not effectively treated.22 23 Engel postulated in 1978 that psychiatric disorder must be considered as an additional risk factor in the presence of factors lowering the threshold for conduction disturbances. This may result in vasodepressor syncope, benign arrhythmias, or even in dangerous arrhythmias and sudden cardiac death.24 Moreover, major depressive disorder was associated with an almost fourfold increase in the relative risk for arrhythmias in the Northwick Park study.23 Although we found a high early recurrence rate of symptoms in patients with major depressive disorder, there was no evidence of arrhythmia in most cases. In contrast to earlier studies, our patients were treated exceptionally with selective serotonin reuptake inhibitors. Treatment with these inhibitors does not have side effects on cardiac impulse conduction. Di Girolamo et al demonstrated in a placebo-controlled study that the selective serotonin reuptake inhibitor paroxetine significantly improved the symptoms of patients with vasovagal syncope.25 This is in contrast to tricyclic antidepressants, which were not used in our study patients.26 Finally, it must be speculated that many syncopal episodes just seem to be a real syncope, since infrequently witnessed. Following this, we think that transient loss of consciousness rather than real syncope is most common among patients with major depressive disorder. Real orthostatic intolerance for ISSUE classification types 3A and 4A seems unlikely and could not be proved. Thus, we relabelled the ISSUE classification type 3A and 4A to “uncertain mechanism”.

The question as to whether psychiatric disorders are the cause or effect of syncope/transient loss of consciousness has not been resolved.27 Deharo et al showed that the heart rhythm observed during spontaneous syncope in vasovagal patients with ILRs does not correlate with head-up tilt test.28 Garcia-Civera et al demonstrated that a negative electrophysiological study and head-up tilt test is of little value to predict the mechanism of syncope.29 Thus, a strategy based on early application of ILRs allows a specific and effective treatment in patients with unexplained syncope.30 This watch and wait strategy with ILRs also appears to be safe, since no patient was handicapped after a traumatic syncope during follow-up in our study. Because syncope may be due to multiple causes in a single patient, monitoring should be continued after the first episode when no arrhythmia is recorded.

Limitations

The power of statistical analysis is related to the non-randomised design and the number of study patients. The prevalence of mixed vasovagal syncope may be underestimated. A population study showed that only 23% of patients with this type of syncope are referred to the hospital.31 For these reasons, the results of the present study reflect the distribution of patients referred for syncope investigation, rather than the distribution of patients with unexplained syncope in the general population. Furthermore, the ISSUE classification must be handled with caution, because blood pressure was not measured during the syncopal recurrence. Thus, real “orthostatic intolerance” cannot be measured. Nevertheless, this classification is an effective tool for discriminating between different suspected mechanisms of syncope. Finally, potential interferences from electronic article surveillance devices can cause artefacts on ILR tracings.32 33

CONCLUSION

Our results demonstrate that application of an ILR leads to a clear diagnosis of heart rhythm during symptoms in the majority of patients irrespective of an underlying heart disease. The mechanism of unexplained syncope and the diagnostic yield are similar in patients with and without SHD. Based on the documented heart rhythm during recurrence, half of all study patients received specific treatment. In patients with numerous syncopal and presyncopal events, psychiatric evaluation should be encouraged. In most of these patients a transient loss of consciousness rather than a real syncopal recurrence with a low incidence of rhythm disturbances can be expected. To exclude a potential arrhythmia selected patients with refractory transient loss of consciousness could represent an expansion of the indications for loop recorder implantation.