Data for this multicenter, retrospective, observational study of 173 patients who received an ILR at 11 hospitals in Korea were gathered from February 2006 to April 2018. Enrolled patients with recurrent US and negative conventional diagnostic examinations received an ILR (Reveal LINQ ICM, Reveal Plus, or Reveal DX-XT; Medtronic Inc., Minneapolis, MN, USA or Confirm^TM; St. Jude Medical, St. Paul, MN, USA). The initial examinations for syncope evaluation included careful history-taking (total number and duration of syncope, prodromal symptoms, syncope-related injury, underlying disease), physical examination, baseline 12-lead ECG, two-dimensional echocardiography, 24-hour Holter, treadmill test, head-up tilt test (HUT), carotid sinus massage, and orthostatic blood pressure measurement. Invasive diagnostic tools, such as an electrophysiological study and coronary angiography were performed if clinically indicated. When a neurologic cause was suspected, brain computed tomography or magnetic resonance imaging was performed. The conventional diagnostic tests performed for syncope evaluation before ILR implantation are shown in Supplementary Table 1.

Any hypertensive medication that could cause symptomatic bradycardia or hypotension, such as a beta-blocker or Ca blocker, was withheld before the conventional evaluation or ILR implantation at the physician's discretion to determine whether they were causing the syncope. If the conventional examination did not reveal the cause of syncope, an ILR was implanted. Some physicians implanted an ILR when the clinical syncope was atypical for reflex syncope or orthostatic hypotension, even when the HUT or orthostatic BP measurement was positive.

Prodromal symptoms were defined as autonomic activation that occurred before unconsciousness, such as dizziness, nausea, vomiting, sweating, palpitation, or chest or abdominal discomfort. Major syncope-related injuries were defined as a major bone segment fracture, such as a skull fracture causing intracranial hemorrhage, and tooth subluxation. Minor syncope-related injuries were defined as swelling, ecchymosis or bruising, abrasion or laceration, and external superficial hematoma or minor bleeding.

Any bundle branch blocks (BBBs) were defined as left bundle branch block (LBBB) or right bundle branch block (RBBB) by a baseline 12-lead ECG, following the recent American Heart Association/American College of Cardiology/Heart Rhythm Society recommendations.22 Structural heart disease was defined as coronary artery disease (CAD), heart failure (HF), cardiomyopathy (CMP), or moderate to severe valvular heart disease (VHD).

Each ILR was positioned subcutaneously in the left parasternal area at the level of the 4th–5th intercostal space under local anesthesia. The ILR stores rhythm strips when the patient presses a button in response to symptom occurrence (patient activation) or when the heart rate meets a preset limit (auto activation). The auto activation setting varied among patients, but it was generally programmed for a ventricular pause of > 3 seconds, a ventricular rate of < 40 beats/min, or a ventricular rate > 180 beats/min for more than 16 beats. Patients were taught to use the device prior to discharge after ILR implantation. Since small, insertable ILR devices (Reveal LINQ^TM ICM; Medtronic Inc., Minneapolis, MN, USA and Confirm Rx^TM; St. Jude Medical, St. Paul, MN, USA) came to the market in 2014 and 2017, respectively, ILR procedures have been simple and rapid and rarely entail major complications.23

Regular ILR follow-up at an outpatient clinic was performed at 3- to 6-month intervals at the physician's discretion. Immediate ILR interrogation was performed in the case of recurrent syncope. The ILR remained implanted until a diagnostic event was recorded or until the end of the battery life. If no diagnosis of syncope was made before the end of the battery life, removal or reinsertion of the ILR was performed at the physician's discretion.

The study endpoint was a symptom-correlated ECG diagnosis by ILR interrogation during the first recurrent syncope after implantation of the ILR. The secondary endpoint was significant arrhythmia irrespective of syncope, defined as a ventricular pause of > 3 seconds, a ventricular rate of < 40 beats/min, or a ventricular rate > 180 beats/min for more than 16 beats.

Continuous variables are expressed as the mean ± standard deviation. A χ² analysis was used for categorical data, independent t-tests for continuous data, and Mann-Whitney tests for nonparametric cases. Cumulative incidence and event-free curves were derived using Kaplan-Meier analyses, stratified by study groups, and compared using the log-rank test. To identify independent predictors for PM implantation, we first conducted univariate analyses and then included predictors with a significance level < 10% in a multivariable Cox proportional hazards model with a 95% confidence interval (CI). All P-values were two-sided, and statistical significance was accepted at P < 0.05. Data were analyzed using Statistical Package for the Social Sciences, version 11.0 (SPSS, Inc., Chicago, IL, USA) with Windows 2000 (Microsoft, Redmond, WA, USA).

This study was approved by each institution's Institutional Review Board (IRB) (Samsung Medical Center, IRB No. 2017-12-120). The requirement for informed consent was waived because we used only anonymized retrospective data that were routinely collected during clinical practice.

The study population is summarized in Fig. 1 and baseline characteristics of the study population are shown in Table 1. A total of 173 recurrent US patients (mean age, 67.6 ± 16.5 years; 62% men) were analyzed. The median number of previous cases of syncope was 3 (interquartile range, 2–5). Hypertension was present in 89 patients (51.4%), and a history of paroxysmal atrial fibrillation (AF) was noted in 41 (24%). Structural heart disease was present in 30 patients (17.3%): CMP in 13 (7.5%), CAD in 9 (5.2%), HF in 5 (2.9%), and VHD in 3 (1.7%). Significant CAD was re-vascularized. The VHD patients were 2 cases with moderate aortic valve stenosis and 1 with a well-functioning prosthetic valve. Those conditions were not directly associated with syncope.

Symptom-correlated arrhythmia and ILR-guided therapy are shown in Table 2. A total of 52 patients (30.1%) experienced a first recurrence of syncope after ILR implantation during the mean follow-up period of 9.1 ± 13.7 months. The cumulative incidence of recurrent syncope was 34.3% and 47.2% at 1 and 2 years, respectively (Fig. 2A). Among those patients, a symptom-correlated ECG diagnosis during syncope was made by the ILR in 34 (65.3%) cases. Bradycardia related arrhythmia diagnoses were as follows: 24 sinus node dysfunction and 2 sudden atrioventricular (AV) blocks. Tachy-arrhythmia syncope diagnoses were as follows: 4 SVT, 2 ventricular tachycardia (VT), and 2 ventricular fibrillation (VF). As ILR-guided therapy, 25 (73.5%) patients underwent PM implantation, and 3 patients underwent implantable cardioverter defibrillator (ICD) implantation.

Significant arrhythmia documented by IRL irrespective of syncope and the corresponding ILR-guided therapies are shown in Supplementary Table 2. Significant arrhythmia was documented by ILR irrespective of syncope in 99 (57.2%) patients after the mean follow-up of 5.5 ± 7.9 months. The cumulative incidence of ILR-guided therapy was 47.8% and 57.1% at 1 and 2 years, respectively. (Fig. 2B). A device (PM or ICD) was implanted if the significant arrhythmia found by ILR interrogation was deemed a possible cause of the clinical syncope by the physician, following the ESC guidelines.24 A PM was implanted in 60 patients (34.7%), 5 (2.9%) received an ICD, and 4 (2.3%) underwent radiofrequency catheter ablation (RFCA).

A total of 60 patients (34.7%) underwent PM implantation during a mean follow-up period of 6.0 ± 9.3 months. Comparisons between patients who received a PM and those who did not are shown in Table 3. The PM group was older, took more antihypertensive medication, had a higher proportion of paroxysmal AF history and relatively short duration of syncope, and were more likely to have had any BBB on the baseline 12-lead ECG. The univariable analyses showed that age > 75 years (hazard ratio [HR], 1.97; 95% CI, 1.18–3.30; P < 0.01), a history of paroxysmal AF (HR, 2.73; 95% CI, 1.62–4.58; P < 0.01), and any BBB on the baseline 12-lead ECG (HR, 2.98; 95% CI, 1.35–6.58; P < 0.01) were significantly associated with later PM implantation (Fig. 3). The multivariable analysis showed that a history of paroxysmal AF (HR, 2.34; 95% CI, 1.33–4.12; P = 0.01) and any BBB on the baseline 12-lead ECG (HR, 2.52, 95% CI, 1.09–5.85; P = 0.03) were significantly associated with PM implantation (Table 4).