A prospective, ambispective, single-center observational study was conducted in patients younger than 18 years who underwent S-ICD implantation. The comparison group included contemporaneous patients who received a TV-ICD. Follow-up adhered to our center's standard protocol, involving an initial in-person visit after hospital discharge (within the first month postimplantation), followed by remote device monitoring using the Latitude transmission system (Boston Scientific, United States). In the event that transmission failed, patients were contacted by telephone to assess their clinical status and resolve any technical issues. In-person evaluations were scheduled based on the patient's underlying heart condition. TV-ICD patients were prospectively included starting in 2016 and retrospectively enrolled as far as 2014. Two patients who were switched from TV-ICD to S-ICD were included in both groups, taking into account the length of time they belonged to each group. The follow-up protocol was consistent, as described above.

The standard procedure in our center for S-ICD implantation involves generator placement between the latissimus dorsi and serratus muscles using a 2-incision technique, with potential modifications based on the patient's specific characteristics. The electrode was situated in either a left or right parasternal position, depending on the screening results. Fluoroscopy was used before the procedure in all cases to verify correct device positioning and mark the skin. The decision to use fluoroscopy during implantation was left to the operator's discretion. TV-ICD implantation followed the standard technique, leaving a loop in the electrode to accommodate body changes as the patient grew. Antibiotic prophylaxis was used in all procedures.

The typical programming mode for S-ICD included 2 detection zones: shock delivery at 250 beats per minute (bpm) and conditional therapy starting at 220 bpm. The SMART-PASS filter was activated (if available), and postshock pacing was implemented. In TV-ICD, the ventricular fibrillation (VF) window started at 220 bpm, and a monitoring or second therapy zone was chosen at the operator's discretion.

The primary outcome measure of the study was a composite of complications and inappropriate shocks. As secondary outcomes, we analyzed these events separately, as well as therapy effectiveness, appropriate shocks, and the need and reasons for device replacement. Events were carefully evaluated in the subgroup of patients with a body mass index (BMI) ≤20, as this factor has been associated with an increase in complications in previous studies.12

Complications were categorized into acute events (occurring during the procedure) and events developing after implantation. We analyzed all complications potentially associated with the procedure, the generator, and the electrode. These included pneumothorax, pericardial effusion, tamponade, surgical wound infection, device infection, device dysfunction, need for nonelective device replacement, hematoma requiring a subsequent intervention, electrode displacement requiring a second procedure, therapy failure, and lack of ventricular arrhythmia detection, as well as complications requiring an initially unplanned surgical procedure or device replacement.

The patients’ relevant demographic and clinical data, as well as events occurring during follow-up were recorded. All data were collected using RedCap (Research Electronic Data Capture) forms.13,14

Qualitative data are presented as frequency and percentage, and quantitative data as median [interquartile range]. Variables were compared using the chi-square test and the Wilcoxon rank-sum test, respectively. A survival analysis based on time to event was conducted for time-dependent variables. Kaplan-Meier survival curves were obtained, and differences were analyzed using the log-rank test. To compare survival at a specific time point (eg, at 5 years), we used a test based on cloglog transformation of the Kaplan-Meier estimators.15 Hazard ratios for the primary and secondary events were not calculated because the study variable ICD type violated the proportional hazards assumption of the Cox regression model. Data analysis was performed using R software,16 version 4.2.3, and the following packages: Hmisc (5.0.1), survival (3.5.3), ComparisonSurv (1.1.1), gtsummary (1.7.0) for tables, and ggplot2 (3.4.1) and survminer (0.4.9) for graphs.17–22

The study adhered to the principles of the Helsinki Declaration and was approved by the Research Ethics Committee (CEIm) of the center. Prospectively included patients provided signed informed consent to participate, whereas those included retrospectively were exempted by the CEIm. The image featuring a patient was published after signed consent was obtained from the patient's parents.

Since January 2016, S-ICDs have been implanted in 26 patients aged 5 to 17 years, with a minimum weight of 20kg; among them, 12 (46%) had BMI ≤20. In 2 patients, S-ICD implantation was performed because of previous complications with TV-ICDs. Screening results were favorable in at least 2 vectors on 1 side in all except 3 patients, who had only 1 suitable vector on each side. Starting in 2014, TV-ICDs have been implanted in 19 patients (aged 5 to 16 years, minimum weight 24kg; 11 [58%] with BMI ≤20). Since 2016, the reasons for TV-ICD implantation have been a pacing requirement or unfavorable screening for S-ICD placement.

Baseline characteristics of the 2 groups are shown in table 1 and in figure 1 of the supplementary data. Body weight was significantly higher in S-ICD-treated patients then in those receiving a TV-ICD (59kg vs 45kg, respectively; P=.027), although there were no differences in BMI (20.2 vs 19.2; P=.32). In the overall cohort, the predominant underlying heart diseases were hypertrophic cardiomyopathy (51%), arrhythmogenic conditions (16%), congenital heart disease (13%), and ion channel disorders (11%).

Implantation data and programming at discharge in the 2 groups are described in tables 2 and 3, and in table 1 of the supplementary data. In total, 41 implants (91%) were performed in the electrophysiology room, and 42 (93%) with the patient under general anesthesia. The predominant approach for S-ICD implantation was the 2-incision technique (n=24 [92%]), with intermuscular placement in 23 (88%) patients and subserratus placement in 3 (12%) patients weighing<30kg (figure 1). Fluoroscopy was used during the procedure in 17 patients (65%) to ensure correct electrode positioning. The A219 model was chosen for implantation in all except the first patient in the series, who received the A209 model. Based on the operator's criteria, defibrillation testing was omitted in 1 case due to the patient's clinical condition, and ventricular fibrillation (VF) could not be induced in another patient.

Figure 1.

Central illustration. S-ICD implants are safe and effective in children; complication rates and inappropriate shock rates are not higher than with transvenous implantable cardioverter-defibrillators. A-D: Kaplan-Meier curves for appropriate shock-free survival (A), inappropriate shock-free survival (B), complication-free or inappropriate shock-free survival (C), and complication-free survival (D). E: Outcome of subserratus S-ICD implantation in the smallest patient in the series (5 years, 112cm, 20kg). S-ICD, subcutaneous implantable cardioverter defibrillator.

(0.28MB).

Among TV-ICD-treated patients, 15 (79%) received a single-chamber device. The fluoroscopy time (4.1 [3.0-8.9] vs 0.3 [0.2-0.5] min; P<.001) and procedure duration (90 [71-112] vs 71 [54-75] min; P=.012) were longer in TV-ICD than in S-ICD procedures. No defibrillation tests were carried out.

In all S-ICD patients, 2 zones were programmed, with the shock window threshold set at 250 bpm, and the conditional window at 230 (220-230) bpm. In TV-ICD patients, the VF zone threshold was set at 220 (220-226) bpm, the monitoring zone at 180 (170-200) bpm in 13 (68%) patients, and a second therapy zone was programmed at 190 (178-200) bpm in 4 (21%) patients.

Overall median follow-up was 3.71 [1.51-5.73] years: 2.70 [1.56-4.77] in the S-ICD group and 5.50 [1.96-8.07] in the TV-ICD group (P=.091). Events are described in table 4, tables 2 to 4 of the supplementary data, figure 1, and figure 2. All patients were alive at completion of the study, and 3 (6.7%) patients had undergone transplantation.

Figure 2.

Graph showing events since implantation. A: Patients with a subcutaneous implantable cardioverter defibrillator. B: Patients with a transvenous implantable cardioverter defibrillator.

(0.25MB).

In the S-ICD group, appropriate shock-free survival at 1 and 5 years was 88% and 54%, respectively (table 2 of the supplementary data). Among the 9 patients (35%) who received appropriate shocks, cardioversion was successful at the first attempt (100% effectiveness). In most cases (78%), shocks were delivered to treat polymorphic ventricular tachycardia (VT). None of the patients required switching to TV-ICD for pacing needs, whether for resynchronization, antitachycardia therapy, or bradyarrhythmia. Inappropriate shock-free survival at 1 and 5 years was 96% and 85%, respectively. Three patients (12%) received inappropriate shocks. These were due to T-wave oversensing (n=2), corrected in both patients by optimizing sensing during exercise, and myopotential oversensing (n=1).

Complication-free survival at 1 and 5 years was 96%. There were no procedure-related complications (table 3 of the supplementary data). During the follow-up period, 5 events occurred in 3 patients (12%) (table 4 of the supplementary data): 2 patients experienced surgical wound infections (1 had infection in both wounds) that were resolved with antibiotics and no need for further surgery, and 1 patient required surgical wound cleaning without device removal. In addition, generator replacement was required in 2 patients due to premature battery depletion (both cases were affected by the 92400926D-FA alert). These events occurred after the fifth year of implantation in both patients. The incidence of complications was not higher in patients with BMI ≤20 (figure 2 of the supplementary data).

In the TV-ICD group, 4 patients (21%) experienced appropriate shocks, all of which were successful at the first attempt (100% effectiveness). The underlying cause in 2 patients was sustained monomorphic VT. These patients experienced additional VT episodes that were suppressed by antitachycardia pacing. Seven complications occurred in 6 patients (32%) (complication-free survival rates at 1 and 5 years were 79% and 57%, respectively). Procedure-related complications included 1 pneumothorax and 1 pericardial effusion, managed conservatively, and 1 bleeding episode requiring surgery (table 3 of the supplementary data). During follow-up, 1 patient had undetected VT below the threshold, and 3 patients had out-of-range shock impedance due to electrode stretching caused by growth. Inappropriate therapy was delivered to 3 patients (16%) due to the following events: T-wave oversensing (n=1), sinus tachycardia (n=1), and atrial flutter (n=1) (table 4 of the supplementary data). In the S-ICD group, there were no signal sensing alterations due to growth-related changes in electrode position in the chest (figure 3), even among the 3 patients with a height increase >10cm.

Figure 3.

Radiologic images showing changes in subcutaneous implantable cardioverter-defibrillator position with developmental patient growth. A: Image at implantation. B: Image at 4 years’ follow-up after>10cm growth.

(0.1MB).

Overall, there were no statistically significant differences between groups in the primary endpoint of the study: event-free survival rates at 1 and 5 years were 92% and 80% in patients receiving S-ICDs, and 73% and 63% in those with TV-ICDs, respectively (figure 1 and table 2 of the supplementary data). At 5 years of follow-up, complication-free survival was higher in the S-ICD group than the TV-ICD group (96% vs 57%; cloglog P=.016), However, later events in the S-ICD group equalized outcomes at the end of follow-up (figure 1D). Differences in complication rates were also observed in the subgroup of patients with BMI<20 (5-year complication-free survival 100% vs 48%; P=.037) (figure 3). Inappropriate shock rates were similar in the 2 groups (figure 1 and table 2 of the supplementary data).

Some limitations of this study should be acknowledged. First, the research was based on a single-center registry from a tertiary center with experience in pediatric device implantation, and procedures were carried out by a small number of operators. These factors may restrict general applicability of the results. Second, the small sample size and existing selection bias (regarding type of device) greatly limit conclusions drawn from the statistical comparisons. Hence, it would be advisable to conduct multicenter studies in this line. In addition, it could be worthwhile to consider including devices implanted in pediatric patients in the National ICD Registry23 published by the Arrhythmia Association of the Spanish Society of Cardiology, to collect data from other centers in Spain. Third, the follow-up time, although longer than in other studies, may have been insufficient to detect some long-term complications, especially concerning device battery life and potential electrode-related events. Lastly, factors that could have an impact on deciding the choice of device, such as quality of life, and others, such as cost-effectiveness, were not assessed in this study.