The extravascular implantable cardioverter–defibrillator: technology, evidence, and clinical perspectives

Abstract

The extravascular implantable cardioverter–defibrillator (EV-ICD) offers a fully extravascular alternative to transvenous ICDs, combining substernal lead placement with defibrillation and limited pacing capabilities. Designed to mitigate intravascular complications while maintaining generator size and longevity comparable to conventional ICDs, the EV-ICD supports antitachycardia pacing, post-shock pacing, and temporary bradycardia management. Procedural experience demonstrates high implantation success, low rates of major complications, and effective arrhythmia termination. Early real-world data indicate that lead dislodgement, pneumothorax, and pocket infection are uncommon, and lead explantation is feasible when required. Antitachycardia pacing effectively terminates a substantial proportion of ventricular tachycardias, while inappropriate therapies have been reduced through optimized lead positioning and advanced detection algorithms. Patient selection should exclude, among others, those with permanent pacing needs, anatomical constraints, and prior radiotherapy or sternotomy. Future directions include optimization of lead design and exploration of alternative implantation sites, with ongoing evaluation of long-term device performance and safety. Current evidence supports the EV-ICD as a safe and effective option in selected patients, including younger populations or those with limited vascular access, although its definitive role in sudden cardiac death prevention requires further long-term study.

Implantable cardioverter–defibrillator (ICD) therapy is a well-established therapy for primary and secondary prevention of sudden cardiac death (SCD) in patients at elevated risk of, or with prior, ventricular tachyarrhythmias due to a variety of acquired or inherited cardiac conditions [1, 2]. Transvenous ICD systems (TV-ICDs) remain the most established devices, pairing reliable arrhythmia detection with both defibrillation and pacing therapies. Contemporary systems provide bradycardia support and can detect and treat ventricular tachycardia using antitachycardia pacing (ATP), thereby obviating high-energy defibrillation or cardioversion shocks in appropriate episodes [3]. The implantation of TV-ICDs, due to the transvenous and intracardiac course of the leads, is associated with a spectrum of acute procedural complications, including cardiac perforation, tamponade, pneumothorax, hemothorax, and vascular injury [4, 5]. Over time, long-term adverse events may also occur, such as central venous thrombosis, venous obstruction or stenosis, tricuspid valve dysfunction, pocket erosion or infection, and lead-related endocarditis. In addition, lead malfunctions, including fractures or insulation defects, represent a particularly serious concern, as they may result in inappropriate shocks or the failure to deliver effective therapy, often necessitating complex and high-risk lead extraction procedures [5, 6]. To address the need for a fully extravascular alternative, the subcutaneous ICD (S-ICD, Boston Scientific, Marlborough, MA, USA) was introduced in 2010, eliminating the necessity for transvenous leads and thus reducing the risk of intravascular complications [7]. The S‑ICD is implanted entirely in the subcutaneous space, with the generator positioned along the left midaxillary line and the electrode tunneled beneath the skin along the sternal bone. This design is particularly advantageous in patient populations at elevated risk, including those with limited venous access, prior tricuspid valve disease, younger age, or heightened susceptibility to infection, such as patients with a history of endocarditis, immunosuppression, dialysis dependence, or diabetes mellitus [8]. Despite its favorable safety profile, the S‑ICD is limited by the absence of continuous bradycardia and antitachycardia pacing, restricting its use in patients who require pacing support. Its pacing function is confined to short periods of transient stimulation following shock delivery. Additionally, because of the extrathoracic position of the lead, defibrillation with the S‑ICD requires a substantially higher energy output of approximately 80 J compared with 20–40 J in contemporary TV-ICDs. The higher energy requirement of the S‑ICD necessitates a larger pulse generator, which has implications for implantation and patient tolerance [7]. This may entail practical considerations in patients with low body mass index, potentially increasing the risk of generator erosion or aesthetic concerns, and requires caution during resuscitation to avoid inadvertent exposure of bystanders to shock energy [9]. Furthermore, S‑ICD requires pre-implantation ECG screening to assure appropriate sensing capabilities and reliable discrimination between QRS complexes and T waves, which is essential to minimize the risk of inappropriate shocks. Previous studies have shown that approximately 85% of patients with a primary or secondary prevention indication for an ICD demonstrate a surface ECG compatible with S‑ICD implantation when assessed using the screening template [10]. The proportion is slightly lower in patients with congenital heart disease (83%) and markedly reduced in those with a left ventricular assist device (73%; [11, 12]). The extravascular ICD (EV-ICD; Aurora, Medtronic, Minneapolis, MN, USA) was developed as an alternative to transvenous systems, designed to mitigate intravascular risks while preserving essential therapeutic functions. Positioned substernally, its lead enables closer proximity to cardiac signals, improving sensing and allowing for defibrillation at energies comparable to TV-ICDs (up to 40 J), alongside pacing functions for tachyarrhythmia termination and transient bradycardia management, thereby addressing a key limitation of the S‑ICD. By avoiding the vasculature, the EV-minimizes complications linked to transvenous hardware while maintaining a generator size and longevity similar to TV-ICDs. It offers a therapeutic option for patients who do not require permanent pacing and may not be suitable candidates for transvenous systems [13, 14]. Data from real-world cohorts have shown high procedural success in terminating induced ventricular arrhythmias and a low incidence of major complications at discharge, in line with prior clinical investigations [15]. This article aims to provide a comprehensive overview of the EV-ICD system, including its design and function, patient selection criteria, implantation technique, and a discussion of safety, complications, and evidence from published studies. By integrating clinical data and practical experience, we highlight the potential role of the EV-ICD in modern SCD prevention strategies.

The EV-ICD system: design and functions

The EV-ICD consists of a pulse generator similar in size and form to a conventional single-chamber ICD and a dedicated quadripolar lead with an epsilon-shaped distal section incorporating two pacing rings (Ring 1 and Ring 2) and two shock coils (Coil 1 and Coil 2; [16]). The substernal implantation of the lead ensures direct proximity to the heart, optimizing sensing and defibrillation performance. This design enables multiple sensing and pacing configurations through the rings and coils, with the standard sensing vector defined as Ring 1 to Ring 2 [17]. Two additional vectors (Ring 1 to can and Ring 2 to can) enable alternative sensing configurations in cases of inadequate signal quality with the standard vector. Reported R‑wave amplitudes range between 1.86 ± 0.93 mV [16] and 3.4 ± 2.0 mV [14]. To ensure reliable detection despite lower absolute amplitudes compared with transvenous systems, sensitivity can be programmed down to 0.075 mV, supported by dedicated algorithms that suppress oversensing and enhance discrimination between supraventricular and ventricular arrhythmias [16]. Several pacing vectors are available, enabling three distinct stimulation modes: antitachycardia pacing for termination of ventricular tachycardias without shock delivery, post-shock pacing to prevent asystole, and temporary antibradycardia pacing. The vectors Ring 1 to Ring 2 and Ring 1 to Coil 2 are configured as low-output pacing vectors, whereas Coil 1 to Coil 2 serves as a high-output vector capable of delivering amplitudes up to 30 V with a pulse width of 10 ms [17]. Mean pacing thresholds were measured at 10.8 ± 2.2 V at a pulse width of 10 ms [18]. Defibrillation and cardioversion for ventricular fibrillation and ventricular tachycardia occur between the coils and the generator. Unlike many TV-ICDs, the EV-ICD cannot charge for defibrillation while delivering ATP. The temporary antibradycardia pacing feature is designed to detect and treat asystole but is not sufficient for long-term pacing. When activated, the device operates in OVO mode, continuously monitoring for pauses in ventricular activity. If a pause exceeding the programmable interval (5–15 s) is detected, the system automatically switches to VVI mode at 40 bpm for 30 s before reverting to OVO mode. This process can be reinitiated in cases of recurrent pauses. Primarily due to lower defibrillation energy requirements, the generator has a projected service life of approximately 11.7 years [17, 19].

Patient selection and preprocedural planning

Implantation of an EV-ICD may be considered in all patients with a primary or secondary prophylactic indication for ICD therapy but without a permanent indication for pacemaker or resynchronization therapy [3, 18]. Extravascular ICD implantation is contraindicated in patients with prior or planned sternotomy, active devices providing unipolar or multi-chamber pacing, or antitachyarrhythmia therapies. Likewise, in patients with sternal deformities or pronounced thoracic anomalies, such as a marked pectus excavatum, those with pericardial disease, recurrent pericarditis leading to effusion or calcification, mediastinitis, or after thoracic radiation therapy, the EV-ICD should currently not be used. However, there have been reports describing successful implantation even in challenging anatomies [20]. Additional exclusion criteria include prior abdominal surgery in the epigastric region, hiatal hernia with distorted mediastinal anatomy, decompensated heart failure, oxygen-dependent chronic obstructive pulmonary disease, and marked hepatosplenomegaly [13]. The presence of a leadless pacemaker does not represent a contraindication [21]. Careful surgical preparation and planning are of particular importance for EV-ICD implantation [22]. While preoperative ECG screening to assess EV-ICD suitability is not required, preoperative thoracic computed tomography can be helpful for planning the procedure by providing detailed visualization of anatomical structures [17]. Of particular relevance are the retrosternal space, the size and position of the right atrial appendage (as it carries a risk of P‑wave oversensing), and the relationship of the lungs, since the lead is implanted in this region [17, 23]. Computed tomography also enables detection of adhesions or thoracic deformities that may complicate the procedure [22, 24]. Importantly, the EV-ICD represents an additional therapeutic option in patients with vascular access challenges or those who have experienced complications from prior transvenous procedures, such as lead-related endocarditis or venous thrombosis, where conventional TV-ICD implantation may be high risk or unfeasible.

Implantation technique and procedural workflow

Implantation of an EV-ICD may be performed in the electrophysiology laboratory or in a hybrid operating room, provided fluoroscopic equipment with a C-arm is available. Immediately before the procedure, the planned positions of the pulse generator along the left midaxillary, carina tracheae, and the superior border of the cardiac silhouette are marked on the skin under fluoroscopic guidance. Key anatomical landmarks, including the xiphoid process, costal margin, and left parasternal border, are also marked. General anesthesia with intubation is currently recommended. Continuous intraoperative monitoring of vital parameters and ECG is obligatory, as is the application of external defibrillation patches outside the sterile field [17, 24]. Standard preparation includes sterile draping of the entire thoracic region and preoperative antibiotic prophylaxis. The procedure begins with a subxiphoid incision of approximately 3 cm toward the left costal margin, followed by preparation of the rectus abdominis fascia. Once the fascia is opened, the retrosternal space is bluntly dissected using the finger. A dedicated tunneling tool with an overlying sheath is then introduced under fluoroscopic guidance into the retrosternal space. The tip of the tunneling tool should remain directly retrosternal at all times to minimize the risk of cardiac or pulmonary injury. Positioning is verified in anteroposterior projection, after which the tool can be carefully advanced in a left anterior oblique (90° LAO) orientation. Tunneling along the left parasternal border is carried out under continuous alignment control and without resistance up to the level of the carina tracheae. After removal of the tunneling tool, the lead is introduced through the sheath. At this point, initial sensing measurements can be obtained, with particular attention to the absence of P‑wave oversensing. If P waves are detected, the lead can be gently repositioned caudally or retunneled. Retunneling is also required if R‑wave sensing is inadequate (< 1 mV). Once a position with acceptable measurements is achieved, the lead is secured to the subxiphoid fascia using non-absorbable suture material with a firm knot (e.g., a constrictor knot) to prevent postoperative dislodgement. Afterward, a subcutaneous generator pocket is created at the previously marked site. With the aid of a second tunneling tool, the lead is tunneled into the generator pocket and connected to the generator (Fig. 1). Intraoperative defibrillation testing of the EV-ICD system is recommended [23]. In the pivotal trial, the median procedure time from skin incision to wound closure was 76.0 ± 32.7 min. Lead positioning was successful at the first attempt in 84% of cases, with up to four attempts required, and the mean R‑wave amplitude at implantation was 2.4 mV [13, 25]. Postoperatively, a chest X‑ray in two projections should be performed to document lead and generator position (Fig. 2). In addition, device interrogation on the following day is recommended, including evaluation of sensing values in supine, left, and right lateral positions. Until successful postoperative assessment, telemetric monitoring is advisable [26].

Fig. 1.

Implantation steps of an extravascular implantable cardioverter–defibrillator. a Substernal skin incision at the previously marked site. Preparation of the rectus sheath followed by blunt finger dissection into the retrosternal space. b Introduction of the tunneling tool. c Release of the lead. f After lead fixation, tunneling into the lateral generator pocket. e Preparation of the generator pocket. d Connection to the generator and placement into the pocket

Fig. 2.

Postoperative chest X‑ray of an extravascular implantable cardioverter–defibrillator (EV-ICD) in two views: a anteroposterior, b lateral. The lead features two stimulation rings (Ring 1 and 2, blue lines) and two shock coils (Coil 1 and 2, orange lines). The generator is visible in the left anterior axillary line

Safety, efficacy, and clinical outcomes

In clinical trials and early real-world experience, the EV-ICD has demonstrated both procedural safety and efficacy. In line with prior clinical investigations [13], early data from real-world cohorts have shown that substernal lead tunneling and placement were successful in 99.6% of implants, with defibrillation testing achieving 99.0% success in tested patients [15]. Electrical measurements obtained at implantation remained stable at discharge. The overall rate of system- or procedure-related major complications was low at 3.9%, consistent with premarket data (2.4%) and comparable to complication rates reported for TV-ICDs in large registries (3.2%; [27]). Given the novelty of the substernal lead location, hands-on training is required for operators, particularly in early experience. Nevertheless, implantation was performed safely and successfully by a broad range of implanters (81 operators across 19 countries). In the pivotal study, an infection rate of 4.1% was observed within 10 months, most of which were successfully treated pharmacologically, with only 1.3% requiring system explantation due to pocket infection [13]. In the early real-world data, the most frequent complications were lead dislodgement (0.9%) and pneumothorax (0.9%). Clinical experience has shown that extravascular lead dislodgement into the pleural space can occur even after apparently uncomplicated implantation. Extraction and retunneling under fluoroscopic guidance are feasible options for pleural dislocation, thus avoiding conversion to alternative ICD strategies [28]. Furthermore, complete lead removal has been demonstrated to be safe and achievable up to 3 years post-implantation, with a success rate of 93.1% and simple traction being sufficient in most cases [29]. Notably, no cases of mediastinitis, endocarditis, tamponade, or complications requiring cardiac intervention or sternotomy occurred in either the pivotal study or the early real-world data [13, 15]. Overall, major intra- or postprocedural complications, such as lead dislodgements, were infrequent, although complications with the EV-ICD may necessitate conversion to a transvenous device. In the pivotal study, eight system revisions were performed, primarily due to infections [13]. Compared with premarket cohorts, patients in the EV-ICD Post Approval Registry (Enlighten) were significantly younger (49.4 vs. 53.4 years; p < 0.001) and healthier, with a higher mean left ventricular ejection fraction (43.5%; p = 0.004) and lower prevalence of cardiomyopathy, coronary artery disease, and hypertension (all p < 0.001; [15]). By comparison, both the Enlighten and the premarket EV-ICD patients are younger and healthier compared with the typical ICD population [30]. This suggests that, in clinical practice, the EV-ICD is being implanted more frequently in younger patients with inherited conditions predisposing to ventricular arrhythmias. Premarket studies excluded patients with a left ventricular ejection fraction less than 20%, which may have influenced the observed complication rates. By contrast, Enlighten reported major complications not previously seen. The pneumothorax rate of 0.9% is similar to that reported for TV-ICDs (0.9%) but higher than for S‑ICDs (0.1%; [31, 32]). Despite these events, overall procedural safety remained favorable, consistent with the pivotal trial [15]. Efficacy outcomes in both pivotal and real-world settings were similarly encouraging. In the pivotal study the system demonstrated high defibrillation efficacy, with success rates of 98.7% at implantation and 100% for discrete spontaneous events [13]. This represents greater efficacy at implantation than historically reported for TV-ICDs (90.5–93.0%; [33, 34]) and comparable outcomes to the S‑ICD (100%; [8]). Sensing and detection performance was reliable, with all induced ventricular tachyarrhythmias appropriately identified during implant testing, maintaining a safety margin at a sensitivity of ≥ 0.2 mV. Antitachycardia pacing successfully terminated 70% of monomorphic ventricular tachycardias, aligning with reported termination rates of 52% for endocardial or coronary vein ATP (Fig. 3; [15, 35]). Inappropriate ICD therapies initially occurred in 9.7% of patients, primarily due to P‑ or T‑wave oversensing and lead noise [13], before electrode handling during implantation was adjusted and the detection algorithm was optimized. This early experience underscores the importance of precise lead positioning to minimize the risk of P‑wave oversensing. In addition, the “Smart Sense” algorithm was subsequently integrated into the commercial EV-ICD device to avoid inappropriate shocks caused by P‑wave oversensing, achieving a 78% reduction [16]. Data from the Enlighten Registry have shown an overall incidence of inappropriate shocks at 6 months of 8.5% [15], higher than current TV-ICD systems [36] but similar to rates observed in early transvenous and subcutaneous devices [37]. Beyond procedural safety and therapy efficacy, patient-reported outcomes also favor the EV-ICD, with high acceptance and low psychological distress linked to its small generator size and pacing capabilities [38]. Pacing function were generally well tolerated, although antibradycardia pacing was deactivated in 4.6% of patients and postshock pacing in 1.8% due to discomfort with the pacing sensation. Antitachycardia pacing was either not programmed or later deactivated in 25.4% of patients for the same reason in the pivotal study, underscoring that while pacing functions broaden the therapeutic profile of the EV-ICD, patient perception of pacing remains an important consideration in long-term management [13].

Fig. 3.

Intracardiac electrograms illustrating successful termination of monomorphic ventricular tachycardia by antitachycardia pacing delivered via the extravascular implantable cardioverter–defibrillator

Future perspectives and clinical outlook

As experience with the EV-ICD continues to expand, focus has shifted to device design, implantation approaches, and patient selection, highlighting its evolving role in contemporary ICD therapy. Lead optimization—including electrode design and energy delivery strategies—may enhance signal amplitudes, reduce pacing thresholds, and thereby extend functionality beyond temporary bradycardia support toward more sustained pacing capabilities. Parallel advances in sensing algorithms are anticipated to refine arrhythmia discrimination and reduce inappropriate therapies [39]. In addition to these refinements, novel approaches to lead placement are under investigation. Additional studies are also needed to evaluate alternative implantation sites, such as intercostal versus substernal positioning, as well as the performance of button-shaped electrodes [40, 42]. Early experience with parasternal intercostal insertion of a DF4-compatible lead, positioned anterior to the pericardium and connected to standard transvenous generators, suggests the possibility of hybrid systems that combine the advantages of extravascular and transvenous technologies [42]. The clinical trajectory of the EV-ICD will also depend on long-term evaluation of lead stability, device longevity, complication rates, and explant safety over periods exceeding 4 years [39]. Early data from the Enlighten Registry confirmed high procedural success and low periprocedural complication rates, consistent with pivotal trial findings [15]. However, the Enlighten cohort was younger and healthier than typical ICD populations, underscoring the need for additional data in older patients with greater comorbidity, as well as in those with advanced structural heart disease [13]. Beyond adult cohorts, the EV-ICD may offer a particularly valuable therapeutic option for pediatric patients at risk of sudden cardiac death, especially when anatomical, clinical, or psychological considerations limit the applicability of transvenous or subcutaneous systems. The combination of effective defibrillation, ATP capability, extravascular lead placement, and compact generator size is particularly relevant in this younger population with long life expectancy [42, 43].

Conclusion

Current evidence supports the extravascular implantable cardioverter–defibrillator as a safe and effective option for selected patients. However, its definitive role in device-based sudden cardiac death prevention will depend on further technical refinement and long-term clinical data.

Declarations

Conflict of interest

D. Duncker received moderate speaker fees, travel grants, and/or a fellowship grant from Abbott, Astra Zeneca, Biotronik, Boehringer Ingelheim, Boston Scientific, Bristol Myers Squibb, CVRx, Daiichi Sankyo, Medtronic, Microport, Pfizer, Sanofi, and ZOLL. E. Angelini and K. Albert declare that they have no competing interests.

This article does not contain any studies with human participants or animals performed by any of the authors.