Subcutaneous and Epicardial Defibrillators

Implantation

Implantation of an S-ICD can be accomplished using anatomic landmarks without the need for fluoroscopy. An approximately 4-cm incision is made, and a subcutaneous pocket is formed in a lateral position along the sixth rib, between the midaxillary and anterior axillary lines. The high-voltage lead is tunneled anteriorly and secured using two small incisions to anchor the proximal electrode at the level of the xiphoid process and the distal electrode at the level of the second intercostal space. This places the coil 1 to 2 cm to the left of and parallel to the sternum ( Fig. 17-2 ). Alternatively, a two-incision implant technique has been proposed. This method uses a peel-away sheath to allow tunneling of the distal electrode and shock coil lateral to the sternum, eliminating the need for an incision and anchoring at the distal electrode. Use of this technique demonstrated that lead position remains stable and may reduce the risk of erosion at the distal tip of the electrode, where subcutaneous tissue is minimal in many patients. Following implantation, defibrillation testing is then usually performed using 65-J shocks to terminate induced ventricular fibrillation (VF). The S-ICD delivers nonprogrammable 80-J shocks, so testing at 65 J ensures an adequate safety margin.

Procedural complications of S-ICD implantation include infection of the subcutaneous pocket and lead migration or dislodgement if the distal electrode is not adequately anchored. Also, erosion of the generator and subcutaneous leads has been reported. Erosion may be caused by the larger size of the generator, small patient body habitus, or improper lead location (i.e., directly over the sternum, where subcutaneous tissue is limited). The rate of pocket infection with the S-ICD is higher than that of transvenous ICD systems, with recent reports ranging from 4% to 6%. This may be due to the use of multiple incisions at implant, as well as to the inherent higher risk of the cohort of patients selected for S-ICD implantation. Although the risk of infection may be similar to or even higher than that associated with transvenous systems, the risk of the infection itself is less because the absence of intravascular leads markedly reduces the risk of serious complications from extraction, and intervascular infections such as endocarditis have not been reported.

Effectiveness and Safety

The ability of the S-ICD to detect and terminate induced VF was initially evaluated in 59 patients, 58 (98%) of whom were successfully defibrillated with 65 J on two successive induced episodes. One patient was defibrillated in the first but not the second induced episode. Following permanent implantation in the 59 patients, there were 12 episodes of spontaneous ventricular tachycardia (VT) that were successfully identified and treated in 10 months of follow-up, demonstrating the effectiveness of the S-ICD for the detection and treatment of ventricular tachyarrhythmias. The successful treatment of VT/VF was reproduced in a single-center experience in which 4 of 31 patients who received an S-ICD implant had VT/VF in 286 days of follow-up, and all VT/VF episodes were successfully terminated.

In the largest prospective multicenter evaluation of the safety and efficacy of the S-ICD system in patients otherwise indicated for an ICD without pacing indications, 304 patients underwent successful implantation and defibrillation threshold (DFT) testing. The primary safety endpoint demonstrated a 180-day complication-free rate of 99.0%, with no cases of lead failures, endocarditis, bacteremia, cardiac tamponade, hemothorax, or pneumothorax. Also, there was no lead or generator migration in 99% of patients during follow-up. There were 18 suspected or confirmed infections in the study population, 4 of which required explantation. There were 14 superficial or incisional infections that were all successfully treated conservatively. The primary efficacy endpoint in this study demonstrated 100% efficacy in acute conversion of VF with 65 J in DFT testing on two consecutive inductions. Chronic DFT testing at 65 J was performed in 75 patients, and a 96% success rate was achieved. In the three patients (4%) in whom 65 J failed, an 80-J shock was successful. The average time to therapy for induced episodes was approximately 15 seconds. A total of 119 VT/VF episodes (38 discrete episodes of VT/VF and 81 VT/VF events in the setting of VT/VF storm) in 21 patients were treated. All discrete episodes were successfully treated, and the S-ICD was successful in converting 92% of the discrete episodes with one shock and 97% with one or more shocks (one episode of monomorphic VT terminated spontaneously during charging for a second shock and in follow-up this same patient had another VT episode that was successfully treated with a single shock). All of the episodes of VT/VF that occurred during VF storm were successfully treated with the S-ICD.

Some authors have suggested that, in early clinical experience, ineffective shock delivery may occur despite successful testing at implant. In a multicenter case-control study, defibrillation efficacy was compared in 69 patients who received S-ICDs to age- and sex-matched control subjects who received standard transvenous ICDs. There was no statistical difference in first-shock efficacy between S-ICDs and transvenous ICDs at a 15-J safety margin (89.5% vs. 90.8%, P = 0.8).

A worldwide experience in the Evaluation of Factors Impacting Clinical Outcome and Cost Effectiveness of the S-ICD (EFFORTLESS S-ICD) Registry has recently been reported. A total of 472 patients were enrolled, with a mean follow-up of 558 days. Implant times were an average of 1 hour, and 87% of implants were performed without the use of fluoroscopic imaging. In the first 30 days, there was a complication rate of 3%, with one hematoma, one pneumothorax, and one pleural effusion. In follow-up, there were no lead fractures. Four patients had documented lead migration, and only two leads required repositioning. At 1 year of follow-up, the complication-free rate was 94%. A total of 18 patients (4%) had documented or suspected infection, 10 of which were serious enough to require system explantation (explant rate of 2.2%). Initial DFT testing in 393 patients demonstrated a successful conversion rate of 99.7%. Of the 91 episodes of spontaneous VT/VF recorded during follow-up, 51 were discrete and 40 occurred during VT/VF storm. The single-shock success rate for the discrete episodes was 88%, and ultimately, with more than one shock and/or spontaneous termination, all 51 episodes were successfully converted. In 4 patients, there were 40 VT/VF storm episodes. Two of these patients ultimately died, one as a result of pump failure and one due to VF that occurred after prolonged bradycardia and failed defibrillation. The mean time to therapy for spontaneous episodes was 17.5 seconds, with a range of 6.0 to 29.4 seconds, reflecting a longer charge time for high energy (80 J) shocks in the ambulatory setting. The overall 95% confidence interval for conversion efficacy of spontaneous episodes was 96.1% (90.8-100%).

More recently, 2-year results of a pooled analysis of the S-ICD System IDE Clinical Study and the EFFORTLESS Registry confirmed the efficacy of the S-ICD system for treating VT/VF, as well as reductions in complications and inappropriate shock rates with increased operator experience and strategic device programming. Table 17-1 summarizes the early clinical trials using the S-ICD system.

Arrhythmia Detection, Sensitivity, and Discrimination

Of particular importance in use of the S-ICD is the ability to identify cardiac signals without interference from other extracardiac signals that may be biologic (i.e., myopotentials) or extrinsic noise. Such real-world circumstances cannot be easily reproduced in the electrophysiology laboratory and highlight the importance of monitoring for effective arrhythmia detection and termination. Cardiac signals are about one order of magnitude smaller when recorded from subcutaneous electrodes compared with intracardiac electrodes, so the lower signal-to-noise ratio creates engineering challenges to appropriate discrimination of events.

Before implantation, patients undergo surface electrocardiogram (ECG) screening in three vectors that mimic the sensing vectors employed by the S-ICD generator and proximal and distal electrodes. An ECG screening tool was developed to detect those patients who may be at increased risk for T-wave oversensing and therefore may not be good candidates for S-ICD implantation due to the risk of inappropriate shocks. In one study, 8% of patients failed this ECG screening. Figures 17-3 and 17-4 depict the screening tool and an example of a failed profile due to large T waves, respectively.

Once implanted, the device automatically selects the optimal sensing vector using any of the three combinations between the proximal and distal sensing electrodes and the pulse generator. The vector that best avoids noise, R (QRS) wave double-counting, and T-wave oversensing is used automatically by the device for rhythm interpretation. This choice can be manually changed, although once programmed it is the only vector used for all episodes. In its current iteration, the S-ICD can be programmed as a single- or dual-zone device. In a single-zone configuration, the shock zone is programmed to deliver therapy above a given rate threshold. In a dual-zone configuration, a conditional shock zone is set (programmable between 170 and 240 beats per minute) where arrhythmia discrimination algorithms are active. These algorithms use a stored QRS:T morphology template to compare detected morphologies in the conditional shock zone and withhold therapy if waveforms match or duration criteria are not met. Following identification of a VT/VF episode, the device charges and confirms a sustained tachyarrhythmia before delivering 80-J shocks. In the event of an unsuccessful initial shock, the device can automatically reverse polarity for subsequent shocks. Due to an average capacitor charge time of 14 ± 2 seconds, the arrhythmia is reconfirmed before defibrillation to avoid treatment for a nonsustained episode. Successful termination of VT/VF resulting in postshock asystole triggers 30 seconds of demand pacing at 50 beats per minute. Twenty-four treated episodes can be stored with up to 120 seconds per episode, including pre-event electrograms to termination. An example of a detected VT episode with an appropriate shock is depicted in Figure 17-5 .

Arrhythmia discrimination requires the analysis of signals very different from transvenous ICD systems, using the three previously described sensing electrodes with signal characteristics closely resembling those of surface ECG recordings. To prevent oversensing of cardiac signals and noise, the sensing algorithm adapts thresholds to the amplitude of the QRS wave and decays over time. Each signal is also sent through three different double-detection algorithms to prevent oversensing caused by double counting (i.e., T-wave oversensing). An example of T-wave oversensing is shown in Figure 17-6 (see Case Study 17-1 ). Once these algorithms are completed, the S-ICD determines the rate of the rhythm to assess for a tachyarrhythmia. Once a tachyarrhythmia is detected, discrimination analyses occur. First, the morphology of the tachyarrhythmia beats is correlated to a baseline template acquired during sinus rhythm. A high degree of correlation (>50%) is categorized as an atrial arrhythmia. If there is low correlation, subsequent tachycardic beats are compared with previous tachycardic morphologies and, if a polymorphism is noted, then ventricular arrhythmia (i.e., polymorphic VT) is suggested. If the beat-to-beat relationship is monomorphic, then discrimination algorithms continue and the QRS width of the tachycardic beat is compared with the baseline template. A wide complex QRS is characterized as ventricular, and a narrow complex is labeled as an atrial tachyarrhythmia. This rhythm analysis uses an 18 of 24 duration criteria, and durations are automatically increased following nonsustained arrhythmias. Finally, a confirmation algorithm is employed following capacitor charging before shock to ensure that a sustained arrhythmia is present.