Introduction

The implantable cardioverter defibrillator (ICD) is routinely employed for primary and secondary prevention of sudden cardiac death [1, 2]. The totally subcutaneous implantable defibrillator (S-ICD, Boston Scientific, Natick, MA, USA) has been introduced as a new alternative to the conventional transvenous defibrillator system. The obvious advantages of this system are described as a reduction of lead complications and systemic infections. The subcutaneous ICD system can particularly be considered in patients with congenital heart disease [3], other rare entities impeding transvenous lead implantation [4], or electrical heart disease [5, 6]. Young patients with hypertrophic cardiomyopathy (HCM) also represent suitable candidates for S-ICD implantation [7].

The early results of the worldwide EFFORTLESS registry suggested an appropriate system performance. Occurrence of arrhythmic events and inappropriate shocks resembled those reported for conventional transvenous ICD systems [8]. Comparable results were described in the 2-year follow-up of the same cohort [9]. Of note, a trend toward a reduction of inappropriate shocks was reported after 2 years follow-up. Nonetheless, common problems in the S-ICD patient population such as inappropriate sensing due to muscular noise and T-wave oversensing [10–14] or high defibrillation thresholds [15] may occur. Especially in patients with HCM, inappropriate shock deliveries due to T-wave oversensing remain a clinical challenge [16]. Data from the EFFORTLESS registry suggest a hazard ration of 4.6 for inappropriate shocks in patients with HCM [16]. In the present study, the single-center experience of 18 HCM patients who received an S-ICD system and had a minimal follow-up duration of 6 months was systematically evaluated.

Methods

The study conforms to the declaration of Helsinki and later amendments. Between July 2010 and June 2015, 102 S-ICD systems were implanted at our institution. Among these, 18 patients with HCM and a minimal follow-up duration of 6 months were analyzed. Patient baseline characteristics are summarized in Table 1. All patients underwent preoperative ECG screening [17, 18] and an intraoperative defibrillation test. An ineffective first shock required further tests in either reverse polarity or after repositioning of the subcutaneous lead and/or the pulse generator. Devices were routinely programmed to a detection rate of 220 bpm (conditional shock zone) and 240 bpm (shock zone) (Table 2).

Table 1 Baseline characteristics
Full size table
Table 2 Ventricular arrhythmias and T-wave oversensing in HCM patients with an S-ICD system (n = 18)
Full size table

Results

Regular follow-up

18 HCM patients of a mean age of 35 ± 19 years received an S-ICD system. The mean follow-up duration was 31.7 ± 15.4 months. 3 patients presented hypertrophic obstructive cardiomyopathy (HOCM), while 15 patients did not display signs of obstruction of the left ventricular outflow tract (HNCM). The mean left ventricular ejection fraction was 62.8 ± 5.7 %. 14 of 18 patients (78 %) received the S-ICD system as a primary prophylaxis of sudden cardiac death, while ventricular arrhythmias had already occurred in 4 patients (22 %). In 9 of 18 patients (50 %), the primary sensing vector was recommended while the secondary sensing vector was chosen in 7 patients (39 %) and the alternative sensing vector in 2 patients (11 %). Ventricular fibrillation could be induced in 17 of 18 patients during intraoperative defibrillation test. In 1 patient, sustained ventricular fibrillation was not inducible and the patient did not accept postoperative defibrillation tests. In 15 of 17 patients (88 %), the first defibrillation test with 65 J was effective. In 2 of 12 patients (12 %), ventricular fibrillation was not terminated by the first internal shock. In these patients, further tests with 65 J and reversed polarity effectively terminated ventricular fibrillation.

T-wave oversensing and inappropriate shocks

In this cohort, T-wave oversensing occurred in 7 of 18 patients (39 %). In 3 patients, these episodes were short and did not lead to inappropriate shock delivery while 4 of 18 patients (22 %) experienced at least one inappropriate shock.

In 2 of these 3 patients without shock delivery, the sensing vector was changed from secondary to primary after occurrence of T-wave oversensing during ergometer test. In the third patient, T-wave oversensing occurred during non-sustained slow ventricular tachycardia. Thereafter, no further episodes occurred and, therefore, no changes have been programmed.

Among patients who encountered inappropriate shocks due to T-wave oversensing, 3 patients experienced more than 1 shock. In one of these patients, supraventricular tachycardia with concomitant T-wave oversensing led to delivery of 3 successive inappropriate shocks. Here, the sensing vector was altered from secondary to primary. In another patient, T-wave oversensing during sexual activity led to delivery of 5 successive shocks. The last shock resulted in ventricular fibrillation that was appropriately detected and terminated by another 80 J shock. Extensive ergometer tests led to programming of the alternative sensing vector. Thereafter, no T-wave oversensing occurred. Another patient experienced one inappropriate shock due to T-wave oversensing during sinus tachycardia as well as another episode with inappropriate shock delivery in the presence of triple count due to p-wave and T-wave oversensing. In this case, oversensing was present in all possible sensing vectors during ergometer test. Therefore, the S-ICD system was explanted and replaced by a conventional transvenous defibrillator system [19].

Detection of ventricular arrhythmias and further adverse events

Ventricular arrhythmias were appropriately detected in 4 patients. In 3 patients, non-sustained fast ventricular tachycardia was detected but the episodes terminated spontaneously before a shock could be delivered. Therefore, the charging process was interrupted. In another patient, one episode of sustained polymorphic ventricular tachycardia was appropriately detected and terminated with the first 80 J shock. Further adverse events included surgical revision due to a mobile sensing electrode and resulting noise detection as well as one case of early battery failure requiring pulse generator change after 32 months without shock delivery.

Discussion

In the present single-center registry, a significant proportion of S-ICD recipients with HCM experienced inappropriate shocks as a result of T-wave oversensing. T-wave oversensing occurred in 39 % of HCM patients. In 1 patient, explantation of the S-ICD system and implantation of a transvenous ICD system were necessary. Ventricular arrhythmias were adequately detected and effectively terminated if necessary.

T-wave oversensing

In the present study, an increased incidence of T-wave oversensing in patients with HCM was observed as compared with the recently published data from the EFFORTLESS registry [16]. In this multi-center registry, 8.3 % of 581 patients experienced inappropriate shocks. However, the follow-up duration of 21 ± 13 months was significantly shorter than in the present registry (31.7 ± 15.4 months). In accordance with the latest results of the EFFORTLESS registry, in most cases, reprogramming of the sensing vector from secondary to primary configuration prevented further T-wave oversensing. However, in one case programming of the alternative sensing vector was necessary.

A low R/T ratio represents a major risk factor for the occurrence of T-wave oversensing in this patient cohort. Alterations of the R/T ratio in the subcutaneous ECG may occur during follow-up and result in T-wave oversensing. These alterations occur most likely during exercise testing [18]. Changing of the sensing vector in patients with inappropriate shock deliveries resulted in an increased R/T ratio, especially during exercise test. Other abnormalities such as bundle branch block or repolarization abnormalities can also play a role. These scenarios should normally be detected during the preoperative screening process. Nonetheless, T-wave alterations are more likely in HCM patients [20, 21] and, therefore, increase the risk of oversensing. Furthermore, a reduced R/T ratio can be observed in the presence of obesity [18]. T-wave alterations may also be a result of drug-induced QT-prolongation [22].

Studies in patients with conventional transvenous defibrillator systems suggest that the rate of inappropriate shocks in usual defibrillator patient cohorts may be lowered to approximately 5 % by programming higher detection rates as well as increased detection intervals [23–25]. Although the detection intervals of the S-ICD system are fixed to 18/24, the longer charging period leads to a comparable time to therapy as conventional transvenous ICD systems. However, in the present study all S-ICD systems were programmed to detection rates of 220 bpm. Therefore, significant modifications of the detection rate to extend time to therapy were not possible.

Implications

According to the significantly elevated incidence of T-wave oversensing and inappropriate shocks in HCM patients, these patients should be selected thoroughly and monitored closely. A recent screening analysis of QRS/T-wave morphology suggested HCM as an independent predictor for non-suitability for S-ICD systems [18]. The results of our registry further support this finding but also suggest that differentiated programming of sensing vectors depending on results of exercise tests may reduce the rate of inappropriate shocks. The majority of patients included in the present registry were implanted long before the latest recommendations were published [16]. At any rate, exercise-based optimization of the S-ICD system should be thoroughly performed in HCM patients to reduce the risk of T-wave oversensing. This should include examination of all three sensing vectors during exercise test as well as acquisition of subcutaneous ECG templates [26]. Exercise testing after S-ICD implantation is of particular interest because surface ECGs employed for preoperative screening cannot be regarded as adequate surrogates for subcutaneous ECGs during exercise [27].

Of note, new sensing algorithms may further reduce the susceptibility to T-wave oversensing [28] although the suggested algorithms have not yet been evaluated in prospective studies. A thorough preoperative screening in combination with these algorithms may solve the problem of T-wave oversensing in the future (Fig. 1).

Fig. 1
figure 1

a Representative example of T-wave oversensing during exercise test. b No T-wave oversensing after changing the sensing vector in the same patient

Full size image

Conclusion

The present study underlines that patients with HCM who receive an S-ICD system possess a significant risk for T-wave oversensing. Often, T-wave oversensing can be avoided by changing the sensing vector. Although preoperative ECG screening is thoroughly performed, T-wave oversensing can still occur and lead to inadequate shock delivery eventually. An increased probability of T-wave oversensing can be observed in the presence of low R-wave signals. This is especially important in patients with HCM because T-wave abnormalities may occur frequently [20] and thereby lead to alterations of the R/T ratio. Therefore, the vector with the highest discrimination between R- and T-waves in different positions and during exercise test should be chosen.

In case of repeated inappropriate shocks and reproducible occurrence of oversensing during exercise tests in all sensing vectors, a switch to a conventional ICD system should be discussed.