Author + information
- Received January 6, 2016
- Revision received February 12, 2016
- Accepted February 18, 2016
- Published online October 1, 2016.
- Tom F. Brouwer, MD∗ (, )
- Kirsten M. Kooiman, PA,
- Louise R. Olde Nordkamp, MD, PhD,
- Vokko P. van Halm, MD, PhD and
- Reinoud E. Knops, MD
- Heart Center, Department of Clinical and Experimental Cardiology and Cardiothoracic Surgery, Academic Medical Center, Amsterdam, the Netherlands
- ↵∗Reprint requests and correspondence:
Dr. Tom F. Brouwer, Department of Cardiology, Academic Medical Center, Room F3-241, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Objectives The study sought to describe the concept of algorithm-based screening with an external subcutaneous implantable cardioverter-defibrillator (S-ICD) to evaluate sensing using the rhythm discrimination algorithm of the device.
Background In a proportion of patients, screening for S-ICD therapy with the dedicated screening tool results in false negative and false positive results.
Methods Both patients who failed the standard screening and who passed with abnormal baseline ECGs were screened with an external S-ICD to evaluate sensing at rest and during exercise in all 3 sensing vectors (algorithm-based screening). Patients with adequate sensing were implanted with an S-ICD. Follow-up data regarding (in)appropriate shocks was collected.
Results Algorithm-based screening was performed in 15 patients. Group 1 consists of 8 who failed standard screening and Group 2 consists of 7 who passed and had abnormal ECGs. Six of 8 who failed standard screening in all sensing vectors demonstrated adequate sensing with the external S-ICD and were implanted with an S-ICD. Of these 6 implanted patients in Group 1, 1 inappropriate shock was observed duration median of 17 months’ follow-up and 2 episodes of ventricular fibrillation were successfully treated. Of the 7 patients in Group 2, who passed standard screening, 2 demonstrated inadequate sensing during additional screening with the external S-ICD. No appropriate or inappropriate shocks were observed in Group 2 during 10 months’ follow-up.
Conclusions Algorithm-based screening with the external S-ICD may improve patient selection and reduce the number of false positive and false negative screening results of the standard screening method.
Conventional transvenous implantable cardioverter-defibrillators (ICDs) show a survival benefit, despite the fact that some patients experience implant-related and long-term complications (1–5). The subcutaneous implantable cardioverter-defibrillator (S-ICD) was introduced to overcome complications associated with transvenous leads (2,6). Therefore the S-ICD may be preferred in specific patient groups with a high risk of transvenous lead related complications, for example, those with a high risk for lead failure, infection, a lack of vascular access, on hemodialysis or with certain anatomical anomalies (7).
The S-ICD system relies on a morphology based heart rhythm discrimination algorithm (6). Therefore, only patients who pass the pre-implantation electrocardiogram (ECG) morphology screening can be considered for S-ICD therapy (7). To determine the suitability of individual patients for S-ICD therapy the morphology of the QRS complexes and T waves are screened prior to implantation with a dedicated screening tool, in order to ensure adequate sensing (8). About 7% to 11% of patients screen out for S-ICD therapy and cannot be implanted because of increased risk of under- or oversensing (10–12). A recent study by Zeb et al. (9) demonstrated high sensitivity of the screening tool compared to the device algorithm to select adults with structurally normal hearts, at the cost of a lower specificity. In patients with congenital heart disease both the sensitivity and specificity were significantly lower.
Generally, patients who screen out, and therefore have a contraindication for S-ICD therapy, will be implanted with a transvenous device. However, some of these patients have a (relative) contraindication for transvenous ICD therapy, for example because of lack of vascular access, prosthetic valves or recurrent transvenous lead-related complications such as lead failure or endocarditis (7). Especially in this patient group, false negative results using the screening tool should be prevented and a more accurate approach to screening is needed, to determine whether these patients are truly not suitable for S-ICD therapy (9). On the contrary, patients with false positive results may be at increased risk of inappropriate shocks by the S-ICD, due to under- or oversensing heart rhythm.
The objective of this study is to describe the use of a novel algorithm-based screening method using an external S-ICD. This was performed in 2 specific patient categories: Group 1 consists of patients who screened out for S-ICD therapy and had a (relative) contraindication for transvenous ICD therapy. Group 2 consists of patients with abnormal ECGs, such as high amplitude T waves, low amplitude QRS complexes, large P waves or changes in QRS morphology during pacing, who screened in for S-ICD therapy but concerns over sensing performance remained.
Study outline and population
This study was conducted in a tertiary care cardiology center between 2009 and 2015. Potential candidates were identified upon presentation to the inpatient and outpatient clinic. Two groups were selected: those with a contraindication for transvenous ICD therapy were selected if they also failed the pre-implant ECG morphology screening using the manufacturer’s dedicated screening tool and those with abnormal ECGs who screened in for S-ICD therapy but concerns over sensing performance remained. Those who screened out and had no (relative) contraindication for transvenous leads were implanted with a transvenous ICD. Patients included in the ongoing PRAETORIAN (A Prospective, Randomized Comparison of Subcutaneous and Transvenous Implantable Cardioverter-Defibrillator Therapy) trial (NCT01296022) were excluded from this analysis (13).
The option of screening with an external S-ICD (described subsequently) was discussed with the patients. All patients were informed that the use of this external S-ICD for screening was off-label. Those who agreed, underwent algorithm-based screening. Patients deemed suitable for S-ICD therapy on the basis of the sensing evaluation with the external S-ICD, were implanted with an S-ICD.
The interventions in this study were carried out as part of clinical care and were not part of any study protocol. The need for informed consent was waived by the institutional ethical review board, because of the observational nature of the study.
Subcutaneous ECG morphology screening tool
The ECG morphology screening has been described elsewhere, but briefly: 4 ECG electrodes were placed on the pre-specified points on the chest wall simulating the sensing vectors of the S-ICD (12). A 10-s ECG strip is recorded in both supine and standing position and screened with the dedicated screening tool (Figure 1). A patient is considered eligible for S-ICD implantation when all QRS complexes and T waves in the screening ECGs are found suitable in at least 1 sensing vector in both a supine and standing position during rest. A patient is considered not suitable for S-ICD therapy if they fail this screening, both in the regular left parasternal and alternative right parasternal position.
Algorithm-based screening with the external S-ICD
The external S-ICD was made by the hospital’s technical department of a SQ-RX Pulse Generator (model A1010) and the Q-TRAK Subcutaneous Electrode (model A3010) (both from Cameron Health, San Clemente, California) with an 80 Ω resistance connecting the shock coil to the pulse generator to approximate the chest wall impedance. During the study period the device algorithm updates were installed on the external S-ICD when released. To mimic the 3 sensing vectors (primary, secondary, and alternate), the 2 S-ICD sensing electrodes of the lead and the can of the pulse generator were connected to standard ECG patches, which were applied to the skin at the same position as in regular subcutaneous ECG (S-ECG) screening (Figure 2). Wireless connection was made between the S-ICD and the Q-Tech programmer (Cameron Health) to visualize sensing. Detection zones were programmed at 170 beats/min (conditional zone) and at 250 beats/min (unconditional zone) to achieve the most sensitive setting for dual zone arrhythmia detection.
Captured ECGs for all sensing vectors (primary, secondary and alternate) at standard gain setting and 2-fold gain setting were collected at rest in both supine and standing positions. The ECG signals from the S-ICD were displayed on the programmer and evaluated in real time for undersensing and oversensing. All 3 vectors were evaluated again during exercise in an upright position on a home trainer. The target heart rate during exercise was >150 beats/min and preferably at the maximum predicted heart rate, calculated as 220 beats/min minus age in years, to check for oversensing during sinus tachycardia.
A vector was considered suitable when there was adequate sensing in both supine and the standing position at rest, and during exercise, and during exercise. If more than 1 vector was found suitable, 1 vector was defined as most suitable. Evaluation of the ECG signals was on the basis of QRS amplitude, stability of the QRS amplitude, R-wave to T-wave (R-T) ratio, and morphology changes.
Implantation and follow-up
Implantation of all S-ICD systems was performed in accordance to the prevailing hospital protocol. The most suitable sensing vector during the algorithm-based pre-implantation screening was programmed manually and a signal template at rest was acquired. Two tachycardia therapy zones were programmed at discretion of the implanting physician. One day post-implantation an automatic setup was performed. The automatic setup of the S-ICD automatically identifies the optimal sensing vector. In our patients the vector selected by the S-ICD during the automatic setup was registered, but not necessarily programmed. S-ECGs in all 3 sensing vectors were recorded. These captured S-ECGs were compared to the pre-implantation screening ECGs. All patients underwent a post-implantation exercise test to evaluate all vectors, select the optimal vector on the basis of the R-T ratio, and optimize sensing of the S-ICD in the programmed vector by acquiring a template during exercise as described elsewhere (14). Follow-up visits for all patients were documented and verified using the hospital patient data registration system up to July 2015.
Statistical analysis was performed in R, version 3.0.3 (R Foundation for Statistical Computing, Vienna, Austria). Continuous, non-normally distributed variables are displayed as median with corresponding interquartile range (IQR) (25th percentile to 75th percentile). Kaplan-Meier estimates for freedom of inappropriate shocks with corresponding 95% confidence interval were calculated at 3-year follow-up. No interferential statistics were used.
Between 2009 and December 2015, a total of 150 S-ICD patients were implanted in our center that were not enrolled in the PRAETORIAN trial.
In Group 1 there were 8 patients in whom S-ICD therapy was highly preferred on the basis of individual patient conditions and who screened out for S-ICD implantation due to an unacceptable S-ECG morphology as determined by using the manual screening tool (Table 1). The median age was 53 years (IQR: 37 to 62 years) and for 5 of these 8 patients the indication for ICD therapy was primary prevention of sudden cardiac death. Furthermore, 4 patients had a concomitant pacemaker or cardiac resynchronization pacemaker in situ. The contraindication for transvenous ICD therapy was limited or fully absent venous access in 2 patients, baffle obstruction in transposition of the great arteries in 2 patients, recurrent transvenous lead and pocket complications (infections and lead fractures) in 2 patients, and refusal of transvenous ICD therapy in 2 patients. All 8 patients were screened using the external S-ICD.
In Group 2 there were 7 patients who screened in using the standard screening method and who were additionally evaluated with the external S-ICD (Table 2). The median age was 30 years (IQR: 28 to 45 years) and for 4 of these 7 patients the indication for ICD-therapy was primary prevention of sudden cardiac death. One patient in this group with transposition of the great arteries had a transvenous pacemaker in situ at the time of screening. The diagnosis of 3 patients was hypertrophic cardiomyopathy and 3 patients had nonischemic cardiomyopathies: 1 toxic, 1 dilated, and 1 genetic of origin. These patients were additionally screened for 3 specific reasons. First, in patients in whom increase of the T-wave during exercise may result in T-wave oversensing (e.g., hypertrophic cardiomyopathy patients). Second, in patients with low-amplitude QRS complexes of which the amplitude may further decrease during exercise. Third, in patients with a concomitant pacemaker in situ that results in different QRS complex morphologies during intrinsic and paced rhythms.
Screening with the external S-ICD
Sensing was analyzed in all the 3 sensing vectors in supine and standing position at rest and in upright position during exercise. Six of 8 patients in Group 1 were deemed suitable for S-ICD implantation on the basis of screening with the external S-ICD and were subsequently implanted with a S-ICD. Table 1 presents the baseline characteristics of the patients in Group 1. Two patients were deemed unsuitable for S-ICD implantation after algorithm-based screening with the external S-ICD. One patient with limited venous access and 1 with recurrent lead failures were successfully implanted with a transvenous ICD. In all suitable patients, the most optimal sensing vector determined by the external S-ICD, remained the sensing vector of choice post-implantation. Figure 3 shows 2 patients who screened out with the standard screening method. One had adequate sensing with the external S-ICD and was implanted with an S-ICD and the other one demonstrated oversensing with the external S-ICD and was not implanted.
In Group 2, 5 of 7 patients who screened in with the standard screening method also screened in when evaluated with the external S-ICD (Table 2). There were 3 patients with hypertrophic cardiomyopathy who, during exercise testing, showed minor changes in T-wave morphology that did not result in QRS T-wave double counting and were deemed suitable for S-ICD therapy. The right panel of Figure 4 shows an example of 1 of the hypertrophic cardiomyopathy patients. The fourth patient with a toxic cardiomyopathy due to chemotherapy, low amplitude QRS T-wave and a small R-T ratio was exercised with the external S-ICD to ensure adequate sensing. As sensing in all 3 vectors was adequate, this patient was implanted with an S-ICD. The fifth patient had a history of transposition of the great arteries and an abdominal single chamber pacemaker in situ for intermittent AV-conduction disease. During exercise testing and simultaneous algorithm-based screening with the external S-ICD this patient newly presented with a sick sinus syndrome and therefore a dual chamber pacemaker was indicated. However this patient did pass screening with the external S-ICD. Due to the sick sinus syndrome he was implanted with a dual-chamber ICD, but would have been eligible for S-ICD therapy in the absence of sick sinus syndrome.
Two patients in Group 2 screened out with the external S-ICD. One patient with genetic cardiomyopathy, low-amplitude QRS complexes, and a low R-T ratio was exercised with the external S-ICD and showed over- and undersensing in all 3 vectors, both in rest and during exercise. Moreover, the QRS complex amplitude in this patient decreased during exercise (Figure 4) (Patient C). The second patient that screened out with the external S-ICD also had a genetic cardiomyopathy, low amplitude QRS complexes and large P waves. This patient was deemed not suitable as the QRS complex amplitude decreased during exercise and the large P waves resulted in oversensing. Both patients were implanted with a transvenous ICD.
The median follow-up of the 6 implanted patients in Group 1 was 17 months (interquartile range: 8 to 35 months). Ventricular arrhythmias were detected in 2 patients: Patient #1 had an untreated nonsustained ventricular tachycardia and Patient #4 received 2 successful shocks for ventricular fibrillation. Patient #8 received an inappropriate shock 34 days post-implantation. The amplitude of the sensed QRS complexes had decreased, resulting in a poor R-T ratio and T-wave oversensing. The pacing configuration was changed from left ventricle tip to can in left ventricle ring to right ventricle ring, which resulted in a better R-T ratio, excellent biventricular pacing, and adequate sensing by the S-ICD. Furthermore, a morphology template during (pacemaker driven) tachycardia of 125 beats/min was stored as reference. The patient remained free of inappropriate shocks during the next 7 months of follow-up. The Kaplan-Meier estimate for freedom of inappropriate shocks in Group 1 was 83.3% (58% to 100%) at 3-year follow-up. In Group 2, the median follow-up was 10 months (interquartile range: 4 to 17 months) and no appropriate or inappropriate shocks occurred during follow-up.
This proof of concept study provides several important findings. First it demonstrates the feasibility of algorithm-based screening using the device’s discrimination algorithm in an external S-ICD in patients who fail standard screening with the manufacturer’s dedicated tool. Six of 8 patients who failed standard screening, were deemed suitable for S-ICD therapy after screening with the external S-ICD. Second, in patients with abnormal baseline ECGs, such as high-amplitude T waves, low-amplitude QRS complexes, large P waves, or changes in QRS morphology during pacing, 2 of 7 were deemed not suitable after algorithm-based screening. Although this is not conclusive evidence, it likely that these 2 patients would have been at increased risk of inappropriate shocks by the S-ICD and are therefore potentially better off with a transvenous ICD.
The third important finding is that algorithm-based screening with the external S-ICD can be performed at rest and during pacemaker or exercise induced sinus tachycardia. Also, the identified optimal vector pre-implantation was predictive of the optimal sensing vector post-implantation. Moreover, only 1 inappropriate shock occurred in the group that initially screened out with standard screening during a median follow-up of 17 months and 2 episodes of ventricular fibrillation were successfully detected and treated.
False positive and false negative screening results
The sensing algorithm of the S-ICD depends on ECG morphology characteristics, specifically the R-wave amplitude, T-wave amplitude, R-T ratio, QRS duration, and QT interval (9). The pre-implant standard screening is used to assure appropriate sensing and consequentially to prevent inappropriate shocks post-implantation. Zeb et al. (9) reported a sensitivity of 95% and specificity of 79% of the screening tool in patients with structurally normal hearts, but significantly lower sensitivity and similar specificity in patients with congenital heart disease (84% and 79%). The S-ICD user’s manual excludes implantation in patients who screen out using the dedicated screening tool, this being the case in 7% to 11% in previously published cohorts (10,11).
Patients with a false negative screening result are potentially wrongfully withheld from S-ICD therapy. We here provide evidence that the number of false negative screening results can be reduced using algorithm-based screening, because 6 of 8 patients with negative screening results were eventually suitable for S-ICD therapy. In this cohort of patients with a (relative) contraindication for transvenous ICD therapy, false negative screening results may unnecessarily increase the risk of ICD therapy, as transvenous ICD therapy remains the only option.
False positive screening outcomes place patients at increased risk of inappropriate shocks. The rate of false positive results with standard screening method is likely to be higher than the rate false negative results due to the low specificity of the screening tool (9). In this study, 7 patients with abnormal baseline ECGs were specifically selected due to concerns over sensing, of which 2 of 7 screened out. If these patients would have been implanted with an S-ICD after standard screening, they would have represented only 1% to 2% of the total group of implanted patients. However, there is no data on the majority of patients who were not additionally screened after passing the standard screening method.
A major advantage of using the external S-ICD is the ability to do algorithm-based screening both at rest and during exercise- or pacemaker-induced tachycardia. It will therefore provide more information about the actual performance of the sensing algorithm than the standard screening tool. Also a template of the ECG morphology during tachycardia can be programmed to further optimize sensing with the external S-ICD (14). The manufacturer’s screening tool can only be used up to a heart rate of 148 beats/min, because the sensing algorithm changes the duration of the refractory period and the amplitude detection threshold above this frequency. With algorithm-based screening device performance can be analyzed during heart rate induced changes in sensitivity of the discrimination algorithm and posture dependent variation of QRS complexes and T waves.
In all 6 patients in Group 1 who received an S-ICD the selected optimal vector pre-implantation was identical to the post-implantation optimal sensing vector. Unfortunately, 1 patient received an inappropriate shock shortly after implantation due to decreased R-T ratio and subsequent T-wave oversensing. The R-T ratio was increased by a different pacing configuration of the concomitantly present cardiac resynchronization pacemaker. This effectively prevented recurrence episode of inappropriate shocks.
Clinical and future perspectives
The screening tool is reported to have a high sensitivity of 95%, and is therefore good at what screening tests are designed for: to rule out patients not suitable for S-ICD therapy (9). However, a second confirmation test is currently lacking. This test should have a high specificity in order to reduce the number of false positive results of the screening test.
Algorithm-based screening with the external S-ICD may hold benefits for the future, as it could potentially reduce the number of false negative and false positive screening results. To do so, the algorithm-based screening must have a higher test accuracy than standard screening method when compared to the gold standard: the implanted S-ICD. Prospective studies are needed to determine the test accuracy of algorithm-based screening and assess whether it will reduce the inappropriate shock rate.
Given the positive experience in this cohort with algorithm-based screening and the potential for improved screening in all ICD patients we encourage the manufacturer of the S-ICD to further investigate this new screening method. The usability would increase if algorithm-based screening can be incorporated into the device programmer to avoid the need for technical modification of S-ICDs for this purpose.
First, it is an observational study describing the experience of 1 hospital. Second, the limited number of patients included could result in unreliable estimates of the inappropriate shock rate reported. Third, algorithm-based screening with the external S-ICD has not yet been validated and was performed off-label. Fourth, the extent to which the transcutaneous and subcutaneous sensed signals correspond is currently unknown and needs further examination. Also the implanted device needs to be in the same position as in which screening was performed. Last, to perform algorithm-based screening, technical modifications to the S-ICD have to be made as described in the Methods section and shown in Figure 2. Because the manufacturer does not supply an S-ICD with these modifications only a hospital with an experienced technical department will be able to modify an S-ICD for this purpose.
Algorithm-based screening with the external S-ICD may improve patient selection by reducing the number of false positive and false negative screening results of standard screening. The performance of the sensing algorithm with the external S-ICD was predictive for actual performance of the sensing algorithm after implant.
COMPETENCY IN MEDICAL KNOWLEDGE: Prerequisite ECG morphology screening for S-ICD therapy with the standard screening tool results in some patients in false negative and false positive test outcomes. Replacement of the current screening tool with algorithm-based screening can potentially reduce the number of false screening results.
TRANSLATIONAL OUTLOOK: The algorithm-based screening method for S-ICD therapy should be further investigated for clinical outcomes such as inappropriate shocks. The usability of algorithm-based screening would increase if it can be incorporated in the S-ICD programmer.
Dr. Brouwer, PA Kooiman, Dr. Olde Nordkamp, and Dr. Knops are investigators of the actively recruiting PRAETORIAN trial. PA Kooiman has received honoraria from Boston Scientific. Dr. Knops has served as a consultant for and received research grant support from Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Brouwer and Kooiman contributed equally to this work.
- Abbreviations and Acronyms
- implantable cardioverter-defibrillator
- interquartile range
- R-wave to T-wave
- subcutaneous implantable cardioverter-defibrillator
- subcutaneous electrocardiogram
- Received January 6, 2016.
- Revision received February 12, 2016.
- Accepted February 18, 2016.
- American College of Cardiology Foundation