Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Boston Scientific Corporation Background The current Subcutaneous ICD (S-ICD) model incorporates SMART Pass (SP) to improve sensing and discrimination capabilities to reduce inappropriate shocks (IAS). SP status is programmable but may also be disabled automatically in the setting of low amplitude signals or low heart rate in order to avoid under-sensing of VT/VF. Objective To evaluate SP impact on IAS, appropriate shocks (AS), complications and mortality in the UNTOUCHED S-ICD trial. Methods Primary prevention patients (pts, n=1111) with ejection fraction ≤35% and no pacing requirement were followed for up to 18 months. SP status during a study visit was programmed ON or OFF and status between visits was either consistently OFF, ON, or automatically disabled (DIS). The impact of SP status on pt outcomes was evaluated using Kaplan-Meier (K-M) analysis. Multivariable proportional hazard analysis identified predictors of IAS and SP disable events. Results Percent of pts with SP always ON, always OFF, ON with DIS, and OFF then ON with no DIS were 56, 16, 15, and 13%, respectively. At least one SP DIS occurred in 177 pts, but only 13% had 2 or more, mostly due to PVCs and low EGM amplitudes. Significant multivariable predictors of SP disable events are history of atrial fibrillation (hazard ratio (HR) 2.49, odds ratio (OR) (1.49-4.16); p=.0005), only one passing vector at S-ICD screening, (HR 1.85, OR (1.10-3.10; p=.0202) and lower left ventricular ejection fraction (HR 1.05, OR (1.01-1.08); p=.0074). K-M IAS rates were highest for pts experiencing DIS (fig 1) and lowest for SP ON. While neither AS (p=0.58) nor complication (p=0.58) rates varied significantly according to SP status, mortality was lower for pts with SP ON during any duration of time (p=0.044) by univariate analysis. Further analysis is planned to better understand the relationship between SP status and mortality. Conclusion Patients in the UNTOUCHED trial with SMART Pass (SP) consistently ON had significantly fewer inappropriate shocks, with no impact on appropriate therapy for VT/VF. Patients with history of atrial fibrillation, lower left ventricular ejection fraction, and only one passing vector at S-ICD screening are at higher risk of SP disable events; therefore, care should be taken for these patients to assess SP status and their higher risk for inappropriate shocks.

Background The electrocardiogram (ECG) is commonly used, but most recent rhythm discrimination algorithms still lack both specificity and sensitivity. Deep learning techniques have shown promising results in the classification of physiological signals like ECGs. Purpose To develop and test a deep learning (DL) model to discriminate between atrial fibrillation (AF) and sinus rhythm (SR). Methods For the development of the DL model we used 1499 ECGs sampled at 500 Hz of patients diagnosed with AF. All ECGs were labeled by two experienced investigators. Only ECGs labeled as SR or AF were included in the dataset. To simplify the learning process, solely the first ECG channel was used. The ECG waveforms were preprocessed using the Fourier cosine series to correct for baseline wander. Input data was generated by normalizing and scaling all different heartbeats by centralizing the R peak, leading to 15744 single heart beat samples of 80 data points (figure A). Multiple feedforward architectures were tested with different numbers of layers, filters and activation functions. The models were trained by equally splitting the data (50%SR, 50%AF) in a training (65%), validation (25%) and test set (15%). The best performing model was chosen based on the accuracy. Results A total of 1469 ECGs (1061 (72%)SR, 408 (28%)AF) were included. The model with the best performance was a feedforward model consisting three dense layers with ReLU activation and four dense layers with Linear activation. Training of the model was performed in 32 epochs. Validation of the model resulted in an accuracy of 96% (figure B), precision of 95% and recall of 96%. Conclusions The morphology based deep learning model developed in this study was able to discriminate atrial fibrillation from sinus rhythm with a fairly high accuracy using a limited size dataset and only one lead.

Funding Acknowledgements Type of funding sources: None. Background Routine defibrillation testing during implant and replacement of the subcutaneous implantable cardioverter-defibrillator (S-ICD) is recommended per current guidelines. Recently, concerns have been raised about an increase in shock impedance and consequent shock failure during defibrillation testing in S-ICD patients undergoing a generator replacement. Purpose We aim to describe the defibrillation success rate in relation to the shock impedance in patients undergoing S-ICD generator replacement in our large tertiary center. Methods In this retrospective analysis, data from replacement procedures were collected from all patients who underwent an S-ICD generator replacement in our center from June 2014 to December 2020. Defibrillation testing was performed with at least one shock of ≤65J, and a successful shock was defined as terminating the ventricular arrhythmia within 5 seconds after the shock. Results A total of 133 patients underwent an S-ICD generator replacement, 5.8 ± 0.9 years after initial implant. Reasons for replacement were: reaching of elective replacement indicator (n = 119), early battery depletion (n = 9), complaints of generator pocket (n = 3) and device malfunction (n = 2). Defibrillation testing was performed in 111 patients (86.5%) undergoing a replacement procedure. Shock impedance data from both the implant and replacement procedure were available in 101 patients. The median shock impedance of these patients during their replacement procedure was significantly higher than during their implant, 79Ω (IQR 66-94) and 66Ω (IQR 57.5-81) respectively (Z = -5.552, p < 0.001). Despite the higher shock impedance, first shock during defibrillation testing was successful in 105/111 patients (94.6%), with a success rate of 97.3% after two attempts. In the remaining three patients, the ventricular arrhythmia could only be terminated with a 80J shock. This was the case during both their initial implant and their replacement procedure. Shock impedance increase between implant and replacement was not significantly higher in patients with a successful first shock compared to patients with an unsuccessful first shock (Δ+11.1 ± 20.0Ω versus Δ+12.7 ± 27.6Ω, p = 0.86). Conclusion In this large retrospective analysis, we have shown a first shock success rate during S-ICD generator replacement of 94.6%, which is similar to the success rate of defibrillation testing after initial implant. After multiple attempts, defibrillation testing success rate was 100%. Even though the median shock impedance during replacement was significantly higher than during the initial implant, there was no difference in impedance increase in patients with a successful shock compared to patients with an unsuccessful shock. Abstract Figure. Defibrillation success

Funding Acknowledgements Type of funding sources: None. Background Implantable cardioverter-defibrillator (ICD) therapy is associated with the risk of inappropriate shocks (IAS). IAS, defined as any shock on a different rhythm than VT or VF,  cause psychological stress, decrease the quality of life and may provoke ventricular arrhythmias. In the subcutaneous ICD (S-ICD) the majority of IAS are caused by T-wave oversensing (TWOS), often during exercise. Exercise-optimized programming during an exercise ECG test (X-ECG) after implantation has shown to be successful in reducing IAS in patients known with TWOS. In recent years, new discrimination algorithms in the latest generation S-ICDs have significantly reduced the risk of TWOS. The benefit of performing an X-ECG in these latest generation S-ICDs to reduce IAS is unclear. Purpose We aim to describe the effect of exercise-optimized programming after S-ICD implantation on inappropriate shock rate in the latest generation S-ICDs. Methods In this retrospective multicenter study, data were collected from two experienced S-ICD hospitals in the Netherlands. All patients underwent an S-ICD implantation of second or third generation between February 2015 and December 2020. Patients younger than 21 years were excluded. Patients with an X-ECG after implantation were compared with patients without X-ECG after implantation. Total number of patients with IAS and cause of the first IAS were evaluated. Results In total, 262 patients were included in the X-ECG group and 61 in the no X-ECG group. The median follow-up time was 22 months in the X-ECG group (IQR 9-33) and 23 months in the no X-ECG group (IQR 12-33, P = 0.9). Mean age was 51 ± 15 years and 61 ± 15 years respectively (P< 0.001). Primary prevention indication was similar in both groups (56% for the X-ECG group versus 49% for the no X-ECG group, P = 0.4). A total of 8 patients (3.1%) experienced IAS in the X-ECG group; 3 first shocks (1.15%) were due to TWOS, 2 (0.8%) were given on a SVT and 3 (1.15%) on other non-cardiac activity. In the no X-ECG group, 6 patients (9.8%) experienced IAS; 1 first shock (1.6%) was due to TWOS, 4 (6.6%) were given on a SVT and 1 (1.6%) on other non-cardiac activity. Patients with an X-ECG had a significantly lower risk of IAS compared to patients  in the no X-ECG group (hazard ratio 0.32; 95% CI 0.1 to 0.9; P = 0.027). The Kaplan-Meier estimate of IAS-free survival for the X-ECG group was 61 months (95% CI 59 to 62) and 50 months (95% CI 46 to 55) for the no X-ECG group. Results are shown in the figure. Conclusion This study shows that, in the latest generation S-ICDs, exercise-optimized programming after S-ICD implantation results in a significantly lower risk of IAS in adults. Patients with an X-ECG after S-ICD implantation were younger,  which may have affected the outcome. Further prospective data with more equal groups is necessary. Until then, we recommend considering exercise testing after S-ICD implantation in the latest generation S-ICDs. Abstract Figure.