Aims Patients with cardiac disease are considered high risk for poor outcomes following hospitalization with COVID-19. The primary aim of this study was to evaluate heterogeneity in associations between various heart disease subtypes and in-hospital mortality. Methods and results We used data from the CAPACITY-COVID registry and LEOSS study. Multivariable Poisson regression models were fitted to assess the association between different types of pre-existing heart disease and in-hospital mortality. A total of 16 511 patients with COVID-19 were included (21.1% aged 66–75 years; 40.2% female) and 31.5% had a history of heart disease. Patients with heart disease were older, predominantly male, and often had other comorbid conditions when compared with those without. Mortality was higher in patients with cardiac disease (29.7%; n = 1545 vs. 15.9%; n = 1797). However, following multivariable adjustment, this difference was not significant [adjusted risk ratio (aRR) 1.08, 95% confidence interval (CI) 1.02–1.15; P = 0.12 (corrected for multiple testing)]. Associations with in-hospital mortality by heart disease subtypes differed considerably, with the strongest association for heart failure (aRR 1.19, 95% CI 1.10–1.30; P < 0.018) particularly for severe (New York Heart Association class III/IV) heart failure (aRR 1.41, 95% CI 1.20–1.64; P < 0.018). None of the other heart disease subtypes, including ischaemic heart disease, remained significant after multivariable adjustment. Serious cardiac complications were diagnosed in <1% of patients. Conclusion Considerable heterogeneity exists in the strength of association between heart disease subtypes and in-hospital mortality. Of all patients with heart disease, those with heart failure are at greatest risk of death when hospitalized with COVID-19. Serious cardiac complications are rare during hospitalization.

Aims Patients with cardiac disease are considered high risk for poor outcomes following hospitalization with COVID-19. The primary aim of this study was to evaluate heterogeneity in associations between various heart disease subtypes and in-hospital mortality. Method and results We used data from the CAPACITY-COVID registry and LEOSS study. Multivariable modified Poisson regression models were fitted to assess the association between different types of pre-existent heart disease and in-hospital mortality. 10,481 patients with COVID-19 were included (22.4% aged 66-75 years; 38.7% female) of which 30.5% had a history of cardiac disease. Patients with heart disease were older, predominantly male and more likely to have other comorbid conditions when compared to those without. COVID-19 symptoms at presentation did not differ between these groups. Mortality was higher in patients with cardiac disease (30.3%; n=968 versus 15.7%; n=1143). However, following multivariable adjustment this difference was not significant (adjusted risk ratio (aRR) 1.06 [95% CI 0.98-1.15, p-value 0.13]). Associations with in-hospital mortality by heart disease subtypes differed considerably, with the strongest association for NYHA III/IV heart failure (aRR 1.43 [95% CI 1.22-1.68, p-value <0.001]) and atrial fibrillation (aRR 1.14 [95% CI 1.04-1.24, p-value 0.01]). None of the other heart disease subtypes, including ischemic heart disease, remained significant after multivariable adjustment. Conclusion There is considerable heterogeneity in the strength of association between heart disease subtypes and in-hospital mortality. Of all patients with heart disease, those with severe heart failure are at greatest risk of death when hospitalized with COVID-19.

Background and purpose The electrocardiogram (ECG) is frequently obtained in the work-up of COVID-19 patients. So far, no study has evaluated whether ECG-based machine learning models have added value to predict in-hospital mortality specifically in COVID-19 patients. Methods Using data from the CAPACITY-COVID registry, we studied 882 patients admitted with COVID-19 across seven hospitals in the Netherlands. Raw format 12-lead ECGs recorded within 72 h of admission were studied. With data from five hospitals (n = 634), three models were developed: (a) a logistic regression baseline model using age and sex, (b) a least absolute shrinkage and selection operator (LASSO) model using age, sex and human annotated ECG features, and (c) a pre-trained deep neural network (DNN) using age, sex and the raw ECG waveforms. Data from two hospitals (n = 248) was used for external validation. Results Performances for models a, b and c were comparable with an area under the receiver operating curve of 0.73 (95% confidence interval [CI] 0.65–0.79), 0.76 (95% CI 0.68–0.82) and 0.77 (95% CI 0.70–0.83) respectively. Predictors of mortality in the LASSO model were age, low QRS voltage, ST depression, premature atrial complexes, sex, increased ventricular rate, and right bundle branch block. Conclusion This study shows that the ECG-based prediction models could be helpful for the initial risk stratification of patients diagnosed with COVID-19, and that several ECG abnormalities are associated with in-hospital all-cause mortality of COVID-19 patients. Moreover, this proof-of-principle study shows that the use of pre-trained DNNs for ECG analysis does not underperform compared with time-consuming manual annotation of ECG features.

Funding Acknowledgements Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Horizon Introduction There is a growing interest in the use of accelerometers and sensors embedded in implantable cardioverter-defibrillators (ICDs) for monitoring patient activity. Despite evidence regarding the potential clinical value of device-measured activity (D-PA), the validity of these measurements has not yet been established. Objective To assess the validity of device-measured activity against a research-grade, widely validated wearable accelerometer. Methods This is a subanalysis of the ongoing multicenter, prospective SafeHeart study. Raw accelerometry data was continuously sampled at 50Hz from a wrist-worn accelerometer (GENEActiv) during 12 months. Days with at least 22 hours of wear time were used to create summary measures of time in activity, daily active volume and total slow walking steps. These measures were compared to D-PA harmonised as percentage of active time per day, from four different ICD vendors’ remote transmission data, using linear mixed effect models. Results Wearable and device-measured activity data in 51 ICD patients rendered 1228 days (mean 24 days ±19) with both wearable and device-measured activity data. There were significant differences between wearable and device-measured accelerometery in the average time active per day (Table 1). For two vendors significant associations between D-PA, daily active volume, and total slow walking steps were observed. Also, associations between D-PA and daily active time and moderate vigorous physical activity were found in a third vendor. For the fourth vendor no association between any wearable activity metric and D-PA was found. Inter-patient differences accounted for 73.1% of the total variance in D-PA. Conclusion Results demonstrate substantial differences in device-measured activity measurements compared to research-grade activity data. This has implications for the utility and generalizability of D-PA as clinical parameter.

Funding Acknowledgements Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Horizon2020 Introduction Wearable devices are gaining interest in the clinical assessment of physical behavior as a marker of disease severity. With the increased use, patient willingness and adherence will be increasingly important. As part of the SafeHeart study, examining the potential of physical behavior as an identifier of clinical deterioration in patients with an implantable cardioverter defibrillator (ICD), we present preliminary results on adherence to a wrist-worn wearable used for physical behavior assessment. Purpose Define the willingness to participate and long-term adherence to wearables in an ICD population. Methods This is a preliminary analysis of the ongoing multicenter, prospective, observational SafeHeart study. SafeHeart is aimed to construct a personalized prediction engine for ICD therapy using wearable-assessed physical behavior, remote ICD monitoring, electronic health records, and patient-reported data. The study will enroll 400 participants with an ICD with or without cardiac resynchronization therapy (CRT-D). In this preliminary analysis, wearable data was analyzed for the first 50 participants, where inclusion required a minimum of 1 month of follow up data. No data from the wearables were provided to the participants. The wrist-worn wearables were used continuously (day and night) for up to 12 months of follow-up. Adherence to the wearable was measured through patient-reported (subjective) adherence and wearable-measured (objective) adherence. Data were extracted from the wearables and non-wear time was detected via open source algorithms. A valid day was set to 22 hours of available wear time with 24-hour periods assessed from 3pm to 3pm for sleep metric capture. The willingness to participate and dropout rates were calculated for the same first 50 patients of the study. Results A total of 50 ICD participants were included in this study. The mean age was 65.1 years, 82 % male, with a mean follow up of 7 weeks, generating 326 patient weeks of data. Regarding patient-reported adherence, participants reported 81.4% full adherence and 18.6 % of participants reported very brief non-wear due to e.g. sauna or surgery. Of those reporting non-wear, 62.5% described one episode only of non-wear lasting 15-75 minutes. Regarding objectively measured adherence from wearable data, full adherence was shown in 91.7% of days. The mean number of valid days per participant was 41.3. Recruitment rates showed a willingness to participate of 50% (50/100) out of eligible subjects invited. No participants were lost to follow Conclusion Results show high adherence and reasonable willingness to participate without wearable adherence dropping over time. Comparison of objectively measured and patient-reported adherence showed similar values.

Funding Acknowledgements Type of funding sources: None. Background Remote monitoring (RM) for implantable cardioverter defibrillators (ICDs) is recommended as the standard of care in clinical guidelines. Presumably, the restrictions on face-to-face visits that were imposed during the coronavirus (COVID-19) pandemic have further accelerated the adoption of RM. However, quantitative real-world data on the uptake of RM during the COVID-19 pandemic is lacking. Purpose To assess the uptake of RM during the COVID-19 pandemic to a pre-COVID-19 period, and compare the arrhythmic burden between the two groups. Methods This is a substudy of the retrospective, observational single-center DISTANT-study. For this substudy, data from patients who were enrolled in the RM program after de novo ICD implantation (single- and dual chamber, biventricular or subcutaneous ICDs) were analysed. The time until RM was initiated per patient was calculated for patient implanted during the COVID-19 pandemic (March 2020-January 2021) and compared to a similar 10-month period pre-COVID-19 (May 2019-March 2020). ICD therapy (shock and/or anti-tachycardia pacing), non-sustained ventricular tachycardia (NSVT), supraventricular tachycardia (SVT) and mortality were registered for each patient. Patients &lt;18 years old at implantation and patients with a follow-up of &lt;6 months were excluded from this analysis. Results A total of 134 patients (72.4 % male, mean age 57.3 ±14.9 years) were eligible for this substudy, of which 61 patients in the COVID-19 group and 73 patients in the pre-COVID-19 group. In both groups there was a similar percentage of primary prevention ICD implantations (COVID-19: 43%, pre-COVID: 44%; p=0.888). During COVID-19, RM was initiated more promptly following ICD implantation compared to pre-COVID-19 (respectively 63 days vs. 131 days; p=0.007). Second, in the COVID-19 group 60.7% patients were enrolled in RM within 30 days following implantation compared to 39.7% in the pre-COVID-19 group (p=0.016). In terms of arrhythmic burden, no differences in the occurrence of ICD therapy (p=0.759), NSVT (p=0.267) and SVT (p=0.454) were observed. Conclusion During the COVID-19 pandemic RM was initiated more promptly following ICD implantation compared to before the pandemic, however, no differences in arrhythmic burden between groups were observed.

Funding Acknowledgements Type of funding sources: Public grant(s) – EU funding. Main funding source(s): Eurostars Introduction Patients at a high risk of sudden cardiac death (SCD) benefit from an implantable cardioverter defibrillator (ICD). However, they remain at a high risk of (inappropriate) shocks, heart failure, mortality and psychological distress. Consumer-level wearable accelerometry as method for recording physical behaviour (PB) has gained popularity over the past years, but so far the clinical potential is largely underinvestigated. The identification of patterns in PB and the association with clinical outcomes may provide a means to improve ICD therapy. Purpose This review addresses the evidence concerning PB in ICD patients and aims to characterise PB patterns associated with clinical outcomes. Methods A systematic review of studies focussing on accelerometer-assessed PB in patients older than 18 years equipped with an ICD, or patients at a high risk of SCD (e.g. advanced heart failure) was performed. PB could be assessed using a wearable accelerometer or an embedded accelerometer in the ICD (i.e. device-measured physical activity (D-PA)). Papers presenting quantitative data in English language peer reviewed journals published between January 2000 and September 2020 were identified via the OVID MEDLINE and OVID EMBASE databases. A study protocol describing study selection, data charting and summarisation of results was developed apriori. Study selection was conducted by two independent reviewers and a third reviewer in case of disagreement. Results A total of 4219 studies were identified, of which 51 were deemed appropriate for this review. Of these studies, 29 examined D-PA (n = 169.742 patients), 19 examined wearable accelerometery (n = 1.601) and 3 validated wearable accelerometry against D-PA (n = 106). The main findings were that (i) a low level of physical activity (PA) after implantation of the ICD and (ii) a decline in physical activity were both associated with an increased risk of ICD shocks, hospitalization and mortality. Second, PB was affected by cardiac factors (e.g. onset of atrial arrhythmias, ICD shocks) and non-cardiac factors (e.g. seasonal differences, pandemic lockdown). Third, PB was related to left ventricular ejection fraction, physical and cognitive function and quality of life. The evidence regarding wearable accelerometry compared to D-PA was scarce and heterogeneous. Conclusion This review demonstrated the potential of PB as an identifier of clinical deterioration in an ICD population. Accelerometer-assessed PB data could improve early warning systems and facilitate preventive and pro-active strategies, especially considering the nature of PB as modifiable risk factor. We suggest two directions for future research: (i) prospective collection of wearable accelerometry data in an ICD population to identify the most clinically relevant behavioural metrics (ii) investigation of preventive measures that can be undertaken once changes in PB are observed. Abstract Figure. Accelerometry-derived physical behaviour

Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NWO Rubicon (452019308) Amsterdam Cardiovascular Sciences Background Left Ventricular Ejection Fraction (LVEF) is suboptimal as a sole marker for predicting sudden cardiac death (SCD) or benefit from an ICD. Machine (ML) and deep (DL) learning provide new opportunities for personalised predictions using complex, multi-modal physiological data. Objective We hypothesise that risk stratification for ICD implantation can be improved by ML and DL models that combine clinical variables with time series features from 12-lead electrocardiograms (ECG). Methods We present a multicentre study of 1010 patients with an ischaemic, dilated or non-ischaemic cardiomyopathy and LVEF≤35% implanted with an ICD between 2007 and 2021 for primary prevention of SCD (64.9 ±10.8 years, 73.2% male) in two academic hospitals. For each patient, raw 12-lead, 10-second ECG-recordings obtained &lt;90 days before ICD implantation and clinical details were collected. Supervised ML and DL models were trained and validated using stratified k-fold cross-validation (5 repeats, k=10) on a development cohort (n=550) from Hospital A. We used this model to predict ICD non-benefit defined as mortality without prior appropriate ICD-therapy on an external dataset from Hospital B (n=460). Kaplan-Meier survival analysis stratified by high vs. low predicted probability using Youden's J statistic as cut-off was performed. Results At 3-year follow-up, 16.0% of patients had died of whom 72.8% met the criteria for ICD non-benefit. Extreme gradient boosting models identified subjects with ICD non-benefit with an area under the receiver operator curve (AUROC) of 0.897 ±0.05 during internal validation (Figure 1, solid line). In the external cohort, the AUROC was 0.793 (95% CI 0.75-0.84) (Figure 1, dashed line). Survival analysis for low vs. high predicted risk indicated ICD non-benefit rates of 6.0% versus 30.3% at 3-year follow-up, respectively (Figure 2). Conclusion A ML model that combined clinical with ECG features better predicted ICD non-benefit at 3-years in a primary prevention population than currently available risk scores such as the MADIT-ICD score. This approach may provide new tools to support personalized decision making for ICD therapy. Prospective validation is needed to assess the real-world clinical performance of the model.