Ectopy-triggering ganglionated plexuses ablation to prevent atrial fibrillation: GANGLIA-AF study

BACKGROUND

The ganglionated plexuses (GPs) of the intrinsic cardiac autonomic system may play a role in atrial fibrillation (AF).

OBJECTIVE

We hypothesized that ablating the ectopy-triggering GPs (ET-GPs) prevents AF.

METHODS

GANGLIA-AF (ClinicalTrials.gov identifier NCT02487654) was a prospective, randomized, controlled, 3-center trial. ET-GPs were mapped using high frequency stimulation, delivered within the atrial refractory period and ablated until nonfunctional. If triggered AF became incessant, atrioventricular dissociating GPs were ablated. We compared GP ablation (GPA) without pulmonary vein isolation (PVI) against PVI in patients with paroxysmal AF. Follow-up was for 12 months including 3-monthly 48-hour Holter monitors. The primary end point was documented ≥30 seconds of atrial arrhythmia after a 3-month blanking period.

RESULTS

A total of 102 randomized patients were analyzed on a per-protocol basis after GPA (n = 52; 51%) or PVI (n = 50; 49%). Patients who underwent GPA had 89 ± 26 high frequency stimulation sites tested, identifying a median of 18.5% (interquartile range 16%-21%) of GPs. The radiofrequency ablation time was 22.9 ± 9.8 minutes in GPA and 38 ± 14.4 minutes in PVI (P < .0001). The freedom from ≥30 seconds of atrial arrhythmia at 12-month follow-up was 50% (26 of 52) with GPA vs 64% (32 of 50) with PVI (log-rank, P = .09). ET-GPA without atrioventricular dissociating GPA achieved 58% (22 of 38) freedom from the primary end point. There was a significantly higher reduction in antiarrhythmic drug usage postablation after GPA than after PVI (55.5% vs 36%; P = .05). Patients were referred for redo ablation procedures in 31% (16 of 52) after GPA and 24% (12 of 50) after PVI (P = .53).

CONCLUSION

GPA did not prevent atrial arrhythmias more than PVI. However, less radiofrequency ablation was delivered to achieve a higher reduction in antiarrhythmic drug usage with GPA than with PVI.

Prognostic Significance of Ventricular Arrhythmias in 13 444 Patients With Acute Coronary Syndrome: A Retrospective Cohort Study Based on Routine Clinical Data (NIHR Health Informatics Collaborative VA-ACS Study)

Background A minority of acute coronary syndrome (ACS) cases are associated with ventricular arrhythmias (VA) and/or cardiac arrest (CA). We investigated the effect of VA/CA at the time of ACS on long-term outcomes. Methods and Results We analyzed routine clinical data from 5 National Health Service trusts in the United Kingdom, collected between 2010 and 2017 by the National Institute for Health Research Health Informatics Collaborative. A total of 13 444 patients with ACS, 376 (2.8%) of whom had concurrent VA, survived to hospital discharge and were followed up for a median of 3.42 years. Patients with VA or CA at index presentation had significantly increased risks of subsequent VA during follow-up (VA group: adjusted hazard ratio [HR], 4.15 [95% CI, 2.42-7.09]; CA group: adjusted HR, 2.60 [95% CI, 1.23-5.48]). Patients who suffered a CA in the context of ACS and survived to discharge also had a 36% increase in long-term mortality (adjusted HR, 1.36 [95% CI, 1.04-1.78]), although the concurrent diagnosis of VA alone during ACS did not affect all-cause mortality (adjusted HR, 1.03 [95% CI, 0.80-1.33]). Conclusions Patients who develop VA or CA during ACS who survive to discharge have increased risks of subsequent VA, whereas those who have CA during ACS also have an increase in long-term mortality. These individuals may represent a subgroup at greater risk of subsequent arrhythmic events as a result of intrinsically lower thresholds for developing VA.

EP-PINNs: Cardiac Electrophysiology Characterisation Using Physics-Informed Neural Networks

Accurately inferring underlying electrophysiological (EP) tissue properties from action potential recordings is expected to be clinically useful in the diagnosis and treatment of arrhythmias such as atrial fibrillation. It is, however, notoriously difficult to perform. We present EP-PINNs (Physics Informed Neural Networks), a novel tool for accurate action potential simulation and EP parameter estimation from sparse amounts of EP data. We demonstrate, using 1D and 2D in silico data, how EP-PINNs are able to reconstruct the spatio-temporal evolution of action potentials, whilst predicting parameters related to action potential duration (APD), excitability and diffusion coefficients. EP-PINNs are additionally able to identify heterogeneities in EP properties, making them potentially useful for the detection of fibrosis and other localised pathology linked to arrhythmias. Finally, we show EP-PINNs effectiveness on biological in vitro preparations, by characterising the effect of anti-arrhythmic drugs on APD using optical mapping data. EP-PINNs are a promising clinical tool for the characterisation and potential treatment guidance of arrhythmias.

Point-of-care screening for heart failure with reduced ejection fraction using artificial intelligence during ECG-enabled stethoscope examination in London, UK: a prospective, observational, multicentre study

Background

Most patients who have heart failure with a reduced ejection fraction, when left ventricular ejection fraction (LVEF) is 40% or lower, are diagnosed in hospital. This is despite previous presentations to primary care with symptoms. We aimed to test an artificial intelligence (AI) algorithm applied to a single-lead ECG, recorded during ECG-enabled stethoscope examination, to validate a potential point-of-care screening tool for LVEF of 40% or lower.

Methods

We conducted an observational, prospective, multicentre study of a convolutional neural network (known as AI-ECG) that was previously validated for the detection of reduced LVEF using 12-lead ECG as input. We used AI-ECG retrained to interpret single-lead ECG input alone. Patients (aged ≥18 years) attending for transthoracic echocardiogram in London (UK) were recruited. All participants had 15 s of supine, single-lead ECG recorded at the four standard anatomical positions for cardiac auscultation, plus one handheld position, using an ECG-enabled stethoscope. Transthoracic echocardiogram-derived percentage LVEF was used as ground truth. The primary outcome was performance of AI-ECG at classifying reduced LVEF (LVEF ≤40%), measured using metrics including the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity, with two-sided 95% CIs. The primary outcome was reported for each position individually and with an optimal combination of AI-ECG outputs (interval range 0–1) from two positions using a rule-based approach and several classification models. This study is registered with ClinicalTrials.gov, NCT04601415.

Findings

Between Feb 6 and May 27, 2021, we recruited 1050 patients (mean age 62 years [SD 17·4], 535 [51%] male, 432 [41%] non-White). 945 (90%) had an ejection fraction of at least 40%, and 105 (10%) had an ejection fraction of 40% or lower. Across all positions, ECGs were most frequently of adequate quality for AI-ECG interpretation at the pulmonary position (979 [93·3%] of 1050). Quality was lowest for the aortic position (846 [80·6%]). AI-ECG performed best at the pulmonary valve position (p=0·02), with an AUROC of 0·85 (95% CI 0·81–0·89), sensitivity of 84·8% (76·2–91·3), and specificity of 69·5% (66·4–72·6). Diagnostic odds ratios did not differ by age, sex, or non-White ethnicity. Taking the optimal combination of two positions (pulmonary and handheld positions), the rule-based approach resulted in an AUROC of 0·85 (0·81–0·89), sensitivity of 82·7% (72·7–90·2), and specificity of 79·9% (77·0–82·6). Using AI-ECG outputs from these two positions, a weighted logistic regression with l2 regularisation resulted in an AUROC of 0·91 (0·88–0·95), sensitivity of 91·9% (78·1–98·3), and specificity of 80·2% (75·5–84·3).

Interpretation

A deep learning system applied to single-lead ECGs acquired during a routine examination with an ECG-enabled stethoscope can detect LVEF of 40% or lower. These findings highlight the potential for inexpensive, non-invasive, workflow-adapted, point-of-care screening, for earlier diagnosis and prognostically beneficial treatment.

Funding

NHS Accelerated Access Collaborative, NHSX, and the National Institute for Health Research.

Wearable In-Ear PPG: Detailed Respiratory Variations Enable Classification of COPD

An ability to extract detailed spirometry-like breath-ing waveforms from wearable sensors promises to greatly improve respiratory health monitoring. Photoplethysmography (PPG) has been researched in depth for estimation of respiration rate, given that it varies with respiration through overall intensity, pulse amplitude and pulse interval. We compare and contrast the extraction of these three respiratory modes from both the ear canal and finger and show a marked improvement in the respiratory power for respiration induced intensity variations and pulse amplitude variations when recording from the ear canal. We next employ a data driven multi-scale method, noise assisted multivariate empirical mode decomposition (NA-MEMD), which allows for simultaneous analysis of all three respiratory modes to extract detailed respiratory waveforms from in-ear PPG. For rigour, we considered in-ear PPG recordings from healthy subjects, both older and young, patients with chronic obstructive pulmonary disease (COPD) and idiopathic pulmonary fibrosis (IPF) and healthy subjects with artificially obstructed breathing. Specific in-ear PPG waveform changes are observed for COPD, such as a decreased inspiratory duty cycle and an increased inspiratory magnitude, when compared with expiratory magnitude. These differences are used to classify COPD from healthy and IPF waveforms with a sensitivity of 87% and an overall accuracy of 92%. Our findings indicate the promise of in-ear PPG for COPD screening and unobtrusive respiratory monitoring in ambulatory scenarios and in consumer wearables.

Structure and function of the ventricular tachycardia isthmus

Catheter ablation of postinfarction reentrant ventricular tachycardia (VT) has received renewed interest owing to the increased availability of high-resolution electroanatomic mapping systems that can describe the VT circuits in greater detail, and the emergence and need to target noninvasive external beam radioablation. These recent advancements provide optimism for improving the clinical outcome of VT ablation in patients with postinfarction and potentially other scar-related VTs. The combination of analyses gleaned from studies in swine and canine models of postinfarction reentrant VT, and in human studies, suggests the existence of common electroanatomic properties for reentrant VT circuits. Characterizing these properties may be useful for increasing the specificity of substrate mapping techniques and for noninvasive identification to guide ablation. Herein, we describe properties of reentrant VT circuits that may assist in elucidating the mechanisms of onset and maintenance, as well as a means to localize and delineate optimal catheter ablation targets.

Neural networks applied to 12-lead electrocardiograms predict body mass index, visceral adiposity and concurrent cardiometabolic ill-health

Background

Obesity is associated with electrophysiological remodeling, which manifests as detectable changes on the surface electrocardiogram (ECG).

Objective

To develop neural networks (NN) to predict body mass index (BMI) from ECGs and test the hypothesis that discrepancies between NN-predicted BMI and measured BMI are indicative of underlying adiposity and/or concurrent cardiometabolic ill-health.

Methods

NN models were developed using 36,856 12-lead resting ECGs from the UK Biobank. Two architectures were developed for continuous and categorical BMI estimation (normal weight [BMI <25 kg/m²] vs overweight/obese [BMI ≥25 kg/m²]). Models for male and female participants were trained and tested separately. For each sex, data were randomly divided into 4 folds, and models were evaluated in a leave-1-fold-out manner.

Results

ECGs were available for 17,807 male and 19,049 female participants (mean ages: 61 ± 7 and 63 ± 8 years; mean BMI 26 ± 5 kg/m² and 27 ± 4 kg/m², respectively). NN models detected overweight/obese individuals with average accuracies of 75% and 73% for male and female subjects, respectively. The magnitudes of difference between NN-predicted BMI and actual BMI were significantly correlated with visceral adipose tissue volumes. Concurrent hypertension, diabetes, dyslipidemia, and/or coronary heart disease explained false-positive classifications (ie, calculated BMI <25 kg/m² misclassified as ≥25 kg/m² by NN model, P < .001).

Conclusion

NN models applied to 12-lead ECGs predict BMI with a reasonable degree of accuracy. Discrepancies between NN-predicted and calculated BMI may be indicative of underlying visceral adiposity and concomitant cardiometabolic perturbation, which could be used to identify individuals at risk of cardiometabolic disease.

A Multicenter External Validation of a Score Model to Predict Risk of Events in Patients With Brugada Syndrome

A multivariate risk score model was proposed by Sieira et al in 2017 for sudden death in Brugada syndrome; their validation in 150 patients was highly encouraging, with a C-index of 0.81; however, this score is yet to be validated by an independent group. A total of 192 records of patients with Brugada syndrome were collected from 2 centers in the United Kingdom and retrospectively scored according to a score model by Sieira et al. Data were compiled summatively over follow-up to mimic regular risk re-evaluation as per current guidelines. Sudden cardiac death survivor data were considered perievent to ascertain the utility of the score before cardiac arrest. Scores were compared with actual outcomes. Sensitivity in our cohort was 22.7%, specificity was 57.6%, and C-index was 0.58. In conclusion, up to 75% of cardiac arrest survivors in this cohort would not have been offered a defibrillator if evaluated before their event. This casts doubt on the utility of the score model for primary prevention of sudden death. Inherent issues with modern risk scoring strategies decrease the likelihood of success even in robustly designed tools such as the Sieira score model.

Classification of Fibrillation Organisation Using Electrocardiograms to Guide Mechanism-Directed Treatments

Background: Atrial fibrillation (AF) and ventricular fibrillation (VF) are complex heart rhythm disorders and may be sustained by distinct electrophysiological mechanisms. Disorganised self-perpetuating multiple-wavelets and organised rotational drivers (RDs) localising to specific areas are both possible mechanisms by which fibrillation is sustained. Determining the underlying mechanisms of fibrillation may be helpful in tailoring treatment strategies. We investigated whether global fibrillation organisation, a surrogate for fibrillation mechanism, can be determined from electrocardiograms (ECGs) using band-power (BP) feature analysis and machine learning. Methods: In this study, we proposed a novel ECG classification framework to differentiate fibrillation organisation levels. BP features were derived from surface ECGs and fed to a linear discriminant analysis classifier to predict fibrillation organisation level. Two datasets, single-channel ECGs of rat VF (n = 9) and 12-lead ECGs of human AF (n = 17), were used for model evaluation in a leave-one-out (LOO) manner. Results: The proposed method correctly predicted the organisation level from rat VF ECG with the sensitivity of 75%, specificity of 80%, and accuracy of 78%, and from clinical AF ECG with the sensitivity of 80%, specificity of 92%, and accuracy of 88%. Conclusion: Our proposed method can distinguish between AF/VF of different global organisation levels non-invasively from the ECG alone. This may aid in patient selection and guiding mechanism-directed tailored treatment strategies.

Multicentre validation of point-of-care screening tool for heart failure: single-lead ECG recorded by smart stethoscope predicts low ejection fraction using artificial intelligence

Background/Introduction

Artificial intelligence (AI) applied to 12-lead ECG can identify left ventricular ejection fraction (EF) ≤35% with a sensitivity and specificity of 86.3% and 85.7%, respectively. Whether AI algorithms trained on 12-lead can accurately predict EF from single-lead ECGs (recorded by a smart stethoscope) remains unknown. This could facilitate point-of-care screening for low EF during routine clinical examination.

Purpose

First independent multicentre real-world UK National Health Service (NHS) prospective validation of 12-lead-ECG-trained AI algorithm applied to single-lead ECG recorded by a smart stethoscope, with AI algorithm tuned to detect EF ≤40%.

Methods

Prospective recruitment of unselected patients attending for echocardiography across six urban NHS hospital sites (UK). In addition to transthoracic echocardiogram (routine care), all participants had 15 seconds of supine, single-lead ECG recorded at six different positions (figure), encompassing standard anatomical positions for cardiac auscultation. A convolutional neural network (CNN) previously trained on 35,970 independent pairings of 12-lead-ECG and echocardiograms was retrained to use the single-lead ECG as input. Accuracy of CNN detection of low EF (binary ≤40%) is reported at a threshold of 0.5 against gold-standard; echo-determined percentage EF.

Results

Among 353 patients recruited (mean age 63±17; 58% male, 43.1% non-white), 309 (87.5%) had an EF &gt;40%, and 44 (12.5%) had EF ≤40%. The best single recording position in isolation was position 3 (sensitivity 57.9% [42.2–73.6], specificity 86.3% [82.2–90.3]). Taking any of the six positions performed during the examination as predicting EF ≤40%, this achieved a sensitivity of 81.2% and specificity of 61.5%.

Conclusion(s)

In this first prospective multicentre validation study the retrained AI algorithm reliably detected low EF from single-lead ECGs acquired using a novel ECG-enabled stethoscope in standard auscultation positions. The ability to identify patients with possible low EF during routine physical examination addresses a significant unmet clinical need in point-of-care ruling in/out of heart failure, and has potential to provide broader population-level screening for asymptomatic cardiovascular disease.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): National Institute of Health Research, Accelerated Access Collaborative & NHSX: Artificial Intelligence in Health & Social Care Award

Triage-HF plus: 12-month study of remote monitoring pathway for triage of heart failure risk initiated during the Covid-19 pandemic

Background

The Covid-19 pandemic necessitated rapid adoption of remote monitoring across cardiovascular patient cohorts. Most patients with cardiac implantable electronic devices (CIEDs) are now able to be remotely monitored using either scheduled, patient- or threshold-triggered transmissions. The validated “Triage Heart Failure Risk Score” (Triage-HFRS) is a medical algorithm within company-specific CIEDs that can risk-stratify patients as low-, medium- or high-risk of worsening heart failure (WHF) in the next 30 days based on integrated monitoring of physiological parameters. Building on a previous proof-of-concept of the Triage-HF Plus pathway, we integrated remote data with simple 5-question telephone triage within a clinical pathway to identify WHF during the first year of the Covid-19 pandemic.

Purpose

Prospective evaluation of clinical remote monitoring pathway integrating Triage-HFRS with protocolised telephone triage (Triage-HF Plus pathway).

Methods

Prospective, real-world evaluation of clinical pathway serving a large urban region over a 12-month period, using data from April 2020 to April 2021 (initiated during the first wave of Covid-19 pandemic in the UK). From a population of 435 patients with CIEDs, 87 “high” Triage-HFRS alerts were received and patients contacted for telephone triage assessment. Screening questions were designed to identify episodes of WHF and non-HF events. Intervention was at discretion of the clinical practitioner and in line with guideline-directed practice. A consecutive sample of 115 “medium” risk scores received the same triage.

Results

Successful contact was made with 72 (82.8%) high-risk patients. Classification for high scoring patients confirmed on triage included isolated heart failure (18.3%), heart failure concurrent to medical problem (5.7%), alternative medical problem (10.3%), and recent hospital admission (8.0%); triage reassured absence of acute cause of high score in 40.2%. The sensitivity and specificity for detection of WHF was 87.9% (0.77–0.99) and 59.4% (0.50–0.69) respectively. Positive and negative predictive values were 40.3% and 94.0%, respectively. Overall accuracy was 66.2%.

Conclusions

The Triage-HF Plus pathway served as a useful remote monitoring tool for identifying patients with WHF whose care had been otherwise disrupted by the Covid-19 pandemic, allowing timely intervention and cementing the longer-term role for such models of care delivery. Crucially, in this multimorbid, high-cost population, relevant non-HF issues were also identified. The high negative predictive value further highlights the potential of proactive surveillance over conventional, periodic follow up.

Funding Acknowledgement

Type of funding sources: Foundation. Main funding source(s): Imperial Health Charity

Left ventricular activation time and pattern are preserved with both selective and nonselective His bundle pacing

BACKGROUND

His bundle pacing (HBP) can be achieved in 2 ways: selective HBP (S-HBP), where the His bundle is captured alone, and nonselective HBP (NS-HBP), where local myocardium is also captured, resulting a pre-excited electrocardiogram appearance.

OBJECTIVE

We assessed the impact of this ventricular pre-excitation on left and right ventricular dyssynchrony.

METHODS

We recruited patients who displayed both S-HBP and NS-HBP. We performed noninvasive epicardial electrical mapping for left and right ventricular activation time (LVAT and RVAT) and pattern.

RESULTS

Twenty patients were recruited. In the primary analysis, the mean within-patient change in LVAT from S-HBP to NS-HBP was -5.5 ms (95% confidence interval: -0.6 to -10.4, noninferiority P < .0001). NS-HBP did not prolong RVAT (4.3 ms, -4.0 to 12.8, P = .296) but did prolong QRS duration (QRSd, 22.1 ms, 11.8 to 32.4, P = .0003). In patients with narrow intrinsic QRS (n = 6), NS-HBP preserved LVAT (-2.9 ms, -9.7 to 4.0, P = .331) but prolonged QRS duration (31.4 ms, 22.0 to 40.7, P = .0003) and mean RVAT (16.8 ms, -5.3 to 38.9, P = .108) compared to S-HBP. Activation pattern of the left ventricular surface was unchanged between S-HBP and NS-HBP, but NS-HBP produced early basal right ventricular activation that was not seen in S-HBP.

CONCLUSION

Compared to S-HBP, local myocardial capture during NS-HBP produces pre-excitation of the basal right ventricle resulting in QRS duration prolongation. However, NS-HBP preserves the left ventricular activation time and pattern of S-HBP. Left ventricular dyssynchrony is not an important factor when choosing between S-HBP and NS-HBP in most patients.

Postinfarct ventricular tachycardia substrate: Characterization and ablation of conduction channels using ripple mapping

BACKGROUND

Conduction channels have been demonstrated within the postinfarct scar and seem to be co-located with the isthmus of ventricular tachycardia (VT). Mapping the local scar potentials (SPs) that define the conduction channels is often hindered by large far-field electrograms generated by healthy myocardium.

OBJECTIVE

The purpose of this study was to map conduction channel using ripple mapping to categorize SPs temporally and anatomically. We tested the hypothesis that ablation of early SPs would eliminate the latest SPs without direct ablation.

METHODS

Ripple maps of postinfarct scar were collected using the PentaRay (Biosense Webster) during normal rhythm. Maps were reviewed in reverse, and clusters of SPs were color-coded on the geometry, by timing, into early, intermediate, late, and terminal. Ablation was delivered sequentially from clusters of early SPs, checking for loss of terminal SPs as the endpoint.

RESULTS

The protocol was performed in 11 patients. Mean mapping time was 65 ± 23 minutes, and a mean 3050 ± 1839 points was collected. SP timing ranged from 98.1 ± 60.5 ms to 214.8 ± 89.8 ms post QRS peak. Earliest SPs were present at the border, occupying 16.4% of scar, whereas latest SPs occupied 4.8% at the opposing border or core. Analysis took 15 ± 10 minutes to locate channels and identify ablation targets. It was possible to eliminate latest SPs in all patients without direct ablation (mean ablation time 16.3 ± 11.1 minutes). No VT recurrence was recorded (mean follow-up 10.1 ± 7.4 months).

CONCLUSION

Conduction channels can be located using ripple mapping to analyze SPs. Ablation at channel entrances can eliminate the latest SPs and is associated with good medium-term results.

In vivo grafting of large engineered heart tissue patches for cardiac repair

Engineered heart tissue (EHT) strategies, by combining cells within a hydrogel matrix, may be a novel therapy for heart failure. EHTs restore cardiac function in rodent injury models, but more data are needed in clinically relevant settings. Accordingly, an upscaled EHT patch (2.5 cm × 1.5 cm × 1.5 mm) consisting of up to 20 million human induced pluripotent stem cell–derived cardiomyocytes (hPSC-CMs) embedded in a fibrin-based hydrogel was developed. A rabbit myocardial infarction model was then established to test for feasibility and efficacy. Our data showed that hPSC-CMs in EHTs became more aligned over 28 days and had improved contraction kinetics and faster calcium transients. Blinded echocardiographic analysis revealed a significant improvement in function in infarcted hearts that received EHTs, along with reduction in infarct scar size by 35%. Vascularization from the host to the patch was observed at week 1 and stable to week 4, but electrical coupling between patch and host heart was not observed. In vivo telemetry recordings and ex vivo arrhythmia provocation protocols showed that the patch was not pro-arrhythmic. In summary, EHTs improved function and reduced scar size without causing arrhythmia, which may be due to the lack of electrical coupling between patch and host heart.

HRS White Paper on Clinical Utilization of Digital Health Technology

This collaborative statement from the Digital Health Committee of the Heart Rhythm Society provides everyday clinical scenarios in which wearables may be utilized by patients for cardiovascular health and arrhythmia management. We describe herein the spectrum of wearables that are commercially available for patients, and their benefits, shortcomings and areas for technological improvement. Although wearables for rhythm diagnosis and management have not been examined in large randomized clinical trials, undoubtedly the usage of wearables has quickly escalated in clinical practice. This document is the first of a planned series in which we will update information on wearables as they are revised and released to consumers.

Rapid Exclusion of COVID Infection With the Artificial Intelligence Electrocardiogram

OBJECTIVE

To rapidly exclude severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection using artificial intelligence applied to the electrocardiogram (ECG).

METHODS

A global, volunteer consortium from 4 continents identified patients with ECGs obtained around the time of polymerase chain reaction-confirmed COVID-19 diagnosis and age- and sex-matched controls from the same sites. Clinical characteristics, polymerase chain reaction results, and raw electrocardiographic data were collected. A convolutional neural network was trained using 26,153 ECGs (33.2% COVID positive), validated with 3826 ECGs (33.3% positive), and tested on 7870 ECGs not included in other sets (32.7% positive). Performance under different prevalence values was tested by adding control ECGs from a single high-volume site.

RESULTS

The area under the curve for detection of acute COVID-19 infection in the test group was 0.767 (95% CI, 0.756 to 0.778; sensitivity, 98%; specificity, 10%; positive predictive value, 37%; negative predictive value, 91%). To more accurately reflect a real-world population, 50,905 normal controls were added to adjust the COVID prevalence to approximately 5% (2657/58,555), resulting in an area under the curve of 0.780 (95% CI, 0.771 to 0.790) with a specificity of 12.1% and a negative predictive value of 99.2%.

CONCLUSION

Infection with SARS-CoV-2 results in electrocardiographic changes that permit the artificial intelligence-enhanced ECG to be used as a rapid screening test with a high negative predictive value (99.2%). This may permit the development of electrocardiography-based tools to rapidly screen individuals for pandemic control.

Determinants of Shielding Behaviour During the COVID-19 Pandemic and Associations with Wellbeing in >7,000 NHS Patients: 17-week Longitudinal Observational Study

BACKGROUND

The UK National Health Service (NHS) classified 2.2 million people as clinically extremely vulnerable (CEV) during the first wave of the 2020 COVID-19 pandemic, advising them to 'shield' - to not leave home for any reason.

OBJECTIVE

The aim of this study was to measure the determinants of shielding behaviour and associations with wellbeing in a large NHS patient population, towards informing future health policy.

METHODS

Patients contributing to an ongoing longitudinal participatory epidemiology study (LoC-19, n = 42,924) received weekly email invitations to complete questionnaires (17-week shielding period starting 9th April 2020) within their NHS personal electronic health record. Question items focused on wellbeing. Participants were stratified into four groups by self-reported CEV status (qualifying condition) and adoption of shielding behaviour (baselined at week 1 or 2). Distribution of CEV criteria is reported alongside situational variables and uni- and multivariable logistic regression. Longitudinal trends in physical and mental wellbeing were displayed graphically. Free-text responses reporting variables impacting wellbeing were semi-quantified using natural language processing. In the lead up to a second national lockdown (October 23rd, 2020), a follow-up questionnaire evaluated subjective concern if further shielding were advised.

RESULTS

7,240 participants were included. Among the CEV (2,391), 1,133 (47.3%) assumed shielding behaviour at baseline, compared with 633 (15.0%) in the non-CEV group. Those CEV who shielded were more likely to be Asian (Odds Ratio OR 2.02 [1.49-2.76]), female (OR 1.24 [1.05-1.45]), older (OR per year increase 1.01 [1.00-1.02]) and live in a home with outdoor space (OR 1.34 [1.06-1.70]) or 3-4 other inhabitants (3 = OR 1.49 [1.15-1.94], 4 = OR 1.49 [1.10-2.01]); and be solid organ transplant recipients (2.85 [2.18-3.77]) or have severe chronic lung disease (OR 1.63 [1.30-2.04]). Receipt of a government letter advising shielding was reported in 1,115 (46.6%) of CEV and 180 (3.7%) of non-CEV and was associated with adopting shielding behaviour (OR 3.34 [2.82-3.95] and 2.88 [2.04-3.99], respectively). In both groups, shielding was longitudinally associated with worse physical and mental wellbeing (p<.05). Access to food and grocery supplies was a more prevalent concern among those shielding (p<.05). Concern for wellbeing if future shielding was required was most prevalent among the CEV who had originally shielded.

CONCLUSIONS

Future health policy must balance the potential protection from COVID-19 against our findings that in this population shielding may have negatively impacted wellbeing and was adopted in many in whom it was not indicated, and variably in whom it was. This therefore also requires clearer public health messaging and support for wellbeing if shielding is to be advised in future pandemic scenarios.

CLINICALTRIAL

Septal late gadolinium enhancement on Cardiac MRI predicts failure to achieve left bundle pacing

Funding Acknowledgements

Type of funding sources: Foundation. Main funding source(s): British Heart Foundation

Background; Left bundle area pacing is a novel technique that provides direct stimulation of cardiac conduction tissue in order to deliver physiological ventricular activation. The approach for left bundle area pacing is transseptal lead implantation, where the lead is advanced from the right ventricular side of the septum to the left ventricular side to capture the proximal left bundle. Observational data suggests that whilst this is a safe and feasible method, implant success rate is not 100%, and appears to be lower in patients with a cardiac resynchronization therapy (CRT) indication rather than a bradycardia indication for pacing. The mechanisms for failure to advance the lead through the ventricular septum are not well understood.

Purpose; We used pre-procedural CMR to determine whether there are features which can help identify patients where lead implantation may be challenging.  We assessed whether the extent and location of septal late gadolinium enhancement identified patients in whom left bundle area pacing will be challenging.  We hypothesized that the presence of extensive scar in the septum impedes advancing the lead to the left ventricular septum and prevents capture of the left bundle.

Methods; Patients underwent cardiac MRI including motion corrected free-breathing late gadolinium enhancement imaging1 before implantation.  Scar was quantified using the full height half maximum method and expressed as the overall proportion of myocardial mass in the basal anteroseptal and basal inferoseptal segments, as shown in Figure 1.  Left bundle area pacing was then attempted in patients with a CRT indication for pacing.  We compared the extent of septal scar between patients in whom left bundle area pacing was achieved and those where there was failure to advance the lead deep into the septum.

Results; 12 patients (11 male, 1 female), with average age 72 (IQR 63 to 78) and LVEF 30% (IQR 26 to 33) were studied.  There was failure to advance the lead deep into the septum in 4 patients.  There was a significantly higher basal septal scar burden in those patients where there was failure to advance the left bundle lead compared to those in which left bundle capture was achieved as shown in Figure 2 (median 55% and 5% respectively, p-value 0.02 by Wilcoxon signed rank test).

Conclusion; The presence and extent of late gadolinium enhancement in the basal septum appears to be an important determinant of successful implantation of left bundle pacing lead using current implant technology. This may be because extensive septal scar prevents advancement of the pacing lead through the septum. Cardiac MRI before left bundle area pacing is likely to be useful in procedural planning.

Increasing adiposity and the presence of cardiometabolic morbidity is associated with increased Covid-19-related mortality: results from the UK Biobank

BACKGROUND

Although obesity, defined by body mass index (BMI), has been associated with a higher risk of hospitalisation and more severe course of illness in Covid-19 positive patients amongst the British population, it is unclear if this translates into increased mortality. Furthermore, given that BMI is an insensitive indicator of adiposity, the effect of adipose volume on Covid-19 outcomes is also unknown.

METHODS

We used the UK Biobank repository, which contains clinical and anthropometric data and is linked to Public Health England Covid-19 healthcare records, to address our research question. We performed age- and sex- adjusted logistic regression and Chi-squared test to compute the odds for Covid-19-related mortality as a consequence of increasing BMI, and other more sensitive indices of adiposity such as waist:hip ratio (WHR) and percent body fat, as well as concomitant cardiometabolic illness.

RESULTS

13,502 participants were tested for Covid-19 (mean age 70 ± 8 years, 48.9% male). 1582 tested positive (mean age 68 ± 9 years, 52.8% male), of which 305 died (mean age 75 ± 6 years, 65.5% male). Increasing adiposity was associated with higher odds for Covid-19-related mortality. For every unit increase in BMI, WHR and body fat, the odds of death amongst Covid19-positive participants increased by 1.04 (95% CI 1.01-1.07), 10.71 (95% CI 1.57-73.06) and 1.03 (95% CI 1.01-1.05), respectively (all p < 0.05). Referenced to Covid-19 positive participants with a normal weight (BMI 18.5-25 kg/m²), Covid-19 positive participants with BMI > 35 kg/m² had significantly higher odds of Covid-19-related death (OR 1.70, 95% CI 1.06-2.74, p < 0.05). Covid-19-positive participants with metabolic (diabetes, hypertension, dyslipidaemia) or cardiovascular morbidity (atrial fibrillation, angina) also had higher odds of death.

CONCLUSIONS

Anthropometric indices that are more sensitive to adipose volume and its distribution than BMI, as well as concurrent cardiometabolic illness, are associated with higher odds of Covid-19-related mortality amongst the UK Biobank cohort that tested positive for the infection. These results suggest adipose volume may contribute to adverse Covid-19-related outcomes associated with obesity.

Explanation-visualised deep learning model for accessory pathway localisation using 12-lead electrocardiography

Funding Acknowledgements

Type of funding sources: Public Institution(s). Main funding source(s): British Heart Foundation Imperial Centre of Research Excellence

Background/Introduction

ECG algorithms for identifying accessory pathway (AP) locations are inaccurate and difficult to use. Human expert interpretation is poorly reproducible. Artificial intelligence (AI) techniques such as machine learning can improve accuracy in classification tasks by eschewing theory-driven predictions. More reproducible and accurate AP localisation could shorten procedure time and personalise ablation consent.

Purpose

We developed a neural network to perform AP localisation using 12-lead ECGs. Its decision-making process was analysed to enable explainability of the neural network.

Methods

A convolutional neural network was trained on raw, digital, intra-procedural 12-lead ECGs of patients with manifest APs who underwent successful ablation. ECGs were labelled with AP locations as left-sided, septal or right-sided using procedure reports, fluoroscopy and electro-anatomical maps. Accuracy of the neural network was assessed via 4-fold cross-validation and was compared to the Arruda algorithm. Five cardiologists were also assessed for their accuracy in determining locations in sub-groups of cases. The neural network was retrospectively analysed to identify areas of ECGs most influential to its predictions using importance mapping.

Results 

In 156 cases, accuracy of the neural network (92.9%) was significantly higher than the Arruda algorithm (76.9%; p &lt; 0.0001) and all five cardiologists (37.5% to 65.9%; p = 0.0001 to 0.0290). Importance mapping demonstrated that the QRS complexes of leads aVL and V1 were perceived as most influential, indicating interrogation of the lateral and anterior-posterior axes respectively.

The figure shows (A) architecture of the neural network, (B) accuracy of the neural network, Arruda algorithm and five cardiologists, (*, p = 0.05 – 0.01; **, p = 0.01 – 0.001; ***, p = 0.001 - 0.0001; ****, p &lt; 0.0001; as compared to the neural network) and (C) example importance maps for 12-lead ECGs of left-sided, septal and right-sided APs (in order from left to right), with darker regions corresponding to greater relative importance. 

Conclusion 

AI ECG interpretation allows accurate, reproducible prediction of AP locations, superior to conventional algorithms and human interpretation. Although AI decision-making is thought of as a ‘black box’, explanation visualisation techniques such as importance mapping allow humans to understand aspects of how a neural network make decisions. A prospectively validated neural network could be integrated into clinical practice to improve pre-procedural AP localisation. Abstract Figure. Summary of results

Laser doppler derived peripheral perfusion can distinguish haemodynamically tolerated VT from haemodynamically compromised VT

Funding Acknowledgements

Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): NIHR Imperial Biomedical Research Centre

Introduction

Implantable Cardioverter-Defibrillators (ICDs) cannot distinguish between ventricular tachycardia (VT) with haemodynamic compromise from haemodynamically tolerated VT to ensure that therapies are delivered only when necessary. Currently, unnecessary therapies are reduced by longer duration thresholds and higher rate thresholds. This can result in ICDs withholding or delaying therapies during haemodynamically compromising VT while potentially still providing therapies during rapid or prolonged VT that is haemodynamically well tolerated.

Laser doppler perfusion monitoring (LDPM) allows assessment of peripheral blood flow as a surrogate for haemodynamic status. We have previously demonstrated that laser doppler perfusion signals, analysed using an electro-mechanical coupling algorithm (SafeShock), can reliably identify loss of perfusion during ventricular fibrillation (VF), as well as discriminate VF from simulated lead fractures and T wave over-sensing. The utility of LDPM signals in VT, however, has not been established.

Purpose

In this study we assessed the utility of LDPM using the SafeShock algorithm to discriminate haemodynamically tolerated VT from VT with haemodynamic compromise.

Methods

Recruited participants underwent a rapid ventricular pacing protocol to simulate VT at different rates. Pacing was performed using the right ventricular lead of an implanted pacing device or a temporary pacing wire in the right ventricular apex. 3-lead ECG, blood pressure (either invasively using a radial artery catheter or non-invasively using beat-by-beat finometry) and LDPM signal were continuously recorded during the protocol. LDPM signals during simulated VT were analysed using the SafeShock electro-mechanical algorithm and compared to blood pressure change from baseline intrinsic rhythm to simulated VT.

Results

We obtained 588 recordings of simulated VT in 56 patients at rates of 100 bpm, 120 bpm, 140 bpm, 160 bpm, 180 bpm and 200 bpm. Percentage change in systolic blood pressure from baseline to VT correlated with LDPM-derived perfusion value during VT (Spearman’s Rho = 0.7786, p &lt; 0.0001).

Using a cut-off of 5 units, perfusion value predicted a 20% drop in systolic blood pressure in VT with an accuracy of 89.4% (sensitivity 94.8%, specificity 83.6%, p value &lt;0.0001).

Conclusions

Peripheral perfusion measurements, analysed using an electro-mechanical algorithm, can accurately discriminate haemodynamically tolerated VT from VT with haemodynamic compromise. ICDs with integrated LDPM sensors and algorithms could make therapy decisions based on the circulatory status of patients with arrhythmias not just rate and duration parameters. This could reduce unnecessary therapies while facilitating prompt treatment of compromising arrhythmias. Abstract Figure 1

Non-selective and selective His bundle pacing both preserve left ventricular activation time and pattern

Funding Acknowledgements

Type of funding sources: Public Institution(s). Main funding source(s): British Heart Foundation

Background: His bundle pacing can be achieved in two ways

selective His bundle pacing, where the His bundle is captured alone, and non-selective His bundle pacing, where local myocardium is also captured resulting a pre-excited ECG appearance. We assessed the impact of this ventricular pre-excitation on left and right ventricular dys-synchrony.

Methods

We recruited patients who displayed both selective and non-selective His bundle pacing. We performed non-invasive epicardial electrical mapping to determine left and right ventricular activation times and patterns.

Results

In the primary analysis (n = 20, all patients), non-selective His bundle pacing did not prolong LVAT compared to select His bundle pacing by a pre-specified non-inferiority margin of 10ms (LVAT prolongation: -5.5ms, 95% confidence interval (CI): -0.6 to -10.4, non-inferiority p &lt; 0.0001). Non-selective His bundle pacing did not prolong right ventricular activation time (4.3ms, 95%CI: -4.0 to 12.8, p = 0.296) but did prolong QRS duration (22.1ms, 95%CI: 11.8 to 32.4, p = 0.0003).

In patients with narrow intrinsic QRS (n = 6), non-selective His bundle pacing preserved left ventricular activation time (-2.9ms, 95%CI: -9.7 to 4.0, p = 0.331) but prolonged QRS duration (31.4ms, 95%CI: 22.0 to 40.7, p = 0.0003) and mean right ventricular activation time (16.8ms, 95%CI: -5.3 to 38.9, p = 0.108) compared to selective His bundle pacing.

Activation pattern of the left ventricular surface was unchanged between selective and non-selective His bundle pacing. Non-selective His bundle pacing produced early basal right ventricular activation, which was not observed with selective His bundle pacing.

Conclusions

Compared to selective His bundle pacing, local myocardial capture during non-selective His bundle pacing produces right ventricular pre-excitation resulting in prolongation of QRS duration. However, non-selective His bundle pacing preserves the left ventricular activation time and pattern of selective His bundle pacing. When choosing between selective and non-selective His bundle pacing, left ventricular dyssynchrony is not an important factor. Abstract Figure: Selective vs Non-Selective HBP