Single-lead ECG Research Articles

Telephone training to improve ECG quality in remote screening for atrial fibrillation.

&#xD;Self-recorded, single-lead ECGs are increasingly used to diagnose arrhythmias. However, they can be of variable quality, affecting the reliability of interpretation. In this analysis of ECGs collected in atrial fibrillation screening studies, our aims were to: (i) determine the quality of ECGs when recorded unsupervised; and (ii) investigate whether telephone training improved ECG quality.&#xD;&#xD;Approach:&#xD;Data was obtained from the Screening for Atrial Fibrillation with ECG to Reduce stroke (SAFER) programme, where participants recorded four single-lead ECG traces per day for three weeks using a handheld device. ECG quality was assessed by an automated algorithm, and participants who recorded >25% poor-quality ECGs from days 4-10 of screening were identified for training to improve ECG recording technique. Training was delivered when research team capacity permitted.&#xD;&#xD;Main results:&#xD;13,741 participants recorded 1,127,264 ECGs, of which 41,288 (3.7%) were poor-quality. Most participants (51.5%) didn't record any poor-quality ECGs. 1,088 (7.9%) participants met the threshold for training. Of these, 165 participants received training and 923 didn't. The median proportion of poor-quality ECGs per participant on days 1-3 was 41.7 (27.3-50.0) % for those who received training and 33.3 (25.0-45.5) % for those who didn't. On days 11-21, the median proportions of poor-quality ECGs per participant were significantly lower (p<0.001) for those who received training, 17.8 (5.0-31.6) %, and those who didn't, 14.0 (4.8-30.2) %. Comparing these groups, the mean (95% confidence interval) reduction in proportion of poor-quality ECGs from days 1-3 to days 11-21 was 20.2 (16.8-23.5) % in those who received training and 16.0 (14.7-17.3) % in those who didn't (p=0.396). &#xD;&#xD;Significance:&#xD;Most participants achieved adequate quality ECGs. For those that didn't, ECG quality improved over time regardless of whether they received telephone training. Telephone training may therefore not be required to achieve improvements in ECG quality during screening.&#xD.

Long-term clinical outcomes informed by artificial intelligence-enabled single-lead ECG detection of reduced ejection fraction

Abstract Background * Reduced left ventricular ejection fraction (LVEF) is linked to poor clinical outcomes (1, 2). * Despite its increasing prevalence, reduced LVEF is often undetected due to resource-limited echocardiography and lack of suitable point-of-care diagnostic tools. * Previous studies have shown that an artificial intelligence-enabled electrocardiogram (AI-ECG) algorithm can accurately identify patients with reduced LVEF, but may also offer novel insights on long-term prognosis (3, 4). Purpose We aimed to determine if an AI-ECG algorithm designed for detecting reduced LVEF can also independently predict 2-year major adverse cardiovascular events (MACE) and all-cause mortality. Method * A retrospective multicentre observational study investigating the ability of AI-ECG to predict long-term clinical outcomes (i.e. over two years' follow up). * Analysis included a total of 1,007 consecutive unselected patients attending for routine echocardiography, with a single-lead ECG recording performed at the same time. * An AI-ECG algorithm, designed to identify impaired LVEF, was applied to each single-lead ECG. * Occurrence of MACE and all-cause mortality associated with AI-ECG results (Figure 1) was investigated using Cox regression; evaluating performance of AI-ECG results as a classifier (0 or 1) and as a probability score (0 to 1). * The study was approved by the UK Health Research Authority (reference 21/LO/0051). Results * The mean age was 62.3 years. 52.4% of patients were male and 57.5% of patients were White. * Appropriately, patients with a positive AI-ECG had a higher MACE rate compared to those with a negative AI-ECG (34.2% vs. 11.9%; adjusted hazard ratio (aHR) 2.02; 95% CI, 1.46 – 2.81; p &amp;lt; 0.005), primarily driven by increased mortality (23% vs. 9.6%; p &amp;lt; 0.0001; aHR 1.65; 95% CI, 1.12 – 2.43) and heart failure hospitalization (14.5% vs. 1%; p &amp;lt; 0.0001) – Figure 2. * AI-ECG probability score (per 10% increase) was significantly associated with MACE (aHR 1.17; 95% CI, 1.09 – 1.25; p &amp;lt; 0.005) and all-cause mortality (aHR 1.11; 95% CI, 1.02 – 1.20; p = 0.01). * Importantly, sub-analysis of AI-ECG for those who have a normal LVEF (i.e. ≥50%) continued to show association between positive AI-ECG and MACE (27.2% vs. 11.9%; p &amp;lt; 0.0001; aHR 1.70, 95% CI 1.17 – 2.48) and all-cause mortality (20.4% vs. 9.6%; p &amp;lt; 0.0001; aHR 1.53, 95% CI 1.00 – 2.36). Conclusion An AI-ECG algorithm designed to identify impaired LVEF can identify patients at risk of MACE and all-cause mortality from single-lead ECG screening – regardless of their LVEF. Such results may enable further risk stratification for cardiovascular investigations through the simple addition of single-lead ECG recording.

Predicting 1-year atrial fibrillation risk from a patch-based ambulatory ECG monitoring without atrial fibrillation

Abstract Introduction Early identification of atrial fibrillation (AF) can enable interventions that reduce cardiovascular events. However, AF remains undiagnosed in about 10-20% of affected individuals, and paroxysmal AF often remains undetected despite the use of in-clinic ECGs, consumer smartwatch ECGs, or ambulatory ECG monitoring. We therefore investigated whether ambulatory ECG patch-based monitoring without AF can predict AF in the subsequent 1-year post monitoring. Purpose This study aims to validate a machine-learning model for AF prediction within 1-year post-wear, based on the observation of non-AF single-lead ECG recordings. Methods We considered a clinically monitored cohort of 67,296 individuals wearing multiple single-lead ECG multi-day ambulatory monitoring patches as part of real-world care. For all selected individuals, the first patch recorded no occurrences of AF and was worn for a minimum of 6.5 days. We assess the risk of future AF based on the raw ECG, heartrate variability, frequency of ectopic beats, age and gender (ECG model), which is then compared with the risk assessment by a model based on age and gender (Baseline model). We then observe AF incidence on subsequent patches worn within one year from the first patch. We evaluated the model's accuracy by observing the number of AF cases identified (or missed) among the individuals with highest risk according to the model. Results In the clinically monitored cohort, a total of 3463 individuals experienced AF within the next year (incidence 5.1%) after a first AF-free monitoring period of a minimum 6.5 days. The receiver operating characteristic (ROC) curve shows an area under the curve (AUC) for the ECG model of 0.75 (95% CI: [0.74, 0.76]), and for the Baseline model of 0.66 (CI: [0.65, 0.67]). When selecting the top 20% (13,459) participants with the highest risk and verifying who developed AF in the following 1-year, we obtain positive prediction value (PPV) or precision of 12% and sensitivity (recall) of 46% for the ECG model (which correctly identifies 1,594 individuals with incident AF), and of 8% and 29%, respectively, for the Baseline model (p-value&amp;lt;0.001). (Figure) Conclusion A machine learning model using as input an AF-free single-lead ECG can quantify the risk of observing AF in a population including individuals with low burdens of paroxysmal or new onset of AF within 1 year. Further research and validation are needed to enhance the interpretability and clinical applicability of the model, while prediction may be further improved with clinical AF diagnoses ascertained from other data sources. The findings of this study have important implications for the early detection and management of AF, potentially enabling earlier intervention and improved outcomes in individuals with undiagnosed AF.

Detection of ATTR cardiac amyloidosis using a novel artificial intelligence algorithm for wearable-adapted noisy single-lead electrocardiograms

Abstract Background Amyloid cardiomyopathy (ATTR-CM) is often underdiagnosed, with very few diagnosed before the onset of symptoms, leading to delayed use of disease-modifying therapies that could have substantial prognostic implications early in the disease course. In the US, the disease has a substantially higher prevalence among Black adults, who are further challenged by lower access to healthcare services and lower preventative care, further reducing the ability for clinician-led diagnosis. We sought to develop an approach for identifying ATTR-CM on noisy single-lead ECGs obtainable on portable and wearable ECG devices, allowing broad use in community screening. Purpose To develop a portable/wearable device-adapted AI-ECG algorithm capable of detecting ATTR-CM on noisy single-lead ECGs. Methods We identified patients with ATTR-CM from 2015 through June 2023 across 5 hospitals of a large U.S.-based hospital system using positive bone scintigraphy scans or pharmacotherapy with an approved transthyretin stabilizer. In the development cohort, we used lead I of 12 lead ECGs, commonly captured by portable and wearable ECG devices, to train a novel noise-adapted model explicitly designed to infer information against simulated real-world noises. For this, ECG signals were augmented using random Gaussian real-world noise within four frequency ranges at multiple signal-to-noise ratios. We tested the model in an independent set of cases and matched controls drawn from July through December 2023. Results The development cohort consisted of 1,011 ECGs from 234 patients (median age 79 years [IQR:70-85], 17% women), with 10:1 age- and sex-matched 10,110 controls with 10,110 ECGs (age 79 [IQR:70-85] years 17.7% female). In the independent test set of 139 ECGs from 47 patients (age 80 [75-86] years, 32% women) and matched controls (1390 ECGs and patients) from July through December 2023, the AUROC (area under the receiver operating characteristic curve) of the AI-ECG model for ATTR-CM was 0.90 [95%CI: 0.88-0.92] (A). The AUPRC was 0.44 [0.36-0.53]. The sensitivity and specificity of the model were 0.85 and 0.80. The positive predictive value ranged from 0.12 at a 3% prevalence level to 0.59 at a 25% prevalence (B). Conclusions A novel portable and wearable-devices adapted model demonstrates promising performance for detecting ATTR-CM on single-lead ECGs, representing a more accessible method for screening in community-dwelling adults.

Flecainide acetate inhalation solution for cardioversion of recent-onset, symptomatic atrial fibrillation: results of the phase 3 RESTORE-1 trial

Abstract Introduction Safe, rapid, and efficacious drugs for acute cardioversion of paroxysmal AF (PAF) are needed. In an open-label, Phase 2 study (INSTANT), orally inhaled flecainide was safe and efficacious for acute conversion of PAF to sinus rhythm (SR).[1, 2] Purpose Evaluate the efficacy and safety of flecainide acetate inhalation solution (FlecIH-103) compared to placebo in patients with short-duration, symptomatic, newly diagnosed, or recurrent PAF. Methods RESTORE-1 was a multicentre, randomized, double-blind, placebo-controlled trial in patients with symptomatic PAF (≤48 hours). Patients received FlecIH-103 at an estimated total lung dose of 120 mg or placebo (3:1) over 8 minutes using a breath-actuated nebulizer and were continuously monitored for 90 minutes (observation period), using a single-lead patch ECG electrode. Results The study was stopped prematurely after 55 of the expected 400 patients had been enrolled due to lower-than-expected flecainide peak plasma levels (197.8 ng/mL; 1.85-fold lower than those observed with the same target dose in INSTANT) and modest efficacy. The Data Monitoring Committee confirmed that there were no safety risks. Fifty-four of the enrolled patients (42 active; 12 placebo) were in AF at the start of inhalation. Baseline characteristics were well-balanced between treatment groups; overall mean age was 59.6 years (SD: 13.5) and 31.5% of patients were female. Cardioversion rates during the observation period in the efficacy population (n=51; 39 active, 12 placebo) were 30.8% (95% CI: 18.5-46.5%) in the active group and 0% (95% CI: 0.0-28.2%) in the placebo group (p=0.04; figure 1). In the active group, the median time to conversion from start of inhalation was 12.8 minutes (IQR: 7.8-25.0 minutes), which includes an 8-minute inhalation period. Among the 12 patients whose AF converted to SR, only 1 (8.3%) still had AF-related symptoms compared to 71.8% of patients whose AF did not convert to SR (p&amp;lt;0.001). Time to discharge-eligible status was 1.6h for patients whose AF cardioverted and 3.3h for those whose AF did not (p=0.002; figure 2). The most common adverse events (AEs; ≥8%) occurring more frequently in the active than the placebo group were the following: dyspnoea (14.3%), dysgeusia (14.3%) and cough (11.9%). No serious AEs or cardiovascular events of special interest (hypotension, ventricular tachycardia, bradycardia, sinus pause, atrial flutter 1:1, other arrhythmias) were reported. Conclusions Orally inhaled flecainide, at approximately half of the targeted dose, was modestly effective and safe. The 1.5-fold lower conversion rate observed in this study, compared to the same dose in the INSTANT trial, is consistent with the lower plasma levels of flecainide achieved. Future studies will be designed to improve drug delivery to yield peak plasma levels of flecainide associated with higher conversion rates.

A multilabel deep learning model for the detection of conduction and rhythm disorders from PDF outputs of a widely available portable ECG Device

Abstract Background Wearable and portable devices with ECG capabilities can provide broad access to screening for cardiovascular disorders. However, most are approved for the detection of only atrial fibrillation. Most algorithms for detecting rhythm and conduction disorders are trained and tested on 12-lead ECGs or the lead I of clinical ECGs, with an unclear role for extending the diagnostic range of wearable and portable devices. Moreover, no current algorithms enable cardiovascular diagnosis on data available to end users, who largely have access to the PDF outputs of their data. Purpose To validate a novel artificial intelligence (AI) model trained to identify conduction disorders and arrhythmias on synthetic ECG outputs from lead I data of clinical ECGs, to detect the same disorders on cardiologist-annotated PDF outputs from the Alivecor KardiaMobile single-lead ECGs (Kardia ECGs). Methods We extracted lead I data from 3,076,352 clinical ECGs across 5 hospitals of a large U.S.-based hospital system from 2000-2024 and simulated wearable PDF outputs from commercial devices using a novel plotting approach that replicated the variations in plotted formats. Gold-standard cardiologist written diagnosis statements were processed to generate 8 clinical labels spanning conduction and rhythm disorders, and we developed a multilabel model to predict these labels using an EfficientNet-B3 convolutional neural network. We examined the performance of the model on 24,808 Kardia ECGs annotated by cardiologists. Results The model achieved a high performance on clinical ECGs in a held-out subset of the synthetic lead I ECG printouts resembling wearable outputs (AUC &amp;gt; 0.9 for all 8 labels). In the independent 24,808 Kardia ECGs, annotated by experts, the model achieved high performance across all labels. For the rhythm disorders atrial fibrillation or atrial flutter, premature atrial contraction, and premature ventricular contraction, the model had AUROCs of 0.99, 0.91, and 0.95, respectively. For the conduction disorders 1st-degree AV block and 2nd or 3rd-degree AV blocks, the model had AUROCs of 0.94 and 0.99, respectively. Finally, for supraventricular tachycardias, wide QRS, and paced ECGs, the model had AUROCs of 0.998, 0.98, and 0.98, respectively. Conclusions We demonstrate that a deep learning model trained to identify rhythm and conduction disorders based on synthetic ECGs where lead I data from clinical 12 lead ECGs are used to replicate real-world PDF outputs of wearable ECGs, generalizes to real-world wearable and handheld single-lead devices.

External validation of a machine learning based classification algorithm for ambulatory heart rhythm diagnostics using smartphone-photoplethysmography - the SMARTBEATS algorithm study

Abstract Introduction Smartphone-photoplethysmography (PPG) can be used for heart rhythm monitoring. According to current guidelines, ECG verification is required for a new diagnosis of atrial fibrillation (AF). Accurate machine learning (ML) based automatic algorithms have the potential to facilitate clinical applications generating large numbers of PPG recordings, by offloading the need for manual over-read. Purpose The aim of this study was to evaluate the diagnostic performance of an automatic ML-based algorithm for heart rhythm diagnostics using smartphone-PPG recorded by patients with AF in an unsupervised ambulatory setting. Methods Patients undergoing direct current cardioversion (DCCV) at a university hospital for treatment of AF or atrial flutter (AFL) were asked to perform one-minute heart rhythm recordings post-cardioversion at least twice daily for 30 days. All participants were provided with an iPhone 7 smartphone running the CORAI Heart Monitor PPG application simultaneously with a single-lead ECG recording (KardiaMobile). The automatic algorithm uses support vector machines (SVM) to classify heart rhythm from smartphone-PPG recordings. The SVM classifier evaluated here was trained on ambulatory PPG recordings made by patients in a separate cardioversion cohort, ensuring complete independence between training and validation sets. PPG recordings in the validation cohort were analyzed by the automatic algorithm and diagnostic performance was calculated by comparing heart rhythm classification output from the SVM classifier to the heart rhythm diagnosis from the simultaneous ECG recordings. ECG recordings were interpreted independently by two cardiologists and were set as gold standard. Results The SVM classifier was trained on data from 153 patients with 11,749 PPG recordings, of which 5,873 (50.0%) were labelled as AF and 5,876 (50.0%) as sinus rhythm (SR), as assessed by manual reading. In the validation cohort, 280 patients, with a median age of 69.0 years (31% women), registered 18,005 simultaneous PPG and ECG recordings. Recordings interpreted as AFL on ECG (2.0%), as having insufficient quality on ECG (4.9%) or PPG (2.8%), and low certainty algorithmic classifications (1.2%) were excluded, leaving 16,057 PPG recordings. Included ECG recordings consisted of 71.2% interpreted as (SR) and 28.8% as AF. Algorithm classification of the PPG recordings in the validation cohort diagnosed AF (sensitivity) in 99.7% and SR (specificity) in 99.7% of the recordings, with an overall accuracy of 99.7%. F1-score was 99.4% and area under the ROC curve (AUC) was 0.999. Conclusion A machine learning based algorithm for automatic heart rhythm diagnostics showed excellent diagnostic performance for smartphone-PPG recordings in an unsupervised ambulatory setting, minimizing the need for ECG verification in cardioversion populations.

Prospective longitudinal evaluation of AI-ECG in newly diagnosed heart failure (PLANE-HF)

Abstract Background * In the United Kingdom, there are one million people suffering from heart failure (HF), with more than 60,000 new cases annually (1). * Monitoring for declining left ventricular ejection fraction (LVEF) could optimise treatment approaches and avert healthcare episodes. * Easy home-based tracking of cardiac function would be transformative. * Previous studies have validated the capability of artificial intelligence-enabled electrocardiogram (AI-ECG) to correlated with changes in LVEF (2). * We hypothesize that longitudinal changes in AI-ECG probability score is associated with changes in the left ventricular (LV) function. Method * A prospective multicentre longitudinal study recruiting newly diagnosed HF with reduced ejection fraction (HFrEF) cases to investigate application of AI-ECG using longitudinal single-lead ECG recordings via patient-administered smart ECG-stethoscope examination – Figure 1. * A total of 36 patients were included in the analysis, with invitation to return for repeat echocardiography at 6-8 weeks. * Adjusted logistic regression was performed to analyze changes in LVEF by 10% as the outcome variable, investigating the relationship with the trajectory of AI-ECG probability scores. This study was approved by the UK Health Research Authority (reference 22/LO/0701). Result * Out of 36 patients, 19 patients had ischaemic and 17 had a non-ischaemic HF. * The mean age was 56.3 years, 97.2% patients were male and 46.6% were White. * 328 longitudinal single-lead ECGs were captured and used in the analysis. * 12 (33.3%) patients had an increase of their LVEF by 10%; preceding AI-ECGs was predictive of this recovery (OR 1.36; 95% CI, 1.05 – 1.65, p value = 0.016 – Table 1) Clinical Importance: * Individualised predictions of deterioration and improvement * Proactive medication adjustment * Enhanced triage for echocardiography Conclusion * AI-ECG designed to identify left ventricular systolic dysfunction (LVSD) can predict changes in LVEF% among newly diagnosed HFrEF patients. * Such results can place AI-ECG as a remote monitoring tool to track improvement or deterioration in patients’ LVEF%, which facilitate a proactive management and prevention strategies.

Atrial Fibrillation (AF) detected by repeat handheld ECG screening and the risk of heart failure hospitalization

Abstract Purpose Heart failure and atrial fibrillation (AF) frequently complicate one another. Risk of heart failure in screen-detected patients was not investigated. We aimed to compare risk of hospitalization for heart failure (HHF) between screen-detected and clinically-diagnosed AF, and to evaluate the impact of AF burden. Methods Consecutive patients aged over 65 attending outpatient clinics were prospectively recruited for AF screening, using handheld single-lead ECG (AliveCor) during December 2014 and December 2017. Patients who had &amp;gt;1 visits received repeat screening. All participants were divided into 5 cohorts: (i) previously known AF at enrolment; (ii) initial screen-detected AF; (iii) subsequent screen-detected AF; (iv) clinically-diagnosed AF during follow-up (FU); (v) No AF. Risk of HHF was estimated by adjusted sub-distribution hazard ratios (aSHR) derived from Fine and Gray regression, accounting for death as competing risk, adjusting for CHA2DS2VASC components and chronic kidney disease, AF diagnosis being handled as time-dependent variable. Patients who participated in repeat screening and previously known AF formed a sub-cohort. Patients with previously known AF who were in AF in the initial screening and those with repeat screen-positivity were defined probably persistent AF, otherwise paroxysmal AF (single screen-positivity) or no AF (never screen-positive). Risk of HHF was compared between groups stratified by AF burden (persistent vs. paroxysmal) and scenario of AF diagnosis (screen-detected vs. clinically-diagnosed). Results Of 11,972 subjects (mean age 76.6±7.8, female 49.2%), 18.7% (n=2,236) had known AF, 1.9% (n=223) initial screen-detected AF, 0.6% (n=71) subsequent screen-detected AF, 7.3% (n=867) clinically-diagnosed AF during FU, and 71.6% (n=8,559) no AF. During a median FU of 6.7 (IQR:4.4-8.0) years, risk of HHF was highest in patients with clinically-diagnosed AF (aSHR=3.5(2.7-4.5)), followed by subsequent screen-detected AF (aSHR=2.8(1.6-5.0)), previously known AF (aSHR=1.9(1.6-2.2)), and initial screen-detected AF (aSHR=1.7(1.1-2.4)) (Figure 1a). Sensitivity analysis excluded those with known heart failure at baseline, after which risk of HHF became comparable between previously known and initial screen-detected AF (Figure 1b). Sub-analysis for participants (32.2% (n=3,853), mean age 76.7±7.3, female 47.5%) of repeat screening showed screen-detected persistent AF (aSHR=3.7(1.9-7.2)) was at highest risk of HHF, followed by clinically-diagnosed AF (AF burden unassessed) (aSHR=3.3(2.2-5.0)), previously known persistent AF (aSHR=3.1(2.5-3.8)), screen-detected paroxysmal AF (aSHR=2.1(1.4-3.4)), previously known paroxysmal AF (aSHR=1.8(1.3-2.4) (Figure 1c). Conclusion Higher AF burden and FU new AF diagnosed by either screening or clinic visit was associated with higher risk of HHF. Screening facilitating early AF diagnosis and early downstream intervention, warranted further study.

Screening for Atrial Fibrillation (AF) by handheld single-lead ECG and risk of all-cause mortality stratified by non-AF cardiovascular disease

Abstract Purpose Screening for Atrial fibrillation (AF) has been suggested as a strategy to increase early detection of silent AF. We aimed to evaluate all-cause mortality between patients with screen-detected and clinically-diagnosed AF, and the effect of oral anticoagulation therapy (OAC) on mortality. Methods We prospectively enrolled and performed AF screening using handheld single-lead ECG(AliveCor) on consecutive patients who attended medical outpatient clinics during December 2014 and December 2017. Repeat screening was performed among patients who had &amp;gt;1 clinic visits. All participants were divided into 5 cohorts: (i) previously known AF at baseline (kAF); (ii) initial screen-detected AF (S1-AF); (iii) subsequent screen-detected AF (FU-sAF); (iv) clinically-diagnosed AF during follow-up (FU-cAF); and (v) no AF. Risk of all-cause mortality was estimated by adjusted hazard ratio (HR) using COX proportional hazards regression, adjusting for age, gender, co-morbidities, concurrent chronic use of medications including OAC, with AF detection or clinical diagnosis and OAC therapy handled as time-varying variable. Subgroup analysis was conducted to identify heterogeneity of effect across with or without non-AF cardiovascular disease (CVD). Results Of 11,972 subjects (mean age 76.6±7.8, female 49.2%), 18.7% (n=2,236) had previously kAF, 1.9% (n=223) S1-AF, 0.6% (n=71) FU-sAF, 7.3% (n=867) FU-cAF, and 71.6% (n=8,559) no AF. During a median FU duration of 6.8(IQR: 5.2-8.1) years, patients with kAF (HR=1.8(1.6-2.0)), S1-AF (HR=1.6(1.3-1.9)) and FU-cAF (HR=1.6(1.4-1.8)) were at higher risk of all-cause mortality, compared to individuals without AF, whereas FU-sAF (HR=1.0(0.7-1.4)) had similar low risk of death as no AF. Heterogeneity of effect was detected in subgroup analysis. Among patients without non-AF CVD, FU-sAF (HR=6.0(1.8-19.8)) and FU-cAF (HR=4.7(3.3-6.7)) were at markedly higher risk of all-cause mortality. DOAC (HR=0.57(0.38-0.87)) but not VKA (HR=0.72(0.43-1.15) reduced mortality. However, in patients with CVD other than AF, risk of FU-sAF (HR=1.1(0.8-1.6)) decreased to as low as individuals without AF (P for interaction =0.0039).Both DOAC (HR=0.51(0.46-0.58)) and VKA (HR=0.87(0.78-0.98) were beneficial. Conclusion Likely delayed AF detection suggested by groups of FU-cAF and FU-sAF may be associated with significant rise in risk of mortality among patients without non-AF CVD. The effect diminished when non-AF CVD as potential competing risk existed. The role of VKA in reducing all-cause mortality warrants more studies.

Validation of diagnostic wearable ECG data from high-risk patients and integration into the atrial fibrillation catheter ablation pathway

Abstract Background Wearables like the Apple Watch (AW) offer non-invasive, long-term monitoring and patient-led, single-lead ECG recordings. Although validated for Atrial Fibrillation (AF) detection in the general population, the accuracy and clinical utility of the single-lead ECG recording feature for the diagnosis of AF has not been evaluated in patients at high risk of atrial arrhythmias.(1) An evaluation of the accuracy of the Wearable-derived labels is important to determine the feasibility of integrating AW-based remote monitoring into clinical pathways. Objective To evaluate the accuracy and data burden from AW-based, patient-led ECG monitoring in patients undergoing AF catheter ablation. Methods Patients referred for first-time AF catheter ablation to the Study Institution between January 2022 and March 2023 were invited to enrol in the AF Follow-up with Apple Watch (AFFU-AW) Trial. Patients randomised to the active arm of the study received 12 months of post-ablation follow-up using a loaned Apple Watch. Participants underwent an education session and were asked to perform daily ECGs; sending through any labelled as AF, inconclusive, or in the context of symptoms. The accuracy of the automated ECG labelling was determined against blinded, independent labelling by two experienced Electrophysiologists. Results 1093 remote transmissions were emailed in for evaluation from 52 patients to the Study team. This was a median of 6 [2, 17] per patient and a receiver burden of 2 [1, 4] ECGs per day. 475 were categorised by the AW as AF with a sensitivity of 0.71 and specificity of 0.96 compared to expert labelling. 342 (31%) were not categorised by the wearable, with an ‘inconclusive’ interpretation (161 [47%]) or ‘high heart rate’ (108 [32%]) as the most cited reasons. 178 (55%) of the uncategorised ECGs were subsequently overread as AF. Accompanying symptoms were reported with 324 (30%) ECGs, most commonly palpitations. Conclusion This is the first reported evaluation of the AF diagnostic feature based on AW ECG recordings as part of a remote patient rhythm surveillance pathway after catheter ablation. The high positive predictive value of ‘Sinus Rhythm’ and ‘AF’ labels and the high specificity of the ‘AF’ label are comparable to diagnostic algorithms in implantable loop recorders.(2) The source code to automate data curation and structured databasing of wearable ECGs has been made available for care providers to integrate the service into device monitoring workflows.(3)

Validation of chest-belt ECG evaluation using artificial intelligence

Abstract Background Simple, cheap wearables for long-term ECG monitoring with automatic analysis of all measurements using a bespoke artificial intelligence (AI) are scarce. Annotated single-lead ECG data from a chest-belt are missing. Methods We have developed a complex software environment to acquire raw ECG data from the Polar H10 chest-belt, and a bespoke, original AI to detect artifacts, QRS complexes and the presence of atrial fibrillation (AF) from chest-belt recordings. Original manually annotated ECG recordings from real patients from cardiology departments using the chest-belt and a public MIT-BIH database as reference were used for validation. Results QRS determination by AI (version Q4/23) in evaluating MIT-BIH database of ECG samples had a 97.67 % sensitivity and a 97.73 % positive predictive value. AF in MIT-BIH database was detected by AI with 90.31 % sensitivity and 92.09 % positive predictive value. To assess AI accuracy in ECG from a chest-belt, 150 consecutive measurements using Polar H10 chest-belt (215 hours of ECG recordings in total) from individual patients were included in the analysis. Patients were not limited in any physical activity during measurements. 7.15 % of the length of all recordings (15.4 hours) was assessed as noise by AI and excluded from the analysis. AF lasting longer than 30 seconds was present in 14 of 150 (9.3 %) manually evaluated recordings. QRS determination by AI in evaluating manually annotated chest-belt data had 99.76 % sensitivity and 99.74 % positive predictive value. AF in manually annotated chest-belt data was detected by AI with 92.86 % sensitivity and 100 % positive predictive value. Conclusion A chest-belt can be used to reliably detect QRS complexes and the presence of atrial fibrillation with high accuracy when using a bespoke artificial intelligence for the evaluation. Commonly available devices capable of long ECG measurements could be used for patient-empowered screening for important arrhythmias.Atrial fibrillation on chest-belt ECG

Development and Multinational Validation of an Ensemble Deep Learning Algorithm for Detecting and Predicting Structural Heart Disease Using Noisy Single-lead Electrocardiograms.

AI-enhanced 12-lead ECG can detect a range of structural heart diseases (SHDs) but has a limited role in community-based screening. We developed and externally validated a noise-resilient single-lead AI-ECG algorithm that can detect SHD and predict the risk of their development using wearable/portable devices. Using 266,740 ECGs from 99,205 patients with paired echocardiographic data at Yale New Haven Hospital, we developed ADAPT-HEART, a noise-resilient, deep-learning algorithm, to detect SHD using lead I ECG. SHD was defined as a composite of LVEF<40%, moderate or severe left-sided valvular disease, and severe LVH. ADAPT-HEART was validated in four community hospitals in the US, and the population-based cohort of ELSA-Brasil. We assessed the model's performance as a predictive biomarker among those without baseline SHD across hospital-based sites and the UK Biobank. The development population had a median age of 66 [IQR, 54-77] years and included 49,947 (50.3%) women, with 18,896 (19.0%) having any SHD. ADAPT-HEART had an AUROC of 0.879 (95% CI, 0.870-0.888) with good calibration for detecting SHD in the test set, and consistent performance in hospital-based external sites (AUROC: 0.852-0.891) and ELSA-Brasil (AUROC: 0.859). Among those without baseline SHD, high vs. low ADAPT-HEART probability conferred a 2.8- to 5.7-fold increase in the risk of future SHD across data sources (all P<0.05). We propose a novel model that detects and predicts a range of SHDs from noisy single-lead ECGs obtainable on portable/wearable devices, providing a scalable strategy for community-based screening and risk stratification for SHD.

Heart rate variability as a dynamic marker of surgeons' stress during vascular surgery.

A surgeon experiences elevated stress levels when operating. Acute stress is linked to cognitive overload, worsening surgical performance. Chronic stress poses a significant risk to a surgeon's health. Identifying intraoperative stress may allow for preventative strategies that reduce surgeons' stress and subsequently improve patient outcomes. The aim of this study was to assess the feasibility of using heart rate variability as a marker of stress during vascular surgery. A total of 11 senior surgeons were evaluated performing three different vascular surgery procedures. Heart rate variability metrics (low-frequency to high-frequency ratio and standard deviation of the normal-normal interval) were determined from single-lead ECG traces at predetermined procedural performance points. State-Trait Anxiety Inventory-6, a validated stress tool, was used to assess surgeon-reported stress. Subjective reports of procedural difficulty were also collected. One-way ANOVA compared heart rate variability at key performance points with baseline. Pearson's coefficient assessed correlation between heart rate variability and subjective stress. Data were collected for six carotid endarterectomies, six open abdominal aortic aneurysm repairs, and five lower limb bypasses. Heart rate variability metrics indicating markedly greater stress were observed at key performance points across all procedures. Peaks in stress were consistent across different surgeons performing the same procedure. A significant correlation was observed between heart rate variability metrics and subjective State-Trait Anxiety Inventory-6 stress reports (r = 0.768, P =<0.001). The most difficult procedural steps reported corresponded with heart rate variability metrics displaying the greatest stress. Heart rate variability may be a viable approach to assess intraoperative stress and cognitive load during vascular surgery and could be used to evaluate whether a theatre intervention (for example timeout) could reduce stress in areas of surgical difficulty.

RawECGNet: Deep Learning Generalization for Atrial Fibrillation Detection From the Raw ECG.

Deep learning models for detecting episodes of atrial fibrillation (AF) using rhythm information in long-term ambulatory ECG recordings have shown high performance. However, the rhythm-based approach does not take advantage of the morphological information conveyed by the different ECG waveforms, particularly the f-waves. As a result, the performance of such models may be inherently limited. To address this limitation, we have developed a deep learning model, named RawECGNet, to detect episodes of AF and atrial flutter (AFl) using the raw, single-lead ECG. We compare the generalization performance of RawECGNet on two external data sets that account for distribution shifts in geography, ethnicity, and lead position. RawECGNet is further benchmarked against a state-of-the-art deep learning model, named ArNet2, which utilizes rhythm information as input. Using RawECGNet, the results for the different leads in the external test sets in terms of the F1 score were 0.91-0.94 in RBDB and 0.93 in SHDB, compared to 0.89-0.91 in RBDB and 0.91 in SHDB for ArNet2. The results highlight RawECGNet as a high-performance, generalizable algorithm for detection of AF and AFl episodes, exploiting information on both rhythm and morphology.

Radiofrequency catheter ablation of persistent atrial fibrillation by pulmonary vein isolation with or without left atrial posterior wall isolation: long-term outcomes of the CAPLA trial.

Posterior wall isolation (PWI) is commonly incorporated into catheter ablation (CA) strategies for persistent atrial fibrillation (AF) in an attempt to improve outcomes. In the CAPLA randomized study, adjunctive PWI did not improve freedom from atrial arrhythmia at 12 months compared with pulmonary vein isolation (PVI) alone. Whether additional PWI reduces arrhythmia recurrence over the longer term remains unknown. In this multicenter, international, randomized study patients with persistent AF undergoing index CA using radiofrequency (RF) were randomized to PVI+PWI versus PVI alone. Patients underwent regular follow-up including rhythm monitoring for a minimum of 3 years post CA. AF burden at 3 years post-ablation was evaluated with either 28-day continuous ambulatory ECG monitoring, twice daily single-lead ECG or from cardiac implanted device. Evaluated endpoints included freedom from any documented atrial arrhythmia recurrence after a single procedure, AF burden, need for redo catheter ablation, rhythm at last clinical follow-up, healthcare utilisation metrics and AF-related quality of life. 333 of 338 (98.5%) patients (mean age 64.3±9.4 years, 23% female) completed 3-year follow-up, with 169 patients randomized to PVI+PWI and 164 patients to PVI alone. At a median of 3.62 years post-index ablation, freedom from recurrent atrial arrhythmia occurred in 59 patients (35.5%) randomized to PVI+PWI vs 68 patients (42.1%) randomized to PVI alone (HR 1.15, 95% CI 0.88-1.51, p=0.55). Median time to recurrent atrial arrhythmia was 0.53 years (IQR 0.34-1.01 years). Redo ablation was performed in 54 patients (32.0%) in the PVI+PWI group vs 49 patients (29.9%, p=0.68) in the PVI alone group. Pulmonary vein reconnection was present in 54.5% (mean number of reconnected PVs 2.2±0.9) and posterior wall reconnection in 75%. Median AF burden at 3 years was 0% in both groups (IQR 0-0.85% PVI+PWI vs 0-1.43% PVI alone, p=0.49). Sinus rhythm at final clinical follow-up was present in 85.1% with PVI+PWI vs 87.1% with PVI alone (p=0.60). Mean AF Effect On Quality-Of-Life (AFEQT) score at 3 years post-ablation was 88.0±14.8 with PVI+PWI vs 88.9±15.4 with PVI alone (p=0.63). In patients with persistent AF, the addition of PWI to PVI alone at index RF catheter ablation did not significantly improve freedom from atrial arrhythmia recurrence at long-term follow-up. Median AF burden remains low and AF quality of life high at 3 years with either ablation strategy.

Study on OSA screening and influencing factors in community-based elderly hypertensive patients based on single-lead wearable ECG devices.

Assessing whether single-lead ECG can be effectively and relatively inexpensively used in large-scale OSA screening, and identifying factors influencing moderate-to-severe OSA among elderly hypertensive patients without atypical symptoms in primary care. The study gathered data from 15 medical institutions in Ningxia between January and December 2022 using cloud platforms. The dataset included basic information and 72-h ECG monitoring for 2573 hypertensive patients over 65. OSA screening was conducted using the single-lead wearable ECG devices based on the ACAT algorithm. A multivariable logistic regression identified the main factors affecting OSA severity in these patients, and the AUC was used to assess the model's predictive accuracy. The study found an OSA detection rate of 87.10%, with 55.42% being moderate to severe cases. Key risk factors associated with developing moderate-to-severe OSA included cardiac irregularities like supraventricular extrasystole and atrioventricular block, male gender, lifestyle factors like alcohol consumption and smoking, and health indicators such as SDNN ≤ 100ms, abnormal LF/HF ratio, BMI, and age. The model's accuracy for predicting OSA, indicated by a ROAUC of 0.625, was moderate. Factors like gender, tea consumption, stroke history, and ventricular tachycardia were also independently linked to OSA severity. This study combines single-lead wearable ECG devices with the ACAT algorithm for OSA screening in Ningxia, China. Initial screening identified 87.10% of participants as having OSA, with 55.42% being moderate to severe cases. This suggests a convenient, low-cost, and repeatable ECG-based method for OSA screening, potentially improving early detection and management of OSA by identifying potential risk factors.

Identifying a Heterogeneous Effect of Atrial Fibrillation Screening in Older Adults: A Secondary Analysis of the VITAL-AF Trial.

One-time atrial fibrillation (AF) screening trials have produced mixed results; however, it is unclear if there is a subset for whom screening is effective. Identifying such a subgroup would support targeted screening. We conducted a secondary analysis of VITAL-AF, a randomized trial of one-time, single-lead ECG screening during primary care visits. We tested two approaches to identify a subgroup where screening is effective. First, we developed an effect-based model using a T-learner. Specifically, we separately predicted the likelihood of AF diagnosis under screening and usual care conditions; the difference in probabilities was the predicted screening effect. Second, we used a validated AF risk model to test for a heterogeneous screening effect. We used interaction testing to determine if observed AF diagnosis rates in the screening and usual care groups differed when stratified by decile of the predicted screening effect and predicted AF risk. Baseline characteristics were similar between the screening (n=15187) and usual care (n=15078) groups (mean age 74 years, 59% female). In the effect-based analysis, in the highest decile of predicted screening effectiveness (n=3026), AF diagnosis rates were higher in the screening group (6.50 vs. 3.06 per 100 person-years, rate difference 3.45, 95%CI 1.62 to 5.28). In this group, the mean age was 84 years and 68% were female. The risk-based analysis did not identify a subgroup where screening was more effective. Predicted screening effectiveness and predicted baseline AF risk were poorly correlated (Spearman coefficient 0.13). In a secondary analysis of the VITAL-AF trial, we identified a small subgroup where one-time screening was associated with increased AF diagnoses using an effect-based approach. In this study, predicted AF risk was a poor proxy for predicted screening effectiveness. These data caution against the assumption that high AF risk is necessarily correlated with high screening effectiveness.

Detection of atrial fibrillation using a nonlinear Lorenz Scattergram and deep learning in primary care

BackgroundAtrial fibrillation (AF) is highly correlated with heart failure, stroke and death. Screening increases AF detection and facilitates the early adoption of comprehensive intervention. Long-term wearable devices have become increasingly popular for AF screening in primary care. However, interpreting data obtained by long-term wearable ECG devices is a problem in primary care. To diagnose the disease quickly and accurately, we aimed to build AF episode detection model based on a nonlinear Lorenz scattergram (LS) and deep learning.MethodsThe MIT-BIH Normal Sinus Rhythm Database, MIT-BIH Arrhythmia Database and the Long-Term AF Database were extracted to construct the MIT-BIH Ambulatory Electrocardiograph (MIT-BIH AE) dataset. We converted the long-term ECG into a two-dimensional LSs. The LSs from MIT-BIH AE dataset was randomly divided into training and internal validation sets in a 9:1 ratio, which was used to develop and internally validated model. We built a MOBILE-SCREEN-AF (MS-AF) dataset from a single-lead wearable ECG device in primary care for external validation. Performance was quantified using a confusion matrix and standard classification metrics.ResultsDuring the evaluation of model performance based on the LS, the sensitivity, specificity and accuracy of the model in diagnosing AF were 0.992, 0.973, and 0.983 in the internal validation set respectively. In the external validation set, these metrics were 0.989, 0.956, and 0.967, respectively. Furthermore, when evaluating the model’s performance based on ECG records in the MS-AF dataset, the sensitivity, specificity and accuracy of model diagnosis paroxysmal AF were 1.000, 0.870 and 0.876 respectively, and 0.927, 1.000 and 0.973 for the persistent AF.ConclusionsThe model based on the nonlinear LS and deep learning has high accuracy, making it promising for AF screening in primary care. It has potential for generalization and practical application.

Wearable ECG Device and Machine Learning for Heart Monitoring.

With cardiovascular diseases (CVD) remaining a leading cause of mortality, wearable devices for monitoring cardiac activity have gained significant, renewed interest among the medical community. This paper introduces an innovative ECG monitoring system based on a single-lead ECG machine, enhanced using machine learning methods. The system only processes and analyzes ECG data, but it can also be used to predict potential heart disease at an early stage. The wearable device was built on the ADS1298 and a microcontroller STM32L151xD. A server module based on the architecture style of the REST API was designed to facilitate interaction with the web-based segment of the system. The module is responsible for receiving data in real time from the microcontroller and delivering this data to the web-based segment of the module. Algorithms for analyzing ECG signals have been developed, including band filter artifact removal, K-means clustering for signal segmentation, and PQRST analysis. Machine learning methods, such as isolation forests, have been employed for ECG anomaly detection. Moreover, a comparative analysis with various machine learning methods, including logistic regression, random forest, SVM, XGBoost, decision forest, and CNNs, was conducted to predict the incidence of cardiovascular diseases. Convoluted neural networks (CNN) showed an accuracy of 0.926, proving their high effectiveness for ECG data processing.