Explore the vast range of titles available.

Recent global developments underscore the prominent role big data have in modern medical science. But privacy issues constitute a prevalent problem for collecting and sharing data between researchers. However, synthetic data generated to represent real data carrying similar information and distribution may alleviate the privacy issue. In this study, we present generative adversarial networks (GANs) capable of generating realistic synthetic DeepFake 10-s 12-lead electrocardiograms (ECGs). We have developed and compared two methods, named WaveGAN* and Pulse2Pulse. We trained the GANs with 7,233 real normal ECGs to produce 121,977 DeepFake normal ECGs. By verifying the ECGs using a commercial ECG interpretation program (MUSE 12SL, GE Healthcare), we demonstrate that the Pulse2Pulse GAN was superior to the WaveGAN* to produce realistic ECGs. ECG intervals and amplitudes were similar between the DeepFake and real ECGs. Although these synthetic ECGs mimic the dataset used for creation, the ECGs are not linked to any individuals and may thus be used freely. The synthetic dataset will be available as open access for researchers at OSF.io and the DeepFake generator available at the Python Package Index (PyPI) for generating synthetic ECGs. In conclusion, we were able to generate realistic synthetic ECGs using generative adversarial neural networks on normal ECGs from two population studies, thereby addressing the relevant privacy issues in medical datasets.

Machine learning (ML) methods for the analysis of electrocardiography (ECG) data are gaining importance, substantially supported by the release of large public datasets. However, these current datasets miss important derived descriptors such as ECG features that have been devised in the past hundred years and still form the basis of most automatic ECG analysis algorithms and are critical for cardiologists’ decision processes. ECG features are available from sophisticated commercial software but are not accessible to the general public. To alleviate this issue, we add ECG features from two leading commercial algorithms and an open-source implementation supplemented by a set of automatic diagnostic statements from a commercial ECG analysis software in preprocessed format. This allows the comparison of ML models trained on clinically versus automatically generated label sets. We provide an extensive technical validation of features and diagnostic statements for ML applications. We believe this release crucially enhances the usability of the PTB-XL dataset as a reference dataset for ML methods in the context of ECG data.

Deep learning-based tools may annotate and interpret medical data more quickly, consistently, and accurately than medical doctors. However, as medical doctors are ultimately responsible for clinical decision-making, any deep learning-based prediction should be accompanied by an explanation that a human can understand. We present an approach called electrocardiogram gradient class activation map (ECGradCAM), which is used to generate attention maps and explain the reasoning behind deep learning-based decision-making in ECG analysis. Attention maps may be used in the clinic to aid diagnosis, discover new medical knowledge, and identify novel features and characteristics of medical tests. In this paper, we showcase how ECGradCAM attention maps can unmask how a novel deep learning model measures both amplitudes and intervals in 12-lead electrocardiograms, and we show an example of how attention maps may be used to develop novel ECG features.

Spinal cord injury (SCI) is characterized by autonomic cardiovascular dysfunction that may contribute to the three‐ to fourfold greater risk of heart disease and stroke compared to non‐injured individuals. While exercise training elicits beneficial changes in autonomic function and vascular structure in healthy individuals, it is unclear if similar adaptations occur in individuals with SCI. Adults with chronic SCI (>1 year post injury) were randomized to 12 weeks of supervised upper‐body rowing exercise (UBROW; 3×/week; n = 8), adhering to current exercise guidelines, or control (CON; n = 9). Autonomic cardiovascular control was assessed by heart rate variability (HRV; electrocardiography) and blood pressure responses to a sit‐up test (finger plethysmography). Brachial (peripheral) and carotid (central) artery diameter and wall thickness (near‐ and far‐wall carotid intima–media‐thickness) were measured using high‐resolution ultrasound. All measurements were conducted at baseline, after 6 and 12 weeks. There was no effect of UBROW on time and frequency domain HRV or blood pressure responses to the sit‐up test (group‐by‐time interactions: P ≥ 0.28; effect sizes: ηp2 ≤ 0.11). For UBROW, brachial artery diameter increased from 4.80 ± 0.72 mm at baseline to 5.08 ± 0.91 mm after 12 weeks (P < 0.05, ηp2 = 0.27). Carotid artery dimensions did not change, and there were no correlations between changes (baseline–12 weeks) in brachial artery diameter and changes in HRV outcomes (r ≤ 0.40, P ≥ 0.14). While upper‐body rowing exercise enlarged brachial artery diameter, carotid artery dimensions and autonomic cardiovascular control did not change, suggesting local vascular remodelling, but no systemic vascular adaptations, in response to a supervised 12‐week exercise intervention in spinal cord‐injured humans. What is the central question of this study? What are the effects of volitional upper‐body rowing exercise on autonomic cardiovascular control and vascular structure in individuals with chronic SCI? What is the main findings and its importance? This study demonstrates that 12 weeks of volitional upper‐body rowing exercise results in localized enlargement of the brachial artery diameter with no influence on carotid artery dimensions (diameter and wall thickness) or autonomic vascular control. These findings provide novel evidence of local (and not systemic) vascular adaptations in response to exercise training in individuals with SCI.

Syncope is a common presentation in the emergency department, yet risk stratification for adverse events remains a substantial clinical challenge. In this study, the use of artificial intelligence (AI) models for identification of patients at high risk of 1-year cardiovascular death was investigated. The study included 39,735 patients aged 18 years or older who presented to a Danish emergency department with a discharge diagnosis of syncope and had an electrocardiogram recorded on the same day. Using multiple electrocardiogram parameters four different models were developed - artificial neural network, logistic regression, random forest, and extreme gradient boosting. Additionally, the models’ ability to stratify patients into low- and high-risk groups of 1-year cardiovascular death was assessed. The mean area under the receiver operating characteristic curves was 0.85 (0.01), while the mean area under the precision-recall curve was 0.11 (0.01). Stratification into low- and high-risk groups showed hazard ratios ranging from 10.80 (95% CI, 8.55–13.63) to 17.97 (95% CI, 12.32–26.20) in univariate Cox Proportional Hazards regression analysis across the four models. These findings demonstrate that AI-based models can successfully distinguish high-risk from low-risk patients with syncope, allowing for targeted evaluation and efficient clinical decision making.

Elderly individuals occupy an increasing part of the general population. Conventional and speckle-tracking transthoracic echocardiography may help guide risk stratification in these individuals. The purpose of this study was to evaluate the potential utility of conventional and speckle-tracking echocardiography in the screening of cardiac abnormalities in the elderly population. Two cohorts of elderly individuals (sample size: 1441 and 944) were analyzed, who were part of a randomized controlled clinical trial (LOOP study) and of an observational study (Copenhagen City Heart Study), recruiting participants from the general population >70 years of age with cardiovascular risk factors (arterial hypertension, diabetes mellitus, heart failure, or prior stroke) and sinus rhythm. Participants underwent a comprehensive transthoracic echocardiographic examination, including myocardial speckle tracking. Cardiac abnormalities were defined according to the ASE/EACVI guidelines. Structural cardiac abnormalities such as left ventricular (LV) remodeling, mitral annular calcification (MAC), and aortic valve sclerosis (with or without stenosis) were highly prevalent in the LOOP study (40%, 39%, and 27%, respectively). Moreover, a high prevalence of functional cardiac alterations such as LV diastolic dysfunction (LVDD), abnormal LV longitudinal systolic strain (GLS), and abnormal left atrial (LA) reservoir strain was present in the LOOP study (27%, 18%, and 9%, respectively). Likewise, the rate of LVDD, abnormal GLS, and abnormal LA reservoir strain was comparable in the validation sample from the Copenhagen City Heart Study. In line with these findings, subjects with LV remodeling, MAC, and aortic valve changes had a higher prevalence of LVDD, abnormal GLS, and abnormal LA reservoir strain than those without structural cardiac alterations. The findings of this study highlight the potential clinical utility of conventional and speckle-tracking echocardiography in the screening of structural and functional cardiac abnormalities in the elderly population. Further studies are warranted to determine the prognostic relevance of these findings.

Aims The present study had two aims: (i) compare echocardiographic parameters in COVID‐19 patients with matched controls and (2) assess the prognostic value of measures of left (LV) and right ventricular (RV) function in relation to COVID‐19 related death. Methods and results In this prospective multicentre cohort study, 214 consecutive hospitalized COVID‐19 patients underwent an echocardiographic examination (by pre‐determined research protocol). All participants were successfully matched 1:1 with controls from the general population on age, sex, and hypertension. Mean age of the study sample was 69 years, and 55% were male participants. LV and RV systolic function was significantly reduced in COVID‐19 cases as assessed by global longitudinal strain (GLS) (16.4% ± 4.3 vs. 18.5% ± 3.0, P < 0.001), tricuspid annular plane systolic excursion (TAPSE) (2.0 ± 0.4 vs. 2.6 ± 0.5, P < 0.001), and RV strain (19.8 ± 5.9 vs. 24.2 ± 6.5, P = 0.004). All parameters remained significantly reduced after adjusting for important cardiac risk factors. During follow‐up (median: 40 days), 25 COVID‐19 cases died. In multivariable Cox regression reduced TAPSE [hazard ratio (HR) = 1.18, 95% confidence interval (CI) [1.07–1.31], P = 0.002, per 1 mm decrease], RV strain (HR = 1.64, 95%CI[1.02;2.66], P = 0.043, per 1% decrease) and GLS (HR = 1.20, 95%CI[1.07–1.35], P = 0.002, per 1% decrease) were significantly associated with COVID‐19‐related death. TAPSE and GLS remained significantly associated with the outcome after restricting the analysis to patients without prevalent heart disease. Conclusions RV and LV function are significantly impaired in hospitalized COVID‐19 patients compared with matched controls. Furthermore, reduced TAPSE and GLS are independently associated with COVID‐19‐related death.

BackgroundPatients with severe mental illness (SMI) have a reduced life expectancy of one to two decades as compared to the general population, with most years of life lost due to somatic diseases. Most previous studies on disorders constituting SMI, e.g. schizophrenia and bipolar disorder, have investigated the disorders separately and hence not compared the disorders in terms of mortality rates relative to the background population.MethodsA register-based cohort study including the entire Danish population comparing mortality rates relative to the background population, controlling for age and sex, i.e. standardized mortality ratios (SMRs) in patients diagnosed with schizophrenia with those in patients diagnosed with bipolar disorder, during the study period from 1995 to 2014.ResultsThe SMR of patients with SMI was significantly higher than one for each calendar year in the study period with an overall SMR of 4.58, 95% CI (4.48–4.69) in patients diagnosed with schizophrenia (n = 38,500) and of 2.57 (95% CI 2.49–2.65) in patients diagnosed with bipolar disorder (n = 23,092). When investigating time trends in SMR for schizophrenia and for bipolar disorder, respectively, an increase in SMR over time was shown with a mean increase of 0.03 per year for schizophrenia and 0.02 for bipolar disorder (p < 0.01 for both disorders). The ratio between SMR for schizophrenia and SMR for bipolar disorder for each calendar year over the study period was constant (p = 0.756).ConclusionsIncreasing SMRs over the last 20 years were found for both patients diagnosed with bipolar disorder and patients diagnosed with schizophrenia. Despite clear differences between the two disorders regarding SMRs, the increases in SMR over time were similar, which could suggest similar underlying factors influencing mortality rates in both disorders.

Persons living with HIV (PLWH) may have increased incidence of cardiovascular events and longer QTc intervals than uninfected persons. We aimed to investigate the incidence and risk factors of de novo major electrocardiogram (ECG) abnormalities and QTc prolongation in well-treated PLWH. We included virologically suppressed PLWH without major ECG abnormalities, who attended the 2-year follow-up in the Copenhagen comorbidity in HIV infection (COCOMO) study. ECGs were categorized according to Minnesota Code Manual. We defined de novo major ECG abnormalities as new major Minnesota Code Manual abnormalities. Prolonged QTc was defined as QTc > 460 ms in females and QTc > 450 ms in males. Of 667 PLWH without major ECG abnormalities at baseline, 34 (5%) developed de novo major ECG abnormalities after a median of 2.3 years. After adjustment, age (RR: 1.57 [1.08–2.28] per decade older), being underweight (RR: 5.79 [1.70–19.71]), current smoking (RR: 2.34 [1.06–5.16]), diabetes (RR: 3.89 [1.72–8.80]) and protease inhibitor use (RR: 2.45 [1.27–4.74) were associated with higher risk of getting de novo major ECG abnormalities. Of PLWH without prolonged QTc at baseline, only 11 (1.6%) participants developed de novo prolonged QTc. Five percent of well-treated PLWH acquired de novo major ECG abnormalities and protease inhibitor use was associated with more than twice the risk of de novo major ECG abnormalities. De novo prolonged QTc was rare and did not seem to constitute a problem in well-treated PLWH.

Heart rate (HR) at rest is a well-known marker of cardiovascular morbidity and mortality. Results on the association between HR and incident atrial fibrillation (AF) have, however, been conflicting. Using digital electrocardiograms from 281,451 primary care patients, we aimed to describe the association between HR at rest and the hazards of incident AF. Secondary end points were death from all causes and pacemaker implantation. Data on drug use, co-morbidity, and outcomes were collected from nationwide administrative health care registries. During a median follow-up time of 8.4 years, 15,666 subjects were observed to develop AF, of which 1,631 were lone AF. A HR at rest from 30 to 51 beats/min was associated with an adjusted hazard ratio of 1.16 (95% CI 1.06 to 1.27) for AF compared with the reference group (66 to 72 beats/min). From 72 beats/min and upward, the hazard ratio of AF increased in a dose-response manner, reaching an adjusted hazard ratio of 1.36 (95% CI 1.26 to 1.46) for HR between 95 and 120 beats/min. Both for low and high HR, the associations were accentuated for the outcome lone AF (adjusted hazard ratios of 1.48, 95% CI 1.19 to 1.84 and 1.84, 95% CI 1.47 to 2.30 for HR between 30 to 51 and 95 to 120 beats/min, respectively). For death from all causes, the hazard increased almost linearly with increasing HR. A HR at rest from 30 to 51 beats/min was associated with an adjusted hazard ratio of 1.80 (95% CI 1.46 to 2.21) for pacemaker implantation. In conclusion, a U-shaped association was found between HR at rest and incident AF, and this association was strongest for the outcome lone AF.