Aims Despite marked progress in the management of atrial fibrillation (AF), detecting AF remains difficult and AF-related complications cause unacceptable morbidity and mortality even on optimal current therapy. Methods and results This document summarizes the key outcomes of the 8th AFNET/EHRA Consensus Conference of the Atrial Fibrillation NETwork (AFNET) and the European Heart Rhythm Association (EHRA). Eighty-three international experts met in Hamburg for 2 days in October 2021. Results of the interdisciplinary, hybrid discussions in breakout groups and the plenary based on recently published and unpublished observations are summarized in this consensus paper to support improved care for patients with AF by guiding prevention, individualized management, and research strategies. The main outcomes are (i) new evidence supports a simple, scalable, and pragmatic population-based AF screening pathway; (ii) rhythm management is evolving from therapy aimed at improving symptoms to an integrated domain in the prevention of AF-related outcomes, especially in patients with recently diagnosed AF; (iii) improved characterization of atrial cardiomyopathy may help to identify patients in need for therapy; (iv) standardized assessment of cognitive function in patients with AF could lead to improvement in patient outcomes; and (v) artificial intelligence (AI) can support all of the above aims, but requires advanced interdisciplinary knowledge and collaboration as well as a better medico-legal framework. Conclusions Implementation of new evidence-based approaches to AF screening and rhythm management can improve outcomes in patients with AF. Additional benefits are possible with further efforts to identify and target atrial cardiomyopathy and cognitive impairment, which can be facilitated by AI.

We present a study on the feasibility of the automatic detection of atrial fibrillation (AF) from cardiac vibration signals (ballistocardiograms/BCGs) recorded by unobtrusive bedmounted sensors. The proposed system is intended as a screening and monitoring tool in home-healthcare applications and not as a replacement for ECG-based methods used in clinical environments. Based on BCG data recorded in a study with 10 AF patients, we evaluate and rank seven popular machine learning algorithms (naive Bayes, linear and quadratic discriminant analysis, support vector machines, random forests as well as bagged and boosted trees) for their performance in separating 30 s long BCG epochs into one of three classes: sinus rhythm, atrial fibrillation, and artifact. For each algorithm, feature subsets of a set of statistical time-frequency-domain and time-domain features were selected based on the mutual information between features and class labels as well as first- and second-order interactions among features. The classifiers were evaluated on a set of 856 epochs by means of 10-fold cross-validation. The best algorithm (random forests) achieved a Matthews correlation coefficient, mean sensitivity, and mean specificity of 0.921, 0.938, and 0.982, respectively.

No abstract

Episodes of hospitalization for heart failure patients are frequent and are often accompanied by fluid accumulations. The change of the body impedance, measured by bioimpendace spectroscopy, is an indicator of the water content. The hypothesis was that it is possible to detect edema from the impedance data. First, a finite integration technique was applied to test the feasibility and allowed a theoretical analysis of current flows through the body. Based on the results of the simulations, a clinical study was designed and conducted. The segmental impedances of 25 patients suffering from heart failure were monitored over their recompensation process. The mean age of the patients was 73.8 and their mean body mass index was 28.6. From these raw data the model parameters from the Cole model were deduced by an automatic fitting algorithm. These model data were used to classify the edema status of the patient. The baseline values of the regression lines of the extra- and intracellular resistance from the transthoracic measurement and the baseline value of the regression line of the extracellular resistance from the foot-to-foot measurement were identified as important parameters for the detection of peripheral edema. The rate of change of the imaginary impedance at the characteristic frequency and the mean intracellular resistance from the foot-to-foot measurement were identified as important parameters for the detection of pulmonary edema. To classify the data, two decision trees were considered: One should detect pulmonary edema (n(pulmonary) = 13, n(none) = 12) and the other peripheral edema (n(peripheral) = 12, n(none) = 13). Peripheral edema could be detected with a sensitivity of 100% and a specificity of 90%. The detection of pulmonary edema showed a sensitivity of 92.31% and a specificity of 100%. The leave-one-out cross-validation-error for the peripheral edema detection was 12% and 8% for the detection of pulmonary edema. This enables the application of BIS as an early warning system for cardiac decompensation with the potential to optimize patient care.