Research Article
Thomas Duning, P. Kirchhof,
Abstract
Background: Atrial fibrillation (AF) – be it permanent or paroxysmal – is the most frequent and most effectively treatable cause of stroke. However, when paroxysmal, even electrocardiography for 24 hours (24h-ECG) misses AF in more than 50% of cases. We assessed whether extended Poincaré analysis of ECG R-R intervals (EPA) can help to identify electrocardiographic remodeling suggestive of paroxysmal AF (PAF). Methods: Twenty-nine patients with previously diagnosed PAF were re-assessed by 24h-ECG using conventional analysis and EPA based on a previously trained algorithm considering among other ratios of R-R interval duration, number of premature atrial complexes, approximate entropy, and standard deviation of Poincaré plots-axes. 24h-ECG from 21 healthy subjects without a history of AF served as negative, and 9 patients with permanent AF as positive controls. Results: PAF during 24h-ECG was detected in 4 out of 29 (14%) patients with a history of PAF by conventional analysis. EPA classified ECGs of these and 22 additional patients with a history of PAF, i.e. a total 90%, as suggestive of PAF. All patients with permanent AF were identified by both tools. EPA additionally classified the ECG in 4 out of 21 control subjects as suggestive of PAF. Conclusions: Extended Poincaré analysis in patients with a history of PAF is more sensitive to electrocardiographic abnormalities than is conventional 24h- ECG analysis. These findings warrant prospective studies of EPA in patients with a high likelihood of PAF, i.e. with stroke of undetermined origin.