Accurate identification of atrial flutter (AFL) types is not available before the ablation procedure. In this study we propose a vectorcardiographic (VCG) approach to characterize distinct AFL types in a non-invasive way.
Methods: We hypothesised that there is a higher similarity among the VCG loops described by the same AFL type and that a slow conduction region, playing an important role for its identification, is consistent with the re-entry mechanisms. Surface atrial waves are isolated by the administration of adenosine during the electrophysiological (EP) study. After the averaging of subsequent atrial cycles, VCG is obtained from the inverse Dower’s transform and resampled. Archetypes are then created for different AFL groups, from the VCG loops from several patients. In this study we considered 4 different AFL groups: common and perimitral AFL types, both including clockwise and counterclockwise variants. The gold standard was the EP study and velocity within the VCG trajectory allowed for the localization of slow regions. Finally, a correlation index for the comparison of two VCGs was defined from the mean correlation point-by-point of the vector described by the VCG loops, compared then to each archetype. Results: The proposed index was predominantly influenced by slow velocity regions presenting a higher density of samples. Correlation was higher when comparing cases belonging to the same AFL group: 0.947 ± 0.021 for common and 0.859 ± 0.053 for perimitral, in contrast to 0.638 ± 0.110 when correlating VCG loops from different AFL groups. No relevant contrasts were highlighted between groups when evaluating other parameters. Conclusions: The hypothesised bearing of the slow conduction regions when characterising AFL was demonstrated and a new vantage point has been stablished in diagnosing through physiological signals analysis. This non-invasive approach arises as a promising technique to identify the reentrant cause of AFL in advance of EP procedure.