Temporal Electrogram Analysis: Algorithm Development

The automatic discrimination of physiological from pathological tachycardias by rate criteria alone lacks adequate specificity. Tachycardia detection algorithms based upon morphological analysis of the endocardial electrogram have been attributed high specificity although their specificity has not been proven. A previous study had shown temporal electrogram analysis (TEA) to be an algorithm of high sensitivity in the detection of ventricular arrhythmias despite low computational demands. In this study, the specificity and potential for automatic implementation have been assessed. Manual adjustment of thresholds for individual patients gave a maximum potential sensitivity of 97% (26/27 arrhythmias correctly recognized as non-sinus). The use of automatic setting of thresholds reduced sensitivity to 81%. The specificity of the algorithm, as assessed by exercise testing, was only 60%.