Accuracy of Rhythm Classification Using a Data Log System in Implantable Cardioverter Defibrillators

Publisher Website
Google Scholar
Abstract
Because the presence or absence of symptoms alone may be insufficient to correctly diagnose the rhythm for which implantable Cardioverter defibrillator therapy is delivered, we hypothesized that the addition of data Jog information available in Telectronics ATP 4210 may improve the accuracy of rhythm classification. With this system the recorded ventricular electrogram cycle length is reported on a beat‐to‐beat basis immediately before, during, and after the tachyarrhythmia is detected. Using this information recorded from the data log in 32 separate tachyarrhythmia episodes in 20 patients, we compared the sensitivity, specificity, and predictive accuracy of rhythm classification on the basis of symptoms alone, data Jog alone, and data Jog combined with symptoms. While classification based on symptoms alone is highly specific (10/10 episodes), it is insensitive and has an overall predictive accuracy of 53%. By contrast, data log is sensitive (90%) and specific (91%) with better predictive accuracy (94%) than symptoms alone (P = 0.002). The addition of symptoms to information on beat‐to‐beat cycle length from data log resulted in a slight increase in predictive accuracy.

This publication has 14 references indexed in Scilit: