Abstract 2079: Automated Rhythm Identification to Guide Defibrillation without Interrupting CPR
Introduction Current versions of automated external defibrillators (AEDs) mandate repetitive interruptions of chest compression for rhythm analyses prior to defibrillation. We sought to achieve rhythm analyses without comprise of the survival value of uninterrupted chest compression.
Methods Electrocardiograms were recorded with conventional right subclavicular and apical electrodes of AEDs during CPR on human victims. Techniques of wavelet-based transformation and shape-based morphology detection were utilized for rhythm identification. Morphological components differentiated between organized rhythms, ineligible for defibrillation, from disorganized rhythms that were eligible for defibrillation and especially ventricular fibrillation and apparent asystole. The morphological consistency of the rhythm was measured by computing the residual differences between autocorrelation and cross-correlation functions. The potential rhythm was then classified based on the residual value as an organized or disorganized rhythm.
Results The algorithm was validated on 31,246 out-of-hospital recordings, including 9,235 corrupted by compression and ventilation artifacts on 185 patients after sudden cardiac death. An organized rhythm was detected with a sensitivity of 90.5%,a specificity of 87.7% and a predictive value of 91%. A disorganized rhythm was detected with a sensitivity of 87.7 %, a specificity of 90.5 % and a predictivity of 87.2%.
Conclusion The advantages of uninterrupted chest compression are therefore likely to outweigh the minimally reduced predictive value of a shockable rhythm. The proposed algorithm therefore would allow AED’s to prompt defibrillation without repetitive interruptions of CPR for rhythm analyses. (US patent 7,039,457 B2 issued in favor of the non-profit Weil Institute, Rancho Mirage, CA)