Abstract 271: Predicting Defibrillation Success Without Interrupting Chest Compression in Cardiac Arrest Patients
Background: In performing CPR on sudden cardiac arrest patients, a balance between chest compressions (CC) and defibrillation is needed. Optimizing the timing of defibrillation shocks may improve outcomes compared to time-based protocols. There exist numerous predictive features computed from the ECG signal before shock delivery to distinguish between rhythms which convert to spontaneous circulation and those which do not. Unfortunately, since CC induce significant artifact on the ECG signal, all these features assume a pause in CC before delivering the shock, so that there is no CC artifact on the ECG signal to be analyzed. However, in the 2010 guidelines, the AHA challenges manufacturers to “seek innovative methods to decrease the amount of time chest compressions are interrupted for AED operation.” It would be beneficial if the predictive features could be calculated while CC are performed and CC artifact is present.
Objective: Examine the effects of CC artifact on shock success predictive features and how such effects can be minimized.
Methods: A total of 111 short waveform segments from cardiac arrest patients with VF rhythm and transition from CC to hands-off were used. Thirteen temporal and spectral predictive features were calculated on a 4-sec “noisy” interval containing CC artifact preceding the transition at the end of CC, and on the adjacent 4-sec “clean” interval with no CC artifact following the transition. We presumed the underlying VF remained constant during both intervals. Predictive features were then calculated on a “filtered” version of the noisy segments with CC artifact removed using a newly developed method.
Results: The Wilcoxon signed-rank paired test showed CC artifact had a significant impact on all predictive features. However, after removal of CC artifact using our filter, the test showed all but one feature gave similar values to those from the “clean” segments.
Conclusion: The results demonstrate that CC artifact has a significant impact on shock success predictive features, but that the effect can be minimized using our CC artifact filtering method. Large prospective studies are needed to show if a combination of our filtering method and predictive features can help minimize CC interruptions and improve resuscitation outcomes.
- Cardiopulmonary resuscitation
- Cardiac arrest
- Ventricular fibrillation
- Electric countershock
- © 2013 by American Heart Association, Inc.