AMSA and Slope Are Predictive of Shock Success and Return of Spontaneous Circulation but Not Arrest Duration Following Ventricular Fibrillation Cardiac Arrest

Introduction Ventricular fibrillation (VF) remains the most common cause of witnessed cardiac arrest. Recent efforts to improve defibrillation efficacy have focused on optimising the timing of defibrillation during cardiopulmonary resuscitation (CPR). Frequency-domain analysis of VF has demonstrated promise in estimating VF duration, predicting shock success and prognosticating on short term survival. Objective To determine the accuracy of amplitude and frequency based measures of VF in predicting defibrillation success and return of spontaneous circulation (ROSC). Methods ECG data were analysed from Heartsine defibrillator recordings of patients suffering VF cardiac arrest between 2006 and 2011. Waveform data was recorded at 170Hz and resampled at 250Hz. A 4.1s window was used to analyse the VF waveform prior to each defibrillation attempt. Amplitude spectral area (AMSA) was calculated as the summed product of frequency and amplitude from 4Hz to 48Hz. Slope was calculated as the median of the absolute value of differences in signal voltage every 12ms within the window. Defibrillation success was defined as an organised rhythm lasting greater than 5s occurring within 10s of defibrillation. Results 41 patients (mean age 63), (76\% male), received 118 shocks for VF. AMSA and slope were strongly predictive of 1st and all shock success (table). An AMSA threshold of \>7.7mVHz predicted defibrillation success with a sensitivity and specificity of 90\% and 60\% respectively. For VF preceding the 1st shock AMSA and slope were also predictive of final ROSC (AUC: 0.762, 0.765). There was no correlation between any of the calculated parameters and time to 1st shock. Conclusion Within this heterogeneous cohort of VF patients AMSA and Slope were strongly predictive of defibrillation success and ROSC. The lack of correlation between VF parameters and arrest duration likely reflects the contribution of CPR and substrate in influencing VF waveform characteristics