Abstract P76: Predictive Characteristics of Six ECG Waveform Measures in Out of Hospital Ventricular Fibrillation Arrest
Background: Although individual quantitative measures of the VF waveform can predict shock outcome, little is known about their comparative performance. We evaluated the association between 6 different VF measures and the outcomes of return of spontaneous circulation (ROSC) and survival to hospital discharge.
Methods: We used the electronic defibrillator recording from 298 out of hospital VF arrests from a metropolitan EMS system to derive 6 waveform measures: logarithm of absolute correlations (LAC), median slope (MS), amplitude spectrum area (AMSA), angular velocity (AV), frequency Ratio (FR), and cardioversion output predictor (COP). Five second ECG epochs just prior to each shock were used to calculate the measures. The defibrillator recording, EMS report, and hospital chart were used to assess ROSC and survival after each shock. We evaluated the predictive relationship between each VF measure and outcome using the t-test and area under the curve (AUC) of receiver operating curve. We evaluated positive predictive value (PPV) after selecting cutoffs to achieve a negative predictive value (NPV) of 0.90
Results: A total of 683 shocks were analyzed. Each measure predicted ROSC and survival (p<0.001 for each association) (Table 1⇓). Using a NPV of 0.9, the PPV for ROSC varied moderately from 0.24 to 0.34 across the VF measures (Table 2⇓).
Conclusion: Although each measure predicted ROSC and survival, prognostic characteristics were not identical. Differences in prognostic characteristics may enable an approach that combines measures to optimize outcome prediction. Such a hypothesis requires investigation.