Abstract 298: Predicting Pulse Status During Cardiopulmonary Resuscitation Using Dimensionality Reduction of Electrocardiogram Wavelet Transforms

Abstract

Listen

Translate article

Background: Current resuscitation protocols require pausing cardiopulmonary resuscitation (CPR) to check for a pulse. However, pausing CPR during a pulseless rhythm worsens patient outcome. We designed an ECG-based algorithm that predicts pulse status during uninterrupted CPR, and evaluated its performance both with and without CPR. Methods: We evaluated 383 patients who were treated for out-of-hospital cardiac arrest using defibrillators with real-time ECG, CPR, and audio recordings. We collected paired adjacent ECG segments with and without CPR during organized rhythms. Segments were collected during the 10-s CPR period just prior to pulse check, and 5-s segments without CPR during the pulse check. ECG segments with or without a pulse were identified by the audio annotation and recorded blood pressures. Patients were randomly divided into 60% (230/383) training and 40% (153/383) test groups. From training data, we developed an algorithm to predict clinical pulse status based on the wavelet transform of the bandpass-filtered ECG, applying principle component analysis (PCA). We then trained a linear discriminant model using 3 principle component modes as input features. Model performance was evaluated on test group segments with and without CPR using receiver operating curves overall and according to initial arrest rhythm. Results: There were 230 patients (540 pulse checks) in the training set and 153 patients (372 pulse checks) in the test set. In both of these sets, about 16% (37/230 and 25/153) of the patients presented with initial non-shockable rhythm . Overall 38% (351/912) of checks had a spontaneous pulse. The areas under the receiver operating characteristic curve (AUCs) for predicting pulse status with and without CPR on test data were 0.84 and 0.89, respectively. Conclusion: A novel ECG-based algorithm demonstrates potential to improve resuscitation by predicting presence of a spontaneous pulse without pausing CPR.