Detection of spontaneous pulse using the acceleration signals acquired from CPR feedback sensor in a porcine model of cardiac arrest

Abstract

BackgroundReliable detection of return of spontaneous circulation with minimal interruptions of chest compressions is part of high-quality cardiopulmonary resuscitation (CPR) and routinely done by checking pulsation of carotid arteries. However, manual palpation was time-consuming and unreliable even if performed by expert clinicians. Therefore, automated accurate pulse detection with minimal interruptions of chest compression is highly desirable during cardiac arrest especially in out-of-hospital settings.ObjectiveTo investigate whether the acceleration (ACC) signals acquired from accelerometer-based CPR feedback sensor can be used to distinguish perfusing rhythm (PR) from pulseless electrical activity (PEA) in a porcine model of cardiac arrest.MethodsCardiac arrest was induced in 49 male adult pigs. ECG, arterial blood pressure (ABP) and ACC waveforms were simultaneously recorded during CPR. 3-second segments containing compression-free signals during chest compression pauses were extracted and only those segments with organized rhythm were used for analysis. PR was defined as systolic arterial pressure >60 mmHg and pulse pressure >10 mmHg, while PEA was defined as an organized rhythm that does not meet the above criteria for PR. Peak correlation coefficient (CCp) of the cross-correlation function between pre-processed ECG and ACC, was used to discriminate PR and PEA.Results63 PR and 153 PEA were identified from the total of 1025 extracted segments. CCp was significantly higher for PR as compared to PEA (0.440±0.176 vs. 0.067±0.042, p<0.01) and highly correlated with ABP (r = 0.848, p<0.001). The area under the receiver operating characteristic curve, sensitivity, specificity and accuracy were 0.965, 93.6%, 97.5% and 96.7% for the ACC-based automatic spontaneous pulse detection.ConclusionsIn this animal model, the ACC signals acquired from an accelerometer-based CPR feedback sensor can be used to detect the presence of spontaneous pulse with high accuracy.

Frontal EEG Changes with the Recovery of Carotid Blood Flow in a Cardiac Arrest Swine Model

“…The frequency distribution of EEG was also affected. As the CBF recovered, the higher frequency components including alpha (8-13 Hz) and beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) increased, whereas the lower frequency components including delta (<4 Hz) and theta (4)(5)(6)(7)(8) decreased. These changes affected the functional dynamics associated with varying amplitudes and multi-frequency responses, including the level of complexity and the amount of energy, diversity, and randomness [33], which could be indicated by the increase of log energy entropy and Rényi entropy, as shown in Figure 4.…”

“…If the ECG rhythm was shockable, a biphasic defibrillation shock of 200 J was applied by the EMT to restart the heart. Monitoring was initiated when a palpable pulse with organized QRS complexes and systolic blood pressure over 60 mmHg appeared [21]. Sustained ROSC was confirmed if spontaneous circulation continued for 20 min [19].…”

Accelerometry-based classification of circulatory states during out-of-hospital cardiac arrest

“…Inspired by the long-known technique of palpating the apex beat on the chest wall, these accelerometers were recently investigated on their capability to detect motions of the heart propagated to the chest wall to provide complimentary information about the mechanical activity of the heart. A proof of concept of this idea was provided in porcine models [16] and with smartphone-seismocardiography in humans [17], where the latter one focused only on the discrimination of the signals by blinded observers, while [16] proposed classifiers only on a single feature. However, a machine-learning-based automatic solution for determining the circulatory state of the heart from accelerometer data does not exist so far.…”

A Machine Learning Framework for Pulse Detection During Out-of-Hospital Cardiac Arrest

“…Recent studies used the regional cerebral oxygen saturation measured through near-infrared spectroscopy to predict ROSC during in-and out-of-hospital cardiac arrest [20], [21]. Acceleration signals measured on the common carotid artery from in-hospital cardiac arrest patients [22] or acquired from accelerometer-based CPR feedback devices in a porcine model of cardiac arrest [23] have also been used to develop pulse detectors. An abrupt rise of end tidal CO 2 (EtCO 2 ) has recently been reported as a specific but non-sensitive predictor of ROSC in OHCA [24], [25].…”