scholarly journals Deep Neural Networks for ECG-Based Pulse Detection during Out-of-Hospital Cardiac Arrest

Entropy ◽  
2019 ◽  
Vol 21 (3) ◽  
pp. 305 ◽  
Author(s):  
Andoni Elola ◽  
Elisabete Aramendi ◽  
Unai Irusta ◽  
Artzai Picón ◽  
Erik Alonso ◽  
...  
Keyword(s):  
Neural Network ◽  
Cardiac Arrest ◽  
State Of The Art ◽  
Early Recognition ◽  
Pulseless Electrical Activity ◽  
Spontaneous Circulation ◽  
Bayesian Optimization ◽  
Pulse Detection ◽  
Electrocardiogram Ecg ◽  
Hospital Cardiac Arrest

The automatic detection of pulse during out-of-hospital cardiac arrest (OHCA) is necessary for the early recognition of the arrest and the detection of return of spontaneous circulation (end of the arrest). The only signal available in every single defibrillator and valid for the detection of pulse is the electrocardiogram (ECG). In this study we propose two deep neural network (DNN) architectures to detect pulse using short ECG segments (5 s), i.e., to classify the rhythm into pulseless electrical activity (PEA) or pulse-generating rhythm (PR). A total of 3914 5-s ECG segments, 2372 PR and 1542 PEA, were extracted from 279 OHCA episodes. Data were partitioned patient-wise into training (80%) and test (20%) sets. The first DNN architecture was a fully convolutional neural network, and the second architecture added a recurrent layer to learn temporal dependencies. Both DNN architectures were tuned using Bayesian optimization, and the results for the test set were compared to state-of-the art PR/PEA discrimination algorithms based on machine learning and hand crafted features. The PR/PEA classifiers were evaluated in terms of sensitivity (Se) for PR, specificity (Sp) for PEA, and the balanced accuracy (BAC), the average of Se and Sp. The Se/Sp/BAC of the DNN architectures were 94.1%/92.9%/93.5% for the first one, and 95.5%/91.6%/93.5% for the second one. Both architectures improved the performance of state of the art methods by more than 1.5 points in BAC.

scholarly journals Do combined ultrasound and electrocardiogram-rhythm findings predict survival in emergency department cardiac arrest patients? The Second Sonography in Hypotension and Cardiac Arrest in the Emergency Department (SHoC-ED2) study

CJEM ◽  
2019 ◽  
Vol 21 (6) ◽  
pp. 739-743 ◽  
Author(s):  
Nicole Beckett ◽  
Paul Atkinson ◽  
Jacqueline Fraser ◽  
Ankona Banerjee ◽  
James French ◽  
...  
Keyword(s):  
Emergency Department ◽  
Cardiac Arrest ◽  
Hospital Discharge ◽  
Point Of Care ◽  
Cardiac Activity ◽  
Pulseless Electrical Activity ◽  
Spontaneous Circulation ◽  
Point Of Care Ultrasound ◽  
Electrocardiogram Ecg ◽  
Higher Sensitivity

ABSTRACTObjectivesPoint-of-care ultrasound (POCUS) is used increasingly during resuscitation. The aim of this study was to assess whether combining POCUS and electrocardiogram (ECG) rhythm findings better predicts outcomes during cardiopulmonary resuscitation in the emergency department (ED).MethodsWe completed a health records review on ED cardiac arrest patients who underwent POCUS. Primary outcome measurements included return of spontaneous circulation (ROSC), survival to hospital admission, and survival to hospital discharge.ResultsPOCUS was performed on 180 patients; 45 patients (25.0%; 19.2%–31.8%) demonstrated cardiac activity on initial ECG, and 21 (11.7%; 7.7%–17.2%) had cardiac activity on initial POCUS; 47 patients (26.1%; 20.2%–33.0%) achieved ROSC, 18 (10.0%; 6.3%–15.3%) survived to admission, and 3 (1.7%; 0.3%–5.0%) survived to hospital discharge. As a predictor of failure to achieve ROSC, ECG had a sensitivity of 82.7% (95% CI 75.2%–88.7%) and a specificity of 46.8% (32.1%–61.9%). Overall, POCUS had a higher sensitivity of 96.2% (91.4%–98.8%) but a similar specificity of 34.0% (20.9%–49.3%). In patients with ECG-asystole, POCUS had a sensitivity of 98.18% (93.59%–99.78%) and a specificity of 16.00% (4.54%–36.08%). In patients with pulseless electrical activity, POCUS had a sensitivity of 86.96% (66.41%–97.22%) and a specificity of 54.55% (32.21%–75.61%). Similar patterns were seen for survival to admission and discharge. Only 0.8% (0.0–4.7%) of patients with ECG-asystole and standstill on POCUS survived to hospital discharge.ConclusionThe absence of cardiac activity on POCUS, or on both ECG and POCUS together, better predicts negative outcomes in cardiac arrest than ECG alone. No test reliably predicted survival.

Resuscitation for Patients with Out-of-Hospital Cardiac Arrest: Singapore

2002 ◽  
Vol 17 (2) ◽  
pp. 96-101 ◽  
Author(s):  
Ghee Hian Lim ◽  
Eillyne Seow
Keyword(s):  
Cardiac Arrest ◽  
Life Support ◽  
Hospital Setting ◽  
Pulseless Electrical Activity ◽  
Spontaneous Circulation ◽  
Very High Frequency ◽  
Traumatic Cardiac Arrest ◽  
Bystander Cardiopulmonary Resuscitation ◽  
Hospital Cardiac Arrest ◽  
Very High

AbstractAim:To evaluate characteristics and outcome of out-of-hospital cardiac arrest (OHCA) patients presenting to the Emergency Department (ED), and to examine factors that could be used to determine to prolong or abort resuscitation for these patients.Method:All OHCA patients presenting to the ED were studied over a three-month period from November 2001 through January 2002. Patient with traumatic cardiac arrest were excluded. Data were collected from the ambulance case records, ED resuscitation charts, and the ED Very High Frequency (VHF) radio case-log sheet. Information collected included the patient's demographic characteristics, timings (time from call to ambulance arrival on scene, time from arrival at scene to departure from scene, time from scene to arrival in the ED) recorded in the pre-hospital setting, the outcome of the resuscitation, and the final outcome for patients who survived ED resuscitation.Results:Ninety-three non-traumatic patients with an OHCA were studied during the three-month period. Of the 93 patients, 15 (16.1%) survived ED resuscitation, and one survived to hospital discharge. There were no statistically significant differences for age, race, or gender with regards to the outcome of the resuscitation. The initial cardiac rhythms were asystole (65), pulseless electrical activity (21), and ventricular fibrillation (7). Fourteen (15%) received bystander cardiopulmonary resuscitation (CPR). All seven patients with return of spontaneous circulation (ROSC) on arrival in the ED survived ED resuscitation. The ambulance took an average of 11.80 ±3.36 minutes for the survivors and 11.8 ±4.22 minutes for the non-survivors from the time of call to get to these patients. The average of the scene times was 12.5 ±4.61 minutes for the survivors and 12.0 ±4.02 minutes for the non-survivors. Transport time from the scene to the ED took an average of 39.1 ±8.32 minutes for the survivors and 37.2 ±9.00 minutes for the non-survivors.Conclusion:The survival rate for patients with OHCA after ED resuscitation is similar to the results from other studies. There is a need to increase the awareness and delivery of basic life support by public education. Automatic External Defibrillators (AED) should be available widely to ensure that the chance of early defibrillation is increased. Prolonged resuscitation efforts appear to be futile for OHCA patients if the time from cardiac arrest until arrival in the ED is ≥30 minutes coupled with no ROSC, and if continuous asystole has been documented for >10 minutes.

Abstract 243: Out-of-hospital Cardiac Arrest in the Outpatient Clinical Environment

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_4) ◽  
Author(s):  
Oscar J Mitchell ◽  
Michael Wang ◽  
Stacie Neefe ◽  
Michael Lynch ◽  
William D Schweickert ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Stress Test ◽  
Pulseless Electrical Activity ◽  
Neurological Status ◽  
Spontaneous Circulation ◽  
University Of Pennsylvania ◽  
Clinical Settings ◽  
Clinical Environment ◽  
Resuscitation Training ◽  
Hospital Cardiac Arrest

Background: Out-of-hospital cardiac arrest (OHCA) remains a leading cause of death in the US, affecting over 400,000 annually. Although outcomes have improved, rates of return of spontaneous circulation (ROSC) and survival are lower from OHCA than from in-hospital cardiac arrest. Clinical emergencies, including OHCA, in outpatient clinical settings are often first attended by rapid response teams (RRT), but the causes and outcomes from OHCA in these environments has not been characterized. An understanding of this population is critical both for RRT quality improvement and resource allocation. Objectives: We aimed to describe OHCA that occurred in outpatient clinical environments. We hypothesized that OHCA in ambulatory settings would be uncommon and would be concentrated in a limited number of higher-risk outpatient areas. Methods: Retrospective analysis of all RRT calls for non-hospitalized adult outpatients occurring between 2012- 2020 at the Hospital of the University of Pennsylvania. Results: There were 7336 RRT calls;25 were for OHCA. Information was available for 24 of these. Mean age was 64 +/- 16.7 y, and 38% were female. Initial rhythm was pulseless electrical activity in 79%, ventricular fibrillation in 13%, asystole in 4%, and was not recorded in 4%. CPR was initiated in all cases prior to the arrival of the RRT. ROSC was obtained in 84%, extracorporeal membrane oxygenation (ECMO) was initiated in 8%, and 8% did not survive. ROSC was obtained within 5 min in 58% of cases. Of those that attained ROSC, 75% survived to discharge, 80% with good neurological status (CPC 1-2). Most events (54%) were judged to be iatrogenic. Of these, 54% were due to anaphylaxis, 15% were during cardiac stress testing, and 31% were peri-procedural. The most common causes of anaphylaxis were chemotherapeutics and radiographic contrast agents. Conclusions: OHCA events are uncommon in the outpatient clinical setting and are frequently iatrogenic. These events are associated with high rates of ROSC and neurologically intact survival. Iatrogenic OHCA occurred during a limited number of clinical settings, including outpatient procedural, , infusion, and stress test locations.. This finding suggests the need to focus resuscitation training in these specific environments.

scholarly journals Towards the Prediction of Rearrest during Out-of-Hospital Cardiac Arrest

Entropy ◽  
2020 ◽  
Vol 22 (7) ◽  
pp. 758
Author(s):  
Andoni Elola ◽  
Elisabete Aramendi ◽  
Enrique Rueda ◽  
Unai Irusta ◽  
Henry Wang ◽  
...  
Keyword(s):  
Machine Learning ◽  
Cardiac Arrest ◽  
Patient Outcomes ◽  
Learning Model ◽  
Spontaneous Circulation ◽  
Operating Characteristics ◽  
Machine Learning Model ◽  
Validation Procedure ◽  
Electrocardiogram Ecg ◽  
Hospital Cardiac Arrest

A secondary arrest is frequent in patients that recover spontaneous circulation after an out-of-hospital cardiac arrest (OHCA). Rearrest events are associated to worse patient outcomes, but little is known on the heart dynamics that lead to rearrest. The prediction of rearrest could help improve OHCA patient outcomes. The aim of this study was to develop a machine learning model to predict rearrest. A random forest classifier based on 21 heart rate variability (HRV) and electrocardiogram (ECG) features was designed. An analysis interval of 2 min after recovery of spontaneous circulation was used to compute the features. The model was trained and tested using a repeated cross-validation procedure, on a cohort of 162 OHCA patients (55 with rearrest). The median (interquartile range) sensitivity (rearrest) and specificity (no-rearrest) of the model were 67.3% (9.1%) and 67.3% (10.3%), respectively, with median areas under the receiver operating characteristics and the precision–recall curves of 0.69 and 0.53, respectively. This is the first machine learning model to predict rearrest, and would provide clinically valuable information to the clinician in an automated way.

scholarly journals A Machine Learning Model for the Prognosis of Pulseless Electrical Activity during Out-of-Hospital Cardiac Arrest

Entropy ◽  
2021 ◽  
Vol 23 (7) ◽  
pp. 847
Author(s):  
Jon Urteaga ◽  
Elisabete Aramendi ◽  
Andoni Elola ◽  
Unai Irusta ◽  
Ahamed Idris
Keyword(s):  
Machine Learning ◽  
Cardiac Arrest ◽  
Electrical Activity ◽  
Area Under The Curve ◽  
Pulseless Electrical Activity ◽  
Learning Model ◽  
Spontaneous Circulation ◽  
Machine Learning Model ◽  
Balance Accuracy ◽  
Hospital Cardiac Arrest

Pulseless electrical activity (PEA) is characterized by the disassociation of the mechanical and electrical activity of the heart and appears as the initial rhythm in 20–30% of out-of-hospital cardiac arrest (OHCA) cases. Predicting whether a patient in PEA will convert to return of spontaneous circulation (ROSC) is important because different therapeutic strategies are needed depending on the type of PEA. The aim of this study was to develop a machine learning model to differentiate PEA with unfavorable (unPEA) and favorable (faPEA) evolution to ROSC. An OHCA dataset of 1921 5s PEA signal segments from defibrillator files was used, 703 faPEA segments from 107 patients with ROSC and 1218 unPEA segments from 153 patients with no ROSC. The solution consisted of a signal-processing stage of the ECG and the thoracic impedance (TI) and the extraction of the TI circulation component (ICC), which is associated with ventricular wall movement. Then, a set of 17 features was obtained from the ECG and ICC signals, and a random forest classifier was used to differentiate faPEA from unPEA. All models were trained and tested using patientwise and stratified 10-fold cross-validation partitions. The best model showed a median (interquartile range) area under the curve (AUC) of 85.7(9.8)% and a balance accuracy of 78.8(9.8)%, improving the previously available solutions at more than four points in the AUC and three points in balanced accuracy. It was demonstrated that the evolution of PEA can be predicted using the ECG and TI signals, opening the possibility of targeted PEA treatment in OHCA.

Abstract 110: The Dynamic Clinical Course of Resuscitation in Children With Cardiac Arrest

Circulation ◽  
2018 ◽  
Vol 138 (Suppl_2) ◽  
Author(s):  
Trond Nordseth ◽  
Dana Niles ◽  
Trygve Eftestøl ◽  
Vinay Nadkarni ◽  
Robert Sutton ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Clinical Course ◽  
Pulseless Electrical Activity ◽  
Spontaneous Circulation ◽  
Clinical State ◽  
State Transitions ◽  
Temporary Increase ◽  
Older Children ◽  
Chest Compressions ◽  
Hospital Cardiac Arrest

Background: During cardiac arrest, a child may be in one of five clinical states (“rhythms”): 1) Bradycardia with poor perfusion; 2) Pulseless Electrical Activity (PEA); 3) Ventricular Fibrillation or Pulseless Ventricular Tachycardia (VF/VT); 4) Asystole; or 5) Spontaneous Circulation (ROSC). The aim of study was to investigate and quantify the dynamic characteristics of this process. Methods: We prospectively acquired data on rhythm and clinical states using recording defibrillators during active CPR. Recordings were analyzed as a multi-state statistical model, focusing on transitions between PEA (including bradycardia with poor perfusion), VF/VT, Asystole, and ROSC (defined as an organized electrical rhythm without chest compressions >= 1 minute). Instantaneous transition rates were obtained by smoothing the Nelson-Aalen estimator of cumulative intensities. Results: In 74 Cardiac Arrest events with evaluable data, median patient age was 15 years [range 1.75 to 22.9; IQR 11 to 17]. Fifty percent had a respiratory etiology and 51 % were female. PEA was the most frequent presenting cardiac arrest rhythm (38 %); followed by VF/VT (24 %), bradycardia (22 %), and asystole (16 %). Starting from time of defibrillator attachment (median 3 minutes into the event) as shown in the figure that shows 45 minutes of CPR, the prevalence of sustained ROSC reached an asymptotic value of 30 % at 20 minutes. We observed a temporary surge of PEA at about 12 minutes, resulting from a doubling (from 0.1 to 0.2 transitions/min) of the instantaneous transition rate of ROSC to PEA during this period. Conclusion: We provide a quantitative overview of the dynamic process of clinical state transitions during in-hospital cardiac arrest and resuscitation in older children and adolescents. A notable feature was a temporary increase in the prevalence of PEA at 12 minutes.

Abstract 10125: Continuous Heart Rate Dynamics Preceding In-Hospital Cardiac Arrest of Respiratory Etiology

Circulation ◽  
2021 ◽  
Vol 144 (Suppl_2) ◽  
Author(s):  
Rongzi Shan ◽  
Xiao Hu ◽  
Noel G Boyle ◽  
Duc H Do
Keyword(s):  
Heart Rate ◽  
Cardiac Arrest ◽  
Sinus Node ◽  
A Priori ◽  
Pulseless Electrical Activity ◽  
Respiratory Arrest ◽  
Abrupt Onset ◽  
Spontaneous Circulation ◽  
Sinus Arrest ◽  
Hospital Cardiac Arrest

Introduction: Respiratory failure is a common cause of pulseless electrical activity (PEA) arrest in hospitalized patients, but how pathophysiologic changes in these conditions affect heart rate (HR) pre-arrest is not well described. We describe HR dynamics prior to in-hospital cardiac arrest (IHCA) among PEA/asystole arrest patients with respiratory etiology. Methods: In this retrospective descriptive study, we evaluated 67 patients with ≥3 hours of continuous ECG data recorded immediately preceding PEA/asystole IHCA in a single institution from 2010-2014. We identified respiratory arrest cases (eg. pneumonia, aspiration, pulmonary embolism, acute respiratory distress syndrome) by chart review and evaluated ECG patterns up to 24 hours prior to arrest to identify patterns of HR increase, HR decrease, sinus arrest, and escape rhythms. Results: We identified 31/67 patients with respiratory etiology (age 59±17 years, 52% male, 83% return of spontaneous circulation, 41% survived to discharge); of these 23/31(74%) fit an a priori model of HR response (Figure). Twelve cases demonstrated clear onset of HR increase at a median of 44 (IQR 28-507) minutes prior to arrest, while the remaining 11 cases started the monitoring period in sinus tachycardia. The mean peak HR was 120±20 bpm. An abrupt onset of HR decrease occurred at a median of 3.4 (IQR 2.3-5.9) minutes prior to arrest. Sinus arrest occurred during the HR decrease phase in 18/23 cases; the first escape rhythm was atrial in 11 (61%), junctional in 2 (11%) and ventricular in 3 (17%) cases. Conclusion: The majority of IHCA due to respiratory etiology (74%) follow a typical model of HR increase due to physiologic compensation to hypoxia, followed by rapid HR decrease leading to PEA arrest, likely from the vagal effect of hypoxia and sinus node suppression from acidosis. Understanding HR trends can aid clinical management as well as development of artificial intelligence models for prediction of IHCA.

Abstract 13102: Outcomes of Primary and Secondary Pulseless Electrical Activity During In-Hospital Cardiac Arrest

Circulation ◽  
2021 ◽  
Vol 144 (Suppl_2) ◽  
Author(s):  
Eirik Unneland ◽  
Anders Norvik ◽  
Shaun McGovern ◽  
David Buckler ◽  
Unai Irusta ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Electrical Activity ◽  
Additive Model ◽  
Pulseless Electrical Activity ◽  
Spontaneous Circulation ◽  
Clinical State ◽  
Time To Event ◽  
Time To Event Data ◽  
Average Patient ◽  
Hospital Cardiac Arrest

Background: Pulseless Electrical Activity (PEA) is common during in-hospital cardiac arrest. We investigated the development of four types of PEA: PEA as presenting clinical state (primary) and PEA secondary to transient return of spontaneous circulation (ROSC), ventricular fibrillation/tachycardia (VF/VT), or asystole (ASY). Methods: We analyzed 660 episodes of cardiac arrest at one Norwegian and three U.S. hospitals. ECG, chest compressions and ventilations were recorded by defibrillators during CPR. Clinical states were annotated using a graphical application. We quantified the transition intensities from PEA to ROSC (i.e. the immediate probability of a transition), and the observed half-lives for the four types of PEA (i.e. how quickly PEA develops into another clinical state), using Aalen’s additive model for time-to-event data. Results: The transition intensities to ROSC from primary PEA (n=386) and secondary PEA after ASY (n=226) were about 0.08 per minute, peaking at 6 and 9 min, respectively (figure, left). Thus, an average patient in these types of PEA has about 8% chance to achieve ROSC in one minute. Much higher transition intensities to ROSC of about 0.20 per min were observed for secondary PEA after transient ROSC (n=209) or VF/VT (n=225), peaking at 10 and 5 min, respectively. Half-live times for the four types of PEA (figure, right) were 8.5 min, 6.8 min, 4.6 min and 1.6 min, for primary PEA, and secondary PEA after ASY, transient ROSC and VF/VT, respectively. Discussion: The observed clinical development of PEA in terms of intensity, peak intensity and half-lives during resuscitation differs substantially between the four types of PEA. The chance of obtaining ROSC is considerably lower in primary PEA or PEA after ASY, compared to PEA following transient ROSC or after VF/VT. This may increase understanding of the nature of PEA and the process leading to ROSC; and allow for simple prognostic assessments during a resuscitation attempt.

Sign in / Sign up

Export Citation Format

Share Document