Correlation of EEG-based brain resuscitation index and end-tidal carbon dioxide in porcine cardiac arrest model

Evaluation and monitoring perfusion of vital organs is important during resuscitation from cardiac arrest. We developed a non-invasive electroencephalogram (EEG) based brain resuscitation index (EBRI) as a physiologic indicator measuring organ perfusion during cardiopulmonary resuscitation (CPR) and evaluated the correlation of EBRI and end-tidal carbon dioxide (ETCO2). A randomized crossover experimental study using a porcine cardiac arrest model was designed. After 1 minute of untreated ventricular fibrillation, 10 periods of higher-quality CPR (compression depth 5 cm and compression rate 100/min) for 50 seconds and lower-quality CPR (compression depth 3 cm and compression rate 60/min) for 50 seconds were performed in alternation. EBRI was calculated from the single EEG channel with the lowest noise. Mixed-model analysis was conducted to compare the differences of hemodynamic parameters, ETCO2, and EBRI between higher-quality CPR periods and lower-quality CPR periods. Pearson’s correlation coefficient was calculated to assess correlation between EBRI and ETCO2. The experiment was performed on 5 female swine (44.6 ± 2.8 kg). Higher-quality CPR showed significantly higher delta EBRI (median [IQR] 0.1 [0.0–0.2]) than did lower-quality CPR (median [IQR] –0.1 [–0.2–0.0], p < 0.01). EBRI had a statistically moderate positive correlation with ETCO2 (r = 0.51). In this porcine cardiac arrest model, EBRI was successfully obtained during resuscitation and had a statistically moderate correlation with ETCO2.

Abstract 287: Bimodal Brain Monitoring Using Portable EEG and Cerebral Oximetry During Cardiopulmonary Resuscitation (CPR): A Pilot Study

Background: There is a need to monitor brain resuscitation during CPR. Furthermore, 10% of CA survivors report cognitive activity and 2-3% report consciousness during CPR. Methods: Using data from a prospective study (AWARE-II) with bimodal brain monitoring during CPR (EEG, [Massimo] and regional cerebral oxygenation [rSO2 normal 60-80%]) using cerebral oximetry [Nonin]), we sought to identify EEG rhythms, their progression over time, and underlying rSO2 thresholds. Inclusion criteria: In hospital CA, ≥ 18 years. Raw EEG data were captured as images (1/second) during CPR pauses, while rSO2 was captured continuously. Results: 38 patients were recruited; mean age 69±16 with duration of CPR (10-60 mins). We captured 362 EEG images. 91 (25%) were artifacts. Among the 271 interpretable images, there were 13 EEG rhythms placed into 6 groups based on their association with normal/near normal, seizures, coma, absence of cortical activity as follows: Group 1 (Non-epileptogenic): Theta, Delta, Alpha, Group 2 (Epileptogenic/Spike and Wave): Generalized Rhythmic Delta Activity (GRDA) + Spike (S), Delta + S, Theta + S, Burst Suppression with GRDA + S, Group 3 (Flat-line): Marked Voltage Attenuation (MVA), Group 4 (Coma State): Burst Suppression (w/ GRDA, Delta, and/or Theta), Group 5: GRDA, Group 6: Generalized Periodic Discharges (GPD). MVA/absence of activity was observed throughout CPR time and across all rSO2 ranges. Normal/near normal (non-epileptogenic) including alpha rhythms (which to our knowledge is described for the first time), and epileptogenic rhythms were observed with rSO2>30% Conclusions: While unclear whether non-epileptogenic rhythms are associated with transient periods of ROSC, however real-time bimodal brain monitoring provides insights regarding brain resuscitation and its dynamic interaction with patient factors. While ischemia may cause epileptogenic activity, there are periods of normal/near-normal cortical activity despite prolonged CPR >45-60 mins. A minimal threshold of brain oxygen delivery (rSO2>30%) may be required for cortical activity. These data raise questions regarding assumptions of irreversible brain damage with prolonged CPR, as well as the possibility of consciousness and cognitive activity during CPR.

Abstract 272: Feasibility of Non-Invasive Cerebral Blood Flow Monitoring During CPR

Introduction: Reliable non-invasive monitoring of cerebral blood flow (CBF) during cardiac arrest would greatly facilitate goal-directed brain resuscitation during CPR. The Ornim c-FLOW™ provides real-time, continuous, non-invasive, direct monitoring of CBF via ultrasound tagged near infrared spectroscopy using adhesive sensors applied to the forehead. Values range from 0-100 units with a reported baseline value of 55±7 units (mean±sd). C-FLOW™ values are refreshed every three seconds for each of two forehead probes. The feasibility of using c-FLOW™ to monitor CBF during cardiac arrest has not been previously reported. Methods: The c-FLOW™ was applied in the ED to adult patients undergoing CPR for cardiac arrest that occurred in the ED or outside the hospital. c-FLOW™ values were continuously recorded during CPR and for up to 6 hours post-ROSC. c-FLOW™ values were correlated with corresponding end-tidal CO 2 (PetCO 2 ) values during CPR. Changes in c-FLOW™ values after vasopressor therapy were also quantified. Results: c-FLOW™ values were continuously recorded on patients undergoing CPR during 10 cardiac arrests. Initial, minimum, maximum, and mean values during CPR were 30.7±12.7, 17.3±15.0, 51.3±15.6, and 31.3±12.6 units, respectively. Maximum values after ROSC and VA ECMO were 43.0±10.9 and 59.0±12.0 units, respectively, and mean values after ROSC and VA ECMO were 24.0±11.7 and 35.3±12.7 units, respectively. The minimum value recorded after cessation of resuscitation efforts was 1.7±3.7 units. There was no significant correlation between c-FLOW™ values and simultaneous PetCO 2 values during CPR (R 2 0.01, p>0.05). c-FLOW™ values increased 7.6±8.5 units after IV/IO epinephrine boluses during CPR, though increased less with each subsequent bolus. Conclusions: Application of the c-FLOW™, a continuous real-time monitor of CBF, during cardiac arrest is feasible in the ED setting. c-FLOW™ values suggest variable and dynamic CBF during CPR. c-FLOW™ values do not appear to correlate with PetCO 2 but appear to detect increases in CBF associated with vasopressor therapy during CPR. Future studies are needed to determine the value of continuous non-invasive CBF monitoring as part of a goal-directed strategy to optimize brain resuscitation during CPR.

Abstract 271: Discrimination Performance of a Non-Invasive and Single-Channel Electroencephalography-Based Brain Cell Viability for Comparing High vs. Low Quality Cardiopulmonary Resuscitation; A Porcine Cardiac Arrest Experimental Study

Background: Maintenance of brain cell viability is the most important goal of cardiopulmonary resuscitation (CPR) in cardiac arrest. We have developed an non-invasive and single-channel EEG-based Brain Resuscitation Index (EBRI) that can be used to measure the brain cell viability. The purpose of this study is to evaluate whether high-quality (HQ) and low-quality (LQ) CPR can be distinguished using the EBRI. Methods: A random case-crossover experimental animal study using porcine cardiac arrest model was performed. After 1 minute of untreated ventricular fibrillation, HQ-CPR (compression depth 5 cm and compression rate 100 per min) and LQ-CPR (compression depth 3 cm and compression rate 60 per min) were alternated every 50 seconds and repeated 10 cycles overall for five pigs. The computing formula of the EBRI was developed to predict the end-tidal CO2 (ETCO2) value using a single-channel EEG signals attached to forehead. The EEG signals were continuously measured and calculated simultaneously with the EBRI. Receiver operating characteristic of the area under the curve (ROC-AUC) was used to compare the discrimination power of HQ-CPR versus LQ-CPR. We compared the ROC-AUCs of two EBRI models reflecting time lags )o seconds to 16 seconds) between CPR change (HQ to LQ or LQ to HQ) and measurement of EBRI) comparing with ETCO2 itself. Results: Demographics of vital signs and hemodynamic parameters were significantly different between HQ- versus LQ-CPR. (Table 1) The ETCO2 and various models of EBRI with different time lags were concordantly displayed according to HQ- and LQ-CPR alteration in Figure 1. When comparing the AUC-ROC of EBRI models with various time lags for HQ-CPR versus LQ-CPR, the EBRI with 8 seconds time lag model showed the highest value; 0.86 (0,86-0,87). The AUC-ROC of ETCO2 itself showed 0.67 (0.66-0.68). Conclusion: A non-invasive and single-channel EEG-based brain resuscitation index than ETCO2 showed the higher performance for discriminating high-and low-quality CPR in a case-crossover animal study. Keywords: Cardiopulmonary Resuscitation, Quality, Electroencephalography, Cell Viability