Abstract 104: The Utility of Cerebral Oximetry (rSO2%) During In-Hospital Cardiac Arrest as a Marker for the Prediction of Return of Spontaneous Circulation (ROSC)
Background: Real time recognition of the quality of brain oxygen delivery is critical during resuscitation. Although end tidal CO2 monitoring (ETCO2) has been proposed as a marker of organ perfusion and return of spontaneous circulation (ROSC), its utility is limited by ventilation:perfusion mismatch and its poor correlation with cerebral blood flow. Non-invasive monitoring of cerebral oxygen content during cardiac arrest, using cerebral oximetry, may be a more sensitive method of assessing cerebral oxygen delivery. The aim of this study was to investigate the utility of cerebral oximetry during CPR and the relationship between regional cerebral oxygen saturation (rSO2) with ROSC during in-hospital cardiac arrest.
Methods: In a prospective cohort study, cerebral oximetry was applied to 122 consecutive in-hospital patients with a cardiac arrest undergoing CPR. Continuous data were collected until either ROSC lasting >20 minutes was achieved or resuscitation efforts were terminated. Logistic regression and receiver operating curves (ROC) were used to evaluate the utility of mean rSO2 as a predictor of ROSC.
Results: Twenty two patients were excluded due to CPR lasting <5 minutes. Of the remaining 100 patients, 55% were male, with 42% achieving ROSC. There was a significant difference in mean rSO2% in patients who achieved ROSC, compared to those who did not (51.8 ± 10.9% vs. 38.0 ± 13.1%, p < 0.0001). The area under the ROC curve (AUC) 95% was 0.78 (0.7-0.9). The cutoff values for mean rSO2 corresponding to 100% sensitivity and 100% specificity were 33.97% and 61.54%, respectively. Using a threshold rSO2 value of 50% led to a sensitivity (CI) 0.6 (0.41-0.54), specificity (CI) 0.8 (0.67-0.89), positive predictive value (95CI) 0.67 (0.5-0.8), and negative predictive value (95CI) 0.72(0.6-0.8). There was a statistically significant difference in the duration of CPR (20.88 mins ± 16.0% vs. 28.3 mins ± 20.6%, p=0.01) and time to initial placement of oximetry sensor (5.4 ± 4.8 mins vs. 8.6 ± 8.6 mins, p= 0.03) between patients who achieved ROSC and those who did not.
Conclusions: Cerebral oximetry appears to have high sensitivity and specificity in predicting ROSC and may have a role as a real-time, non-invasive predictor of ROSC, as well as futility, during cardiac arrest.
- © 2013 by American Heart Association, Inc.