PREDICTIVE VALUE OF CLINICAL MEASURES FOR RSO2 DURING IN-HOSPITAL CARDIAC ARREST

Abstract

TOPIC: Critical Care TYPE: Original Investigations PURPOSE: Cerebral oximetry is a non-invasive method of monitoring regional cerebral oxygenation (rSO2). rSO2 has been proposed as an alternative to end-tidal CO2 (ETCO2) as a predictor of return of spontaneous circulation (ROSC) and end-organ perfusion. Previously, we demonstrated that higher rSO2 is associated with higher rates of ROSC and improved neurologic recovery in in-hospital cardiac arrest (IHCA) patients. However, there are currently no studies examining patient-specific, peri-cardiac arrest variables that correlate with higher rSO2. Identifying modifiable variables can help tailor patient-specific interventions in addition to ACLS to optimize rSO2 and improve survival and neurologic recovery from IHCA. The aim of this study was to identify which clinical variables, particularly modifiable ones, are predictive of a higher rSO2. METHODS: We performed an ancillary cohort analysis of data collected as part of an ongoing multicenter prospective study of adult IHCA in 15 hospitals across the US and UK. rSO2 values were collected in real-time using a portable near-infrared spectroscopy device (SedLine, Nonin). Demographics, vitals, and labs were collected via chart review. Among 323 adult IHCA patients, under a multivariable regression model, we investigated the predictive value of age, hemoglobin (hgb), partial pressure of oxygen (PaO2), epinephrine dose, and APACHE II score for mean rSO2. Using two-sample t-tests, we also investigated the difference in mean rSO2 in relation to sex, shockable versus non-shockable rhythm, and manual versus mechanical CPR. RESULTS: We demonstrated a small, statistically significant, negative correlation for hgb (coefficient = -0.17, p = 0.03). Due to the unexpected negative correlation, an individual linear regression was performed, demonstrating no statistically significant correlation (coefficient = -0.14, p = 0.07). We did not find any statistically significant correlation with the other variables. We also did not find a statistically significant difference in mean rSO2 in relation to sex, initial rhythm, and type of CPR. CONCLUSIONS: Our study demonstrates no statistically significant correlation between PaO2, hgb, epinephrine dose, APACHE II scores, and rSO2. The lack of correlation for hgb is particularly interesting, as hgb has been demonstrated to positively correlate with rSO2 in healthy individuals. However, there are limitations to this study. Intra-cardiac paO2 and hgb were generally high in our sample size (92 mmHg and 14.0 g/dL, respectively), which may explain the lack of statistical significance. In addition, rSO2 data from each individual patient across the duration of CPR is variable, and utilizing mean rSO2 may be an oversimplification. CLINICAL IMPLICATIONS: The current standard of care is to apply a single ACLS protocol to an inherently heterogeneous group of patients undergoing cardiac arrest. As a result, there is ongoing interest in identifying modifiable variables to optimize survival and neurologic recovery. Our study demonstrates that clinical measures we expect to correlate with ROSC and neurologic recovery may not actually have a very large impact on clinical outcome. However, given the limitations of this study, further studies need to be conducted to identify and tailor patient-specific approaches to enhance resuscitative care. DISCLOSURES: No relevant relationships by Hon Chau, source=Web Response No relevant relationships by Elise Huppert, source=Web Response No relevant relationships by Sam Parnia, source=Web Response No relevant relationships by Thaddeus Tarpey, source=Web Response No relevant relationships by Jing Xu, source=Web Response