Abstract 416: Four Score Remains Effective Prognostic Tool in Cardiac Arrest After Eliminating Respiratory Component

Abstract

Introduction: The FOUR (Full Outline of UnResponsiveness) score was developed as a more effective alternative to the Glasgow Coma Scale for prognostication of critically-ill neurology patients. Current research on the FOUR score in cardiac arrest is limited, but suggests that it may be useful. Given that whether or not a patient over-breathes the ventilator can often be confounded by the set respiratory rate, hypocarbia, and other factors, we evaluated the FOUR score with and without the respiratory component (FOUR-). Methods: We retrospectively studied 83 cardiac arrest patients treated at an urban hospital from 2011-2018 from the Multimodal outcome characterization in comatose cardiac arrest (MOCHA) registry. FOUR and FOUR- score within first 24 hours (day 1) and 72-96 hours (day 4) after arrest were evaluated for ability to predict in-hospital mortality and survival to discharge. Results: Day 1 FOUR score < 4 had 78% (67%-87%) sensitivity and 57% (29%-82%) specificity, while FOUR- score < 4 had 84% (73%-92%) sensitivity and 50% (23%-77%) specificity for predicting in-hospital mortality. Day 4 FOUR and FOUR- scores < 4 had higher specificities (both 94% [71%-100%]) but lower sensitivities (63% [45%-79%] and 69% [51%-83%], respectively) for mortality than day 1. With outcome changed to survival to discharge, day 1 FOUR score > 8 had 29% (8%-58%) sensitivity and 97% (90%-100%) specificity, while FOUR- > 8 had 21% (5%-51%) sensitivity and 100% (95%-100%) specificity. Day 4 FOUR and FOUR- scores > 8 had lower specificities (89% [73%-97%] and 91% [77%-98%], respectively), but higher sensitivities (53% [28%-77%]) and 47% [23%-72%], respectively) than day 1. There were no differences in mortality between FOUR and FOUR- < 4 on day 1 (p=0.89) or 4 (p=0.95), and no differences in survival between FOUR and FOUR- > 8 on day 1 (p=0.26) or 4 (p=0.85). Conclusions: Both the FOUR and FOUR- scores had high specificity for mortality and survival, which is important given that incorrectly predicting a bad outcome could lead to premature withdrawal of life support. The absence of a significant difference between the FOUR and FOUR- and the stronger prognostic ability of the FOUR- suggest that the respiratory component may not provide additional prognostic utility in cardiac arrest.