Abstract 12142: Are Outcomes After Cardiac Arrest Influenced by the Quality of ICU Care?

Abstract

Objective: To determine the association of modifiable hemodynamic and biochemical factors with outcome after cardiac arrest. Hypothesis: Optimal control of modifiable post-arrest factors for secondary injury improves outcome after cardiac arrest. Methods: We identified 222 consecutive cardiac arrest patients who received targeted temperature management in the Maine Medical Center registry from 1/2013 to 8/2015. We abstracted demographics, intra-arrest factors, and modifiable post-arrest factors [e.g. the absence of hypotension (MAP<60mm Hg in the first 6 hours), normoglycemia (median glucose 80-180mg/dl), normoxemia (no PaO2>300 or <60mm Hg), and normocapnia (median PCO2 30-50mm Hg)] during the first 24 hours after return of spontaneous circulation. Primary outcome was cerebral performance category (CPC) at discharge. We compared pre-arrest, intra-arrest, and modifiable post-arrest factors in patients with good and poor outcomes using the Wilcoxon or Fischer-exact test, as appropriate. We also constructed logistic regression models to determine the independent effects of these factors on outcome (SAS 9.3, Cary, NC). Results: Thirty percent of patients had a CPC 1-2 at discharge. In univariate analysis, pre-arrest factors associated with poor outcome were history of COPD, cardiovascular disease, hypertension. A shockable rhythm was associated with good outcome. Of the studied modifiable post-arrest factors, normoglycemia was significantly associated with good outcome (p=0.002). In multivariate analysis that included pre-arrest and intra-arrest variables, optimal control of two or more post-arrest factors was associated with improved outcome (OR 7.2, p=0.0007, AUC for model 0.842). Conclusions: Optimal control of blood glucose, oxygen, carbon dioxide and blood pressure in the post-arrest period is associated with improved outcome after cardiac arrest. The combined endpoint of 2 or more post-arrest modifiable targets controlled was the strongest post-arrest predictor of good outcome in our model. While previous studies have reported these factors to be individually associated with poor outcome, further studies evaluating how the combination of these factors contribute to outcome is warranted.