Predicting Outcome With Diffusion-Weighted Imaging in Cardiac Arrest Patients Receiving Hypothermia Therapy: Multicenter Retrospective Cohort Study

Abstract

Objectives: Many comatose patients following cardiac arrest have ischemic brain injury. Diffusion-weighted imaging is a sensitive tool to identify hypoxic-ischemic brain injury. The accurate prediction of the prognosis for comatose cardiac arrest survivors has been challenging, and thus, a multimodal approach, combining diffusion-weighted image findings, could be feasible. The aim of this study was to assess regional brain injury on diffusion-weighted imaging and to test the potential association with its neurologic outcome in patients treated with target temperature management after out-of-hospital cardiac arrest. Design and Setting: A multicenter, registry-based, retrospective cohort study was conducted using cases from 24 hospitals across South Korea. Of the 930 adult ( 18 yr) nontraumatic out-of-hospital cardiac arrest patients treated with target temperature management between January 2007 and December 2012 at these hospitals, we included the patients who underwent brain diffusion-weighted imaging in the first week after cardiac arrest. The brain regions examined included the four cerebral lobes, basal ganglia-thalamus, brain stem, and cerebellum. Imaging results were compared between a good neurologic outcome, defined as a cerebral performance category score of 1 or 2, and a poor neurologic outcome (cerebral performance category score 3). Measurement and Main Results: Poor neurologic outcome occurred in 118 of the 172 patients analyzed (68.6%). Positive diffusion-weighted image findings, defined as any regional brain injury lesion in diffusion-weighted imaging, were present in 106 patients. Positive diffusion-weighted image findings had 93% sensitivity, 86% specificity, 76% positive predictive value, and 96% negative predictive value for a poor neurologic outcome. The poor outcome group had higher numbers of affected brain lesions than the good outcome group (3.8 1.9 vs 0.1 +/- 0.6; p < 0.01). By multivariate analysis, positive diffusion-weighted image findings (odds ratio, 58.2; 95% CI, 13.29-254.91) and lack of a shockable rhythm (odds ratio, 0.13; 95% CI, 0.03-0.57) were associated with a poor neurologic outcome. Conclusions: Diffusion-weighted imaging allows reliable prediction of poor neurologic outcome in comatose patients treated with target temperature management after out-of-hospital cardiac arrest. Further prospective validation study will be required to generalize this result.