Objective Prediction of the prognosis of comatose survivors after cardiopulmonary arrest (CPA), so-called post-resuscitation encephalopathy (PRE), relies on neurological examination findings. Early laboratory indicators of poor prognosis (vegetative state/death) are not sensitive enough.
Methods We analyzed the results of magnetic resonance (MR) imaging with fluid-attenuated inversion recovery (FLAIR) and diffusion-weighted imaging (DWI) in 22 consecutive patients with PRE. Clinical details such as arrest place and anoxia time along with neurological examination findings including items of Glasgow coma scale (GCS) and the Full Outline of UnResponsiveness (FOUR) score were determined. Receiver Operator Characteristics (ROC) curves were produced to determine prognostic yield of the parameters studied.
Results Prognosis was classified as 'poor' (Glasgow-Pittsburg Cerebral Performance —CPC-score 4 or 5) in 16 and 'better' (CPC score 1-3) in 6 patients. The lower limit of confidence interval (CI) of the area under the curve (AUC) of the ROC was higher than 0.5 for visual, motor and total scores of GCS and FOUR score. Presence of a lesion pattern of multilobar, or diffuse, cortical involvement, termed as "extensive cortical lesion pattern" in MR imaging was a very good predictor of poor prognosis with an AUC of ROC of 0,937. Sensitivity of GCS motor part score and MR was 87.5% (95% CI: 61.6%-92.6%). Motor part of the FOUR score has a slightly lower sensitivity (68.7% with 95% CI from 41.4% to 88.9%). Incorporating of MR to the motor scores (either GCS or FOUR score) improved sensitivity to 100 % (95% CI: 79.2%-100%). AUC of the ROC was 1.000 (95%CI: 0.844-1.000) for the combination of MR and GCS motor score.
Conclusion This study provides the preliminary evidence that MRI, when used in conjunction with a neurological examination, may have potential in terms of predicting outcome in patients with PRE.
2009 by The Japanese Society of Internal Medicine