Skip to content


  • Poster presentation
  • Open Access

Can we predict neurological prognoses with computed tomography just after resuscitation?

  • 1,
  • 1,
  • 2 and
  • 1
Critical Care201014 (Suppl 1) :P330

  • Published:


  • White Matter
  • Grey Matter
  • Cardiac Arrest
  • Unfavorable Outcome
  • Hounsfield Unit


Noncontrast computed tomography (CT) is often used to evaluate a primary catastrophic brain injury after cardiac arrest. On day 3 after resuscitation, cerebral damage presents as swelling on CT; however, there is no strong evidence associated with how prognosis is predicted on CT performed just after resuscitation. We assessed whether the ratio (in Hounsfield units (HU)) of grey matter (GM) to white matter (WM) on CT performed just after resuscitation can be used as a predictor of patient outcome.


We utilized the electronic medical record to examine admission to the ICU after resuscitation from out-of-hospital cardiac arrest. Brain CTs performed just after resuscitation were re-evaluated. Patients were excluded if they had head trauma, cerebral vessel disease and were 16 years of age or younger. Clinical outcome was assessed at the time of discharge using the Pittsburgh cerebral performance category (CPC). On noncontrast CTs, HU were measured as GM in the regions of bilateral caudate head, putamen, globus pallidaus and cortical layer of the precentral sulcus, and as WM in the regions of the periventricular WM. We defined good neurological prognosis as CPC 1 to 2 and an unfavorable prognosis as CPC 3 to 5. Each ratio of average GM to WM was calculated, and they were compared with a good outcome and an unfavorable outcome.


In 1,057 episodes of out-of-hospital cardiac arrest, 296 were resuscitated and admitted to our ICU; 64 were excluded, and the final study sample consisted of 232 patients. CPC 1 to 5 at hospital discharge was 16%, 6%, 6%, 12% and 60%. The mean time from arrival to perform brain CT was 66 (25 to 330) minutes. The HU of caudate head, putamen and globus pallidaus were much higher in good outcome than in unfavorable outcome (P < 0.001, t test). The HU of cortical layer of the precentral sulcus and periventricular WM were not significantly different. The caudate head/WM ratio and globus pallidaus/WM ratio were much higher in good outcome (P < 0.01, t test). Using ROC analysis, we determined that a caudate head/WM ratio >1.21 or a globus pallidaus/WM ratio >1.17 was 90% predictive of a good outcome in this study.


The caudate head/WM ratio and globus pallidaus/WM ratio by brain CT just after resuscitation are useful to predict the post-resuscitation neurological prognosis.

Authors’ Affiliations

St Luke's International Hospital, Tokyo, Japan
St Luke's Life Science Institute, Tokyo, Japan


  1. Wijdicks EFM, et al.: Neurology. 2006, 67: 203-210. 10.1212/01.wnl.0000227183.21314.cdView ArticleGoogle Scholar


© BioMed Central Ltd. 2010