Neuron Specific Enolase as a Biomarker Predicting Neurological Outcome after Cardiac Arrest in Patients Treated by Therapeutic Hypothermia

Jeung, Yu Jin; Lee, Byung Kook; Lee, Hyoung Youn; Jeung, Kyung Woon; Ryu, Hyun Ho; Chun, Byeong Jo; Moon, Jeong Mi; Heo, Tag; Min, Yong Il

Abstract

PURPOSE
Neurological outcome prediction is an important aspect of post-resuscitation care in cardiac arrest survivors. The appearance of high serum neuron specific enolase (NSE) is known to be associated with ischemic brain injury and poor neurological outcome. The application of therapeutic hypothermia to cardiac arrest survivors has been shown to improve neurological outcomes. As such, we investigated the predictive value of serial serum NSE levels in patients who were resuscitated from cardiac arrest.
METHODS
This study included 123 cardiac arrest survivors who were treated by therapeutic hypothermia from January 2008 to June 2011. Blood samples used for evaluating NSE were collected at return of spontaneous circulation (ROSC) at 6, 24 and 48 hours after initiation of therapeutic hypothermia. Neurological outcome was graded as 'good' or 'poor' at discharge and assessed according to the Cerebral Performance Category scale (CPC). A poor outcome was defined as a CPC value of 3-5.
RESULTS
Receiver operating characteristic (ROC) analysis revealed NSE cut-off values of 53.9 microg/L (sensitivity 14.6%), 48.5 microg/L (sensitivity 30.6%), 80.0 microg/L (sensitivity 40.0%), and 52.7 microg/L (sensitivity 55.5%) for poor outcomes with a specificity of 100%, measured at ROSC of 6, 24 and 48 hours after initiation of therapeutic hypothermia, respectively. The poor outcome group showed significant change in NSE concentration over time (p=0.002), while the good outcome group did not.
CONCLUSION
Detection of NSE at the cut-off value, 48 hr after initiation of therapeutic hypothermia was a specific but moderately sensitive marker of poor outcome at discharge. Single measurements of NSE should be cautiously interpreted, but NSE change over time was helpful in predicting the neurologic outcome.