Open Access Highly Accessed Open Badges Research article

Serum neuron-specific enolase as predictor of outcome in comatose cardiac-arrest survivors: a prospective cohort study

Cédric Daubin1*, Charlotte Quentin2, Stéphane Allouche34, Olivier Etard5, Cathy Gaillard6, Amélie Seguin1, Xavier Valette1, Jean-Jacques Parienti67, Fabrice Prevost1, Michel Ramakers1, Nicolas Terzi89, Pierre Charbonneau1 and Damien du Cheyron110

Author Affiliations

1 Department of Medical Intensive Care, CHU de Caen, Caen, F-14000, France

2 Department of Medical Intensive Care, Mémorial France-Etats-Unis Hospital, Saint-Lô, France

3 Department of Biochemistry, CHU de Caen, Caen, F-14000, France

4 UPRES EA 3919, CHU de Caen, Caen University, Caen, F-14000, France

5 Laboratory of Neurological Functional Exploratory, CHU de Caen, Caen, F-14000, France

6 Department of Biostatistics and Clinical Research, CHU de Caen, Caen, F-14000, France

7 INSERM, UMR-S 707, Paris, F-75012, France

8 INSERM, ERI27, Caen, F-14000 France; Univ Caen, Caen, F-14000 France; CHRU Caen, Department of Medical Intensive Care, Caen, F-14000, France

9 E.A. 4497, Université de Versailles-Saint Quentin en Yvelines, 92380 Garches, France

10 UPRES, EA 2128, Caen, F-14000, France

For all author emails, please log on.

BMC Cardiovascular Disorders 2011, 11:48  doi:10.1186/1471-2261-11-48

Published: 8 August 2011



The prediction of neurological outcome in comatose patients after cardiac arrest has major ethical and socioeconomic implications. The purpose of this study was to assess the capability of serum neuron-specific enolase (NSE), a biomarker of hypoxic brain damage, to predict death or vegetative state in comatose cardiac-arrest survivors.


We conducted a prospective observational cohort study in one university hospital and one general hospital Intensive Care Unit (ICU). All consecutive patients who suffered cardiac arrest and were subsequently admitted from June 2007 to February 2009 were considered for inclusion in the study. Patients who died or awoke within the first 48 hours of admission were excluded from the analysis. Patients were followed for 3 months or until death after cardiopulmonary resuscitation. The Cerebral Performance Categories scale (CPC) was used as the outcome measure; a CPC of 4-5 was regarded as a poor outcome, and a CPC of 1-3 a good outcome. Measurement of serum NSE was performed at 24 h and at 72 h after the time of cardiac arrest using an enzyme immunoassay. Clinicians were blinded to NSE results.


Ninety-seven patients were included. All patients were actively supported during the first days following cardiac arrest. Sixty-five patients (67%) underwent cooling after resuscitation. At 3 months 72 (74%) patients had a poor outcome (CPC 4-5) and 25 (26%) a good outcome (CPC 1-3). The median and Interquartile Range [IQR] levels of NSE at 24 h and at 72 h were significantly higher in patients with poor outcomes: NSE at 24 h: 59.4 ng/mL [37-106] versus 28.8 ng/mL [18-41] (p < 0.0001); and NSE at 72 h: 129.5 ng/mL [40-247] versus 15.7 ng/mL [12-19] (p < 0.0001). The Receiver Operator Characteristics (ROC) curve for poor outcome for the highest observed NSE value for each patient determined a cut-off value for NSE of 97 ng/mL to predict a poor neurological outcome with a specificity of 100% [95% CI = 87-100] and a sensitivity of 49% [95% CI = 37-60]. However, an approach based on a combination of SSEPs, NSE and clinical-EEG tests allowed to increase the number of patients (63/72 (88%)) identified as having a poor outcome and for whom intensive treatment could be regarded as futile.


NSE levels measured early in the course of patient care for those who remained comatose after cardiac arrest were significantly higher in patients with outcomes of death or vegetative state. In addition, we provide a cut-off value for NSE (> 97 ng/mL) with 100% positive predictive value of poor outcome. Nevertheless, for decisions concerning the continuation of treatment in this setting, we emphasize that an approach based on a combination of SSEPs, NSE and clinical EEG would be more accurate for identifying patients with a poor neurological outcome.