BACKGROUND

Neuron-specific-enolase (NSE) is frequently used to predict the neurologic outcome in persistently unconscious patients after cardiopulmonary resuscitation (CPR). However, its predictive value is unclear in the setting of veno-arterial extracorporeal membrane oxygenation therapy (ECMO). Aim of this project is to evaluate the predictive value of NSE in ECMO patients.

METHODS

NSE was measured after 24, 48, and 72 h in post-CPR ECMO patients. Neurologic status was evaluated using the best Cerebral Performance Categories Score (CPC) during the hospital stay. Patients who deceased within the first 24 h and patients who were awake during the first 24 h were excluded. ROC curves were calculated to assess the discriminative ability of single NSE measurements. Trajectories of serial NSE values were investigated using latent class mixed models.

RESULTS

The derivation cohort consisted of 65 patients, 30-day all-cause mortality was 47.7% and a poor neurological outcome with a CPC score of 4-5 was seen 30.7%. NSE measurement after 48 h showed the best discrimination for poor neurological outcome (AUC of 0.87 in the ROC curve; cut-off value of 70 μg/L). Specificity was highest if using serial NSE measurements at all three time points. These results could be validated in an external cohort of 64 patients.

CONCLUSION

In post-CPR patients on ECMO, NSE can be used to assess the neurologic outcome. Importantly, specificity was highest if using serial NSE measurements. Further research using prospective datasets is needed to verify these findings.