Neuron-specific enolase measured in serum compared to plasma for neuroprognostication in out-of-hospital cardiac arrest

Neuron-specific enolase (NSE) is an acknowledged biomarker for prognosticating neurological outcome after cardiac arrest, with elevated concentrations associated with poor outcome. This study assesses and compares the prognostic performance of NSE measured in serum and plasma for 1-year all-cause mortality among patients resuscitated from out-of-hospital cardiac arrest (OHCA). This investigation is based on post hoc analyses of the Blood Pressure and Oxygenation Targets After Out-of-Hospital Cardiac Arrest (BOX) trial, performed in patients remaining comatose after resuscitation. NSE was measured 48 h after admission using three distinct methods; 1) Serum-NSE measured in fresh serum samples, 2) frozen-plasma-NSE, measured in freeze–thaw EDTA–plasma from stored biobank samples, and 3) in a subset of the samples also as frozen-serum-NSE, measured in freeze–thaw serum from stored biobank samples. A total of 381 comatose OHCA patients were included, with an overall one-year mortality of 33.1%. Serum-NSE concentrations were significantly higher than frozen-plasma-NSE, with median concentrations of 21.2 µg/L (IQR: 15.7–45.5) versus 19.1 µg/L (IQR: 11.2–39.6), p < 0.001, respectively. Notably, the difference between serum-NSE and frozen-plasma-NSE increased with higher NSE concentrations. The mean difference was 65.8 µg/L with 95% limits of agreement of +/- 125.75 µg/L among patients with NSE > 60 µg/L. For predicting 1-year all-cause mortality, the AUROC for serum-NSE was 0.93 and 0.83 for frozen-plasma-NSE with a significant difference in AUROC of 0.10 (CI: 0.06; 0.14), p < 0.001. In a sub-group analysis (n = 67), there was no significant difference when comparing AUROC between serum-NSE and frozen-serum-NSE (difference of 0.03 [CI: −0.04; 0.09], p = 0.44). However, within this sub-group, frozen-serum-NSE performed better than frozen-plasma-NSE with an AUROC difference of 0.08 (CI: −0.15; −0.01), p = 0.02. Serum-NSE had greater accuracy in predicting 1-year mortality than frozen-plasma-NSE.