Background: Hemolysis, a common adverse event associated with veno-arterial extracorporeal membrane oxygenation (VA-ECMO), may affect neuron-specific enolase (NSE) levels and potentially confound its prognostic value in predicting neurological outcomes in resuscitated patients without return of spontaneous circulation (ROSC) that require extracorporeal cardiopulmonary resuscitation (eCPR). Therefore, a better understanding of the relationship between hemolysis and NSE levels could help to improve the accuracy of NSE as a prognostic marker in this patient population.
Methods: We retrospectively analyzed the records of patients who received a VA-ECMO for eCPR between 2004 and 2021 and were treated in the medical intensive care unit (ICU) of the University Hospital Jena. The outcome was measured clinically by using the Cerebral Performance Category Scale (CPC) four weeks after eCPR. The serum concentration of NSE (baseline until 96 h) was analyzed by enzyme-linked immunosorbent assay (ELISA). To evaluate the ability of individual NSE measurements to discriminate, receiver operating characteristic (ROC) curves were calculated. Serum-free hemoglobin (fHb, baseline until 96 h) served as a marker for identifying a confounding effect of parallel hemolysis.
Results: 190 patients were included in our study. A total of 86.8% died within 4 weeks after ICU admission or remained unconscious (CPC 3-5), and 13.2% survived with a residual mild to moderate neurological deficit (CPC 1-2). Starting 24h after CPR, NSE was significantly lower and continued to decrease in patients with CPC 1-2 compared to the group with an unfavorable outcome of CPC 3-5. In addition, when evaluating on the basis of receiver operating characteristic curves (ROC), relevant and stable area under the curve (AUC) values for NSE could be calculated (48 h: 0.85 // 72 h: 0.84 // 96 h: 0.80; p < 0.01), and on the basis of a binary logistic regression model, relevant odds ratios for the NSE values were found even after adjusting for fHb regarding the prediction of an unfavorable outcome of CPC 3-5. The respective adjusted AUCs of the combined predictive probabilities were significant (48 h: 0.79 // 72 h: 0.76 // 96 h: 0.72; p ≤ 0.05).
Conclusions: Our study confirms NSE as a reliable prognostic marker for poor neurological outcomes in resuscitated patients receiving VA-ECMO therapy. Furthermore, our results demonstrate that potential hemolysis during VA-ECMO does not significantly impact NSE's prognostic value. These findings are crucial for clinical decision making and prognostic assessment in this patient population.
Keywords: CPR; Cerebral Performance Category Scale; ECMO; hemolysis; hypoxic brain injury; intensive care; neuron specific enolase.