Serum neurofilament light chain and multimodal neuroprognostication after cardiac arrest - A retrospective cohort study.

Background And Purpose: Most patients remain comatose within the first days after cardiac arrest (CA) and resuscitation. Guidelines recommend multimodal neuroprognostication including neuron specific enolase (NSE) as serum biomarker. Neurofilament light-chain (NFL) may have higher prognostic accuracy earlier after cardiac arrest and a lower risk of confounders. This study investigates the prognostic value of serum NFL in clinical routine compared to established prognostic tests.

Methods: Monocentric retrospective observational study of patients with serum NFL between 24-96 h after CA. Neurological outcome was evaluated at hospital discharge via the Cerebral Performance Category score (CPC), dichotomized as good (CPC 1-3) and poor (CPC 4-5). Prognostic performance for good and poor neurological outcome prediction was analysed for NFL, NSE, electroencephalography (EEG), somatosensory evoked potentials (SSEP), and head computed tomography (CT). NFL was measured using the SIMOA Quanterix assay.

Results: 152 patients were included, median age was 61 years, 24% were female. 10 patients were discharged in vegetative state or comatose (7%), and 78 died before discharge (51%). NFL > 2000 pg/ml predicted poor outcome with 53% (43-63%) sensitivity and 100% (94-100%) specificity. Most patients (69%) with NFL < 55 pg/ml had a good outcome. Predictive accuracy was similar to other neuroprognostic tests (AUC 0.89, 0.84-0.94). In head-to-head comparisons with the other neuroprognostic tests, NFL identified 16-41% additional poor outcome patients.

Conclusion: NFL (SIMOA) > 2000 pg/ml predicts poor neurological outcome with high specificity, while low concentrations strongly argue against severe HIE. Adding NFL to established neuroprognostication tests increases sensitivity of poor outcome prediction.