Background: In comatose post-cardiac arrest patients, a serum neuron-specific enolase (NSE) level of >33 μg/L within 72 h was identified as a reliable marker for poor outcome in a large Dutch study (PROPAC), and this level was subsequently adopted in an American Academy of Neurology practice parameter. Later studies reported that NSE >33 μg/L is not a reliable predictor of poor prognosis. To test whether different clinical laboratories contribute to this variability, we compared NSE levels from the laboratory used in the PROPAC study (DLM-Nijmegen) with those of our hospital's laboratory (ARUP) using paired blood samples.
Methods: We prospectively enrolled cardiac arrest patients who remained comatose after resuscitation. During the first 3 days, paired blood samples for serum NSE were drawn at a median of 10 min apart. After standard preparation for each lab, one sample was sent to ARUP laboratories and the other to DLM-Nijmegen.
Results: Fifty-four paired serum samples from 33 patients were included. Although the serum NSE measurements correlated well between laboratories (R = 0.91), the results from ARUP were approximately 30% lower than those from DLM-Nijmegen. Therapeutic hypothermia did not affect this relationship. Two patients had favorable outcomes after hypothermia despite NSE levels measured by DLM-Nijmegen as >33 μg/L.
Conclusions: Absolute serum NSE levels of comatose cardiac arrest patients differ between laboratories. Any specific absolute cut-off levels proposed to prognosticate poor outcome should not be used without detailed data on how neurologic outcomes correspond to a particular laboratory's method, and even then only in conjunction with other prognostic variables.