LAUSR

We thank Dr Cho et al for their interest in our paper. We respond to their concerns of self-fulfilling prophecy and distortive effects of hypothermia and sedative medication on the electroencephalogram (EEG). First, we acknowledge the risk of self-fulfilling prophecy bias. Potential self-fulfilling prophecy has been a limitation in all studies on EEG, somatosensory evoked potential, imaging, and blood biomarkers for outcome prediction of comatose patients after cardiac arrest. Of studies that form the basis of international treatment guidelines, only 9 of 73 (12%) report complete blinding of the treating team to the predictor under study. Treatment suspension policies are described in 37 of 73 (51%), with short minimal treatment durations (3–7 days). All studies, including ours, acknowledge and discuss this important limitation. To minimize self-fulfilling prophecy in the current study, national guidelines for decisions on withdrawal of life-sustaining treatment were strictly followed. Herein, EEG in the first 3 days is not included. There was no blinding for the EEG, but EEGs were reviewed by neurologists, whereas patients were treated by intensivists. Only upon detection of epileptiform patterns were intensivists informed. Final (predictive) EEG classifications were assigned offline after the recordings, blinded for clinical state, medication, and outcome. Prevention of self-fulfilling prophecy bias would require complete blinding of the treating team to any test result and prolonged life-support in patients with persistent coma. This is difficult. Some predictors, such as clinical examination, cannot be kept hidden, and others, such as EEG, can reveal potentially treatable complications. Also, indefinite support care is associated with ethical and financial problems. Nonetheless, we agree with Cho et al that future studies would benefit from detailed descriptions of criteria for withdrawal of life-sustaining treatment. Second, we stress that suppressed or synchronous patterns on a suppressed background cannot be solely induced by hypothermia, propofol, or midazolam in the administered dosages. We also emphasize that all studies on EEG in comatose patients after cardiac arrest have shown that differences between patients with poor and good outcome are largest in the first 24 hours after cardiac arrest. All available data show that predictive values are high in the first 24 hours, despite hypothermia or sedative medication, and decrease thereafter. Apparently, in postanoxic encephalopathy, early brain activity reflects the extent of hypoxic–ischemic injury. With persistent suppression or synchronous patterns on a suppressed background, prognosis is invariably poor. Otherwise, with appearance of continuous patterns within 12 hours, the probability of a good neurological outcome is large.