Quality of evidence in studies evaluating neuroimaging for neurologic prognostication in adult patients resuscitated from cardiac arrest

Aim: Neuroimaging has been proposed as a predictor of neurologic outcome in comatose survivors of cardiac arrest. We reviewed the quality and level of evidence of the current neuroimaging literature for predicting neurologic outcome in cardiac arrest patients treated with or without therapeutic hypothermia (TH).

Data sources: Medline, EMBASE, and Cochrane Databases were searched using the terms "cardiac arrest," "cardiopulmonary resuscitation," "brain hypoxia," "brain anoxia," "brain hypoxia-ischaemia," "neuroimaging," and "prognosis." Eligible studies were reviewed and classified by level of evidence and methodological quality as defined by the International Liaison Committee on Resuscitation (ILCOR).

Results: 928 studies were identified, 84 of which met inclusion criteria: 74 were supportive of neuroimaging to predict outcome, eight unsupportive, and two equivocal. Several studies investigated more than one imaging modality: 27 investigated computed tomography (CT), 46 magnetic resonance imaging (MRI), and 18 alternate imaging modalities, including positron emission tomography and single photon emission computed tomography. No randomized controlled trials were identified. Seven cohort and case control studies were identified, only one of which was graded "good" quality, two were "fair" and four were "poor."

Conclusion: Neuroimaging is an evolving modality as a prognostic parameter in cardiac arrest survivors. However, the quality of the available literature is not robust, highlighting the need for higher quality studies before neuroimaging can be supported as a standard tool for prognostication in the patient population.