Addressing an Inherent Bias in Neuroprognostication: A Collaboration Between the US and Brazil to Reduce the Impact of Self-Fulfilling Prophecy in Cardiac ARrEst (SPARE) - PROJECT ABSTRACT Hypoxic-ischemic brain injury in cardiac arrest survivors ranges in severity from mild cognitive impairment to per- sistent vegetative state. Precise estimation of neurologic outcome (neuroprognostication) in this setting is crucial to guide families and physicians in goals-of-care decisions and to improve allocation of critical care resources. Outcome prediction tools include neurologic examination, neuroimaging, biochemical markers, and electrophysi- ologic studies. Accurate neuroprognostication remains challenging, as previous studies assessing the prognostic accuracy of these tools have been affected by withdrawal of life support due to perceived poor neurologic prog- nosis (WLS-N), and thus subject to self-fulfilling prophecy bias. This bias, along with early pessimistic prognosti- cation, most often leads to premature death of those dependent on life support, including some with potential for good neurologic recovery. Consequently, the natural history of hypoxic-ischemic brain injury unbiased by the effect of WLS-N has not been adequately investigated. In Brazil, the severity of hypoxic-ischemic brain injury is assessed by prognostic tools similar to those used in the US; however, life-sustaining therapies are maintained for prolonged periods due to cultural norms. This difference in practice, unbiased by WLS-N, allows for a unique opportunity to observe the natural history of hypoxic-ischemic brain injury, and to identify true predictors of poor outcome. The elucidation of the impact of WLS-N on the performance of neuroprognostic tools is a vital step toward more accurate outcome prediction. This will be the first study to combine the advanced expertise from leading international sites analyzing neuroimaging and electrophysiologic data applied to a population unbiased by WLS-N, and to unveil the impact of self-fulfilling prophecy bias on the natural history of hypoxic-ischemic brain injury and on the prediction performance of guideline-recommended neuroprognostic tools. We will prospectively collect data on demographics, premorbid characteristics, details of cardiac arrest and resuscitation, post-cardiac arrest care, detailed neurological examination findings, electrophysiologic studies and neuroimaging at standard- ized time points, and assess functional outcomes at discharge and at 3, 6 and 12 months. Our goals are to: 1) Determine the independent effect of WLS-N on mortality and overall outcomes, and 2) Demonstrate the prognostic accuracy of current guideline-recommended tools in an uncensored population. As an ex- ploratory aim, we will also develop independent, blinded multimodal neuroprognostic models for the US and Brazil cohorts, and compare their performance. By completing this innovative project, we will understand the impact of WLS-N on the natural history of hypoxic-ischemic brain injury and better characterize the timing of recovery after cardiac arrest. These data will influence future neuroprognostication studies and trials of neuro- therapeutics, as well as optimize the use of critical care resources. This study is the stepping stone to strength- ening data on neuroprognostication by improving their reliability, thereby ultimately easing undue patient and family suffering from prognostic uncertainty, and directing medical resources most efficiently.
