Neuroprognostication Bias: A Collaboration to Reduce the Impact of Self-fulfilling Prophecy in Cardiac ARrEst

Cardiovascular disease remains the leading cause of death in the United States. Mortality rates of cardiac arrest range from 60-85%, and approximately 80% of survivors are initially comatose. Of those who survive, 50% are left with a permanent neurological disability, and only 10% are able to resume their former lifestyle. Early prognosis of comatose patients after cardiac arrest is critical for management of these patients, yet predicting outcome for these patients remains quite challenging.

The primary study objective of SPARE is to assess the value of using a systematic, multi-modal approach for neuroprognostication in the unconscious post-cardiac arrest population. We hypothesize that prognostication using this approach will be significantly improved compared to historical controls. This approach will be novel because:

All patients who are unconscious at least 24 hours post-cardiac arrest, whereas previous studies on neurologic outcome tended to have restrictive inclusion criteria, such as no pre-existing neurologic impairment (e.g. dementia or prior cerebrovascular injury), or included an unduly restrictive population, such as patients with a strictly comatose state.

The prognostic modalities used to assess patients will be applied at specific time points that will maximize their utility.

Patients' families and clinicians will be encouraged to provide adequate time to allow for a delayed recovery, especially in cases of uncertain outcome, thus minimizing the self-fulfilling prophesy bias of early withdrawal of life-sustaining therapies (WLST). This will be particularly pertinent in the comparison of US and Brazil/Italy patients, as the Brazilian and Italian populations are not commonly exposed to premature WLST (as can be the case in the US), one of the major sources of biases in prognostication studies of cardiac arrest due to the self-fulfilling prophecy.

SPARE is a multi-center, international, prospective registry designed to evaluate the use of a multi-modality approach to neuroprognostication after cardiac arrest. Subjects will be evaluated with standard, accepted, and widely available assessment modalities, including clinical examination, neurophysiologic (electroencephalography and evoked potentials (per site standard of care), serum biomarkers (per site standard of care), and neuroimaging testing.

The purpose of this study is to collect data from a prospective large-scale cohort involving cardiac arrest survivors, and the clinical characteristics and prognostic features that affect their neurologic outcome. The ultimate goal is to derive a prediction model for neuroprognostication in cardiac arrest, using multiple clinical modalities that are already clinically in use, but in a standardized fashion. We hypothesize that by using a multimodal approach combining clinical assessment tools obtained at standardized time points, we will improve the accuracy of neuroprognostication in initially unconscious cardiac arrest survivors. The US and non-US populations will be compared, as the non-US population is less exposed to early WLST, thus eliminating the self-fulfilling prophecy bias that has plagued all CA studies to date.

Outcomes will be assessed at discharge, at 3 months post-arrest, 6 months, and annually up to 5 years afterwards. The primary outcome will be the proportion of subjects with good versus poor outcome, with a dichotomized approach of the modified Rankin Scale (mRS): good outcome defined as mRS scores of 0-3, and poor outcome as mRS scores of 4-6. Secondary outcome measures include overall scores on the Cerebral Performance Category Scale (CPC), Cerebral Performance Category - Extended (CPC-E), and Montreal Cognitive Assessment (MOCA) (or Telephone Montreal Cognitive Assessment (T-MOCA)).