By Peter B. Forgacs, MD

Assistant Professor of Neuroscience and Neurology, Feil Family Brain & Mind Research Institute and Department of Neurology, Weill Cornell Medical College; Adjunct Assistant Professor of Clinical Investigation, The Rockefeller University

Dr. Forgacs reports no financial relationships relevant to this field of study.

SYNOPSIS: Studies assessing EEG reactivity in comatose patients after severe brain injuries are highly variable and almost never provide replicable definitions for presence or absence of EEG reactivity, even though it is used increasingly as a prognostic measure.

SOURCE: Admiraal MM, van Rootselaar AF, Horn J. Electroencephalographic reactivity testing in unconscious patients: A systematic review of methods and definitions. Eur J Neurol 2017;24:245-254. doi: 10.1111/ene.13219.

Electroencephalography (EEG) traditionally is used to identify potentially reversible causes of altered levels of alertness in hospitalized patients, such as seizures or status epilepticus and metabolic encephalopathy. More recently, however, some EEG features also have been used to assess prognosis of neurological outcomes after severe brain injuries. Notably, the use of EEG to aid prognostication is increasingly explored in comatose post-cardiac arrest patients, especially since various targeted temperature management (TTM) protocols introduced continuous EEG monitoring as standard of care in these patients. In particular, the absence of EEG reactivity is considered to be one of the indicators of poor outcome. Recently, several guidelines, such as the American Heart Association Post-Cardiac Arrest Care Guideline and the European Resuscitation Council and European Society of Intensive Care Medicine Guideline, advised testing of EEG reactivity for consideration as a prognostic marker after cardiac arrest. However, none of these guidelines and not even the American Clinical Neurophysiology Society Standardized Critical Care EEG Terminology provides specific descriptions of stimulus administration and precise definitions for presence or absence of EEG reactivity.

Admiraal et al addressed this shortfall via a systematic review of all studies between 1970 and May 2016 that assessed EEG reactivity. The inclusion criteria were reports of original research studies in English or Dutch that involved adult patients who were unconscious at the time of testing as a result of a severe brain injury from any etiologies. The authors identified 40 articles based on these criteria and assessed their methodological quality using the QUality In Prognostic Studies (QUIPS) tool. In addition, the authors also rated the descriptions of stimulus protocol and reactivity definitions using a four-category grading scale based on the level of reproducibility.

According to the QUIPS assessments, most studies had reasonably good overall quality. However, descriptions of stimulation protocols and definitions of EEG reactivity definitions were highly variable. While 33/40 studies did specify their stimulation protocol, only three described it in sufficient detail to be exactly replicable. All studies used at least auditory stimulation, but most studies also reported highly variable combination of additional stimuli, including noxious, visual, and sensory stimulations. Of the 27/40 articles that reported their definition of EEG reactivity, none reported it in sufficient detail to be exactly replicable. The authors concluded that EEG reactivity testing clearly is not standardized and descriptions of stimulus protocols and precise definitions of EEG reactivity are highly insufficient in the current literature.


The major finding of this review is that methods of stimulation to elicit EEG reactivity are highly variable and the studies examined almost never provide replicable definitions. As new therapeutic interventions and advancements in acute medical care have led to improved survival and neurological outcomes after severe brain injuries, many previously accurate predictive indicators of outcomes in these patients have become less reliable. As a result, major efforts are now devoted to finding early but accurate prognostic factors applicable to everyday clinical practice, which often is strained to simultaneously provide the best possible treatment to those who have potential for good recovery while limiting care that is futile. However, a major concern is that even though prognostication criteria for outcomes are still evolving in some instances, withdrawal of life-sustaining therapy (WLST) decisions continue to drive mortality in patients who do not readily regain consciousness after severe brain injuries. In addition, in some cases, even well-designed, large, multicenter trials, have high rates of WLST, which carry the risk of self-fulfilling prophecies that may affect outcomes. These considerations highlight the need for rigorous examination of studies assessing accuracy of prognostic factors, as done in this current review, and put emphasis on the ethical obligation to include only clearly defined and highly reproducible prognostic markers in future guidelines.

Based on the findings of this review, future studies aimed to assess EEG reactivity in comatose patients should use a clear protocol for stimulus delivery, use multiple types of stimuli (i.e., auditory, noxious, visual), and apply them multiple times to ensure reproducibility. They should also describe the exact timing of stimulation in relation to time of injury; document concurrent use of sedative medications or other treatments that may affect EEG reactivity (such as hypothermia); and define clear thresholds for changes in EEG frequency, amplitude, and waveforms (e.g., stimulus induced discharges). Importantly, large, prospective, multicenter trials should be conducted based on such reproducible criteria before EEG reactivity is included in prognostication guidelines.

This study also strongly highlights the general need for precise definitions in assessments of any prognostic factor that is intended to be used in prognostication of comatose patients with severe brain injuries. This is especially important as many times these factors influence withdrawal of life-sustaining therapy decisions. Future prospective studies aimed to develop or validate prognostication guidelines should include only highly accurate and reproducible clinical or diagnostic assessments.