Developing electroencephalography as a tool to predict delirium and long-term cognitive outcome after traumatic brain injury.

Approximately 1 million patients are hospitalised every year in Europe after Traumatic Brain Injury (TBI).

The most severe TBI patients are admitted to Intensive Care Units (ICU) and receive high-dose sedation to prevent and reduce brain oedema. Clinicians stop sedation when an assessment of neurological recovery is required.

While a few patients remain comatose or only partially regain awareness, the majority of patients recover but many experience delirium on recovery.

Delirium can present as an altered mental state with confusion, disorientation, poor attention, amnesia and motor agitation (hyperactive subtype), drowsiness (hypoactive subtype), or both (mixed subtype).Delirium is frequent in the general ICU population (up to 87%) and is associated with longer ICU stay and increased adverse outcomes.

Due to the heterogeneity of symptoms, the poor reproducibility of screening tools and lack of reliable prediction models, delirium is often under-recognised and difficult to predict.The role of malfunctioning brain networks in the pathophysiology of delirium development is being progressively supported by electroencephalography (EEG) and brain imaging findings.

TBI patients are at the highest risk of brain network dysfunction; the injury itself leads to structural damage of brain connections and neurotransmitter imbalance triggered by high-dose sedative drugs disrupts the brain's functional connectivity.Delirium in ICU isone of the top three priorities for intensive care research and its management represents a significant unmet need for patients who survive TBI.

During this fellowship, my ultimate aim is to develop a bedside tool based on EEG functional connectivity to identify patients at high risk of developing delirium following TBI.

I will achieve this goal by performing: 1) an international audit of >2000 ICU patients, 2) an observational EEG study in 150 ICU patients following TBI and 3) an anaesthesia EEG study in healthy volunteers.This will allow me to a) describe the prevalence and clinical features of delirium;b) allow early prediction of delirium by separating the EEG signatures of prolonged sedation from those of persistent neurological impairment andc) understand brain mechanisms underlying each delirium subtype and their impact on long-term outcome.The international audit on the practices of sedation, analgesia and delirium management in ICU (SAnDMAN study) is funded and underway.

It will identify the prevalence of delirium and its subtypes and provide an international collaborative framework for future validation of the developed EEG metrics in a larger sample of ICU patients.

In the observational patient and healthy volunteer studies, I will test my hypotheses that: 1) the degree of impaired information processing (measured as EEG functional connectivity) correlates with delirium severity in TBI patients in ICU; 2) persistent impaired connectivity at two weeks is associated with long-term cognitive impairment; 3) evoked-responses to tactile and auditory stimuli of patients with hyperactive and hypoactive delirium resemble the responses elicited in healthy volunteers receiving excitatory (ketamine) and depressant (dexmedetomidine) drugs, respectively.

The proposed research project will be carried out in the Anaesthesia Neuroimaging Group, based in the University of Oxford and the John Radcliffe Hospital.