College answer

General 

• Testing too early (esp. before 72 hrs) is unreliable
• Hypothermia and sedative/ relaxants confound most tests
• Associated organ impairments (renal, hepatic) may delay sedative drug clearance and cause encephalopathy
• Seizures (convulsive or non-convulsive)
• Many studies done were not blinded – risk of self-fulfilling prophesy

Clinical: 

• Pupil responses may be underestimated cf. pupilometer
• Pre-existing ocular pathology – e.g. cataracts, blindness
• Recent use of high dose adrenaline, eye drops 
• Corneal reflex- Less specific than pupil response
• Motor responses before 72 hrs unreliable
• Status myoclonus is poorly defined
• Lance-Adams syndrome of awake myoclonus not predictive
• Pre-existing weakness or other pathologies

Electrophysiological: 

• Background signal noise may cause false positives.
• Lack of standardisation in measurement
• Electrode placement may be inconsistent
• Poorly defined endpoints

Radiology

• Brain imaging studies are substantially effected by timing of the study as changes evolve over time. All imaging studies limited by small sample size and selection bias.

Biomarkers: 

• Threshold values for and timing not well established
• Measurement and other tissue confounders not well established e.g. in haemolysis. Poorly defined endpoints

Minimising the confounders

• Define the context/ exclude other causes of unconsciousness
• Caution with renal or hepatic impairment
• Knowledge of any pre-existing pathologies from history
• Waiting at least 72 hrs longer before testing if hypothermia/ sedation/ relaxant
• Multiple modality testing is more reliable than single tests
• Repeated observation especially when patient is hypothermic/ recent sedation or there is doubt Most unconscious patients will recover within 5 days and nearly all by 8 days.
• Skill in interpretation is required for most test especially electrophysiology and imaging
• Use of TOF to exclude paralysis
• Be aware of the risk of self-fulfilling prophesy.

Examiners Comments:

Overall poorly answered with limited detail and little attention paid to the factors which confound prognostication.

Discussion

This question has apppeared many times before in varying forms. This incarnation most closely resembles Question 4 from the second paper of 2014. This time, of the confounding factors,  the college also asked "and how they can be minimised". The table below was plagiarised from the chapter on prognostication after cardiac arrest, with little modification