Approach to the unconscious patient in the ICU


This is all about generating a broad range of differentials.

One's assessment of the unconscious patient searches for focal neurological signs and meningism. If these are absent, one is left looking for subtle clues in the examination which may explain the decreased level of consciousness. Apart from our very own Oh's Manual, one may also examine an excellent article by David E Bateman from 2001, which details the neurological assessment of coma. A suggested pattern of examination for the unconscious patient in the CICM exam is available elsewhere.

SAQs requiring the candidates to generate a list of differentials for unconsciousness have included the following:

Generating a list of differentials for the decreased level of consciousness

Differential Diagnosis of Unconsciousness

With focal  signs

Vascular causes:

  • Stroke
  • Vascular insufficiency of the brain, eg. critical vessel stenosis of some specific vessel
  • Intracranial haemorrhage

Infectious causes:

  • Brain abscess
  • Meningoencephalitis with focal cranial nerve damage

Neoplastic causes

  • Space-occupying tumour

Intrinsic neurological causes

  • Pre-existing focal neurology, superimposed on an acute unconscious state

Autoimmune causes

  • Cerebral vasculitis

Traumatic causes

  • Focal neurological injury due to head trauma
    • That includes surgical injury due to retraction of structures, as in elective neurosurgery
  • Increased intracranial pressure, giving rise to false localising signs

Without focal  signs

Vascular causes:

  • Brainstem stroke, resulting in damage to the reticular activating system
  • Vascular insufficiency of the brain, eg. diffuse cerebral small vessel disease

Infectious causes:

  • Encephalitis without meningism
  • Neurological sequelae of systemic infection, eg. septic encephalopathy

Drug-related causes:

  • Persisting effects of sedatives in the context of diminished clearance

Intrinsic neurological cause

  • Delirium of prolonged intensive care stay - a "hypoactive" form thereof
  • Seizures (post
  • Non-convulsive status epilepticus
  • Raised intracranial pressure

Autoimmune causes

  • Cerebral vasculitis

Traumatic causes

  • Sequelae of diffuse brain injury, eg. diffuse axonal injury
  • Hypothermia or hyperthermia

Endocrine and metabolic causes:

  • Hypoadrenalism
  • Hypothyroidism
  • Hepatic encephalopathy
  • Uremic encephalopathy
  • Wernicke's encephalopathy

With meningism

Vascular causes:

  • Subarachnoid haemorrhage

Infectious causes:

  • Meningitis


Unconsciousness with focal signs

This is typically easy. Something is structurally wrong with the brain.

There is either a stroke, bleeding, pus or tumour.

Unconsciousness with features of meningism

Again, this is easy. Not many things can irritate the meninges.

One must consider either meningitis or a subarachnoid haemorrhage.

Unconsciousness without any specific features

This is much more difficult. One has to cast one's net widely.

Vascular causes: brainstem infarct or haemorrhage; alternatively, a stroke which has wiped out a large portion of the brain mass (eg. a hemispheric infarct)

Infectious causes: encephalitis due to an infectious agent

Neoplasm: more likely, multiple metastases

Drugs: of these, there is more than a handful.

Autoimmune: Lupus cerebral vasculitis springs to mind, but here are several others

Trauma: concussion, cerebral contusions or a diffuse axonal injury can give rise to a loss of consciousness without any focal features.

Endocrine and metabolic causes: these are numerous:

  • Hyper and hyponatremia
  • Hyper and hypoglycaemia
  • Hypoxia
  • Hypercapnia
  • Myxoedema
  • Uraemia
  • Hepatic encephalopathy
  • Adrenal insufficiency
  • Non-convulsive status epilepticus

Assessing the level of consciousness

Many scales are available, but most people rely on the old and trusted Glasgow Coma Scale.

The GCS has a few important problems, and a good editorial on this topic is available, which summarises the problems pretty well.

In spite of its shortcomings, the GCS should form the first steps of a neurological assessment, particularly where it comes to an undifferentiated unconscious patient.

Assessing the eyes

Testing the pupils gives one some reassurance regarding the entire optic pathway. The retina, the optic nerve, the chiasm, the midbrain and third nerve - all play a role in the light response. In general, the presence of full conjugate movement on oculovestibular stimulus suggests that much of the brainstem is intact.

The possible causes of abnormal pupils are discussed in the following table:

Unilateral miosis Bilateral miosis Unilateral mydriasis

Bilateral mydriasis

Horners syndrome

Damaged sympathetics



Cholinergic agents


Pontine lesions

Thalamic lesisons


Uncal herniation

Midbrain lesions

Hypoxic brain injury

Bilateral midbrain lesion

Sympathomimetic drugs

Anticholinergic drugs

From this table, one should carry away the awareness that midbrain lesions cause a dilated pupil, whereas pontine lesions cause a constricted one (as the sympathetic outflow is interrupted by pontine lesions, but only the third nerve is affected by midbrain lesions).

This whole eye examination thing can really get out of hand. If one is interested in an extensive discussion of eye signs, one can pursue their demented lust for cranial nerve physiology in the chapter on examination of the cranial nerves in the critically ill patient.

The "blown pupil"

Uncal herniation causes a third nerve palsy by stretching the oculomotor nerve over the petroclinoid ligament, a delicate band of dura stretching between the clinoid process and petrous portion of the temporal bone (purists may remark that it is an extension of the tentorium cerebelli). As the brainstem is pushed down through the foramen magnum, so the oculomotor nerve is dragged down with it, and the ligament becomes the fulcrum point where it is compressed. The parasympathetic fibres in the nerve no longer supply the iris in such circumstances, and there is unopposed sympathetic input into the pupil's diameter.

This pattern occurs early in uncal herniation because the third nerve is stretched before any major structures are being crushed. In contrast, when the herniation is central, the pupils are usually small - their parasympathetic supply remains undisturbed while the sympathetic fibers in the brainstem are being squashed. Late in the game, Cheyne-Stokes respiration and fixed mid-dilated pupils develop.

Eye movements

As far as the eye movements go, there are several pearls one should recall:

  • Presence of spontaneous eye movements excludes brainstem pathology as the cause of coma. The 3rd, 4th and 6th nerves are all engaged, and the medial longitudinal fasciculus coordinates conjugate eye movements.
  • Horisontal movements of the eyes are coordinated by the contralateral hemisphere, via the ipsilateral pons. The right hemisphere commands the left pons to look left. Thus, in hemispheric cortical pathology, the eyes will look towards the lesion, whereas in pontine pathology they will deviate away from it.
  • Vertical movements of the eyes are coordinated by both hemispheres.

Assessing the motor score

Any spontaneous movement is good. Even abnormal movements are good -abnormal movements are a focal sign and may give one some clue as to where the lesion is:

  • Chorea and hemiballismus: basal ganglia lesion
  • Myoclonic jerks: more likely due to a global hypoxic injury

Abnormal posturing

This is a part of the motor score in the GCS, and if not present spontaneously, it can be elicited by a painful stimulus.

  • Decerebrate posturing: rigidity with arms inwardly rotated; this is generally held to be a brainstem problem and is supposed to be due to an exaggeration of the reflexes normally relied upon to defeat gravity when standing upright.
  • Decorticate posturing sees the patient assume a "pugilistic" pose, with arms flexed as if ready to commence a boxing match. This is generally thought of as a cerebral problem (i.e. a lesion above the brainstem).

The red nucleus of the rostral midbrain is responsible for this physical sign. The rubrospinal tract, when unopposed by cortical activity, will cause a flexion of the upper limbs, resulting in the characteristic posture of the decorticate patient. Lesions below the red nucleus prevent this flexor response, and in response to pain the patient will tense and extend the limbs- this is called reactive extensor postural synergy.  Spinal reflex arcs are still present, and this extension will produce a reflexive flexion of the opposing muscle groups, which results in the observed whole-body rigidity.

Each seems as bad as the other, but there is a reason why they are given different ratings in the Glasgow Coma Scale. There is an empirically observed survival difference between the two findings. The creators of the scale (in 1974) found that head-injured Scottish drunks with decorticate posturing recovered in 37% of cases, whereas the decerebrate drunks only recovered in 10% of cases.

Confusingly, metabolic states can result in this sort of finding. One article from the 1960s reports on both being present (and being reversible) in a patient with hepatic encephalopathy.

Influence of coma aetiology on the respiratory pattern

A decreased respiratory rate makes one think of opiates first. However, hypothyroidism can also be to blame. Additionally, the decreased respiratory rate can be the cause of coma, as it can cause hypercapnia due to an otherwise trivial non-anaesthetic opiate dose.

An increased respiratory rate is sometimes seen in midbrain lesions and in various sorts of metabolic encephalopathy. Certainly, these various metabolic problems are rarely without an acid-base component, and the disturbance is almost always an acidosis.

Cheyne-Stokes respiration is an "oscillatory" pattern of breathing, with rate alternating between fast and slow- and occasionally with apnoea. An ancient paper on the subject (from 1909, co-authored by J.S. Haldane!) describes this pattern very well for those who have never seen it. It is seen mainly in patients with metabolic encephalopathy and (nocturnally) in heart failure; however, the mechanism can extend to patients with increased intracranial pressure due to some sort of diffuse insult. The issue seems to be still debated. In general, the mechanism is thought to be some sort of dysregulation of the central control of breathing, with periodic hyperventilatory overcompensation for hypoxia and hypercapnia leading to periods of apnoea, and thus more hypoxia and hypercapnia.

Oh's Manual also mentions "apneustic breathing" where one pauses in inspiration (apparently a pontine problem) and "ataxic breathing" which is totally disorganised and apparently a feature of pre-terminal medullary dysfunction.

"Will he ever wake up, doctor?"

Prognostication in coma is a delicate thing, particularly if one is not sure what has caused it.

At least where it comes to hypoxic brain injury, we have some clinical predictors to fall back on. More of this is explored in the chapter on the prognostication of neurological recovery after cardiac arrest. In short, the following is a list of poor prognostic features:

  • Non-CPR downtime of over 8 minutes
  • ROSC after more than 30 minutes
  • Absent pupillary responses after 72 hours
  • Poor motor response after 72 hours (anything worse than withdrawal)
  • Absent spontaneous eye movements after 24 hours

Of course, these findings are all from studies which were done before the era of therapeutic hypothermia. As we are reminded by Venkatesh, who the hell knows what the outcomes are for cooled patients. It is possible that excessive negativity and pessimism on our part is leading to a number of unfair decisions to withdraw care.

Approach to the undifferentiated coma

Though the list of differentials is broad, a stereotypical response should be reproduced for any case of unexplained coma.

  • Control the airway.
  • Administer glucose.
  • Administer thiamine.
  • At least think about antidotes to the most obvious drugs. Naloxone comes to mind.
  • Briefly examine the motor response and the pupils.
    • If there are any features of increased intracranial pressure or herniation, administer a rescue dose of hyperosmolar therapy, be it hypertonic saline or 20% mannitol.
  • Order some basic bloods, of which the renal and hepatic function will be the most helpful. Serum osmolality and blood ketones could also be informative.A toxicology screen would never be a bad idea.
  • Organise a CT scan. In some cases, an MRI will usually follow (eg. if a brainstem lesion or acute infarction are suspected).
  • Perform a lumbar puncture. Of course, if none of the other investigations have yielded an explanation, then the lumbar puncture may not add very much. Typically, it is a means of confirming a CNS infection, and culturing the organism.
  • An EEG is occasionally useful. Typically, the report says "diffuse cerebral dysfunction". There are only a few situations when it would actually help in the diagnostic process:
    • In hepatic encephalopathy (there are characteristic triphasic waves)
    • In herpes encephalitis (there are apparently characteristic "spikes")
    • In non-convulsive status epilepticus: one study has discovered that among patients with an otherwise unexplained coma, 8% turned out to have undiagnosed status epilepticus after an EEG.
    • To talk about the use of EEG in diagnosis and prognostication of hypoxic brain injury seems somewhat irrelevant, given the minimal weighting these findings are given by the ANZIC statement on brain death. However, at least one CICM fellowship question has touched on this topic. Thus, I suppose we candidates should remember that severe hypoxic encephalopathy is associated with the following EEG features:
      • Presence of theta activity
      • Diffuse slowing
      • Burst suppression
      • Alpha coma



Oh's Intensive Care manual: Chapter 49   (pp. 549) Disorders  of  consciousness  by Balasubramanian  Venkatesh

Plum, Fred, and Jerome B. Posner. The diagnosis of stupor and coma. Vol. 19. Oxford University Press, 1982. - warning! This link takes you to a download of the whole 9.0Mb file, which represents the entire volume of the 1980 second edition.

Maramattom, Boby Varkey, and Eelco FM Wijdicks. "Uncal herniation."Archives of neurology 62.12 (2005): 1932-1935.

Skrzat, J., et al. "The clinical significance of the petroclinoid ligament." Folia morphologica 66.1 (2007): 39-43.

Conomy, John P., and Michael Swash. "Reversible decerebrate and decorticate postures in hepatic coma." New England Journal of Medicine 278.16 (1968): 876-879.

Bateman, David E. "Neurological assessment of coma." Journal of Neurology, Neurosurgery & Psychiatry 71.suppl 1 (2001): i13-i17.

Feldman, Martin H., and Shirley Sahrmann. "The Decerebrate State in the Primate: II. Studies in Man." Archives of neurology 25.6 (1971): 517-525.

Teasdale, Graham, and Bryan Jennett. "Assessment of coma and impaired consciousness: a practical scale." The Lancet 304.7872 (1974): 81-84.

Dowell, Anthony R., et al. "Cheyne-Stokes respiration: a review of clinical manifestations and critique of physiological mechanisms." Archives of internal medicine 127.4 (1971): 712.

Andreas, Stefan, et al. "Cheyne-Stokes respiration and prognosis in congestive heart failure." The American journal of cardiology 78.11 (1996): 1260-1264.

Douglas, C. Gordon, and J. S. Haldane. "The causes of periodic or Cheyne-Stokes breathing." The Journal of physiology 38.5 (1909): 401-419.

Tomlinson, B. E. "Brain-stem lesions after head injury." Journal of Clinical Pathology. Supplement (Royal College of Pathologists). 4 (1970): 154.

Perman, Sarah M., et al. "Timing of neuroprognostication in postcardiac arrest therapeutic hypothermia." Critical care medicine 40.3 (2012): 719.

Towne, A. R., et al. "Prevalence of nonconvulsive status epilepticus in comatose patients." Neurology 54.2 (2000): 340-340.

TEASDALE, GRAHAM, and BRYAN JENNETT. "Assessment of coma and severity of brain damage." Anesthesiology 49.3 (1978): 225.

Green, Steven M. "Cheerio, laddie! Bidding farewell to the Glasgow Coma Scale." Annals of emergency medicine 58.5 (2011): 427-430.

Gill, Michelle R., David G. Reiley, and Steven M. Green. "Interrater reliability of Glasgow Coma Scale scores in the emergency department." Annals of emergency medicine 43.2 (2004): 215-223.

Riechers, Ronald G., et al. "Physician knowledge of the glasgow coma scale."Journal of neurotrauma 22.11 (2005): 1327-1334.