Neurology: A Clinician's Approach (Cambridge Medicine (Paperback)), 1st Ed.

2. Coma


Coma is a state of eyes-closed unresponsiveness in which even the most vigorous stimulation fails to arouse the patient.1 Because comatose patients cannot communicate, the history must be assembled from family members, emergency service records, and hospital notes. Clues to the etiology of coma obtained from the history include the presence of trauma, evidence of intoxication, and history of cardiac, pulmonary, hepatic, and renal disease. The tempo of coma onset may also be helpful: sudden onset in the absence of trauma favors a cardiogenic source or intracranial hemorrhage, whereas gradual onset is more consistent with a metabolic cause or a slowly expanding mass lesion. In many cases, the history contains few details beyond the patient being “found down,” and the evaluation of the comatose patient quickly shifts to physical examination and diagnostic testing.


Mental status examination

The purpose of the mental status examination of the comatose patient is to verify that they are actually comatose rather than just encephalopathic. Before beginning the examination, make sure that any short-acting sedatives such as midazolam or propofol are discontinued. By definition, a comatose patient’s eyes should be closed and they should appear as if they are sleeping. If gently calling out their name does not produce any response, yell out their name or gently squeeze their hand. Attempt to awaken them with increasingly noxious stimuli: severely encephalopathic patients may respond to painful maneuvers such as rubbing the sternum, applying nailbed pressure, or pinching the areola. Comatose patients will not. Document the reaction to each stimulus and also note what happens when it is withdrawn.

Pupillary reactions

Abnormal pupillary reactions may provide insight into structural causes of coma involving the thalamus and brainstem. A quick rule of thumb (with several important exceptions) is that symmetric pupils, even if abnormal in size and unreactive, are more likely due to toxic or metabolic lesions, while asymmetric pupils are more likely due to structural lesions. Before assigning too much weight to an abnormal pupillary examination, exclude preexisting pupillary irregularities, such as those that might be due to prior cataract surgery. Chapter 7 contains a more detailed discussion of pupillary neuroanatomy and function. The following patterns of pupillary reactions are the most important ones in comatose patients:

1. Normal-sized pupils with normal reactions (Figure 2.1A). This pattern argues against a structural source of coma and is more suggestive of a toxic or metabolic disturbance.

2. Small, reactive pupils (Figure 2.1B). Although thalamic lesions may produce this pattern, patients with small, reactive pupils are more likely to have toxic or metabolic disturbances.

3. Unreactive midsized pupils (Figure 2.1C). Midbrain lesions may produce unreactive midsized pupils. More commonly, however, this pattern is the result of toxic or metabolic disturbances.

4. Unreactive pinpoint pupils (Figure 2.1D). Pontine lesions classically produce pinpoint pupils. It is more common, however, for patients with unreactive pinpoint pupils to have a toxic or metabolic disturbance, particularly opioid intoxication.

5. Asymmetric pupils, abnormal pupil is dilated (Figure 2.1E). The most common causes of this pattern in comatose patients are uncal herniation and ruptured posterior communicating artery aneurysm. Both conditions are true neurological

Figure 2.1

Figure 2.1 Pupillary reactions in coma. See text for details. PCOM, posterior communicating artery.

emergencies that require rapid evaluation and treatment.

6. Asymmetric pupils, abnormal pupil is constricted (Figure 2.1F). Coma accompanied by Horner’s syndrome points to lateral brainstem damage.

7. Fixed and dilated pupils (Figure 2.1G). This pattern suggests severe coma or brain death.

Blink reflexes

The sensory nerve fibers responsible for the blink reflex originate in the cornea and travel in the ophthalmic branch of the trigeminal nerve (Figure 2.2). These trigeminal nerve fibers synapse in the ipsilateral principal sensory nucleus of the trigeminal nerve and the nucleus of the spinal trigeminal tract in the pons and medulla. Neurons originating from these trigeminal nuclei send axons to both the ipsilateral and contralateral facial nuclei. The facial nucleus gives rise to the facial nerve, which innervates the ipsilateral orbicularis oculi, contraction of which produces blinking.

To assess the corneal reflex, peel both eyelids open and gently stroke the sclera and cornea with a wisp of cotton or sterile gauze. Both eyes should blink in response to this stimulus. Test the blink reflex in each eye in sequence. The following are the important blink reflex patterns found in comatose patients (Figure 2.3):

Figure 2.2

Figure 2.2 Schematic of the blink reflex. See text for details.

1. Normal responses in both eyes point to preserved integrity of the blink reflex pathways in the pons and medulla (Figure 2.3A).

2. Stimulation of the right eye produces no blink in either eye, while stimulation of the left eye produces normal blink responses in both eyes (Figure 2.3B). This pattern points to dysfunction of the right trigeminal nerve or nuclei in the pons and medulla.

Figure 2.3

Figure 2.3 Common patterns on blink reflex testing in patients in coma. See text for details.

3. Stimulation of either eye fails to produce a blink response in the right eye (Figure 2.3C). The lesion in this case is in the right facial nucleus or nerve.

4. Stimulation of either eye produces a blink in the ipsilateral eye, but not in the contralateral eye (Figure 2.3D). This pattern suggests dysfunction of the pathways connecting the trigeminal nuclei to the contralateral facial nucleus in the pons and medulla.

5. Bilaterally absent blink responses point to severe brainstem dysfunction, which may be due to either structural or metabolic processes (Figure 2.3E). Bear in mind that patients who wear contact lenses may also lose their blink responses.

Eye position

Horizontal eye position

The frontal eye fields (FEF) in the frontal lobes are the most important structures in the supranuclear control of horizontal eye movements. Projections from the FEF synapse with the contralateral abducens nucleus. Thus, activation of the FEF or inhibition of the abducens nucleus produces contralateral eye deviation, while inhibition of the FEF or activation of the abducens nucleus leads to ipsilateral eye deviation. Supranuclear, nuclear, and infranuclear lesions may lead to abnormal eye positions, which can help to localize the process responsible for coma. In all cases, it is helpful to interpret eye deviation in the context of any associated hemiparesis (Figure 2.4):

1. Destructive right frontal lesions such as strokes or tumors produce rightward deviation of the eyes accompanied by left hemiparesis (Figure 2.4A).

2. Irritative right frontal lesions such as seizures produce leftward deviation of the eyes. There may or may not be a left hemiparesis (Figure 2.4B).

3. Right thalamic lesions produce “wrong-way eyes” that are deviated to the left and are accompanied by left hemiparesis (Figure 2.4C).2

4. Right pontine lesions produce leftward eye deviation. Left hemiparesis may or may not be present (Figure 2.4D).

In addition, horizontal dysconjugate gaze abnormalities (the eyes look in different directions) are often helpful in localizing coma. Common patterns include:

1. Exodeviation of both eyes. This is the pattern seen in many patients with coma, and usually does not have localizing value.

2. Hypo- and exodeviation of one eye (“down and out”) secondary to ipsilateral third-nerve palsy.

3. Esodeviation of one eye secondary to sixth-nerve palsy.

Vertical eye position

The supranuclear control of vertical eye movements is more complex, and involves the bilateral frontal lobes and structures within the brainstem including the vestibular nuclei and interstitial nucleus of Cajal in the midbrain. The important abnormalities of vertical ocular eye position in coma include:

1. Downward deviation of the eyes, which suggests a severe dorsal midbrain lesion.3

2. Vertical ocular misalignment pointing to skew deviation from a brainstem lesion on to fourth-nerve palsy.

3. Hypo- and exodeviation of one eye (“down and out”) secondary to ipsilateral third-nerve palsy.

Spontaneous eye movements

The spontaneous eye movements of comatose patients are usually slow and roving or absent altogether. Absent eye movements suggest a greater depth of coma, and possibly brain death, but do not have particular localizing value. Ocular bobbing is characterized by quick

Figure 2.4

Figure 2.4 Patterns of horizontal eye deviation and hemiparesis in patients with coma.

downward eye movements, which are followed by a slower return back to the primary position, and classically reflect pontine damage.4 Dipping refers to slow downward eye movements with a quicker upward return: this finding has less localizing value than bobbing. Bobbing and dipping may also have inverse forms, in which the first movement is upwards rather than downwards.

Oculocephalic responses

The oculocephalic response may be assessed by the head thrust maneuver or by cold caloric testing. To test the oculocephalic response via the head thrust maneuver, grasp the head by the forehead and chin. Hold the eyes open and turn the head briskly to one side. In comatose patients with intact brainstem function, the eyes should turn in the direction opposite to head rotation. Patients with structural brainstem lesions may demonstrate dysconjugate eye movements. Those who are deeply comatose or brain dead have no eye movements at all. Do not use the head thrust maneuver in patients with possible cervical spine instability, as neck manipulation may worsen motor deficits and even lead to paralysis.

Because the head thrust maneuver is only a weak stimulus to eye movement, most comatose patients require cold caloric testing to properly assess the oculocephalic response. To test cold caloric responses, place the head of the bed at 30° above the horizontal, thereby aligning the horizontal semicircular canal parallel to the ground. Examine the auditory canal to ensure that excessive cerumen accumulation will not interfere with the test, and disimpact the ears as necessary. Fill a 60 ml syringe with ice water and attach the syringe to a short piece of intravenous tubing. Place the tubing into the ear and infuse the ice water slowly over 5 minutes. If brainstem function is intact, then the eyes should deviate towards the side of ice water infusion. After performing the test on one ear, wait approximately 5 minutes for the vestibular system to reset and then test the opposite ear. Important patterns of oculocephalic response testing are shown in Figure 2.5:

1. Cold water placed in either ear produces ipsilateral eye deviation (Figure 2.5A). This is the expected response in a patient with a metabolic encephalopathy.

2. Cold water placed in the right ear produces no response (Figure 2.5B). Cold water placed in the left ear produces tonic ipsilateral eye deviation. This is the pattern seen in patients with right vestibular nerve or right lateral pontine damage.

3. Cold water placed in the right ear produces rightward eye deviation of the right eye only (Figure 2.5C). Cold water placed in the left ear produces leftward eye deviation of the left eye only. This is consistent with a midline lesion of the midbrain and pons producing bilateral internuclear ophthalmoplegia (Chapter 6).

Figure 2.5

Figure 2.5 Important patterns of cold caloric testing in patients with coma.

4. Cold water placed in either ear produces no response (Figure 2.5D). This occurs with severe coma or brain death.

Motor examination

The motor examination helps to determine the presence and severity of coma and in some cases localizes the responsible lesion. Movements may be divided into the following four categories.

Spontaneous and purposeful

Spontaneous, purposeful movements indicate that the patient is not comatose, and should prompt evaluation for encephalopathy as discussed in Chapter 1.

Spontaneous but nonpurposeful

For comatose patients, the most important spontaneous, nonpurposeful movement is polymyoclonus caused by anoxic brain injury. These movements are characterized by brief muscle jerks of the arms, legs, and face, followed by relaxation of the involved muscles. Sometimes they may take the form of violent jaw closure and result in tongue laceration or even severing of a mechanical airway (Chapter 14).


Comatose patients may demonstrate one of several reflex movements. In most patients, these are limited, local, nonpurposeful movements. They may be differentiated from normal movements by their lack of habituation to repeatedly applied, painful stimuli. The most widely known reflex movements in coma are decorticate and decerebrate posturing. In decorticate posturing, a painful stimulus causes flexion at the elbows, wrists, and fingers, and adduction of the arms. In decerebrate posturing, a painful stimulus causes internal rotation of the arms with extension at the elbows and flexion–pronation at the wrists. In both decorticate and decerebrate posturing, there is extension at the hips, extension at the knees, and plantarflexion at the ankles. The anatomical basis of decorticate and decerebrate posturing is less well defined in humans than in laboratory animals, and both may result from nonstructural, metabolic processes or from brainstem pathology. Both forms of posturing are associated with a poor outcome, with decerebrate posturing portending a worse prognosis.


If the patient does not move spontaneously or in response to verbal command, compress the fingernail or toenail bed with the handle of a reflex hammer. Severely encephalopathic patients but not those in coma may purposefully move the hand or foot away from such a stimulus. In some patients, a painful stimulus may produce only a facial grimace or heart rate elevation with no visible motor response in the limbs. This lack of a motor response in the presence of a preserved autonomic response is due either to severe brain damage or to neuromuscular dysfunction (Chapter 12). Absent movements with no change in heart rate suggest brain death.

Respiratory patterns

Abnormal respiratory patterns may suggest specific anatomic localizations of coma.1 The classic patterns are often not observed while the patient is intubated, sedated, and paralyzed, but may become obvious if mechanical ventilation is discontinued temporarily. Cheyne–Stokes breathing is characterized by hyperpneic phases that build to a crescendo and then taper to apneic periods lasting for 10–20 seconds. This pattern usually implies bilateral frontal lobe pathology, and is common in metabolic encephalopathies. Hyperventilation is another pattern associated with toxic and metabolic encephalopathies, generally those that produce metabolic acidosis. Central neurogenic hyperventilation is rare, and is usually seen in the context of brainstem glioma or lymphoma.5 Apneustic breathing is characterized by 2- or 3-second pauses that occur at the end of inspiration and expiration, and reflects pontine damage. Ataxic breathing has an irregular, gasping quality, and is secondary to lower pontine or upper medullary dysfunction.

Investigation of impaired consciousness and coma

Approximately two-thirds of the causes of coma are due to medical conditions, such as metabolic abnormalities and toxins, while the remaining one-third are due to structural causes such as trauma, brain hemorrhage, and tumor.1Because the number of potential causes of coma is quite large, I find it helpful to divide the investigation into three phases based on the frequency of the responsible causes and the ease of obtaining diagnostic testing:

Phase 1: history, examination, and basic studies

By the time a neurologist is consulted, a basic metabolic workup and CT scan of the brain is usually available. Combined with a careful history and physical examination, these data help to establish one of the following coma diagnoses:

• Trauma. Patients with head trauma sufficient to cause coma almost always have abnormal head CT scans. In addition to skull fractures, abnormalities following trauma include epidural, subdural, intraparenchymal, and subarachnoid hemorrhages. Some patients have no clear evidence of fracture or hemorrhage, but the CT scan (or, more likely, an MRI) shows evidence of diffuse axonal injury.

• Intracranial mass lesion. Bilateral frontal or brainstem lesions including tumors, abscesses, and intracranial hemorrhages may all lead to coma.

• Subarachnoid hemorrhage. Aneurysmal rupture leading to subarachnoid hemorrhage is an important cause of coma that may be detected with a CT scan (Chapter 19).

• Hypoxic–ischemic injury. Whether due to anoxia following cardiac arrest or to severe hypoxia secondary to pulmonary disease, irreversible brain damage occurs after just minutes of global ischemia and is among the most serious causes of coma.

• Toxic or metabolic disturbances with normal imaging studies. Comatose patients with normal CT scans usually have a toxic or metabolic disturbance, often more than one. Routine laboratory testing is generally sensitive to the conditions listed in Table 2.1, many of which are reversible.

Phase 2: MRI, electroencephalography, and lumbar puncture

Although the history, CT scan, and basic laboratory studies often disclose the etiology of coma, further evaluation is necessary should these initial investigations fail to identify the responsible process.

MRI serves several purposes in patients with coma of unclear etiology. Diffusion-weighted MRI identifies hypoxic–ischemic changes several days before they are noted on a routine head CT. MRI may also disclose two specific infarctions that may be poorly visualized by CT. The first is infarction of the intralaminar nuclei of the thalamus and rostral midbrain6 (supplied by the paramedian thalamic arteny; Figure 2.6). The second is infarction of the base of the pons leading to the locked-in state. Other causes of coma that may be missed by CT but detected by MRI include occult encephalitis or posterior reversible encephalopathy syndrome (PRES) (Chapter 1).

EEG helps to establish the presence of severe encephalopathy or brain death in unclear cases of coma. The real value of EEG in comatose patients, however, actually lies in its ability to detect nonconvulsive

Table 2.1 Medical causes of coma

Table 2.1

Figure 2.6

Figure 2.6 Axial fluid attenuation inversion recovery (FLAIR) MRI in a patient with paramedian thalamic artery infarction. This is an uncommon cause of coma, but it is important to recognize because it is easy to overlook. The bilateral thalamic hyperintensities are quite symmetric, and it is easy to misdiagnose this finding as an artifact.

status epilepticus (NCSE), a potentially reversible condition that often eludes clinical diagnosis (Chapters 1 and 20). Continuous rather than routine EEG should be employed if NCSE is suspected, as the first electroencephalographic evidence of seizures may appear only after more than 24 hours of monitoring in 20% of patients.7

Lumbar puncture should be performed to evaluate for CNS infections, particularly bacterial meningitis and herpes encephalitis (Chapter 1).

Phase 3: uncommon etiologies and coma mimics

If the diagnosis remains unclear after an initial panel of investigations, MRI, EEG, and lumbar puncture, then consider less common toxins, neuromuscular mimics of coma, and psychogenic unresponsiveness.

Some toxins that may not be detected by routine toxicology screens are listed in Table 2.2. Consultation with a toxicologist is often helpful when considering these less common agents.

Severe neuromuscular disorders may lead to a state of profound weakness that mimics coma. Conditions such as Guillain–Barré syndrome, myasthenia gravis, and botulism are usually diagnosed before weakness reaches this severity, but on some occasions, motor function declines so precipitously that the initial phase of weakness may go unrecognized. Rapidly progressive weakness and difficulty weaning from the ventilator due to neuromuscular disease acquired in the intensive care unit are discussed further in Chapter 12.

Psychogenic unresponsiveness secondary to conversion disorder, malingering, or catatonia may be profound to the point that it mimics a comatose state. Obviously, exhaustive medical evaluation must be conducted before these possibilities are even considered. For patients with conversion disorders or malingering, cold caloric testing may clinch the diagnosis and cure the coma by inciting violent nausea and vomiting. Although patients with psychiatric disorders do not have an organic explanation for coma, they require attention and life support that is just as careful as that provided to patients with organic neurological disorders.

Prognostication in coma

The ability to accurately predict the outcome of a comatose patient is essential, as it provides families with reasonable expectations about the potential for

Table 2.2 Less common toxins that lead to coma

Table 2.2

recovery and advisability of continuing life support. Prognostication is based on the proximate cause of coma, the neurological examination, and, in some instances, diagnostic test results. In general, patients with reversible causes of coma such as hypoglycemia or uremia have better prognoses than those with severe head trauma or hypoxic–ischemic injury.

The largest body of data concerning coma prognostication comes from patients with coma after cardiac arrest. Often, the goal in this patient population is to define which group of patients has a zero chance of a good neurological outcome. If any reasonable chance of a good outcome remains, aggressive supportive care must be continued. The following examination and laboratory results predict grave prognosis after cardiac arrest, and should be applied only to patients with this specific etiology8,9

• absent brainstem reflexes

• myoclonic status epilepticus at day 1 (24 hours after cardiac arrest)

• evoked potentials showing absent N20 responses at days 1–3

• serum neuron-specific enolase >33 µg/l at days 1–3

• absent pupil or corneal response at day 3

• no motor response or extensor (decerebrate) posturing at day 3

This list suggests that, using the clinical examination alone, it is unwise to deliver an authoritative statement about coma prognosis until at least 72 hours after cardiac arrest. Unless supplemental tests are available, be cautious in providing too much information before this time. Keep several additional caveats in mind when using this list. First, confounders such as hypothermia, sedatives, and other toxins must be excluded as responsible for the comatose state. Secondly, N20 evoked potentials may be difficult to measure in the intensive care unit due to electrical noise. Thirdly, the turnaround time for neuron-specific enolase results is usually several weeks and almost always prevents it from being used in a timely fashion. Finally, the now commonplace practice of inducing mild-to-moderate hypothermia after cardiac arrest improves long-term neurological outcome, and it is not clear how this will change prognostication.10,11

The persistent vegetative state

After several days to a few weeks of deep coma, patients may appear to awaken and enter a vegetative state, which is given the name persistent vegetative state (PVS) if it lasts for at least 1 month.12 This state is characterized by roving or tracking eye movements and what appears to be an irregular sleep–wake cycle. However, these patients do not interact with their environment in a meaningful way. They may grunt or moan, but they do not speak or comprehend. A PVS is the result of bilateral cortical damage with relatively preserved diencephalic and brainstem function. Although it may seem that PVS is a more favorable state than coma, its ultimate prognosis is still quite poor: of patients with nontraumatic PVS, only 1% have a good outcome at 1 year, while for patients with traumatic PVS, only 7% will have a good outcome at 1 year.13 A PVS should be considered permanent if it persists for 3 months in patients with nontraumatic conditions and 12 months in patients with traumatic brain injuries.14 News stories of patients recovering after years of coma or PVS are exceptional and should not be used to give family members false hope.

Brain death

Brain death is defined as the complete loss of brain function despite preserved cardiac function. It is particularly important to recognize brain death in order to allow decisions about withdrawing aggressive medical support and to plan for organ procurement. Before a patient is diagnosed with brain death, all potentially reversible causes of coma must be corrected. Sedatives such as midazolam and propofol must be discontinued and the patient’s core temperature should be raised to at least 97°F. Next, the patient must be carefully examined, often using an institution-specific brain death protocol. The patient must be unarousable to any stimulus, lack pupillary and corneal reflexes, have no cold caloric responses, and not gag or cough when suctioned. Deep tendon reflexes may be (and often are) preserved. Many institutions require that the examination be confirmed several hours or a day after the initial assessment.

The apnea test is used to confirm brain death. Hypercarbia is a profound stimulus to breathe, and when it fails to produce a respiratory effort, it indicates severe brain damage incompatible with life. Before performing the apnea test, obtain a baseline arterial blood gas sample and note the partial pressure of carbon dioxide (PCO2). Next, preoxygenate the patient with 100% oxygen for at least 10 minutes. Following preoxygenation, discontinue mechanical ventilation while continuing to provide oxygen via a face mask. The apnea test is positive (i.e. consistent with brain death) if no respiratory efforts are visible after 10 minutes of ventilator discontinuation. Before reconnecting the patient to the ventilator, draw a repeat arterial blood gas sample to confirm the adequacy of hypercarbia: the PCO2 level must be at least 60 mmHg or 20 mmHg greater than the baseline level.

In some patients, difficulties with interpreting the neurological examination or minor metabolic abnormalities prevent airtight confirmation of brain death. In such cases, several supplementary diagnostic tests establish brain death by showing the absence of cerebral electrical activity or cerebral blood flow. These include:

• EEG showing electrocerebral silence

• somatosensory evoked potentials showing absent N20 responses

• absent cerebral blood flow as determined by:

• transcranial Doppler ultrasound

• cerebral angiography

• nuclear scintigraphy

Increased intracranial pressure

Expansion of the intracranial contents is limited by the rigid confines of the skull. Blood, tumor, abscess, and edema are tolerated to a limit extent before symptoms and signs of increased intracranial pressure develop. In its earliest stage, increased intracranial pressure causes nonspecific headaches and visual blurring. Recognition of increased intracranial pressure at this stage may allow the responsible process to be diagnosed and reversed, thereby preventing additional neurological deterioration. Further increases in intracranial pressure lead to encephalopathy, seizures, and a variety of focal neurological findings. The most devastating consequences of increased intracranial pressure are the herniation syndromes in which brain tissue is displaced from its normal location, compressing or damaging otherwise healthy structures. The most important of these syndromes are uncal and transtentorial herniation.

Uncal herniation

A hemispheric mass or edema may cause expansion of one cerebral hemisphere relative to the other, leading to herniation of the uncus of the temporal lobe medially and inferiorly into the tentorial notch.1 The earliest signs of uncal herniation are ipsilateral pupillary dilation produced by stretching or compression of the third nerve and a decrease in consciousness, which results from compression of the upper brainstem. Pupillary dilation in the presence of preserved consciousness, however, is essentially never due to uncal herniation.1 As uncal herniation progresses, hemiparesis develops ipsilateral to the herniating mass as the contralateral cerebral peduncle is compressed, the so-called Kernohan’s notch phenomenon. Less commonly, the ipsilateral cerebral peduncle is compressed leading to a contralateral hemiparesis. Thus, pupillary dilation is more reliable than hemiparesis in lateralizing uncal herniation. Shearing or compression of the posterior cerebral arteries in the tentorial notch may lead to cortical blindness. Because uncal herniation may progress rapidly to a state of irreversible neurological compromise or death, it must be identified as quickly as possible to allow the responsible source to be treated.

Transtentorial herniation

An expanding midline lesion may cause herniation downwards through the tentorium, compressing the thalamus and brainstem.1 In early transtentorial herniation, the patient appears to be sleepy with small, minimally reactive pupils. It is very easy to misdiagnose the patient with a metabolic encephalopathy, and a high index of suspicion must be maintained in order to make the diagnosis at this stage, as further progression leads to a poor outcome. As herniation continues, the midbrain is compressed, leading to paresis of upgaze, unresponsiveness, and decorticate posturing. Continued downward herniation compromises the pons, resulting in loss of lateral eye movements and motor unresponsiveness or decerebrate posturing. In the final stage of transtentorial herniation, medullary compression produces irregular breathing, flaccidity, and eventually death.

Management of increased intracranial pressure

The first step in managing increased intracranial pressure is to improve cerebral venous drainage by placing the head of the patient’s bed at a 45° angle. Next, intubate the patient and hyperventilate them to a PCO2 of 25 mmHg. This decrease in PCO2 produces cerebral vasoconstriction and increases the volume available for the brain parenchyma to occupy. Next, administer the osmotic diuretic mannitol (1–1.5 g/kg IV bolus and then 0.25–0.5 g/kg IV boluses q6h as needed). Measure the serum osmolality with each dose of mannitol, and discontinue the mannitol if the osmolality exceeds 315 mOsm. Consider intracranial pressure monitoring in comatose patients in whom the neurological examination cannot be monitored reliably. Other treatments including barbiturates, corticosteroids, and therapeutic hypothermia probably do not improve outcome in patients with increased intracranial pressure and should be avoided. The definitive treatment of increased intracranial pressure is removing the proximate cause, whether it is a hemorrhage, edema, tumor, or abscess. In patients in whom the relevant source cannot be addressed, consider craniotomy and temporal lobectomy to reduce rapidly increasing intracranial pressure. Because most patients who reach the stage at which this intervention is considered have a poor prognosis, decisions about neurosurgical intervention should be made very carefully.


1. Posner JBSaper CBSchiff NDPlum, FPlum and Posner’s Diagnosis of Stupor and ComaOxford: Oxford University Press; 2007.

2. Fisher CM. Some neuro-ophthalmological observations. J Neurol Neurosurg Psychiatry 1967;30:383–392.

3. Baloh RWFurman JMYee RD. Dorsal midbrain syndrome: clinical and oculographic findings. Neurology 1985;35:54–60.

4. Fisher CM. Ocular bobbing. Arch Neurol 1964;11:543–546.

5. Tarulli AWLim CBui JDSaper CBAlexander MP. Central neurogenic hyperventilation: a case report and discussion of pathophysiology. Arch Neurol 2005;62:1632–1634.

6. Barth ABogousslavsky JCaplan LR. Thalamic infarcts and hemorrhages. In: Bogousslavsky J, Caplan LR, eds. Stroke Syndromes. Cambridge: Cambridge University Press; 2001; 461–468.

7. Claassen JMayer SAKowalski RGEmerson RGHirsch LJ. Detection of electrographic seizures with continuous EEG monitoring in critically ill patients. Neurology 2004;62:1743–1748.

8. Levy DEBates DCaronna JJ, et al. Prognosis in nontraumatic coma. Ann Intern Med 1981;94:293–301.

9. Wijdicks EFHijdra AYoung GB, et al. Practice parameter: prediction of outcome in comatose survivors after cardiopulmonary resuscitation (an evidence-based review). Neurology 2006;67:203–210.

10. Bernard SAGray TWBuist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–563.

11. The Hypothermia After Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–556.

12. Jennett BPlum F. Persistent vegetative state after brain damage: a syndrome in search of a name. Lancet 1972;1:734–737.

13. The Multi-Society Task Force on PVS. Medical aspects of the persistent vegetative state – second of two parts. N Engl J Med 1994;330:1572–1579.

14. Giacino JT. The vegetative and minimally conscious states: consensus-based criteria for establishing diagnosis and prognosis. Neurorehabilitation 2004;19:293–298.