Chapter 67. Coma, Persistent Vegetative State, and Brain Death

• The neuroanatomy of coma can be divided into three major categories: diffuse brain dysfunction or bithalamic injury, primary brain stem disorders, and secondary brain stem compression from supratentorial and infratentorial mass lesions.
• Most cases of coma are due to metabolic disorders or exogenous drug intoxication.
• Patient evaluation must follow an orderly sequence, beginning with vital signs, general physical examination, and neurologic examination.
• The most important single sign distinguishing toxic-metabolic coma from primary brain disease is the presence of pupillary light responses.
• The neurologic examination of the patient in coma is brief and focuses on (1) level of consciousness, (2) pupils, (3) eye movements, (4) motor responses, and (5) respiratory pattern.
• Computed tomographic (CT) scanning of the brain is the most valuable acute test to rule out structural causes of coma.
• Hypoxic-ischemic encephalopathy after cardiopulmonary arrest or shock states may be ameliorated by aggressive measures to increase cerebral blood flow after resuscitation.
• Serial neurologic examination over the first 72 hours is most helpful to determine the prognosis for patients with atraumatic coma; for anoxic brain injury, failure to recover pupillary responses or corneal reflexes in the first 24 hours is a poor prognostic sign.
• As therapies aimed at cerebral resuscitation and preservation following acute injury are developed and proved efficacious, prior guidelines for determining prognosis will require redefinition and reconfirmation.
• The Uniform Determination of Death Act states that “an individual who has sustained either (1) irreversible cessation of circulatory and respiratory functions, or (2) irreversible cessation of all functions of the entire brain, including the brain stem, is dead."
• The determination of death by brain criteria is based on clinical examination, and in most cases does not require confirmatory tests. However, the cause of coma must be known, and the cause must be sufficient to explain irreversible cessation of whole brain function.

Consciousness is a difficult term to define, and even more complicating is the fact that many different meanings and classification systems exist for the various states of decreased level of consciousness, such as drowsiness, stupor, and coma. For practical reasons, however, in the evaluation of consciousness most clinicians give greater weight to the patient's responses and behavior than to what the patient says. Hence, consciousness can be defined in its simplest form as the patient's awareness of self and environment and the responsiveness to his or her needs and external stimulation. The level of consciousness used in clinical practice refers to the state of arousal and should be separated from the content of consciousness, which describes various forms of cognitive behaviors and thinking. An awake person is fully responsive (alert) to stimuli and is able to specify an awareness of self and environment.

Impaired consciousness is generally categorized by the level of responsiveness to external and internal stimuli (Table 67-1). Drowsiness (“lethargy") is a state of reduced physical and mental activity and a drowsy person can often not sustain wakefulness without external stimulation. It is similar in appearance to light sleep and almost always accompanied by reduced attention and concentration span and mild confusion. A stuporous patient requires repeated, stronger stimuli to arouse and quickly drifts back to persistent inactivity when the stimulus is withdrawn. When aroused, this patient may or may not open his or her eyes and partially respond to requests. At times restlessness and motor stereotypes are observed. Sopor, the Latin word for deep sleep and a term used in some European countries, denotes an intermediate state between stupor and coma. Coma (from the Greek komas, or deep sleep) is a state of unresponsiveness in which the patient is incapable of arousing to external or internal stimuli (lack of alertness). The degree of coma can vary from lighter stages (also denoted as semicoma) with observed changes in autonomic function or brief moaning to strong stimulation, to the deepest stage with absence of any brain stem responses (e.g., pupillary and corneal reflexes), cyclic autonomic activity, and motor tone. Sleep and pathologic states of consciousness undeniably share some common features; for example, the sleeping person is not aware of him- or herself and in this respect is unconscious. Of course, the important difference is that a sleeping person can be aroused to full consciousness. Furthermore, electrodiagnostic evaluation differs in the two conditions (see below) and cerebral glucose uptake does not decrease during sleep but does so in coma. For in-depth discussions of this topic the reader is referred to the major textbooks in neurology and to the seminal work by Plum and Posner.1

A vegetative state can follow coma and identifies a state in which brain stem and diencephalic (thalamic) activity is present to a degree that clinical signs of arousal are observed.2 This condition is without meaningful mental function. The patients often show blink responses to light; intermittent eye movements (sometimes erroneously interpreted as following objects or looking at family members); stimulus-sensitive automatisms such as swallowing, bruxism, and moaning, as well as primitive motor responses and cycles of sleeping and waking. However, despite the appearance of wakefulness, the patient's cerebral function remains severely impaired with total inattention and unawareness. If this state is long-lasting it is referred to as persistent vegetativestate (PVS) and is used as a descriptive clinical syndrome rather than a disease-specific entity. The most common causes include cardiac arrest, head trauma, and various causes of thalamic necrosis. Vegetative states can also be seen in the terminal phase of degenerative illnesses such as Alzheimer's disease. Ambiguous terms for PVS such as apallic syndrome and neocortical death should be avoided.3,4 Minimal conscious state can be diagnosed in patients displaying inconsistent behavioral evidence of awareness of the environment, but they cannot communicate and are unable to follow instructions reliably5 (Table 67-2).

Clinical practice teaches that consciousness should be viewed as a continuum between different pathological conditions and not as an all-or-none phenomenon,1 and that it is frequently difficult to identify definite signs of conscious perception of environment and self in patients with severe brain injuries. The latter limits the diagnostic certainty of remaining brain function on clinical grounds since the identification of consciousness relies purely on the deduction whether consciousness is present or absent in a particular patient.2

Recent data obtained in patients with PVS helps to reveal the critical brain areas involved in consciousness. Extensive loss and severe neuronal damage is causally related to loss of consciousness in patients with PVS, a finding supported by a study identifying decreased cortical radiolabeled flumazenil uptake (a benzodiazepine antagonist and neuronal marker).6 In corroboration, several studies consistently identified decreased cerebral glucose metabolism and blood flow.7,8 However, brain regions with the most consistent decrease in cerebral glucose consumption in patients in PVS are the polymodal association areas of the frontal, temporal, and parietal lobes9; interestingly, normal subjects display higher brain metabolism during wakefulness than in sleep in similar regions.10 Furthermore, thalamocortical disconnections can be identified using auditory and sensory external stimulations.11

Some carefully performed studies identified patients with PVS who activate primary and associative cortex, depending on the complexity and familiarity of the test stimuli.12,13 It is, however, unclear whether the identified activations represent consciousness since no conclusions to the connectivity of thalamocortical brain regions and larger neuronal networks can be drawn. Similarly, brain plasticity with recovery of functional thalamocortical connections and reestablishment of neuronal networks may allow certain patients to regain consciousness after severe brain injury and PVS.9 In recent years we have gained much insight into which brain areas seem necessary for conscious experience; however, future research should broaden our knowledge about what form of brain activity in these areas confers consciousness.

Because coma is a sleeplike state, it is not surprising to find that the neuroanatomy of coma is closely related to brain stem centers that regulate daily cycles of wakefulness and sleep: the reticular activating system (RAS). In animals the RAS lies within the center of the brain stem, extending from the midbrain into the hypothalamus and thalamus.14,15 Electrical stimulation of this region in a sleeping animal produces arousal, and ablation causes coma. Lesions in the pathways of the brain stem reticular formation or RAS have the greatest impact on changes in consciousness.

The RAS is a loosely organized core of polysynaptic neurons reaching in the brain stem from the lower medulla through the paramedian pons to central midbrain. At the diencephalic level it includes several functionally related nuclei in the thalami (especially the medial thalamus) and cerebral projections are prominent to the inferomedial frontal lobes, but also reach almost all cerebral cortex. The essential role of the RAS is arousal and maintenance of wakefulness; injury leads to reduction or failure of arousal. As the brain stem RAS receives direct spinothalamic information, incoming sensory stimulations are not only projected to the sensory cortices, but are also needed to activate the brain stem RAS for the maintenance of consciousness. Within the brain stem and thalamus the RAS is confined to rather small anatomic areas; therefore even small lesions can severely impair arousal and consciousness. In contrast, RAS fibers are sparse and spread out as they move towards the cerebral hemispheres, hence only larger cortical lesions will lead to impaired consciousness at that level.

Lesions in the medulla and lower pons need to be quite large to induce significant coma (loss of awareness) since the RAS is rather thinned in these regions. More commonly, lesions in the ventral pons (such as basilar artery occlusions) lead to severe motor pathway injury sparing somatosensory and ascending RAS (arousal) systems. This state is referred to as locked-in syndrome or de-efferented state, as the patient has preserved consciousness. However, the patient cannot speak or respond and is unable to move cranial, trunk, or extremity muscles, but retains the ability for vertical gaze and eye blinking. In contrast, vascular occlusions of the top of the basilar artery or the posterior cerebral arteries result in injuries to the midbrain or medial thalami which lead to significant impairment of consciousness but no or only minimal focal neurologic findings.

Traumatic brain shearing injury can lead to slit-like hemorrhages and ischemia or inflammatory to necrotizing lesions of the midbrain, resulting in impaired consciousness. Cerebral masses can produce either direct or indirect (uncal or tentorial herniation) displacement and torsions of the midbrain and reduced alertness. Akinetic mutism refers to patients who are silent and apathetic with greatly reduced activity and no focal findings on examination. A number of pathologic lesions have been described, often associated with inferior frontal lobe damage bilaterally. Sudden injury to either or both cerebral hemispheres may produce impaired consciousness, indicating that wakefulness has no hemispheric dominance and requires some cerebral function. Bilateral, extensive acute or subacute damage to the cortex and white matter, for example due to trauma, hypoxia, or infection, impairs activation of the upper RAS, leading to impaired consciousness. Notably, even large cortical (lobar) areas can be injured and initially not affect consciousness at all until secondary injury from swelling and bleeding occurs, as evidenced by patients with penetrating wounds of the cerebral hemispheres who remain fully awake. Similarly, degenerative disorders such as Alzheimer's disease generally do not or only minimally affect the RAS since these patients remain fully awake. Hypersomnia refers to a condition of excessive drowsiness and sleep. It may occur in the setting of narcolepsy, hypothalamic disorders, sleep disorders, or psychiatric illness.

Several categories of consciousness-impairing mechanisms and lesions can be defined. First, one cause is an easily identifiable mass lesion compressing the upper RAS either directly or indirectly (such as tumor, abscess, meningitis, or hemorrhage); second, discrete lesions of the upper brain stem (examples are outlined above) may be the cause; and third, a larger group of patients includes those in whom suppression of the RAS is induced by metabolic derangements, toxic states, or seizures. Such functional causes of coma may be reversible by correcting the underlying metabolic derangement or removing the offending drug. “Metabolic” coma is likely the most common etiologic category resulting in impaired consciousness in medical critical care units.

The neuroanatomy of consciousness is not linked to any specific motor, sensory, or cognitive function. Coma and specific focal signs of neurologic damage may occur together or independently. Therefore, patients in coma can be divided into two major clinical categories: coma with or coma without focal neurologic signs. This simple division is often the best way to identify primary neurologic disease as the cause for coma.

Acute depression in level of consciousness is a critical, life-threatening emergency that requires a systematic approach for evaluation of etiology. The variety of causes of coma are myriad. Therefore, a reliable history should be obtained from family, witnesses, or medical personnel, and examination should seek representative localizing neurological and general physical findings.

Historical Features

Clues can be ascertained from the onset of coma. An acute onset in a previously healthy individual may indicate a cerebral vascular etiology (i.e., subarachnoid hemorrhage, intracerebral hemorrhage, or hemispheric or brain stem stroke), generalized epileptic activity, traumatic brain injury, or drug overdose. Likewise, a subacute deterioration may point to systemic illness, evolving intracranial mass, or a degenerative infectious or paraneoplastic neurologic disorder. Moreover, the duration of a comatose state should be documented because it may have predictive value for prognosis in certain causes.

General Clinical Features and Protocol

Frequently the etiology of acute depression in consciousness in the hospitalized patient includes sepsis, acid-base and electrolyte disorders, or hepatic, renal, or cardiac failure. Therefore careful physical examination is performed with attention to vital signs, spontaneous breathing pattern, and careful auscultation of the lungs and heart (Table 67-3). Airway patency and protective reflexes should be assessed. Emergency measures should be taken to ensure vital functions continue despite an obscure diagnosis. Furthermore, laboratory studies should be obtained to exclude metabolic and endocrine causes.

During evaluation, core body temperature is an important clue. Hypothermia can be seen in drug overdose, brain death, or acute spinal cord transection. Moreover, hyperthermia can be seen in infection; traumatic brain injury; subarachnoid, intracerebral, or pontine hemorrhage; and hypothalamic dysfunction.

Neurologic Examination

The neurologic exam in a patient with depressed level of consciousness can be a valuable tool to localize the etiology. The important neurologic features include (Table 67-4): (1) respiratory pattern, (2) pupillary size and reactivity, (3) eye position and movements, (4) corneal reflexes, and (5) motor function.

Assessment of Consciousness

The determination of the level of consciousness depends on analyzing arousability and content (see Table 67-1). Initially, observe whether the patient appears asleep or wakeful with spontaneous eye opening. In a sleeping patient, quantify how much stimulation is required to arouse the patient. Attempts should be made to elicit a behavioral response by verbal command alone. If no response is obtained, then physical stimulation should be used, first by shaking the patient. Then noxious stimulation can be applied by digital pressure to the supraorbital nerves or nailbeds of the fingers or toes. Purposeful attempts by the patient to remove the offending stimulus indicate preservation of brain stem function and intact connections to the appropriate cerebral hemisphere. Eye opening, either spontaneous or in response to stimulation, indicates preserved function of the RAS in the upper brain stem and hypothalamus. Once aroused, the patient's ability to remain wakeful and respond coherently is determined.

Lethargy (or drowsiness), stupor, and coma represent different points on a continuum of decreasing levels of consciousness. Patients in these states appear to be sleeping with eyes closed. In contrast, patients with akinetic mutism and locked-in syndrome appear to be awake with eyes opened.

The Glasgow Coma Scale (Table 67-5) is used to assign a numerical description of consciousness. The scale was devised to evaluate patients with head injury and is most reliable and reproducible in trauma patients.16,17 Its application in nontraumatic conditions is less reliable, but it is still the most widely used clinical scale to evaluate the level of consciousness. Furthermore, it provides a reproducible tool to monitor progression.

Respiratory Control

The cerebral cortex and forebrain are important in the control of regular respiration. Patients with isolated brain injury uncomplicated by other critical medical illnesses may have characteristic breathing patterns that aid in neuroanatomic localization (Fig. 67-1). However, these patterns are not reliable in patients with multiple organ system failure who are receiving mechanical ventilation. Nevertheless, a discussion is warranted.

Cheyne-Stokes respiration is a periodic breathing pattern in which periods of hyperpnea regularly alternate with apnea in a smooth crescendo-decrescendo pattern. This neurogenic respiratory alteration occurs with damage to the cortex and forebrain bilaterally, or secondary to cardiac or respiratory failure. It is the result of the loss of frontal lobe control over respiratory patterns with excessive dependence on blood CO2 levels to trigger brain stem respiratory centers.

Midbrain and upper pontine lesions may cause a central neurogenic hyperventilation syndrome with persistent deep hyperventilation. It can only be diagnosed with arterial blood gas measurements, since hyperventilation also occurs secondary to hypoxemia and acidemia. Likewise, metabolic disorders, especially the early stages of hepatic coma, cause central neurogenic hyperventilation.

Lesions of the middle or lower pons are characterized by deep prolonged inspiration followed by a long pause referred to as apneustic breathing. Most patients with this respiratory pattern require early intubation and mechanical ventilation.

Ataxic and irregular periodic breathing occurs with lesions in the dorsomedial medulla and may be accompanied by hypersensitivity to respiratory depressants. These patterns are not compatible with sustained life.

When assessing a comatose patient, the rate and pattern of respiration should be observed. In addition, vomiting and hiccups should be noted because they may result from intrinsic brain stem pathology or transmitted pressure on the brain stem. Furthermore, spontaneous yawning may occur in comatose patients. The neurogenic networks for this complex respiratory response are integrated in the lower brain stem.

Pupillary Size and Reactivity

In one study of 346 comatose patients, the pupillary reflex was shown to be the strongest prognostic variable for awakening when compared with evoked-potential studies.18 Pupillary size is controlled by the autonomic nervous system and is dictated by the balance between sympathetic and parasympathetic input to the pupillary dilators and constrictors, respectively. The parasympathetic efferents to the pupil originate from the Edinger-Westphal nucleus in the upper midbrain and travel with the ipsilateral third cranial nerve (oculomotor). Dysfunction within this pathway will produce unopposed sympathetic input to the pupil and relative pupillary dilation ipsilateral to the lesion. The sympathetic efferents to the pupil originate in the hypothalamus, descend through the brain stem and cervical spinal cord, and exit the upper thoracic spinal cord (T1 to T3 levels). From this point they ascend the carotid sheath and follow the vasculature to the pupil. Any disruption of the sympathetic fibers along this loop can lead to unopposed parasympathetic pupillary activity and subsequently an ipsilateral small (miotic) pupil.

A light stimulus to one eye produces constriction of the ipsilateral pupil (direct response) and contralateral pupil (consensual response), through a network of connections. Table 67-6 summarizes the pupillary changes commonly seen in coma and their significance.

Small reactive pupils may be due to a toxic-metabolic disturbance. Very small pupils (pinpoint) that react to naloxone are characteristic of narcotic overdose. Pinpoint pupils that are poorly reactive are characteristic of pontine dysfunction. Lesions rostral or caudal to the midbrain may disrupt descending sympathetics and produce small pupils.

Bilateral, widely dilated, fixed pupils are due to sympathetic overactivity from an endogenous cause (seizures or severe anoxic ischemia) or exogenous catecholamines (dopamine or norepinephrine) or atropine-like drugs.

Since the midbrain is the one location in the brain stem where parasympathetic and sympathetic pupillary fibers are adjacent, a midbrain lesion classically results in intermediate pupil size. Such pupils are seen in brain death and severe midbrain injuries.

A unilaterally dilated, unreactive pupil in a comatose patient may be caused by herniation of the ipsilateral temporal uncus through the tentorium, which compresses the ipsilateral oculomotor nerve and its parasympathetic fibers. In this setting, the large pupil is eventually accompanied by other evidence of cranial nerve (CN) III disruption (i.e., ipsilateral eye deviation inferolaterally). In the setting of head trauma, this implies an ipsilateral epidural, subdural, or intracerebral hematoma. In nontraumatic conditions, it usually occurs with large cerebral infarcts, spontaneous intracerebral hematoma, or supratentorial brain tumors.

Eye Position and Movement

The eye muscles are controlled by three sets of cranial nerves, CN III, CN IV (trochlear), and CN VI (abducens), their nuclei being located in the upper midbrain, lower midbrain, and pontomedullary junction, respectively. Proper eye movement control requires a network of interconnections between these nuclei so that the eyes move conjugately. This interconnection is referred to as the medial longitudinal fasciculus (MLF), which is also integrated with the vestibular nuclei and allows for reflex conjugate eye movement in response to positional head changes.

Figure 67-2 displays the relevant anatomy accounting for horizontal conjugate eye movements. Each frontal eye field controls gaze to the contralateral side by stimulating the contralateral pontine paramedian reticular formation (PPRF) at the pontomedullary junction. Lesions of the frontal eye fields or the PPRFs lead to conjugate eye deviation, provided that the MLF is intact. Therefore, a lesion of the right frontal eye field or left PPRF impairs leftward gaze, and thus the eyes conjugately deviate to the right. In short, the eyes turn toward the lesion with frontal eye field dysfunction and away from the lesion with PPRF dysfunction. In contrast, MLF lesions are manifested as poor adduction of the eye ipsilateral to the MLF lesion. Spontaneous “roving” eye movements in all directions in the comatose patient demonstrate integrity of a significant portion of the brain stem. If no spontaneous eye movements are observed, the intactness of the interconnections responsible for eye control is in question. Since comatose patients are unable to follow commands, maneuvers that take advantage of vestibular input to ocular control must be utilized.

An oculocephalic reflex (doll's eye maneuver) is performed by rapidly rotating the head from side to side and observing the patient's eye positional changes (Fig. 67-3). The normal response in the comatose patient with intact brain stem is for the eyes to remain fixed on the same point in space. Thus when the head is turned rightward, the eyes move to the left. When the head is turned leftward, the eyes move conjugately to the right. If a comatose patient does not have normal doll's eyes, a disruption of brain stem ocular and vestibular connections may be present. Of course, in the setting of trauma, the head should not be rotated due to the possibility of cervical spine injury. In this situation or when doll's eye maneuvers are inconclusive, cold water calorics are helpful.

Oculovestibular reflexes (cold water caloric testing) depend on vestibular system stimulation by altering endolymphatic flow in the semicircular canals. The change in endolymphatic flow is achieved by instilling ice-cold water in the external auditory canal, thereby cooling the mastoid process, and in turn the semicircular canal. Prior to performing this test, the external auditory canal should be examined to confirm intactness of the tympanic membrane and remove any impacted cerumen. The head should then be elevated 30 degrees. A functional apparatus for instilling the water is a butterfly catheter (with the needle removed) connected to a syringe containing approximately 100 mL of cold water.

The responses to cold water in patients with various lesions are summarized in Fig. 67-3. In normal wakeful patients the response is horizontal nystagmus, with the slow phase toward and the fast phase away from the stimulated side and with minimal eye movement from the midline. With diminishing consciousness in patients without structural brain stem damage, the fast phase of the nystagmus disappears, and the eyes tend to deviate conjugately toward the stimulated side. Structural brain stem disease eliminates the caloric response, as does inner ear disease, deep drug coma, and anticonvulsant drug overdose. In order to ensure proper interpretation of cold water caloric testing, the opposite side should not be stimulated until 5 minutes after the initial side.

Corneal Reflex

The corneal reflex is an important protective mechanism for the cornea. it is a blinking reflex triggered when the cornea is presented with a noxious stimulus. The afferent limb is via the trigeminal nerve (CN V), and the efferent limb is via the facial nerve (CN VII). Although corneal reflexes assess brain stem function, they have limited localizing value.

Motor Function

The corticospinal tract predominantly originates from the frontal cortex and descends ipsilaterally through the corona radiata, the posterior limb of the internal capsule, and the cerebral peduncle of the midbrain and consolidates in the pyramids, the ventral swellings of the medulla. The pyramidal fibers decussate to the contralateral side at the junction of the medulla and spinal cord to form the lateral corticospinal tract.

Observation is the key to the motor examination in the comatose patient. The patient is observed for spontaneous movements or maintenance of particular postures. Lesions involving the corticospinal tract generally lead to diminished contralateral spontaneous activity. Upper midbrain or more rostral lesions may lead to decorticate posturing characterized by flexion of the contralateral arm at the elbow and hyperextension of the leg. Central midbrain and high pontine lesions, with a relatively intact brain stem inferiorly, may lead to decerebrate posturing characterized by contralateral arm and leg extension. Such posturing may also be caused by structural lesions or metabolic insults and it is often mistaken for seizure activity. The patient should be observed for the presence of tremor, myoclonus, or asterixis, because these may be associated with toxic-metabolic encephalopathies.

After observing for spontaneous movements and posturing, motor tone should be assessed by passive flexion and extension of the extremities. Tone may be increased or decreased, depending on the location of the motor system involvement. Afterward, noxious stimuli should then be applied to each limb and the supraorbital regions. Purposeful movement upon noxious stimulation suggests intactness of motor tracts to that limb, whereas decorticate or decerebrate posturing in response to noxious stimuli has the localizing significance mentioned above.

Acute corticospinal tract lesions may cause hyporeflexia because hyperreflexia may not occur for days to weeks after the injury. However, a Babinski sign, which is characterized by extension of the great toe and fanning of the other toes upon lateral plantar stimulation, may be present acutely with corticospinal tract lesions. Complete bilateral paralysis without any response to noxious stimuli usually indicates a grave prognosis. However, spinal cord injury and neuromuscular transmission blockade must be excluded because they may produce a similar state of complete paralysis.

Alcoholism, cerebral trauma, and cerebrovascular diseases account for a majority of comatose patients. Other major causes for admission include epilepsy, drug intoxication, diabetes, and severe infection. In the university hospital setting Plum and Posner1 found one-quarter of comatose patients to have cerebrovascular disease, 6% were the consequence of trauma, all “mass lesions” (i.e., tumors, abscesses, hemorrhages, and infarcts) accounted for less than one-third, while subarachnoid hemorrhage, meningitis, and encephalitis accounted for another 5%. The majority of cases were the consequence of exogenous (drug overdose) and endogenous (metabolic) intoxications and hypoxia.

When the examination leads to the conclusion that the central neuraxis is affected at multiple levels or diffusely, the etiology is most likely of toxic-metabolic origin. Metabolic causes and toxic ingestions account for the largest number of patients with depressed consciousness (Table 67-7). The single most important sign distinguishing metabolic from structural coma is the presence of the pupillary light response.19,20 In metabolic coma, confusion and stupor commonly precede symmetrical motor signs. Asterixis, myoclonus, tremor, and seizures are common. Central hyperventilation occurs frequently.

Supratentorial mass lesions causing compression or displacement of the upper brain stem (Table 67-8) usually present with focal neurologic signs that are asymmetrical. Neurologic dysfunction usually progresses in a rostral-caudal fashion, and the examination usually points to one anatomic area of the brain stem at a given point in time.

Subtentorial masses or destructive lesions causing coma usually are associated with brain stem dysfunction or sudden onset of coma. Brain stem signs always precede or accompany the onset of coma and always include abnormalities of eye movements. Cranial nerve palsies are usually present. Irregular respiratory patterns are common and usually appear at the onset of coma.

Selected Causes of Coma

Metabolic derangements, cerebrovascular diseases, head trauma, and drug intoxications are reviewed in Chaps. 62, 63, 93, and 102.

Hypoxic-ischemic encephalopathy is the most common and most devastating cause of coma in most critical care units. The term is a pathologic diagnosis that refers to the effects of various degrees of global brain ischemia, sometimes complicated by hypoxemia. Cardiac arrest is the most common cause of global brain ischemia.21–24 Each year there are 200,000 attempts at cardiac resuscitation, and only 70,000 are successful. Of the early survivors, 30% leave the hospital, but only 10% resume their former lifestyle because of permanent neurologic deficits. Other important causes of global brain ischemia include severe hypotension, cardiac failure, strangulation, cardiopulmonary bypass, status epilepticus, diffuse cerebral arteriosclerosis, increased intracranial pressure, cerebral arterial spasm, closed head trauma, and hyperviscosity.

The degree of neuronal injury in this condition depends on the degree of mismatch between metabolic demand and delivery of substrate (oxygen and glucose) to the brain.25 For example, the brain can tolerate 45 minutes of total circulatory arrest with complete recovery if hypothermia to 18°C is induced. Conversely, a brain with metabolic activity that is eightfold above normal during status epilepticus will sustain neuronal damage after 2 hours, even with perfect maintenance of arterial oxygenation, glucose, and blood pressure.

The neurologic syndromes that follow cardiac arrest and resuscitation are diverse and depend on the duration of ischemia (time from arrest to cardiopulmonary resuscitation [CPR]), adequacy of CPR, underlying cardiovascular disease, degree of arterial atherosclerosis, adequacy of postresuscitation cerebral perfusion, and whether the patient is cooled following arrest (see Chap. 16). Patients in coma less than 12 hours after resuscitation usually make an excellent recovery. Those in coma more than 12 hours have permanent neurologic deficits due to focal or multifocal infarcts of the cerebral cortex in arterial border zones. They may be left with permanent amnesia, dementia, bibrachial or quadriparesis, cortical blindness, seizures, myoclonus, and ataxia.26,27 If the coma persists for 1 week, recovery is rare, and most patients remain in a PVS due to laminar necrosis of the cerebral cortex with preservation of brain stem function.28,29

Early in the course of hypoxic-ischemic encephalopathy, specific neurologic signs can predict outcome with a high degree of reliability. In a study of 210 patients, the absence of pupillary responses on the first day after CPR predicted poor outcome.30 None of 52 patients recovered, and only 3 regained consciousness. No patient who lacked corneal reflexes after the first day ever regained consciousness. After 3 days, a lack of purposeful motor responses predicted poor outcome in all patients (PVS or severe disability). Certain early signs were associated with good recovery. At 1 day, the following signs were associated with at least a 50% chance of regaining independent function: verbal responses of any type, purposeful eye movements or motor responses, normal ocular reflexes, and response to verbal commands.

The cause of hypoxic-ischemic encephalopathy is not a simple response to circulatory arrest. Evidence has accumulated that the brain can tolerate a longer period of ischemia than previously thought if certain conditions are met. After a 10-minute period of global brain ischemia, if the circulation is adequately restored, there is marked hyperemia, subsequently followed by a delayed, progressive fall in cerebral blood flow to levels considerably below prearrest values.31 In some cases, cerebral metabolism remains high in the face of low blood flow, a situation that worsens the effects of the initial ischemia. There is experimental evidence suggesting that if cerebral blood flow can be maintained at a hyperemic level, the brain can recover.32 It appears that some of the damage is due to a lack of adequate reperfusion after resuscitation, the no-reflow phenomenon. The mechanisms of delayed hypoperfusion and no-reflow are poorly understood but may be due to diffuse arterial spasm, calcium influx, vasoconstrictor prostaglandins, and intravascular coagulation. Even when adenosine triphosphate levels recover promptly as evidence of successful reperfusion, neurons show progressive morphologic damage, suggesting that injury may be in large part due to reperfusion rather than ischemia, per se. Persistently impaired protein synthesis following reperfusion may prevent the cell from repairing damage. In addition, apoptotic pathways appear to be activated, triggering neuronal death. In addition, other biochemical changes are initiated after ischemia that can cause delayed neuronal injury: elevation of intracellular calcium, release of neurotoxic excitatory amino acids (glutamate and aspartate), reoxygenation injury from superoxide formation, and brain lactic acidosis.33,34

The pathophysiology of altered cerebral blood flow and metabolism following CPR is complex and points to several windows of opportunity where the devastating effects of global brain ischemia may be ameliorated. Calcium channel blockers such as nimodipine, nicardipine, and lidoflazine are being investigated, experimentally and clinically, to determine if they can prevent neuronal damage following global brain ischemia and improve cerebral perfusion after resuscitation. Excitatory amino acid antagonists (MK-801) may be effective at preventing delayed neuronal injury after ischemia.31 Free radical scavengers and excitatory amino acid antagonists have also been considered as potentially useful to improve outcome after brain injury; to date, despite some promise in animal models of central nervous system (CNS) injury, these agents have not shown benefits in clinical trials. The mechanism of benefit from induced hypothermia following anoxic brain injury is unknown, and this topic is further considered in Chap. 16.

Diagnostic Procedures in Evaluating the Comatose Patient

Keeping in mind the preceding differential diagnosis for states of coma, the sequence of diagnostic studies becomes clear. Rapid identification of metabolic or toxic causes of coma is determined by laboratory testing of blood, urine, gastric aspirate, and cerebrospinal fluid (CSF) (Table 67-9).

A variety of diagnostic tests are useful in evaluating comatose patients. Computed tomographic (CT) scanning of the head is useful in detecting changes in brain density. It is utilized mainly to detect focal brain disease, being most sensitive for diagnosing acute hemorrhage, which appears as an area of increased density. Conversely, infarction shows up as an area of lucency, subtly apparent at least several hours after onset. Neoplasms and abscesses are lucent on CT but often accumulate IV contrast material due to alterations in the blood–brain barrier. CT is the test of choice in acute trauma because it provides detailed images of bony structures of the skull base.35 It may also show parenchymal shifts and effacements of CSF spaces, suggesting the presence of increased intracranial pressure or the presence of hydrocephalus. In addition, CT can be performed easily and rapidly in critically ill, intubated, and artificially ventilated patients.

Magnetic resonance imaging (MRI) is the most sensitive method to image the brain and define diverse pathologies. The image resolution is much better than CT, and it allows images of the central nervous system in multiple planes. MRI is particularly helpful for imaging posterior fossa structures, which often are poorly visualized on CT due to bony artifact. MRI frequently displays pathologic processes earlier than CT does, which may be critical for prompt initiation of appropriate therapy, as in herpes encephalitis.

Lumbar puncture and CSF examination are essential in the diagnosis of meningitis and encephalitis. On occasion, CSF analysis is more sensitive than CT in documenting subarachnoid hemorrhage. The major contraindication to performing a spinal tap is cerebral edema. Since processes associated with cerebral edema represent several of the etiologic considerations in the comatose patient, CT should generally be performed prior to the spinal tap in comatose patients. If it is essential to obtain CSF in states associated with intracranial masses and intracranial hypertension, ventriculostomy may be performed safely and provides an excellent tool for subsequent monitoring and treatment of intracranial hypertension.

Electroencephalography (EEG) measures brain wave activity. It is useful in detecting focal cerebral dysfunction, seizures, encephalitis, and diffuse metabolic encephalopathy, although it is nonspecific with regard to cause (i.e., uremia vs. hepatic failure).36,37 A rare patient in coma from nonconvulsive status epilepticus will be identified by EEG.38 It can be used as objective verification of brain death, but this application is less popular in contemporary practice.39

Evoked-potentialstudies (visual, auditory, and somatosensory) have been used extensively for assessment of brain function in comatose patients with acute brain injury, with the hope that they would provide information about prognosis.37,39 In general, they are no more useful than the clinical examination, except for somatosensory evoked potential (SSEP). The absence of cortical waves in SSEPs performed early after onset of hypoxic-ischemic coma has a high specificity for predicting the likelihood of nonawakening. SSEPs are resistant to sedative drug intoxication and may be present in drug overdose states that cause an “isoelectric” EEG. The persistence of cortical SSEPs in comatose patients with head trauma predicts potential recovery in about one-third of patients; the absence of SSEPs predicts poor outcome in the vast majority.37,40 Recently one study showed promising potential with more specific tests looking at the role of late auditory (N100), cognitive evoked potentials (mismatch negativity; MMN), and middle latency auditory evoked potentials (MLAEPs) for prognosis of awakening in a cohort of 346 comatose patients.18 The strongest prognostic variable was pupillary reflex (estimated probability 79.7%), and the estimated probability rose to 87% when N100 was present, and to 89.9% when MLAEPs were present. Interestingly, when MMN was present, 88.6% of patients awakened and no patient with MMN became permanently vegetative.

In a recent small study of 34 patients with anoxic coma admitted over a 2-year period to an intensive care unit, the predictive value of combined clinical exam, SSEP, and EEG was evaluated.41 On day 3 or thereafter, patients with extensor motor response to pain or worse and a “malignant” EEG (low amplitude, less than 50 μV, delta rhythm, nonreactive; or burst-suppression; or suppression <20 μV, alpha/theta coma, nonreactive; or epileptiform discharges with burst-suppression), or those patients with flexor posturing or worse and bilaterally absent SSEPs invariably had poor outcome. However, some patients with initially malignant EEG and normal SSEPs may recover and should be supported until their prognosis becomes more definitive.

Treatment must be instituted immediately, even when the diagnosis is uncertain, to prevent further brain damage secondary to complications. Oxygenation must be ensured, and airway protection is essential. All patients in coma should have a cuffed endotracheal tube placed quickly; the need for mechanical ventilation is determined by the degree of spontaneous breathing and the need for therapeutic hyperventilation. Trauma patients with suspected cervical spine injuries may need emergency tracheostomy to avoid extension of the neck during endotracheal tube placement (see Chap. 35). Subsequent ventilator adjustments are determined by arterial blood gases.

Arterial blood pressure must be maintained. Hypotension causes secondary brain ischemia. Intravascular volume replacement with blood or isotonic solutions often requires hemodynamic monitoring. The goals are to attain normal intravascular volume with normal central venous pressure, mean arterial pressure, and blood osmolality. Excessive volume replacement can aggravate intracranial hypertension, especially if hypotonic solutions are administered and serum osmolality falls. Inadequate intravascular volume will cause a fall in cardiac output and worsen brain ischemia. Inotropic and vasopressor drugs are administered as needed.

As part of the initial management of all patients with coma, glucose should be given (50 mL of 50% glucose) as soon as blood is sent to the laboratory. Although there is a theoretical risk of hyperglycemia causing brain lactic acidosis during ischemia, the risk of damage from hypoglycemic coma is much greater and requires emergency treatment. Thiamine (100 mg IM or IV) is administered with the glucose to avoid Wernicke-Korsakoff syndrome. Seizures, regardless of cause, must be stopped (see Chap. 64). Intracranial hypertension, if present, should be treated aggressively using intracranial pressure monitoring as a guideline for treatment.35,42 Details of intracranial pressure monitoring and treatment are described in Chap. 65.

Systemic infections, especially gram-negative sepsis, can cause stupor or coma on a toxic basis and must be promptly treated. Severe acid-base disorders, while rarely responsible by themselves for coma, can worsen the overall situation by causing secondary cardiovascular and respiratory failure. Rapid correction of severe acidosis or alkalosis may be helpful.

Hyperthermia can accompany a variety of pathologic states, either infectious or secondary to hypothalamic or brain stem damage. A 1°C elevation in body temperature will increase cerebral tissue metabolic demand by 10%. Therefore, hyperthermia can itself exacerbate the harmful effects of ischemia, hypoxia, or hypoglycemia. Hyperthermia should be aggressively treated. Mild therapeutic hypothermia was first proposed as a treatment for brain trauma in the late 1950s. Hypothermia can lower ICP, alter chemical pathways that could contribute to injury, and modulate apoptosis, but the mechanism of benefit seen in animal models remains unknown. Large trials in patients with traumatic brain injury have been conflicting43,44 while benefit seems established following out-of-hospital cardiac arrest.45,46

Specific antidotes may be effective in coma secondary to drug intoxication. Details of treatment for drug overdose are described in Chapter 102. Other than general measures, as outlined above, treatment depends on accurate diagnosis of cause and specific directed treatment.

While it is possible to describe broad prognostic likelihoods for patients with nontraumatic coma, particularly postanoxic brain injury (see above), there are no precise tools to determine which patients with coma will evolve to vegetative state. Generally, the clinical course and long-term outcome of vegetative state depends in part on the etiology of the brain injury. A useful classification in this regard describes three broad categories: (1) acute traumatic and nontraumatic brain injuries, (2) degenerative and metabolic brain disorders, and (3) severe congenital malformations of the CNS. Recovery of consciousness from a posttraumatic PVS is unlikely after 12 months for both children and adults. For nontraumatic PVS, recovery after 3 months is exceedingly rare. Patients with degenerative disorders or congenital abnormalities are very unlikely to recover consciousness after several months of PVS. For all patients remaining in PVS, life span is substantially reduced and generally ranges from 2 to 5 years. Survival beyond 10 years is extremely unusual. Life expectancy is largely influenced by the complications accompanying chronic care of these individuals.

Patients who recover consciousness after severe head trauma are often left with marked disabilities. The overall main costs of disability from brain injury are loss of employment. Only about half of all patients with full time employment prior to the severe head injury event were able to return to full time employment,47 most of them at lower occupational status. In a group of severely brain-injured young men, however, the re-employment rate can be significantly improved by using on-site training accompanied by a job coach, as well as continuing long-term support with use of behavioral and cognitive training strategies.48,49

In recent years, the success rate of organ transplantation has increased dramatically, and transplantation has become standard therapy for patients with end-stage kidney, heart, and liver disease. Unfortunately, thousands of critically ill patients will not receive needed organs due to a lack of understanding of the concepts and criteria for the declaration of death. We are wasting a rare and precious resource because health professionals, as well as the public, have been misinformed about definitions and procedures necessary to declare death in the setting of massive, irreversible brain damage50,51 (see also Chap. 91).

Although specific protocols for the determination of death vary slightly from institution to institution, guidelines in the United States were firmly established by the President's Commission in 1981 and have been uniformly accepted by the American Medical Association, American Academy of Neurology, and American Bar Association.52–54

The Uniform Determination of Death Act states that “an individual who has sustained either (1) irreversible cessation of circulatory and respiratory functions, or (2) irreversible cessation of all functions of the entire brain, including the brain stem, is dead. A determination of death must be made in accordance with accepted medical standards.” In 1987, the state of New York adopted the preceding statement as the legal definition of death. Since that time, this definition has been widely adopted throughout North America and Europe.55,56

However, confusion still surrounds (1) the definition of death and (2) the criteria for determining death in a patient who has stable cardiovascular function but irreversible cessation of brain function. Part of this confusion arises from the continued use of the ambiguous term “brain death,” implying that there is more than one type of death. If health professionals are confused about this concept, it is no surprise that the public remains uncertain about the terminology.

The medical and legal definitions of death are clear: brain death and cardiac death are the same. Dissenting opinions from the strict religious orthodoxy of Roman Catholicism and Orthodox Judaism persist, however, due to individual interpretations and applications of religious beliefs and laws. It is unlikely that there will ever be a unanimously accepted definition of death, due to diverse religious and ethical opinions. However, time has brought about increasing acceptance that the irreversible cessation of brain function is a mechanism of death.

The core of the clinical diagnosis of brain death is to establish unresponsiveness and brain stem areflexia. The preconditions are that (1) the cause of coma is known and (2) the cause is adequate to explain the irreversible cessation of whole brain function.57,58 In almost all cases, cerebral circulatory arrest from intracranial hypertension is the mechanism. Confirmatory tests are appropriate in selected cases, particularly for posterior fossa processes that can cause devastating brainstem dysfunction without hemispheric injury. Careful attention to these preconditions will alert the intensivist to special circumstances that require re-evaluation and confirmatory tests. In the press for organ procurement, the cautious physician may be perceived as delaying the transplant process to ensure that the diagnosis of death is unequivocal.

The President's Commission's recommendations grew out of studies at Harvard Medical School in the 1960s and extensive collaborative studies sponsored by the National Institute of Neurological Disorders and Stroke in the years 1971 and 1972.59,60 These clinical studies were performed before the availability of CT imaging, making the cause of coma and anatomic extent of brain damage uncertain in many cases. Therefore, the use of routine EEG became a standard part of many determination-of-death protocols. With the ability to directly visualize the extent of brain damage in comatose patients by CT, the determination of irreversibility became more precise. To this day, however, there continues to be unnecessary reliance on confirmatory laboratory tests rather than the use of objective clinical criteria. In contemporary practice, the use of transcranial Doppler to verify cerebral circulatory arrest is more relevant than EEG.

In most states in the United States and many countries around the world, clinical criteria are sufficient to diagnose brain death. Confirmatory tests are reserved for situations in which there is uncertainty regarding cause and reversibility of coma. However, there is no doubt that the diagnosis of brain death requires diagnostic respect from the physician and its proper pronouncement is best verified by physicians experienced with the diagnosis.

Clinical Diagnosis

It is important to keep a structured approach in mind when approaching a patient with the possible diagnosis of brain death. First, imperative prerequisites that should be fulfilled prior to determination of brain death include (1) irreversible brain catastrophe which involves both the cerebral hemispheres and the brain stem; (2) core temperature >32°C; (3) no evidence of intoxication, poisoning, or the use of paralytics, anesthetics, or sedatives; and (4) no confounding medical conditions such as severe endocrine, electrolyte, and acid-base disturbances. CT scan may or may not show abnormalities consistent with brain death. For example, following cardiopulmonary arrest the CT may show abnormalities only visible to the expert and may be interpreted as normal by those who are less experienced. On the other hand, a significant and easily identifiable mass lesion seen on CT scan does not necessary imply brain death. As a rule, if the brain CT is discrepant with the clinical diagnosis of brain death, a repeat study is warranted; if the repeat CT scan remains discrepant, the search for other confounding factors should be intensified.

The examination of a potentially brain dead patient should be methodologic as well as documented in detail: Identify the lack of consciousness (coma), verify the patient's core temperature and ventilator dependency, test and describe that no movements and brain stem reflexes are present (including absence of pharyngeal and tracheal reflexes), and demonstrate the lack of spontaneous respiration to increasing arterial carbon dioxide levels (apnea test). The apnea test demonstrates the lack of ventilatory drive while the patient remains oxygenated; a patient fails the apnea test when an increase of arterial CO2 to 60 mm Hg (from a baseline range of 35 to 45 mm Hg) or a 20 mm Hg increase from the pre–apnea test baseline arterial CO2 has been documented.61 The successful performance of an apnea test requires a specific methodology to minimize hypoxemia during testing and proactively manage hypotension to assure study completion. For these reasons, it should only be done by physicians experienced with the test. Generally, failure to adequately perform an apnea test (e.g., because of hypotension or a marked drop in oxygenation) should indicate the need for another confirmatory test.

Certain spinal movements and reflexes can be observed in brain-dead patients without casting doubt on the diagnosis. They may be especially prominent during apnea testing, and physical stimulation of the patient. They are often short-lasting and symmetric. Well-known but rather uncommon in its complete form is a brief attempt of the body to sit up to about 60° with raising of both arms (Lazarus sign). A convenient classification of spinal movements and reflexes seen in brain-dead patients61 by body region includes but is not limited to: cervical region (tonic neck reflexes or head turning), upper extremities (flexion-withdrawal or extension pronation or flexion), trunk (opisthotonic posturing, flexion, or abdominal reflexes), and lower extremities (plantarflexion, triple flexion, or Babinski sign). Observation of any of these movements should lead to verification that they are compatible with the diagnosis of brain death (by an experienced physician) and preparation of the family of their mechanism and relevance.

Confirmatory Testing

Generally, confirmatory testing is recommended in children less than 1 year old and in situations in which adequate clinical testing cannot be performed. Lack of cerebral blood flow, that is, cerebral circulatory arrest, can be documented by standard arterial angiography, transcranial Doppler, or radionuclide scan, and there is some, but limited, experience with magnetic resonance imaging/angiography and CT angiography. Four-vessel cerebral angiography may be performed in clinically dead patients who have an uncertain diagnosis.57 Complete absence of cerebral circulation is an absolute confirmation of brain death. The major use of the technique is for rapid diagnosis of death in patients whose clinical examination is obscured by hypothermia or drug intoxication. Transcranial Doppler sonography is a noninvasive bedside technique that can measure, in a qualitative fashion, blood flow in the proximal portions of the main cerebral arteries. An ultrasonic probe is placed over the temporal bones (usually) and the direction and velocity of blood flow can be measured. A number of investigators have shown that absence or reversal of diastolic flow in the cerebral arteries can be demonstrated in patients who meet the traditional criteria for brain death;62 the operator-dependent sensitivity ranges from 90% to 99%, and the specificity is 100%.63 About 10% of patients cannot be insonated because of excessive skull thickness. Radioisotope brain scanning can accurately document absent brain blood flow in the cerebral hemispheres, but not in the vertebrobasilar circulation.64 Some institutions have portable units that can be brought to the bedside in the ICU. Because this technique does not image the posterior circulation, the clinical diagnosis of brain stem areflexia becomes even more important. In situations of suspected or known drug intoxication, isotope brain scanning is not helpful, since it will not answer the question of whether brain stem areflexia is due to drug effect or irreversible damage. Xenon-enhanced CT and 99mT-HMPAO single photon emission computed tomography (SPECT) are noninvasive techniques for accurately measuring brain blood flow in all arterial territories. It has been used in young children and infants with great reliability to confirm the clinical criteria of death.65 Particularly in children it can overcome many of the problems associated with EEG and cerebral angiography. Unfortunately, the technique is only available in a referral center.

Absence of any electrical activity on a 30-minute EEG with increased sensitivity settings is consistant with brain death and has a reported sensitivity and specificity of about 90%.66 If the cause of coma is clearly established from anatomic imaging studies of the brain, and clinical criteria are met that all brain functions are absent, EEG confirmation is not necessary.67 Apneic coma and an isoelectric EEG in the face of normal brain imaging studies are strongly suggestive of sedative drug intoxication, and appropriate toxicology studies must be performed.

Neurologic States Resembling Brain Death

As correct identification and verification of brain death is rather difficult and complex and misjudgment may have great negative impact, not only on the patient, but also on the family and the diagnosing medical doctor. The diagnosis should only be made by a physician with experience in careful neurologic examinations, interpretations of brain imaging studies, and skilled evaluation of confirmatory studies used to diagnose brain death. Misleading diagnoses may lead the superficial examiner to believe that the patient is brain dead, but often the history and examination together reveal details inconsistent with the diagnosis of cerebral perfusion arrest. For details, the reader is referred to current textbooks in neurology and a recent monograph by Wijdicks.61 Examples of disorders potentially mimicking brain death include but are not limited to severe hypothermia, acute metabolic coma (e.g., endocrine and organ failure, among others); poisoning (via drugs such as antidepressants, anticonvulsants, analgesics/sedatives, and many more, or via toxins and poisons); locked-in syndrome, akinetic mutism, and possible PVS, as well as peripheral nerve disorders in which the patient may appear brain dead, but is fully awake.

It is well known that brain injury can have immediate and delayed systemic effects on multiple organs. Hence, it is not surprising that brain-dead patients will experience significant effects on overall body function. Generally, vascular motor tone and cardiac stability and performance decreases, pulmonary edema, disseminated intravascular coagulation, and hypothermia may occur, and severe electrolyte and fluid balance disturbances are noted. Up to one-quarter of all potential donor organs are rejected because of the detrimental impact of these changes, which are commonly associated with inadequate medical management (see Chap. 91).

Since the determination of any brain dead patient as a medically suitable organ donor is only done by organ procurement organizations, all brain dead patients should be managed as potential organ donors. This requires proactive anticipatory care to avoid the expected systemic complications that can sabotage the success of organ donation once a family approves. A brief discussion of some of the more important management issues follows.

1. Maintain hemodynamic stability. The initial response of brain injury is massive discharge of catecholamines leading to hypertension and tachycardia and not infrequently myocardial injury. In brain death, hypotension and hypovolemia associated with electrolyte and temperature disturbances will lead to systemic vascular instability. Target hemodynamic goals in these patients could be simplified to maintain systolic blood pressure at about 100 mm Hg, heart rate <100 beats per minute, and urine output >100 mL/h. For fluid resuscitation, boluses of either 5% albumin or crystalloid infusion are recommended, the particular choice depends on the treating physician's preference and individual patient's sensitivity to low osmotic pressure–mediated tissue edema. For heart-lung donors we prefer use of colloids to reduce the risk of precipitating acute heart failure, especially in patients with brain injury–related myocardial abnormalities. Inotropic support is indicated for a volume-resuscitated patient with systolic blood pressures <100 mm Hg; principally, any vasopressor agent can be used; however, ideally use the lowest dose possible to avoid further organ impairments. We commonly initiate treatment with low-dose dopamine to stabilize cardiac function and blood pressure while dilating renal, mesenteric, and coronary vasculature. However, if myocardial contractility is already impaired, dobutamine in combination with dopamine would be an appropriate first step. Blood loss is treated with transfusions to a target hematocrit of about 30%. Consumption of blood factors and platelets due to disseminated intravascular coagulation (aggravated by massive release of brain thromboplastin) can be rapid and replacement may become increasingly difficult.

2. Manage diabetes insipidus. Posterior lobe pituitary injury and necrosis leading to diabetes insipidus (DI) in brain-injured and brain-dead patients is common and should be recognized immediately. Helpful clinical parameters supporting the diagnosis of DI include (1) hypotonic polyuria, identified by urine output >4 mL/kg per hour and urine specific gravity <1.005; (2) urine osmolality <300 mOsm/L; and (3) plasma osmolality >300 mOsm/L (may be confounded by the use of osmotic diuretics). Diabetes insipidus will invariably lead to dehydration, hypernatremia, and vascular collapse. Therapy is simplified by early recognition of DI and includes hypotonic fluid resuscitation on a volume-to-volume basis, and desmopressin acetate, for example, as an initial bolus of 0.3 μg IV, and then as an adjusted dose guided by clinical and laboratory parameters (e.g., titrate urine output to 2 to 3 mL/kg), given approximately every 6 hours to a total 24-hour dose ranging from 10 to 40 μg. Infusion of aqueous vasopressin can be used alternatively (start at 3 U/h); however, the infusion should be stopped prior to surgical organ recovery to diminish the dose-dependent effects on systemic vascular constriction. Recently, the addition of thyroid hormone replacement was shown to have a vasopressor-supporting effect.68 The management of diabetes insipidus can be complex, and is best guided by those experienced with brain death management.

3. Maintaining normothermia. Adverse effects of hypothermia are well known and include reduction of cardiac output and peripheral vascular resistance leading to arrhythmia and hypotension, hypoxia, hyperglycemia, and coagulopathy. Brain-dead patients have lost their ability to control body temperature and are hence dependent on ambient temperature and the temperature of the infusion products they receive. For these reasons (as well as for the establishment of the diagnosis of brain death) the core temperature should be kept constantly above 36°C using conventional methods.

4. Maintain glucose and electrolyte balance. Hypernatremia (e.g., from untreated DI) in excess of 155 mEq/dL may lead to a higher incidence of graft loss after liver transplantation and should therefore be treated aggressively.69 Certainly, electrolytes need to be evaluated and supplemented as needed, at least every 2 to 4 hours, especially with aggressive fluid resuscitation and treatment of DI, as hypernatremia, hypokalemia, hypocalcemia, and hypomagnesemia are common. Hypophosphatemia is frequently observed in brain-dead patients, and if untreated, may lead to hemolysis, rhabdomyolysis, and platelet dysfunction. Glucose, potassium, and ketones should also be checked regularly, as hyperglycemia due to relative insulin resistance, and use of steroid-, pressor-, and glucose-containing solutions is frequent in these patients. Often an insulin infusion is required to maintain blood glucose levels between 150 and 200 mg/dL.

The proper diagnosis and management of brain death is complex and requires expertise and respect for the diagnosis as a real mechanism of death. The need for proper diagnosis supersedes whether or not the patient will become an organ donor. However, the sensitive and thoughtful management of the patient and family can at least keep the opion of organ donation open; a decision to be finalized between organ procurement, and, usually, families. This is one situation in which proper management can potentially save several lives.