Coma

Published on 07/03/2015 by admin

Filed under Critical Care Medicine

Last modified 22/04/2025

Print this page

This article have been viewed 3371 times

Coma

Axel Rosengart, Sea Mi Park, Karen Berger and Igor Ougorets

Chapter Outline

Concept and Terminology of Impaired Consciousness

Alteration of consciousness is a frequent admission diagnosis to critical care services. Most patients require immediate and often extensive diagnostic workup as both time to diagnosis and treatment initiation are decisive factors for brain recovery. To the public, the portrayal of coma is overly optimistic as 89% of “comatose” patients in a meta-analysis of 64 characters in American soap dramas regained consciousness and around 92% “survived” coma.¹ In contrast, medical experience does not compare as favorably, as only about 50% of patients survive in an unselected coma population.^2–4 Therefore, defining coma in a succinct and operational manner will provide a deeper understanding of the depth of brain injury and avoid miscommunications and unrealistic prognostications.

Consciousness can be viewed as the patient’s responses to internal and external stimulation and his or her awareness of self and the environment. These two basic components of consciousness, arousal and awareness, are clinically used to classify the state of consciousness, but, strictly speaking, the qualitative examination of arousal should be separated from the quantitative assessment of the patient’s self- and environmental awareness and cognitive function. These interdependent components are regulated by different neuroanatomic domains within the central nervous system. Arousal and ensuing wakefulness are tightly dependent on the network function of the ascending reticular activating system (ARAS), namely, the midbrain reticular formation, mesencephalic nucleus and tegmentum, thalamic intralaminar (centromedian) nucleus, and dorsal hypothalamus. These build a network of brainstem and diencephalic centers with a strong connection to the cerebral cortices. Self-awareness and the spectrum of cognitive interactions with the environment are the purview of the hemispheres. Simplistically, the hemispheres are structured into three functional systems: units that receive, process, and store sensory information; units that generate and regulate motor activity; and units responsible for programming, regulating, and verifying actions. These three functional systems operate as an intercortical network; relate, as a whole, to the experience of awareness; and allow cognitive performance. The partition of arousal network and intercortical (cognitive) network implies that the level of self-awareness and cognitive performance cannot be examined in a patient with lack of arousal. In other words, in a person with a quantitatively reduced state of consciousness, the qualitative assessment of the content of consciousness is not possible. Practically speaking, the assessment of content (i.e., testing a patient’s cognitive abilities and thinking) should be done after the best level of arousal has been achieved. A fully aroused person can express completely normal or deficient awareness and cognitive performance. Conversely, failure of arousal renders it impossible to test awareness and cognition.

When evaluating a patient’s mental status, the state of consciousness should be viewed as a continuum ranging from the patient who has full alertness and cognitive lucidity to the deeply comatose patient; it is not an all-or-nothing phenomenon.⁵ An all-or-nothing approach limits the interpretation of remaining brain function and, hence, diagnostic certainty. Furthermore, the duration and time development of coma are helpful diagnostic features and should complement the quantitative and qualitative assessment of consciousness. Standard classification systems categorize consciousness based on systematic testing of arousal, awareness, and content. Table 61.1 classifies various abnormalities of consciousness, and the following section delineates clinically important syndromes.

Table 61.1

Classifications and Descriptions of Abnormalities of Consciousness

RAS, reticular activating system.

In the awake state, the examination allows the clinician to test a person for the degree of self-awareness and intactness of cognitive function. Drowsiness, sleepiness, and lethargy resemble reduced spontaneous physical and mental activities in a person who cannot sustain wakefulness without repeated external stimulation. They are somewhat comparable to the experience of lighter sleep, though drowsy patients almost always have reduced attention and concentration and some degree of associated mild confusion. Although sleep and pathologic states of consciousness share common features—for instance, the sleeping person is unaware of his or her resembling unconsciousness—quick reversal to full consciousness is a classic feature of sleep.

A stuporous patient (from Latin stupure, meaning “insensible”) requires repeated, stronger stimuli to show some arousal, and these patients may or may not open their eyes. Full arousal and alertness are not achieved and with continuing external stimulation, restlessness and stereotypic motor responses are observed, but without the appropriate cognitive interactions. A deep stuporous state closer to coma is differentiated by European clinicians and is referred to as sopor (from Latin sopor, meaning “deep sleep”).

Coma (from Greek κµα, or “deep sleep”) is a continuous state of unresponsiveness identified by an inability to arouse to vigorous (noxious) external or internal stimuli. The degree of coma can differ; lighter stages (sometimes denoted as semicoma) can be identified by brief moaning to strong stimulation and associated observed changes in autonomic function; whereas the deepest coma examination shows an absence of any response, including brainstem reflex responses (i.e., lack of oculo- and pupillomotoric responses). Some cyclic autonomic activity such as the sleep-wake cycle and changes in motor tone may coexist. A detailed discussion of coma is provided in the classic textbook by Plum and Posner.⁶

The vegetative state is characterized by the complete absence of behavioral evidence for self- or environmental awareness. This state can follow coma and identifies a state in which brainstem and diencephalic (thalamic) activity is present to a degree that clinical signs of spontaneous or stimulus-induced arousal and sleep-wake cycles are observed. 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; or primitive motor responses. If this state lasts longer than 30 days, it is referred to as persistent vegetative state (PVS) and is used as a descriptive clinical syndrome rather than a disease-specific entity. The most common causes include cardiac arrest, head trauma, severe brain infections, and various causes of thalamic injury. 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.2,7

Minimal conscious state can be diagnosed in patients displaying some but often inconsistent behavioral evidence of awareness of the environment, but they cannot communicate their content and are unable to follow instructions reliably.⁸ It describes a large group of patients who are different from vegetative patients in that they demonstrate some signs of awareness of themselves and their surroundings, albeit inconsistently.

As mentioned earlier, the time progression and persistence of the impairment can be used to further classify abnormalities of consciousness. Delirium is classified as a mental disorder because it involves a fluctuating level of consciousness and pervasive impairment in mental, behavioral, and emotional functioning. It is commonly acute in onset and short in duration and is frequently correlated to a specific etiology such as medications, anesthesia, or sleep deprivation. The criteria for delirium listed in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) include a disturbance of consciousness with reduced the ability to sustain or shift attention and focus. In our experience, attention and concentration deficits are always present in delirious patients and accompanied by a variable degree of cognitive dysfunction unattributed to preexisting dementia. The dysfunctions of this acute syndrome include memory deficits, disorientation, disturbance of language and situational perception, and disorganized behavior. Other delirium markers are its acute onset and tendency to fluctuate along a spectrum of hyperactive (i.e., agitation, restlessness, emotional labiality), hypoactive (withdrawal, flat affect, apathy, decreased responsiveness), or mixed presentations. Concurrent or sequential appearance of features of both hyper- and hypoactive states is not unusual. The history, physical examination, and ancillary test results frequently indicate that the disturbance is caused by an underlying medical or situational condition.

Dementia is a chronic condition in which content of consciousness is affected initially, but in advanced stages reduced levels of arousal are seen. Dementia is generally progressive and affects memory and at least two other cognitive domains such as language, executive function, planning motor tasks, and recognition. Dementing illnesses include a variety of intrinsic degenerative diseases of the cerebral hemispheres, but they also result from many other causes such as traumatic brain injury and hydrocephalus. The underlying etiology can be reversible (i.e., after correction of a vitamin deficiency) or irreversible (Lewy body dementia). Additionally, disease progression can be highly variable, lasting weeks to years, often with fluctuating severity, or be static after the initial impact injury. Anxiety, depression, and agitation may alternate and greatly complicate the disease course. The anatomically affected neurologic structures are commonly a result of bihemispheric cortical dysfunction; however, subcortical structures (i.e., vascular dementia from multifocal small vessel disease) can also represent the predominant pathology.

Akinetic mutism is a silent, alert-appearing immobility of a patient with injury to the hypothalamus or basal forebrain. It manifests as apparent depressed levels of consciousness in a patient with well-formed sleep-wake cycles and with little or no evidence of awareness or spontaneous motor activity. Various etiologies can present or lead to an akinetic mute state, and it is imperative to have a rigorous neurologic examination and careful review of neuroimaging and electroencephalography (EEG). The term locked-in syndrome was introduced by Plum and Posner to describe the quadriplegia and anarthria resulting from the disruption of corticospinal and corticobulbar pathways, respectively. The extent of the deficits in patients with locked-in syndromes can vary, but these patients are aware of the environment and can hear and understand what is discussed around them. They may retain some ability of oculomotor (i.e., ability to open eyes) or other residual cranial nerve function. These patients are also anarthric or severely dysarthric with varying degrees of motor deficits ranging from complete quadriplegia to quadriparesis. As consciousness is preserved, communication is possible using vertical or lateral eye movements or blinking of the eyelid to signal yes/no responses. The majority of cases are caused by basilar artery occlusion leading to brainstem infarction in the ventral pons, yet numerous other etiologies have been described. Abulia (from the Greek meaning “lack of will”) is an apathetic state in which the patient is awake, has normal sleep-wake cycle, and is very slow to respond to stimuli. Mental function is usually normal when tested with sufficient stimulation. It is secondary to bilateral frontal lobe disease and in severe instances may mimic or progress to akinetic mutism.

Neuroanatomy, Neurotransmitter, and Pathology

Consciousness, which consists of three main parts, largely depends on the integrity of brain structure for arousal (Fig. 61.1). The ascending reticular activating system (ARAS) is the lowest order arousal system, consisting of a set of brainstem nuclei located within the brainstem and interconnected by neuronal circuits. The ARAS relays arousal signals to the more rostral thalami, which, in turn, act as hemispheric gatekeepers of arousal and regulate consciousness and sleep-wake transitions. In turn, thalamic activated cerebral cortices allow cognitive processing, which then determines the overall content of consciousness.⁶ Structural damage to or metabolic-chemical derangements of any or all of these elements will affect consciousness. Important components of the ARAS^9–11 include the midbrain reticular formation, mesencephalic nuclei (dorsal raphe nucleus, pedunculopontine tegmental nucleus, locus coeruleus), and ventral tegmental areas. Those areas have rich connections to thalamic intralaminar nuclei and further frontally reaching projections to the tuberomammillary nucleus, lateral hypothalamus, and basal forebrain.

Figure 61.1 Main centers involved in providing arousal and wakefulness and their respective predominant neurotransmitters.

The ascending arousal system of the brainstem contains different neurotransmitter systems. Main cholinergic projections include the ascending mesopontine tegmental and basal forebrain pathways projecting to the thalami and to virtually all subcortical and cortical structures. These projections usually function as synaptic facilitators but at times can act as depressors of transmission.¹² Glutamate is the main transmitter influencing the firing patterns of tegmental cholinergic neurons.¹³ The adrenergic component of the ARAS is closely associated with the noradrenergic neurons of the midbrain locus coeruleus. It runs in parallel with cholinergic projections rostrally to the cortical areas but also descends caudally within the spinal cord.¹⁴ Hypocretin/orexin neurons within the hypothalamus activate both adrenergic and cholinergic pathways and coordinate activity of the entire ARAS system, enhancing complementary and synergistic control of arousal and locomotion.¹⁵ Histamine-releasing cells from the hypothalamus and tuberomammillary neurons are active during the qualitative activation needed for cognition and EEG activation, but they remain silent during sleep.^16–18

The main inhibitor of the arousal system is γ-aminobutyric acid (GABA), without which normal sleep does not occur. The GABAergic circuitry is located in the midbrain and pons, but its activity is regulated and sustained by forebrain structures.¹⁹ It is thought that the GABAergic system’s main function is to contain and channel the spread of arousal, both spontaneously and in response to a stimulus. GABAergic processes are responsible for the occurrence of paradoxical sleep (deep sleep characterized by a brain wave pattern similar to that of wakefulness, rapid eye movements, and heavier breathing but without motor responses). The name “waking neurons” has been given to the serotonergic dorsal raphe nucleus in the midbrain,²⁰ which receives convergent excitatory input from the noradrenaline, histamine, and hypocretin/orexin arousal systems^21,22 as the main activator of the ARAS. Cortical signals that pass through the striatum (caudate and putamen) are refined by the action of dopamine. Dopaminergic neurons are tonically active and stimulate or inhibit cortical neurons depending on various environmental influences (pain, hunger, etc.).^23,24 It is thought that many sensory signals that reach conscious awareness pass through this basal ganglia loop with dopamine as the modulating neurotransmitter. Koch and colleagues suggested that coalitions of neurons compete to dominate conscious thoughts and activities at any given time point and, hence, dopaminergic innervation may facilitate the transient domination of conscious awareness by certain sets of neuron coalitions.²⁵ Therefore, it seems that arousal is an event with a defined time distribution, which is tightly orchestrated by a few key neurotransmitters. As we know, there is no singular “arousal neurotransmitter.”^10,26

Pathology Seen in Patients with Impaired Consciousness

There are four major pathologies that can cause severe, acute, and global reductions of consciousness.6,27 (1) One of these pathologies is the presence of diffuse, global, or extensive multifocal bilateral dysfunction of the cerebral cortex. In this injury mechanism, the cortical gray matter is diffusely impaired and so are cortical-subcortical excitatory feedback loops. The clinical examination will reveal disinhibited autonomic brainstem reflexes, which have also been described as “reticular shock.” (2) Another is injury to the paramedian gray matter from the level of the nucleus parabrachialis of the pons (tegmentum) and reaching to the midbrain pretectal areas and ventral posterior hypothalamus. An injury of this type damages the ascending arousal system and normal cortical activation. The affected structures are predominantly the paramedian gray matter, extending rostrally from the level of the nucleus parabrachialis of the pons (tegmentum) and reaching rostrally as far as the adjacent pretectal area and ventral posterior hypothalamus. (3) The widespread disconnection of the cortex from subcortical activating mechanisms acts pathophysiologically to produce effects similar to both of the previously mentioned conditions. (4) Finally, cortical and subcortical arousal mechanisms can be affected by a variety of diffuse disorders and to various degrees—for example, metabolic encephalopathy in a patient with acute liver failure.

Approach to Coma

The initial approach to stupor and coma is based on the principle that all alterations in arousal are acute, life-threatening emergencies. Urgent steps are required to prevent or minimize permanent brain damage from reversible causes. Patient evaluation and treatment must occur simultaneously. Serial examinations with accurate documentation are necessary to determine a change in the patient’s status. Accordingly, management decisions (therapeutic and diagnostic) must be made. The clinical approach to an unconscious patient entails the following steps: (1) emergency treatment; (2) history (from relatives, friends, and emergency medical personnel); (3) general physical examination; (4) neurologic profile, the key to categorizing the nature of coma; and (5) specific management.

Emergency Management

Initial assessment must focus on vital signs to determine appropriate resuscitation measures; diagnostic processes can occur later. Urgent and sometimes empiric therapy must be given to avoid additional brain insults.

Oxygenation and Intubation

Oxygenation must be ensured by establishing an airway for ventilation of the lungs. The threshold for intubation should be low in a comatose patient, even if respiratory function is sufficient for proper ventilation and oxygenation. The level of consciousness may deteriorate, and breathing may decompensate suddenly and unexpectedly. An open airway must be ensured and protected from aspiration of vomitus and blood. If severe neck injury is a possibility or has not been excluded, intubation should be performed by the most skilled practitioner without extension of the patient’s neck. Mask-bag-assisted ventilation should continue during the examination if necessary. Signs of arousal include dilated, reactive pupils, copious tears, diaphoresis, tachycardia, and systemic hypertension.

A brief neurologic examination is mandatory before any sedative or paralytic required for intubation is administered. The key points of a rapid neurologic examination are the following:²⁸

• Assessment of level of arousal and level of consciousness (i.e., coma, stupor, lethargy)

• Pupillary size and response to light

• Abnormalities of eye movements (i.e., disconjugate eyes; unilateral gaze paralysis; lack of voluntary movements)

• Grimacing and motor responses to noxious stimulation

• Abnormal plantar response (Babinski’s sign)

The preferred route of emergency intubation is orotracheal, which can be performed rapidly, safely, and reliably with inline stabilization of the neck in patients with suspected cervical spine injury. Intubation causes intense reflexive cardiovascular stimulation that may lead to a deleterious elevation of intracranial pressure (ICP). Therefore, the patient should be adequately sedated. Etomidate is the preferred sedative in patients with suspected ICP elevation as it reliably facilitates induction in less than 1 minute with a duration of action of 4 to 6 minutes. Propofol is another anesthetic agent that does not increase ICP; however, its hypotensive effects, which can ultimately decrease cerebral perfusion pressure, often limit its use. Both medications result in a dose-dependent decrease of cerebral metabolic rate that reduces cerebral blood flow and ICP. Ketamine has a fast onset, but it may elevate ICP and should generally be avoided. Midazolam can alter the patient’s mental status and may prohibit a postintubation examination.

Body Temperature

Hyperthermia is detrimental in brain injury as it increases brain metabolic demands and facilitates secondary brain injury.³⁰ Elevated temperature greater than 104° F (40° C) requires immediate, lifesaving cooling measures, even before the underlying cause is determined and treated. In 2002, two research groups independently published that lowering the body temperature to 33° C for 12 or 24 hours in comatose survivors of cardiac arrest resulted in nearly doubling the number of patients being discharged home or to rehabilitation. Generally speaking, no patient with acute brain injury should be allowed to be hyperthermic, and modern technology provides a variety types of noninvasive and invasive cooling equipment suitable to maintain any target core temperature desired. Core temperatures of less than 93° F (34° C) on admission should be slowly elevated to above 35° C to reduce the risks of unwanted side effects from uncontrolled, persistent hypothermia.

History

Once the patient’s condition and vital signs have been stabilized, clues to the cause of coma should be sought by interviewing relatives, friends, bystanders, or medical personnel who may have observed the patient before or during the decline in consciousness. The history should address witnessed events, evolution of coma, recent medical and surgical histories, and medication/drug intake.

Physical Examination

A systematic, detailed examination is necessary when approaching a comatose patient (Box 61.1, Table 61.2). Important findings include evidence of trauma, acute or chronic medical illnesses, ingestion of drugs (needle marks, alcohol breath), and the presence of nuchal rigidity.

Box 61.1 Neurologic Profile

A Modified Glasgow Coma Scale

Table 61.2

Correlation Between Levels of Brain Function and Clinical Signs

Structure	Function	Clinical Sign
Cerebral cortex	Conscious behavior	Speech (including any sounds) Purposeful movement Spontaneous To command To pain
Brainstem activating and sensory pathways (reticular activating system)	Sleep-wake cycle	Eye opening Spontaneous To command To path
Brainstem motor pathways	Reflex limb movements	Flexor posturing (decorticate) Extensor posturing (decerebrate)
Midbrain CN III	Innervation of ciliary muscle and certain extraocular muscles	Pupillary reactivity
Pontomesencephalic MLF	Connects pontine gaze center with CN III nucleus	Internuclear ophthalmoplegia
Upper pons
CN V	Facial and corneal	Corneal reflex-sensory
CN VIII	Facial muscle innervation	Corneal reflex-motor response Blink Grimace
Lower pons
CN VIII (vestibular portion) connects by brainstem pathways with CN III, IV, VI	Reflex eye movements	Doll’s eyes Caloric responses
Pontomedullary junction pressure	Spontaneous breathing Maintained blood pressure	Breathing and blood pressure do not require mechanical or chemical support
Spinal cord	Primitive protective responses	Deep tendon reflexes Babinski response

CN, cranial, nerve; MLF, medial longitudinal fasciculus.

As outlined earlier, an in-depth physical and neurologic examination followed by serial neurologic evaluations (i.e., hourly) is the most important, indispensable, and readily available method of assessing a comatose patient.³¹ However, in coma patients the examination remains limited in detecting changes in brain function. Nevertheless, the astute clinician will carefully evaluate for changes in findings (i.e., the appearance of new asymmetry in examination or focal deficits, progressive loss of brainstem reflexes, or loss of reflex motor responses). Finding progression of impairment is of immense clinical value as it indicates a new or worsening “intracranial emergency” demanding immediate clarification and stabilization. Skilled neurologic examination is challenging and provides a limited surveillance window of brain tissue at risk. Intracranial monitoring and, if needed, repeated head imaging have become the standard of care in neurocritical care units.

The differential diagnosis of a comatose patient is supported by the patient’s risk factor profile, activity at symptom onset, and disease progression. Acute, devastating headaches with nausea and vomiting quickly followed by impaired consciousness in a hypertensive patient with a history of smoking should be rapidly worked up for subarachnoid or intracerebral hemorrhage. Deterioration of consciousness hours after head injury necessitates immediate exclusion of newly emerging, delayed intracranial hematoma with intracranial pressure crisis. In contrast, the slow progression of a focal neurologic deficit evolving to generalized depressed consciousness in an elderly individual directs toward an intracranial metastasis with secondary seizures. Increased intracranial pressure is a common denominator of many acute and subacute central nervous system (CNS) injuries. Typical manifestations include headaches from the dural stretch of trigeminal (V) sensory fibers; intractable nausea and recurrent vomiting often associated with visual disturbances; some neurologic deficits on detailed examination; and, in later stages, reduction in level of alertness.

When performing the neurologic examination, monitoring vital signs can be utilized as part of the autonomic nervous system evaluation. For example, appearance of systolic hypertension, vagal bradycardia, and respiratory irregularity is known as Cushing’s triad and indicates increased ICP. In addition, changes in the pattern of respiration in a spontaneously breathing patient can help localize the level of injury. Abnormalities of respiration can range from apnea to hyperpnea with undulating crescendo-decrescendo patterns as well as complete irregularity of breathing with erratic pauses that finally terminate in complete apnea. This “autonomic survey” and observing for spontaneous patient movements are followed by the assessment of a patient’s level of arousal. The reduction in response to external stimulations provides the basis for commonly used classification schemes such as the Glasgow Coma Scale (GCS). If the patient’s level of alertness permits, then orientation, attention and concentration span and intactness of cognitive functions are examined next. This is followed by a detailed assessment of cranial nerve, motor, and peripheral reflex status. Optic nerve head edema (papilledema) is an important and, if present, a reliable manifestation of raised ICP. Additional findings of increased ICP on fundoscopy include venous engorgement, loss of venous pulsation, optic disk hemorrhage with increased disk diameter, and blurring of its margins. Compression of the third nerve is a result of a variety of intracranial processes, including uncal herniation and aneurysms of the posterior communicating artery. It is identified by pupillary dilatation, ptosis, and various degrees of ophthalmoparesis except for abduction and intorsion-depression. A sixth nerve palsy, typically a nonlocalizing sign, is identified by abduction deficit and is seen in patients with elevated ICP or hydrocephalus and stretching of the nerve but is also present in patients with intrinsic brainstem injury (i.e., strokes). Patients with acute intracranial mass lesions and increased ICP commonly develop herniation syndromes—that is, shifting of brain tissue from areas with higher regional ICP toward areas with normal ICP. If the mass lesion is predominantly located within one hemisphere, the main force vector will shift the brain primarily laterally across the midline and toward the opposite hemisphere as it impresses on the upper brainstem and thalamic structures during lateral displacement. More centrally located mass lesions (i.e., acute, obstructive hydrocephalus or global brain swelling as seen in fulminant encephalitis or hepatic encephalopathy) will to the contrary lead primarily to downward herniation with the main force vector impressing upon the brainstem. Different herniation syndromes have different clinical presentations and therapeutic approaches. When intracranial hypertension remains uncontrolled and escalates uniformly central, downward herniation occurs involving increasingly more brainstem structures in a rostrocaudal fashion. The terminal event is complete brainstem destruction resulting invariably in brain death. During this course, progressive loss of brainstem reflexes can be determined on examination and at this stage, the patient is deeply comatose. Injury to the fifth cranial nerve nucleus within the pons results in loss of the corneal reflex. A vestibulocochlear nerve nucleus compromise results in loss of the oculocephalic reflex and manifests as “doll’s eyes” or the inability to maintain eye position as the head moves. Medullary injury with damage to the ninth and tenth nerves results in loss of lower cranial nerve reflexes including the absence of gag and cough reflexes.

Ipsilateral hemiparesis indicates the possibility of uncal herniation causing compression of the contralateral cerebral peduncle as temporal lobe tissue herniates into the space between the tentorium and brainstem. For example, a left-sided subdural hematoma causing temporal herniation pushes the opposite cerebral peduncle and its motor fibers against the right tentorial edge (right-sided Kernohan’s notch) leading to a hemiparesis on the same side as the hemispheric lesion. Posturing is involuntary flexion or extension of the arms or legs spontaneously or elicited by stimulation, and it is included in the Glasgow Coma Scale as a measure of the severity of brain injury. Three types of posturing can be observed depending on the level of brain injury: decorticate (flexor), decerebrate (extensor), and opisthotonos (body arching along the craniospinal axis) posturing. Injury involving the brainstem above or below the red nucleus in the midbrain leads to decorticate and decerebrate posturing, respectively. Of note, flexion or extension can be witnessed either unilaterally or bilaterally, is sometimes seen only intermittently, and can involve one or all extremities. Decorticate posturing commonly indicates damage of the cerebral hemispheres, the internal capsule, and the thalamus, possibly also involving the uppermost brainstem. Decerebrate posturing is involuntary extension of both upper (elbow) and lower extremities indicating brainstem damage below the midbrain. Progression from decorticate to decerebrate posturing is indicative of progressive transtentorial herniation. Opisthotonic posturing is an infrequently encountered sign seen with severe brainstem injury or extrapyramidal lesions involving the axial muscles. Importantly, increasing downward pressure leads to dysfunction first at the level of the diencephalon (i.e., thalami), next affecting the upper and middle sections of the brainstem (midbrain and pons), and ultimately impeding medullary function leading to deep coma without any reflexes or motor responses (flaccidity). Once complete rostrocaudal herniation has occurred, the patient is brain dead.

Assessment of Coma: General Aspects

Correctly characterizing disorders of consciousness continues to pose interesting clinical questions and diagnostic challenges with important ethical consequences (Box 61.2). Not only may an individual patient acutely fluctuate in his or her examination findings, but recovery may take place over prolonged time periods, necessitating constant reassessments. To address this uncertainty, standardized neurobehavioral assessments have become the best tool for categorizing coma and its transitional stages. Predicting long-term outcome has significantly improved when utilizing standardized assessments. Generally speaking, clinical and electrophysiologic markers of coma and its transitional states remain unsatisfactory. The reader is reminded that larger observational studies identified a high rate of misdiagnoses (>40%),³² especially in patients in vegetative state, when assessment is made on clinical grounds only.

Box 61.2 Characteristics of Categories in Coma

Supratentorial Mass Lesion Affecting the Diencephalon/Brainstem

Initial focal cerebral dysfunction

Dysfunction progresses rostral to caudal

Signs reflect dysfunction at one level

Signs often asymmetric

Infratentorial Structural Lesion

Symptoms of brainstem dysfunction or sudden onset coma

Brainstem signs precede/accompany coma

Cranial nerve and oculovestibular dysfunction

Early onset of abnormal respiratory patterns

Metabolic-Toxic Coma

Confusion/stupor precede motor signs

Motor signs usually symmetric

Pupil responses generally preserved

Myoclonus, asterixis, tremulousness, and generalized seizures common

Acid-based imbalance common with compensatory ventilatory changes

Psychogenic Coma

Eyelids squeezed shut

Pupils reactive or dilated, unreactive (cycloplegics)

Oculocephalic reflex unpredictable; nystagmus on caloric tests

Motor tone normal or inconsistent

No pathologic reflexes

Awake-pattern EEG

EEG, electroencephalogram.

Standardized Neurobehavioral Assessment

The behavioral diagnosis of the level of consciousness in severely brain-damaged patients remains challenging. To improve differential diagnostic evaluations, evidence-based recommendations endorse neurobehavioral characteristics using standardized tools.³³ The Coma Recovery Scale-Revised (CRS-R) has excellent content validity and is a solid tool addressing all established differential diagnostic criteria. To differentiate vegetative, minimally conscious state (MCS) or emerging MCS, several other tools have demonstrated good validity, including the Sensory Modality Assessment Technique (SMART), Sensory Stimulation Assessment Measure (SSAM), Wessex Head Injury Matrix (WHIM), and Stimulation Profile (WNSSP). The FOUR and CRS-R tools showed substantial evidence of good interpreter reliability. Our recommendations are to use the Coma Recovery Scale-Revised (CRS-R)³² as a valid and reliable tool in most situations for differential diagnostic assessments. With reservations, the SMART, WNSSP, SSAM, WHIM, and DOCS can be used. The lack of evidence with respect to content validity, standardization, and reliability make the FOUR, INNS, Glasgow Coma Scale, Swedish Reaction Level Scale-1985, Loewenstein Communication Scale, and CLOCS undesirable.

Neuroimaging

Novel neuroimaging modalities have not only provided additional insight into residual brain function in patients with disorders of consciousness but have also raised ethical discussions concerning the clinical management of these patients. Neuroimaging studies have provided considerable evidence for the existence of major functional differences in patients with vegetative state (VS) and minimally conscious state (MCS).34–37 For example, positron emission tomography (PET) and functional MRI (fMRI) studies delineated more extended cerebral processing in response to various external stimuli of modalities in patients with MCS.^35,36,38 Resting state fMRI studies identified that MCS patients have preservations of functional connectivity to higher-order associative cortices, such as default networks.^37,39 Complementary to these results, brain tissue integrity, as assessed by diffusion weighted imaging and gray matter volumetry, is more preserved in MCS as compared to VS.^34,40 These findings correlate to the relative improved long-term prognosis in MCS.⁴¹

Currently, MR imaging is the procedure of choice for structural and functional imaging of the brain. Further experience with MR tractography (DTI), estimating the extent of brain connectivity, and MR spectroscopy, evaluating regional neuronal chemistry, will enhance outcome predictions in patients with severe brain injury. In addition, passive, active-response, and resting-state functional MRI paradigms are currently being validated to help in differentiating impaired states of consciousness. The combination of clinical examination and standardized coma assessment with structural and functional imaging currently seems to provide the best estimates for long-term outcome.

Conventional Magnetic Resonance Imaging (MRI)

Commonly available morphologic MRI acquisitions include non-contrast-enhanced sagittal T1, axial diffusion, axial fluid attenuated inversion recovery (FLAIR), axial T2, coronal T2* (or gradient echo [GRE]) sequences as well as 3D T1-weighted volume acquisitions.

FLAIR and T2-SE sequences are commonly used to delineate, measure, and monitor the extent and impact of acute brain injury. Common pathologies seen include brain edema, contusions, intracerebral, dural or subarachnoid blood, herniation syndromes, and hydrocephalus. T2* (also called GRE) and susceptibility-weighted imaging (SWI) are helpful for delineating blood product and detecting hemorrhagic diffuse axonal injuries (DAI). In traumatic Brain Injury (TBI) patients, the total number of lesions detected on FLAIR and T2* has been shown to inversely correlate with admission GCS score.42,43 Three-dimensional rendering allows evaluating and quantifying brain atrophy after injury.⁴⁴ Conventional MRI is useful in TBI patients. Lesions in the pons, midbrain, and basal ganglia, especially when bilateral, correlate with poor outcome. However, the contrary may not be true as MRI may fail to explain why some vegetative state or marked cognitively impaired patients have minimal or no lesions. Further advanced imaging studies would be especially helpful for those patients (discussed later). Diffusion tensor imaging (DTI) can be viewed as an extension of diffusion-weighted imaging (DWI) where diffusion restricted water molecule movement (i.e., water trapped in ischemic cells or, in DTI, confined to axonal spaces) is preferentially visualized. As with most conventional MRI techniques, obtaining results with DTI is independent of the coma or sedation status of a patient.⁴⁵ Serial changes in white matter DTI metrics can be used as surrogate markers to evaluate treatment responses even when clinical scores cannot be obtained.⁴⁶ Whole brain white matter and regional DTI measures have become sensitive markers of TBI impact correlating with acute and discharge injury severity, especially when adjusted for age, gender, and admission GCS score.^47,48

Proton Magnetic Resonance Spectroscopy (MRS)

MRS reveals noninvasively regional, metabolic information of targeted brain regions correlating with the functional neuronal state of the examined sample region. To assess brain function, comprehensive MRS protocols looking for changes in axial chemical shift imaging at the level of the cortical white matter,49,50 splenium,⁵¹ basal ganglia (thalamus),⁵² insular cortex, and pons⁴² have been used. Using this approach, it is possible to detect posttraumatic neurochemical damage and neuronal dysfunction in brain areas where conventional MR imaging techniques show no abnormalities.⁵³ In a TBI case-control study, the N-acetylaspartate (NAA)/creatinine (Cr) ratio correlated as a good neuronal marker with the overall recovery (and therefore lack as reliable index of unfavorable outcome after TBI), whereas less consistency was found using the choline (Cho)/Cr⁵⁴ or NAA/Cho ratios.⁵⁵ Moreover, early posttraumatic spectroscopic assessment of axonal damage is clinically relevant for prognostication, and significant increases in brain lactate levels seem to be a dependable index of injury severity and outcome in TBI.⁵⁶

Positron Emission Tomography (PET)

18-fluorodeoxyglucose (FDG)-PET delineates glucose utilization deficiencies and, hence, it provides insights to neuronal energy metabolism correlating to regional, functional performance. For example, in acute vegetative state, overall glucose utilization was significantly reduced to 50% to 70% in comparison with age-matched, healthy controls.⁵⁷ As expected, patients with locked-in syndrome (i.e., pontine lesions) showed normal cortical metabolism. TBI patients with diffuse axonal injury had neuronal hyperglycolysis and metabolic depression.⁵⁸ 1C-Flumazenil (FMZ)-PET measures the density of benzodiazepine receptors (BZRs) providing an estimate of neuronal integrity.¹⁵ O-H₂O PET imaging delineates activation-induced changes in regional cerebral blood flow in response to passive external (auditory, somatosensory or visual) stimulation. The combination of preserved activation of primary cortices (i.e., visual) but lack of co-activation of secondary processing areas provides evidence that “higher-order” associative cortical networks are disconnected from the primary sensory cortex.^38,59–63 Thalamocortical disconnections have also been reported using this technique.⁶⁴

Functional MRI (fMRI)

Functional MR imaging provides activity measures of brain regions at rest and after both passive (internal, though-induced) and active (external, command-induced) stimulation.⁶⁵ Such acquisitions can obtain estimates of average metabolic activity in relatively short time periods (i.e., 10 minutes) compared to PET (~30 minutes).⁶⁶ fMRI has been shown to be a good predictor of outcome in patients with impaired consciousness.^67–72 The previously discussed PET evidence of preserved but isolated (disconnected) cortical activation areas in a patients in vegetative state was corroborated by fMRI.^39,73–76

Electrophysiology

EEG

Although a single electroencephalography (EEG) study can be a helpful diagnostic tool (i.e., to identify seizures), diagnostic certainty improves with continuous or serial EEG monitoring. Reactivity that is the increase in brain wave frequencies with patient stimulation has been used to assess thalamocortical function. For example, a patient with renal or hepatic dysfunction may have slowed background rhythms with triphasic waves and preserved or absence reactivity. Cerebral hypoxia, drug intoxications, and sedative medications as well as other CNS depressants produce abnormal EEG patterns including burst suppression, alpha coma, and spindle coma. Focal EEG slowing with or without evidence of abnormal excitability (i.e., focal spikes) can be found in comatose patients with structural lesions, either focal cortical, focal deep-seated diencephalic or mesencephalic reticular formation, or diffuse bihemispheric. Furthermore, EEG is the primary diagnostic tool to identify patients with impaired consciousness due to nonconvulsive seizures or status epilepticus.

Interpreting EEG can be challenging for anyone not trained in neurophysiology. However, despite the numerous etiologies leading to coma, EEG in coma patients is dominated by a few stereotypic patterns generally providing good indicators of the depth of the coma. Separating three common themes of interpretations will aid the nonexpert in utilizing EEG for diagnosis and prognosis in patients with disorders of consciousness. EEG findings can be categorized into responsiveness (reactivity to acoustic, painful, and photic stimuli), brain wave frequency abnormalities (focal or global), and specific EEG patterns (i.e., seizure discharges). Keeping this categorization in mind, commonly encountered EEG findings will be discussed next.

In patients with alpha coma the underlying brain wave activity is predominantly in the (normal) alpha range, from 8 to 12 Hz. In contrast to the normal EEG in a healthy, awake person, the EEG pattern found in comatose patients lacks alpha reactivity, implying a significant disturbance of thalamocortical pathways. In 36 patients in alpha coma,⁷⁷ one third showed anterior predominance and two thirds (66%) showed uniform distribution of alpha activity. Three major brain-injury categories have been observed in alpha coma:^78,79 status postcardiorespiratory arrest (anoxic encephalopathy), toxic encephalopathies, and de-efferentiated (locked-in) state (i.e., pontine hemorrhage).

Theta coma and alpha-theta coma patients show slower brain wave activities in the theta 4- to 7-Hz range (theta coma), which in some is superseded by a predominance of alpha activity (alpha-theta coma).⁸⁰ In many coma patients, the alpha and theta activities will change after about 5 days to a more permanent pattern at which time the EEG may be uniformly slow and depressed. EEG reactivity may emerge in subsequent studies as a relatively favorable sign; however, resolution into a burst-suppression pattern without reactivity signifies unfavorable outcome after anoxic encephalopathy.⁸⁰

Beta coma consists of fast activities in the 12- to 16-Hz range, usually with frontal predominance. Interspersed alpha, theta, delta, and spindle (sleep)-like activities can be found. The most frequent cause for beta coma is sedative-hypnotic overdose from substances such as benzodiazepines and barbiturates.⁸¹

Intermittent rhythmic delta activity (IRDA) consists of low amplitude, slow 2- to 3-Hz sinusoidal waves. Those waves can appear intermittently during the early stages of impairment of consciousness and soon after loss of alpha rhythm.⁷⁸ This nonspecific pattern can be seen in many metabolic and toxic encephalopathies, hemispheric lesions, acute brain injuries, and other conditions affecting cortical and subcortical structures diffusely. The prognosis of IRDA depends on the underlying etiology.

Continuous high-voltage delta activity consists of arrhythmic, high-amplitude, polymorphic 1- to 2-Hz delta activity. This pattern is seen in later stages of coma and may evolve out of theta coma. This contrasts with the slow, nonreactive theta and delta coma EEG of low amplitude (<20 µV) and without reactivity.

Spindle coma consists of an EEG pattern predominated by 11- to 14-Hz spindle discharges, which are generally symmetric and synchronous, appearing as sleeplike oscillations on a background rhythm dominated by theta and delta activities. It resembles the EEG of a sleeping person and additional findings include anterior, low-voltage rhythmic bursts, K complexes, vertex waves, and delta slowing. However, in these patients, wakefulness is not regained after stimulation and depth of consciousness does not fluctuate in spindle coma. These patterns indicate dysfunction at the brainstem level. Traumatic brain injury, anoxic encephalopathy, infectious encephalitis, drug-induced and metabolic encephalopathy, as well as postictal states can lead to this EEG finding.82–84 The prognosis of spindle coma is uncertain, as both positive^85–87 and negative outcomes⁸⁴ have been reported.

Periodic lateralized epileptiform discharges (PLEDs) occur asymmetrically (lateralized) and are episodic in nature. Morphologically, variations include sharp waves, spikes, slow waves, or a combination of these followed by a slow wave with a periodicity ranging from 0.3 to several seconds.⁸⁸ PLEDs are nonspecific and can be seen in a wide variety of brain injuries such as after cortical ischemia and with brain tumors. They are also associated with recent seizures, alcohol withdrawal, and toxic-metabolic encephalopathy among many others.⁸⁹ In contrast, generalized periodic epileptiform discharges (GPEDs) are seen bilaterally and typically occur at a frequency of <1 Hz. Subclassifications define the interval between discharges and include periodic short-interval diffuse discharges (PSIDDs), periodic long-interval diffuse discharges (PLIDDs), and suppression burst patterns.⁹⁰ GPEDs are seen after global brain injury—for example, after hypoxic or hepatic encephalopathy, drug intoxication, but also with various degenerative disorders.⁹⁰

Bilateral independent periodic lateralized epileptiform discharges (BIPLEDs) are slow (<1 Hz), bilateral independent epileptiform activity ⁹¹ seen in coma due to bihemispheric lesions. For example, after anoxic encephalopathy and infections, they are associated with a poorer prognosis⁹¹ and about two times higher overall mortality than patients with PLEDS.^91,92

The burst-suppression pattern identifies an EEG demonstrating generalized, synchronous bursting of slow but mixed-frequency waves of high-voltage alternating periodically interspersing within longer periods of isoelectric (flat) EEG. It identifies abnormal cortical excitability,⁹³ and with deepening coma, the isoelectric periods become longer. Main etiologies include anoxic encephalopathy, severe intoxication, or sedative/anesthesia-induced coma states. The overall prognosis is determined by the underlying etiology.⁹⁴

Triphasic waves are distinct, blunt delta waves (~2 to 3 Hz) consisting of a high-voltage positive wave preceded and followed by lower amplitude negative waves. This nonspecific EEG can be seen in various brain injuries, among them serotoninergic syndrome, Creutzfeldt-Jacob disease, lithium and baclofen toxicity, hepatic and renal insufficiency, and neuroleptic malignant syndrome.⁹⁵

Studies have identified that different EEG patterns seen in patients with hypoxic (postarrest) encephalopathy include suppression, burst suppression, alpha and theta coma, complete generalized suppression (<10 µV), and generalized periodic complexes in combination with burst suppression. Unfortunately, common to all of them is the poor correlation to long-term outcome.⁹⁴

Event-Related Potentials (ERPs)

These EEG potentials are identifiable after signal-averaging EEG-derived potentials evoked by a specific sensory, cognitive, or motor stimulus. Under normal circumstances, the brain generates event-related, neuronal (electrical) responses to stimulation. This neuronal activation can be detected by surface EEG; however, it requires specific mathematical processing of the EEG signal to identify brain oscillations that are time and phase locked to particular events (stimuli). Detection of such event-related potentials in patients with impaired consciousness has been interpreted as the brain’s ability to process stimulus information and for cortical movement preparation. Advantages of this approach include detected signals that have an excellent time-stimulus correlation within the millisecond range helpful to relate a particular stimulus to certain cortical processing. The method is low cost and noninvasive, and it can be used to monitor the brain function. Fast processors have allowed the development of brain-computer interfaces based on event-related signal responses that assist communication with brain-injured patients. Several ERP paradigms have been developed allowing identification of specific response signals. The N1-P2 complex has been identified in visual ERPs and its presence interpreted as a negative correlation to repressiveness.⁹⁶ The P300 (or P3) is an average, positive potential with a latency of about 300 milliseconds. It is generated when a patient mentally reacts to a defined cognitive stimulus and is related to positive outcomes in coma.⁹⁷ Cognitive responses to command have been verified in patients with complete locked-in syndrome after basilar artery thrombosis.⁹⁸

Evoked Potentials

Brainstem auditory evoked responses (BAER) are a series of seven potentials recorded at specific time intervals at the skull after defined auditory stimulation employing headphones. Based on the redundancy and interconnections of the auditory pathways, BAER are robust and can be obtained in many patients with depressed level of consciousness, including patients under anesthesia or deep sedation. This implies that the absence of BAER, or its lemniscal complexes III-IV reflecting the integrity of upper brainstem, uniformly infers a poorer prognosis in coma (i.e., 3 to 6 days after head injury).⁹⁹ In addition, BAER have been employed as a “brainstem monitoring” tool to identify progressive downward brain herniation. Frequently, in clinical settings, BAER monitoring is combined with somatosensory evoked potential (SEP) and EEG monitoring. Somatosensory evoked potentials (SSEP) are a series of peaks recorded over the neck and scalp in response to electrical stimulation of a peripheral nerve, usually the median nerve. The N13 wave is a negative peak at 13 milliseconds reflecting activity at the cervical cord, whereas the N20 is a negative peak at 20 milliseconds reflecting the earliest cortical activity. Longer latency responses are also seen at 25, 35, 45, 70, and 95 milliseconds, reflecting higher cortical processing. Multimodality evoked potentials (MEP) combine simultaneous brainstem auditory, somatosensory, and visual evoked potential monitoring¹⁰⁰ in order to enhance the predictive value of evoked potential monitoring—for example, in TBI patients¹⁰¹ or during induced barbiturate coma where examinations cannot be obtained.¹⁰²

Cerebral Blood Flow Monitoring

Transcranial Doppler (TCD) records the velocity and pulsatility of cerebral blood flow, allowing hemodynamic analysis, despite the use of sedatives or moderate hypothermia. TCD has been utilized for serial examinations in comatose patients after initial recovery from cardiac arrest in order to detect and treat early cerebral hemodynamic changes. If a patient has an appropriate temporal ultrasound window, TCD can determine cerebral perfusion arrest and, hence, can be used as an ancillary study to determine brain death.¹⁰³ Further, velocity and pulsatility indices correlate as a noninvasive measure with ICP elevations and also identify the delayed presence of a hyperemic flow pattern—that is, high velocities with low pulsatility as seen in patients with intracranial hypertension from hyperemia.¹⁰⁴

Xenon-Enhanced CT Scanning

Xenon CT utilizes the diffusion of xenon into the CNS to determine cerebral blood flow (CBF)¹⁰⁵ in order to assess perfusion deficits after a traumatic brain injury.¹⁰⁵ Using hyperventilation, xenon CT can be used to estimate cerebral vasoreactivity, a marker correlating with outcome in comatose TBI patients.¹⁰⁶

General Treatment Approach

The immediate goals in a comatose patient are stabilization of vital signs, rapid diagnosis, immediate treatment to reverse the primary injury (i.e., recanalization of cerebral blood flow), and minimization of further nervous system damage from secondary injury (i.e., brain swelling). These general goals are greatly facilitated by a team approach of medical and nursing staff with expertise in acute brain injuries. Successive care steps in managing coma of unknown etiology include (1) immediate exclusion of most harmful brain injuries (i.e., intracranial bleeding from an unsecured aneurysm), (2) sequentially ruling out alternative diagnoses, (3) determining the precise injury type leading to coma, (4) stabilization and reversal of primary brain injury, and (5) intensive brain monitoring (serial examinations, imaging, ICP, etc.) with watchful anticipation and early treatment of secondary injuries. It is of utmost importance to remember that the “time is brain” concept holds true for each of these care steps. Acute coma should be viewed as a “brain arrest” and approached with similar urgency and diligence as cardiac arrest.

If known, initial coma evaluation must take the injury mechanism into account. For example, in trauma victims, empiric cervical spine stabilization and early detection of an unstable c-spine injury by reconstructed cervical CT will be needed (i.e., before evaluating oculocephalic responses). Endotracheal intubation is frequently required due to the risk of aspiration and the need to maintain an open upper airway in an unconscious patient. Prehospital endotracheal intubation was associated with a significant decrease in mortality from 36% to 26% in patients with blunt injury and a GCS ≤ 8, especially those with severe head injury by anatomic criteria.¹⁰⁷ Furthermore, mechanical ventilation is required to avoid oxygen desaturations (of the brain) or to induce therapeutic hypocapnia to acutely lower uncontrolled ICP. Early tracheotomy is often indicated in moderate to severe brain injuries. For example, a GCS of ≤4 on day 3 of injury or a Simplified Acute Physiology Score (SAPS)¹⁰⁸ of >15 on day 4 can indicate tracheotomy with 84% positive predictive value.¹⁰⁹ There are many benefits of early tracheotomy including increased patient comfort and subsequent reduction in sedatives and analgesics and ease of oral and pulmonary care, and its association with shorter duration of mechanical ventilation and ICU length of stay.¹¹⁰ At our institution we commonly perform bedside tracheotomy, which is a safe bedside procedure that circumvents the need to transport patients with brain injury and reduces overall costs.¹¹¹

Any correctable coma etiology such as abnormal glucose and electrolyte levels, hypotension, hypoxemia, or abnormal core temperatures should be managed immediately. Intravenous access is immediately established for the administration of resuscitation fluids, administration of contrast, and therapeutic care. Infusion of hypotonic solutions should be avoided to prevent deterioration of intracranial swelling, and there is a definite role for immediate ICP monitoring (i.e., using an external ventricular drain, which also allows the release of CSF) in any acute brain injury with the potential for ICP. The management of raised ICP is discussed in Chapter 16.

Traumatic and Nontraumatic Coma

Coma is a descriptive term based on examination findings and can result from a range of heterogeneous etiologies that are classically grouped into traumatic versus nontraumatic coma. As outlined earlier, other stages of impaired consciousness, including vegetative state, typically follow nonremitting acute coma. Usually after 1 month of coma, the term persistent vegetative state is used for both nontraumatic and traumatic etiologies; however, some authors prefer the term permanent vegetative state 3 months after nontraumatic and 1 year after traumatic injury. Notably, these terms do not imply irreversibility and do not exclude the potential for further recovery.¹¹²

Dividing coma into traumatic and nontraumatic etiologies is useful as each group is predominated by different demographics and disease prognosis. Often patients with traumatic brain injury achieve greater functional and cognitive improvements than patients suffering from acquired cerebrovascular or anoxic injury. In a longitudinal, prospective study of 192 traumatic, 104 cerebrovascular, and 33 anoxic brain injury patients, male gender was prevalent in all etiologies. Patients with traumatic brain injury were younger and had shorter admission/rehabilitation intervals, greater functional and cognitive outcomes, and a higher frequency of returning home. In contrast, patients with anoxic brain injury achieved the lowest grade of functional and cognitive recovery.

Coma after Severe Traumatic Brain Injury

Therapeutic interventions following severe traumatic brain injury (TBI; GCS ≤ 8) should focus on blocking the primary injury process and preventing or reducing secondary brain injury. Primary injury in severe TBI includes hematoma formation, diffuse axonal injury after high-velocity injury (i.e., vehicle collisions, falls from great height), and ischemic brain injury due to traumatic vascular lesions (i.e., arterial dissection in the neck). Immediate arrest of ongoing primary injury (i.e., by acute recanalization of a dissected arterial occlusion or reversing anticoagulation during ongoing intracerebral bleeding) should be attempted whenever possible. Unfortunately, little can be done to treat or reverse the primary injury in TBI. Secondary brain injuries after TBI are common and include increased brain swelling leading to ICP elevations, brain tissue shifts (herniation) and critical reduction in cerebral perfusion pressure, formation of new hematoma, prolonged seizures, hypoxia, neurogenic/hypovolemic hypotension, uncontrolled hyperventilation, hypoglycemia, thromboembolism from venous thrombosis, and so on. Therefore, early TBI management strategies should focus on prevention and early recognition of such complications and a combination of intracranial (i.e., ICP, brain temperature, brain oxygen tension and perfusion) and systemic monitoring as well as repeated neuroimaging is indicated in the patients.

Noncontrast CT is the most common imaging evaluation tool in TBI patients. A cervical spine CT with reconstructed views and head and neck CT angiogram (CTA) are often added during initial evaluation, if injury to the neck and arteries, respectively, is suspected. Bone and brain windows should be reviewed separately and in experienced hands. Head imaging has a good sensitivity for the presence of pathologic air, skull fractures, hemorrhagic lesions, abnormalities of the brain parenchyma (i.e., shift), and obstruction of CSF spaces (hydrocephalus). Injury to the neuroanatomic structures of consciousness or the presence of additional etiologies for coma (i.e., alcohol) should be considered when disparities between CT findings and examination exist. In this setting, appropriate additional screening and further imaging diagnostics (i.e., MRI) must be initiated.¹¹³ Compared to CT, MRI has a higher sensitivity to detect diffuse axonal injury (DAI); however, this advantage has not translated into improved clinical outcomes as there is currently no treatment for this condition. However, both the detection of DAI and delineating injury of major motor (i.e., corticospinal) tracts on diffusion tensor imaging (DTI) can be utilized together with the clinical course to prognosticate TBI.¹¹⁴ MR spectroscopy (MRS) provides additional “chemical” information, especially of the white matter injury process.¹¹⁵

The presence and severity of secondary brain injury should constantly be assessed and managed throughout the acute phase of TBI. Its time course can be variable but is somewhat predictable: for example, ranging from hours (i.e., development of new hematoma) to 1 to 3 days for brain edema development to several days as seen in late brain ischemia from vasospasm due to traumatic subarachnoid blood. There is only limited evidence to support the use of ICP monitoring. ICP monitoring has evolved as an essential tool and also one of the earliest monitoring tools in clinical practice; thus many believe clinical trials cannot ethically be justified. Current TBI management guidelines recommend its use in severe TBI. Of note, head imaging cannot predict ICP. Furthermore, treating for elevated ICP (other than empirically in the emergency setting) and not monitoring ICP can be deleterious and result in a poor outcome.¹¹⁶ In addition to ICP, other brain monitoring techniques utilized in modern TBI management include jugular venous and transcranial oxygen saturations (SjvO₂; NIRS), brain tissue oxygen tension (btO₂), brain tissue metabolism (microdialysis) and temperature, direct brain perfusion, transcranial Doppler sonography, and electrophysiologic studies such as EEG and evoked potentials. A discussion of these techniques is beyond the scope of this chapter and the reader is referred to recent reviews.

The management of severe TBI includes emergent decompressive craniectomy in controlling intractable ICP, especially if deterioration is associated with hematoma evacuation and there is some evidence that there will be positive outcomes in surviving patients.117–128 Although timing of decompressive craniectomy remains controversial, early referral seems most intuitive. The Brain Trauma Foundation’s Guidelines for Management of Severe Traumatic Brain Injury regard craniectomy as an option after TBI to manage patients with intracerebral hemorrhage (ICH) and the resultant increased ICP refractory to medical management.¹¹⁶ Intravenous mannitol and hypertonic saline solutions are a main building block in the management of increased ICP.^129–141 The superiority of one over the other is uncertain, and both approaches are routinely used in the management of severe TBI.¹⁴² Therapeutic hypothermia is postulated to exert neuroprotective effects by several mechanisms, including reducing cerebral metabolic rate/requirement, modulating apoptosis or programmed cell death, reducing intracellular calcium/toxic excitatory neurotransmitters/inflammation, and preserving protein synthesis.^{116,143–151} However, the reported impact on neurologic outcome in severe TBI remains inconsistent and a meta-analysis of randomized controlled trials suggests its lack of benefit in severe TBI.^143,144 We utilize hypothermia as last resort therapy and commonly, together with pentobarbital therapy and ICP-matching, CPP elevation to protect the brain from otherwise intractable ICP elevations. Until well-designed, randomized controlled trials are available, its use must be weighed against potential side effects such as an increased incidence in pneumonia and coagulation abnormalities, among others.^143,151

The initial GCS seems to be a reasonable marker to predict mortality from severe TBI. To further improve prediction, models include a combination of clinical data, imaging results, GCS score, and patient demographics.¹⁵² In a large, traumatic coma databank study,¹⁵³ patients with severe injury (GCS 3 to 8) had approximately 36% mortality. In 170 patients, the GCS 5 hours postinjury was compared to the Glasgow Outcome Scale (GOS) at 1 month and positive recovery was seen in 99% of mild TBI (GCS 13 to 15), 71% of moderate TBI(GCS 9 to 12), and in only 35% of severe TBI. In contrast, only 2% to 3% of patients with mild TBI (GCS 13 to 15) died. Generally speaking, clinical outcome after TBI is more variable than in nontraumatic coma (i.e., after hypoxia).¹⁵⁴ Furthermore, outcome predictions for patients with intermediate GCS scores identify that a midline shift of less than 4.1 mm on initial CT scanning had a significantly higher favorable outcome rate compared with patients with a larger shift.¹⁵⁵

Biomarkers after traumatic brain injury (TBI), such as cellular glial fibrillary acidic protein (GFAP) and S100B released with astrocyte injury, may offer diagnostic and prognostic tools in addition to clinical indices. In 79 patients with TBI (GSC ≤ 12), identified intracranial mass lesion, pupillary reactivity, GFAP, and S100B were the strongest predictors of death and unfavorable outcome, with S100B as the strongest single predictor of unfavorable outcome with 100% discrimination.¹⁵⁶ Somatosensory evoked potentials in 58 TBI patients unconscious for >30 days were compared with outcome at 12 months¹⁵⁷ and showed an excellent prediction for outcome (persistent coma or MCS, 92% and 86%, respectively) at 12 months when combined with age, sex, and GCS.¹⁵⁷ Furthermore, among electrophysiologic testing, SEP and BAEP are more sensitive than EEG (i.e., 45% to 60% sensitive versus 35%) and if EEG and SEP show opposing results, generally speaking, prognosis is linked to the better test result.¹⁵⁸ The results of multimodality evoked potentials can be used as a reliable prognostic indicator (91% accuracy) in TBI patients.¹⁰¹

The International Mission for Prognosis and Analysis of Clinical Trials in TBI (IMPACT) is a prognostic model predicting 6-month functional outcome (Glasgow Outcome Scale [GOS]) and mortality in 587 patients with severe TBI (mean age 38 ± 17 years). At the 6th month, the median GOS was 3 (IQR 3) and 6-month mortality 41%, and age, motor score, pupillary reactivity, head CT findings, secondary insults, and laboratory values displayed good prediction ability for unfavorable outcome and mortality.¹⁵⁹ In another study of 428 isolated, older adult TBI cases, the in-hospital death rate was 28%, and increasing age, decreasing GCS, and injury type were significant independent predictors of in-hospital mortality. At the 6-month follow-up, age < 75 years and with systolic blood pressure (BP) on arrival at hospital of 131 to 150 mm Hg had higher likelihood of living independently at follow-up. No patients with a GCS < 9 had a favorable 6-month outcome, and most of them died, with the survival rate for brainstem injury at 21%.¹⁶⁰

Coma after Cardiac Arrest

Cardiac arrest affects more than 300,000 people each year in the United States alone and anoxic-ischemic encephalopathy is among the leading causes of mortality and morbidity. Since the publication of two pivotal trials in 2002 utilizing systemic cooling, the outcome for patients in coma after out-of-hospital ventricular fibrillation arrest has significantly improved. Simultaneously, prognostication after arrest in patients has become more complicated as the literature is lacking new prospective outcome studies integrating the advantages of hypothermia treatment.

The retrospective landmark study by Levy and colleagues,¹⁶¹ which reviewed postcardiac arrest outcome of 210 patients studied in the mid-1970s, is still commonly applied for prognostication. However, several inherent epidemiologic shortcomings of this study exist in addition to the fact that postcardiac arrest management has obviously advanced. The American Academy of Neurology published guidelines in 2006¹⁶² based on an in-depth review of the available literature; however, most studies included were performed without the use of hypothermia. The guideline defines poor outcome either as coma or death at 1 month or severe disability at 6 months. Further, the absence of brainstem reflexes or absence of extensor motor responses at 72 hours was shown to be highly specific (or have a false positive rate near zero) and so were absence of somatosensory evoked responses (SSEPs) and a high level of neuron-specific enolase (NSE).

Hypothermia has multiple actions not only within the CNS but also systemically. Among them are depression of neuronal activity, amelioration of intracellular postischemic reactions mitigating secondary cell injury, cardiac arrhythmias, hypocoagulability, decrease in immune defense, and reduction in drug metabolism and clearance. Notably, the latter can lead to misinterpretation of the prolonged effects of sedatives and paralytics. Therefore, interpretation of the neurologic examination and prognostication in patients undergoing hypothermia or rewarming is complex and challenging. Generally speaking and without the use of hypothermia, absence of brainstem reflexes implies that the cortex is severely affected as the brainstem neurons are more resistant to hypoxia-ischemia. Absence of pupillary reflexes at 72 hours predicts poor outcome with high specificity; however, the opposite, the presence of reflexive pupils, only predicts good outcome in about 21%. Studies performed after the introduction of hypothermia 163–165 delineate predictive findings for absent papillary responses but delineate that the corneal reflex seems to be more vulnerable to prolonged sedative and hypothermic effects. With respect to motor responses, studies performed prior to hypothermia identify a high predictive value for poor outcome for the absence of motor responses at 72 hours, a clinical finding that outperformed even SSEP prediction.¹⁶⁶ On the contrary, studies suggest that the lack of motor responses does not invariably predict poor outcome in patients treated with hypothermia (i.e., positive predictive values 92% and negative predictive value 85%) and exceptions have been reported.^{163–165,167}

Electroencephalography (EEG) is used in postcardiac arrest patients to determine outcome but also to differentiate myoclonus from shivering and detect seizures. Generalized EEG suppression, burst suppression with or without epileptiform activity, and a flat background with intermittent generalized periodic complexes are indicators of poorer prognosis. The opposite is true for finding reactivity and variability of the EEG pattern in patients treated with or without cooling. Larger prospective studies are currently lacking; however, some reports identify some hypothermia-treated patients with postanoxic status epilepticus who can achieve good outcome,¹⁶⁸ especially when the status evolves from a continuous EEG pattern and not from burst suppression.¹⁶⁹ As seizures and status epilepticus are seen in up to 65% of patients with less severe cortical damage, continuous EEG is indicated to guide management and prognostication in many postarrest patients.

According to current practice parameters from 2006, the absence of the N20 (primary cortical somatosensory) wave on SSEPs in the first 3 days after cardiac arrest identifies a high likelihood for poor outcome. In contrast, only about 41% of patients with present N20 wave will recover. Prognosticating based on SSEP findings in patients treated with hypothermia has only been reported in a few smaller studies with uncertain overall results 170,171; however, SSEP likely remains useful as evoked potentials are relatively resistant to metabolic and temperature derangements. Also, the NSE is a frequently used and validated biochemical marker used for prognostication. Levels >33 mcg/L at 72 hours or earlier were found to have zero false predictive rate in patients treated without hypothermia. However, hypothermia decreases NSE levels. Some studies have shown that a similar cutoff also holds true for the hypothermia-treated patient population,^172–174 whereas others found unacceptable false-positive rates for the cutoff value 33 mcg/L.^175,176 With respect to imaging, certain MRI findings such as severe cortical abnormalities on diffusion-weighted imaging have been shown to correlate to poor outcome,¹⁷⁷ but imaging studies often report on smaller patient numbers and are performed unblinded, reducing their value for prediction. Some studies identify that most patients who regained consciousness had normal cortical structures, some of them despite involvement of the deep gray nuclei.^178,179

Myoclonic status epilepticus are generalized, multifocal, repetitive muscle jerks involving some or all parts of the body. They are frequently seen within 24 hours after arrest, and EEG often shows burst suppression or bursts of generalized spikes. Whether myoclonic discharges actually represent seizures remains controversial, and as patients may have severe cortical damage and clinical myoclonus at the same time, it is postulated that myoclonus status epilepticus is subcortical in origin. In contrast, postanoxic status epilepticus is cortical in origin (but some overlap to myoclonic status may exist), whereas the posthypoxic myoclonus (Lance-Adams syndrome) is generally seen after primary respiratory failure. It is triggered by startle, often subsides over time, and the patient may be aware and distressed by it. Myoclonic status epilepticus has been associated with poor prognosis and remains poor even in the setting of hypothermia.163,165

Given these results it seems prudent to forecast the postarrest patient treated with hypothermia 72 hours after the event. Examination should take place at least 12 hours after termination of all sedatives and later if concomitant renal and liver damage are confounding factors. Prognostic variables associated with but not invariably predictive of poor outcomes are (1) absences of brainstem reflexes, (2) absence of motor responses or extensor posturing, (3) EEG abnormalities including generalized suppressions or periodic complexes on suppressed background and a burst suppression pattern with generalized epileptiform activity, (4) myoclonic status epilepticus, (5) MRI with widespread cortical injury on DWI or CT with loss of gray-white matter differentiation and sulcal effacement, (6) high NSE levels, and (7) absence of N20 on SSEP.

Ethical Considerations

A surrogate decision maker is a person who directs care when the patient is unable to provide consent. Legally and ethically, surrogate decisions should follow the expressed wishes of the patient. In the absence of known prior wishes, surrogates should utilize the best standard with respect to what the patient would do when confronted by the circumstance. A surrogate designated by the patient has precedence over other potential decision makers and is delineated by an advanced directive, also called a health care proxy, health care agent, or durable power of attorney for health care. In contrast, a living will expresses patient preferences but does not authorize a surrogate decision maker or spokesperson. If there is no designated surrogate, family members and close friends are selected in order of their relationship to the patient; the highest priority is given to the spouse, then parents, then children, then siblings followed by other relatives and finally friends. In a comatose patient, surrogates or family members may authorize a do no resuscitate (DNR) order or request limitations of care (i.e., no hemodialysis). Approximately two thirds of DNR orders are put into place by surrogate decision makers.¹⁸⁰

It is important to strive for the best possible diagnostic and prognostic correctness when discussing a DNR order or withdrawal of care. Understandably, relatives and surrogates place a great deal of decision making on the significance of conscious experience and the level of expected patient interactions with the environment (recovery). Therefore, the diagnostic accuracy of a patient’s present level of consciousness is requested from the treating physician as well as the best estimate of improvement of consciousness.¹⁸¹ Identifying medical futility in a patient with markedly impaired consciousness allows the surrogate and physician to withdraw life-supporting therapy. However, surrogate decision making may change if a patient is in a minimal conscious state, as this disorder implies different prognostication, especially in TBI patients.¹⁸² Further, it is important for physicians to reiterate to the family that coma progression into the vegetative state is not considered an improvement. This progression is seen in most coma patients and sleep-wake cycles, roving eye movements, startle, and other witnessed phenomena are merely reflexive and do not indicate consciousness or awareness. Of note, “persistence” is a diagnosis given to a patient who remains in a vegetative state for greater than 1 month. This state is considered permanent when it persists for 3 months after anoxic brain injury and 12 months after TBI.² Greater prognostic uncertainties exist for patients in a state of minimal consciousness, as these patients have consistent evidence of consciousness, awareness, and, in some, the ability to communicate. Because these patients have an unpredictable recovery, it is necessary to continue standardized assessment and therapy.

Physicians and care teams should utilize the latest medical evidence when discussing the probability of recovery from impaired consciousness with the family and surrogate decision maker. Further, it should be delineated that prognostication is inherently difficult. Outcome studies include a broader patient spectrum with a range of brain injury mechanisms. Furthermore, recovery is a dynamic and fluctuating process with marked individual variations. Such discussions require a delicate balance of perspectives on the need for clinical decision making and the inevitable fact of remaining medical uncertainties. Discussions with families and surrogates should include a thorough investigation about what the patient would have decided facing the inevitable scenario of permanent impaired (cognitive and physical) existence. This includes the importance of respecting a surrogate’s decision that ongoing life support would have been unacceptable with the patient’s wishes.

Key Points

• Coma is a state of pathologic unresponsiveness from which the patient cannot be aroused. The eyes are closed, and only reflex responses or no responses at all can be elicited from the patient when applying vigorous stimulation.

• The cause of coma is diverse and follows neuronal injury, reversible or irreversible, to both cerebral hemispheres or to the reticular activating system in the diencephalon and brainstem.

• A supratentorial mass legion located within one hemisphere can induce a lateral herniation syndrome with lateral torsion of the midbrain plus uncal herniation. In contrast a mass lesion located within the brain midline (i.e., acute hydrocephalus) primarily lead to downward herniation with rostocaudal compression of the brainstem.

• An acute mass lesion below the tentorium (i.e., a cerebellar hemorrhage) early on compresses the brainstem, leading to cranial nerve deficits, obstructed CSF outflow resulting in hydrocephalus, and, if large enough, herniation of brain tissue upward through the tentorium and downward through the foramen magnum. On examination these patients may present with a complete lack of all cranial nerve and motor function, yet hemispheric function may be preserved.

• Initial attention must focus on the restoration of respiratory, hemodynamic, and metabolic homeostasis in parallel with a search to identify the cause and acuity of the coma state. Generally, a comatose patient’s recovery depends on appropriate treatment of the underlying disorder but also very much on monitoring and stabilizing brain function (ICP, perfusion, etc.).

• Status epilepticus is characterized by either clinically detectable (convulsive) or EEG-identified (nonconvulsive) seizure activity. Both require immediate and aggressive treatment to avoid permanent brain damage.

• Alteration in cognitive function (especially attention and concentration deficit) is the earliest manifestation of metabolic encephalopathy. Some disease processes may provide specific diagnostic clues.

• The neurologic examination consists of an assessment of the level of arousal determined by eye opening, verbal responses, and reflex or purposeful movements in response to noxious stimulation of the face, arms, and legs. Neuro-ophthalmologic function is evaluated by spontaneous eye movements, pupillary size and response to light, oculocephalic responses (doll’s eyes), and oculovestibular (ice water caloric) responses. Autonomic function is assessed by the respiratory pattern and observing the monitored vital sign changes before and during examination.