The normal unconsciousness of sleep is characterized by prompt reversibility on threshold sensory stimulation, and maintenance of wakefulness following arousal. The degree of stimulation required depends on the stage of sleep (stage IV non–rapid eye movement sleep is the deepest) and the sensory stimulation used.

Consciousness is deemed depressed when suprathreshold sensory stimulation is required for arousal and wakefulness cannot be maintained unless the stimulation is continuous [1,2]. Responsible specific lesions involve the ARAS or both cerebral hemispheres; the former by brainstem damage, or compression due to masses situated in other compartments, and the latter by multifocal insults or unilateral lesions with associated major mass effect. In addition, a wide array of metabolic derangements, toxins, or diffuse injuries may depress consciousness by affecting the ARAS, the cerebral hemispheres, or both. The spectrum of depressed states—lethargy, hypersomnolence, obtundation, stupor, and coma—is defined by the level of consciousness observed on examination. The etiologies are diverse (Table 169.1), with the degree of depression dependent on the nature of the insult, its duration, and the location and extent of the brain injury.

The first signs of brain dysfunction may be mild and barely noticeable. The patient may be described initially as confused or drowsy before progressing to lethargy or hypersomnolence and eventually to a more depressed state. Hypersomnolent patients maintain arousal only with vigorous and continuous sensory stimulation; while awake, however, they may be oriented and make appropriate responses. The most common cause of hypersomnolence in the hospital is sleep deprivation, mostly iatrogenic, especially in the around-the-clock care setting of the intensive care unit (ICU). Patients with discrete diencephalic or midbrain tegmentum lesions may also present with hypersomnolence [3,4]. Because these lesions affect the ARAS and spare the cerebral hemispheres, cognitive content is usually preserved. Rostral extension of a midline lesion may involve thalamic structures (especially the dorsomedial nuclei) and cause difficulties with the ability to store new memories. Other mesencephalic structures may be affected and cause abnormalities of pupillary function, internuclear ophthalmoplegia, and third nerve dysfunction.

Obtunded patients usually can be aroused by light stimuli but are mentally dulled and unable to maintain wakefulness. Stuporous patients can be aroused only with vigorous noxious stimulation. While awake, neither obtunded nor stuporous patients demonstrate a normal content of consciousness, but both may display purposeful movements, attempting to ward off painful stimuli or to remove catheters, endotracheal tubes, or intravenous lines.

Patients in coma are unresponsive to suprathreshold sensory stimulation, including noxious stimulation that is strong enough to arouse a deeply sleeping patient but not strong enough to cause physical injury. Although the patient usually lies motionless, movements such as stereotyped, inappropriate postures (decerebration and decortication) and spinal cord reflexes (triple flexion and Babinski responses) may occur. Whatever the etiology, the duration of coma is typically no longer than 2 to 4 weeks, after which one of the three conditions supervenes: arousal to full or partial recovery, a vegetative state, or death.

Most of the literature on prognosis of comatose patients comes from nontraumatic coma, largely anoxic–ischemic brain injury. A landmark paper by Levy, Plum, and associates from 1981 established the neurological examination – particularly brainstem reflexes including pupillary, corneal, and oculocephalic reflexes – as important predictors of poor outcome in nontraumatic coma [5]. Multiple studies followed which confirmed the importance of motor responses in addition to brainstem examination, and some diagnostic tests were established as useful in predicting outcomes; these are well summarized in the American Academy of Neurology Practice Parameter by Wijdicks et al., published in 2006 [6]. Given the life-or-death responsibility of the physician providing a prognosis, only clinical indicators or diagnostic tests that are highly specific with a near zero false–positive rate are utilized. A poor outcome is predicted by the absence of pupillary and corneal reflexes, absent or extensor motor responses, absent responses to caloric testing of the oculovestibular reflex at day 3 post-arrest, and the presence of myoclonic status epilepticus on day 1 post-arrest. The absence of N20 responses on somatosensory evoked potential (SSEP) testing, and the finding of serum neuron-specific enolase levels more than 33 μg per L on days 1 to 3 post-arrest also indicate a poor prognosis (Fig. 169.1). Prognostication must include consideration of the etiology of the disease process, the clinical examination findings, and radiological evidence of damage to the upper pons, midbrain, diencephalon, and other vital structures for arousal.

Patients in psychogenic coma appear comatose but have clinical and laboratory evidence of wakefulness [1]. Psychogenic unresponsiveness may be suggested by active resistance or rapid closure of the eyelids, pupillary constriction to visual threat, fast phase of nystagmus (i.e., a saccade) on oculovestibular or optokinetic testing, and avoidance of self-injury (e.g., by averting an arm dropped toward the patient’s face) or annoying stimulation such as a nasal tickle (moving head away from stimulus). Caloric testing with ice water irrigation of the ear will elicit a
normal nystagmoid response with the fast or corrective component directed away from the irrigated ear and possibly some nausea and vomiting. Deep tendon reflex examination is often normal but can be voluntarily suppressed. EEG alpha waves that attenuate with eye opening are inconsistent with coma or sleep. Most diagnostic tests will be unrevealing. Psychiatric conditions that may be associated with psychogenic coma are conversion reactions secondary to hysterical personality, severe depression, or acute situational reaction, catatonic schizophrenia, dissociative or fugue states, severe psychotic depression, and malingering.

The locked-in state is a nearly total paralysis without loss of consciousness [7,8]. Because the most common cause of this state is destruction of the base of the pons, the patient is completely paralyzed except for muscles subserved by midbrain structures (i.e., vertical eye movements and blinking). Consciousness is preserved because the ARAS is located in the tegmentum of the pons, dorsal to the damaged area. The most frequent cause is cerebrovascular such as cerebral infarction from a basilar thromboembolism or pontine hemorrhage from uncontrolled hypertension; less frequent etiologies of the syndrome are acute polyneuropathy (Guillain-Barré syndrome), acute poliomyelitis, toxins that block transmission at the neuromuscular junction, and myasthenia gravis. It is important to note that locked-in patients are capable of hearing, seeing, and feeling external stimuli and pain. Adequate analgesia and anxiolysis should be provided despite the absence of external signs of pain and anxiety. A 5- to 10-year survival has been reported in as high as 80% of patients in some series and a surprising 58% of patients surveyed reported satisfaction with life despite their disability in a small case series [8].

The term brain death refers to a determination of physical death by brain-based, rather than cardiopulmonary-based, criteria [9]. Brain death is the irreversible destruction of the brain, with the resulting total absence of all cortical and brainstem function, although spinal cord reflexes may remain [10,11]. It is not to be confused with severe but incomplete brain damage with a poor prognosis or with a vegetative state, conditions in which some function of vital brain centers still remains. In brain death, support of other organs is futile for the patient, whereas when there is some residual brain or brainstem function, or a vegetative state, decisions regarding ongoing life support clearly depend on the wishes of the patient or his or her proxy.

In brain death, pupils are mid-position and round (not oval), and apnea persists even when arterial carbon dioxide tension (PCO₂) is raised to levels that should stimulate respiration. Table 169.2 summarizes the guidelines used in the United States. Brain death may be simulated by drug intoxications and cannot be evaluated when toxic drugs are present; depending on preserved renal and hepatic function most such toxic
effects do not persist longer than 36 hours. Hypothermia also precludes a diagnosis of brain death, and the patient must be brought to normal temperature prior to declaring death. Brain death is a clinical diagnosis, but ancillary tests such as an EEG and blood flow studies (transcranial Doppler, technetium-99 m scan, or conventional cerebral angiography) may be useful where the clinical examination is compromised by sedating medications. Unresponsiveness that can mimic brain death may occur with extensive brainstem destruction, for example, after basilar artery thrombosis. Despite absent brainstem reflexes, continued cortical activity on the EEG and persistent cerebral blood flow would demonstrate that the patient is not brain dead.

The American Academy of Neurology has published practice parameters for the determination of brain death. The criteria take into account etiology, performance of two separate clinical examinations 6 hours apart, and include the method of apnea testing with preoxygenation and oxygen [11]. Since criteria for brain death vary from state to state, and procedures to determine brain death differ among institutions, it is important to be familiar with the guidelines in your institution [12]. The occurrence of brain death provides the opportunity for organ donation, and most institutions have a protocol that includes informing organ bank organizations to facilitate this.

When the cerebral hemispheres are insulted by toxic, metabolic, anoxic, structural, or infectious processes, the patient may appear acutely confused [13,14]. Poor arousal and an abnormal content of consciousness may contribute to the clinical presentation, and the etiologies are legion (Table 169.3). Patients with clouded consciousness are easily distracted or startled by environmental stimuli. Their processing of information is slow and effortful, arousal fluctuates from drowsiness to hyperexcitability, and poor attention span impairs recall and recent memory. If sensorial clouding becomes more advanced, sensory input is increasingly misinterpreted, daytime drowsiness alternates with nocturnal agitation, disorientation for place and time becomes apparent, and repeated prompting is required for a response to even the simplest commands.

Delirious patients typically manifest acutely fluctuating confusion, with psychomotor overactivity, agitation, autonomic instability, and often visual hallucinations. Clinical observations frequently suggest that the disturbance of cognition or perception is directly related to a potentially reversible general medical condition rather than to an evolving dementia. Hyperexcitability may alternate with periods of drowsiness or relative lucidity. Signs of autonomic overactivity include pupillary dilatation, diaphoresis, tachycardia, and hypertension. Patients with delirium may not sleep, sometimes for periods of several days; the success of treatment can be judged by the development of normal sleep. Delirium tremens, the most serious consequence of ethanol withdrawal, is perhaps the best-known example of this state. Because the routine Mini-Mental State Examination often cannot be administered to unstable, intubated patients, alternative screening tools have been developed for early detection and monitoring of delirium in the ICU [15,16]. Validated tools such as the Confusion Assessment Method, or CAM-ICU scale, have the advantage of being simple and easy to administer, highly reliable and applicable in patients who are intubated. Systematic screening may help detect early delirium and allow prompt, cost-effective treatment. Delirium has been linked to prolonged ICU stay and ventilator days, and is associated with postdischarge cognitive dysfunction and worse 6-month mortality outcomes [16,17]. The use of interventions that reduce delirium in the ICU include reduction and intermittent use of sedatives, or spontaneous awakening trials, as well as sedation with alpha adrenergic medications such as dexmedetomidine [18,19].

In beclouded dementia, confusion is superimposed on an underlying subacute or chronic cognitive disorder. The preexisting cerebral dysfunction may be mental retardation, dementia, or the deficits from a vascular, neoplastic, or demyelinative process. In some cases, the underlying disorder is not diagnosed until the confusion appears during an intercurrent illness (e.g., sepsis or infection, congestive heart failure, surgical procedures, anemia, drug overdose, or intolerance).

Patients with dementia have subacute or chronic intellectual dysfunction unaccompanied by a reduction in arousal [20]. The patient exhibits a decline in multiple cognitive functions, including memory, language, spatial orientation, personality, abstract thinking, and insight. The ability to carry out testing requires relative preservation of attention and language
comprehension. The causes of dementia include degenerative processes (Alzheimer’s disease, Pick’s disease, Huntington’s disease), metabolic and nutritional disorders (hypothyroidism, pellagra, vitamin B₁₂ deficiency), infectious diseases (subacute spongiform encephalopathy, acquired immunodeficiency syndrome dementia, neurosyphilis, chronic meningitis, progressive multifocal leukoencephalopathy), cerebrovascular disorders (multi-infarct dementia, anoxia-ischemia), hydrocephalus with normal or increased intracranial pressure, and toxins.