Laboratory examinations are highly informative in the investigation of the acquired metabolic diseases. In patients with symptoms suggestive of a metabolic encephalopathy the following determinations are usually made: serum Na, K, Cl, Ca, Mg, glucose, HCO3, renal function tests (blood urea nitrogen [BUN] and creatinine), liver function tests (aspartate aminotransferase [AST], alanine aminotransferase [ALT], bilirubin, NH3), thyroid function tests (T4 and thyroid-stimulating hormone [TSH]), and osmolality and, in certain cases, oxygen saturation and blood gas determinations. These are almost always supplemented by toxicology tests and measurement of the serum concentrations of relevant medications as discussed in the next chapter. Serum osmolality can be measured directly or calculated from the values of Na, glucose, and BUN (in mg/dL), using the following formula:
Here the basic disorder is a lack of oxygen and of blood flow to the brain, the result of failure of the heart and circulation or of the lungs and respiration. Often, both are responsible and one cannot say which predominates; hence the dually ambiguous allusions in medical records to "ischemic-hypoxic" encephalopathy. This combined encephalopathy in various forms and degrees of severity is one of the most frequent and disastrous cerebral disorders encountered in every general hospital.
Reduced to the simplest formulation, a deficient supply of oxygen to the brain is either the result of a failure of cerebral perfusion (ischemia) or of a reduced amount of circulating arterial oxygen, of diminished oxygen saturation, or of insufficiency of hemoglobin (hypoxia). Although they are often combined, the neurologic effects of ischemia and hypoxia are subtly different. The medical conditions that most often lead to it are as follows:
A global reduction in cerebral blood flow (myocardial infarction, ventricular arrhythmia, aortic dissection, external or internal blood loss, and septic or traumatic shock)
Hypoxia from suffocation (drowning, strangulation, or aspiration of vomitus, food, or blood; from compression of the trachea by a mass or hemorrhage; tracheal obstruction by a foreign body, or a general anesthesia accident)
As a subset of the above, diseases that paralyze the respiratory muscles (Guillain-Barré syndrome, amyotrophic lateral sclerosis, myasthenia, and, in the past, poliomyelitis) or damages the medulla and leads to failure of breathing
The special case of carbon monoxide (CO) poisoning (nonischemic hypoxia)
The product of blood oxygen content and the cardiac output is the ultimate determinant of the adequacy of oxygen supply to the organs. When blood flow is stable, the most important element in the delivery of oxygen is the oxygen content of the blood. This is the product of hemoglobin concentration and the percentage of oxygen saturation of the hemoglobin molecule. At normal temperature and pH, hemoglobin is 90 percent saturated at an oxygen partial pressure of 60 mm Hg and still 75 percent saturated at 40 mm Hg; i.e., as is well known, the oxygen saturation curve is not linear.
Physiology of Ischemic and Hypoxic Damage
A number of physiologic mechanisms of a homeostatic nature protect the brain under conditions of both ischemia and hypoxia. Through a mechanism termed autoregulation, there is a compensatory dilatation of resistance vessels in response to a reduction in cerebral perfusion, which maintains blood flow at a constant rate, as noted in Chap. 34. When the cerebral blood pressure falls below 60 to 70 mm Hg, an additional compensation in the form of increased oxygen extraction allows normal energy metabolism to continue. In total cerebral ischemia, the tissue is depleted of its sources of energy in about 5 min, although longer periods are tolerated under conditions of hypothermia. Also, energy failure because of hypoxia is counteracted by an autoregulatory increase in cerebral blood flow; at a PO2 of 25 mm Hg, the increase in blood flow is approximately 400 percent. A similar increase in flow occurs with a decrease in hemoglobin to 20 percent of normal.
In most clinical situations in which the brain is deprived of adequate oxygen, as already commented, there is a combination of ischemia and hypoxia, with one or the other predominating. The pathologic effects of ischemic brain injury from systemic hypotension differ from those caused by pure anoxia. Under conditions of transient ischemia, one pattern of damage takes the form of incomplete infarctions in the border zones between major cerebral arteries (Chap. 34). With predominant anoxia, neurons in portions of the hippocampus and the deep folia of the cerebellum are particularly vulnerable. More severe degrees of either ischemia or hypoxia, or the combination, lead to selective damage to certain layers of cortical neurons, and if more profound, to generalized damage of all the cerebral cortex, deep nuclei, and cerebellum. The nuclear structures of the brainstem and spinal cord are relatively resistant to anoxia and hypotension and stop functioning only after the cortex has been badly damaged.
The cellular pathophysiology of neuronal damage under conditions of ischemia is discussed in Chap. 34. One mechanism of injury is an arrest of the aerobic metabolic processes necessary to sustain the Krebs (tricarboxylic acid) cycle and the electron transport system. Neurons, if completely deprived of their source of energy, are unable to maintain their integrity and undergo necrosis. However, neuronal cell death occurs through more than one mechanism. The most acute forms of cell death are characterized by massive swelling and necrosis of neuronal and nonneuronal cells (cytotoxic edema). Short of immediate ischemic necrosis, a series of internally programmed cellular events may also propel the cell toward death in a delayed fashion, a process for which the term apoptosis has been borrowed from embryology. There is experimental evidence that certain excitatory neurotransmitters, particularly glutamate, contribute to the rapid destruction of neurons under conditions of anoxia and ischemia (Choi and Rothman); the pertinence of these effects to clinical situations is uncertain. Ultimately, this process may be affected by massive calcium influx through a number of different membrane channels, which activates various kinases that participate in the process of gradual cellular destruction. Free radical generation appears to play a role in membrane dissolution as a result of these processes. As shown in experimental models, one of the reasons for the irreversibility of ischemic lesions may be swelling of the endothelium and blockage of circulation into the ischemic cerebral tissues, the "no-reflow" phenomenon described by Ames and colleagues. There is also a poorly understood phenomenon of delayed neurologic deterioration after anoxia; this may be a result of the blockage or exhaustion of some enzymatic process during the period when brain metabolism is restored.
Clinical Features of Anoxic Encephalopathy
Mild degrees of hypoxia without loss of consciousness induce only inattentiveness, poor judgment, and incoordination; in our experience, there have been no lasting clinical effects in such cases, although Hornbein and colleagues found a slight decline in visual and verbal long-term memory and mild aphasic errors in Himalayan mountaineers who had earlier ascended to altitudes of 18,000 to 29,000 ft. These observations make the point that profound anoxia may be well tolerated if arrived at gradually. For example, we have seen several patients with advanced pulmonary disease who were fully awake when their arterial oxygen pressure was in the range of 30 mm Hg. This level, if it occurs abruptly, causes coma. An important derivative observation is that degrees of hypoxia that at no time abolish consciousness rarely, if ever, cause permanent damage to the nervous system.
In the circumstances of severe global ischemia with prolonged loss of consciousness, the clinical effects can be quite variable. Following cardiac arrest, for example, consciousness is lost within seconds but recovery can be complete if breathing, oxygenation, and cardiac action are restored within 3 to 5 min. Beyond 5 min there is usually permanent injury. Clinically, however, it is often difficult to judge the precise degree and duration of ischemia, because slight heart action or an imperceptible blood pressure may have served to maintain the circulation to some extent. Hence some individuals have made an excellent recovery after cerebral ischemia that apparently lasted 8 to 10 min or longer. Subnormal body temperatures, as might occur when the body is immersed in ice-cold water, greatly prolong the tolerable period of hypoxia. This has led to the successful application of moderate cooling after cardiac arrest as a technique to limit cerebral damage (see further on in Treatment of Hypoxic-Ischemic Encephalopathy section).
Generally speaking, anoxic patients who demonstrate intact brainstem function as indicated by normal pupillary light and ciliospinal responses, induced by passive head turning (doll's eye movements), and other vestibulo-ocular reflexes have a more favorable outlook for recovery of consciousness and perhaps of all mental faculties. Conversely, the absence of these brainstem reflexes even after circulation and oxygenation have been restored, particularly pupils that fail to react to light, implies a grave outlook as elaborated further on. If the damage is almost total, coma persists, decerebrate postures may be present spontaneously or in response to painful stimuli, and bilateral Babinski signs can be evoked. In the first 24 to 48 h, death may terminate this state in a setting of rising temperature, deepening coma, and circulatory collapse, or the syndrome of brain death intervenes, as discussed below.
Most patients who have suffered severe but lesser degrees of hypoxia will have stabilized their breathing and cardiac activity by the time they are first examined; yet they are comatose, with the eyes slightly divergent and motionless but with reactive pupils, the limbs inert and flaccid or intensely rigid, and the tendon reflexes diminished. Within a few minutes after cardiac action and breathing have been restored, generalized convulsions and isolated or grouped myoclonic twitches may occur. Either of these phenomena are poor prognostic signs. With severe degrees of injury, the cerebral and cerebellar cortices and parts of the thalami are partly or completely destroyed but the brainstem-spinal structures survive. Tragically, the individual may survive for an indefinite period in a state that is variously referred to as cortical death, irreversible coma, or persistent vegetative state (see discussion of these subjects in Chap. 17). Some patients remain mute, unresponsive, and unaware of their environment for weeks, months, or years. Long survival is usually attended by some degree of improvement but the patient appears to know nothing of his present situation and to have lost all past memories, cognitive function, and capacity for meaningful social interaction and independent existence (a minimally conscious state, actually a severe dementia; see Chap. 17). One has only to observe such patients and their families to appreciate the gravity of the problem, the family's anguish, and the tremendous expense of medical care. The only person who does not appear to suffer is the patient.
With lesser degrees of anoxic-ischemic injury, the patient improves after a period of coma lasting hours or less. Some of these patients quickly pass through this acute post-hypoxic phase and proceed to make a full recovery; others are left with varying degrees of permanent disability.
The findings on imaging studies vary. The most common early change in cases of severe injury is a loss of the distinction between the cerebral gray and white matter (Fig. 40-1); patients with this finding are invariably comatose and few awaken with a good neurologic outcome. With less severe and predominantly hypotensive-ischemic events such as cardiac arrest, watershed infarctions become evident in the border zones between the anterior, middle, and posterior cerebral arteries (Fig. 40-2). The clinical syndromes associated with watershed infarction are discussed below. Yet another pattern of brain destruction, seen at times also in CO poisoning, consists of striatal damage that is evident more by imaging than by clinical features (Fig. 40-3).
CT without contrast infusion after 1 day cardiac arrest demonstrating the loss of distinction between gray and white matter throughout the cerebral hemispheres. The patient remained comatose and became vegetative.
Watershed infarction between the middle and posterior cerebral arteries after brief cardiac arrest. The patient had Balint syndrome.
T2-weighted MRI of striatal damage after anoxia from hanging. The pallidum is spared, in contrast to typical cases of carbon monoxide poisoning (see Fig. 40-5).
(See Chap. 17 for a full discussion)
This represents the most severe degree of hypoxia, usually caused by circulatory arrest; it is manifest by a state of complete unawareness and unresponsiveness with abolition of all brainstem reflexes. Natural respiration cannot be sustained; only cardiac action and blood pressure are maintained. No electrical activity is seen in the EEG (it is isoelectric). At autopsy one finds that most, if not all, the gray matter of cerebral, cerebellar, and brainstem structures—and in some instances, even the upper cervical spinal cord—has been severely damaged.
One must always exercise caution in concluding that a patient has this form of irreversible brain damage, because anesthesia, intoxication with certain drugs, and hypothermia may also cause deep coma and an isoelectric EEG but permit recovery. Therefore it is often advisable to repeat the clinical and laboratory tests after an interval of a day or so, during which time the results of toxic screening also become available. The authors' experience corroborates the general notion that the vital functions of patients with the brain death syndrome usually cannot be sustained for more than several days; in other words, the problem settles itself. In exceptional cases, however, the provision of adequate fluid, vasopressors, and respiratory support allows preservation of the body in a comatose state for longer periods.
Posthypoxic Neurologic Syndromes
The permanent neurologic sequelae or posthypoxic syndromes observed most frequently are as follows:
Persistent coma or stupor, described above
With lesser degrees of cerebral injury, dementia with or without extrapyramidal signs
Extrapyramidal (parkinsonian) syndrome with cognitive impairment (discussed in relation to CO poisoning)
Intention or action myoclonus (Lance-Adams syndrome)
An amnesic state
If hypoperfusion dominates, the patient may also display the manifestations of watershed infarctions that are situated between the end territories of the major cerebral vessels. The main syndromes that become evident soon after the patient awakens are:
Visual agnosias including Balint syndrome and cortical blindness (Anton Syndrome) (see Chap. 22), representing infarctions of the watershed between the middle and posterior cerebral arteries (see Fig. 40-2)
Proximal arm and shoulder weakness, sometimes accompanied by hip weakness (referred to as a "man-in-the-barrel" syndrome), reflecting infarction in the territory between the middle and anterior cerebral arteries. These patients are able to walk, but their arms dangle and their hips may be weak.
The two watershed syndromes may rarely coexist. The interested reader may consult the appropriate chapter in the text on neurologic intensive care by Ropper and colleagues for further details. There are also watershed areas in the spinal cord (Chap. 44).
Seizures may or may not be a problem, and they are often resistant to treatment. Well-formed motor convulsions are infrequent. Myoclonus is more common and may be intermixed with fragmentary convulsions. Myoclonus is a grave sign in most cases but it generally recedes after several hours or a few days. These movements are also difficult to suppress, as noted further on.
Delayed Postanoxic Encephalopathy and Leukoencephalopathy
This is a relatively uncommon and unexplained phenomenon. Initial improvement, which appears to be complete, is followed after a variable period of time (1 to 4 weeks in most instances) by a relapse, characterized by apathy, confusion, irritability, and occasionally agitation or mania. Most patients survive this second episode, but some are left with serious mental and motor disturbances (Choi; Plum et al). In still other cases, there appears to be progression of the initial neurologic syndrome with additional weakness, shuffling gait, diffuse rigidity and spasticity, sphincteric incontinence, coma, and death after 1 to 2 weeks. Exceptionally, there is yet another syndrome in which an episode of hypoxia is followed by slow deterioration, which progresses for weeks to months until the patient is mute, rigid, and helpless. In such cases, the basal ganglia are affected more than the cerebral cortex and white matter as in the case studied by our colleagues Dooling and Richardson. Instances have followed cardiac arrest, drowning, asphyxiation, and carbon monoxide poisoning.
The imaging features of the white matter disorder can be quite striking (Fig. 40-4). A mitochondrial disorder has been suggested, on uncertain grounds, as the underlying mechanism.
MRI fluid-attenuated inversion recovery (FLAIR) sequence of delayed postanoxic leukoencephalopathy in a patient who had recovered after drowning and deteriorated 2 weeks later.
Prognosis of Hypoxic-Ischemic Brain Injury (See also "Prognosis in Coma" in Chap. 17)
Several validated models have been developed to predict the outcome of anoxic-ischemic coma. All of them incorporate simple clinical features involving loss of motor, verbal, and pupillary functions in various combinations. The most often cited study of the prognostic aspects of coma following cardiac arrest is the one by Levy and colleagues of 210 patients, which provided the following guidelines: 13 percent of patients attained a state of independent function within 1 year; at the time of the initial evaluation, approximately 25 percent of patients had absent pupillary light reflexes, none of whom regained independent function; by contrast, the presence on admission of reactive pupils, eye movements, and any motor response, a configuration displayed in approximately 10 percent, was associated with a better prognosis in almost 50 percent of cases. The absence of neurologic function in any of these spheres at 1 day after cardiac arrest, unsurprisingly, was associated with an even poorer outcome. Similarly, Booth and colleagues analyzed previously published studies and determined that 5 clinical signs at 1 day after cardiac arrest predicted a poor neurologic outcome or death: (1) absent corneal responses, (2) absent pupillary reactivity, (3) no withdrawal to pain, and (4) the absence of any motor response. The use of somatosensory evoked potentials in the prognostication of coma is discussed in Chaps. 2 and 17.
Most workers in the field of coma studies have been unable to establish signs that confidently predict a good outcome. The role of somatosensory evoked potentials in prognosis of coma has been addressed in Chap. 17.
In any such case, concurrent intoxication must, of course, be excluded.
The question of what to do with patients in such states of protracted coma is a societal as much as a medical problem. The neurologist can be expected to state the level and degree of brain damage, its cause, and the prognosis based on his own and published experience. One prudently avoids heroic, lifesaving therapeutic measures once the nature of this state has been determined with certainty.
Treatment of Hypoxic-Ischemic Encephalopathy
Treatment is directed initially to the prevention of further hypoxic injury. A clear airway is secured, cardiopulmonary resuscitation is initiated, and every second counts in their prompt utilization. Oxygen may be of value during the first hours but is probably of little use after the blood becomes well oxygenated. Once cardiac and pulmonary function are restored, there is experimental and clinical evidence that reducing cerebral metabolic requirements by inducing hypothermia may have a slight beneficial effect on outcome and may prevent the delayed worsening referred to above, though a recent clinical trial brings this into question (see further on). The use of high-dose barbiturates has not met with the same success.
Much attention was drawn to the randomized trials conducted by Bernard and colleagues and by the Hypothermia After Cardiac Arrest Study Group, of mild hypothermia applied to unconscious patients immediately after cardiac arrest. They reduced the core temperature to 33°C (91°F) within 2 h of the arrest and sustained this level for 12 h in the first trial, and between 32°C and 34°C for 24 h in the second study. Both trials demonstrated improved survival and better cognitive outcome in survivors, compared to leaving the patient in a normothermic state and this led to the development of guidelines and a change in clinical practice in the U.S. and elsewhere after 2002. The outcomes were evaluated by coarse measures of neurologic function. Implementing and sustaining hypothermia, either by external cooling, infusion of cooled normal saline, or intravenous cooling devices is difficult, and the iatrogenic problems of hypotension, bleeding, ventricular ectopy and infection have sometimes arisen, although this mild degree of temperature reduction is usually well tolerated. A third larger trial conducted by Nielsen and colleagues compared temperature maintenance after cardiac arrest at 33°C to maintenance of 36°C and found no difference in the rate of death or in neurological outcome. The results of this third study are still being discussed and it is not clear if it should be interpreted as demonstrating that hypothermic treatment is ineffective or if the avoidance of even mild hyperthermia, observed in the control groups of previous trials, was the important factor in improving outcome. At the time of this writing, it seems to us that induced hypothermia is not obligatory after cardiac arrest but that attempts should be made to keep the body temperature from rising above normal.
Vasodilator drugs, glutamate blockers, opiate antagonists, and calcium channel blockers have been of no proven benefit despite their theoretical appeal and some experimental successes. Corticosteroids ostensibly help to allay brain (possibly cellular) swelling, but, again, their therapeutic benefit has not been evident in clinical trials.
Seizures should be controlled by the methods indicated in Chap. 16. If convulsions are severe, continuous, and unresponsive to the usual medications, continuous infusion of a drug such as midazolam or propofol, and eventually the suppression of convulsions with neuromuscular blocking agents may be required. Often the seizures cease after a few hours and are replaced by polymyoclonus. For the latter, clonazepam, 8 to 12 mg daily in divided doses may be useful but the commonly used antiepileptic drugs have little effect. A state of spontaneous and stimulus-sensitive myoclonus as well as persistent limb posturing usually presages a poor outcome. The striking disorder of delayed movement-induced myoclonus and ataxic tremor that appear after the patient awakens from an anoxic episode (Lance-Adams myoclonus) is a special issue, which is discussed in Chap. 6. Its treatment usually requires the use of multiple medications. Fever is treated with antipyretics or a cooling blanket combined with neuromuscular paralyzing agents.
Carbon Monoxide Poisoning
Strictly speaking, CO is an exogenous toxin, but it is considered here because it produces a characteristic cerebral injury and is frequently associated with delayed neurologic deterioration. The extreme affinity of CO for hemoglobin (more than 200 times that of oxygen) drastically reduces the oxygen content of blood and subjects the brain to prolonged hypoxia and acidosis. Cardiac toxicity and hypotension generally follow. Whether CO also has a direct toxic action on neuronal components is not settled. The effects on the brain for the most part simulate those caused by cardiac arrest. Neurologists are likely to encounter instances of CO poisoning in burn units and in patients who have attempted suicide or have been exposed accidentally to a faulty furnace or to car exhaust in a closed garage. A contemporary review of the subject has been given by Weaver.
Early symptoms include headache, nausea, dyspnea, confusion, dizziness, and clumsiness. These occur when the carboxyhemoglobin level reaches 20 to 30 percent of total hemoglobin. Exposure to relatively low levels of CO from faulty furnaces and gasoline engines should be suspected as the cause of recurrent headaches and confusion that clear upon hospitalization or other change of venue. A cherry-red color of the skin may appear, but is actually an infrequent finding; cyanosis is more common. At slightly higher levels of carboxyhemoglobin, blindness, visual field defects, and papilledema develop, and levels of 50 to 60 percent are associated with coma, decerebrate or decorticate posturing, seizures in a few patients, and generalized slowing of the EEG rhythms. The initial CT scanning is normal or shows mild cerebral edema; later scans may show a characteristic lesion in the pallidum, as described below. Only if there has been associated hypotension does one see the same types of vascular border-zone infarctions that appear after cardiac arrest.
Delayed neurologic deterioration 1 to 3 weeks (sometimes much longer) after CO exposure occurs more frequently than with other forms of cerebral hypoxia. In Choi's survey, this feature was observed in 3 percent of 2,360 cases of CO poisoning and in 12 percent of those ill enough to be admitted to a hospital. Extrapyramidal features (parkinsonian gait and bradykinesia) predominated. Three-quarters of such patients were said to recover within a year. Discrete lesions centered in the globus pallidus bilaterally and sometimes the inner portion of the putamina are characteristic of CO poisoning that had produced coma (Fig. 40-5), but similar focal destruction may be seen after drowning, strangulation, and other forms of anoxia. The common feature among the delayed-relapse patients is a prolonged period of pure anoxia (before the occurrence of ischemia). Basal ganglia lesions may be quite prominent on CT scans even when delayed neurologic sequelae do not occur but they are invariably present between 1 and 4 weeks in patients who develop the delayed extrapyramidal syndrome. In less-severely affected patients we have seen such lesions resolve entirely on CT and MRI and there is no resultant movement disorder.
Unenhanced CT of the brain of a 30-year-old woman who attempted suicide by carbon monoxide inhalation. The only neurologic residua were a mild defect in retentive memory and areas of decreased attenuation in the pallidum bilaterally (arrows).
The initial treatment for carbon monoxide exposure is with inspired oxygen. Because the half-life of CO (normally 5 h) is greatly reduced by the administration of hyperbaric oxygen at 2 or 3 atmospheres, this additional treatment is recommended when the carboxyhemoglobin concentration is greater than 40 percent or in the presence of coma or seizures (Myers et al). According to a trial conducted by Weaver and colleagues, this treatment reduces the incidence of cognitive sequelae from 46 to 25 percent. They administered 3 hyperbaric sessions in the first 24 h after exposure to CO.
High-Altitude (Mountain) Sickness
Acute mountain sickness is another special form of cerebral hypoxia. It occurs when a sea-level inhabitant abruptly ascends to a high altitude. Headache, anorexia, nausea and vomiting, weakness, and insomnia appear at altitudes above 8,000 ft; on reaching higher altitudes, there may be ataxia, tremor, drowsiness, mild confusion, and hallucinations. At 16,000 ft, according to Griggs and Sutton, 50 percent of individuals develop asymptomatic retinal hemorrhages, and it has been suggested that such hemorrhages also occur in the cerebral white matter. Extreme altitude sickness may result in fatal cerebral edema. The overexpression of vascular endothelial growth factor (VEGF), a protein originally noted for its effects on vascular permeability, has been implicated as the cause of cerebral edema in the experiments of Schoch and colleagues. With more prolonged exposure at these altitudes or with further ascent, affected individuals suffer mental impairment that may progress to coma. Hypoxemia at high altitudes is intensified during sleep, as ventilation normally diminishes and also by pulmonary edema, another manifestation of mountain sickness. Reference was made earlier to the observation of Hornbein and colleagues of a mild, but possibly lasting, memory impairment even in acclimated mountaineers who had been exposed to extremely high altitudes for several days. Hackett and Roach have reviewed the treatments for altitude illness.
Chronic mountain sickness, also called Monge disease (after the physician who described the condition in Andean Indians of Peru), is observed in long-term inhabitants of high-altitude mountainous regions. Pulmonary hypertension, cor pulmonale, and secondary polycythemia are the main features. There is usually hypercarbia as well, with the expected degree of mild mental dullness, slowness, fatigue, nocturnal headache, and, sometimes, papilledema (see below). Thomas and colleagues have called attention to a syndrome of burning hands and feet in Peruvians at high altitude, apparently a maladaptive response to chronic hypoxia.
Sedatives, alcohol, and a slightly elevated Pco2 in the blood all reduce tolerance to high altitude. Dexamethasone and acetazolamide prevent and counteract mountain sickness to some extent. The most effective preventive measure is acclimatization by a 2- to 4-day stay at intermediate altitudes.
Hypercapnic Pulmonary Disease
Chronic obstructive pulmonary disease such as emphysema, fibrosing lung disease, neuromuscular weakness, and, in some instances, inadequacy of the medullary respiratory centers each may lead to persistent respiratory acidosis, with elevated of Pco2 and reduced in arterial PO2. The complete clinical syndrome of chronic hypercapnia described by Austen, Carmichael, and Adams comprises headache, papilledema, mental dullness, drowsiness, confusion, stupor and coma, and asterixis. More typically, only some of these features are found. Some patients have a fast-frequency tremor. The headache tends to be generalized, frontal, or occipital and can be quite intense, persistent, steady, and aching in type; nocturnal occurrence is a feature of some cases. The papilledema is bilateral but may be slightly greater in one eye than in the other, and hemorrhages may encircle the choked disc (a later finding). The tendon reflexes are lively and plantar reflexes may be extensor. Intermittent drowsiness, inattentiveness, reduction of psychomotor activity, inability to perceive all the items in a sequence of events, and forgetfulness constitute the more subtle manifestations of this syndrome and may prompt the family to seek medical help. Such symptoms may last only a few minutes or hours, and one cannot count on their presence at the time of a particular examination. In fully developed cases, the CSF is under increased pressure; Pco2 may exceed 75 mm Hg, and the O2 saturation of arterial blood ranges from 85 percent to as low as 40 percent. The EEG shows slow activity in the delta or theta range, which is sometimes bilaterally synchronous.
The mechanism of the cerebral disorder is from a direct CO2 narcosis, but the biochemical details are not known. Normally the CSF is slightly acidotic in comparison to the blood and the Pco2 of the CSF is about 10 mm Hg higher than that of the blood. With respiratory acidosis, the pH of the CSF falls (into the range of 7.15 to 7.25) and cerebral blood flow increases as a result of cerebral vasodilatation. However, the brain rapidly adapts to respiratory acidosis through the generation and secretion of bicarbonate by the choroid plexuses. Brain water content also increases, mainly in the white matter. In animal models of hypercarbia, blood and brain NH3 is elevated, which may explain the similarity of the syndrome to that of hyperammonemic liver failure (Herrera and Kazemi).
The most effective therapeutic measures are positive-pressure ventilation, using oxygen if there is also hypoxia. Oxygen supplementation is, of course, used cautiously in these patients in order to avoid suppressing respiratory drive; marginally compensated patients treated with excessive oxygen have lapsed into coma. Treatment of heart failure, phlebotomy to reduce the viscosity of the blood, and antibiotics to suppress pulmonary infection may be necessary. Often these measures result in a surprising degree of improvement, which may be maintained for months or years.
Unlike pure hypoxic encephalopathy, prolonged coma because of hypercapnia is relatively rare and in our experience has not led to irreversible brain damage. Papilledema, myoclonus, and especially asterixis are important diagnostic features. If aminophylline is administered for the treatment of the underlying pulmonary airway disease, it may produce high blood levels and a tendency for it to produce seizures.
This condition is now relatively infrequent but is an important cause of confusion, convulsions, stupor, and coma; as such, it merits separate consideration as a metabolic disorder of the brain. The essential biochemical abnormality is a critical lowering of the blood glucose. At a level of about 30 mg/dL, the cerebral disorder takes the form of a confusional state and one or more seizures may occur; at a level of 10 mg/dL, there is coma that may result in irreparable injury to the brain if not corrected immediately by the administration of glucose. As with most other metabolic encephalopathies, the rate of decline of blood glucose is a factor in both the depression of consciousness and residual dementia.
The normal brain has a glucose reserve of 1 to 2 g (30 mmol/100 g of tissue), mostly in the form of glycogen. Because glucose is utilized by the brain at a rate of 60 to 80 mg/min, the glucose reserve may sustain cerebral activity for 30 min or less once blood glucose is no longer available. Glucose is transported from the blood to the brain by an active carrier system. Glucose entering the brain either undergoes glycolysis or is stored as glycogen. During normal oxygenation (aerobic metabolism), glucose is converted to pyruvate, which enters the Krebs cycle; with anaerobic metabolism, lactate is formed. The oxidation of 1 mole of glucose requires 6 mole of O2. Of the glucose taken up by the brain, 85 to 90 percent is oxidized; the remainder is used in the formation of proteins and other substances, notably neurotransmitters and particularly gamma-aminobutyric acid (GABA).
When blood glucose falls, the central nervous system (CNS) can utilize nonglucose substrates to a variable extent for its metabolic needs, especially keto acids and intermediates of glucose metabolism, such as lactate, pyruvate, fructose, and other hexoses. In the neonatal brain, which has a higher glycogen reserve, keto acids provide a considerable proportion of cerebral energy requirements; this also happens after prolonged starvation. However, in the face of severe and sustained hypoglycemia, these alternative substrates are inadequate to preserve the structural integrity of neurons, and eventually adenosine triphosphate (ATP) is depleted as well. If convulsions occur, they usually do so during a period of confusion; the convulsions have been attributed to an altered integrity of neuronal membranes and to elevated NH3 and depressed GABA and lactate levels (Wilkinson and Prockop).
The brain is the only organ besides the heart that suffers severe functional and structural impairment under conditions of severe hypoglycemia. Beyond what is described above, the pathophysiology of the cerebral disorder has not been fully elucidated. It is known that hypoglycemia reduces O2 uptake and increases cerebral blood flow. As with anoxia and ischemia, there is experimental evidence that the excitatory amino acid glutamate is involved in the process. The levels of several brain phospholipid fractions decrease when animals are given large doses of insulin. However, the suggestion that hypoglycemia results in a rapid depletion and inadequate production of high-energy phosphate compounds has not been corroborated; some other glucose-dependent biochemical processes must be implicated.
The most common causes of hypoglycemic encephalopathy are: (1) accidental or deliberate overdose of insulin or an oral diabetic agent; (2) islet cell insulin-secreting tumor of the pancreas; (3) depletion of liver glycogen, which occasionally follows a prolonged alcoholic binge, starvation, or any form of severe liver failure; (4) glycogen storage disease of infancy; and (5) an idiopathic hypoglycemia in the neonatal period and infancy; (6) subacute and chronic hypoglycemia from islet cell hypertrophy and islet cell tumors of the pancreas, carcinoma of the stomach, fibrous mesothelioma, carcinoma of the cecum, and hepatoma. Purportedly, an insulin-like substance is elaborated by these nonpancreatic tumors. In the past, hypoglycemic encephalopathy was a not infrequent complication of "insulin shock" therapy for schizophrenia. In functional hyperinsulinism, as occurs in anorexia nervosa and dietary faddism, the hypoglycemia is rarely of sufficient severity or duration to damage the CNS.
The initial symptoms appear when the blood glucose has descended to about 30 mg/dL, nervousness, hunger, flushed facies, sweating, headache, palpitation, trembling, and anxiety. These gradually give way to confusion and drowsiness or occasionally, to excitement, overactivity, and bizarre or combative behavior. Many of the early symptoms relate to adrenal and sympathetic overactivity and some of the manifestations may be muted in diabetic patients with neuropathy. In the next stage, forced sucking, grasping, motor restlessness, muscular spasms, and decerebrate rigidity occur, in that sequence. Myoclonic twitching and convulsions develop in some patients. Rarely, there are focal cerebral deficits, the pathogenesis of which remains unexplained; according to Malouf and Brust, hemiplegia, corrected by intravenous glucose, was observed in 3 of 125 patients who presented with symptomatic hypoglycemia.
Blood glucose levels of approximately 10 mg/dL are associated with deep coma, dilatation of pupils, pale skin, shallow respiration, slow pulse and hypotonia, what had in the past been termed the "medullary phase" of hypoglycemia. If glucose is administered before this level has been attained, the patient can be restored to normal, retracing the aforementioned steps in reverse order. However, once this state is reached, and particularly if it persists for more than a few minutes, recovery is delayed for a period of days or weeks and may be incomplete as noted below.
The EEG is altered as the blood glucose falls, but the correlations are imprecise. There is diffuse slowing in the theta or delta range. During recovery, sharp waves may appear and coincide in some cases with seizures.
The major clinical differences between hypoglycemic and hypoxic encephalopathy lie in the setting and the mode of evolution of the neurologic disorder. The effects of hypoglycemia usually unfold more slowly, over a period of 30 to 60 min, rather than in a few seconds or minutes. The recovery phase and sequelae of the 2 conditions are quite similar.
A large dose of insulin, which produces intense hypoglycemia, even of relatively brief duration (30 to 60 min), is more dangerous than a series of less-severe hypoglycemic episodes from smaller doses of insulin, possibly because the former impairs or exhausts essential enzymes, a condition that cannot then be overcome by large quantities of intravenous glucose. Reflecting the benignity of repeated minor occurrences, the Epidemiology of Diabetes Interventions and Complications Study Research Group have demonstrated that recurrent hypoglycemic episodes in the course of treatment of diabetes over many years are very well tolerated and do not lead to cognitive decline.
A severe and prolonged episode of hypoglycemia may result in permanent impairment of intellectual function as well as other neurologic residua, like those that follow severe anoxia. We also have observed states of protracted coma, as well as relatively pure Korsakoff amnesia. However, one should not be hasty in prognosis, for we have observed slow improvement to continue for 1 to 2 years.
Recurrent hypoglycemia from an islet cell tumor may masquerade for some time as an episodic confusional psychosis or convulsive illness; diagnosis then awaits the demonstration of low blood glucose or hyperinsulinism in association with the neurologic symptoms. We saw a man in the emergency department whose main complaint was episodic inability to dial a touchtone telephone and a mild mental fogginess; he was found to have an insulinoma.
Functional or reactive hypoglycemia is the most ambiguous of all syndromes related to low blood glucose. This condition is usually idiopathic but may precede the onset of diabetes mellitus. The rise of insulin in response to a carbohydrate meal is delayed but then causes an excessive fall in blood glucose, to 30 to 40 mg/dL. The symptoms are malaise, fatigue, nervousness, headache and tremor, which may be difficult to distinguish from anxious depression. Not surprisingly, the term functional hypoglycemia has been much abused, being applied indiscriminately to a variety of complaints that would now be called chronic fatigue syndrome or an anxiety syndrome. In fact, a syndrome attributable to functional or reactive hypoglycemia is infrequent and its diagnosis requires the finding of an excessive reaction to insulin, low blood glucose during the symptomatic period, and a salutary response to oral glucose.
In all forms of hypoglycemic encephalopathy, the major damage is to the cerebral cortex. Cortical nerve cells degenerate and are replaced by microglia cells and astrocytes. The distribution of lesions is similar, although probably not identical to that in hypoxic encephalopathy. The cerebellar cortex is less vulnerable to hypoglycemia than to hypoxia. Auer has described the ultrastructural changes in neurons resulting from experimental hypoglycemia; with increasing duration of hypoglycemia and EEG silence, there are mitochondrial changes, first in dendrites and then in nerve cell soma, followed by nuclear membrane disruption leading to cell death.
Treatment of all forms of hypoglycemia obviously consists of correction of the hypoglycemia at the earliest possible moment. It is not known whether hypothermia or other measures will increase the safety period in hypoglycemia or alter the outcome. Seizures and twitching may not stop with antiepileptic drugs until the hypoglycemia is corrected.
Two syndromes have been defined, mainly in diabetics: (1) hyperglycemia with ketoacidosis and (2) hyperosmolar nonketotic hyperglycemia.
In diabetic acidosis, the familiar picture is one of dehydration, fatigue, weakness, headache, abdominal pain, dryness of the mouth, stupor or coma, and Kussmaul type of breathing. Usually the condition has developed over a period of days in a patient known or proven to be diabetic. Often, the patient had failed to take a regular insulin dose. The blood glucose level is found to be more than 400 mg/dL, the pH of the blood less than 7.20, and the bicarbonate less than 10 mEq/L. Ketone bodies and β-hydroxybutyric acid are elevated in the blood and urine, and there is marked glycosuria. The prompt administration of insulin and repletion of intravascular volume correct the clinical and chemical abnormalities over a period of hours.
A small group of patients with diabetic ketoacidosis, such as those reported by Young and Bradley, develop deepening coma and cerebral edema as the elevated glucose is corrected. Mild cerebral edema is commonly observed in children during treatment with fluids and insulin (Krane et al). Prockop attributed this condition to an accumulation of fructose and sorbitol in the brain. The latter substance, a polyol that is formed during hyperglycemia, crosses membranes slowly, but once it does so, is said to cause a shift of water into the brain and an intracellular edema. However, according to Fishman (1974), the increased polyols in the brain in hyperglycemia are not present in sufficient concentration to be important osmotically; though they may induce other metabolic effects related to the encephalopathy. These are matters of conjecture, as the increase of polyols has never been found. The brain edema in this condition is probably a result of reversal of the osmolality gradient from blood to brain, which occurs with rapid correction of hyperglycemia.
The pathophysiology of the cerebral disorder in diabetic ketoacidosis is not fully understood. No consistent cellular pathology of the brain has been identified in the cases we have examined. Factors such as ketosis, tissue acidosis, hypotension, hyperosmolality, and hypoxia have not been identified. Attempts at therapy by the administration of urea, mannitol, salt-poor albumin, and dexamethasone are usually unsuccessful, though recoveries are reported.
In hyperosmolar nonketotic hyperglycemia, the blood glucose is extremely high, more than 600 mg/dL, but ketoacidosis does not develop, or if it does develop, it is mild. Osmolality is usually in excess of 330 mOsm/L. There is also hemoconcentration and prerenal azotemia. Appreciation of the neurologic syndrome is generally credited to Wegierko, who published descriptions of it in 1956 and 1957. Most of the patients are elderly diabetics but some were not previously known to have been diabetic. An infection, enteritis, pancreatitis, dehydration, or a drug known to upset diabetic control (thiazides, corticosteroids, and phenytoin) leads to polyuria, fatigue, confusion, stupor, and coma. Often the syndrome arises in conjunction with the combined use of corticosteroids and phenytoin (which inhibits insulin release), for example, in elderly patients with brain tumors. The use of osmotic diuretics enhances the risk.
Seizures and focal signs such as a hemiparesis, a hemisensory defect, choreoathetosis, or a homonymous visual field defect are more common than in any other metabolic encephalopathy and may erroneously suggest the possibility of a stroke. Fluids should be replaced cautiously, using isotonic saline and potassium. Correction of the markedly elevated blood glucose requires relatively small amounts of insulin, since these patients often do not have a high degree of insulin resistance.
Hepatic Stupor and Coma (Hepatic, or Portal–Systemic Encephalopathy)
Chronic hepatic insufficiency with portosystemic shunting of blood is punctuated by episodes of stupor, coma, and other neurologic symptoms—a state referred to as hepatic stupor, coma, or encephalopathy. It was clearly delineated by Adams and Foley over 50 years ago. This state complicates all varieties of liver disease and is unrelated to jaundice or ascites. Any form of shunting, even without hepatic disease, such as surgical portal–systemic shunt (Eck fistula) is attended by the same clinical picture (see further on). There are also a number of hereditary hyperammonemic syndromes, usually first apparent in infancy or childhood (discussed extensively in Chap. 37) that lead to episodic coma with or without seizures. In all these states, it is common for an excess of protein derived from the diet or from gastrointestinal hemorrhage to induce or worsen the encephalopathy. Additional predisposing factors are hypoxia, hypokalemia, metabolic alkalosis, excessive diuresis, use of sedative hypnotic drugs, and constipation. A special from of the syndrome in epilepsy patients exposed to valproate; confusion and ataxia may occur acutely or subacutely in these patients (Gomcelli et al). Reye syndrome following viral infections in children, now infrequent, was also associated with very high levels of ammonia in the blood and encephalopathy (see further on).
The clinical picture of acute, subacute, or chronic hepatic encephalopathy consists of a derangement of consciousness, presenting first as mental slowing and confusion, occasionally with hyperactivity, followed by progressive drowsiness, stupor, and coma. The confusional state is combined with a characteristic intermittency of sustained muscle contraction; this phenomenon, which was originally described in patients with hepatic stupor by Adams and Foley and called asterixis (from the Greek sterixis, a "fixed position"). It is now recognized as a sign of various metabolic encephalopathies but is most prominent in this disorder (see Chap. 6). It is conventionally demonstrated by having the patient hold his arms outstretched with the wrists extended, but the same tremor can be elicited by any sustained posture, including that of the protruded tongue. We have seen patients in whom asterixis of the large antigravity muscles (e.g., iliopsoas or quadriceps) causes falling. A variable, fluctuating rigidity of the trunk and limbs, grimacing, suck and grasp reflexes, exaggeration or asymmetry of tendon reflexes, Babinski signs, and focal or generalized seizures round out the clinical picture in a few patients.
The EEG is a sensitive and reliable indicator of impending coma, becoming abnormal during the earliest phases of the disordered mental state. Foley, Watson, and Adams noted an EEG abnormality consisting of paroxysms of bilaterally synchronous slow or triphasic waves in the delta range, which at first predominate frontally and are interspersed with alpha activity and later, as the coma deepens, displace all normal activity (see Fig. 2-5H). A few patients show only random high-voltage asynchronous slow waves.
This syndrome of hepatic encephalopathy is remarkably diverse in its course and evolution. It usually appears over a period of days to weeks and may terminate fatally; or, with appropriate treatment, the symptoms may regress and then fluctuate in severity for several weeks or months. Persistent hepatic coma of the latter type proves fatal in about half of patients (Levy et al). In many patients, the syndrome is relatively mild and does not evolve beyond the stage of mental dullness and confusion, with asterixis and EEG changes. In yet others, a subtle disorder of mood, personality, and intellect may be protracted over a period of many months or even years; this chronic but nevertheless reversible mental disturbance need not be associated with overt clinical signs of liver failure (jaundice and ascites) or other neurologic signs. Characteristically in these patients, an extensive portal–systemic collateral circulation can be demonstrated (hence the term portal–systemic encephalopathy) and an association established between the mental disturbance and an intolerance to dietary protein as well as raised blood ammonia levels (Summerskill et al).
The diversion of blood from the portal system into the vena cava after ligation of the portal veins was first performed in dogs by Eck in 1877. Probably the first and certainly most striking example in man was the case of pure "Eck" fistula reported by McDermott and Adams, in which a portacaval shunt was created during the removal of a pancreatic tumor. The liver was normal. Episodic coma occurred thereafter whenever dietary protein increased. Consciousness was restored on a protein-free diet, and coma could be induced again by ammonium chloride. Postmortem examination 2 years later confirmed the normal liver and showed cerebral changes of hepatic encephalopathy, as described below.
Finally, there is a group of patients (most of whom have experienced repeated attacks of hepatic coma) in whom an irreversible mild dementia and a disorder of posture and movement (grimacing, tremor, dysarthria, ataxia of gait, choreoathetosis) gradually appear. This condition of chronic acquired hepatocerebral degeneration must be distinguished from other dementing and extrapyramidal syndromes (see further on). A few cases of isolated spastic paraplegia (so-called hepatic myelopathy, or more correctly hepatic paraplegia) have been described. Pant and Richardson attributed the syndrome to a loss of Betz cells in the frontal cortex; in other words, a restricted encephalopathy seen in some patients with portal-systemic encephalopathy. Indeed, the spasticity of the legs, increased tendon reflexes, and Babinski signs that are found with PSE, suggest that "hepatic paraparesis" is a common feature of the encephalopathy.
MRI in PSE often demonstrates high signal intensity in the globus pallidus, likely the result of manganese deposition.
The concentrations of blood NH3, particularly if measured repeatedly in arterial blood samples, usually are well in excess of 200 mg/dL, and the severity of the neurologic and EEG disorders roughly parallels to the ammonia levels. With treatment, a fall in the NH3 levels precedes clinical improvement.
The striking finding by Adams and Foley in patients who died in a state of hepatic coma was a diffuse increase in the number and size of the protoplasmic astrocytes in the deep layers of the cerebral cortex, lenticular nuclei, thalamus, substantia nigra, cerebellar cortex, and red, dentate, and pontine nuclei, with little or no visible alteration in the nerve cells or other parenchymal elements. With periodic acid-Schiff (PAS) staining, the astrocytes were seen to contain glycogen inclusions. These abnormal glia cells are generally referred to as Alzheimer type II astrocytes, having been described originally in 1912 by von Hosslin and Alzheimer in a patient with Westphal-Strümpell pseudosclerosis (or Wilson disease). These astrocytes have been studied by electron microscopy in rats with surgically created portacaval shunts (Cavanagh; Norenberg); the cells show a number of striking abnormalities—swelling of their terminal processes, cytoplasmic vacuolation (distended sacs of rough endoplasmic reticulum), formation of folds in the basement membrane around capillaries, and an increase in both the number of mitochondria and enzymes that catabolize ammonia. Also, some degeneration in myelinated nerve fibers in the neuropil and an increase in the cytoplasm of oligodendrocytes are seen. In chronic cases, neuronal loss in the deep layers of the cerebral and cerebellar cortex and lenticular nuclei is found, as well as vacuolization of tissue (possibly astrocytic) resembling the lesions of Wilson disease.
The ubiquitous astrocytic alterations occur to some degree in all patients who die of progressive liver failure and the degree of glial abnormality corresponds generally to the intensity and duration of the neurologic disorder. Possibly, the astrocytic changes affect the synaptic activities of the neurons. The clinical and EEG features of hepatic encephalopathy as well as the astrocytic hyperplasia are more or less specific features of this metabolic disorder. Nevertheless, taken together in a setting of liver failure, they constitute a distinctive clinicopathologic entity.
Pathogenesis of Hepatic Encephalopathy
The most plausible hypothesis relates hepatic coma to an abnormality of nitrogen metabolism, wherein ammonia, which is formed in the bowel by the action of urease-containing organisms on dietary protein, is carried to the liver in the portal circulation but fails to be converted into urea because of hepatocellular disease, portal–systemic shunting of blood, or both. As a result, excess NH3 reaches the systemic circulation, where it interferes with cerebral metabolism in a way that is not fully understood. The ammonia theory best explains the basic neuropathologic change. Because the brain is lacking in urea cycle enzymes, Norenberg has proposed that the hypertrophy of the astrocytic cytoplasm and proliferation of mitochondria and endoplasmic reticulum, as well as the increase in the astroglial glutamic dehydrogenase activity, all reflect heightened metabolic activity of these systems within astrocytes associated with ammonia detoxification. Removal of brain ammonia depends on the formation of glutamine, a reaction that is catalyzed by the ATP-dependent enzyme glutamine synthetase, which is compartmentalized to astrocytes. It has been shown in experimental animals that hyperammonemia leads to a depletion of ATP in midbrain reticular nuclei. Whether this is the primary cause of cerebral dysfunction has not been resolved.
Numerous alternative theories have been suggested. One is that CNS function in cirrhotic patients is impaired by phenols or short-chain fatty acids derived from the diet or from bacterial metabolism of carbohydrate. Another theory holds that biogenic amines (e.g., octopamine), which arise in the gut and bypass the liver, act as false neurotransmitters, displacing norepinephrine and dopamine (Fischer and Baldessarini). Zieve has presented evidence that mercaptans (methanethiol, methionine), which are also generated in the gastrointestinal tract and removed by the liver, act in conjunction with NH3 to produce hepatic encephalopathy. This theory and others have been largely discounted; they are the subject of reviews by Butterworth and coworkers, by Zieve, by Rothstein and Herlong, and by Jones and Basile, to which the reader is referred for detailed information.
Also, manganese has emerged as a potential neurotoxin in the pathogenesis of hepatic encephalopathy (Kreiger et al; Pomier-Layrargues et al). In patients with chronic liver disease and with spontaneous or surgically induced portal–systemic shunts, manganese accumulates in the serum and in the brain, more specifically in the pallidum. This accumulation is readily discernible as a pallidal signal hyperintensity on T1-weighted MRI. Following liver transplantation, there is normalization of the MRI changes and of the associated extrapyramidal symptoms. The effects of manganese chelation on such patients have not been well studied and the mechanisms of accumulated manganese in the pathogenesis of hepatic encephalopathy are not known. It is clear, therefore, that any theory of hepatic encephalopathy must incorporate the cerebral effects of hyperammonemia.
For some time, it has been known that hepatic encephalopathy is associated with increased activity of the inhibitory transmitter GABA in the cerebral cortex. It has also been observed that increased gabanergic neurotransmission may result from substances that inhibit the binding of endogenous benzodiazepine-like compounds to their receptors (Basile et al). Furthermore, these antagonists are found to have some clinical effect—transient arousal in patients with hepatic encephalopathy. The actions of benzodiazepines are mediated by these receptors; hence the designation GABA-benzodiazepine theory. The practicality of using benzodiazepine receptor antagonists, which are short-acting and reversible (e.g., flumazenil) in the treatment of hepatic encephalopathy, remains to be determined (see Mullen), but they offer an interesting diagnostic test.
Until recently, the ammonia and the gabanergic-benzodiazepine hypotheses of the pathogenesis of hepatic encephalopathy had appeared to be unrelated. However, there is evidence, reviewed by Jones and Basile, that ammonia, even in the modestly elevated concentrations that occur in liver failure, inhibits metabolism of GABA by the astrocytes and enhances gabanergic neurotransmission—a concept that unifies hyperammonemia and the neurotransmitter change. Furthermore, the aforementioned glial abnormality may be the explanation for the disorder of the blood-brain barrier that leads to the brain swelling that is seen in some rapidly developing cases of PSE, the prototype for which is the now infrequent Reye syndrome. A parallel astrocytic dysfunction may lead to disruption of the blood–brain barrier and the brain swelling that is known to occur in cases of acute liver failure.
Despite the incompleteness of our understanding of the role of disordered ammonia metabolism in the genesis of hepatic coma, an awareness of this relationship has provided the few effective means of treating this disorder: restriction of dietary protein; reduction of bowel flora by oral administration of neomycin or kanamycin, which suppresses the urease-producing organisms in the bowel; and the use of enemas. The mainstay of treatment has been oral lactulose, an inert sugar that is metabolized by colonic bacteria that produce hydrogen ions and shifts ammonia to ammonium, a nontoxic product that is eliminated in the stool. The past use of oral neomycin carried a risk of renal damage and ototoxicity and has therefore been replaced by rifaximin, a minimally absorbed antibiotic that has less risk. This antibiotic has also been shown by Bass and coworkers to be highly effective in preventing episodic hepatic encephalopathy in tenuously compensated patients. The salutary effect of these therapeutic measures, the common attribute of which is the lowering of the blood NH3, further supports the theory of ammonia intoxication. Ultimately, in cases of intractable liver failure, transplantation becomes a treatment of last resort.
Other treatments with lesser value include bromocriptine, the aforementioned diazepine antagonist flumazenil, and keto analogues of essential amino acids. Theoretically, the keto analogues should provide a nitrogen-free source of essential amino acids (Maddrey et al), a treatment that has been largely abandoned, and bromocriptine, a dopamine agonist, should enhance dopaminergic transmission (Morgan et al) but its mechanism is not known. Administration of branched-chain amino acids may result in improvement in mental status but their effects have been variable and associated with an increased mortality (Naylor et al). The transient beneficial effect of the benzodiazepine antagonist flumazenil has already been mentioned; it is used as well as a diagnostic test.
Fulminant Hepatic Failure and Cerebral Edema
In acute hepatitis, confusional, delirious, and comatose states also occur but their mechanisms are still unknown. Blood NH3 may be elevated but usually not to a degree that would be expected to cause encephalopathy. Severe acute hepatic failure may cause hypoglycemia, which contributes to the encephalopathy and often presages a fatal outcome but the levels of glucose typically detected do not provide an explanation for the encephalopathy.
Cerebral edema is a prominent finding in cases of fulminant hepatic failure and is the main cause of death in patients awaiting liver transplantation. The cerebral edema in these circumstances appears to be related to the rapidity of rise of blood ammonia, but it probably depends as well on additional metabolic derangements that complicate acute liver failure including glial cell failure with consequent incompetence of the blood-brain barrier. The combination of rapidly evolving hepatic failure and massive cerebral edema is similar to that observed in the Reye syndrome, described below.
CT and MRI are effective means of detecting cerebral edema in patients with fulminant hepatic failure, and according to Wijdicks and colleagues, the degree of cerebral swelling is roughly proportional to the severity of encephalopathy. Because patients with fulminant hepatic failure can survive liver transplantation with few or no neurologic deficits, it is important to recognize cerebral edema before the stage of stupor and increased intracranial pressure has been established. Short of transplantation, death in these cases may sometimes be prevented by monitoring the intracranial pressure (as outlined by Lidofsky et al) and administering osmotic diuretics and hyperventilation, as detailed in Chaps. 31 and 35 for the treatment of intracranial hypertension. Some survivors are nonetheless left with cerebral damage from raised intracranial pressure.
An additional issue that arises in assessing cerebral dysfunction in patients with liver disease is the possibility of adverse effects of medications. Individuals with hepatitis C who are treated with interferon-alpha may develop a spectrum of problems ranging from subtle cognitive impairment to a subacutely worsening headache, vomiting, altered consciousness, and focal neurologic findings. The milder syndromes are associated with no or few MRI-visible lesions but the severe ones are usually accompanied by signal changes in the white matter of the occipital lobes and elsewhere (posterior leukoencephalopathy; see Fig. 43-1).
Reye Syndrome (Reye-Johnson Syndrome)
This is a special type of nonicteric hepatic encephalopathy occurring in children and adolescents and characterized by acute brain swelling in association with fatty infiltration of the viscera, particularly the liver. Although individual cases of this disorder had been described for many years, its recognition as a clinical-pathologic entity dates from 1963, when a large series was reported from Australia by Reye and colleagues and from the United States by Johnson and coworkers. The disorder tended to occur in outbreaks (286 cases were reported to the Centers for Disease Control during a 4-month period in 1974). Mainly, these outbreaks were observed in association with influenza B virus and varicella infections, but a variety of other viral infections were implicated (influenza A, echovirus, reovirus, rubella, rubeola, herpes simplex, Epstein-Barr virus). Later it became apparent that the toxic or adjuvant effects of aspirin given during these infections played an important role in producing the disease. Today, only occasional instances of Reye syndrome are observed now that the association with aspirin administration has become widely known and its use in children with viral infections has been interdicted.
Most patients are children, boys and girls being equally affected, but rare instances are known in infants (Huttenlocher and Trauner) and young adults. In most cases, the encephalopathy is preceded for several days to a week by fever, symptoms of upper respiratory infection, and protracted vomiting. These are followed by the rapid evolution of stupor and coma, associated in many cases with focal and generalized seizures, signs of sympathetic overactivity (tachypnea, tachycardia, mydriasis), decorticate and decerebrate rigidity, and loss of pupillary, corneal, and vestibuloocular reflexes. One or two such cases were included in the series of acute "toxic encephalopathy" reported by Lyon and colleagues (see "Acute Toxic Encephalopathy" in Chap. 32). In infants, respiratory distress, tachypnea, and apnea are the most prominent features.
The liver may be greatly enlarged, often extending to the pelvis and providing an important diagnostic clue as to the cause of the cerebral changes. Initially there is a metabolic acidosis, followed by a respiratory alkalosis (rising arterial pH and falling Pco2). The CSF is usually under increased pressure and is acellular; glucose values may be low, reflecting the hypoglycemia. The serum ALT, coagulation times, and blood ammonia are increased, sometimes to an extreme degree. The EEG is characterized by diffuse arrhythmic delta activity, progressing to electrocerebral silence in patients who fail to survive. CT and MRI show the cerebral swelling but are difficult to interpret in these young individuals, who lack any adult brain atrophy.
The major pathologic findings are cerebral edema, often with cerebellar herniation, and infiltration of hepatocytes with fine droplets of fat (mainly triglycerides); the renal tubules, myocardium, skeletal muscles, pancreas, and spleen are infiltrated to a lesser extent. There are no inflammatory lesions in the brain, liver, or other organs. There is not full agreement as to the pathogenesis of this disorder and the mechanism of aspirin toxicity but mitochondrial dysfunction has been implicated.
In a series of children with blood ammonia levels greater than 500 mg/dL who were treated during the years 1967 to 1974, Shaywitz and colleagues reported a mortality of 60 percent. Once the child became comatose, death was almost inevitable. In more recent years, early diagnosis and initiation of treatment before the onset of coma have reduced the fatality rate to 5 to 10 percent. Treatment consists of the following measures: temperature control with a cooling blanket; nasotracheal intubation and controlled ventilation to maintain Pco2 below 32 mm Hg; intravenous glucose covered by insulin to maintain blood glucose at 150 to 200 mg/dL; administration of lactulose, neomycin enemas, and hemodialysis to directly lower the NH3 concentration; control of intracranial pressure by means of continual monitoring and the use of hypertonic solutions (see Chap. 30); and the maintenance of fluid and electrolyte balance (Trauner). Upon recovery, cerebral function returns to normal unless there had been deep and prolonged coma or protracted elevation of intracranial pressure.
Episodic confusion and stupor and other neurologic symptoms may accompany any form of severe renal disease—acute or chronic. The cerebral symptoms attributable to uremia (first described by Addison in 1832) are discerned in normotensive individuals in whom renal failure develops rapidly. Apathy, fatigue, inattentiveness, and irritability are usually the initial symptoms; later, there is confusion, dysarthria, tremor, and asterixis. Infrequently, this takes the form of a toxic psychosis, with hallucinations, delusions, insomnia, or catatonia (Marshall). These symptoms characteristically fluctuate from day to day, or even from hour to hour. In some patients, especially those who become anuric, symptoms may come on abruptly and progress rapidly to a state of stupor and coma. In others, in whom uremia develops more gradually, mild visual hallucinations and a disorder of attention may persist for several weeks in relatively pure form. The EEG becomes diffusely and irregularly slow and may remain so for several weeks after the institution of dialysis. The CSF pressure is normal and the protein is not elevated unless there is a uremic or diabetic neuropathy. In several reports, meningismus and a low-grade mononuclear pleocytosis have been mentioned (Merritt and Fremont-Smith), but we have not encountered this.
In acute renal failure, clouding of the sensorium is practically always associated with a variety of motor phenomena. These usually occur early in the course of the encephalopathy, sometimes when the patient is still mentally clear. The patient begins to twitch and jerk and may convulse. The myoclonic twitches involve parts of muscles, whole muscles, or entire limbs and are lightning-quick, arrhythmic, and asynchronous on the two sides of the body; they are incessant during both wakefulness and sleep. At times the movements resemble those of chorea or an arrhythmic tremor; asterixis is also readily evoked. The motor phenomena are often difficult to classify. Our predecessor authors described the condition as a uremic twitch-convulsive syndrome.
The resemblance of uremic encephalopathy to hepatic and other metabolic encephalopathies has been stressed by Raskin and Fishman, yet we are more impressed with differences than with similarities. When the twitch-convulsive syndrome is observed in association with other diseases such as widespread neoplasia, delirium tremens, diabetic coma, and lupus erythematosus, the causative factor of renal failure is usually discovered.
As the uremia worsens, the patient lapses into a quiet coma. Unless the accompanying metabolic acidosis is corrected, Kussmaul breathing appears and gives way to Cheyne-Stokes breathing and death.
Encephalopathy and coma in the patient with renal failure may, of course, be a result of disorders other than uremia itself. Because of the similarity of this syndrome to tetany, measurement should be made of serum calcium and magnesium—and, of course, hypocalcemia and hypomagnesemia do occur in uremia. But often the values for these ions are normal or near normal, and the administration of calcium and magnesium salts has little effect. The altered excretion of drugs leads to their accumulation, sometimes evoking excessive sedation. Subdural and intracerebral hemorrhages may complicate uremia (and dialysis) because of clotting defects and hypertension; and uremic patients are prone to infections, including meningitis.
Because chronic uremia is so frequently associated with hypertension, a major problem also arises in distinguishing the cerebral effects of uremia from those of severe and accelerated hypertension. Volhard was the first to make this distinction; he introduced the term pseudouremia to designate the cerebral effects of malignant hypertension and to separate them from true uremia. The preferable term, hypertensive encephalopathy, was first used by Oppenheimer and Fishberg. However, the myoclonic-twitch syndrome is not a component of hypertensive encephalopathy. The clinical picture of the latter disorder and its pathophysiology are discussed in "Hypertensive Encephalopathy and Eclampsia" in Chap. 34.
Opinions vary widely as to the biochemical basis of uremic encephalopathy and the twitch-convulsive syndrome. Restoration of renal function completely corrects the neurologic syndrome, attesting to the absence of structural change and a functional disorder of subcellular type. Whether caused by the retention of organic acids, elevation of phosphate in the CSF (claimed by Harrison et al), or the action of urea or other toxins, among them parathyroid hormone, has never been settled. The data supporting the causative role of urea itself are also ambiguous, just as they are for other putative endogenous agents (see Bolton and Young and the review by Burn and Bates). However, it can be stated that urea itself is not the sole inductive agent, as its infusion does not produce the syndrome in humans or animals.
It would appear that every level of the CNS is affected in uremia, from spinal cord to cerebrum. Cellular changes in the brain or spinal cord are limited to mild hyperplasia of protoplasmic astrocytes in some cases, but never of the degree observed in hepatic encephalopathy. Cerebral edema is notably absent. In fact, CT scans and MRI regularly show an element of cerebral shrinkage. A peripheral neuropathy is also a common complication of uremia and is considered in Chap. 46.
Improvement of encephalopathic symptoms may not be evident for a day or two after institution of dialysis. Convulsions, which occur in about one-third of cases, often preterminally, may be resistant to treatment until the uremia is addressed. However, some seizures may be suppressed with relatively low plasma concentrations of antiepileptic drugs, the reason being that serum albumin is depressed in uremia, increasing the unbound, therapeutically active portion of a drug. If there are severe associated metabolic disturbances, such as hyponatremia, the seizures may be difficult to control. One must be cautious in prescribing any of a large number of drugs in the face of renal failure, for inordinately high, toxic blood levels may result. Examples that affect the nervous system are aminoglycoside antibiotics (vestibular damage); furosemide (cochlear damage); and nitrofurantoin, isoniazid, and hydralazine (peripheral nerve damage).
"Dialysis Disequilibrium" Syndrome
This term refers to a group of symptoms that may occur during and following hemodialysis or peritoneal dialysis as a byproduct of some degree of cerebral edema. The symptoms include headache, nausea, muscle cramps, nervous irritability, agitation, drowsiness, and convulsions. The headache, which may be bilateral and throbbing and resemble common migraine, develops in approximately 70 percent of patients, whereas the other symptoms are observed in 5 to 10 percent, usually in those undergoing rapid dialysis or in the early stages of a dialysis program. The symptoms tend to occur in the third or fourth hour of dialysis and last for several hours. Sometimes they appear 8 to 48 h after the completion of dialysis. Originally, these symptoms were attributed to the rapid lowering of serum urea, leaving the brain with a higher concentration of urea than the serum and resulting in a shift of water into the brain to equalize the osmotic gradient (reverse urea syndrome). Now it is believed that the shift of water into the brain is akin to water intoxication and is a result of the inappropriate secretion of antidiuretic hormone.
The symptoms of subdural hematoma, which in some series had in the past occurred in 3 to 4 percent of patients undergoing dialysis, now being less frequent, may be mistakenly attributed to the disequilibrium syndrome.
Dialysis Encephalopathy (Dialysis Dementia)
This is a subacutely progressive syndrome that in the past complicated chronic hemodialysis. Characteristically, the condition begins with a hesitant, stuttering dysarthria and aphasia, to which are added facial and then generalized myoclonus, focal and generalized seizures, personality and behavioral changes, and intellectual decline. The EEG is invariably abnormal, taking the form of paroxysmal and sometimes periodic sharp-wave or spike-and-wave activity (up to 500 mV and lasting 1 to 20 s), intermixed with abundant theta and delta activity. The CSF is normal except occasionally for increased protein.
At first the myoclonus and speech disorders are intermittent, occurring during or immediately after dialysis and lasting for only a few hours, but gradually they become more persistent and eventually permanent. Once established, the syndrome is usually steadily progressive over a 1- to 15-month period (average survival of 6 months in the 42 cases analyzed by Lederman and Henry). A characteristic feature is a transient improvement in speech with the administration of intravenous diazepines.
The neuropathologic changes are said to be subtle and consist of a mild degree of microcavitation of the superficial layers of the cerebral cortex. Although the changes are diffuse, they have been found in one study to be more severe in the left (dominant) hemisphere than in the right and more severe in the left frontotemporal operculum than in the surrounding cortex (Winkelman and Ricanati). The disproportionate affection of the left frontotemporal opercular cortex putatively explains the distinctive disorder of speech and language. In the one case we have studied carefully, we could not be certain of any microscopic changes.
The most plausible view of the pathogenesis of dialysis encephalopathy is that it represented a form of aluminum intoxication (Alfrey et al), the aluminum being derived from the dialysate or from orally administered aluminum gels. In recent years, this disorder has disappeared, the result, in all likelihood, of the universal practice of purifying the water used in dialysis and thereby removing aluminum from the dialysate. This subject has been reviewed by Parkinson and coworkers.
Complications of Renal Transplantation
The risk in immunosuppressed persons of developing a primary lymphoma of the brain or progressive multifocal leukoencephalopathy is well known and has been mentioned in previous chapters. An entirely different encephalopathy that is marked by widespread visual symptoms and edema of the cerebral white matter, evident on the MRI, but mainly occipital, occurs after the administration of cyclosporine and other immunosuppressant drugs. These imaging features of reversible posterior leukoencephalopathy or posterior reversible encephalopathy syndrome (PRES; see Chaps. 34 and 43), are not specific, being seen also in patients with hypertensive encephalopathy, eclampsia, intrathecal methotrexate administration, and other conditions (see Table 43-1 and Fig. 43-1) all of which are probably linked by endothelial dysfunction of cerebral vessels. Systemic fungal infections had in the past been found at autopsy in approximately 45 percent of patients who had had renal transplants and long periods of immunosuppressive treatment; in about one-third of these patients, the CNS was involved. Cryptococcus, Listeria, Aspergillus, Candida, Nocardia, and Histoplasma were the usual organisms. Recent experience suggests a far lower rate of infection but it remains a threat. Other CNS infections that have complicated transplantation are toxoplasmosis and cytomegalovirus (CMV) inclusion disease.
We have found examples of Wernicke-Korsakoff disease and central pontine myelinolysis in uremic patients. A bleeding diathesis may result in subdural or cerebral hemorrhage, as already mentioned.
Encephalopathy Associated with Sepsis and Burns ("Septic Encephalopathy")
Bolton and Young have drawn attention to the frequent occurrence, in severely septic patients, of a drowsy or confusional state that is reversible and not explained by hepatic, pulmonary, or renal failure, electrolyte imbalance, hypotension, drug intoxication, or a primary lesion of the brain. They called the condition "septic encephalopathy." According to their surveys, 70 percent of patients become disoriented and confused within hours of the onset of severe systemic infection; in a few cases, this state may progress to stupor and coma. Notably there are no signs of asterixis, myoclonus, or focal cerebral disorder but paratonia is common, as is the later development of a polyneuropathy. Rapid changes in water balance may occur, leading to the type of osmotic demyelination discussed below.
The encephalopathic state that occurs with severe systemic infection may also develop independently of sepsis, as a component of a syndrome of multiple organ failure and, according to some authors, a complication of widespread cutaneous burns (Aikawa et al). Others have questioned the validity of this last category and have instead found explanatory electrolyte disorders (particularly hyponatremia), sepsis, or multiple brain abscesses.
It has been useful in clinical work to distinguish these encephalopathies of infection and multiorgan failure from those caused by isolated hepatic or renal disease. The lack of a biochemical marker and the confounding effects of hypotension during sepsis (septic shock) leave doubt as to pathogenesis. Altered phenylalanine metabolism and circulating cytokines have been proposed as causes, without firm evidence. Of interest in two of our fatal cases was the presence of brain purpura, but this has otherwise been an infrequent finding. Here, the white matter of the cerebrum and cerebellum was speckled with myriad pericapillary hemorrhages and zones of adjacent necrosis. This pathologic reaction is nonspecific, having also been seen in cases of viral pneumonia, heart failure with morphine overdose, thrombotic thrombocytopenic purpura and arsenic poisoning.
Disorders of Sodium, Potassium, and Water Balance
Drowsiness, confusion, stupor, and coma, in conjunction with seizures and sometimes with other neurologic deficits, may have as their basis a more or less pure abnormality of electrolyte or water balance. Only brief reference is made here to some of these, such as hypocalcemia, hypercalcemia, hypophosphatemia, and hypomagnesemia, as they are considered in other parts of the text.
Hyponatremia and Syndrome of Inappropriate Antidiuretic Hormone
Hyponatremia is defined as a serum sodium level below 135 mEq/L. The hyponatremic state may be isotonic, hypertonic, or hypotonic, depending on the mechanism of reduced sodium concentration. The hypotonic variety is most common in neurologic practice but one also encounters cases of pseudohyponatremia caused by hyperlipidemia or hyperproteinemia (isotonic), hyperglycemic or mannitol-induced hyponatremia (hypertonic), and also cases of water intoxication. The last of these may be associated with systemic hypovolemic (blood loss, salt wasting), hypervolemic (edematous states such as renal, hepatic or heart failure), or isovolemic states (retention of free water).
Hypotonic isovolemic hypernatremia is most often a result of the syndrome of inappropriate antidiuretic hormone secretion (SIADH). This state is of special importance because it complicates neurologic diseases of many types: head trauma, bacterial meningitis and encephalitis, cerebral infarction, subarachnoid hemorrhage, cerebral and systemic neoplasm, Guillain-Barré syndrome and the effects of certain medications. SIADH is the result of excretion of urine that is hypertonic relative to the plasma.
As the hyponatremia develops, there is a decrease in alertness, which progresses through stages of confusion to coma, often with convulsions. As with many other metabolic derangements, the severity of the clinical effect is related to the rapidity of decline in serum Na. Lack of recognition of this state may allow the serum Na to fall to dangerously low levels, 100 mEq/L or lower.
Most instances of hyponatremia have developed slowly, allowing for maintenance of brain volume by the extrusion from cells of various osmotic substances. Rapid correction of sodium in these circumstances risks a reversal of osmotic gradient and a reduction in brain volume. This, in turn, is associated with a special type of central nervous system demyelination ("osmotic demyelination" and central pontine myelinolysis) discussed below. One's first impulse is to administer NaCl intravenously, but this must be done cautiously to avoid these complications. Most cases of SIADH respond to the restriction of fluid intake—to 500 mL per 24 h if the serum Na is less than 120 mEq/L and to 1,000 mL per 24 h if less than 130 mEq/L. Even when the Na reaches 130 mEq/L, the fluid intake should not exceed 1,500 mL per 24 h. In extreme and rapidly developing (less than 48 h) hyponatremia with stupor or seizures, the mechanisms for maintaining cerebral cellular volume have not yet been engaged and therefore infusion of NaCl is necessary to prevent cerebral edema. The amount of NaCl to be infused can be calculated from the current and the target levels of serum Na by assuming that the infused sodium load is distributed throughout the total body water content (0.6 × weight in kilograms):
The desired volume of normal saline can then be determined by keeping in mind that its sodium concentration is 154 mEq/L and that of 3 percent (hypertonic) saline solution is 513 mEq/L. If hypertonic saline is administered, it is usually necessary to simultaneously reduce intravascular volume with furosemide, beginning with a dose of 0.5 mg/kg intravenously, and to increase the dosage until a diuresis is obtained. (As a rule of thumb, 300 to 500 mL of 3 percent saline, infused rapidly intravenously, will increase the serum sodium concentration by about 1 mEq/L/h for 4 h.) Guidelines to prevent an overly rapid correction of Na are elaborated further on in relation to central pontine myelinolysis (no more rapidly than 10 mmol/L in the first 24 h). Although the syndrome of SIADH is usually self-limiting, it may continue for weeks or months, depending on the type of associated brain disease.
Not all patients with neurologic disorders who manifest hyponatremia have SIADH. Diuretic excess, adrenal insufficiency, and salt wasting also produce hypovolemic hyponatremia as a result of natriuresis. When renal salt wasting is seen in the context of a central neurologic disorder, the process has been termed "cerebral salt wasting" (Nelson et al). Sodium loss in these circumstances is attributable to the production by the heart or brain of a potent polypeptide natriuretic factor. As discussed in Chap. 34, under "Subarachnoid Hemorrhage," the distinction between SIADH and cerebral salt wasting is of more than theoretical importance, insofar as fluid restriction to correct hyponatremia may be dangerous in patients with salt wasting, particularly in those with vasospasm after ruptured intracranial aneurysms.
Arieff emphasized the hazards of postoperative hyponatremia in a series of 15 patients, all of them women, in whom severe hyponatremia followed elective surgery. About 48 h after these patients had recovered from anesthesia, their serum Na fell markedly, at which point generalized seizures occurred, followed by respiratory arrest. We are more familiar with acutely developing hyponatremia in the context of prostate surgery, where large amounts of hypotonic fluids are routinely administered both intravenously and intravascularly. A similar syndrome is known in instances of overly zealous fluid resuscitation in children with diabetic ketoacidosis. The mechanisms of neurologic deterioration in all of these cases is likely to be brain edema.
An important consideration in the management of severe hyponatremia, as mentioned earlier, is the rapidity with which the abnormality is corrected and the danger of provoking central pontine myelinolysis and related brainstem, cerebellar, and cerebral lesions (extrapontine myelinolysis; osmotic demyelination). These issues are considered below, in the section "Central Pontine Myelinolysis."
Hypernatremia (Na >155 mEq/L) and dehydration are observed in diabetes insipidus, the neurologic causes of which include head trauma with damage to the pituitary stalk (Chap. 27), and in nonketotic diabetic coma, protracted diarrhea in infants, and the deprivation of fluid intake in the stuporous patient. The last condition is usually associated with a brain lesion that impairs consciousness. Exceptionally, in patients with chronic hydrocephalus, the hypothalamic thirst center is rendered inactive and severe hypernatremia, stupor, and coma may follow a failure to drink. In hypernatremia from any cause, the brain volume is manifestly reduced. Retraction of the cerebral cortex from the dura has been known to rupture a bridging vein and cause a subdural hematoma.
As is true for hyponatremia, the degree of CNS disturbance in hypernatremia is generally related to the rate at which the serum Na rises. Slowly rising values, to levels as high as 170 mEq/L, may be surprisingly well tolerated. Rapid elevations of sodium shrink the brain, especially in infants. Extremely high levels cause impairment of consciousness with asterixis, myoclonus, seizures, and choreiform movements. In addition, muscular weakness, rhabdomyolysis, and myoglobinuria have been reported.
In hypernatremia with hyperosmolality, the brain retains its volume more effectively than do other organs by a compensatory mechanism that has been attributed to the presence of "idiogenic osmoles," possibly glucose, glucose metabolites, and amino acids. The impairment of neuronal function in this state is not understood. Theoretically one would expect neuronal shrinkage and possibly alteration of the synaptic surface of the cell.
The main clinical effect of hypokalemia (≤2.0 mEq/L) is generalized muscular weakness (see Chap. 48). A mild confusional state had been alluded to in the literature but must be very infrequent. The electrolyte condition is readily corrected by adding K to intravenous fluid and infusing it at no more than 4 to 6 mEq/h. Hyperkalemia (>7 mEq/L) also may manifest itself by generalized muscle weakness, although the main effects are changes in the electrocardiogram (ECG), possibly leading to cardiac arrest.
Other Metabolic Encephalopathies
Limitation of space permits only brief reference to other metabolic disturbances that may present as episodic confusion, stupor, or coma. The most important members of this group are summarized below.
This is defined as an elevation of the serum calcium concentration greater than 10.5 mg/dL. If the serum protein content is normal, Ca levels greater than 12 mg/dL are required to produce neurologic symptoms. However, with low serum albumin levels, an increased proportion of the serum Ca is in the unbound or ionized form (upon which the clinical effects depend), and symptoms may occur with total serum Ca levels as low as 10 mg/dL.
In young persons, the most common cause of hypercalcemia is hyperparathyroidism (either primary or secondary); in older persons, osteolytic bone tumors, particularly meta-static carcinoma and multiple myeloma, are often causative. Less-common causes are vitamin D intoxication, prolonged immobilization, hyperthyroidism, sarcoidosis, and decreased calcium excretion (renal failure).
Anorexia, nausea and vomiting, fatigue, and headache are usually the initial symptoms, followed by confusion (rarely a delirium) and drowsiness, progressing to stupor or coma in untreated patients. A history of recent constipation is common. Diffuse myoclonus and rigidity occur occasionally, as do elevations of spinal fluid protein (up to 175 mg/100 mL). Convulsions are uncommon.
The usual manifestations are paresthesias, tetany, and seizures. With severe and persistent hypocalcemia, altered mental status in the form of depression, confusion, dementia, or personality change can occur. Anxiety to the point of panic attack is also known. Even coma may result, in which case there may be papilledema as a result of increased intracranial pressure. Aside from the raised pressure, the CSF shows no consistent abnormality. This increase in intracranial pressure may be manifest by headache and papilledema without altered mentation or with visual obscurations. Hypoparathyroidism is discussed again further on, under "Acquired Metabolic Diseases Presenting as Progressive Extrapyramidal Syndromes."
Other Electrolyte and Acid–Base Disorders
Severe metabolic acidosis from any cause produces a syndrome of drowsiness, stupor, and coma, with dry skin and Kussmaul breathing. The CNS depression does not correlate with the concentration of ketones. Possibly, there are associated effects on neurotransmitters. It is often not possible to separate the effects of acidosis from those caused by an underlying condition or toxic ingestion.
In infants and children, acidosis may occur in the course of hyperammonemia, isovaleric acidemia, maple syrup urine disease, lactic and glutaric acidemia, hyperglycinemia, and other disorders, which are described in detail in Chap. 37. High-voltage slow activity predominates in the EEG, and correction of the acidosis or elevated ammonia level restores CNS function to normal provided that coma was not prolonged or complicated by hypoxia or hypotension. In uncomplicated acidotic coma, no recognizable neuropathologic change has been observed by light microscopy.
Encephalopathy as a consequence of Addison disease (adrenal insufficiency) may be attended by episodic confusion, stupor, or coma without special identifying features; it is usually precipitated in the addisonian patient by infection or surgical stress. Hemorrhagic destruction of the adrenals in meningococcal meningitis (Waterhouse-Friderichsen syndrome) is another cause. Hypotension and diminished cerebral circulation and hypoglycemia are the most readily recognized metabolic abnormalities; measures that correct these conditions reverse the adrenal crisis in some instances. Laureno (1993) reviewed the various neurologic syndromes that result from electrolytic disorders.
Central Pontine Myelinolysis and Other Patterns of Osmotic Demyelination
Adams, Victor, and Mancall observed a rapidly evolving quadriplegia and pseudobulbar palsy in a young alcoholic man who had entered the hospital 10 days earlier with symptoms of alcohol withdrawal. Postmortem examination several weeks later disclosed a large, symmetrical, essentially demyelinative lesion occupying the greater part of the base of the pons. Over the next 5 years, 3 additional cases (2 alcoholic patients and 1 with scleroderma) were studied clinically and pathologically, and in 1959 these 4 cases were reported by Adams and colleagues under the heading of central pontine myelinolysis (CPM). This term was chosen because it reflects both the main anatomic localization of the disease and its essential pathologic attribute: the remarkably dissolution of the sheaths of myelinated fibers and the sparing of neurons. Once attention was focused on this distinctive lesion, many other reports appeared and it became apparent that other areas of myelin in the brain could be similarly affected. The exact incidence of this disease is not known, but in a series of 3,548 consecutive autopsies in adults, the typical lesion was found in 9 cases, or 0.25 percent (Victor and Laureno).
One is compelled to define this disease in terms of its pathologic anatomy because this stands as its most characteristic feature, but it has been appreciated in recent years that the pons is not the only structure that may be affected. Transverse sectioning of the fixed brainstem discloses a grayish discoloration and fine granularity in the center of the base of the pons. The lesion may be only a few millimeters in diameter, or it may occupy almost the entire ventral pons. There is always a rim of intact myelin between the lesion and the surface of the pons. Posteriorly, it may reach and involve the medial lemnisci and, in the most advanced cases, other tegmental structures as well. Very rarely, the lesion encroaches on the midbrain but inferiorly it does not extend as far as the medulla. Identical extrapontine myelinolytic foci in the internal capsule, deep cerebral white matter and corpus callosum may occur independently ("extrapontine myelinolysis"). Exceptionally, symmetrically distributed lesions are found in the thalamus, subthalamic nucleus, striatum, amygdaloid nuclei, lateral geniculate body, white matter of the cerebellar folia (Wright et al).
Microscopically, the fundamental abnormality consists of destruction of the myelinated sheaths throughout the lesion, with relative sparing of the axons and intactness of the nerve cells of the pontine nuclei. These changes always begin and are most severe in the geometric center of the pons, where they may proceed to frank necrosis of tissue. Reactive phagocytes and glia cells are in evidence throughout the demyelinative focus, but oligodendrocytes are depleted. Signs of inflammation are conspicuously absent.
This constellation of pathologic findings provides easy differentiation of the lesion from infarction and the inflammatory demyelination of multiple sclerosis and postinfectious encephalomyelitis. Microscopically, the lesion resembles that of Marchiafava-Bignami disease (Chap. 41), with which it is rarely associated. In the chronic alcoholic, Wernicke disease is often associated with osmotic demyelination, but the lesions bear no resemblance to one another in terms of topography and histology.
The two sexes are affected equally, and the patients do not fall into any one age period. Whereas the cases first reported had occurred in adults, there are now many reports of the disease in children, particularly in those with severe burns (McKee et al). More than half the cases have appeared in the late stages of chronic alcoholism, often in association with Wernicke disease and polyneuropathy. Most cases occur in the context of other serious medical conditions, and diseases with which osmotic demyelination has been conjoined are chronic renal failure being treated with dialysis, hepatic failure, advanced lymphoma, cancer, cachexia from a variety of other causes, severe bacterial infections, dehydration and electrolyte disturbances, acute hemorrhagic pancreatitis, and pellagra. The changes in serum sodium concentration, with which the process is closely aligned, are discussed below.
In many patients there are no symptoms or signs that betray the pontine lesion, presumably because it is so small, extending only 2 to 3 mm on either side of the median raphe and involving only a small portion of the corticopontine or pontocerebellar fibers. In others, its presence is obscured by coma from a metabolic or other associated disease. Prior to the inception of MRI only a minority of cases, exemplified by the first patient observed by Adams, Victor, and Mancall, were recognized during life. In this patient, a serious alcoholic with delirium tremens and pneumonia, there evolved, over a period of several days, a flaccid paralysis of all 4 limbs and an inability to chew, swallow, or speak (thus simulating occlusion of the basilar artery). Pupillary reflexes, movements of the eyes and lids, corneal reflexes, and facial sensation were spared. In some instances, however, conjugate eye movements are limited, and there may be nystagmus. With survival for several days, the tendon reflexes become more active, followed by spasticity and extensor posturing of the limbs on painful stimulation. Some patients are left in a state of mutism and paralysis with relative intactness of sensation and comprehension (pseudocoma, or locked-in syndrome).
The capacity of CT and especially MRI to visualize the pontine lesion has greatly increased the frequency of premortem diagnoses. The MRI discloses a characteristic "batwing" lesion of the pons in typical cases (Fig. 40-6), although this change may become evident only several days after the onset of symptoms. Brainstem auditory evoked responses also disclose the lesions that encroach upon the pontine tegmentum.
T2-weighted MRI showing the typical lesion of central pontine myelinolysis in an alcoholic patient.
Variants of this syndrome are being encountered with increasing frequency. Two of our elderly patients, with confusion and stupor but without signs of corticospinal or pseudobulbar palsy, recovered; however, they were left with a severe dysarthria and cerebellar ataxia lasting many months. After 6 months, these patients' nervous system function was essentially restored to normal. In reference to the pathogenesis of this lesion, originally both patients had serum Na levels of 99 mEq/L, but information about the rate of correction of serum Na was not available. Another of our patients developed a typical locked-in syndrome after the rapid correction of a serum sodium of 104 mEq/L. He showed large symmetrical lesions of the frontal cortex and underlying white matter but no pontine lesion (by MRI).
Brainstem infarction caused by basilar artery occlusion may be simulated by pontine myelinolysis. Sudden onset or step-like progression of the clinical state, asymmetry of long tract signs, and more extensive involvement of tegmental structures of the pons as well as the midbrain and thalamus are the distinguishing characteristics of vertebrobasilar thrombosis or embolism. On MRI studies, an evolving infarction shows signal changes on diffusion-weighted imaging, while the primary finding in osmotic demyelination is brightness of the T2-weighted images. Massive pontine demyelination in acute or chronic relapsing multiple sclerosis rarely produces a pure pontine syndrome. The clinical features and context provide the clues to correct diagnosis.
Etiology and Pathogenesis
As mentioned in the section on hyponatremia, a rapid rise in serum osmolality to normal or higher-than-normal levels is an almost obligate antecedent of this process. One encounters this most commonly in the rapid correction of hyponatremia. In cases related to the correction of hyponatremia, the initial serum sodium concentration is less than 130 mEq/L and usually much lower; this was the case in all the patients reported by Burcar and colleagues and by Karp and Laureno. Laureno (1983) demonstrated the importance of serum sodium in the pathogenesis of this disease experimentally. Dogs made severely hyponatremic (100 to 115 mEq/L) had the electrolyte disorder corrected rapidly by infusion of hypertonic (3 percent) saline; this led to spastic quadriparesis and pontine and extrapontine lesions were found at autopsy, indistinguishable in their distribution and histologic features from those of the human disease. Hyponatremia alone or slowly corrected hyponatremia (<15 mEq/dL in the initial 24 h) did not produce the disease.
McKee and colleagues adduced that in burn patients, extreme serum hyperosmolality was the important factor in the pathogenesis of demyelination. They found the characteristic pontine and extrapontine lesions in 10 of 139 severely burned patients who were examined after death. Each of their patients with CPM had suffered a prolonged, nonterminal episode of severe serum hyperosmolality, which coincided temporally with the onset of the lesion, as judged by its histologic features. Hyponatremia was not prominent and no other independent features could explain the changes. These observations suggest that rapidly rising osmolarity may be a cause of the osmotic demyelination syndromes.
At the present time all one can say is that specific myelinated regions of the brain, most often but not exclusively the center of the base of the pons, have a susceptibility to rapid increase in serum osmolality.
Karp and Laureno, on the basis of their experience and that of Sterns and colleagues, have suggested that the hyponatremia be corrected by no more than 10 mEq/L in the initial 24 h and by no more than about 21 mEq/L in the initial 48 h.