Review Article

Metabolic
Address correspondence to
Dr Jennifer Frontera, Mount
Sinai School of Medicine,
Department of Neurology,

Encephalopathies in One Gustave Levy Place, Box

1136, New York, NY 10029,
Jennifer.Frontera@

the Critical Care Unit mountsinai.org.

Relationship Disclosure:
Dr Frontera reports no
disclosure.
Jennifer A. Frontera, MD Unlabeled Use of
Products/Investigational
Use Disclosure:
ABSTRACT Dr Frontera reports no
disclosure.
Purpose of Review: This article summarizes the most common etiologies and ap- * 2012, American Academy of
proaches to management of metabolic encephalopathy. Neurology.
Recent Findings: Metabolic encephalopathy is a frequent occurrence in the intensive
care unit setting. Common etiologies include hepatic failure, renal failure, sepsis,
electrolyte disarray, and Wernicke encephalopathy. Current treatment paradigms typ-
ically focus on supportive care and management of the underlying etiology. Directed
therapies that target neurochemical and neurotransmitter pathways that mediate
encephalopathy are not currently available and represent an important area for future
research. Although commonly thought of as reversible neurologic insults, delirium
and encephalopathy have been associated with increased mortality, prolonged length
of stay and hospital complications, and worse long-term cognitive and functional
outcomes.
Summary: Recognition and treatment of encephalopathy is critical to improving out-
comes in critically ill patients.

Continuum Lifelong Learning Neurol 2012;18(3):611–639.

INTRODUCTION important indicators of metabolic en-

Metabolic encephalopathy, often re- cephalopathy include acute or subacute
ferred to as toxic-metabolic encepha- onset concordant with organ system
lopathy, encompasses a diverse array of dysfunction, a fluctuating course, lack
systemic conditions that result in global of a fixed focal deficit, tremor, asterixis
cerebral dysfunction in the absence of (typically bilateral; unilateral asterixis
structural brain injury. Metabolic en- suggests an underlying structural ab-
cephalopathies are common in the inten- normality), multifocal myoclonus, and
sive care unit (ICU) setting, particularly relative sparing of the cranial nerves.
in the context of conditions such as li- Preservation of pupillary function is a
ver or kidney failure, sepsis, electrolyte hallmark of metabolic encephalopathy,
imbalance, or endocrine dysfunction. although the pupils are often involved in
Encephalopathy is often multifactorial anticholinergic drug toxicity, glutethi-
because of the interaction of different mide ingestion, and advanced Wernicke
organ systems. Metabolic encephalop- or hepatic encephalopathy. It is impor-
athy can present as an acute confusional tant to note that both hypoglycemia
state, decreased level of arousal, delir- and hyperglycemia can cause focal def-
ium (which is characterized by fluctuat- icits that mimic a structural lesion. Addi-
ing levels of attentiveness), or coma. It is tionally, metabolic disarray can unmask
important to distinguish metabolic en- or worsen old or subclinical focal neu-
cephalopathies from structural or elec- rologic deficits, such as deficits from an
trophysiologic neurologic injuries. Some old stroke (sometimes this phenomenon
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 611

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINTS
h Preservation of pupillary is referred to as ‘‘peeling the onion’’). serotonin, dopamine, +-aminobutyric
function is a hallmark Decorticate and decerebrate posturing acid [GABA], tryptophan, melatonin,
of metabolic can occur with advanced encephalopathy. and glutamate) and cytokines (such as
encephalopathy. interleukins and interferons) have also
EPIDEMIOLOGY been implicated in the pathogenesis of
h Focal examination
findings should alert the The overall prevalence of metabolic encephalopathy. Because the patho-
practitioner to the encephalopathy is difficult to estimate, physiology of encephalopathy is com-
possibility of an but it is usually inferred from the litera- plex, the etiology is often multifactorial
underlying structural ture on delirium, defined as fluctuations and many integrated regions of the
lesion, although in attentiveness. Delirium is common in brain may be simultaneously affected.
hypoglycemia and hospitalized patients, occurring in up to
hyperglycemia can 70% of ICU patients, 16% of postYacute ETIOLOGY
cause focal deficits. care patients, 10% of emergency depart- Major causes of metabolic encephalop-
h Neurotransmitter ment patients, and 42% of hospice pa- athy are listed in Table 7-1 and Table 7-2.
dysfunction and tients.1 As many as 50% of older patients Further discussion of several of these
interruption of the develop delirium at some point during etiologies follows below.
ascending reticular their hospital stay.2 The prevalence of
activating system are delirium in general ICU patients is as Hepatic Encephalopathy
main factors mediating
high as 80% in prospective studies.3 The acuity with which liver failure devel-
encephalopathy.
ops is directly related to the develop-
PATHOPHYSIOLOGY ment of encephalopathy. Patients with
The pathophysiology of metabolic ence- acute liver failure (ALF) are particularly at
phalopathy depends in part on the eti- risk. Hepatic encephalopathy grading is
ology (which will be discussed below). shown in Table 7-3.4 Although the
In general, a disruption in the attention pathogenesis of hepatic encephalop-
and arousal centers of the brain is the athy is not entirely understood, it is
cause of encephalopathy in all patients, thought to be a reversible form of neu-
regardless of etiology. The area primar- rologic dysfunction primarily due to
ily responsible for arousal and attentive- ammonia-induced neurotoxicity. Arte-
ness is the ascending reticular activating rial ammonia levels are often elevated
system (ARAS), which encompasses an in the context of ALF, but a normal am-
area between the midpontine tegmen- monia level does not preclude the di-
tum and the anterior cingulate gyrus. agnosis. Ammonia is produced either
Widespread input reaches the ARAS by glutamine metabolism at the mito-
from the cortex, spinal cord, visual and chondrial level or by catabolism of ni-
auditory centers, thalamus, hypothala- trogenous sources. Its main neurotoxic
mus, and hippocampus. Interruptions effect is related to astrocyte swelling and
or disturbances along any of these tracts dysfunction.5 Additionally, stimulation
may lead to encephalopathy. Additional of NMDA receptors with subsequent
regions that are critical to attention, nitric oxide release and vasodilatation
memory, and executive function, such can be caused by the metabolism of
as the cingulate gyrus, anterior and glutamine into glutamate and ammonia.
medial thalamic nuclei, orbital frontal Hyperemia and cerebral edema may in
cortex, frontal projections to the thala- turn result from ammonia-induced vaso-
mus, fornix, mammillary bodies, hippo- dilatation.6 Patients with fulminant he-
campus, and caudate, can be patic failure have been found to have
compressed or injured, leading to ence- impaired cerebral autoregulation.7 Other
phalopathy. Abnormalities in neuro- mechanisms, such as activation of the
transmitters (such as acetylcholine, aquaporin-4 water channel protein on
612 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

TABLE 7-1 Etiology of Metabolic Encephalopathy in
Adult Populations

b Organ Failure b Infectious

Hepatic encephalopathy Sepsis-associated encephalopathy
Renal failure/uremia b Toxins
Pulmonary failure/ Sedatives
hypoxemia/hypercapnia
Anticholinergics
b Electrolyte Abnormalities
Salicylates
Hypernatremia/hyponatremia
Arsenic
Hypercalcemia/hypocalcemia
Lead
Hypermagnesemia/hypomagnesemia
Ethylene glycol
Hypophosphatemia
Methanol
Hyperglycemia/hypoglycemia
Cyanide
b Endocrine
Carbon monoxide
Hyperthyroid/hypothyroid
Organophosphates
Hyperparathyroid/hypoparathyroid
b Systemic
Adrenal insufficiency
Hypertensive crisis/posterior
b Nutritional reversible encephalopathy
syndrome
Acute thiamine deficiency
(Wernicke encephalopathy) Hyperviscosity syndromes
Vitamin B12 deficiency b Withdrawal States
Niacin deficiency (pellagra) Alcohol/benzodiazepine
withdrawal
Folate deficiency
Opiate withdrawal
Marchiafava-Bignami disease
Acute refeeding syndrome
(hypophosphatemia)

astrocytes, oxindole (a tryptophan me- occurs in 80% of comatose patients with

tabolite) abnormalities, and cate- acute hepatic failure and is the leading
cholamine and other neurotransmitter cause of death in these patients.9
abnormalities, may be involved in gen- Between 86% and 95% of patients with
erating hepatic encephalopathy. Addi- grade 3 or 4 encephalopathy have ele-
tionally, inflammatory markers such as vated ICP, primarily related to cerebral
tumor necrosis factor have been noted edema.10 Because standard head CT is
to be elevated in patients with hepatic known to be insensitive for estimating
encephalopathy, particularly those with ICP,11 ICP monitoring should be con-
elevated intracranial pressure (ICP).8 sidered in patients with grade 3 or
All of the above factors create an grade 4 encephalopathy and in all pa-
abnormal neurochemical milieu leading tients unable to follow commands, as
to cerebral edema. Cerebral edema shown in Case 7-1. The only accurate
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 613

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

TABLE 7-2 Etiology of Metabolic Encephalopathy in

Pediatric Populations

b Organ Failure Lysosomal storage disorders

Hepatic encephalopathy Purine and pyrimidine disorders
Renal failure/uremia Metal metabolism defects
Pulmonary failure/hypoxemia/ Porphyrias
hypercapnia
b Infection
b Electrolyte Abnormalities
Sepsis-associated encephalopathy
Hypernatremia/hyponatremia
Parainfectious encephalopathy/acute
Hypercalcemia/hypocalcemia disseminated encephalomyelitis
Hypermagnesemia/hypomagnesemia Reye syndrome
Hypophosphatemia b Toxins
Hyperglycemia/hypoglycemia Sedatives
b Inborn Errors of Metabolism Anticholinergics
Defects in gluconeogenesis Salicylates
Fatty acid oxidation disorders Arsenic
Amino acid disorders Lead
Organic acidemias Ethylene glycol
Urea cycle disorders Methanol
Carbohydrate disorders Cyanide
Mitochondrial disorders Carbon monoxide
Peroxisomal disorders Organophosphates

way to assess ICP and the efficacy of support the use of ICP monitoring are
cerebral edema treatment is with direct lacking, although some studies suggest
ICP monitoring. This is particularly that ICP monitoring not only can iden-
important in patients with marginal tify subclinical ICP spikes but also can
neurologic examinations or patients lead to therapeutic interventions and
receiving sedation, as the response to provide useful prognostic information.10
medical treatment of edema may be In one study, elevated ICP occurred in
impossible to measure clinically. 95% of patients with fulminant liver
Indeed, the US Acute Liver Failure failure. Eighty-two episodes of elevated
Study Group recommends ICP monitor- ICP (with a median value of 33 mm Hg
ing in grade 3 or grade 4 patients who for a median duration of 60 minutes)
are candidates for orthotopic liver trans- occurred in 21 patients who were
plant (OLT) and in some patients who monitored for a median of 6 days.10
are not candidates for OLT but who This study suggests that ICP eleva-
may have survival benefit with aggres- tions are both severe and frequent and
sive treatment of advanced cerebral may require evaluation and medical
encephalopathy.11 Randomized trials to management.
614 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 a
TABLE 7-3 Hepatic Encephalopathy Grade

Level of Consciousness/ Psychiatric Neuromuscular

Grade Cognitive Function Symptoms Function
1 Sleep disturbance Euphoria/ Tremor
depression
Mild confusion Incoordination
Impaired computations +/- Asterixis
2 Inattentive Irritability Asterixis
Moderate confusion Decreased Slurred speech
inhibitions
Disorientation to time Personality Impaired handwriting
changes
3 Marked confusion Anxiety or apathy Slurred speech
Completely disoriented Inappropriate or Ataxia
bizarre behavior
Lethargic but arousable Paranoia, anger, Asterixis
or rage
Command following Nystagmus
Hypoactive or
hyperactive reflexes
4 NonYcommand Coma Dilated pupils, loss of
following cranial nerve reflexes
Coma Signs of herniation
Flexor or extensor
posturing
Loss of reflexes
a
Reprinted with permission from Frontera JA, Kalb T. Neurological management of fulminant hepatic failure.
Neurocrit Care 2011;14(2):318Y327.4 www.springerlink.com/content/86j22313r6228703/.

Case 7-1
A 23-year-old woman with no significant medical history was admitted to the medical intensive care unit
after swallowing a bottle of acetaminophen. She was intubated for airway protection. Her examination
off sedation was notable for an inability to follow commands, reactive pupils, intact corneal and
oculocephalic reflexes, extensor posturing of both upper extremities, and triple flexion of both lower
extremities. She had elevated bilirubins and transaminases. Her international normalized ratio (INR) was
6.8, fibrinogen was 415 mg/dL, platelet count was 30,000/2L, blood urea nitrogen was 70 mg/dL, and
creatinine was 3.6 mg/dL. She was anuric. Her head CT demonstrated global cerebral edema with
effacement of the basal cisterns (Figure 7-1). She was diagnosed with fulminant liver failure, given
N-acetylcysteine, and evaluated as a liver transplant candidate. Her coagulopathy was reversed. A
parenchymal intracranial pressure (ICP) monitor was placed, and her ICP was found to be 35 mm Hg. The
head of her bed was elevated; she received adequate pain control and sedation; her cerebral perfusion
pressure was maintained at greater than 60 mm Hg; and she was normothermic and normocarbic.

Continued on page 616

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 615

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

Continued from page 615

Despite this, her ICP

remained elevated.
Osmotherapy was
initiated with a
bolus of 23% saline
(30 mL). Because
she had anuric
renal failure, she
was placed on
dialysis (continuous
veno-venous
hemofiltration)
with adjustment of
the replacement
fluid (hypertonic
saline) to achieve
a sodium of
150 mEq/L to
155 mEq/L and FIGURE 7-1 Noncontrast head CT demonstrating global cerebral edema, effacement of basal
cisterns (red arrow), and loss of sulci (blue arrows).
titrated to an ICP
less than 20 mm Hg.
Comment. Because of the patient’s poor neurologic examination and head CT findings, it is
important both to measure and to treat her elevated ICP and cerebral edema. The first factor to
consider is placement of an ICP monitor, which will require reversal of the patient’s coagulopathy. A
parenchymal ICP monitor offers the most reliable option for tracking ICP with a minimal risk of
hemorrhage. Reasonable options to reverse coagulopathy include 10 mg of IV vitamin K, prothrombin
complex concentrates (PCCs), and 15 mL/kg to 20 mL/kg of IV fresh frozen plasma (for an INR greater
than 2.0 to replace factor V not found in PCC). Although recombinant factor VIIa is an alternative to
reverse an elevated prothrombin time or INR, it is important to remember that factors II, V, IX, and X will
not be repleted with factor VIIa alone. Additionally, the rates of arterial thrombosis are as high as
8.5% with recombinant factor VIIa. Combining factor VIIa and PCCs or fresh frozen plasma is not
advised as this can substantially increase the risk of disseminated intravascular coagulopathy. If her
fibrinogen had been less than 100 mg/dL, 10 units of cryoglobulin should have been administered. A
platelet transfusion should be given to reach a goal platelet count of 50,000/2L, and desmopressin can
be administered to treat platelet dysfunction due to uremia.

The choice of ICP monitor has limited and many centers may not
evolved over time. Epidural monitors, stock such ICP monitors. The major
which now have an orphan indication theoretical benefit of epidural monitors
for ICP monitoring in patients with is a reduced risk of hemorrhage with
coagulopathy, were commonly used in insertion. In a 1992 survey of centers
patients with fulminant hepatic failure in treating patients with ALF, epidural
the past. The major limitations of epi- ICP monitors were associated with a
dural monitors are accuracy and drift.13 3% risk of hemorrhage compared to
Additionally, since the indications for 18% with subdural monitors and 13%
epidural ICP monitoring are few, ex- with ‘‘parenchymal’’ monitors.14 In this
perience in insertion may be similarly study ‘‘parenchymal’’ monitors typically

616 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

referred to intraventricular drains rather has been recommended for coagulop-
than ICP monitors placed solely in the athy reversal prior to ICP monitor place-
parenchyma. This study did not provide ment12; however, use of rFVIIa alone
any information about coagulopathy does not replete other depleted coagu-
reversal prior to ICP monitor place- lation factors (notably II,V, IX, and X)
ment. In a separate study of patients and is associated with a higher risk of
with ALF, parenchymal ICP monitoring DIC than other reversal agents. Arterial
(not intraventricular ICP monitoring) thrombotic events are more common
was associated with a 14% hemorrhage with rFVIIa than with other agents and
rate, although death or neurologic have been reported to occur in as many
impairment did not occur in any patient as 8.5% of patients.17 Alternatives to
as a result of ICP monitorYrelated hem- rFVIIa include fresh frozen plasma (FFP)
orrhage.10 Whether the risk of hemor- and unactivated prothrombin complex
rhage after adequate correction of concentrates (PCCs), sometimes
coagulopathy is higher with intraparen- referred to as factor IX concentrates.
chymal monitors than with epidural ICP PCCs come as a powder that requires
monitors remains unclear. At my institu- reconstitution in 20 mL of sterile water
tion, we typically use intraparenchymal per vial and are typically dosed as
monitors because they are more accu- 50 international units (IU)/kg. A dose
rate.15 Intraventricular monitor place- of 1 IU/kg will lead to a 1% correction in
ment is not recommended because of coagulation factors. Although some
the increased risk of bleeding.12 advocate a lower dose of 20 IU/kg, a
Transcranial Doppler assessment of 50% correction is considered necessary
pulsatility index (peak systolic flow to restore hemostasis in most coagulo-
velocity minus end diastolic flow veloc- pathic patients with an elevated interna-
ity divided by mean flow velocity) may tional normalized ratio (INR).18 Most
be useful in patients who are not PCC moieties include varying amounts
candidates for ICP monitor placement. of factors II, VII, IX, and X and protein C
Pulsatility indices greater than 1.5 are and S. A common criticism of available
considered abnormal and can give a US versions of PCCs is the relatively low
rough gauge of the likelihood of ele- dose of factor VII. However, reversal of
vated ICP but not a quantified estimate coagulopathy can be achieved quickly
of ICP. Additionally, transcranial Dop- with minimal volume and a lower
pler does not provide a means of expense than rFVIIa. Because PCCs do
continuous ICP monitoring and may not contain factor V, FFP may be
have suboptimal sensitivity and specific- indicated to replace this factor (particu-
ity, according to some studies.16 larly if the INR is greater than 2.0). It is
Aggressive coagulopathy reversal important to recognize that multiple
prior to ICP monitor placement is crit- doses of rFVIIa or combining rFVIIa
ical. Liver failure can lead to coagulation with PCCs can dramatically increase
abnormalities at multiples levels of the the risk of DIC and DIC-related compli-
cascade. Synthesis disorders can lead to cations and is therefore not advised. An
coagulation factor and fibrinogen defi- alternative to rFVIIa and PCCs is FFP
ciencies. Additionally, thrombocytopenia dosed at 15 mL/kg to 20 mL/kg. Limi-
due to splenic sequestration, platelet tations of FFP include a longer time to
dysfunction due to uremia, and dissemi- administration (because of the need for
nated intravascular coagulopathy (DIC) thawing) and a higher volume load,
can occur in the context of hepatic which can be problematic in patients
failure. Recombinant factor VIIa (rFVIIa) with liver failure, particularly if the liver
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 617

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINTS
h Cerebral edema and failure is complicated by acute renal veno-venous hemofiltration (CVVH) is
elevated intracranial failure. Whether rFVIIa, PCCs, or FFP is used or by adjusting the sodium con-
pressure are common in administered, all patients should receive centration in the dialysis bath. The US
patients in fulminant 10 mg of IV vitamin K.12 Additionally, Acute Liver Failure Study Group recom-
liver failure with hepatic cryoprecipitate is indicated if fibrinogen mends early introduction of renal
encephalopathy and levels are less than 100 mg/dL. Uremia- replacement therapy in patients with
should be treated induced platelet dysfunction in the progressive oliguria, volume overload,
aggressively. context of renal insufficiency can be or electrolyte disarray. CVVH is the pre-
h Intracranial pressure corrected with 0.3 2g/kg of desmopres- ferred renal replacement modality be-
monitoring is sin. Coagulopathy can be considered cause it causes fewer precipitous fluid
recommended in adequately corrected for placement of shifts, has a smoother hemodynamic
patients with grade 3 or an intracranial monitor when laboratory profile, and is able to rapidly and con-
4 encephalopathy, in values are as follows: INR less than 1.5, tinuously address electrolyte disturb-
patients unable to platelets greater than 50,000/mm3, fibri- ance and manage osmotic therapy.12
follow commands who
nogen greater than 100 mg/dL, and a Beyond facilitating hyperosmotic ther-
are candidates for liver
normal partial thromboplastin time. apy, renal replacement therapy may
transplant, and in some
patients who are not
Whether coagulation factors need to have other ICP-lowering effects. For
orthotopic liver be corrected for the entire time an ICP example, in a porcine model study of
transplant candidates monitor is in place or correction is nec- fulminant liver failure, albumin dialysis
but who may have essary only during placement and was shown to mitigate ICP elevations
survival benefit with removal of devices is a matter of de- independent of its effects on cerebral
aggressive treatment bate.10 The expense of continued edema.21
of advanced cerebral aggressive correction can be consider- Aside from ICP management, detec-
encephalopathy. able, and medical complications such as tion and management of seizures is
volume overload, thrombosis, and DIC another important consideration in
can occur with ongoing transfusions. patients with hepatic encephalopathy.
Additionally, continued coagulation cor- Although the true incidence of seizures
rection may mask spontaneous liver re- in patients with liver failure is not clear,
covery. Plasma exchange has been patients with hepatic encephalopathy
shown to be effective and may be an should be considered for EEG monitor-
option for patients with persistent coa- ing because seizures can elevate ICP.
gulopathy.19 In patients who already Because fulminant liver failure is asso-
have a dialysis catheter in place and ciated with a relatively GABA-ergic state,
who can tolerate an interruption from with correspondingly low NMDA activ-
renal replacement therapy to undergo ity, seizures primarily due to liver failure
plasma exchange, this may be an espe- are thought to be rare. However, acute
cially attractive option. hyperammonemia can be associated
Some evidence exists that using a with mitochondrial dysfunction and lead
protocol for ICP management can im- to seizures, as it does in Reye syndrome
prove neurologic outcomes.10 Figure 7-24 and other inborn errors of metabo-
details a flow diagram for management lism.22 It is also important to keep in
of elevated ICP in patients with fulmi- mind that comorbidities of liver failure,
nant hepatic failure. Because 50% of such as hypoglycemia, hyponatremia,
patients with fulminant liver failure have uremia, and intracranial hemorrhage,
concomitant acute renal failure,20 dialy- are common causes of seizure. In one
sis can be an effective way to maintain a series of 42 patients with ALF, 32% had
hyperosmolar state and treat elevated seizure activity.23 This study did not
ICP. This is achieved by using hyper- provide any data on imaging, medica-
tonic replacement fluid if continuous tions, or metabolic disarray that would
618 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 FIGURE 7-2 Management of increased intracranial pressure or malignant edema.

a
Caution with low ejection fraction as there is a risk of flash pulmonary edema. Do not use with peripheral
lines.
b
Caution with renal failure.
c
Can use a bag valve mask on patient for immediate effect. Prolonged use can cause ischemia and
attenuation of effect.
d
Caution as there is a risk of intensive care unit myopathy/neuropathy.
Reprinted with permission from Frontera JA, Kalb T. Neurological management of fulminant hepatic failure. Neurocrit Care 2011;14(2):
4
318Y327. www.springerlink.com/content/86j22313r6228703/.

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 619

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

help identify the etiology for seizure. acetaminophen metabolite N-acetyl-

Additionally, some patients included in p-benzoquinone imine is normally
this study were posttransplant, a state detoxified. NAC may have additional
that is associated with acute GABA- antioxidant and vasoactive benefits. IV
ergic withdrawal and represents a NAC should also be strongly consid-
high-risk time for seizure activity. Be- ered for patients with early-stage nonY
cause hairline EEG was used rather acetaminophen-induced ALF, since
than a full montage and some of the NAC has a generally favorable safety
patients were paralyzed (thereby limit- profile and increasing evidence sug-
ing the detection of clinical seizures), gests that NAC administration is asso-
the actual seizure rate quoted in this ciated with improved transplant-free
study may be an underestimate. A survival among patients with other
separate study identified a seizure rate etiologies of fulminant liver failure.28
of 25% in patients with ALF but sim- Treatment should begin immediately
ilarly did not use continuous EEG with a loading dose of 150 mg/kg in
monitoring or account for confounding 500 mL dextrose 5% over 30 minutes,
factors that could lead to seizure.24 A followed by a maintenance dose of
suggestion would be to use full mon- 50 mg/kg over 4 hours and then
tage continuous EEG monitoring in all 125 mg/kg in 1000 mL dextrose 5%
grade 3 or grade 4 patients, as well as in over 19 hours. The NAC infusion is typi-
liver failure patients with intracranial cally continued until the INR is less
hemorrhage or clinical seizure epi- than 1.5. In those who develop a hyper-
sodes. It is important to keep in mind sensitivity reaction to IV NAC, cortico-
that nonconvulsive seizures occur in steroids and antihistamines should be
10% to 20% of critically ill patients.25Y27 administered and the infusion rate re-
To allow for detection of subtle clinical duced by 50%.
spells, paralysis should be avoided, if Although lactulose and nonabsorb-
possible. The utility of antiepileptic or able antibiotics such as neomycin and
phenytoin prophylaxis is unclear.23,24 rifaximin are frequently used in chronic
Patients with clinical or electrographic liver failure and can help modulate the
seizures should receive antiepileptic production and absorption of nitroge-
treatment, and prophylaxis can be con- nous moieties that contribute to hepatic
sidered for those who have evidence encephalopathy, the use of either lac-
of elevated ICP, very severe cerebral tulose or nonabsorbable antibiotics in
edema, or intracranial hemorrhage, in ALF is controversial. Lactulose has the
whom a seizure might cause herniation. side effect of producing gaseous ab-
Triphasic waves are a common EEG dominal distention, which may obscure
finding among patients with hepatic the operating field during OLT. In
encephalopathy. rare cases, lactulose can lead to mega-
Liver transplant is the definitive treat- colon and bowel ischemia, as well as
ment for ALF and hepatic encepha- excessive diarrhea and intravascular vol-
lopathy; however, other supportive ume depletion. Lactulose failed to dem-
measures can be enacted while awaiting onstrate an impact on outcomes in one
transplant. N-acetylcysteine (NAC) is retrospective study of patients with
commonly administered for acetami- ALF.12 Similarly, data are insufficient to
nophen overdose (via either oral or IV support the use of either rifaximin or
routes) and acts to replenish gluta- neomycin in ALF. Neomycin is not
thione stores. Glutathione conjugation recommended because of the risks of
is the mechanism by which the harmful nephrotoxicity.12
620 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINT
The most commonly used prognos- and creatinine are not directly associated h Uremic encephalopathy
tic classification scheme for ALF, the with encephalopathy. Many of the os- typically occurs when
King’s College Hospital criteria, was motically active toxins that contribute to the glomerular filtration
originally derived from a cohort of pa- uremic encephalopathy have not been rate is less than 10% of
tients with acetaminophen-induced ALF identified, although presumably they are normal and is more
and uses simple measures that nega- small- to moderate-sized water-soluble fulminant in patients
tively predict transplant-free survival molecules since dialysis improves en- with acute renal failure.
(Table 7-4).4,29 cephalopathy. Some potential culprit
entities that have been examined in-
Uremic Encephalopathy clude phenylethanolamine, 3-carboxy-
Renal insufficiency that results in a failure 4-methyl-5-propyl-2-furanpropionate,
to clear nitrogenous waste products can hippurate, methylguanidine, indoleace-
lead to uremic encephalopathy. Typi- tate, and altered levels of norepinephrine,
cally, the glomerular filtration rate is less acetylcholine, and GABA. Acute renal
than 10% of normal when uremic en- failure has been associated with a pro-
cephalopathy occurs. Uremic encephal- inflammatory state and breakdown of
opathy is usually more severe in patients the blood-brain barrier, which may con-
with acute rather than chronic renal in- tribute to encephalopathy. Electrolyte
sufficiency. Patients with hepatorenal abnormalities that can occur in the con-
syndrome are particularly prone to en- text of renal insufficiency, including
cephalopathy because of the inability of hyponatremia, hyperkalemia, secondary
the kidney to clear plasma ammonia. Al- hyperthyroidism, and acidosis, can also
though the severity of renal insufficiency contribute to encephalopathy. Uremic en-
correlates with the risk of uremic en- cephalopathy is treatable by hemodialysis
cephalopathy, absolute levels of urea and electrolyte correction or transplant.

TABLE 7-4 Predictors of Increased Mortality Without Emergent

Transplant: King’s College Hospital Criteriaa,b

Acetaminophen-Induced NonYAcetaminophen-Induced Acute

Acute Liver Failure Liver Failure
Arterial pH G7.30 PT 9100 seconds (international normalized
ratio [INR] 96.5)
OR
OR 3 of the following:
Grade 3 or grade 4 encephalopathy
Patient age G10 or 940 years
AND
Hepatitis due to non-A/non-B virus,
Serum creatinine 9300 2g/mL
halothane, or drug reaction
(3.4 mg/dL) AND prothrombin
time (PT) greater than Delayed onset of encephalopathy (91 week
100 seconds (INR 96.5) after onset of jaundice)
Serum total bilirubin 918 mg/dL (308 mmol/L)
PT greater than 50 seconds (INR 93.5)
a
Modified from O’Grady JG, Alexander GJ, Hayllar KM, et al. Early indicators of prognosis in fulminant hepatic
failure. Gastroenterology 1989;97(2):439Y445.29 B 1989, with permission from Elsevier. www.gastrojournal.org/
article/0016-5085(89)90081-4/abstract.
b
Reprinted with permission from Frontera JA, Kalb T. Neurological management of fulminant hepatic failure.
Neurocrit Care 2011;14(2):318Y327.4 www.springerlink.com/content/86j22313r6228703/.

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 621

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

Seizures occur in 14% to 33% of pa- be dosed 3 times a day because of its
tients with uremia and may be general- shorter half-life in the context of ure-
ized or focal (when an underlying brain mia. Phenytoin is minimally dialyzable.
injury such as a subdural hematoma is Other antiepileptic options include val-
present). Seizures are more common in proic acid and levetiracetam (500 mg
acute renal failure and are exacerbated to 750 mg twice daily), which can also
by electrolyte imbalances such as hy- ameliorate myoclonus.
ponatremia. Seizures must be distin- Dialysis disequilibrium syndrome is a
guished from the asterixis, myoclonus, distinct entity from uremic encephalop-
and multifocal myoclonus that can oc- athy and is characterized by confusion,
cur with uremia. EEG is essential for headache, nausea, vomiting, tremor,
this differentiation. Other typical EEG and myoclonus that occur when nitro-
changes include diffuse slowing, frontal genous products are cleared quickly
intermittent rhythmic theta (Figure 7-3), through dialysis. A urea gradient may
and paroxysmal bilateral high-voltage be established during rapid dialysis
delta and triphasic waves. Phenytoin is causing brain urea content to exceed
frequently used in patients with uremia blood urea concentrations. This may
to control seizures, but the low serum cause an influx of water into brain cells,
albumin that can occur in the context resulting in cerebral edema. Alternately,
of renal insufficiency can increase the sodium and potassium cations may be
fraction of active, nonYprotein-bound displaced from organic anions because
phenytoin. As a result, free phenytoin of a relatively acidotic milieu established
levels are more useful for monitoring during dialysis. These osmotically active
than total serum levels. Phenytoin should moieties may lead to cerebral edema.

FIGURE 7-3 EEG demonstrating frontal intermittent rhythmic slowing.

622 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINTS
Dialysis disequilibrium can occur up to 8 SAE should not be confused with pri- h Any patient in renal
hours after dialysis and can take several mary CNS infections such as abscess, failure with focal deficits
days to spontaneously resolve. meningitis, and encephalitis. The etiol- should undergo brain
Dialysis dementia, thought to be re- ogy of SAE is likely multifactorial and due imaging to evaluate for
lated to aluminum toxicity incurred to inflammatory mediators, cytokines, intracranial
through dialysis, causes progressive per- reduction in monamine transmitters, hemorrhage because
manent memory loss, dysarthria, facial disruption of the blood-brain barrier, subdural hemorrhage is
grimacing, myoclonus, psychiatric se- and microcirculatory abnormalities. common in patients in
quelae (including depression and mood Interestingly, the stress response that renal failure with brain
and personality changes), and mutism. occurs in sepsis is thought to originate atrophy, and uremic
platelet dysfunction can
Most patients have undergone at least in a brain-signaling process that begins
predispose these
2.5 years of dialysis prior to onset, and in both the vagal nerve signaling path-
patients to hemorrhage.
the progression to death is relatively ways and the circumventricular organs.
rapid over 6 months. Dialysis using a The vagal nerve signaling pathways h Sepsis-associated
encephalopathy is
water supply with greater than 20 2g/L of detect systemic inflammation through
mediated by
aluminum has been linked to dialysis de- the afferent pathway and signal the nu-
inflammatory cells,
mentia.30 Dialysis dementia is now rare cleus tractus solitarius and paraven- cytokines, alteration in
because of changes in dialysis technology. tricular nuclei that control vasopressin neurotransmitters,
Cerebral atrophy is common in pa- secretion and the adrenal axis via the endothelial
tients with chronic renal failure and can efferent pathway. The circumventricular abnormalities, and
predispose patients to subdural hema- organs are located near neuroendocrine disruption of the
toma. Additionally, patients with renal centers and have an incomplete blood- blood-brain barrier.
failure may be more susceptible to intra- brain barrier allowing the direct detec-
cranial hemorrhage because of uremia- tion of circulating systemic blood mole-
induced platelet dysfunction. Any renal cules, particularly peptide hormones.35
failure patient with focal deficits should Following stimulation of the vagal path-
undergo brain imaging. MRI in renal fail- way or circumventricular organs, a va-
ure may be normal, may demonstrate riety of cytokines, chemokines, and
atrophy, or may show white matter prostaglandins are released, impacting
hypodensities. A creatinine clearance the cerebral milieu and neurotrans-
of 15 mL/min to 50 mL/min has been mission, particularly beta-adrenergic,
associated with increased white matter GABA-ergic, and serotoninergic release
hyperintensity volume.31 The Lentiform and receptor expression. Glutamate ex-
Fork sign, characterized by T2 hyper- citotoxicity also occurs in SAE. The
intensities in the putamen resembling a subsequent effect is encephalopathy.
fork, has been associated with uremic Mitochondrial dysfunction, oxidative
encephalopathy and metabolic acido- stress, and apoptosis seem to play a
sis.32 A rare syndrome of acute bilat- role in SAE-related brain injury, as do
eral basal ganglia lesions, presenting endothelial activation with subsequent
with parkinsonism and encephalopathy, alteration in vascular tone, microcircu-
can occur in patients with diabetes and latory dysfunction, and activation of the
renal failure.33 coagulation cascade.35 Microthrombo-
sis and infarction can occur as a result
Sepsis-Associated of endothelial disruption. Blood-brain
Encephalopathy barrier breakdown also can occur as a
Sepsis-associated encephalopathy (SAE) result of alteration of the endothelium,
occurs in up to 70% of patients with bac- leading to direct passage of neurotoxic
teremia and is one of the most common molecules into the brain. Regional loss
types of metabolic encephalopathy.34 of autoregulation may occur. Indeed,
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 623

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINTS
posterior reversible encephalopathy syn- Hyponatremia and
h Seizures, both
convulsive and drome, thought to be related to poor Hypernatremia
nonconvulsive, occur in autoregulation in the posterior circu- Hyponatremia is commonly caused by
up to 10% of patients lation, can occur in the context of sepsis. the syndrome of inappropriate secretion
with sepsis. Practitioners Markers of brain injury such as ele- of antidiuretic hormone (SIADH) (which
should have a low vated serum levels of neuron-specific can be seen in the context of brain
threshold for continuous enolase and S-100 protein have been injury, certain medications, and malig-
EEG monitoring. detected in patients with SAE.36 Neuro- nancy) and cerebral salt wasting (which
h Worsening EEG findings pathologic specimens have shown hem- is also common in patients with brain
in sepsis-associated orrhages related to DIC, multifocal injuries, particularly those with subar-
encephalopathy are necrotizing leukoencephalopathy, and achnoid hemorrhage). The mechanism
associated with microabscesses.37 MRI in SAE has re- for cerebral salt wasting is not fully
progressively worse vealed both ischemic lesions and white understood, but some propose that it is
prognosis. matter lesions of unknown etiology.38 due to impaired sympathetic neural in-
h The syndrome of It is important to be vigilant for the put; impaired renin and aldosterone
inappropriate secretion development of seizures (both convul- release; failure of sodium, uric acid, and
of antidiuretic hormone sive and nonconvulsive) in patients with water absorption at the proximal tubule;
and cerebral salt SAE. The high glutamate levels that or a brain natriuretic peptideYinduced
wasting result in similar
occur in patients with SAE may account decrease in sodium reabsorption and
laboratory values. The
for the lower seizure threshold. Seizures renin release. Cerebral salt wasting and
only reliable way to
distinguish the two is by
appear to be common in the context of SIADH can produce nearly identical
estimation of volume sepsis, occurring in up to 10% of crit- laboratory values, including hyponatre-
status. ically ill patients with sepsis.27 Patients mia, high urine osmolarity, low serum
who do not return to a normal cogni- osmolarity, urine sodium greater than
tive state within 30 minutes to 1 hour 40 mEq/L, and low serum uric acid. The
of a convulsive seizure should be con- major distinguishing feature between
sidered for continuous EEG monitoring these two entities is volume status. Pa-
because progression from convulsive tients with cerebral salt wasting have
to nonconvulsive seizures is common. greater urinary losses than intake and
Similarly, SAE patients with unex- are clinically hypovolemic, whereas pa-
plained alteration in mental status tients with SIADH are euvolemic. Other
should be considered for continuous causes of hyponatremia that should be
EEG monitoring to screen for non- evaluated include effective circulating
convulsive seizures. volume depletion due to heart failure,
The presence and severity of SAE has cirrhosis, volume loss, or diuresis, and
been correlated with mortality in pa- hormonal imbalances that occur with
tients with sepsis. The presence of coma hypothyroidism, adrenal insufficiency,
(Glasgow Coma Scale score less than 8) and pregnancy.
in a patient with SAE has been associated Encephalopathy due to hyponatremia
with a mortality rate of 63%.39 Mortality depends on the rate of development of
in patients with bacteremia can be pre- hyponatremia and the absolute sodium
dicted by the EEG. In one study, mor- level. Hyponatremia that develops rap-
tality rates were 0% for patients with no idly (over 12 to 24 hours) and serum
EEG abnormality, 19% for patients with sodium levels less than 120 mEq/L are
theta predominance, 36% for patients more likely to result in symptoms of con-
with delta predominance, 50% for fusion, seizures (generalized tonic clonic),
patients with triphasic waves, and 67% muscle cramps, generalized weakness,
for patients with EEG evidence of burst and, in severe cases, coma. Hyponatremia
suppression.40 that persists for longer than 24 to 48
624 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINTS
hours is associated with the generation ing Levels of Tolvaptan in Hyponatremia h The treatment of acute
of idiogenic osmoles that are toxic to (SALT) 1 and SALT 2 randomized trials and chronic hyponatremia
brain myelin. Rapid correction of chronic of 448 patients with SIADH, euvolemic differs. Patients with
hyponatremia can result in central pon- hyponatremia, cirrhosis, or heart failure. acute symptomatic
tine or extrapontine myelinolysis (os- Compared with placebo, tolvaptan led hyponatremia
motic demyelination) characterized by to a significant increase in serum so- (developing over less
headache, lethargy, coma, seizures, and dium after 4 days of treatment (135 than 24 hours) can
focal deficits. MRI in these cases demon- mEq/L versus 130 mEq/L) and after 30 undergo relatively rapid
strates white matter changes on T2 or days of treatment (136 mEq/L versus correction at 1.5 mEq/L
to 2 mEq/L per hour until
fluid-attenuated inversion recovery 131 mEq/L).41 Conivaptan was studied
symptoms resolve,
(FLAIR) sequences, typically in the pons in a randomized placebo-controlled trial
whereas those with
and other parts of the brainstem. of 84 patients with euvolemic or hyper- chronic hyponatremia
Treatment for symptomatic patients volemic hyponatremia with a baseline (developing over 24 to
with acute hyponatremia (developing sodium of 115 mEq/L to 130 mEq/L. 48 hours) should be
over less than 24 hours) consists of ra- Compared with placebo, conivaptan sig- corrected no faster than
pid correction of sodium at a rate of 1.5 nificantly raised sodium by 6.3 mEq/L 8 mEq/L to 10 mEq/L per
mEq/L to 2 mEq/L per hour until symp- using a 40 mg/d dose and 9.4 mEq/L day to avoid osmotic
toms resolve (usually to a sodium level using a 80 mg/d dose after 4 days of demyelination.
of about 120 mEq/L) (Case 7-2). Pa- treatment.42 h In cerebral salt wasting,
tients with chronic hyponatremia (de- The treatment of cerebral salt wast- it is important to replace
veloping over 24 to 48 hours) are at the ing varies significantly from that of both salt and water
greatest risk for osmotic demyelination SIADH. In cerebral salt wasting, it is because patients are
if the sodium is corrected more than important to replace both salt and water both salt and volume
20 mEq/L per day or if the baseline because patients are both salt and depleted.
sodium is less than 105 mEq/L. Patients volume depleted. Options include sal-
with chronic hyponatremia should be ine infusion (the hypertonicity depends
corrected with no more than 8 mEq/L on the baseline sodium), oral replace-
to 10 mEq/L of sodium per day or 18 ment of salt and water using salt tabs or
mEq/L of sodium in 48 hours. Options hypertonic beverages, and fludrocorti-
for the treatment of SIADH include free sone. It is advisable to replace orally
water restriction, hypertonic saline, in- consumed free water (eg, coffee, tea,
creased oral sodium intake in the form juice, and water) with isotonic or hy-
of salt tablets or high sodium food, and pertonic fluids such as tomato juice,
V2 (vasopressin) receptor antagonists, sports drinks, or soup. Fluid restriction
which produce a pure aquapheresis. of patients with cerebral salt wasting,
Currently, the US Food and Drug Ad- subarachnoid hemorrhage, and vaso-
ministration has approved tolvaptan spasm can lead to devastating infarcts
(an oral selective V2 receptor antago- and should be avoided.43 In all cases
nist) for euvolemic hyponatremia, of hyponatremia, the underlying eti-
SIADH, and hyponatremia associated ology should be addressed. Patients
with heart failure and cirrhosis, and who are actively seizing may receive
conivaptan (an IV mixed V2 and V1a antiepileptics, although the best treat-
receptor antagonist) for euvolemic ment for hyponatremia-related seizures
hyponatremia and SIADH. Some have is sodium correction.
expressed concerns that the V1a antag- Hypernatremia can be due to water
onism of conivaptan may cause hypo- losses, diabetes insipidus, or infusions
tension, increased risk of variceal of saline or hypertonic saline. Slowly
bleeding, or renal dysfunction. Tolvap- developing hypernatremia is not likely
tan was studied in the Study of Ascend- to produce symptoms because brain

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 625

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINT
h Dural sinus thrombosis
is the most feared
Case 7-2
A 35-year-old woman was competing in her first long-distance bicycle race.
complication of
Her physician instructed her to ingest as much water as possible, even if
hypernatremia, but
she was not thirsty. Over the course of 2 hours, she drank 6 liters of water.
it occurs in the
Her friends noticed she became disoriented, and she had a generalized
context of volume
tonic-clonic seizure in the ambulance on the way to the hospital. On
depletionYinduced
examination, her pupils were sluggish, and she had bilateral extensor
hypernatremia rather
posturing. Her sodium level on admission to the hospital was 103 mEq/L.
than hypernatremia due
Comment. Given the patient’s poor mental status, she should be
to saline or hypertonic
intubated immediately. A head CT should be performed to assess for
saline infusion.
cerebral edema, and intracranial pressure monitoring should be
considered. Because her hyponatremia is acute, she can be corrected
rapidly with hypertonic saline at a rate of 1.5 mEq/L to 2 mEq/L per hour
until her sodium is around 120 mEq/L, after which slower correction can
occur. A reasonable choice for correction might be 3% saline beginning
at a rate of 30 mL/h (this must be administered through a central line).
Alternately, 2% saline at a higher infusion rate (such as 50 mL to 75 mL/h)
could be used. Sodium values should be checked frequently (every 4 to
6 hours) to ensure an appropriate rate of correction. Continuous EEG
monitoring should be considered because nonconvulsive seizures are a
risk. A short-term course of antiepileptics can be considered, although the
ultimate treatment for her seizures is sodium correction.

cells can maintain their volume through resultant cerebral edema that may be as
the generation of idiogenic osmoles. responsible for poor outcome as the
Patients with serum osmolality greater absolute sodium value. Additionally, hy-
than 350 mOsm/kg or with serum so- pernatremia may be a marker for ag-
dium levels greater than 160 mEq/L may gressive osmotherapy and may reflect
develop lethargy, confusion, decreased poor ICP control. Finally, hypernatre-
mental status, and, occasionally, sei- mia in the context of mannitol use typi-
zures. Dural sinus thrombosis is the cally occurs when renal insufficiency is
most feared complication of hyperna- present, suggesting that hypernatremia
tremia, but it occurs in the context of may be a marker for organ failure rather
volume depletionYinduced hypernatre- than the direct causal factor for poor out-
mia rather than hypernatremia due to come. There has been one case report of
saline or hypertonic saline infusion. In a fatal hypernatremia in the context of
retrospective study of 339 patients with accidental ingestion with a sodium level
serum sodium levels greater than of 209 mEq/L.45 Correction of hyperna-
150 mEq/L (excluding patients classified tremia should take into account the
as dead by cessation of brain function), underlying etiology and whether hyper-
the most common causes of hyper- natremic osmotic therapy is being used
natremia were mannitol therapy and to treat cerebral edema or elevated ICP.
diabetes insipidus. In a multivariate In the context of hypertonic saline os-
analysis, the authors found that sodium motic therapy, a typical sodium target
levels greater than 160 mEq/L were is 150 mEq/L to 155 mEq/L, although
an independent predictor of mortality some patients may require higher so-
whereas lower quartiles of sodium were dium values to control edema and ICP.
not.44 However, the authors did not When weaning patients off hypertonic
adjust for ICP and shifts in sodium and saline, care should be taken to avoid
626 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINTS
rebound cerebral edema. In the case of with alcohol-related deaths.48,49 Wernicke h The classic triad of
hypernatremia due to volume depletion, encephalopathy is most common in the symptoms for Wernicke
fluid correction with oral free water or context of chronic alcoholism, but poor encephalopathy
hypotonic saline can be used to correct nutrition, malabsorption, renal losses (ataxia, confusion,
at a rate of 1 mEq/L to 2 mEq/L per hour. of thiamine (in patients on dialysis), and oculomotor
and increased metabolic thiamine re- abnormalities) is the
Wernicke Encephalopathy quirement in systemic illness can all be exception rather than the
Wernicke encephalopathy is a compli- causes of thiamine deficiency. Other rule. Encephalopathy
cation of thiamine (vitamin B1) defi- conditions associated with Wernicke may be present in
ciency characterized by a classic triad of encephalopathy include bariatric or isolation. A high index of
clinical suspicion must
encephalopathy, gait ataxia, and ocular gastrointestinal surgery, systemic malig-
be maintained and
motor dysfunction, including ophthal- nancy, transplant, dialysis, AIDS, an-
treatment initiated for
moparesis and nystagmus. Often, not orexia, and hyperemesis of pregnancy. any of these symptoms
all elements of the triad are found Thiamine is a cofactor for enzymes in the proper clinical
together. In one series, only 17% of pa- involved in glycolysis and the citric acid context.
tients with MRI-proven Wernicke ence- cycle, namely pyruvate dehydrogenase,
h Wernicke encephalopathy
phalopathy presented with all three transketolase, and !-ketoglutarate dehy- is most common in those
findings.46 Confusion is the most com- drogenase. Glucose boluses in suscep- with chronic alcoholism
mon symptom, followed by ataxia and tible patients with marginal thiamine but can also occur in
ocular abnormalities. Because the clas- stores may lead to acute thiamine de- anorexia, bariatric
sic triad is the exception rather than pletion and precipitation of Wernicke surgery, dialysis,
the rule, a high index of suspicion must encephalopathy. Thus, it is critical to hyperemesis gravidarum,
be maintained, particularly in confused provide thiamine supplementation and other diseases that
patients with a history of alcoholism prior to a glucose load in at-risk patient lead to malnutrition.
who might otherwise be thought to populations. Neuropathologic speci- h Thiamine is a cofactor
have alcohol intoxication or withdrawal. mens reveal symmetric neuronal loss, for enzymes that are
Horizontal gaze-evoked nystagmus is the petechial hemorrhage, demyelination, integral to glycolysis and
most common ocular finding. Lateral rec- gliosis, and atrophy of the mammillary the citric acid cycle. In
tus palsy, when present, is almost always bodies and structures surrounding the patients with low
bilateral. Pupillary abnormalities may be third ventricle, aqueduct, and fourth ven- thiamine stores, a
glucose challenge can
present, and patients with advanced tricle, including the dorsomedial thal-
rapidly deplete thiamine
Wernicke encephalopathy may have amus, periaqueductal gray, ocular motor
and precipitate
complete ophthalmoplegia and fixed, nuclei, locus ceruleus, and vestibular Wernicke encephalopathy.
miotic pupils. Gait ataxia involves a nuclei. MRI findings include restricted Thiamine should always
combination of vestibular and cerebellar diffusion-weighted imaging and increased be administered prior to,
dysfunction and peripheral neuropathy. T2 and decreased T1 signal in these areas. or concurrent with,
Because cerebellar pathology is limited T2 signal can be reversed with thiamine glucose in at-risk
to the vermis, appendicular ataxia is treatment, and MRI improvement usually patients.
uncommon. The Caine criteria47 pro- correlates with symptomatic improve-
pose that Wernicke encephalopathy can ment. Patients with chronic Wernicke
be diagnosed with 85% sensitivity when encephalopathy typically demonstrate
two of the following are present: dietary mammillary body atrophy.
deficiency, oculomotor abnormalities, Although thiamine deficiency can be
cerebellar dysfunction, and either altered reliably detected by measurement of
mental status or memory impairment. erythrocyte thiamine transketolase be-
On the basis of autopsy series, fore and after the addition of thiamine
Wernicke encephalopathy occurs in pyrophosphate, treatment should typi-
12.5% of those who abuse alcohol and cally precede diagnostic measures given
in up to 59% of those with alcoholism the high morbidity of delayed treatment.
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 627

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINT
h Immediate thiamine Indeed, response to treatment can pro- cephalopathy can progress to coma and
replacement takes vide a diagnosis. Immediate IV thiamine death. Residual deficits, even with treat-
precedence over replacement takes precedence over lab- ment, are common. Progression to
laboratory or imaging oratory or imaging diagnosis. The rec- Korsakoff amnestic syndrome can occur.
diagnosis. ommended regimen is 500 mg of IV
thiamine over 30 minutes 3 times a day DIAGNOSIS AND EVALUATION
for 2 days followed by 500 mg IV or IM Diagnostic tools have been developed
each day for 5 days along with other to assist in the recognition of encephal-
B vitamins.50 Initial oral repletion of opathy and delirium. The Confusion
thiamine is considered inadequate. Assessment Method for the ICU (CAM-
However, after parenteral replacement ICU) scale51 was specifically developed
patients should be continued on 100 mg to identify delirium or fluctuating levels
a day of thiamine as outpatients. Prompt of attentiveness in the ICU as defined
administration of thiamine typically by Diagnostic and Statistical Manual
leads to symptom improvement within of Mental Disorders (DSM-IV) guide-
hours to days. Untreated Wernicke en- lines (Table 7-5).52 The CAM-ICU scale

a,b
TABLE 7-5 Confusion Assessment Method for the Intensive Care Unit

b Feature 1: Acute Onset and Fluctuating Course

Identify an acute change in mental status from the baseline examination.
OR
Identify fluctuating changes in mental status or behavior over the past
24 hours that may vary in severity.
AND
b Feature 2: Inattention
Identify an inability to focus attention, easy distractibility, or inability to process
components of conversation (eg, count backward, say months backward).
AND
b Feature 3: Disorganized Thinking
If the patient is verbal (and not aphasic): Identify illogical or incoherent
thought processes, inability to understand proverbs, or inability to perform
simple calculations (eg, How many things are in a dozen? Where does a
cactus grow?).
If the patient is intubated or nonverbal (and not aphasic): Use yes/no
questions or a letter board to identify illogical or incoherent thought
processes (eg, Can a cat sing? Does wool come from an alligator?).
OR
b Feature 4: Altered Level of Consciousness
Identify if the patient’s level of consciousness is anything other than alert
(ie, drowsy, lethargic, stuporous, comatose, or agitated/combative).
a
Data from Ely EW, et al, JAMA.51 jama.ama-assn.org/content/286/21/2703.long.
b
Reprinted with permission from Frontera JA. Delirium and sedation in the ICU. Neurocrit Care 2011;
14(3):463Y474.52 www.springerlink.com/content/513556774740p05j/.

628 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

was studied in a group of 96 mechan- define delirium as a fluctuating level of
ically ventilated ICU patients with a attentiveness with a reduced ability to
Richmond Agitation Sedation Scale focus, sustain, or shift attention.
(RASS) score of j3 to +4 (Table 7-6)53 The CAM-ICU scale has limitations,
and compared to the criterion standard however. Most notably, it allows for the
DSM-IV definition of delirium. Using the determination of delirium to be made in
CAM-ICU scale, delirium was diagnosed the context of sedative use. Because the
in 83% of ICU patients for a mean of CAM-ICU scale is meant to be used
2.4 days. The authors found 93% to repeatedly, any variations in the amount
100% sensitivity and 98% to 100% spe- of sedation administered at the time of
cificity for the CAM-ICU scale when com- the examination will lead to fluctuating
pared to the DSM-IV guidelines, which CAM-ICU scores. Additionally, because

a
TABLE 7-6 Richmond Agitation Sedation Scale

Score Rating Description

+4 Combative Overtly combative or violent; immediate danger to staff
+3 Very agitated Pulls on or removes tubes or catheters or has aggressive behavior toward staff
+2 Agitated Frequent nonpurposeful movement or patient-ventilator dyssynchrony
+1 Restless Anxious or apprehensive but movements not aggressive or vigorous
0 Alert and calm
-1 Drowsy Not fully alert, but has sustained (more than 10 seconds) awakening with eye
contact to voice
-2 Light sedation Briefly (less than 10 seconds) awakens with eye contact to voice
-3 Moderate Any movement (but not eye contact) to voice
sedation
-4 Deep sedation No response to voice, but any movement to physical stimulation
-5 Unarousable No response to voice or physical stimulation
Procedure
(1) Observe patient. Is patient alert and calm (score 0)?
(2) If patient is not alert, in a loud speaking voice state patient’s name and direct patient to open eyes and look
at speaker. Repeat once if necessary. Can prompt patient to continue looking at speaker.
Patient has eye opening and eye contact, which is sustained for more than 10 seconds (score -1).
Patient has eye opening and eye contact, but this is not sustained for 10 seconds (score -2).
Patient has any movement in response to voice, excluding eye contact (score -3).
(3) If patient does not respond to voice, physically stimulate patient by shaking shoulder and then rubbing
sternum if there is no response to shaking shoulder.
Patient has any movement to physical stimulation (score -4).
Patient has no response to voice or physical stimulation (score -5).
a
Reprinted from Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients.
Am J Respir Crit Care Med 2002;166(10):1338Y1344.53 B 2002, with permission from American Thoracic Society. ajrccm.atsjournals.org/content/166/10/
1338.long.

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 629

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINT
h Although the CAM-ICU sedation medication causes encephalop- condition. Although one of the benefits
remains a useful athy or delirium in up to 30% of cases, of the CAM-ICU is that it is a stan-
research and clinical tool using the CAM-ICU in sedated patients dardized examination that can be quickly
and can be used in will limit the ability to identify other and reproducibly performed by nurses,
conjunction with the underlying etiologies of delirium.54 More physicians, and other health care staff,51
neurologic examination, concerning, serious neurologic events medical staff can similarly become pro-
the additional may occur and go undetected and un- ficient at an abbreviated neurologic
components of the treated when examinations are con- examination (Table 7-7).52 Although
neurologic examination ducted in patients receiving sedation. the CAM-ICU remains a useful research
are necessary in order to Serial neurologic examinations per- and clinical tool and can be used in
identify the presence of
formed off sedation with an adequate conjunction with the neurologic exami-
encephalopathy or
sedation washout period are the best nation, the additional components of
delirium and narrow
down the etiology.
way to identify both encephalopathy the neurologic examination are neces-
and focal neurologic deficits, which sary in order to identify the presence of
may hint at an underlying neurologic encephalopathy or delirium and narrow

TABLE 7-7 Neurologic Assessment ofa,bDelirium and Encephalopathy

in the Intensive Care Unit

Step 1: Assess for Fluctuations in Mental Status

Mental status assessmentc
Arousal Spontaneously awake, eyes open and alert
Opens eyes to voice
Opens eyes to physical stimuli
No eye opening
Attentiveness Able to say months backward from December to January
Able to count backward from 20 to 1 but cannot say
months backward
Able to count from 1 to 10 but cannot do either of
the above
Able to follow complex 2- to 3-step verbal commands but
cannot do any of the above
Able to follow only simple verbal commands
Able to follow only mimicked commands
Cannot follow commands but can track visual stimuli
Cannot track visual stimuli but saccades to voice
Does not saccade to voice but saccades to physical stimuli
No response to examiner
Orientation Oriented to person, place, and time
Oriented to any two of the above
Oriented to self only
Continued on next page

630 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

TABLE 7-7 Continued

Step 2: Identify Localizing Abnormalities In The Neurological Examinationd

Cranial nerve (CN) examination
CN II/III Pupil symmetry and reactivity to light
Visual field assessment, blink to threat (II, VII, visual
pathways)
CN III/IV/VI/VIII Extraocular movements
If patient cannot track, test oculocephalic reflex (doll’s
eyes maneuver)
Vestibulo-ocular reflex, hearing
CN V/VII Corneal response, facial sensation, facial movement
CN IX/X/XII Gag, tongue movement
c
Motor examination
Strength Test all four limbs:
No drift holding arm or leg out for 10 seconds
Drift; limb moves before 10 seconds but does not hit
bed or support
Able to move limb against gravity
Unable to move limb against gravity but some
movement if force of gravity eliminated
No movement to command
Response to If patient cannot comply with above motor examination,
physical stimuli test response to physical stimulie:
Localizes to physical stimuli
Withdraws from physical stimuli
Flexor posturing (decorticate)
Extensor posturing (decerebrate)
Plegia (no movement to physical stimuli) or triple
flexion (stereotyped flexion at hip, knee, and ankle)
a
Reprinted with permission from Frontera JA. Delirium and sedation in the ICU. Neurocrit Care 2011;14(3):
463Y474.52 www.springerlink.com/content/513556774740p05j/.
b
Delirium is present when there is fluctuation in level of arousal, attentiveness, or orientation. The cranial nerve
and motor examination help to identify and localize an underlying neurologic injury that may be contributing to
delirium.
c
Within each category responses are listed from best to worst.
d
Examination should be performed in the patient’s native language after sedation has been discontinued for a
reasonable amount of time.
e
Patient can have mixed responses.

down the etiology. Additionally, a thor- 127 ICU patients in whom a neurologic
ough neurologic examination can rule consultation was conducted for an iso-
out serious structural causes of ence- lated change in mental status, the neu-
phalopathy. In a retrospective study of rologic examination had a 97% negative

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 631

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINT
h Common EEG findings predictive value for detecting acute and blood and sputum cultures should
in the context of neurologic injury and identified is- be performed in the context of fe-
metabolic encephalopathy chemic stroke in 7% of patients and ver or hypothermia. Thyroid function
include triphasic waves subarachnoid hemorrhage in 1% of pa- tests, serum cortisol, thiamine, and
and frontally predominant tients.55 Furthermore, neurologic ill- vitamin B12 levels can be checked if
rhythmic delta activity. In nesses are not uncommon causes of endocrine or nutritional abnormalities
extreme cases of encephalopathy and warrant thorough are suspected.
metabolic encephalopathy evaluation. For example, the rates of Head CT or MRI and vascular imag-
a burst-suppression nonconvulsive status epilepticus are as ing, such as CT or MR angiography, can
pattern can be seen. high as 35% in a neuro-ICU setting,25,56 be considered on the basis of the exam-
and 8% to 10% of medical ICU patients ination and laboratory results. Patients
with no underlying neurologic diagnosis with focal deficits should undergo some
have seizures, most of which are non- form of imaging. In patients with sus-
convulsive.26,27 Seizures in the medical pected flow failure or vasculopathy,
ICU population are particularly com- perfusion imaging (CT or MR) may be
mon in patients with sepsis.27 Similarly, useful. The yield of neuroimaging de-
stroke, which may be associated with pends on the underlying admitting diag-
encephalopathy or delirium, is not un- nosis. In one study, new CT findings
common among ICU patients with were present in 26 of 123 (21%) medical
primary medical diagnoses.57 ICU patients with the diagnosis of
The following is a rational approach ‘‘altered mental status,’’ including ische-
to evaluating metabolic encephalopathy mic infarction in 13 (11%), intracerebral
in the ICU in both general critical care hemorrhage in 2 (2%), and tumor in
and neurocritical care patients. First, a 3 (2%).58
complete history including a thorough Continuous EEG to evaluate for sei-
medical history and details of possible zures or nonconvulsive seizures should
toxin exposure should be performed. be considered in all patients with unex-
Next, sedating and toxic medications plained encephalopathy. In command-
should be discontinued, if possible. The following patients, 95% of seizures will
neurologic examination can be con- be captured with 24 hours of monitor-
ducted after an appropriate sedative ing, but among comatose patients
washout period to identify any focal only 80% of seizures are detected after
features that might offer clues to local- 24 hours of monitoring.25 Therefore,
ization or etiology. Basic serum labora- 48 hours of continuous EEG may be
tory tests to identify metabolic disarray required in comatose patients. Seizures
(eg, hypoxia, hypercarbia, hypoglycemia, appear to be common in the context of
uremia, hyperammonemia, and endo- sepsis, occurring in up to 10% of crit-
crine dysfunction) should be performed. ically ill patients with sepsis.27 Common
These studies should include a basic EEG findings in the context of metabolic
chemistry panel (sodium, potassium, encephalopathy include triphasic waves
chloride, bicarbonate, blood urea nitro- and frontally predominant rhythmic
gen, creatinine, and glucose), magne- delta activity (Figure 7-3). In extreme
sium, ionized calcium, phosphate, liver cases of metabolic encephalopathy a
function tests, arterial blood gas, CO- burst-suppression pattern can be seen
oximetry (if carbon monoxide exposure (Figure 7-4). Although these patterns
or methemoglobinemia is suspected), suggest metabolic encephalopathy, they
cardiac enzymes, complete blood count, are not specific for any given etiology.
and coagulation studies. A urine toxicol- EEG can also be useful in determining
ogy screen can be sent in select cases, whether myoclonus represents a seizure
632 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINTS
h Multimodal monitoring,
which is becoming more
common in neurocritical
care patients, is
particularly useful in
comatose patients and
can supplement the
neurologic examination
and provide insights into
the etiology of fluctuating
mental status.
h It has been increasingly
recognized that
critically ill patients with
EEG demonstrating burst-suppression pattern. metabolic encephalopathy
FIGURE 7-4
are often left with
long-term neurocognitive
deficits.
or has a cortical basis. In general, full- apy, blood pressure, or ventilator man-
montage continuous EEG monitoring is agement) or may support the need for
preferable to a 30-minute EEG because neurosurgical intervention.
spot EEG monitoring has a sensitivity The differential diagnosis for
of only 21% for convulsive seizure and metabolic encephalopathy is listed
only 11% for nonconvulsive seizure. in Table 7-8.
Similarly, hairline EEG or bispectral in-
dex monitoring has inadequate sensitiv- TREATMENT
ity and specificity for seizure detection.59 Current treatment paradigms typically
Other diagnostic evaluations, such as focus on supportive care and treatment
lumbar puncture, can be considered in of the underlying etiology. Directed
patients with fever and meningismus or therapies that target neurochemical and
in patients predisposed to CNS infection neurotransmitter pathways that mediate
(such as those who have undergone a encephalopathy are not currently avail-
neurosurgical procedure). It should be able and represent an important area
noted, however, that new-onset menin- for future research. Etiology-specific
gitis is extremely uncommon in hospi- treatments are detailed above. Myoclo-
talized patients.60 nus can be treated symptomatically with
Multimodal monitoring, which is levetiracetam or valproic acid.
becoming more common in neurocrit-
ical care patients, is particularly useful in PROGNOSIS
comatose patients and can supplement Classically, the brain dysfunction that
the neurologic examination and provide occurs with metabolic encephalopathy
insights into the etiology of fluctuating was thought to be completely reversi-
mental status. Changes in cerebral he- ble. It has been increasingly recognized,
modynamics, perfusion, and neuro- however, that critically ill patients with
chemistry can be detected using brain metabolic encephalopathy are often left
tissue oxygen monitoring, jugular bulb with long-term neurocognitive defi-
oxygen monitoring, ICP monitoring, cits.61 Persistent neurologic and psy-
intracranial blood flow monitoring, and chiatric deficits occur in up to 32% of
microdialysis. These devices may assist medical ICU survivors, including psy-
in medical management (osmotic ther- chomotor slowing and reduced verbal

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 633

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

TABLE 7-8 Differential Diagnosis for Metabolic Encephalopathy

Differential Diagnosis Examples Characteristic Features

Structural brain lesion Ischemic stroke Focal deficits (often fixed)
Intracerebral hemorrhage Cranial nerve involvement
Traumatic brain injury Imaging findings
Subdural hematoma
Subarachnoid hemorrhage
Brain tumor
Cerebral sinus thrombosis
Anoxic brain injury Cardiac arrest Clinical context
Near-drowning Cranial nerve involvement
Airway obstruction
Infection Meningitis Meningismus
Encephalitis Nuchal rigidity
Abscess Photophobia
Focal examination findings
CSF findings
Fever
Elevated white blood cell count
Seizures Focal seizures Rhythmic myoclonus, often not multifocal
Complex partial seizures EEG findings
Generalized seizures
Nonconvulsive seizures
Degenerative diseases Creutzfeldt-Jakob disease Subacute progression
Fatal familial insomnia Myoclonus
EEG findings
MRI findings
CSF 14-3-3 protein
Genetic testing
Psychiatric symptoms
Insomnia/sleep disturbances
Trauma Shaken baby syndrome Retinal hemorrhages
Diffuse axonal injury Imaging findings
Social history
Clinical history

634 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 KEY POINT
fluency, visual and working memory, could receive when evaluated. Coma h An abundance of
and visuoconstruction skills.62 Specific was defined as response to physical evidence links delirium
etiologies of metabolic encephalopathy stimuli but no eye opening, or no re- with poor outcomes in
are associated with distinct long-term sponse to physical stimuli (RASS score ICU survivors, including
sequelae. Patients with Wernicke en- -4 or -5). Patients who were classified increased morbidity and
cephalopathy often go on to develop as ‘‘comatose’’ or ‘‘delirious’’ had in- mortality, prolonged
Korsakoff syndrome, characterized by creased mortality rates and longer length of stay, and
anterograde and retrograde amnesia lengths of stay compared to those with increased hospital
with confabulation. Additionally, per- just one of the diagnoses. Overall, complications such as
manent nystagmus and ataxia are com- 18.5% of patients in this study remained fever, tachycardia, and
hypertension.
mon after Wernicke encephalopathy. in a persistent ‘‘coma’’ and died, al-
Similarly, survivors of hepatic encephal- though neither the cause of ‘‘coma’’
opathy may have permanent cognitive nor death was reported. Importantly,
or motor deficits, and patients with the use of lorazepam and the cumula-
hypoglycemic encephalopathy may incur tive lorazepam dose were directly asso-
cortical injury, mesial temporal injury, ciated with the presence of delirium,
and long-term memory impairment.63,64 but this was not the case for other types
While the prognosis of metabolic en- of sedatives, such as propofol, fentanyl,
cephalopathy largely depends on the and morphine. On the basis of these
underlying systemic illness, the impact data, it is possible that delirium was a
of delirium (often due to metabolic marker for lorazepam use. Indeed,
disarray or multiorgan dysfunction) on lorazepam use itself may have been
outcome has been well studied. responsible for increased delirium, in-
An abundance of evidence links creased mortality, and prolonged length
delirium with poor outcomes in ICU of stay.
survivors, including increased morbidity In a substudy of the Safety and Ef-
and mortality, prolonged length of stay, ficacy of Dexmedetomidine Compared
and increased hospital complications With Midazolam (SEDCOM) trial66 (a
such as fever, tachycardia, and hy- trial that randomized 375 medical and
pertension.65 In a prospective study of surgical ICU patients to receive either
275 mechanically ventilated patients in dexmedetomidine or midazolam), mor-
a medical and cardiac ICU setting, tality was significantly higher in those
delirium occurred in 82% of patients with delirium compared to those with-
and was significantly associated with out (30.3% versus 11.9%, PG.001). The
6-month mortality (hazard ratio 3.2; duration of mechanical ventilation and
95% CI, 1.4-7.7; P=.008).3 The risk length of stay were shorter in those
of death increased 10% for each ICU without delirium (PG0.001).67 There
day spent in delirium (hazard ratio 1.1; was a dose-response effect for the
95% CI, 1.0-1.3, P=.03). Additionally, duration of time spent in delirium and
delirium was associated with prolonged the risk of prolonged ventilation time,
length of hospital stay (hazard ratio 2.0; increased length of stay, and mortality.
95% CI, 1.4-3.0; PG.001) and length of Interestingly, there was no difference
post-ICU stay (hazard ratio 1.6; 95% CI, in the 30-day mortality rates between
1.1-2.3; P=.009). In this study, delirium the dexmedetomidine and midazolam
was diagnosed in patients who were groups in the larger SEDCOM66 study
CAM-ICU51 positive (Table 7-5) with a (22.6% versus 25.4%, P=.60), implying
RASS score53 of -3 to +4 (Table 7-6). that the association of delirium and mor-
There was no stipulation as to the tality cannot be explained by a sedation
amount or type of sedation a patient effect alone. Indeed, after adjusting for
Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 635

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

KEY POINT
h Aside from an age, Acute Physiology and Chronic arrest or neurologic deficits that pre-
association with Health Evaluation II (APACHE II) score, vented them from living independently
mortality and hospital sepsis, shock, pneumonia, type of ICU, (eg, large stroke, severe dementia)
complications, delirium and randomization assignment, the du- were excluded from this study. Also,
has been found to have ration of time spent delirious predicted the authors did not account for any
long-term sequelae, mortality, mechanical ventilation time, new neurologic injury that may have
particularly cognitive and ICU length of stay. occurred during the ICU stay and po-
dysfunction and Aside from an association with mor- tentially contributed to delirium. In
increased risk of tality and hospital complications, delir- fact, many patients in this study were
institutionalization. ium has been found to have long-term at high risk for adverse neurologic
sequelae, particularly cognitive dysfunc- complications of their primary illness.
tion and increased risk of institution- Half of the patients in this study had
alization. A substudy of the Awakening severe sepsis, which is a risk factor
and Breathing Controlled (ABC) trial68 for seizures and status epilepticus.27
prospectively followed a cohort of 99 Another 20% had an admitting diagno-
mechanically ventilated ICU patients sis of myocardial infarction (MI) or con-
surviving longer than 3 months and gestive heart failure, which increases
assessed cognitive outcomes at 3 and the risk of ischemic stroke fivefold
12 months. A neuropsychologist who during the first month after diagnosis.57
was unaware of patient status adminis- Similarly, in the first 42 months after
tered a battery of neuropsychological MI, the risk of stroke is 4.6%70 and the
tests measuring attention and concen- rate of seizures is nearly doubled com-
tration, information processing speed, pared to age- and sex- matched con-
verbal memory, visual-spatial construc- trols.71 Some data suggest that delirium
tion and delayed visual memory, exec- is associated with brain abnormalities
utive function, language, and global seen on MRI, and this may partially
mental status. The authors found im- explain findings of long-term cognitive
paired cognition in 79% of patients at sequelae in patients with encephalop-
3 months and in 71% at 12 months. A athy. In one study, MRI FLAIR, gradient
higher number of days of delirium was echo, and T2 sequences were performed
associated with worse cognitive scores in a group of eight delirious medical
at 3 months (P=.02) and 12 months ICU patients without focal neurologic
(P=.03) after adjusting for age, educa- deficits, meningitis, anoxic brain injury,
tion, baseline cognitive status, APACHE previously identified brain lesions, base-
II scores, severe sepsis, and exposure line cognitive deficits, coma, elevated
to sedative medications in the ICU. intracranial pressure, or metastatic ma-
Relatively small increases in the dura- lignancy.72 White matter hyperintensi-
tion of delirium (from 1 day to 5 days) ties (75%) and atrophy (12%) were the
were associated with a decline in cog- most common MRI abnormalities. Six of
nitive scores by half an SD at 3 months 8 patients had severe abnormalities on
and with an even greater decline at 12 3-month neuropsychiatric testing, al-
months.69 This study is unique because though no patient had MRI findings
the association of the duration of time that altered treatment. The authors did
spent delirious and worse cognitive not report the time from diagnosis of
outcomes was preserved even after ad- delirium to MRI imaging, no premorbid
justing for causes of delirium such as MRIs were available for comparison,
sepsis, exposure to sedative medica- and diffusion-weighted imaging sequen-
tion, and severity of illness. Patients ces were not performed on any of the
with an admitting diagnosis of cardiac patients. Despite these limitations, this
636 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

study suggests that atrophy and white with ammonia-induced brain edema.
J Hepatol 2001;34(4):548Y554.
matter changes on MRI may be related to
poor cognitive outcomes in delirious and 7. Strauss G, Hansen BA, Kirkegaard P, et al.
Liver function, cerebral blood flow
encephalopathic patients. The pathology autoregulation, and hepatic encephalopathy
associated with long-term sequelae of in fulminant hepatic failure. Hepatology
metabolic encephalopathy remains to 1997;25(4):837Y839.
be elucidated and merits further study. 8. Jalan R, Olde Damink SW, Hayes PC, et al.
Some limitations of the delirium lit- Pathogenesis of intracranial hypertension in
acute liver failure: inflammation, ammonia
erature should be noted. The aforemen- and cerebral blood flow. J Hepatol 2004;
tioned studies all excluded patients with 41(4):613Y620.
primary neurologic diagnoses. Hence, 9. Ostapowicz G, Fontana RJ, Schiodt FV, et al.
this literature is not directly generalizable Results of a prospective study of acute liver
to the neurocritical care population. Ad- failure at 17 tertiary care centers in the
United States. Ann Intern Med 2002;137(12):
ditionally, none of the studies parse out 947Y954.
the specific etiology of delirium, account 10. Raschke RA, Curry SC, Rempe S, et al. Results
for potential neurologic ailments that of a protocol for the management of
could lead to delirium, or directly exam- patients with fulminant liver failure. Crit
Care Med 2008;36(8):2244Y2248.
ine the impact of different types of me-
tabolic encephalopathy on outcome. 11. Munoz SJ, Robinson M, Northrup B, et al.
Elevated intracranial pressure and computed
tomography of the brain in fulminant
CONCLUSIONS hepatocellular failure. Hepatology 1991;
Metabolic encephalopathy is common 13(2):209Y212.
in the ICU setting, and current manage- 12. Stravitz RT, Kramer AH, Davern T, et al.
ment focuses on addressing the un- Intensive care of patients with acute liver
failure: recommendations of the U.S. Acute
derlying medical problem. Given the Liver Failure Study Group. Crit Care Med
impact of encephalopathy and delirium 2007;35(11):2498Y2508.
on functional and cognitive outcome, 13. Bacher A. Intracranial hypertension in
future research is needed to identify fulminant hepatic failure. Transplant Proc
more directed treatment strategies to 2006;38(3):783Y785.
limit neurologic sequelae. 14. Blei AT, Olafsson S, Webster S, et al.
Complications of intracranial pressure
monitoring in fulminant hepatic failure.
REFERENCES Lancet 1993;341(8838):157Y158.
1. Kiely DK, Bergmann MA, Murphy KM, et al. 15. Gray WP, Palmer JD, Gill J, et al. A clinical
Delirium among newly admitted postacute study of parenchymal and subdural miniature
facility patients: prevalence, symptoms,
strain-gauge transducers for monitoring
and severity. J Gerontol A Biol Sci Med Sci intracranial pressure. Neurosurgery
2003;58(5):M441YM445. 1996;39(5):927Y931; discussion 931Y932.
2. Francis J. Delirium in older patients. J Am
16. Figaji AA, Zwane E, Fieggen AG, et al.
Geriatr Soc 1992;40(8):829Y838. Transcranial Doppler pulsatility index is not
3. Ely EW, Shintani A, Truman B, et al. Delirium a reliable indicator of intracranial pressure
as a predictor of mortality in mechanically in children with severe traumatic brain
ventilated patients in the intensive care unit. injury. Surg Neurol 2009;72(4):389Y394.
JAMA 2004;291(14):1753Y1762. 17. Mayer SA, Brun NC, Begtrup K, et al. Efficacy
4. Frontera JA, Kalb T. Neurological and safety of recombinant activated factor
management of fulminant hepatic failure. VII for acute intracerebral hemorrhage.
Neurocrit Care 2011;14(2):318Y327. N Engl J Med 2008;358(20):2127Y2137.

5. Albrecht J, Norenberg MD. Glutamine: a 18. Makris M. Optimisation of the prothrombin

Trojan horse in ammonia neurotoxicity. complex concentrate dose for warfarin
Hepatology 2006;44(4):788Y794. reversal. Thromb Res 2005;115(6):451Y453.

6. Larsen FS, Gottstein J, Blei AT. Cerebral 19. Munoz SJ, Ballas SK, Moritz MJ, et al.
hyperemia and nitric oxide synthase in rats Perioperative management of fulminant

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 637

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

 Metabolic Encephalopathies

and subfulminant hepatic failure with involvement in a diabetic uremic patient.

therapeutic plasmapheresis. Transplant Proc Clin Neurol Neurosurg 2009;111(5):477Y479.
1989;21(3):3535Y3536.
34. Young GB, Bolton CF, Austin TW, et al. The
20. Caraceni P, Van Thiel DH. Acute liver failure. encephalopathy associated with septic
Lancet 1995;345(8943):163Y169. illness. Clin Invest Med 1990;13(6):297Y304.
21. Sen S, Rose C, Ytrebo LM, et al. Effect of 35. Iacobone E, Bailly-Salin J, Polito A, et al.
albumin dialysis on intracranial pressure Sepsis-associated encephalopathy and its
increase in pigs with acute liver failure: a differential diagnosis. Crit Care Med 2009;
randomized study. Crit Care Med 2006;34(1): 37(10 suppl):S331YS336.
158Y164.
36. Nguyen DN, Spapen H, Su F, et al. Elevated
22. Felipo V, Butterworth RF. Mitochondrial serum levels of S-100beta protein and
dysfunction in acute hyperammonemia. neuron-specific enolase are associated with
Neurochem Int 2002;40(6):487Y491. brain injury in patients with severe sepsis
and septic shock. Crit Care Med 2006;34(7):
23. Ellis AJ, Wendon JA, Williams R. Subclinical
1967Y1974.
seizure activity and prophylactic phenytoin
infusion in acute liver failure: a controlled 37. Sharshar T, Annane D, de la Grandmaison
clinical trial. Hepatology 2000;32(3):536Y541. GL, et al. The neuropathology of septic
shock. Brain Pathol 2004;14(1):21Y33.
24. Bhatia V, Batra Y, Acharya SK. Prophylactic
phenytoin does not improve cerebral 38. Sharshar T, Carlier R, Bernard F, et al. Brain
edema or survival in acute liver failureVa lesions in septic shock: a magnetic resonance
controlled clinical trial. J Hepatol 2004;41(1): imaging study. Intensive Care Med 2007;
89Y96. 33(5):798Y806.
25. Claassen J, Mayer SA, Kowalski RG, et al. 39. Eidelman LA, Putterman D, Putterman C,
Detection of electrographic seizures with et al. The spectrum of septic encephalopathy.
continuous EEG monitoring in critically ill Definitions, etiologies, and mortalities.
patients. Neurology 2004;62(10):1743Y1748. JAMA 1996;275(6):470Y473.
26. Towne AR, Waterhouse EJ, Boggs JG, et al. 40. Young GB, Bolton CF, Archibald YM,
Prevalence of nonconvulsive status et al. The electroencephalogram in
epilepticus in comatose patients. Neurology sepsis-associated encephalopathy. J Clin
2000;54(2):340Y345. Neurophysiol 1992;9(1):145Y152.
27. Oddo M, Carrera E, Claassen J, et al. 41. Schrier RW, Gross P, Gheorghiade M, et al.
Continuous electroencephalography in the Tolvaptan, a selective oral vasopressin
medical intensive care unit. Crit Care Med V2-receptor antagonist, for hyponatremia.
2009;37(6):2051Y2056. N Engl J Med 2006;355(20):2099Y2112.
28. Lee WM, Hynan LS, Rossaro L, et al. 42. Zeltser D, Rosansky S, van Rensburg H, et al.
Intravenous N-acetylcysteine improves Assessment of the efficacy and safety of
transplant-free survival in early stage intravenous conivaptan in euvolemic and
non-acetaminophen acute liver failure. hypervolemic hyponatremia. Am J Nephrol
Gastroenterology 2009;137(3):856Y864. 2007;27(5):447Y457.
29. O’Grady JG, Alexander GJ, Hayllar KM, et al. 43. Wijdicks EF, Vermeulen M, Murray GD,
Early indicators of prognosis in fulminant et al. The effects of treating hypertension
hepatic failure. Gastroenterology 1989; following aneurysmal subarachnoid
97(2):439Y445. hemorrhage. Clin Neurol Neurosurg 1990;
92(2):111Y117.
30. Fraser C. Neurologic manifestations of the
uremic state. Boston, MA: Little Brown, 1992. 44. Aiyagari V, Deibert E, Diringer MN.
Hypernatremia in the neurologic intensive
31. Khatri M, Wright CB, Nickolas TL, et al. care unit: how high is too high? J Crit Care
Chronic kidney disease is associated with 2006;21(2):163Y172.
white matter hyperintensity volume: the
Northern Manhattan Study (NOMAS). Stroke 45. Moder KG, Hurley DL. Fatal hypernatremia
2007;38(12):3121Y3126. from exogenous salt intake: report of a
case and review of the literature. Mayo Clin
32. Kumar G, Goyal MK. Lentiform Fork sign: a Proc 1990;65(12):1587Y1594.
unique MRI picture. Is metabolic acidosis
responsible? Clin Neurol Neurosurg 2010; 46. Harper CG, Giles M, Finlay-Jones R. Clinical
signs in the Wernicke-Korsakoff complex: a
112(9):805Y812.
retrospective analysis of 131 cases diagnosed
33. Yoon CH, Seok JI, Lee DK, et al. Bilateral at necropsy. J Neurol Neurosurg Psychiatry
basal ganglia and unilateral cortical 1986;49(4):341Y345.

638 www.aan.com/continuum June 2012

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

47. Caine D, Halliday GM, Kril JJ, et al. delirium and fever in the hospitalized
Operational criteria for the classification of elderly. Arch Fam Med 1993;2(3):293Y297.
chronic alcoholics: identification of
61. Hopkins RO, Jackson JC. Long-term
Wernicke’s encephalopathy. J Neurol
neurocognitive function after critical illness.
Neurosurg Psychiatry 1997;62(1):51Y60.
Chest 2006;130(3):869Y878.
48. Skullerud K, Andersen SN, Lundevall J.
62. Jackson JC, Hart RP, Gordon SM, et al.
Cerebral lesions and causes of death in male
Six-month neuropsychological outcome of
alcoholics. A forensic autopsy study. Int J
medical intensive care unit patients. Crit
Legal Med 1991;104(4):209Y213.
Care Med 2003;31(4):1226Y1234.
49. Torvik A, Lindboe CF, Rogde S. Brain lesions
63. Sharma P, Eesa M, Scott JN. Toxic and
in alcoholics. A neuropathological study
acquired metabolic encephalopathies: MRI
with clinical correlations. J Neurol Sci 1982;
appearance. AJR Am J Roentgenol 2009;
56(2Y3):233Y248.
193(3):879Y886.
50. Cook CC, Hallwood PM, Thomson AD.
64. Whitmer RA, Karter AJ, Yaffe K, et al.
B vitamin deficiency and neuropsychiatric
Hypoglycemic episodes and risk of dementia
syndromes in alcohol misuse. Alcohol
in older patients with type 2 diabetes
Alcohol 1998;33(4):317Y336.
mellitus. JAMA 2009;301(15):1565Y1572.
51. Ely EW, Inouye SK, Bernard GR, et al.
65. Inouye SK, Rushing JT, Foreman MD, et al.
Delirium in mechanically ventilated patients:
Does delirium contribute to poor hospital
validity and reliability of the confusion
outcomes? A three-site epidemiologic study.
assessment method for the intensive care
J Gen Intern Med 1998;13(4):234Y242.
unit (CAM-ICU). JAMA 2001;286(21):
2703Y2710. 66. Riker RR, Shehabi Y, Bokesch PM, et al.
Dexmedetomidine vs midazolam for
52. Frontera JA. Delirium and sedation in the
sedation of critically ill patients: a
ICU. Neurocrit Care 2011;14(3):463Y474.
randomized trial. JAMA 2009;301(5):489Y499.
53. Sessler CN, Gosnell MS, Grap MJ, et al. The
67. Shehabi Y, Riker RR, Bokesch PM, et al.
Richmond Agitation-Sedation Scale: validity
Delirium duration and mortality in lightly
and reliability in adult intensive care unit
sedated, mechanically ventilated intensive
patients. Am J Respir Crit Care Med 2002;
care patients. Crit Care Med
166(10):1338Y1344.
2010;38(12):2311Y2318.
54. Francis J. Drug-induced delirium: diagnosis
and treatment. Auckland, New Zealand: 68. Girard TD, Kress JP, Fuchs BD, et al.
Efficacy and safety of a paired sedation
Mairangi Bay, 1996.
and ventilator weaning protocol for
55. Benbadis SR, Sila CA, Cristea RL. Mental mechanically ventilated patients in intensive
status changes and stroke. J Gen Intern Med care (Awakening and Breathing Controlled
1994;9(9):485Y487. trial): a randomised controlled trial. Lancet
56. Vespa PM, O’Phelan K, Shah M, et al. Acute 2008;371(9607):126Y134.
seizures after intracerebral hemorrhage: a 69. Girard TD, Jackson JC, Pandharipande PP,
factor in progressive midline shift and et al. Delirium as a predictor of long-term
outcome. Neurology 2003;60(9):1441Y1446. cognitive impairment in survivors of
57. Alberts VP, Bos MJ, Koudstaal PJ, et al. critical illness. Crit Care Med 2010;38(7):
Heart failure and the risk of stroke: the 1513Y1520.
Rotterdam Study. Eur J Epidemiol 2010;25(11): 70. Loh E, Sutton MS, Wun CC, et al. Ventricular
807Y812. dysfunction and the risk of stroke after
58. Salerno D, Marik PE, Daskalakis C, et al. The myocardial infarction. N Engl J Med 1997;
role of head computer tomographic scans 336(4):251Y257.
on the management of MICU patients with
71. Adelow C, Andersson T, Ahlbom A, et al.
neurological dysfunction. J Intensive Care Prior hospitalization for stroke, diabetes,
Med 2009;24(6):372Y375. myocardial infarction, and subsequent risk
59. Kolls BJ, Husain AM. Assessment of hairline of unprovoked seizures. Epilepsia 2011;
EEG as a screening tool for nonconvulsive 52(2):301Y307.
status epilepticus. Epilepsia 2007;48(5):
72. Morandi A, Gunther ML, Vasilevskis EE, et al.
959Y965.
Neuroimaging in delirious intensive care unit
60. Warshaw G, Tanzer F. The effectiveness of patients: a preliminary case series report.
lumbar puncture in the evaluation of Psychiatry (Edgmont) 2010;7(9):28Y33.

Continuum Lifelong Learning Neurol 2012;18(3):611–639 www.aan.com/continuum 639

Copyright © American Academy of Neurology. Unauthorized reproduction of this article is prohibited.