Sepsis-associated encephalopathy and its differential diagnosis

Emanuele Iacobone, MD; Juliette Bailly-Salin, MD; Andrea Polito, MD; Diane Friedman, MD;
Robert D. Stevens, MD; Tarek Sharshar, MD, PhD

Sepsis is often complicated by an acute and reversible dete-
rioration of mental status, which is associated with increased
mortality and is consistent with delirium but can also be revealed
by a focal neurologic sign. Sepsis-associated encephalopathy is
accompanied by abnormalities of electroencephalogram and so-
matosensory-evoked potentials, increased in biomarkers of brain
injury (i.e., neuron-specific enolase, S-100 ␤-protein) and, fre-
quently, by neuroradiological abnormalities, notably leukoen-
cephalopathy. Its mechanism is highly complex, resulting from
both inflammatory and noninflammatory processes that affect all
brain cells and induce blood-brain barrier breakdown, dysfunc-
tion of intracellular metabolism, brain cell death, and brain inju-
ries. Its diagnosis relies essentially on neurologic examination
that can lead one to perform specific neurologic tests. Electroen-
cephalography is required in the presence of seizure; neuroimag-
ing in the presence of seizure, focal neurologic signs or suspicion
of cerebral infection; and both when encephalopathy remains
unexplained. In practice, cerebrospinal fluid analysis should be
performed if there is any doubt of meningitis. Hepatic, uremic, or
respiratory encephalopathy, metabolic disturbances, drug over-
dose, withdrawal of sedatives or opioids, alcohol withdrawal
delirium, and Wernicke’s encephalopathy are the main differential
diagnoses of sepsis-associated encephalopathy. Patient manage-
ment is based mainly on controlling infection, organ system
failure, and metabolic homeostasis, at the same time avoiding
neurotoxic drugs. (Crit Care Med 2009; 37[Suppl.]:S331–S336)
KEY WORDS: sepsis; encephalopathy; delirium; blood brain bar-
rier; neuroimaging

S epsis is often associated with
an acute and reversible deteri-
oration of mental status, which
affects preferentially con-
sciousness, awareness, cognition, and be-
havior and, therefore, matches with cur-
rent criteria for delirium (1). There are
various denominations for this disorder,
among which the most frequent used are
sepsis-associated encephalopathy (SAE),
sepsis-associated delirium, or brain dys-
function. The term sepsis encephalopa-
thy is misleading as it refers to a direct
cerebral infection, whereas SAE is con-
sidered a diffuse cerebral dysfunction as a
consequence of the systemic inflamma-
tory response to an infection with no
direct central nervous system infestation
From the General Intensive Care Unit (EI), Fermo
Hospital Ancona, Italy; General Intensive Care Unit
(JB-S, AP, DF, TS), Raymond Poincaré Teaching Hos-
pital (AP-HP), University of Versailles Saint-Quentin en
Yvelines, Garches, France; Departments of Intensive
Care Medicine (SS), Hospital of Sud Essonne, Etampes,
France; and the Departments of Anesthesiology and
Critical Care Medicine, Neurology, and Neurosurgery
(RDS), Johns Hopkins University School of Medicine,
Baltimore, MD.
The authors have not disclosed any potential con-
flicts of interest.
For information regarding this article, E-mail:
tarek.sharshar@rpc.aphp.fr
Copyright © 2009 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/CCM.0b013e3181b6ed58
(1). SAE is accompanied by abnormalities
of electroencephalography (EEG) and so-
matosensory-evoked potentials, increase
in biomarkers of brain injury (i.e., neu-
ron-specific enolase, S-100 ␤-protein)
and, frequently, by neuroradiological ab-
normalities. Occurrence of encephalopa-
thy in a septic patient implies a system-
atic diagnostic approach of all potential
factors, in addition to sepsis, that contrib-
utes to the brain dysfunction including
drug toxicity or metabolic disturbances.
First, meningitis or brain abscess should
be ruled out if there is any doubt. The
purpose of this review is to address the
pathophysiology, differential diagnosis,
and outcome of SAE.
PATHOPHYSIOLOGY
Brain Signaling in Sepsis
The response to stress is physiologi-
cally triggered by an activating signal
that is mediated by two pathways (2, 3).
The first one is the vagus nerve, which
can detect visceral inflammation through
its axonal cytokines receptors. The vagus
nerve nucleus is connected to various
brainstem autonomic nuclei, notably the
nucleus tractus solitarius that integrates
the baroreflex, but also the paraventricu-
lar nuclei that control adrenal axis and
vasopressin secretion. The second path-
way involves the circumventricular or-
gans, which are strategically located near
neuroendocrine and neurovegetative nu-
clei, are deprived of a blood-brain barrier
(BBB), and express components of innate
and adaptive immune systems. Once vis-
ceral or systemic inflammation is de-
tected by the first or the second pathway,
the activating signal will spread to behav-
ioral, neuroendocrine, and neurovegeta-
tive centers. Various mediators are in-
volved including proinflammatory and
anti-inflammatory cyokines, nitric oxide,
and prostaglandins but also chemokines,
carbon monoxide. It will affect directly or
indirectly migroglial cells, astrocytes, and
neurons, ending up in a modulation of
neurosecretion and neurotransmission.
It is therefore conceivable that this brain-
activating signal can become pathogenic
and induce such an alteration of neuro-
transmission, which is the pathophysio-
logical substratum of encephalopathy. In
a clinical point of view, it must be em-
phasized that response to stress-related
changes in behavior can be similar to
those observed during an encephalopa-
thy. Furthermore, behavioral response is
mainly controlled by the amygdala and
hippocampus, which are liable to hemo-
dynamic and metabolic (i.e., hypoxemia
and hypoglycemia) insults. Therefore,
modification of behavior can be adaptive
and physiologic, i.e., behavioral features

Crit Care Med 2009 Vol. 37, No. 10 (Suppl.) S331


Hemodynamic failure
Brain vascular dysfunction
Coagulopathy Infection
Drugs toxicity
Metabolic disorders
Ischemia
Hypoxemia
Dysglycemia
Brain
damage
Apoptosis
Figure 1. Mechanisms of sepsis-associated encephalopathy. BBB, blood-brain barrier.
of the response to stress, or maladaptive
and pathophysiological, i.e., conse-
quences of SAE (4).
In addition to these two pathways, en-
dothelial cells play a major role in sepsis-
associated brain inflammation. Sepsis in-
duces their activation, which results in
release of various mediators into the brain
but also in BBB dysfunction (Fig. 1).
Alteration of Neurotransmission
The cholinergic (5), brain ␤-adrener-
gic, ␥-aminobutyric acid, and serotonin-
ergic release or receptors expression (6,
7) are impaired during experimental sep-
sis. There phenomena predominate in
cortex and in hippocampus (8) and may
be mediated by nitric oxide, cytokines,
and prostaglandins (7, 9). Neurotransmit-
ter synthesis is also altered by ammonium
and tyrosine, tryptophan and phenylala-
nine, whose plasma levels are increased
secondary to liver dysfunction and muscle
proteolysis (10). Their neurotoxic effect
might be potentiated by the decrease in
branched-chain amino acids (10).
Mitochondrial Dysfunction,
Oxidative Stress, and Apoptosis
Various brain areas of septic rats, es-
pecially in the hippocampus and cortex,
are liable to an early but transitory oxi-
dative stress (11, 12), which can be in-
duced by nitric oxide, via formation of
peroxinitrite (13) but also a decrease in
antioxidant factors (heat shock factor,
ascorbate) (14) imbalance between super-
S332
Endothelial activation
Intense activating brain
signal
BBB breakdown
Inflammatory Neurotoxic
mediators substances
Oxydative stress
Cellular dysfunction Neurotransmitters
ENCEPHALOPATHY
oxide dismutase and catalase activity (11)
as well as mitochondrial dysfunction
(12). It can be also due to hyperglycemia
and hypoxemia. It has been experimen-
tally demonstrated that expression of in-
ducible nitric oxide synthase (iNOS), dys-
function of respiratory chain, and
formation of superoxide anions within a
medullar autonomic center sequentially
precede hypotension (13). However, it is
interesting to note that previous studies
have not shown any decrease in brain en-
ergy during sepsis (15).
One major consequence of oxidative
stress is apoptosis. Mitochondrial-medi-
ated apoptosis has been evidenced in sep-
tic rats’ brains (16) that might be related
to a decrease in intracellular proaptoptic
(bcl-2) and an increase in antiapoptotic
(bax) factors (16, 17). In patients who had
died from septic shock, neuronal and mi-
croglial apoptosis has been detected in
neurovegetative and neuroendocrine nu-
clei as well as in amygdala (18). Intensity
of apoptosis correlated with expression of
endothelial iNOS. In addition to nitric
oxide, other proapoptotic factors have
been incriminated, such as glutamate, tu-
mor necrosis factor (TNF)-␣ and glucose.
The excito-neurotoxicity of glutamate has
been extensively investigated in various
neurologic disorders including acute
stroke and chronic neurodegeneration
(19). Interestingly, recycling and gluta-
mate export of ascorbate by astrocytes are
inhibited in sepsis (20, 21) and both
plasma and cerebrospinal fluid ascorbate
levels are decreased in patients with SAE
(22). Large amounts of glutamate are also
released by activated microglia, cell acti-
vation that is consistently observed dur-
ing sepsis (23).
The proapoptotic role of TNF-␣, along
with other proinflammatory cytokines, is
supported by the report of multifocal ne-
crotizing leukoencephalopathy in a pa-
tient who had died of septic shock (24).
This disease is characterized by apoptotic
and marked inflammatory lesions of the
pons and excessive systemic inflamma-
tory response. On the other hand, no
correlation was found between TNF-␣ ex-
pression and neuronal or microglial apo-
ptosis (18). Finally, hyperglycemia in-
creases microglial vulnerability to
lipopolysaccharide (LPS)-mediated toxic-
ity (25).
It must be remembered that there are
highly complex interactions between
neurons and glial cells that determine
brain cell survival. Thus, activated micro-
glial cells can become neuroprotective or
neurotoxic. Furthermore, we should not
always consider apoptosis as a deleterious
phenomenon. For instance, it has been
demonstrated to facilitate plasticity.
Endothelial Activation and BBB
Breakdown
LPS and proinflammatory cytokines
trigger the endothelial expression of
CD40, E-selectin, vascular adhesion mol-
ecule-1, and intercellular adhesion mole-
cule-1, activating both the endothelial cy-
clooxygenase 2 synthesis and the I␬B-␣/
nuclear factor-⌲b (NF-␬B) pathway. LPS
also induces endothelial expression of in-
terleukin-1 and TNF-␣ receptors, produc-
tion of interleukin-1␤, TNF-␣, and inter-
leukin-6 as well as the activation of both
endothelial and iNOS (2, 3). Then, this
endothelial activation relays into the
brain the inflammatory response by re-
leasing proinflammatory cytokines and
nitric oxide that are able to interact with
surrounding brain cells.
Endothelial activation alters vascular
tone and induces both microcirculatory
dysfunction and coagulopathy, which will
favor ischemic or hemorrhagic lesions.
However, changes in cerebral flow and its
autoregulation during sepsis are contro-
versial. Thus, both human and experi-
mental sepsis studies found an increased,
unchanged, or reduced cerebral flow (2)
as well as an altered or preserved autoreg-
ulation (2). However, ischemia is consis-
tently observed in brain areas susceptible to
low cerebral flow (24). Furthermore, it has
Crit Care Med 2009 Vol. 37, No. 10 (Suppl.)
been recently shown that sepsis-associ-
ated delirium is more likely associated
with disturbed autoregulation than with
altered cerebral flow or tissue oxygen-
ation (26). Neuropathological examina-
tion reveals hemorrhages in about 9% of
patients who died from septic shock,
which were always associated with clot-
ting disorders (24). However, clotting
tests, platelets count, and occurrence of
disseminated intravascular coagulopathy
were reported to not statistically differ
between septic shock patients with and
without hemorrhages (24).
The other consequences of endothelial
activation include impairment of oxygen,
nutrient, and metabolite movements but,
most importantly, BBB breakdown,
which will facilitate the passage of neu-
rotoxic factors. Interestingly, glutamate
plays an important role in BBB perme-
ability (27). Breakdown of BBB has been
evidenced in an experimental model of
sepsis (28, 29) but also in septic patients,
with help of brain magnetic resonance
imaging (30, 31). BBB breakdown can be
localized around the Virchow-Robin
spaces or diffuse in the whole white mat-
ter. It can also predominate in posterior
lobes, being consistent with a posterior
reversible encephalopathy syndrome.
FEATURES, PREVALENCE, AND
OUTCOME
Its clinical characteristics are changes
in mental status, especially of awareness/
consciousness and cognition. Alteration
of the former range from alteration to
sleep/wake cycle to deep coma and that of
the latter include symptoms of delirium,
i.e., disorientation, hallucination, im-
paired attention, or disorganized think-
ing. It is often associated with changes in
motor activity, ranging from agitation to
hypoactivity (2). Agitation and somno-
lence can occur alternatively. Other but
less frequent motor symptoms include
paratonic rigidity, asterixis, tremor, and
multifocal myoclonus. Alterations of EEG
activity are present most of the time and,
according to severity, include excessive
theta, predominant delta, triphasic waves
and suppression or burst suppression
(32). It is also associated with alteration
of somatosensory-evoked potentials la-
tency or amplitude (33, 34). Elevated
plasma levels of neuron-specific enolase
and S-100 ␤-protein have been reported
in patients with SAE (35) but seem not to
be correlated with clinical or EEG sever-
ity (36). As forementioned, brain mag-
Crit Care Med 2009 Vol. 37, No. 10 (Suppl.)
Sedation necessary
Non sedated Sedated What
strategy?
Discontinuation of sedation
whenever possible
EXAMINATION Normal Monitoring
LUMBAR PUNCTURE IF ANY DOUBT
Focal sign Brain imaging
Myoclonus EEG - Biochemistry - Drugs Brain imaging
Antagonist?
Antagonist?
Coma EEG - Biochemistry - Drugs Brain imaging
Delirium Biochemistry - Drugs EEG Brain imaging
Agitation
UNCONTROLLED SEPSIS?
Figure 2. Proposed decision tree for assessment of brain dysfunction in a septic patient. EEG,
electroencephalogram.
netic resonance imaging can reveal cere- Detection
bral infarctions, posterior reversible
encephalopathy syndrome, or localized to First of all, detection should be based
diffuse leukoencephalopathy (30, 31). on daily neurologic assessment of mental
However, brain magnetic resonance im- status (2, 3). Physicians have at their dis-
aging may fail to detect some brain le- posal validated scores for detecting delir-
sions observed in neuropathological stud- ium and monitoring awareness. For de-
ies, such as hemorrhages related to lirium, the Confusion Assessment
disseminated intravascular coagulopathy, Method for the Intensive Care Unit (39)
microabscesses or multifocal necrotizing or the Intensive Care Unit Delirium
leukoencephalopathy (24). Screening Checklist (40) can be pro-
Sepsis is undoubtedly a cause of agi- posed. For awareness, Glasgow Coma
tation (37) and delirium (3). About a Scale or scores used for monitoring seda-
third of septic patients had a Glasgow tion, such as the Richmond Agitation and
Coma Scale of ⬍12, and alteration of Sedation Scale or the Assessment to In-
alertness and consciousness is an inde- tensive Care Environment (41), can be
pendent prognosis factor, increasing used. Once delirium is identified, an ex-
mortality rate up to 63% when Glasgow haustive neurologic examination assess-
Coma Scale drops to ⬍8 (38). Mortality ing neck stiffness, motor responses, mus-
also increases with severity of electro- cular strength, plantar and deep tendon
physiological abnormalities, ranging reflexes and cranial nerves is mandatory.
from 0 when EEG is interpreted as nor- In case of focal neurologic sign, brain
mal to 67% when it shows burst sup- imaging will be indicated. In case of sei-
pressions. zure (including palpebral myoclonus) or
Surprisingly, the mortality rate also neck stiffness, an EEG and a lumbar
increases with plasma levels of biomark- puncture will be required, before or after
ers, although they do not correlate with neuroimaging (Fig. 2). The main limita-
clinical and neurophysiological severity tion of clinical detection is sedation, as
(3). There are long-term cognitive and sedatives alter awareness and cognition,
psychological sequelae of SAE are un- even after their discontinuation. Despite
known. Although SAE is considered a re- this, we think that a neurologic examina-
versible phenomenon, neuroradiological tion should be performed daily in sedated
and neuropathological findings can plau- patients (Table 1). Occurrence of sudden
sibly cause disability. changes in mental status unexplained by
modification of sedative infusion rate, fo-
DIAGNOSTIC APPROACH cal neurologic sign, seizure, or neck stiff-
ness should lead the physician to discuss
Prevalence and severity of encephalop- neuroimaging, EEG, and lumbar punc-
athy in septic patients should prompt ture. Beside clinical examination, neuro-
physicians to detect a brain dysfunction, physiological testing and assessment of
diagnose its causes, and finally, whenever biomarkers have been proposed for de-
possible, propose a treatment. tecting brain dysfunction in sedated sep-
S333
Table 1. Neurologic approach in sedated patients
Assessment Symptoms
Motor activity Agitated
Awarenesss/awakeness Comatose
Cognition Delirium
Motor and brainstem 1. Modification not
responses explained by sedation
2. Focal signs
3. Myoclonus
MRI, magnetic resonance imaging; EEG, electroencephalogram.
a
Biological, neuroradiological, and electrophysiological complementary investigations must be
discussed; buse of antagonist of benzodiazepines or opioids can be discussed.
tic patients. In contrast to EEG, somato-
sensory-evoked potentials may be useful
as they are not affected by sedative drugs
(33). Even if they can detect subclinical
focal signs and their subcortical and cor-
tical peak latencies are correlated with
severity of the systemic inflammatory re-
sponse syndrome, they are too cumber-
some to be performed routinely at the
patient’s bedside. The usefulness of
bispectral index is controversial in criti-
cally ill patients (42). Plasma levels of
biomarkers of brain cell damages, such as
neuron specific enolase and S-100 ␤-pro-
tein, can be more easily performed. How-
ever, their specificity and sensitivity for
detecting sepsis-associated brain dysfunc-
tion or lesions have not been demon-
strated sufficiently to recommend their
measurement routinely. Neither neuro-
imaging nor cerebrospinal fluid analysis
can be considered a monitoring tool.
Complementary Investigations
After neurologic examination has been
performed, specific tests need to be ad-
dressed. In routine practice, cerebrospi-
nal fluid analysis should be always done
whenever meningitis is suspected. EEG
should be performed if there is seizure
(including palpebral myoclonus) or if al-
tered neurologic status remains unex-
plained as it may be secondary to non-
convulsive epilepsy, which is treatable. In
septic patients, evidences for focal neuro-
logic signs or seizure are undisputable
criteria for brain imaging. It should be
also considered when no cause for en-
cephalopathy has been identified. Mag-
netic resonance imaging is more helpful
than computed tomography scan for de-
tecting recent ischemic stroke, white
matter lesions, or BBB rupture, and for
exploring cerebral arteries. Before trans-
porting the patient, risk and benefit of
S334
Interruption
of Sedation Investigation
Careful If no obvious explanationa
Yes If persistenta,b
Yes If persistenta
1. Yes 1. If persistenta
2. Not necessary 2. Any focal sign (MRI)
3. No 3. Necessary (EEG, MRI . . . )
brain imaging should be balanced. In ad-
dition to etiological diagnosis for the phy-
sicians and the relatives of the patient,
assessment of the nature and extent of
brain damage may also influence the pa-
tient’s treatments. For example, evidence
for brain hemorrhage should lead to dis-
continuation of any drug with an antico-
agulant activity.
Beside these specific tests, standard
laboratory tests must be done to detect
metabolic disturbances, liver dysfunc-
tion, and renal insufficiency as well as
drug chart screening to identify neuro-
toxic treatment that could be discontin-
ued or tapered according to their plasma
levels or to the liver and/or renal func-
tion. In addition to sedatives and analge-
sics, many classes of drugs currently ad-
ministered in critically ill patients are
neurotoxic, notably a number of antibiot-
ics and cardiac drugs. It must be empha-
sized that encephalopathy is often multi-
factorial; therefore, a physician should
not stop his/her query once one factor
has been identified. Finally, reappearance
or persistence of encephalopathy may in-
dicate that sepsis is not controlled and is
a classic feature of deep abscess.
Differential Diagnosis
As aforementioned, a brain infection
must always be suspected and relevance
of brain imaging and lumbar puncture
must always addressed in a septic patient
with any central neurologic symptoms.
Encephalopathy in critically ill patients
can result from various causes, which can
be entwined, masked, or worsened by a
septic process. Sepsis can be the trigger-
ing and aggravating factor of hepatic or
uremic encephalopathy. Anamnesis, re-
spective severity of sepsis and liver or
renal failure can help to discriminate be-
tween these processes.
Septic patients are also susceptible to
drug overdose but also drug withdrawal,
notably of benzodiazepines and opioids as
they are often prescribed to these pa-
tients. The chronological link and the
neurologic improvement after their read-
ministration are arguments for a with-
drawal syndrome. Tobacco dependency is
a risk factor for delirium in critically ill
patients whose dependency can be re-
versed by a nicotine patch.
Alcohol withdrawal delirium is often
evoked in an alcoholic patient who devel-
ops an encephalopathy. Delirium tre-
mens is a potentially fatal complication
that occurs, however, in only 5% of hos-
pitalized alcohol-dependent patients and
usually within 48 hrs to 72 hrs of the last
drink. The predominance of psychomotor
agitation, zoopsias, and autonomic signs
(hyperpyrexia, tachycardia, hypertension,
and diaphoresis) are suggestive. In mal-
nourished or alcoholic patients, Wer-
nicke’s encephalopathy must always be
evoked and treated, especially if there is
an ophthalmoplegia or ataxia. Thiamine
deficiency can be aggravated by infusion
of glucose. Encephalopathy is a sign of
respiratory failure, induced by hypoxemia
and/or hypercapnia, and indicates respi-
ratory assistance. Finally, air embolism is
an iatrogenic cause of sudden coma, ag-
itation, seizure, or focal neurologic signs,
and for which hyperbaric oxygen is rec-
ommended. Medical history of the patient
should also be taken into account as var-
ious diseases can affect the central ner-
vous system.
TREATMENT AND
THERAPEUTIC PERSPECTIVE
There is no specific treatment for sep-
sis-associated delirium yet. Thus, treat-
ment should focus on the underlying sys-
temic illness and supportive measures,
such as control of sepsis, management of
organ failure and metabolic disturbances,
and avoidance of neurotoxic drugs.
Among treatments currently used in
septic patients, intensive insulin therapy,
activated protein C, and steroids may
have some effects on the main and sup-
posed pathophysiological mechanisms of
SAE, which are endothelial activation,
BBB breakdown, neuroinflammation, ox-
idative stress, and apoptosis. However,
there is no evidence from clinical trials
that insulin therapy, activated protein C,
or steroids reduce the incidence or sever-
ity of SAE.
Crit Care Med 2009 Vol. 37, No. 10 (Suppl.)
 Insulin therapy could be neuroprotec-
tive as hyperglycemia experimentally in-
duces oxidative stress and apoptosis.
However, the potential benefit of insulin
therapy in brain-injured patients (43) has
not been confirmed in patients with an
acute stroke (44). In septic patients, it is
impossible to anticipate its effect on brain
function as cerebral metabolism of glu-
cose is highly complex and its distur-
bances in sepsis are insufficiently under-
stood. On one hand, LPS increased
permeability of the BBB to insulin, then
maximizing its central effects (45); on the
other hand, sepsis may alter brain cell
expression of glucose transporters, nota-
bly of glucose transporter 4, which is
modulated by insulin (46). Therefore,
neuronal sensitivity to hypoglycemia and
hyperglycemia might be deeply modified
during sepsis. One must remember that
hypoglycemia is harmful to the brain,
notably the hippocampus.
Experimental models showed that ste-
roids decrease systemic inflammation, re-
duce brain edema (47), improve BBB
functions (48) but also modulate micro-
glia functions (49). The expression of ste-
roid receptors in hippocampus, reduction
of N-methyl-D-aspartate receptors ex-
pression in hippocampus (50), and pre-
vention by hydrocortisone of posttrau-
matic stress syndrome (51) suggest that
they are involved in behavioral response
to stress. The anti-inflammatory and en-
dothelial effects of activated protein C in
sepsis and experimental brain ischemia
are also appealing.
Historically, inhibition of iNOS was
one of the first to be tested because of its
pathogenic role. However, iNOS inhibi-
tion does not improve consciousness but
it does reduce LPS-induced neuronal ap-
optosis in septic rats (52), aggravate brain
ischemia (53), and increase cardiovascu-
lar deaths (54). Serum amyloid P or mag-
nesium (55, 56), which have been shown
to reduce BBB permeability in septic an-
imals, and antioxidant molecules have
also been tested. For instance, riluzole, a
glutamate release inhibitor, diminishes
brain edema as well as BBB permeability,
oxidative damage, and brain injury in
septic rats (the latter two being also re-
duced by silymarin) (27, 57). If infusion
of branched-chain amino acids can re-
duce aromatic amino acids in the brain
by competing with their transport, its
effect on SAE has not been demonstrated
(58). Calcium-channel blockers and re-
moval of circulating cytokines, notably by
coupled plasma filtration adsorption,
Crit Care Med 2009 Vol. 37, No. 10 (Suppl.)
have also been proposed for reducing
BBB permeability (59). Modulation of
brain signaling is another option. It has
been shown that pharmacologic inhibi-
tion of the cholinergic pathway im-
proves sickness behavior (60).
CONCLUSIONS
The diagnosis of encephalopathy in
septic patients is crucial as it is a frequent
and severe complication. Its mechanism
is highly complex, resulting from both
inflammatory and noninflammatory pro-
cesses that affect all brain cells and in-
duce BBB breakdown, dysfunction of in-
tracellular metabolism, brain cell death,
and brain injuries. Its diagnosis relies es-
sentially on neurologic examination that
can lead the physician to perform specific
neurologic tests. EEG and neuroimaging
are required in the presence of seizure,
focal neurologic signs, or suspicion of
cerebral infection and when encephalop-
athy remains unexplained. Cerebrospinal
fluid analysis is indisputably required
when meningitis is suspected. SAE re-
sults from various and combined factors
that must systematically be screened. In
practice, brain infection must be always
evoked; then hepatic, uremic, or respira-
tory encephalopathy, metabolic distur-
bances, drug overdose, withdrawal of sed-
atives or opioids, alcohol withdrawal
delirium, or Wernicke’s encephalopathy
are the main differential diagnoses of
SAE. Currently, its treatment consists
mainly of controlling sepsis. Effects of
insulin therapy, activated protein C, and
steroids on SAE need to be assessed. Sep-
sis-induced BBB dysfunction, brain oxi-
dative stress, and inflammation but also
brain signaling are highly interesting
therapeutic but still experimental targets.
REFERENCES
1. Pandharipande P, Cotton BA, Shintani A, et
al: Motoric subtypes of delirium in mechan-
ically ventilated surgical and trauma inten-
sive care unit patients. Intensive Care Med
2007; 33:1726 –1731
2. Ebersoldt M, Sharshar T, Annane D: Sepsis-
associated delirium. Intensive Care Med
2007; 33:941–950
3. Siami S, Annane D, Sharshar T: The enceph-
alopathy in sepsis. Crit Care Clin 2008; 24:
67– 82
4. Dantzer R: Cytokine-induced sickness behav-
iour: A neuroimmune response to activation
of innate immunity. Eur J Pharmacol 2004;
500:399 – 411
5. Semmler A, Frisch C, Debeir T, et al: Long-
term cognitive impairment, neuronal loss
and reduced cortical cholinergic innervation
after recovery from sepsis in a rodent model.
Exp Neurol 2007; 204:733–740
6. Kadoi Y, Saito S: An alteration in the gamma-
aminobutyric acid receptor system in experi-
mentally induced septic shock in rats. Crit
Care Med 1996; 24:298 –305
7. Kadoi Y, Saito S, Kunimoto F, et al: Impair-
ment of the brain beta-adrenergic system
during experimental endotoxemia. J Surg
Res 1996; 61:496 –502
8. Hellstrom IC, Danik M, Luheshi GN, et al:
Chronic LPS exposure produces changes in
intrinsic membrane properties and a sus-
tained IL-beta-dependent increase in
GABAergic inhibition in hippocampal CA1
pyramidal neurons. Hippocampus 2005; 15:
656 – 664
9. Pavlov VA, Ochani M, Gallowitsch-Puerta M,
et al: Central muscarinic cholinergic regula-
tion of the systemic inflammatory response
during endotoxemia. Proc Natl Acad Sci
U S A 2006; 103:5219 –5223
10. Basler T, Meier-Hellmann A, Bredle D, et al:
Amino acid imbalance early in septic enceph-
alopathy. Intensive Care Med 2002; 28:
293–298
11. Barichello T, Fortunato JJ, Vitali AM, et al:
Oxidative variables in the rat brain after sep-
sis induced by cecal ligation and perforation.
Crit Care Med 2006; 34:886 – 889
12. d’Avila JC, Santiago AP, Amancio RT, et al:
Sepsis induces brain mitochondrial dysfunc-
tion. Crit Care Med 2008; 36:1925–1932
13. Chan JY, Chang AY, Wang LL, et al: Protein
kinase C-dependent mitochondrial transloca-
tion of proapoptotic protein Bax on activa-
tion of inducible nitric-oxide synthase in
rostral ventrolateral medulla mediates car-
diovascular depression during experimental
endotoxemia. Mol Pharmacol 2007; 71:
1129 –1139
14. Christians ES, Yan LJ, Benjamin IJ: Heat
shock factor 1 and heat shock proteins: crit-
ical partners in protection against acute cell
injury. Crit Care Med 2002; 30:S43–S50
15. Hotchkiss RS, Karl IE: Reevaluation of the
role of cellular hypoxia and bioenergetic fail-
ure in sepsis. JAMA 1992; 267:1503–1510
16. Messaris E, Memos N, Chatzigianni E, et al:
Time-dependent mitochondrial-mediated
programmed neuronal cell death survival in
sepsis. Crit Care Med 2004; 32:1764 –1770
17. Semmler A, Okulla T, Sastre M, et al: Sys-
temic inflammation induces apoptosis with
variable vulnerability of different brain re-
gions. J Chem Neuranat 2005; 30:144 –157
18. Sharshar T, Gray F, Lorin de la Grandmaison
G, et al: Apoptosis of neurons in cardiovas-
cular autonomic centres triggered by induc-
ible nitric oxide synthase after death from
septic shock. Lancet 2003; 362:1799 –1805
19. Villmann C, Becker CM: On the hypes and
falls in neuroprotection: Targeting the
NMDA receptor. Neuroscientist 2007; 13:
594 – 615
20. Wilson JX, Dragan M: Sepsis inhibits recy-
cling and glutamate-stimulated export of
S335
 ascorbate by astrocytes. Free Radic Biol Med
2005; 39:990 –998
21. Korcok J, Wu F, Tyml K, et al: Sepsis inhibits
reduction of dehydroascorbic acid and accu-
mulation of ascorbate in astroglial cultures:
Intracellular ascorbate depletion increases
nitric oxide synthase induction and gluta-
mate uptake inhibition. J Neurochem 2002;
81:185–193
22. Voigt K, Kontush A, Stuerenburg HJ, et al:
Decreased plasma and cerebrospinal fluid
ascorbate levels in patients with septic en-
cephalopathy. Free Radic Res 2002; 36:
735–739
23. Brown GC, Bal-Price A: Inflammatory neu-
rodegeneration mediated by nitric oxide, glu-
tamate, and mitochondria. Mol Neurobiol
2003; 27:325–355
24. Sharshar T, Annane D, de la Grandmaison G,
et al: The neuropathology of septic shock.
Brain Pathol 2004; 14:21–33
25. Wang JY, Yang JM, Tao PL, et al: Synergistic
apoptosis induced by bacterial endotoxin li-
popolysaccharide and high glucose in rat mi-
croglia. Neurosci Lett 2001; 304:177–180
26. Pfister D, Siegemund M, Dell-Kuster S, et al:
Cerebral perfusion in sepsis-associated delir-
ium. Crit Care 2008; 12:R63
27. Toklu HZ, Uysal MK, Kabasakal L, et al: The
effects of riluzole on neurological, brain bio-
chemical, and histological changes in early
and late term of sepsis in rats. J Surg Res
2008; 9:9
28. Papadopoulos MC, Lamb FJ, Moss RF, et al:
Faecal peritonitis causes oedema and neuro-
nal injury in pig cerebral cortex. Clin Sci
(Lond) 1999; 96:461– 466
29. Handa O, Stephen J, Cepinskas G: Role of
eNOS-derived nitric oxide (NO) in activation
and dysfunction of cerebrovascular endothe-
lial cells during early onsets of sepsis. Am J
Physiol Heart Circ Physiol 2008; 22:22
30. Sharshar T, Carlier R, Bernard F, et al: Brain
lesions in septic shock: A magnetic reso-
nance imaging study. Intensive Care Med
2007; 33:798 – 806
31. Bartynski WS, Boardman JF, Zeigler ZR, et
al: Posterior reversible encephalopathy syn-
drome in infection, sepsis, and shock. AJNR
Am J Neuroradiol 2006; 27:2179 –2190
32. Young GB, Bolton CF, Archibald YM, et al:
The electroencephalogram in sepsis-associ-
ated encephalopathy. J Clin Neurophysiol
1992; 9:145–152
33. Zauner C, Gendo A, Kramer L, et al: Impaired
subcortical and cortical sensory evoked po-
tential pathways in septic patients. Crit Care
Med 2002; 30:1136 –1139
34. Zauner C, Gendo A, Kramer L, et al: Meta-
S336
bolic encephalopathy in critically ill patients
suffering from septic or nonseptic multiple
organ failure. Crit Care Med 2000; 28:
1310 –1315
35. Nguyen DN, Spapen H, Su F, et al: Elevated
serum levels of S-100beta protein and neu-
ron-specific enolase are associated with brain
injury in patients with severe sepsis and sep-
tic shock. Crit Care Med 2006; 34:1967–1974
36. Piazza O, Russo E, Cotena S, et al: Elevated
S100B levels do not correlate with the sever-
ity of encephalopathy during sepsis. Br J An-
aesth 2007; 24:24
37. Jaber S, Chanques G, Altairac C, et al: A
prospective study of agitation in a medical-
surgical ICU: Incidence, risk factors, and out-
comes. Chest 2005; 128:2749 –2757
38. Eidelman LA, Putterman D, Putterman C, et
al: The spectrum of septic encephalopathy.
Definitions, etiologies, and mortalities. JAMA
1996; 275:470 – 473
39. Ely EW, Inouye SK, Bernard GR, et al: De-
lirium in mechanically ventilated patients.
Validity and reliability of the Confusion As-
sessment Method for the Intensive Care Unit
(CAM-ICU). JAMA 2001; 286:2703–2710
40. Bergeron N, Dubois MJ, Dumont M, et al:
Intensive Care Delirium Screening Check-
list: Evaluation of a new screening tool. In-
tensive Care Med 2001; 27:859 – 864
41. de Jonghe B, Cook D, Griffith L, et al: Adap-
tation to the Intensive Care Environment
(ATICE): Development and validation of a
new sedation assessment instrument. Crit
Care Med 2003; 31:2344 –2354
42. Ely EW, Truman B, Manzi DJ, et al: Con-
sciousness monitoring in ventilated patients:
Bispectral EEG monitors arousal not delir-
ium. Intensive Care Med 2004; 30:1537–1543
43. Van den Berghe G, Schoonheydt K, Becx P, et
al: Insulin therapy protects the central and
peripheral nervous system of intensive care
patients. Neurology 2005; 64:1348 –1353
44. Bruno A, Kent TA, Coull BM, et al: Treatment
of hyperglycemia in ischemic stroke (THIS):
A randomized pilot trial. Stroke 2008; 39:
384 –389
45. Xaio H, Banks WA, Niehoff ML, et al: Effect of
LPS on the permeability of the blood-brain
barrier to insulin. Brain Res 2001; 896:1–2
46. Maher F, Vannucci SJ, Simpson IA: Glucose
transporter proteins in brain. FASEB J 1994;
8:1003–1011
47. Kaal EC, Vecht CJ: The management of brain
edema in brain tumors. Curr Opin Oncol
2004; 16:593– 600
48. Bauer B, Hartz AM, Fricker G, et al: Modu-
lation of p-glycoprotein transport function at
the blood-brain barrier. Exp Biol Med (May-
wood) 2005; 230:118 –127
49. Sierra A, Gottfried-Blackmore A, Milner TA,
et al: Steroid hormone receptor expression
and function in microglia. Glia 2008; 56:
659 – 674
50. Weiland NG, Orchinik M, Tanapat P: Chronic
corticosterone treatment induces parallel
changes in N-methyl-D-aspartate receptor
subunit messenger RNA levels and antago-
nist binding sites in the hippocampus. Neu-
roscience 1997; 78:653– 662
51. Schelling G, Roozendaal B, Krauseneck T, et
al: Efficacy of hydrocortisone in preventing
posttraumatic stress disorder following crit-
ical illness and major surgery. Ann N Y Acad
Sci 2006; 1071:46 –53
52. Kadoi Y, Goto F: Selective inducible nitric
oxide inhibition can restore hemodynamics,
but does not improve neurological dysfunc-
tion in experimentally-induced septic shock
in rats. Anesth Analg 2004; 99:212–220
53. Li H, Forstermann U: Nitric oxide in the
pathogenesis of vascular disease. J Pathol
2000; 190:244 –254
54. Lopez A, Lorente JA, Steingrub J, et al: Mul-
tiple-center, randomized, placebo-con-
trolled, double-blind study of the nitric oxide
synthase inhibitor 546C88: Effect on survival
in patients with septic shock. Crit Care Med
2004; 32:21–30
55. Veszelka S, Urbanyi Z, Pazmany T, et al:
Human serum amyloid P component atten-
uates the bacterial lipopolysaccharide-
induced increase in blood-brain barrier per-
meability in mice. Neurosci Lett 2003; 352:
57– 60
56. Esen F, Erdem T, Aktan D, et al: Effect of
magnesium sulfate administration on blood-
brain barrier in a rat model of intraperitoneal
sepsis: A randomized controlled experimen-
tal study. Crit Care 2005; 9:R18 –R23
57. Toklu HZ, Tunali Akbay T, Velioglu-Ogunc A,
et al: Silymarin, the antioxidant component
of Silybum marianum, prevents sepsis-
induced acute lung and brain injury. J Surg
Res 2008; 145:214 –222
58. Hasselgren PO, Fischer JE: Septic encepha-
lopathy. Etiology and management. Inten-
sive Care Med 1986; 12:13–16
59. Wratten ML: Therapeutic approaches to re-
duce systemic inflammation in septic-
associated neurologic complications. Eur J
Anaesthesiol Suppl 2008; 42:1–7
60. Johnson DR, O’Connor JC, Dantzer R, et al:
Inhibition of vagally mediated immune-to-
brain signaling by vanadyl sulfate speeds re-
covery from sickness. Proc Natl Acad Sci
U S A 2005; 102:15184 –15189
Crit Care Med 2009 Vol. 37, No. 10 (Suppl.)