ness may be confusing, and the prognosis may
be much better than would be indicated by the
lack of these brainstem reflexes, particularly
if the patient receives early plasmapheresis or
intravenous immune globulin. If breathing is
also affected by the Guillain-Barre´ syndrome,
the picture may even simulate brain death.
149
This condition must be considered among
the reversible causes of coma that require ex-
clusion before brain death is declared (see
Chapter 8).
Isolated unilateral or bilateral abducens
palsy may be seen in some patients with in-
creased intracranial pressure, even due to non-
focal causes such as pseudotumor cerebri.
150
It
may also occur with low CSF pressure, with a
spontaneous leak, or after lumbar puncture.
151
In rare cases the trochlear nerve may also be
involved.
152
Motor Responses
Patients with metabolic coma may have para-
tonia and/or extensor plantar responses. How-
ever, spastic rigidity should not be present.
Rarely, patients with metabolic causes of coma,
particularly hypoglycemia,
153
will present with
asymmetric motor responses or even hemi-
plegia (see Chapter 5). Some have suggested
that the focal signs represent the unmasking of
subclinical neurologic impairment. It is true
that most metabolic causes of coma may ex-
acerbate a pre-existing neurologic focal find-
ing, but the presence and even the distribution
of focal findings in patients with hypoglycemia
may vary from one episode to the next, so that
the evidence for a structural cause is not con-
vincing. Furthermore, focal signs caused by
hypoglycemia are more common in children
than adults, again suggesting the absence of an
underlying structural lesion. Similarly, focal
deficits are observed with hypertensive en-
cephalopathy, but in this case imaging usually
identifies brain edema consistent with these
focal neurologic deficits. Cortical blindness is
the most common of these deficits; edema
of the occipital white matter is seen on mag-
netic resonance images, the so-called posterior
leukoencephalopathy syndrome.
154
A number
of severe metabolic causes of coma, especially
hepatic coma, may also cause either decere-
brate or decorticate posturing. In general, al-
though it is important to be alert to the pos-
sibility of false localizing signs in patients with
metabolic causes of coma, unless a structural
lesion can be ruled out, it is still usually nec-
essary to proceed as if the coma has a structural
cause, until proven otherwise.
MAJOR LABORATORY
DIAGNOSTIC AIDS
The neurologic examination, as described
above, is the cornerstone for the diagnosis of
stupor and coma. It can be done at the bedside
within a matter of a few minutes, and it pro-
vides critical diagnostic clues to determine the
tempo of the further evaluation. If focal find-
ings are seen, it may be necessary to institute
treatment even before the remainder of the di-
agnostic testing can be completed. The same
may be true for some types of metabolic coma,
such as meningitis or hypoglycemia. On the
other hand, if the evidence from a nonfocal
examination points toward a diffuse metabolic
encephalopathy, the examiner usually has time
to employ additional diagnostic tools.
Blood and Urine Testing
Because of the propensity for some metabolic
comas to cause focal neurologic signs, it is
important to perform basic blood and urine
testing on virtually every patient who presents
with coma. It is important to draw blood for
glucose and electrolytes, and to do toxic and
drug screening almost immediately. The blood
should not be drawn in a limb with a running
intravenous line, as this may alter the glucose
or electrolytes. Blood gases should be drawn
if there is any suspicion of respiratory insuf-
ficiency or acid-base abnormality. Urine can
then be collected for urinalysis and screening
for toxic substances or drugs (which may no
longer be detectable in the bloodstream). In a
woman of reproductive age, pregnancy test-
ing should also be done as this may affect the
evaluation (e.g., MRI scan may be preferable
to CT, if there is a choice). A bedside mea-
surement of blood glucose is sufficiently ac-
curate to rule out hypoglycemia and obviate
the need for giving glucose. However, if glu-
cose is given, 100 mg of thiamine should be
given as well to prevent precipitating Wernicke
encephalopathy (see Chapter 5).
Examination of the Comatose Patient
77
Computed Tomography Imaging
and Angiography
CT scanning is now ubiquitous, and it should
be applied to any patient who does not have an
immediately obvious source of coma (e.g., a
hypoglycemic patient who arouses with injec-
tion of IV glucose). However, it is still neces-
sary to complete the examination first, as a pa-
tient who is in incipient uncal herniation, or
whose fourth ventricle is compressed by a mass
lesion, may die even during the few minutes it
takes to get a scan, and may need to be treated
emergently first. Similarly, for comatose pa-
tients in whom meningitis is suspected, it is
now standard practice to give IV antibiotics
first, before taking the patient for a CT scan, to
rule out a mass lesion prior to doing a lumbar
puncture (but see discussion on lumbar punc-
ture below and on meningitis in Chapters 4
and 5).
Emergency CT scans done for diagnostic
purposes in patients with a depressed level
of consciousness may appear to be simple to
interpret. This is certainly the case for large,
acute hemorrhages or extensive infarcts. How-
ever, subacute infarction may become isodense
with brain during the second week, and hem-
orrhage may be isodense during the third week
after onset. Acute infarcts may be difficult to
identify, and if there is bilateral edema, it may
be quite difficult to distinguish from ‘‘hyper-
normal brain’’ (i.e., small ventricles and gen-
eral decrease in prominence of the sulci, which
may be seen in young normal brains, particu-
larly if the scan is not of good quality).
In such cases, it may be useful either to ob-
tain a CT scan with contrast, or to have an MRI
scan done (see below). Current-generation CT
scanners are fast enough that it is rarely nec-
essary to sedate a patient to eliminate motion
artifact. However, many MRI examinations still
take significantly longer, and they may be com-
promised if the patient moves. Such patients
may be sedated with a short-acting benzodiaze-
pine, which can be reversed if necessary with
flumazenil. However, conscious sedation should
only be done under the continuous supervision
of a physician who is capable of intubating the
patient if respirations are depressed or com-
promised.
Computed tomography angiography (CTA)
involves reconstruction of images of the in-
tracranial circulation from images acquired
during an intravenous bolus injection of con-
trast dye. Perfusion CT may also identify areas
of decreased perfusion, even in cases where
the plain CT does not yet show an infarct (see
Figure 2–11). CTA is highly accurate for dem-
onstrating occlusions or stenoses of intracra-
nial vessels, but does not give the resolution
of conventional direct imaging angiography.
The images can be acquired quickly and the
method is applicable to patients (see below)
who may not be eligible for magnetic reso-
nance angiography (MRA). However, extract-
ing the vascular images currently requires
more user interaction and takes longer than
MRA. The use of large amounts of contrast dye
can also be a drawback if the patient’s history
of dye reaction and renal function are not
available.
Magnetic Resonance Imaging
and Angiography
MRI scans take substantially longer than CT
scans, and they are often less available for
emergency scanning. Hence, they are less of-
ten used for primary scanning of patients with
coma. However, in many cases, it is necessary
to obtain an MRI scan if a significant question
remains about the origin of the coma after the
CT imaging. Diffusion-weighted imaging may
demonstrate an infarct that otherwise cannot
be documented acutely. Additional sequences
that measure the apparent diffusion coefficient
of water in the brain (ADC mapping) and per-
fusion with blood can be used in cases where
the standard diffusion imaging is confounded
by background T2 bright lesions. This in turn
may lead to a lifesaving intervention (e.g., intra-
arterial tPA in the case of basilar artery oc-
clusion). MRA may also demonstrate arterial
occlusion noninvasively, and MR venography
may identify a dural sinus thrombosis. While
T1 and T2 MRI sequences are not as sensitive
as CT scanning for identifying acute blood,
the combination of fluid-attenuated inversion
recovery (FLAIR) and gradient echo T2* se-
quences is at least as sensitive in acute sub-
arachnoid hemorrhage and may be more sen-
sitive if the bleeding is subacute.
155
On the other hand, MR scanning has sig-
nificant limitations for its use in many coma-
tose patients. Because MRI scanners use a
78
Plum and Posner’s Diagnosis of Stupor and Coma