In this setting, it is sometimes
easier to approach the differential
diagnosis for several of the key
symptoms and signs of the patient.
One possibility is to approach the
patient's thrombocytopenia, and
come up with a differential diagnosis.
Possibilities here include decreased
production that could arise from
a viral infection, bone marrow infiltration
by malignancy or granulomatous inflammation,
or bone marrow failure in the setting
of myelodysplastic syndrome or acute
leukemia. More commonly, thrombocytopenia
arises from increased destruction,
such as occurs from ITP (which can
be idiopathic or drug associated),
infection (either overwhelming sepsis
or viral infection), malignancy,
or consumption. Platelet consumption
occurs by either disseminated intravascular
coagulation or a microangiopathic
hemolytic anemia. A bone marrow
exam, to look for increased or decreased
megakaryocytes, or a platelet transfusion,
to indicate shortened platelet survival,
is sometimes necessary to determine
the etiology of thrombocytopenia.
Thus, it is difficult to narrow
down this patient's differential
diagnosis from her thrombocytopenia.
Alternatively,
we should approach this patient's
differential diagnosis by examining
the possible causes of her anemia.
The differential diagnosis for anemia
can be divided into hypoproliferative
states and hyperproliferative states.
Hypoproliferative anemias are generally
characterized by a low reticulocyte
count. The differential diagnosis
includes deficiencies of the building
blocks of red blood cells (Vitamin
B-12, folate or iron), parvovirus
infection (which specifically targets
erythroid precursors), bone marrow
failure or infiltration (such as
with leukemia or granulomatous disease,
respectively) or chronic inflammation
which gives rise to the anemia of
chronic disease. Hyperproliferative
anemias are characterized by increased
reticulocyte counts that cannot
compensate for increased destruction
or blood loss. The patient shows
no evidence of blood loss through
the gastrointestinal tract or elsewhere.
Hyperproliferative anemias resulting
from red blood cell destruction,
or hemolysis, can be either extravascular
or intravascular. Extravascular
destruction occurs in the spleen,
as antibodies coat red blood cells
leading to phagocytosis in the red
pulp. This type of hemolysis is
associated with a positive Coombs
test and microspherocytes on peripheral
smears; the latter results from
the pinching off of fragments of
cell membrane by the spleen's sinusoidal
macrophages. Intravascular causes
of hemolysis are associated with
increased LDH and AST (which are
released from red blood cells) and
decreased haptoglobin levels. Intravascular
hemolysis is often associated with
low platelet counts (consumptive
thrombocytopenia). Intravascular
hemolysis is characterized by schistocytes
on peripheral smear.
Examination
of the provided peripheral blood
smear helps greatly to characterize
this patient's anemia. The first
finding of note is the markedly
decreased platelet count. On a blood
smear, each platelet seen in a high
power field correlates with approximately
10,000 platelets in the peripheral
circulation. Hence, since the normal
platelet count is approximately
140,000, one should see approximately
14 platelets in a high power field,
and we see far fewer than that in
the provided peripheral smear images.
We note that the few white blood
cells seen in the smear appear normal,
which correlates with the normal
white count noted in the peripheral
smear. The most striking finding
in the blood smear, however, relates
to the red blood cells. We see indirect
evidence of reticulocytosis in the
marked polychromasia of the red
blood cells. While we see a few
microspherocytes, the dominant finding
is that of numerous schistocytes
and helmet shaped cells. The presence
of these numerous schistocytes establishes
that this patient has a microangiopathic
hemolytic anemia.
Microangiopathic
hemolytic anemias occur from mechanical
shearing of the red cells, and consumption
of platelets, and the differential
for this syndrome is limited. One
of the more common causes is disseminated
intravascular coagulation (DIC).
In a patient with a skin rash, as
the current patient had, meningococcal
sepsis leading to DIC is a possibility.
However, patients with DIC typically
have abnormal coagulation factors
(increased PT and aPTT, decreased
fibrinogen), which this patient
did not have. Patients with sepsis
also typically have fever and hypotension.
Another possibility is the HELLP
(hemolysis, elevated liver enzymes,
low platelets) syndrome, which may
develop in pregnant patients and
is associated with marked liver
function abnormalities. Neither
of these was true of the current
patient. Another possibility is
the "waring blender syndrome",
in which hemolysis is associated
with heart valve abnormalities.
The absence of a cardiac murmur
or prior valve replacement argues
strongly against this possibility.
Occasionally, patients with vasculitis,
lupus or antiphospholipid syndrome
present with microangiopathic hemolytic
anemia; however, this patient did
not demonstrate any signs of these
disorders. Yet another possibility
is malignant hypertension, though
this is eliminated by the normal
blood pressure.
The
final category of microangiopathic
hemolytic anemia includes the distinct
but related, thrombotic syndromes,
thrombotic thrombocytopenic purpura
(TTP) and the hemolytic-uremic syndrome
(HUS). The syndromes are characterized
by thrombocytopenia, hemolytic anemia,
fever, renal abnormalities, and
neurologic abnormalities. HUS is
characterized by greater renal dysfunction,
less severe thrombocytopenia, less
elevation of LDH, and fewer schistocytes
than TTP, while TTP is characterized
by predominant neurologic abnormalities,
but these disorders may have significant
clinical overlap. While most cases
are idiopathic, these disorders
have several associations. TTP can
be familial, associated with HIV
infection, or associated with bone
marrow transplants or medications
(such as cyclosporine and ticlopidine).
There is an association of HUS with
the Shiga toxin produced by E. coli
0157:H7. The classic pentad associated
with TTP is microangiopathic hemolytic
anemia, thrombocytopenia, fever,
mental status changes and renal
dysfunction. This patient lacked
significant fever, but it is important
to know that approximately 50% of
patients with TTP will not have
all of the five signs of the pentad
of TTP. Other subtle features of
TTP include the fact that the white
cell blood cell count is typically
normal, liver and lung functions
are typically normal, and the mental
status involvement is not chronic,
as patients may rapidly return to
normal CNS function after being
essentially comatose.
Given
this patient's consolation of clinical
findings, the most likely diagnosis
is TTP. The treatment of TTP involves
plasmapheresis, which has been thought
to work by removing an abnormal
factor which promotes coagulation
from the blood and/or restoring
an absent factor to it. This patient
was treated for TTP with plasmapheresis,
but despite therapy expired suddenly.
An autopsy was performed.
The autopsy findings confirmed the
diagnosis of TTP. Examination of
multiple organs revealed small arterioles
with organizing platelet thrombi.
These were most evident in the heart,
where they were associated with
small foci of myocyte loss with
replacement fibrosis (Figure 1).
Higher power examination of these
thrombi revealed evidence of organization
with endothelial cell proliferation,
consistent with this patient's onset
of symptoms approximately 8-9 days
before demise (Figure 2).
Figure
1
Figure
2
A
recent autopsy series from this
hospital has compared and contrasted
the autopsy findings in a large
series of patients with TTP and
HUS. The study found numerous similar
clinical features between the two
types of patients, including similar
age, presence of purpura, anemia
and thrombocytopenia. However, several
key differences, which reached statistical
significance, were noted. As was
seen in the current case, patients
with TTP were far more likely to
have thrombi affecting the heart.
Thrombi in the brain (Figure 3)
were more common in TTP, which correlates
with the clinical presentation with
mental status changes. The pancreas
was more likely to be involved with
TTP, and thrombi classically affect
the small venules within islets
of Langerhans (Figure 4, Figure
5). The adrenal was also more frequently
involved in TTP, and the platelet
thrombi characteristically affect
the peripheral zona glomerulosa
of the adrenal cortex (Figure 6).
The platelet-rich composition of
the thrombi in TTP can be confirmed
by special stains. The thrombi are
strongly immunoreactive for von
Willebrand factor by immunohistochemistry
(Figure 7), but do not label well
with the PTAH stain, which is a
good marker for fibrin (Figure 8).
In contrast, the thrombi in HUS
are more likely to be true fibrin
thrombi, which label with the PTAH
stain (Figure 9). Clinically, HUS
was far more likely to be associated
with acute renal failure than TTP
(Figure 10).
Figure
3
Figure
4
Figure
5
Figure
6
Figure
7
Figure
8
Figure
9
Figure
10
Recently,
the pathogenesis of TTP has become
more apparent. It had long been
known that patients with TTP had
unusually large von Willebrand factor
multimers in their plasma, which
were postulated to promote abnormal
platelet aggregation and cause thrombi.
Non-familial acquired TTP has been
found to be associated with an acquired
inhibitory antibody against the
von Willebrand factor-cleaving metalloprotease,
which normally cleaves such multimers.
Familial forms of TTP are associated
with a constitutional deficiency
of this protease. In contrast, the
HUS is not associated with these
antibodies, and thus this single
laboratory test may help us distinguish
2 syndromes that overlap clinically.
It is hoped that this new knowledge
may lead to more rational therapies
for TTP in the future. One possibility
is the direct administration of
the inhibited protease (also known
as ADAMTS 13).
References
New
England Journal of
Medicine, 2002;
347:589-600.
New England Journal
of Medicine,
1998; 339:1585-1594.
New England Journal
of Medicine,
1998; 339:1578-1584.
Current Opinion
of Pediatrics
2000; 12:23-28.
Modern Pathology
2002; 15: 5A (abstract
#7).
|
|