Professional Documents
Culture Documents
466
The role and effectiveness of the
consultant
A hematology consultant provides expert advice to requesting
physicians and other health care providers about the diagno-
sis and management of benign or malignant hematologic
disorders. A consultation request might involve an adult
general medical patient, a child or adolescent, a pregnant
woman, a perioperative patient, or an individual who is crit-
ically ill. Within a health care community, the hematologist
might also be asked to serve as a specialty consultant on com-
mittees that maintain a formulary, develop clinical practice
guidelines, establish policies and procedures for transfusion
services, or monitor quality and efciency. Although these
latter roles will not be specically addressed in this chapter,
the data management, organizational, and communication
skills that are critical for patient consultation are also useful
when working within advisory groups.
Like the consultant for general medical conditions, a
clinical hematologist must understand the principles of effec-
tive consultation and the importance of interphysician com-
munication (Table 18-1). These skills may not have been
learned during training, but they are easily acquired and their
application ensures maximal compliance with recommenda-
tions and the highest quality of multidisciplinary patient care.
Consultation for surgery and invasive
procedures
Preprocedure assessment for bleeding risk
Avoiding hemorrhage is a prime concern during surgery,
invasive procedures, and spinal anesthesia. The risk of
hemorrhage is proportional to the depth and extent of the
intervention, the sensitivity of the involved tissue to blood
loss (ie, vascularity, potential for local brinolysis, and effects
of pressure from a hematoma), the ability to achieve surgical
hemostasis, and the possibility that elements of the proce-
dure might induce a hemostatic defect (eg, CPB). One must
also consider whether bleeding could occur because of an
associated anatomic defect (eg, invasive tumor or vascular
malformation) and whether hemorrhage could lead to severe
or permanent functional impairment (particularly within
the central nervous system [CNS], spinal canal, or eye).
Preoperative assessment should include a detailed personal
and family history, physical examination, clear understand-
ing of the planned procedure, and knowledge of the relevant
tests. Abnormalities on preoperative hemostatic testing
Consultative hematology
Keith R. McCrae, Dana C. Matthews, and Michael Linenberger
The role and effectiveness of the
consultant, 466
Consultation for surgery and invasive
procedures, 466
Intensive care unit, 473
Consultation for hematologic
complications of solid-organ
transplantation, 477
Obstetrics and gynecology, 480
Selected outpatient hematology topics, 489
Consultation in pediatric patients, 494
Bibliography, 499
CHAPTER
18
You are consulted to manage the anticoagulation therapy
for a 55-year-old man who is scheduled for cardiac surgery.
He suffered an extensive right lower extremity deep venous
thrombosis (DVT) 4 weeks ago and is currently on therapeutic
warfarin and aspirin, 81 mg/d. You confer with the surgeon
and learn that aortic valve replacement and coronary
revascularization are planned, with cardiopulmonary bypass
(CPB) support intraoperatively. You advise the surgeon to
discontinue aspirin at least 3 days preoperatively because of
the increased risk of platelet-related bleeding with CPB. You
also outline a strategy of bridging anticoagulation with
low-molecular-weight heparin (LMWH) preoperatively and
unfractionated heparin (UFH) for 57 days postoperatively
because of the high risk of recurrent thromboembolism in a
patient with a recent deep venous event.
Clinical case
Consultation for surgery and invasive procedures
| 467
screening tools for procedure-induced bleeding, so they do
not perform well in that setting. For example, a false- negative
APTT result might fail to identify a patient at risk with mild
von Willebrand disease (vWD), particularly if that patients
history was uninformative because he or she had never been
subjected to a prior hemostatic stress. Alternatively, the
APTT may yield a false-positive result that is either inconse-
quential or due to a lupus anticoagulant.
The hematologist is most likely to be consulted about
patients who have a suspicious history of bleeding, abnormal
preprocedure hemostatic screening results, a preexisting char-
acterized bleeding disorder, and/or exposure to drugs that
affect hemostasis. The approach to a preoperative patient with
either a suggestive history or abnormal screening tests should
mimic the approach to the bleeding patient (see Chapter 16).
This begins with completing the historical, clinical, and labo-
ratory database and determining whether a suspected defect
involves primary hemostasis (eg, with a history of immediate
and/or mucocutaneous bleeding) or secondary hemostasis
(eg, with a history of delayed and/or deep tissue bleeding).
The risks for patients with known hemostatic disorders
must be determined based on the nature of the defect and
the specic procedure that is planned. Prolonged, complex,
and deeply invasive procedures present greater risks than
brief, supercial interventions. For high-risk procedures,
preoperative factor replacement therapy is indicated to cor-
rect the PT international normalized ratio (INR) to 1.5,
the APTT to 1.5-times control (in the case of hemophilia,
a factor level of 80100%), and the brinogen to 1 g/L. For
low- to moderate-risk procedures, a PT/INR of 2 is usually
and bleeding postprocedure are most strongly predicted by a
history of abnormal bleeding from any source (skin, mucous
membranes, gastrointestinal tract, menses, and childbirth);
family members with bleeding disorders; use of antiplatelet,
anticoagulant, or other medications that could affect hemo-
stasis (eg, chronic antibiotics); prior bleeding associated with
hemostatic challenges (surgery, trauma, and dental proce-
dures); and/or known medical comorbidities associated with
bleeding diathesis (eg, renal failure, hepatic failure, moderate
to severe thrombocytopenia). By comparison, patients with
no suspicious history and no obvious comorbidities rarely
have abnormal screening tests. When abnormal hemostatic
results are found in such patients, they do not correlate with
procedure-related bleeding or the need for postprocedure
blood transfusion. This applies even to certain high-risk pro-
cedures, including liver biopsy, kidney biopsy, prostate sur-
gery, coronary artery bypass surgery, tonsillectomy, and
gynecologic surgery, where screening hemostatic test results
have low sensitivity and poor positive predictive value.
Despite the lack of evidence to support laboratory screen-
ing in asymptomatic patients, a limited hemostatic evalua-
tion is recommended for patients undergoing moderate- or
high-risk procedures that involve vital organs, deep or exten-
sive dissection, and/or a likelihood of bleeding 500 mL.
These tests include the prothrombin time (PT), activated
partial thromboplastin time (APTT), and platelet count.
Tests for platelet function, including the bleeding time and
platelet function analyzer (PFA-100), do not predict the
bleeding risk of invasive procedures. It is important to
remember that the coagulation tests were not developed as
Table 18-1 Principles of effective consultation and inter-physician communication.*
Principle Comment
Determine the question that is being asked The consultant must clearly understand the reason for the consultation
Establish the urgency of the consultation and
respond in a timely manner
Urgent consultations must be seen as soon as possible (communicate any expected delays
promptly); elective consultations should be seen within 24 hours
Gather primary data Conrm the database personally; do not rely on secondhand information
Communicate as briey as appropriate Compliance is optimized when the consultant addresses specic questions with 5 succinct
and relevant recommendations
Make specic recommendations Identify major issues; limit the diagnostic recommendations to those most crucial; and provide
specic drug doses, schedules, and treatment guidelines
Provide contingency plans Address alternative diagnoses briey; anticipate complications and questions
Understand the consultants role The attending physician has primary/ultimate responsibility; the consultant should not assume
primary care or write orders without permission from the attending
Offer educational information Provide relevant evidence-based literature and/or guidelines
Communicate recommendations directly to
the requesting physician
Direct verbal contact (in person or by phone) optimizes compliance and minimizes confusion
or error
Provide appropriate follow-up Continue involvement and progress notes as indicated; ofcially sign off the case or provide
outpatient follow-up
*Adapted from Goldman L, Lee T, Rudd P. Ten commandments for effective consultations. Arch Int Med. 1983;143:17535 and Sears CL,
Charlson ME. The effectiveness of a consultation. Compliance with initial recommendations. Am J Med. 1983;74:8706.
|
Consultative hematology 468
adequate. Patients with thrombocytopenia should achieve a
platelet count of 50,000/L before moderate- to high-risk
procedures; however, a platelet count of 100,000/L is
desired before neurosurgical interventions. Patients with
immune thrombocytopenia often tolerate surgery well even
with platelet counts 50,000/L. Evidence to support the
use of platelet transfusions in this setting is not available.
Patients with normal baseline hemostasis who are taking
antiplatelet or anticoagulant drugs must be evaluated in the
context of the indications for the antithrombotic therapy
and the bleeding risks associated with the specic procedure.
For example, consumption of aspirin or clopidogrel within
the week prior to coronary artery bypass graft surgery is
associated with signicant increases in perioperative bleed-
ing, transfusion requirement, and reoperation. Thus, aspirin
and other antiplatelet agents should be suspended at least
3 days and perhaps as much as 7 days before elective surgery
when platelet dysfunction is induced by CPB. Patients who
require emergent surgery with CPB while on antiplatelet
agents may require platelet transfusions to minimize periop-
erative bleeding. A similar approach applies to patients
receiving abciximab or other potent, specic inhibitors of
platelet glycoprotein (GP) IIb/IIIa. However, because the
GPIIb/IIIa antagonists have a short half-life, delaying surgery
for 12 hours after stopping the drug may be a safe alternative
approach.
Patients who undergo elective surgery without CPB do
not need to discontinue aspirin or clopidogrel preoperatively.
Aspirin can also be safely continued in outpatients undergo-
ing cataract surgery, simple dental extractions, and certain
dermatologic procedures. However, aspirin, clopidogrel, and
dipyridamole are usually discontinued prior to invasive
biopsies, arthroscopies, traumatic dental surgeries, and out-
patient procedures with a high risk of signicant bleeding.
For patients on chronic therapeutic warfarin, management
around invasive procedures should be guided by evidence-
based recommendations for either temporary interruption or
bridging anticoagulation (Table 18-2). Patients on LMWH
should stop therapy 612 hours before surgery or epidural
anesthesia and not resume it until 2 hours after the proce-
dure. Because dosage is difcult to calculate, protamine sulfate
is not generally used to reverse LMWH. UFH should be
stopped 6 hours before surgery. If surgery must take place in
4 hours, UFH can be reversed with protamine sulfate.
Perioperative hemorrhage, hemostatic agents,
and transfusion
Intraoperative hemorrhage
Intraoperative hemorrhage may be due to inadequate local
hemostasis and/or a systemic coagulation defect. Local
hemostasis may be impaired because of friable vessels or
Table 18-2 Managing oral anticoagulation during invasive procedures.*
Clinical situation Guidelines
Low risk for thromboembolism Discontinue warfarin therapy about 4 days before surgery, and allow the INR to return to
near-normal levels; if intervention increases a risk for thrombosis, begin short-term low-dose
heparin therapy (5000 IU SC) or prophylactic-dose LMWH and resume warfarin therapy
Intermediate risk for thromboembolism Discontinue warfarin therapy about 4 days before surgery, and allow the INR to fall; about
2 days preoperatively, give either low-dose heparin (5000 IU SC) or a prophylactic dose of
LMWH; postoperatively, give low-dose heparin (or LMWH) and resume warfarin therapy
High risk for thromboembolism Discontinue warfarin therapy about 4 days before surgery, and allow the INR to return to normal
range; about 2 days preoperatively, as the INR falls, give full-dose heparin or full-dose LMWH
Heparin can be given SC on an outpatient basis and as a continuous IV infusion after hospital
admission, and discontinued 6 hours before surgery; alternatively, continue SC heparin or
LMWH therapy until 12 to 24 hours before surgery
Low risk for bleeding Lower the warfarin dose 4 or 5 days before surgery to reach an INR of 1.3 to 1.5 at time of
surgery; resume warfarin therapy postoperatively and supplement, if necessary, with low-dose
heparin (5000 IU SC) or prophylactic LMWH
Dental procedures Patients at high risk for bleeding: discontinue warfarin therapy
Patients not at high risk for bleeding: warfarin therapy should not be discontinued
Dental procedures, with need to control
local bleeding
Patients at high risk for VTE: continue warfarin and administer a mouthwash of -aminocaproic
acid
*Adapted from: Ansell J, Hirsh J, Poller L, et al. The pharmacology and management of the vitamin K antagonists: the Seventh ACCP Conference
on Antithrombotic and Thrombolytic Therapy. Chest. 2004;126(suppl):S20433.
No VTE for 3 months; atrial brillation without history of stroke or other risk factors; bileaet mechanical cardiac valve in the aortic position.
VTE in 3 months; mechanical cardiac valve in the mitral position; old-model cardiac valve (ball/cage).
INR international normalized ratio; LMWH low-molecular-weight heparin; SC subcutaneous; VTE venous thromboembolism.
Consultation for surgery and invasive procedures
| 469
compromised tissue integrity related to nutritional deciency,
prior radiation therapy, necrosis, malignant degeneration,
and chronic illness. Preexisting or acquired coagulation
defects should be considered if the patient has a history of a
hemorrhagic diathesis, bleeding from previous surgery, or
bleeding/oozing from multiple sites, including outside of the
operative eld. The consultant may be called to assist the sur-
geon in making a rapid decision regarding the possibility of a
coagulation defect and advisability of reexploring the wound.
Reexploration is appropriate when hemorrhage occurs from
a single surgical site and a rapid assessment of PT/INR, APTT,
platelet count, and brinogen show no evidence of signicant
abnormalities. Potential causes of intraoperative coagulo-
pathy and generalized bleeding include unrecognized preex-
isting defects (eg, vWD, mild hemophilia, factor XI deciency,
thrombocytopathy), hypothermia, hyperbrinolysis, drugs,
uremia, disseminated intravascular coagulopathy (DIC), and
postransfusion (dilutional) coagulopathy.
Hemorrhage occurs in 625% of patients undergoing CPB
surgery. Factors that contribute to hemorrhage include pre-
operative medications (eg, antiplatelet agents, heparin),
wound factors, acidosis, hypothermia, exposure of blood in
the extracorporeal oxygenator with platelet injury, loss of
coagulation factors, brinolysis, and postoperative heparin
therapy. Liver transplantation carries specic risks due to
temporary cessation of coagulation factor synthesis and
enhanced brinolysis. During reperfusion, tissue-type plas-
minogen activator (tPA) is released into the circulation and
proteolysis of vWF occurs. In addition to serial measure-
ments of brinogen, thromboelastography can provide a
real-time evaluation of brinolysis during and after the sur-
gical procedure. Factor replacement, antibrinolytic therapy,
and, in selected patients with active bleeding, recombinant
factor VIIa (rFVIIa) may be used on an empiric basis. Of
note, 2 recent randomized, placebo- controlled trials showed
that single- or multiple-dose prophylactic rFVIIa does not
decrease the number of perioperative red cell transfusions
required by patients undergoing orthotopic liver transplan-
tation for end-stage liver disease. Fibrin sealant has been
used on the liver surface. Finally, neurosurgery involving the
brain carries an increased risk of DIC.
Hemostatic agents
Intraoperative hemorrhage can be prevented or controlled
under certain circumstances by the use of hemostatic agents.
Such agents include conjugated estrogens, desmopressin
(DDAVP), lysine analogues, aprotinin, rFVIIa, and brin
sealant. Oral or intravenous conjugated estrogens, given for
5 to 7 days preoperatively, may decrease platelet-related
bleeding in patients with chronic renal failure. Platelet dys-
function and primary hemostatic defects due to uremia,
myeloproliferative disorders, type I vWD, and congenital
thrombocytopathy can also improve after intravenous or
intranasal administration of DDAVP. Nonspecic prophy-
lactic use of DDAVP is potentially dangerous, particularly in
patients with coronary artery disease. A recent metaanalysis
of 17 randomized, double-blind, placebo- controlled trials
showed no signicant impact of DDAVP on transfusion
requirements among patients undergoing CPB, whereas
treated patients suffered a 2.4-fold increased incidence of
perioperative myocardial infarction (MI).
The lysine analogues, -aminocaproic acid and tranexamic
acid, inhibit plasmin activation and prevent lysis of clots.
These agents are useful in coagulation-factor-decient hem-
orrhage involving mucous membranes where local brinoly-
sis adds to the hemostatic impairment (eg, dental procedures,
epistaxis, and menorrhagia). Lysine analogues have also been
shown to reduce postoperative bleeding after cardiac surgery,
liver transplantation, and prostatectomy. A recent observa-
tional study of 4374 patients undergoing coronary revascu-
larization surgery with CPB showed that -aminocaproic acid
and tranexamic acid were as effective as aprotinin in signi-
cantly reducing surgical blood loss, and neither agent was
associated with an increased risk of thromboembolism or
other complications. The efcacy and safety of - aminocaproic
acid in this trial concurs with prior metaanalyses that revealed
a 30% to 40% reduction in bleeding after cardiac surgery.
Tranexamic acid is no longer commercially available in the
United States; therefore, -aminocaproic acid is the hemo-
static agent of choice for routine use in cardiac surgery
with CPB.
Aprotinin is a bovine lung-derived serine protease inhibi-
tor of plasmin, kallikrein, and brinolysis. It has been used
most extensively for prophylaxis of intraoperative and
postoperative hemorrhage in patients undergoing coronary
artery bypass graft surgery with CPB. This agent also signi-
cantly reduces bleeding associated with liver transplantation,
noncardiac thoracic surgeries, and orthopedic surgeries. Pre-
clinical data and early clinical experience with aprotinin
raised concerns about potential nephrotoxicity and intra-
vascular thrombosis. In 2006, a large observational study
revealed that aprotinin use in patients undergoing coronary
revascularization and CPB was associated with a 1.5- to
2.5-fold increased risk of renal failure, MI, heart failure,
stroke, and encephalopathy, compared with rates in control
patients and among those who received -aminocaproic acid
or tranexamic acid (Mangano et al., 2006). In addition, a
propensity-score, case-control comparison of aprotinin and
tranexamic acid in high-transfusion-risk cardiac surgery
patients showed that aprotinin may be associated with wors-
ening renal function among patients with preexisting renal
dysfunction (Karkouti et al., 2006). End-organ complica-
tions related to aprotinin are believed to be due to ischemic
|
Consultative hematology 470
microthrombotic tissue injury, particularly in the renal
microvasculature. These data raise new concerns about the
safety of aprotinin as well as signicant questions about the
risks of off-label uses of it and suggest that lysine analogues
should be the hemostatic agents of choice for bleeding pro-
phylaxis in coronary artery surgery patients.
rFVIIa is approved for the treatment of bleeding in hemo-
philia A or B patients with acquired inhibitors. Off-label
indications, however, account for over 90% of rFVIIa usage
in US academic medical centers. Most of these indications
are for the prevention and treatment of bleeding associated
with invasive procedures and surgeries in patients with
hepatic or renal insufciency and coagulopathy. Small non-
randomized studies have previously shown that rFVIIa can
reduce bleeding and/or transfusion requirements in patients
undergoing retropubic prostatectomy, noncoronary cardiac
surgery with CPB, and orthotopic liver transplantation.
However, the benet in these settings appears to be small.
More recently, 2 randomized, double-blind, placebo-
controlled trials observed no benet of either a single preop-
erative dose of rFVIIa or multiple perioperative doses with
regard to the amount of perioperative red blood cell (RBC)
transfusion support required by patients undergoing ortho-
topic liver transplantation.
A multicenter assessment of rFVIIa usage for all off-label
indications revealed that a single dose can effectively correct
the INR in over 80% of cases and that hemostasis can be
achieved in roughly 50% of bleeding patients within 6 hours
of treatment. However, rebleeding and hemorrhagic deaths
still occur in roughly one fourth and one third of cases,
respectively. Thromboembolism, line clot, MI, ischemic
stroke, and other potential rFVIIa-associated adverse events
(ie, occurring within 24 hours of the dose) occurred in 9.8%
of patients. A recent survey of the Food and Drug Adminis-
trations (FDAs) Adverse Event Reporting System (AERS)
revealed that roughly two thirds of the thromboembolic
events with rFVIIa from 1999 through 2004 occurred among
patients receiving the drug for off-label indications, most
commonly for surgical bleeding or prophylaxis. Arterial
and venous events, which were equally prevalent, resulted
in signicant morbidity. In addition, thromboembolic
complications were the probable cause of death in roughly
three fourths of the case fatalities. A prospective randomized
trial that compared rFVIIa with placebo for nonsurgical
intracranial hemorrhage observed a 3-fold incidence of
serious thrombotic events (predominantly MI and cerebral
infarction) among the predominantly elderly patients receiv-
ing rFVIIa (7% versus 2%; P NS). Despite the lack of
statistical signicance in this trial, these observations com-
pelled the FDA to issue a warning regarding off-label use of
rFVIIa and thrombotic risk. At this point, the riskbenet
ratio and cost-effectiveness of this very expensive therapy for
high-risk surgical patients remain to be determined by well-
designed controlled trials. For surgical patients who poten-
tially might benet from off-label use, the consultant must
consider patient age, history of atherosclerotic disease, other
medical comorbidities and concurrent medications that
might contribute to a thromboembolic complication.
Fibrin sealant (eg, Tisseel, Hemaseel), also known as brin
glue, is composed of puried, virally inactivated human
brinogen and human thrombin. Tisseel also contains
aprotinin, which may be undesirable in certain settings
(eg, middle ear surgery). Although randomized clinical trial
data and evidence-based guidelines are lacking, brin sealant
has proven effective in cardiac surgery, urologic procedures,
orthopedic surgery, dental procedures, trauma, and neuro-
surgery, where it is used to seal dural leaks and repair otic
ossicles and bony defects. It is useful for achieving local
hemostasis in patients with factor VIII deciency or congen-
ital hemostatic defects.
Transfusion
The collection and transfusion of autologous blood in the
perioperative setting was most popular and widespread in
the early 1990s when concerns about the medical and legal
risks of homologous transfusion were greatest. With the rou-
tine use of nucleic acid testing (NAT), a unit of homologous
blood currently conveys a risk of 1/1,000,000 to 2,000,000
of transmitting human immunodeciency virus (HIV) and
hepatitis C, and this risk will be even smaller if individual
donor NAT replaces minipool NAT methods. The risk of
transmission of hepatitis B with serologic screening is
currently 1/6000 to 1/150,000, and this risk will also decrease
substantially with future NAT. Although preoperative
autologous blood donation (PABD) can eliminate the risks
of transmitting these viral infections, autologous donors are
still susceptible to other transfusion-related hazards, includ-
ing complications from the donation itself. Autologous
donors receive less homologous blood perioperatively, but
they are more likely to be transfused with their own blood.
They are therefore subject to clerical errors, bacterial con-
tamination, and complications related to blood processing.
Additional disadvantages of PABD include a higher likeli-
hood of perioperative anemia (because of the autologous
donation) and a 6/100,000 risk that the autologous collection
will incur a severe adverse reaction requiring hospital admis-
sion, particularly if this is the donors rst experience. It is
recommended that PABD not be used if the likelihood of
perioperative transfusion is 10%.
An effective alternative to PABD is acute normovolemic
hemodilution (ANH). ANH is a procedure in which blood
is removed from the patient immediately prior to surgery,
replaced with saline, and returned to the patient when surgery
Consultation for surgery and invasive procedures
| 471
is completed. ANH can be considered when the anticipated
blood loss is 1500 mL or more; the preoperative hemoglobin
will be at least 12 g/dL after correction of normovolemia; and
the patient has normal cardiovascular function and no major
medical comorbidities, including restrictive or obstructive
lung disease, infection, uncontrolled hypertension, and sig-
nicant renal, liver, or coagulation abnormalities. Oral (not
parenteral) iron supplementation is advised only to raise low
iron levels or to replace iron removed in PABD. In 2 random-
ized trials, ANH was equivalent to PABD in reducing exposure
to homologous blood. Either PABD or ANH can be combined
with recombinant erythropoietin to enhance erythropoiesis.
The addition of preoperative recombinant erythropoietin to
ANH results in the highest postoperative hemoglobin levels,
but is most costly. Studies in orthopedic patients showed that
preoperative recombinant erythropoietin decreased, but did
not eliminate, homologous transfusion.
Predetermined hemoglobin triggers should not be used
in perioperative transfusion management. The decision to
transfuse should be based on individualized assessment and
evidence-based guidelines. A number of studies have demon-
strated that inappropriate perioperative red cell and/or plate-
let transfusions correlate with increased risks of cardiac, renal,
pulmonary, and infectious complications, in addition to a
higher mortality. Controlled trials have shown no improve-
ment in outcome from maintaining a critically ill patients
hemoglobin above an arbitrary value (see the Intensive care
unit section later in the chapter).
Postoperative bleeding
Surgical reexploration is urgently indicated in the presence
of bright red bleeding in a normothermic, nonacidotic post-
operative patient with hemostatically adequate values of PT,
APTT, and platelet count. In this case, a technical or vascular
etiology is most likely. Alternatively, a postoperative patient
with persistent oozing or low-volume bleeding could have a
mild, undiagnosed hemostatic defect, such as vWD or mild
hemophilia, despite a normal APTT (factor VIII and IX levels
as low as 15% can be associated with a normal APTT). In
such patients, activity levels of factor VIII, vWF, and factor
IX are indicated. Older men with mild factor VIII or IX de-
ciency and no history of pathologic bleeding may develop
severe hemorrhage after dental extraction or prostate surgery.
Congenital factor XI deciency would be another consider-
ation, although the baseline APTT is usually prolonged in
clinically affected patients. Isolated prolongation of the PT/
INR suggests vitamin K deciency (eg, from malabsorption
or antibiotics), undiagnosed liver disease, or warfarin inges-
tion. Multiple hemostatic abnormalities and thrombocyto-
penia should prompt consideration of DIC and dilutional
coagulopathy due to hypertransfusion.
DDAVP, cryoprecipitate, or factor concentrate can be used
to correct various degrees of factor VIII deciency. Fresh fro-
zen plasma or factor IX concentrates are required to control
the bleeding in factor IX deciency. Hemophilia-related
bleeding from mucous membranes may also improve with
the use of -aminocaproic acid. Transfusion of plasma, cryo-
precipitate, and/or platelets is indicated to correct the defects
and stop the bleeding with vitamin K deciency, thrombocyto-
penia, and DIC. -Aminocaproic acid and, for highly selected
patients, rFVIIa could be considered with persistent or
uncontrolled bleeding.
Thromboprophylaxis and postoperative
thrombosis
Prophylaxis for thromboembolism
Postoperative thromboembolism complicates a variety of
procedures, particularly large joint surgery or joint replace-
ment, thoracic or abdominal procedures lasting 30 min-
utes under general anesthesia, cancer surgery, major trauma,
and spinal cord injury. The risks are increased in patients
who are immobilized and those with active cancer, a past his-
tory of venous thromboembolism (VTE), hypercoagulable
conditions, or recent VTE. Without postoperative anticoagu-
lation, 50% of patients who have surgery within a month of
a DVT will recur and 6% of those events will be fatal. Antico-
agulation can reduce the risk of recurrence by 90%, but 3%
of patients have severe postoperative hemorrhage (with the
possibility of permanent disability) and 3% of patients
(0.09% overall) will die of bleeding. Despite these concerns,
the risk of VTE-related fatality is much greater without anti-
coagulation (6%) than with it (1%).
The risk of VTE after surgery can be estimated based on the
type of procedure, the age of the patient, and presence of
additional medical comorbidities (Table 18-3). After the risk
of perioperative VTE is determined, a prophylactic strategy
can be recommended according to evidence-based guidelines.
Prophylactic measures include lower extremity intermittent
pneumatic compression devices or graduated compression
stockings, UFH, LMWH, fondaparinux, and oral anticoagu-
lation (Table 18-3).
The timing of the initiation of prophylaxis varies based on
the procedure and regional practice patterns. For hip replace-
ment surgery, fewer DVTs occur when LMWH is started
between 26 hours postoperatively, compared with starting
warfarin or LMWH 12 hours after surgery. For other proce-
dures, the timing of initiation of prophylaxis does not signi-
cantly correlate with outcomes. In Europe, LMWH therapy is
usually started at half doses 12 hours before surgery, whereas
in the United States it is common to start full doses (eg, 30 mg
of enoxaparin every 12 hours) at 1224 hours after surgery.
|
Consultative hematology 472
The bleeding rates are low, 0.9% and 3.5%, respectively. Pro-
phylactic warfarin begun just before or immediately after sur-
gery is less commonly associated with hemorrhage into the
replaced knee but is also less effective in preventing DVT. A
metaanalysis showed that hip surgery is more commonly
associated with symptomatic DVT than is knee surgery.
Another metaanalysis showed that hip replacement under
spinal anesthesia is less often followed by DVT than when it is
performed under general anesthesia. Thromboprophylaxis is
generally continued for 710 days postoperatively. Extended
therapy, for up to 2835 days, should be considered for hip
replacement or hip fracture surgery, cancer surgery involving
the chest or abdomen, or postoperative patients with spinal
cord injury.
The pentasaccharide fondaparinux is an acceptable
alternative to UFH or LMWH and should be used in patients
with a history of heparin-induced thrombocytopenia (HIT).
Recent large randomized clinical trials have indicated that
prophylactic fondaparinux (started 48 hours after surgery)
may be more effective than LMWH (ie, dalteparin or enoxa-
parin) for certain high-risk patients, including those under-
going abdominal cancer surgery and major orthopedic
procedures (knee or hip arthroplasty or hip fracture sur-
gery). Similar to LMWH, extended duration fondaparinux
(for 1923 days) after hip fracture surgery further reduces
the risk of VTE (by 96%), compared with therapy for only
68 days, without a signicant increase in bleeding.
Additional considerations for thromboprophylaxis in sur-
gical patients include the potential need for dose adjustments
or avoidance of LMWH and fondaparinux in patients with
renal failure and measures to avoid perispinal hematoma in
patients receiving neuraxial anesthesia/analgesia (Table 18-4).
Treatment of postoperative deep venous thrombosis
When VTE occurs in a postoperative patient, the consultant
may be asked to prescribe treatment recommendations. For
most low-risk procedures, full anticoagulation can be safely
initiated within 1224 hours after surgery. The agent of choice
in the immediate postoperative period is continuous-infusion
UFH because of the short half-life and ability to reverse rapidly
with protamine if bleeding develops. Relative contraindica-
tions to immediate postoperative anticoagulation include
certain neurosurgical or ophthalmologic procedures in which
bleeding would risk permanent disability or blindness. Abso-
lute contraindications to anticoagulation require placement
of a temporary or permanent inferior vena cava lter.
By the third postoperative day after uncomplicated surgery,
the risks of anticoagulation are comparable to those in a non-
surgical patient. Initial therapy in such patients could include
Table 18-3 VTE risk stratication for surgical patients and evidence-based guidelines for prophylactic anticoagulation.*
Level of risk Denition VTE risk (%) Recommended prophylaxis
Low risk Minor surgery in patients 40 years old
with no additional risk factors
Calf DVT: 2
Proximal DVT: 0.4
Clinical PE: 0.2
Fatal PE: 0.01
Early and aggressive ambulation
Moderate risk Minor surgery with additional risk factors or
Surgery at 4060 years old with no additional
risk factors or
Major surgery in 40 years old with no
additional risk factors
Calf DVT: 1020
Proximal DVT: 24
Clinical PE: 12
Fatal PE: 0.10.4
LDUH (every 12 hours) or LMWH, or GCS or IPC
device (if bleeding risk)
High risk Surgery at 60 years old or
Surgery at 4060 years old with additional risk
factors
Calf DVT: 2040
Proximal DVT: 48
Clinical PE: 24
Fatal PE: 0.41
LDUH (every 8 hours) or LMWH or IPC device
(if bleeding risk)
Highest risk Major surgery at 40 years old with additional
risk factors or
Hip or knee arthroplasty or
Hip fracture surgery or
Major trauma or
Spinal cord injury
Calf DVT: 4080
Proximal DVT: 1020
Clinical PE: 410
Fatal PE: 0.25
LMWH or fondaparinux or oral anticoagulation
(INR 23) or GCS/IPC device LDUH (every 8
hours) or GCS/IPC device LMWH
Consider extended prophylaxis (ie, 710 days) for
cancer surgery, hip arthroplasty, hip fracture
surgery, or spinal cord injury
*Adapted from Geerts WH, Pineo GF, Heit JA, et al. Prevention of venous thromboembolism: the Seventh ACCP Conference on Antithrombotic
and Thrombolytic Therapy. Chest. 2004;126(suppl):S338400.
Additional risk factors: prior VTE, obesity, heart failure, cancer, hypercoagulable state.
DVT deep venous thrombosis; GCS graduated compression stocking; IPC intermittent pneumatic compression; LDUH low dose
unfractionated heparin; LMWH low-molecular-weight heparin; PE pulmonary embolism; VTE venous thromboembolism.
LMWH and warfarin as prescribed in standard doses for non-
surgical patients with DVT. The patients prior experience
with therapy for thrombosis may inuence the recommenda-
tion. Postphlebitic syndrome is a distressing and morbid con-
dition. To prevent it, thrombolytic therapy may be considered
for carefully selected postsurgical patients with massive ilio-
femoral DVT. The consultant should bear in mind that hem-
orrhagic stroke, active intracranial neoplasm, intracranial
surgery or trauma within the previous 2 months, and active
recent or internal bleeding are absolute contraindications
to thrombolytic therapy. Relative contraindications include
a bleeding diathesis, uncontrolled severe hypertension, cardio-
pulmonary resuscitation, nonhemorrhagic stroke within
2 months, surgery within 10 days, and thrombocytopenia. It is
also worth noting that clinical trials of thrombolytic agents
have shown marginal improvement in survival and long-term
morbidity despite the immediate discernable benets of
thrombolysis. The use of elastic compression stockings (pres-
sure of 30 to 40 mm Hg) during 2 years after a DVT has been
shown to decrease the incidence and severity of postphlebitic
syndrome. The duration of anticoagulation with therapeutic
warfarin after an uncomplicated postoperative VTE is usually 3
months; however, longer treatment would be indicated in the
setting of an ongoing hypercoagulable condition (eg, antiphos-
pholipid antibody syndrome or active cancer), history of
multiple thromboembolic episodes, or persistent venous
obstruction with severe postphlebitic syndrome.
Table 18-4 Recommendations to minimize the risk of perispinal
hematoma in surgical patients undergoing neuraxial anesthesia/
analgesia.*
Avoid neuraxial anesthesia/analgesia in patients who have bleeding
disorders
Avoid neuraxial anesthesia/analgesia in patients who have drug-related
preoperative impairment of hemostasis
Delay catheter insertion/spinal needle for 812 hours after a
subcutaneous dose of heparin or a twice-daily dose of LMWH
Delay catheter insertion/spinal needle for 18 hours after a once-daily
injection of LMWH
Delay anticoagulant prophylaxis if a hemorrhagic aspirate (ie, bloody
tap) is seen during initial insertion
Remove epidural catheter when the anticoagulant effect is at a
minimum (ie, just before next scheduled dose)
Delay anticoagulant prophylaxis at least 2 hours after catheter is
removed
With warfarin prophylaxis, continuous epidural analgesia should be
limited to 12 days and the PT/INR should be 1.5 at the time of
catheter removal
The safe usage of fondaparinux prophylaxis with continuous epidural
anesthesia is under investigation
*Adapted from Horlocker TT, Wedel DJ, Benzon H, et al. Regional
anesthesia in the anticoagulated patient: dening the risks (the
second ASRA Consensus Conference on Neuraxial Anesthesia and
Anticoagulation). Reg Anesth Pain Med. 2003;28:17297 and Geerts
WH, Pineo GF, Heit JA, et al. Prevention of venous thromboembolism:
the Seventh ACCP Conference on Antithrombotic and Thrombolytic
Therapy. Chest. 2004;126(suppl):S338400.
INR international normalized ratio; LMWH low-molecular-weight
heparin; PT prothrombin time.
Preoperative hemostatic screening test abnormalities and
postoperative bleeding risk are most strongly associated with a
personal or family history of bleeding, use of antiplatelet or
anticoagulant agents, prior procedure-related bleeding, and
known medical comorbidities associated with bleeding
diathesis.
Intraoperative hemorrhage may be due to inadequate local/
surgical hemostasis and/or a preexisting or acquired systemic
coagulation defect.
Hemostatic agents such as conjugated estrogens,
desmopressin (DDAVP), lysine analogues, aprotinin, rFVIIa, and
brin sealant may be useful for selected surgical patients with
an increased risk of bleeding; however, potential adverse effects
and cost must also be considered.
PABD and ANH, with or without adjunctive recombinant
erythropoietin support, may be considered as alternatives to
homologous blood transfusion for selected surgical patients.
The risk of VTE after surgery and the appropriate
anticoagulant prophylaxis strategy should be determined
according to the type of procedure, the age of the patient, and
presence of additional medical comorbidities.
Management of acute VTE in a postoperative patient is similar
to the approach in a nonsurgical patient, except that bleeding
risks and contraindications to anticoagulation and thrombolysis
must be carefully considered in the context of the procedure.
Key points
Intensive care unit
Anemia
Many patients hospitalized in intensive care units (ICUs) for
acute medical illnesses have anemia at the time of admission,
and most of those who do not will develop it. The patho-
physiology of anemia in acutely ill patients is multifactorial,
reecting both blunted erythropoietin production as well as a
blunted response to erythropoietin. Anemia is exacerbated by
bleeding and frequent phlebotomy. A clinical trial examining
the role of transfusion in 838 critically ill patients with hemo-
globin values 9.0 g/dL randomized patients to a liberal
transfusion strategy, in which the hemoglobin was main-
tained between 10 and 12 g/dL, or a restrictive group, in which
the hemoglobin was maintained between 7 and 9.9 g/dL.
In-hospital mortality was signicantly lower in the restricted
group, though no differences were observed in 30-day
Intensive care unit
| 473
|
Consultative hematology 474
mortality. This study suggests that acutely ill patients should
not be transfused based on the hemoglobin value alone and
that other factors should be considered when weighing the
potential benets of transfusion. Despite this report, recent
surveys suggest that physicians who care for patients in the
intensive care setting continue to transfuse at hemoglobin lev-
els of approximately 9.0 g/dL. Patients on mechanical ventila-
tion receive approximately 75% of all red cell transfusions.
Recent studies have demonstrated that the use of recombi-
nant erythropoietin in patients with acute illness who are
expected to remain in intensive care for a prolonged period
can diminish the use of transfusion by 1020%, though these
studies have not been adequately powered to assess effect on
survival. The cost effectiveness of erythropoietin in this
setting has also not been established.
Thrombocytopenia
Approximately 40% of patients in medical or surgical
ICUs develop a platelet count below 100,000/L; 1020%
develop severe thrombocytopenia, with a platelet count below
50,000/L. The spectrum of disorders that may cause throm-
bocytopenia in this setting is extensive. Some causes of
thrombocytopenia in this setting are summarized in Table
18-5, several of which are described in detail in this section.
Evaluation of the patient with thrombocytopenia should
include a detailed history, thorough physical examination,
and close inspection of the peripheral blood lm. A history of
chronic or prior thrombocytopenia and the setting in which
it occurred should be queried. Underlying marrow disease, as
well as preexisting morbidities that may induce chronic
thrombocytopenia such as liver disease, neoplasia, or immune
thrombocytopenic purpura (ITP), may be largely excluded by
the documentation of normal hematologic parameters prior
to the onset of the acute illness. Abnormalities on the periph-
eral blood lm that may aid in reaching a diagnosis include
(i) platelet clumping when blood is collected in EDTA-
containing tubes, suggesting pseudothrombocytopenia; (ii)
schistocytes, which may suggest an underlying thrombotic
Table 18-5 Diagnosis of thrombocytopenia in the intensive care unit.
Diagnosis\nding Clinical ndings Special LDH APTT
APL
antibody
Platelet
morphology
Red cell
morphology
Pseudo-
thrombocytopenia
None Platelet count normal
in citrate or heparin
Normal Normal Absent Clumps Normal
Drug-induced
thrombocytopenia
None, bleeding/petechiae
if severe
Higher incidence with
certain drugs
Normal Normal Absent Normal Normal
ITP None, bleeding/petechiae
if severe
Often associated with
prior history
Normal Normal Absent Normal or
mildly
enlarged
Normal
Heparin-induced
thrombocytopenia
May be associated with
thrombosis
514 days after starting
heparin unless
reexposure, may have
DIC as well
Normal Prolonged
due to
heparin
Absent Normal Normal
DIC Underlying illness usually
identiable
Increased D-dimer or
FDP, decreased
brinogen in some,
mildly prolonged PT,
some microangiopathic
hemolytic anemia
Normal
or mildly
increased
Usually
normal
Absent Normal Schistocytes
in some
Catastrophic aPL
syndrome
Infection, fever, at least 3
organs involved
Small vessel
microthrombi
Normal
or mildly
increased
Normal or
increased
Present Normal Normal
TTPHUS Thrombocytopenia and
microangiopathic
hemolytic anemia
without other apparent
cause; may also have
renal dysfunction,
neurologic changes,
fever
Microangiopathic
hemolytic anemia,
ADAMTS 5% may
help conrm diagnosis
but is not readily
available, is not
standardized, and must
be sent to specialized
labs
Increased Normal Absent Normal or
mildly
enlarged
Schistocytes
aPL antiphospolipid; APTT activated partial thromboplastin time; DIC disseminated intravascular coagulation; FDP brin
degradation product; ITP idiopathic thrombocytopenic purpura; LDH lactate dehydrogenase; PT prothrombin time; TTP-HUS =
thrombotic thrombocytopenic purpura-hemolytic uremic syndrome.
microangiopathy (TMA) such as thrombotic thrombocyto-
penic purpura (TTP), the hemolytic uremic syndrome (HUS),
or DIC; (iii) poikilocytes or nucleated RBCs, which may
reect a myelophthisic process; or (iv) abnormal leukocytes,
which may suggest a hematologic malignancy, myelodyspla-
sia, or a syndrome of congenital thrombocytopenia. Bone
marrow aspiration and biopsy should be pursued when an
underlying marrow disorder is suspected.
Drug-induced thrombocytopenia
Acutely ill patients are commonly treated with several drugs,
many of which may be capable of inducing thrombocytopenia
through a variety of mechanisms. A partial list of offending
agents includes heparin, trimethoprim/sulfamethoxazole,
beta-lactam antibiotics, vancomycin, cephalothin, carbam-
azepine, hydrochlorothiazide, nonsteroidal antiinamma-
tory drugs, phenytoin, procainamide, quinidine and quinine,
rifampin, sulfasalazine, sulfonylureas, and valproic acid. An
expanded list of drugs implicated in drug-induced thrombo-
cytopenia is maintained at http://moon.ouhsc.edu/jgeorge.
The diagnosis of drug-induced thrombocytopenia is a
diagnosis of exclusion. A clear understanding of the temporal
relationship between the institution of the drug and the onset
of thrombocytopenia is essential. The decision of whether to
discontinue the administration of a potentially offending
drug depends on the level of suspicion that this is the offend-
ing agent and the relative importance of that drug to the
patients treatment plan. In some circumstances, measure-
ment of drug-dependent antiplatelet antibodies may be also
useful, though these assays are not widely available. Following
discontinuation of an offending drug, the time to resolution
of thrombocytopenia depends on the rate of clearance of that
drug and its metabolites from plasma or tissue and the mech-
anism by which thrombocytopenia is induced. However,
most cases of drug-induced thrombocytopenia resolve within
710 days.
Heparin is the most common cause of drug-induced, anti-
body-mediated thrombocytopenia and is commonly used in
ICU patients; 1 to 2% of heparin-treated patients develop
isolated thrombocytopenia (HIT), and in approximately
30% of these individuals thrombocytopenia is accompanied
by thrombosis (HITT). HIT/T (either asymptomatic throm-
bocytopenia or thrombocytopenia and thrombosis) occurs
with 510-fold greater frequency in patients treated with
UFH versus LMWH. HIT/T has not been reported in patients
exposed to fondaparinux, although patients treated with this
agent developed antibodies reactive with heparin (unfrac-
tionated or low-molecular-weight)platelet factor 4 (PF4)
complexes. The diagnosis of HIT/T is clinical and should be
considered in a patient with an otherwise unexplained fall in
the platelet count of at least 50% occurring 514 days after
initiation of heparin therapy. In patients recently exposed to
heparin, however, HIT may occur soon after reexposure
(median time 10.5 hours).
In patients with HITT, venous thrombi are more common
than arterial thrombi, though this ratio may be reversed in
patients with underlying arterial vascular disease. HITT is
associated with a mortality of approximately 20%, with 10%
of such patients requiring limb amputations. Upon suspicion
of HITT, heparin should be discontinued and alternative anti-
coagulation initiated. Because approximately 50% of patients
with HIT (but not clinically evident thrombosis) will develop
thrombosis after discontinuation of heparin, these individuals
should receive alternative anticoagulation as well. Conrma-
tory laboratory testing should be requested in all patients with
suspected HIT or HITT and is based upon immunologic
measurement of antibodies against heparinPF4 complexes
or the ability of such antibodies to activate platelets.
Alternative anticoagulants used in patients with HIT/T
include the direct thrombin inhibitors lepirudin and argatro-
ban. Lepirudin is approved for treatment of HITT, renally
excreted, and monitored by adjusting the partial thrombo-
plastin time to 1.5 to 2.5 times control. Argatroban is a 509-d,
arginine-based molecule approved for treatment of HITT as
well as for thromboprophylaxis in patients with HIT. Argatro-
ban is metabolized in the liver and adjusted to maintain the
partial thromboplastin time at 1.5 to 3.0 times the baseline
value.
Thrombocytopenia has also been observed in 0.51.0% of
patients treated with the GPIIb/IIIa inhibitor abciximab.
Abciximab-induced thrombocytopenia often develops
within the rst 24 hours after an initial infusion. If this
diagnosis is suspected, abciximab should be stopped and
platelets transfused if the platelet count falls below 20,000/
L. Abciximab may also induce pseudothrombocytopenia;
in this situation, repeating the platelet count in heparinized
blood usually yields a normal value. Ticlopidine and, less
commonly, clopidogrel are thienopyridine adenosine diphos-
phate (ADP)-receptor antagonists that may induce a TTP
HUS-like syndrome in which microangiopathic hemolysis
accompanies the low platelet count.
Infection and disseminated intravascular coagulation
Infection is another common cause of thrombocytopenia in
critically ill patients. The mechanisms of infection-induced
thrombocytopenia include DIC, enhanced clearance of plate-
lets coated by antiplatelet antibodies or nonspecically bound
immunoglobulin, accelerated platelet phagocytosis induced by
high concentrations of macrophage colony- stimulating factor,
and infection of bone marrow stromal cells and megakaryo-
cytes with viruses such as HIV. Management consists of treat-
ing the underlying infection and administering platelet
transfusions as needed. Though platelet transfusion thresholds
as low as 10,000/L are adequate for chemotherapy-induced
Intensive care unit
| 475
|
Consultative hematology 476
thrombocytopenia, higher thresholds (15,00020,000/L) are
appropriate for critically ill patients with fever, infection, and/
or DIC. Management of the coagulopathy of DIC may require
infusion of fresh frozen plasma (FFP) and cryoprecipitate for
severe hypobrinogenemia (brinogen levels 100 mg/dL).
Several uncommon infections may also be associated with
severe thrombocytopenia. Hantavirus is a u-like illness
endemic in the western United States that is associated with
rapid development of noncardiac pulmonary edema. Ehrlich-
iosis is a tick-borne infection that occurs in the southern and
upper midwestern United States, associated with high fever,
headache, myalgias, and CNS symptoms.
Massive transfusion-induced thrombocytopenia
Transfusion of more than 1520 U of packed RBCs may lead
to a dilutional thrombocytopenia. Hypothermia, platelet
dysfunction, and a dilutional coagulopathy (if concurrent
plasma is not administered) may also contribute to bleeding
in massively transfused patients.
Thrombotic thrombocytopenic purpura and the
hemolytic uremic syndrome
TTP and HUS are thrombotic microangiopathies that share
the central processes of microangiopathic hemolytic anemia
and thrombocytopenia. Each of these disorders may be
accompanied by other manifestations such as neurologic
abnormalities, renal dysfunction, and fever. In current prac-
tice, however, microangiopathic hemolytic anemia and
thrombocytopenia without another apparent cause are con-
sidered sufcient criteria for institution of treatment. In gen-
eral, patients with TTP have a higher incidence of neurologic
manifestations, whereas those with HUS have more renal
involvement. The presence of severely decreased activity of
the vWF-cleaving protease, ADAMTS13, is common in clas-
sic TTP, but not HUS. However, the clinical manifestations
of these disorders overlap extensively, and they are consid-
ered by some as a single disorder, TTPHUS.
There are 2 major variants of HUS. The most common
follows enteric infection with verotoxin-producing Esche-
richia coli (VTEC). Infection is followed by abdominal pain
and bloody diarrhea. Up to 20% of patients with bloody
diarrhea progress to HUS and acute renal failure within the
next 56 days. This variant of HUS is most common in chil-
dren, often occurs in epidemics, and is generally self-limited,
though affected patients may require extensive support
including dialysis. The second variant of HUS is referred to
as sporadic and is most common in adults in the postpar-
tum period. The TMA that may complicate stem cell trans-
plantation, chemotherapy, and immunosuppressive
medication resembles this variant. Some apparently sporadic
cases may reect a familial pattern of transmission associ-
ated with a deciency of complement factor H.
The pathogenesis of TTP may involve deciency of the
vWF-cleaving protease, ADAMTS13, leading to increased lev-
els of ultra-large vWF multimers (ULvWF) that are particu-
larly active in binding to platelets and inducing platelet
agglutination. ADAMTS13 deciency is most often due to the
presence of neutralizing antibodies against the protease,
although a congenital variant of TTP may also result from
deciency or dysfunction of the enzyme. The utility of
ADAMTS13 measurement in the diagnosis and management
of TTP remains uncertain due to limited availability and lack
of assay standardization. Decreased ADAMTS13 levels also
occur in situations unrelated to TTP, such as liver disease,
pregnancy, and DIC. Levels in these patients generally remain
above 5%, and only levels below this range appear to have high
specicity for TTP.
Although untreated TTP is fatal in more than 90% of cases,
plasma exchange induces remission in approximately 85% of
patients. The role of corticosteroids in acute TTP is contro-
versial. Although HUS does not appear to respond as well to
plasma exchange as TTP, a trial of this therapy is indicated, in
part due to the difculty in clearly distinguishing these 2 dis-
orders in some patients. Approximately 30% of successfully
treated TTP cases relapse within 12 months, and some patients
experience multiple relapses. Retrospective studies suggest
that splenectomy may reduce the incidence of relapse. Ritux-
imab has been reported to be effective in many cases of refrac-
tory TTP, and its use in this disorder is increasing.
Catastrophic antiphospholipid antibody syndrome
The catastrophic antiphospholipid antibody syndrome is a
rare but serious complication that may occur in patients with
antiphospholipid (aPL) antibodies. This disorder may on
occasion be difcult to distinguish from TTPHUS because
it is characterized by multiorgan involvement by micro-
thrombi and thrombocytopenia. However, only approxi-
mately 15% of patients with the catastrophic antiphospholipid
syndrome have schistocytes on the peripheral blood lm,
whereas schistocytes are present in essentially all patients
with TTPHUS. A consensus conference has dened guide-
lines for the catastrophic antiphospholipid syndrome that
include (i) the presence of lupus anticoagulant activity and/
or positive tests for antiphospholipid (aPL) antibodies,
(ii) involvement of 3 or more organs, (iii) development of
manifestations within 1 week or less, and (iv) conrmation of
histopathology of small vessel occlusion in at least 1 organ.
This disorder is most often precipitated by infection, but other
triggers such as trauma, withdrawal of anticoagulation, or
neoplasia have been described. Due to the rarity of the catas-
trophic antiphospholipid syndrome, recommendations for
therapy are anecdotal. However, plasma exchange, aggressive
anticoagulation, and antibiotics are thought to be of benet.
Platelet transfusion
Platelet transfusion should be used judiciously and, except in
the case of emergencies, only after a clinical diagnosis has
been established. Platelet transfusion in individuals with
chronic thrombocytopenic syndromes increases the risk of
alloimmunization, leading to platelet refractoriness.
In certain conditions, such as TTPHUS and HIT, transfu-
sion of platelets has been followed by a precipitous decline in
the patients clinical status. There has been much discussion
as to the platelet transfusion threshold in severely thrombo-
cytopenic patients. Some experts have suggested that a plate-
let transfusion threshold of 10,000/L be employed in patients
with stable thrombocytopenia without risk factors for bleed-
ing, or active bleeding, whereas a threshold of 20,000/L has
been suggested in patients who are unstable or have signi-
cant risk factors for bleeding. For patients with active bleed-
ing or those undergoing invasive procedures, a platelet count
of 50,000/L has been suggested, although in specic situ-
ations such as the administration of epidural anesthesia an
expert panel has recommended that the platelet count be at
least 80,000/L.
Consultation for hematologic
complications of solid-organ
transplantation
Anemia develops in almost all patients treated in the intensive
care setting for prolonged periods. A prospective randomized
study demonstrated that patients managed using a conservative
transfusion strategy aimed at maintaining the hemoglobin
between 7 and 9.9 g/dL had reduced in-hospital mortality
compared with those managed using a more liberal transfusion
strategy. However, these ndings have not been widely adopted
into clinical practice.
Recombinant erythropoietin can reduce the use of red cell
transfusion in ICU patients, though its cost effectiveness has not
been established.
Thrombocytopenia in ICU patients is often multifactorial and
may result from drugs, infection, or other causes. A history of a
normal platelet count prior to the onset of acute illness
generally suggests that the cause of thrombocytopenia is
related to the acute illness or other intervention.
The diagnosis of HIT remains primarily clinical, and
laboratory testing should be regarded as conrmatory. In
patients with suspected HIT, even in the absence of thrombosis,
heparin should be discontinued and alternative anticoagulation
initiated.
First-line therapy for TTPHUS remains plasma exchange.
Several small reports suggest that rituximab may be useful in
some patients failing plasma exchange, or with relapses.
Key points
You are asked to evaluate a 34-year-old woman with
pancytopenia. She underwent a successful orthotopic liver
transplant 30 days ago. Over the last few days, she has
developed fever and an erythematous rash. The white blood
cell (WBC) count has progressively declined to 1000/L, and
the absolute neutrophil count has dropped to 100/L. The
platelet count is 15,000/L, and hematocrit is 25%. Among
other medications, she is receiving azathioprine, prednisone,
and tacrolimus. There have been no clinical or laboratory signs
of acute graft rejection. She denies bleeding. Blood cultures
have been negative. The patient is cytomegalovirus (CMV)
seronegative; however, the organ was procured from a CMV-
seropositive donor. You recommend a plasma CMV DNA assay,
peripheral blood lymphocyte chimerism study, and marrow
aspirate and biopsy. You also recommend discontinuing
azathioprine and other nonessential medications and starting
empiric antibiotics for febrile neutropenia.
Clinical case
Transfusion support
The hematology consultant may be asked to assist with
transfusion support for solid-organ transplant recipients.
Patients undergoing liver transplantation are particularly chal-
lenging because of the associated coagulopathy. Liver trans-
plant patients require a median of 1012 U each of red cells,
FFP, and platelets. Because of the concern of postoperative
thrombosis of the vascular anastomoses, platelet transfusions
are used sparingly. Fewer blood products are needed for heart
lung and heart transplants, whereas most single lung, kidney,
and kidneypancreas transplants require no blood transfu-
sions. Important considerations for transplant patients include
the potential risks of inducing HLA alloimmunization prior to
transplantation, transmission of CMV, and development of
transfusion-associated graft-versus-host disease (GVHD).
Pretransplant transfusion was formerly a standard method
of inducing immunologic tolerance and reducing the risk of
renal allograft rejection. However, a recent prospective
randomized trial found that transfusion did not improve out-
comes at 1 or 5 years after cadaveric renal transplant, indicat-
ing that immunosuppressive regimens adequately replace the
tolerizing benet of transfusion. Importantly, exposure to
allogeneic blood leukocytes can induce anti-HLA antibodies.
Pretransplant HLA sensitization, as revealed by a positive
panel reactive antibody (PRA) study, is associated with higher
risks of acute and chronic rejection and compromised sur-
vival among recipients of renal, cardiac, and lung transplants.
Consultation for hematologic complications of solid-organ transplantation
| 477
|
Consultative hematology 478
To avoid pretransplant sensitization, renal, heart and lung
transplant candidates should receive only leukocyte-reduced
cellular blood components. Conicting data exist regarding a
potential adverse role of HLA alloimmunization in liver
transplantation; therefore, leukoreduction is considered
optional for potential liver recipients. Organ transplant can-
didates with positive PRA or major ABO incompatibility with
high isohemagglutinin titers against donor A or B antigens
can be treated with plasmapheresis and/or additional immu-
nosuppression (ie, intravenous immunoglobulin [IVIg] or
rituximab) to minimize the risk of hyperacute rejection.
CMV infection is the most frequent infectious complication
after solid-organ transplantation. Although the greatest risk
involves reactivation in a seropositive recipient, the incidence
of transfusion-transmitted CMV in seronegative recipients
of organs from seronegative donors ranges from 6% to 15%.
Therefore, such patients should receive screened cellular
blood products from CMV- seronegative donors or CMV-
safe products that have been leukocyte-reduced by ltration.
A few cases of transfusion-associated GVHD have been
reported among renal, liver, and heart transplant recipients.
However, a clear causal association with cells from the
transfused blood products was convincingly shown in only
2 reports. Most cases of GVHD in solid-organ recipients are
due to transfer of alloreactive donor T cells along with the
transplanted graft (see discussion in the following section).
Because of the rarity of transfusion-associated GVHD, irra-
diation of cellular blood products is not routinely indicated
for solid-organ transplant recipients.
Posttransplant lymphoproliferative disorders
Posttransplant lymphoproliferative disorders (PTLD) com-
prise a morphologically diverse group of lymphoid prolifera-
tions that usually consist of reactive or clonal B cells (roughly
85% of cases) and less commonly T cells, natural killer (NK)
cells, or immune cell mixtures. In contrast to PTLD that arises
from donor B cells after hematopoietic stem cell transplanta-
tion, PTLD in organ recipients most commonly derives from
recipient B cells. Over 80% of B-cell PTLD cases are triggered
by EpsteinBarr virus (EBV) infection. These are usually due
to primary infection rather than viral reactivation. Unre-
strained EBV infection and B-cell proliferation are facilitated
by impaired cytotoxic T-cell (CTL) immune response, micro-
environmental factors, cytokines/lymphokines, and viral
oncogene expression (eg, latent membrane protein 1). T-cell
PTLD is not associated with active EBV infection but may be
caused by human T-cell lymphotropic virus (HTLV), particu-
larly in endemic regions, such as the Far East.
Risk factors for the development of PTLD in solid-organ
transplantation include (i) EBV-seronegative recipient,
particularly if the organ is from an EBV-seropositive donor;
(ii) primary or reactivated EBV infection; (iii) high-dose and/
or prolonged immunosuppression, particularly with potent
anti-T-cell antibodies given for rejection and/or calcineurin
inhibitors (ie, cyclosporine and tacrolimus); (iv) pediatric age
group; (v) CMV infection; and (vi) hepatitis C infection. The
relative risks of PTLD in adult heartlung, lung, liver, pan-
creas, and cadaveric renal transplant recipients are 240-, 59-,
30-, 35- and 13-fold, respectively. The incidence rates in pedi-
atric recipients are 2- to 5-fold higher than rates in adults. A
lower incidence of PTLD has been associated with use of pro-
phylactic antiviral therapies, such as acyclovir and ganciclo-
vir, in higher risk children who are seronegative for EBV prior
to transplant. Plasma EBV RNA levels are moderately elevated
in posttransplant subjects without PTLD and extremely high
in subjects with PTLD. A very short EBV RNA doubling time
occurring within a few months of transplant should raise sus-
picions of emerging PTLD and prompt consideration of pre-
emptive therapy.
The World Health Organization (WHO) has classied
PTLD into 3 major categories (Table 18-6). Histopathologic
interpretations are best conrmed by a referral center that
specializes in the diagnosis of these disorders. Limited cyto-
genetic studies have identied chromosomal abnormalities
in T-cell and most B-cell cases of monomorphic PTLD,
including trisomies of 9 and/or 11 and rearrangements of
8q24.1 (involving MYC), 3q27, and 14q32.
Reactive or polyclonal PTLD usually occurs within 1 year
of transplant and is often accompanied by a rapidly rising
blood EBV RNA titer. Clonal and monomorphic PTLD also
most commonly occur in the early posttransplant course, but
the risk continues over time. Overall incidence rates at years 1,
Table 18-6 World Health Organization classication of posttransplant
lymphoproliferative disorders (PTLD).*
Category Subtype
Early lesions Reactive plasmacytic hyperplasia
Infectious mononucleosis-like
Polymorphic PTLD Polyclonal
Monoclonal
Monomorphic PTLD B-cell lymphomas
Diffuse large B-cell lymphoma
(immunoblastic, centroblastic, anaplastic)
Burkitt/Burkitt-like lymphoma
Plasma cell myeloma
T-cell lymphomas
Peripheral T-cell lymphoma
Other types (-, hepatosplenic, T/NK)
Other types Hodgkin disease-like lesions
Plasmacytoma-like lesions
*Adapted from Harris NL, Jaffe ES, Diebold J, et al. The World
Health Organization classication of neoplastic diseases of the
haematopoietic and lymphoid tissues: report of the clinical advisory
committee meeting, Airlie House, Virginia, November 1997.
Histopathology. 2000;36:6986.
NK natural killer.
2, and 6 after solid-organ transplantation are 0.224%, 0.054%,
and 0.031%, respectively. Fever, weight loss, and constitu-
tional symptoms are common at diagnosis. Rare patients may
present with a sepsis-like syndrome. Roughly 50% of PTLD
tumor masses are localized at presentation to a single extra-
nodal site (kidney, bowel, liver, skin, lungs, tonsils, oral
mucosa, or CNS), whereas 15% are localized to a single nodal
region. One third of patients present with multifocal or dis-
seminated disease involving nodes and/or multiple extra-
nodal tissues. PTLD may be seen in or near the transplanted
organ in 14% of renal transplants and in 2050% of lung
transplant cases. Standard staging studies used for non-
Hodgkin lymphoma (NHL) should be carried out for patients
with PTLD. Limited experience indicates that uorodeoxy-
glucose positron emission tomography (FDG PET) can be
useful in the staging and follow-up of PTLD.
Treatment options for PTLD involve 2 approaches:
(i) interventions to enhance recipient CTL response and (ii)
direct antitumor therapies. Recipient CTL response can be
enhanced by (i) reduction of immunosuppression, which
may induce a remission in up to 60% and 85% of polymor-
phic PTLD cases in adults and children, respectively; (ii)
interferon-alfa (for polyclonal disease when immunosup-
pression reduction is ineffective); (iii) antibodies against
interleukin-6 (IL-6; investigational); and (iv) adoptive T-cell
therapy using EBV-specic donor CTL lines (investigational).
Direct antitumor therapies include (i) surgical excision of
localized disease, (ii) radiotherapy of local disease (particu-
larly in the CNS and as adjuvant therapy after surgery),
(iii) multiagent chemotherapy using regimens effective for
lymphomas in nontransplant patients (eg, CHOP, EPOCH,
DHAP, ProMACE-CytaBOM), (iv) rituximab as a single
agent, (v) chemotherapy in combination with rituximab
(investigational), and (vi) autologous stem cell transplant
with myeloablative conditioning (investigational).
Combination chemotherapy or rituximab are indicated
for clonal PTLD when reduction of immunosuppression is
ineffective. Most chemotherapy regimens require supportive
care with hematopoietic growth factors. Despite reduced
immunosuppression, allograft rejection is uncommon
because of the immunosuppressive effect of chemoimmuno-
therapy. CNS chemoprophylaxis should be a part of treat-
ment with Burkitt and Burkitt-like histologies. Although
variable, the complete remission rate with combination
chemotherapy is roughly 50%; 2030% of patients survive
5 years. Rituximab monotherapy, consisting of 4 or more
doses, induces a response in 4464% of patients, with the
majority being complete remissions. The 1-year survival rate
after rituximab is 5065%; however, 5-year survival may be
only 35%. Given that rituximab monotherapy and multi-
agent chemotherapy yield the same long-term survival, cur-
rent trials are investigating the safety and efcacy of combining
chemotherapy with rituximab.
Drug-related abnormalities
Immunosuppressive agents may cause myelosuppression and
resultant cytopenias. Azathioprine deserves special mention
because it causes cytopenias in more than 10% of patients,
some of whom are affected after a long period of tolerance.
Idiosyncratically, it may cause a syndrome of red cell aplasia;
less frequently, it may cause severe marrow hypoplasia. The
half-life of azathioprine and that of its principal myelotoxic
metabolite, 6-mercaptopurine, is prolonged in renal dysfunc-
tion. Thus, the myelotoxicity of azathioprine will abruptly
increase with rejection of a transplanted kidney or renal dys-
function from any other cause. Allopurinol prolongs the half-
life of azathioprine and 6-mercaptopurine, as well. Although
transplanters are likely to avoid allopurinol in patients taking
azathioprine or to make appropriate dose reductions, other
caregivers who are unaware of the interaction may prescribe
full doses without informing the transplant physician. The
hematologic toxicity of azathioprine is also markedly increased
by other medications including trimethoprim/sulfamethoxa-
zole and angiotensin-converting enzyme (ACE) inhibitors.
Leukopenia and pancytopenia in patients who have
received solid-organ transplants may be due to infection
with CMV or its treatment with ganciclovir. Oral valganci-
clovir may also cause leukopenia, but apparently at a lower
rate than with ganciclovir. Sirolimus and mycophenolate
mofetil can also cause leukopenia and neutropenia.
Thrombotic microangiopathy (TMA), a syndrome resem-
bling TTPHUS, is seen in roughly 23% of renal transplant
patients receiving a calcineurin inhibitor. The incidence is 50%
if the transplant was performed in an adult (but not a child)
with end-stage renal disease induced by HUS. Infrequent cases
of TMA have also been reported following liver, heart, lung,
bowel, and pancreas transplants. The syndrome is characterized
by fragmentation hemolysis, unexplained thrombocytopenia,
and elevated lactate dehydrogenase (LDH), with or without
progressive renal insufciency. Difculty may be encountered in
distinguishing between calcineurin inhibitor-induced TMA and
renal allograft rejection vasculitis. This differential is important
because drug withdrawal is required for TMA whereas dose
increases may be warranted for rejection. The appearance of
focal or diffuse glomerular microangiopathy on renal biopsy
conrms TMA. In one third of renal transplant TMA cases, his-
tologic changes are seen in the graft with minimal or no blood
manifestations. Calcineurin inhibitor dose reduction, a change
from cyclosporine to tacrolimus, or drug withdrawal may resolve
the disorder. However, systemic TMA in renal transplant recipi-
ents leads to dialysis and graft loss in 54% and 38% of cases,
respectively. TMA localized to the graft often responds to drug
manipulation alone. Plasma exchange is often used for persis-
tent or progressive cases; however, the rationale for this treat-
ment in patients without a pretransplant history of TTP is
unfounded and its benet in this setting is unproven.
Consultation for hematologic complications of solid-organ transplantation
| 479
|
Consultative hematology 480
Alloimmune complications
Donor immune cells and marrow-derived cells that accom-
pany the solid organ into the transplant recipient may tran-
siently engraft and give rise to clinical problems related to
antihost cell reactions. These so-called passenger lympho-
cytes can give rise to 2 acquired cellular immune reactions:
(i) GVHD and (ii) immune hemolysis, also referred to as
passenger lymphocyte syndrome.
GVHD is a relatively rare complication of solid-organ
transplantation. One institution reported a 1% incidence in
liver recipients; the greatest risk was among those with a sin-
gle HLA class I mismatch. Because of the high number of
immunocompetent lymphocytes in intestinal tissue, donors
of small bowel grafts are often treated prior to harvest with
antilymphocyte antibodies or corticosteroids to minimize
the transmission of donor T cells. GVHD often presents at
26 weeks posttransplant with unexplained fever, rash, and
diarrhea followed by profound neutropenia and thrombo-
cytopenia. In this setting, the differential diagnosis includes
allograft rejection, drug reaction, and CMV infection. Mar-
row hypoplasia/aplasia and demonstration of large numbers
of circulating donor T lymphocytes strongly support the
diagnosis (donor T lymphocytes are usually undetectable by
conventional methods after 23 weeks posttransplant). The
mortality rate of GVHD in liver recipients is 80%, with
death due to sepsis and multisystem failure. Overall, GVHD
accounts for 0.08% of deaths in posttransplant patients.
Treatment includes aggressive supportive care, high-dose
corticosteroids, anti-T-cell antibodies, and/or a trial of
reduced immunosuppression (to facilitate rejection of the
donor T cells). Investigational therapies include allogeneic
stem cell transplantation and infusion of ex vivo generated
alloreactive antidonor host T cells.
Passenger lymphocyte syndrome is characterized as allo-
immune hemolysis due to donor-derived antibodies against
host red cell antigens. This complication occurs in roughly
10%, 30%, and 70% of kidney, liver, and heartlung recipi-
ents, respectively. Most cases are due to isohemagglutinins
produced by B cells from minor ABO-incompatible donors.
The syndrome has also been associated with donor antibodies
against Rh D, c, e, Jk(a), K, and Fy(a). Hemolysis is usually
abrupt in onset, occurring at 3 to 24 days posttransplant, with
a drop in hemoglobin, elevation in serum bilirubin and LDH,
and a positive DAT. Although the hemolysis may be severe
and life threatening, it is usually mild to moderate and self-
limited. Transfusion support includes group O red cells and
plasma or platelet components that are ABO compatible with
both recipient erythrocytes and donor organ tissue cells (to
avoid transfusing antibody that might contribute to hemoly-
sis or graft rejection). Excessive and/or persistent hemolysis
may require plasma exchange (to decrease antibody titer) or
red cell exchange (to replace host red cells with group O
erythrocytes). High-dose corticosteroids are often used, but
the benet may be outweighed in this setting by the patients
lack of tolerance for additional immunosuppression. Fortu-
nately, the syndrome typically persists for only a few weeks or
months, presumably because the donor lymphocytes have
limited survival in the recipient.
Considerations for transfusion support of transplant patients
include the risks of inducing HLA alloimmunization
preoperatively, transmission of CMV, and the small risk of
transfusion-associated GVHD postoperatively.
PTLD most commonly occurs within the rst year after a
solid-organ transplant, but the risk of developing clonal PTLD
persists for many years.
Roughly 50% of PTLD tumor masses at presentation are
localized to a single extranodal site, 15% are localized to a
single nodal region, and one third present with multifocal or
disseminated involvement. PTLD occurs in or near the
transplanted organ in a signicant number of cases.
Leukopenia and pancytopenia after solid-organ
transplantation may be caused by drugs, infection with CMV, or
GVHD due to transmission of donor alloreactive T cells with the
transplanted organ.
Hemolytic anemia in a posttransplant patient may be due to
TMA or, in the setting of a minor ABO mismatch between
donor and recipient, the passenger lymphocyte syndrome.
Key points
You are asked to see a 19-year-old woman who is 24 weeks
pregnant. A complete blood cell (CBC) count in clinic today
revealed a WBC count of 65,000/L, a platelet count of 25,000/
L, and hematocrit of 19%. Her prenatal course is reportedly
uncomplicated, but she has been noncompliant. Her last clinic
visit was over 3 months ago. She has no signicant past medical
history and is taking only prenatal vitamins. She notes recent
easy bruising and gum bleeding after brushing her teeth. She
complains of fatigue and dyspnea on exertion. Fetal movement
has been normal. Physical examination reveals ecchymoses on
the extremities, shotty cervical lymphadenopathy, and painless
hepatosplenomegaly. Review of the peripheral blood smear
reveals hypochromic red cells; decreased platelets; normal
granulocytes; and a predominant population of homogenous,
immature, agranular leukocytes with a high nuclear:cytoplasmic
ratio and prominent nucleoli. Coagulation studies are normal.
Blood chemistries reveal elevated liver enzymes and lactate
dehydrogenase.
Clinical case
Obstetrics and gynecology
Anemia
During normal pregnancy, the plasma volume expands by
4060%, whereas the red cell mass expands by 2050%.
Thus, a physiologic anemia develops, leading to a normal
hematocrit value of 3032%. A hemoglobin below 10 g/dL
suggests the possibility of a pathologic process such as nutri-
tional deciency. Anemia is surprisingly well tolerated during
pregnancy, and major fetal and maternal complications do
not become frequent until the hemoglobin falls to approxi-
mately 6 g/dL. Iron deciency accounts for 75% of cases of
nonphysiologic anemia in pregnancy, and the incidence of
iron deciency anemia in the United States during the third
trimester may exceed 50%. Current recommendations sug-
gest that pregnant patients receive 1530 mg/d of supple-
mental elemental iron, although studies examining the
efcacy of iron supplementation during pregnancy have not
shown a clear benet to pregnancy outcomes. Iron supple-
mentation may also be given parenterally to patients who do
not absorb or tolerate oral iron. Iron dextran is categorized
as a pregnancy class C drug (safety uncertain) whereas the
newer preparations, iron sucrose and ferric gluconate, are
pregnancy class B (presumed safe based on animal studies).
The majority of megaloblastic anemias during pregnancy
are due to folate deciency, whereas vitamin B
12
deciency is
rare. Folate requirements increase from 50 g/d in the non-
pregnant female to at least 150 g/d during pregnancy, and
the Centers for Disease Control recommends supplementa-
tion with 400 g/d of folate to prevent neural tube defects.
Folate deciency is most precisely diagnosed by measuring
plasma levels of homocysteine and methylmalonic acid.
Thrombocytopenia
Thrombocytopenia affects 10% of pregnant women and
results from several disorders that may or may not be specic
to pregnancy. Pregnant patients may present with isolated
thrombocytopenia or develop thrombocytopenia as a com-
ponent of a systemic disorder that may be unique to preg-
nancy. A summary of causes of thrombocytopenia in
pregnancy is presented in Table 18-7.
Isolated thrombocytopenia most commonly results from
gestational or incidental thrombocytopenia of pregnancy.
Incidental thrombocytopenia usually develops during the
second or third trimester in otherwise healthy pregnant
women. The degree of thrombocytopenia is usually mild, and
by denition the platelet count does not fall below 70,000/L.
This disorder may be a more pronounced variant of the usual
10% decrease in platelet count that occurs during normal
pregnancy. Incidental thrombocytopenia does not affect
pregnancy outcome, does not cause thrombocytopenia in the
offspring of affected women, and requires no specic treat-
ment. However, incidental thrombocytopenia may be impos-
sible to distinguish from ITP or may be an early manifestation
of a more serious disorder that develops later in pregnancy.
Thus, patients with suspected incidental thrombocytopenia
should be monitored closely throughout pregnancy.
ITP affects approximately 1 in 10,000 pregnancies. In
contrast to gestational thrombocytopenia, ITP usually
becomes apparent during the rst trimester. A prior history
of thrombocytopenia or autoimmune disease preceding
pregnancy or during previous pregnancies is useful in making
a diagnosis of ITP. Patients with ITP generally present with
Table 18-7 Differential diagnosis of thrombocytopenia in pregnancy.
MAHA
Thrombo -
cytopenia Coagulopathy Hypertension Liver disease Renal disease CNS disease Time of onset
ITP Mild-severe Anytime,
common in 1st
trimester
Gestational Mild 2nd3rd trimester
Preeclampsia Mild Mild-moderate Absent-mild Moderate-severe Protein Seizures in
eclampsia
3rd trimester
HELLP Moderate-
severe
Moderate-severe May be present
(mild)
Absent-severe Moderate-
severe
Absent-
moderate
Absent-
moderate
3rd trimester
HUS Moderate-
severe
Moderate-severe Absent Absent-mild Absent Moderate-
severe
Absent-mild Postpartum
TTP Moderate-
severe
Severe Absent Absent Absent Absent-
moderate
Absent-severe 2nd3rd trimester
AFLP Mild Mild-moderate Severe Absent-mild Severe Absent-mild Absent-mild 3rd trimester
AFLP acute fatty liver of pregnancy; CNS central nervous system; HELLP hemolysis, elevated liver function tests, low platelets;
HUS hemolytic uremic syndrome; ITP idiopathic thrombocytopenic purpura; MAHA microangiopathic hemolytic anemia;
TTP thrombotic thrombocytopenic purpura.
Obstetrics and gynecology
| 481
|
Consultative hematology 482
more severe thrombocytopenia than those with incidental
thrombocytopenia, but the 2 disorders may be indistinguish-
able when ITP is mild. Therapy for ITP in pregnant patients
is similar to that in patients who are not pregnant; however,
greater emphasis should be placed on avoiding corticoste-
roids, particularly at high doses, because these agents may
cause toxicities unique to pregnancy, such as pregnancy-
induced hypertension. Prednisone exposure in the rst tri-
mester may also be teratogenic, with an increased incidence
of cleft palate. Some have advocated the use of IVIg to mini-
mize corticosteroid exposure; success with anti-D has also
been reported anecdotally. Patients with severe refractory
thrombocytopenia should be considered for splenectomy,
which may be performed most safely in the mid second tri-
mester, when the risk of inducing premature labor is mini-
mized and the gravid uterus does not yet obscure the surgical
eld. Approximately 10% of the offspring of patients with
ITP will also be thrombocytopenic, and 5% will have platelet
counts below 20,000/L. However, no maternal interven-
tions have been convincingly shown to increase the fetal
platelet count. Likewise, cesarean delivery of the offspring of
mothers with ITP has not been shown to reduce the risk of
fetal intracranial hemorrhage, a rare complication affecting
1% of these infants, at delivery. These considerations, and
the appreciation that the risk of fetal platelet count determi-
nation by percutaneous umbilical cord blood sampling
(PUBS) is likely greater than that of fetal intracranial hemor-
rhage associated with vaginal delivery, explain why the utili-
zation of PUBS has decreased signicantly in recent years.
Thrombocytopenia may occur in patients with preeclamp-
sia. Preeclampsia affects 6% of all rst pregnancies and usu-
ally develops in the third trimester; its diagnostic features
include hypertension and proteinuria (300 mg/24 hours).
Preeclampsia occurs most commonly in primigravidas, who
are either 20 or 30 years of age. Eclampsia, dened by the
presence of grand mal seizures accompanying preeclampsia,
complicates 1% of preeclamptic pregnancies. Some studies
suggest that preeclampsia may be more common in patients
with thrombophilia, particularly those with aPL antibodies.
Thrombocytopenia develops in up to 50% of patients with
preeclampsia, with its severity generally related to that of the
underlying disease. The pathogenesis of thrombocytopenia
in preeclampsia is not well understood, although platelet
consumption related to subtle activation of the coagulation
system has been suggested. However, because thrombocyto-
penia may in some cases precede other manifestations of pre-
eclampsia, it is unlikely that activation of the coagulation system
is the primary cause of thrombocytopenia in all patients.
The HELLP (hemolysis, elevated liver function tests, low
platelets) syndrome affects 0.10.9% of all live births and is
associated with a maternal mortality of 04%. HELLP and
preeclampsia share many clinical features, though HELLP
occurs in a slightly older population (mean age of approxi-
mately 25 years). Because approximately 10% of patients with
HELLP develop hypertension and proteinuria, it is often con-
sidered a preeclampsia variant. The major diagnostic criteria
for HELLP include microangiopathic hemolytic anemia, levels
of serum aspartate aminotransferase exceeding 70 U/L, and
thrombocytopenia, with a platelet count below 100,000/L.
Microangiopathic hemolysis is accompanied by schistocytes
on the peripheral blood lm and an elevated LDH; some
experts suggest that a minimal LDH of 600 U/dL should be
required for diagnosis. In these cases, HELLP may be very dif-
cult to distinguish from TTPHUS. Because many patients
with HELLP may present with isolated right upper quadrant
and epigastric pain in the absence of hypertension and pro-
teinuria, patients may be misdiagnosed as having primary
gastrointestinal disease and referred for surgical consider-
ation. HELLP is associated with signicant maternal and fetal
morbidity and mortality, which exceeds that associated with
preeclampsia; therefore, prompt diagnosis and treatment are
essential. The offspring of patients with both preeclampsia
and HELLP may also become thrombocytopenic.
Therapy for HELLP and preeclampsia is directed toward
stabilization of the mother, followed by expeditious delivery,
after which these disorders usually remit. However, HELLP,
in particular, may occasionally worsen or even develop post-
partum. Pre- or postnatal corticosteroids have been sug-
gested in several small randomized studies (not placebo
controlled) to hasten resolution of the biochemical abnor-
malities and thrombocytopenia associated with HELLP,
though these studies have not been sufciently powered to
demonstrate an effect on maternal or fetal mortality.
TTP and HUS are discussed earlier in this chapter. The
incidence of TTP appears to be increased during pregnancy,
and pregnant women comprise 1020% of the patients in
many large TTP series. Whether the decreased levels of
ADAMTS13 that occur in pregnancy predispose to the devel-
opment of thrombotic microangiopathies in this setting is
uncertain. Though some reports suggest that most cases of
TTP develop in the late second trimester, recent studies dem-
onstrate that TTP occurs commonly near term. The manifes-
tations of TTP in pregnant and nonpregnant women are
identical, and pregnant patients respond equally well to
plasma exchange. Although delivery of the fetus is the stan-
dard approach to management of patients with preeclamp-
sia, this intervention does not ameliorate the course of TTP.
The incidence of HUS is also increased in pregnancy.
Although some cases of HUS develop near term, the major-
ity of cases develop 3 to 4 weeks postpartum. These cases
resemble most closely the sporadic forms described previ-
ously, with renal failure as the predominant manifestation.
The prognosis of postpartum HUS is poor, with a mortality
rate as high as 25%. Although responses to plasma exchange
have been reported, the response of HUS in this setting is not
as robust as that in TTP; nevertheless, a trial of plasma
exchange is indicated, particularly given the difculty in dis-
tinguishing these disorders.
Last, there are several additional causes of thrombocyto-
penia that the hematologist must consider in pregnant
patients. DIC may accompany severe preeclampsia and also
be initiated by processes such as retained fetal products, pla-
cental abruption, or amniotic uid embolization. It is com-
monly associated with mild thrombocytopenia, and the
degree of microangiopathic hemolysis is generally not as
severe as that observed in preeclampsia, HELLP, or the pri-
mary thrombotic microangiopathies. Acute fatty liver of
pregnancy (AFLP) usually occurs in the third trimester and
affects primarily primiparas and women with twin gesta-
tions. Symptoms include nausea, vomiting, right upper
quadrant pain, anorexia, jaundice, and cholestatic liver dys-
function. Most patients have DIC, and diabetes insipidus
and hypoglycemia are present in more than 50% of cases.
Thrombocytopenia is usually mild, but maternal bleeding is
common due to the accompanying coagulopathy resulting
from diminished hepatic synthesis of coagulation proteins,
as well as DIC and acquired antithrombin deciency. Some
cases of AFLP and possibly HELLP may result from fetal
mitochondrial fatty acid oxidation disorders (FAODs). The
most common is a deciency of long-chain 3-hydroxyacyl-
CoA dehydrogenase. Drug-induced thrombocytopenia has
been discussed previously, and HIV infection may also cause
an immune-mediated thrombocytopenia due to antibodies
reactive with amino acids 4966 of GPIIIa, which cross-react
with HIV Nef. Impaired megakaryopoiesis also occurs in
HIV-infected patients due to infection of megakaryocytes by
the virus. Finally, patients with HIV also have an increased
incidence of TTP that has occurred in pregnant individuals.
Congenital thrombocytopenias, such as the MayHegglin
anomaly or other disorders associated with mutations in the
MYH9 gene, may occasionally escape diagnosis until adult-
hood. Type IIB vWD may induce thrombocytopenia due to
enhanced platelet agglutination caused by increased levels of
mutant vWF in pregnancy. The mutant vWD in these patients
binds its platelet receptor GPIb with increased avidity, and
the resultant platelet agglutination may lead to accelerated
platelet clearance and thrombocytopenia.
von Willebrand disease
vWD is the most common inherited bleeding disorder and
is found in approximately 1% of the general population.
There are several types of vWD. Type I results from a partial
quantitative deciency of vWF and accounts for 50% of all
vWD cases. Type II vWD consists of several subtypes: type
IIN is characterized by a loss-of-function mutation in the
FVIII binding site in vWF, type IIM by decreased afnity of
vWF for its platelet receptor GPIb, and type IIA by a selective
decrease in the high-molecular-weight multimers of vWF.
Type IIB has been discussed previously. Type III vWD is
characterized by a severe deciency in vWF and a corre-
sponding deciency of FVIII.
Levels of vWF and FVIII increase during pregnancy, gen-
erally beginning in the early second trimester. Levels of these
proteins peak around term and decline postpartum. Most
patients with type I vWD normalize their levels of vWF and
FVIII during pregnancy, making the diagnosis of vWF in
these individuals difcult, although those with more severe
disease may not. Although levels of vWF may also increase in
patients with type II vWD, functional levels may not be sig-
nicantly enhanced due to the production of a functionally
decient protein. Levels of vWF generally do not increase
during pregnancy in patients with type III vWD.
The decline in vWF levels after delivery is unpredictable
and subject to signicant variability among individuals.
Although levels may decline gradually over 12 weeks in
some patients, in others they decrease precipitously. Thus, if
a diagnosis of vWD is initially suspected in a pregnant
patient, this individual should be reassessed 68 weeks post-
partum for a diagnosis to be conrmed.
Bleeding may occur during either pregnancy or postpar-
tum in patients with vWD. Antepartum hemorrhage is
unusual, but when it does occur is more common during the
rst trimester, prior to the increase in vWF levels. Postpar-
tum hemorrhage is more common and may be primary
(occurring within the rst 24 hours after delivery) or sec-
ondary (occurring later). In many cases, vWD may be rst
diagnosed following an episode of postpartum hemorrhage.
Although the frequency of postpartum hemorrhage is not
well dened, the ndings of several retrospective studies sug-
gest that it affects 3050% of women with vWD and is more
common in patients with type II vWD and those with vWF
levels 50%, underlying the recommendation by some that
vWF levels be maintained at 50% or higher during delivery.
However, there are no strict guidelines as to the levels of vWF
required for safe delivery and/or regional anesthesia.
Several therapeutic options are available for therapy for
postpartum hemorrhage in patients with vWD. DDAVP has
been used successfully, primarily in patients with type I vWD.
Although theoretical concerns about the vasoconstrictive
effects of DDAVP on placental perfusion, as well as the
potential oxytocic effect of DDAVP, have been raised, these
have not been borne out in practice. However, because preg-
nant patients often receive large volumes of uids at the time
of delivery, DDAVP should be used with caution over this
period because signicant uid retention and hyponatremia
have been reported anecdotally. For patients requiring
replacement therapy, puried vWFFVIII concentrates are
Obstetrics and gynecology
| 483
|
Consultative hematology 484
preferred over cryoprecipitate because they have undergone
virucidal treatment. Because the rate of fall in vWF levels
after delivery varies substantially among individuals, and
patients may be at risk for bleeding for several weeks, the dura-
tion of treatment must be determined on an individual basis.
Thrombophilia and thromboembolism during
pregnancy
Women with thrombophilia are at increased risk for thrombo-
embolism during pregnancy and the puerperium, as well as
fetal loss. Compared to the nonpregnant state, the relative risk
of VTE during pregnancy and the postpartum period is
approximately 4.2, with an overall incidence of 199.7/100,000
in a primarily Caucasian population. The rate of VTE is
approximately 5-fold higher in the postpartum period than
during pregnancy. The increased risk of VTE during preg-
nancy is signicantly amplied in patients with thrombo-
philia; indeed, approximately half of all thromboemboli in
pregnant women are associated with inherited thrombo-
philia. Current guidelines suggest that in patients with VTE
during pregnancy or in the postpartum period, fetal loss that
is recurrent and/or late in pregnancy, or a history of cerebral
venous thrombosis during oral contraceptive use or hormone
replacement therapy should be evaluated for thrombophilia.
The increased risk of thrombosis in pregnancy reects in
part the normal physiologic hypercoagulability that occurs in
this setting. Pregnancy is associated with increased levels of
factors VII, VIII, X, brinogen, and vWF, as well as increases
in plasminogen-activator inhibitor type 1. In parallel, a sub-
stantial decrease in the levels of protein S occurs. In addition
to these physiologic changes, other factors associated with
pregnancy such as immobilization during complicated preg-
nancies and following cesarean delivery or compression of
the iliac veins by the gravid uterus may predispose to the
development of thrombosis.
Most lower extremity VTEs in pregnant patients involve
the left leg. The diagnosis of VTE is difcult in this popula-
tion, due to the lack of validation of standard diagnostic stud-
ies. A recent systematic review concluded, however, that an
abnormal compression ultrasound (CUS) is sufcient for a
diagnosis of DVT during pregnancy. By comparison, a nor-
mal CUS does not reliably exclude DVT because of the low
sensitivity for isolated iliac DVT; thus, patients with a normal
CUS should undergo repeat testing 68 days later, or earlier if
indicated by symptoms. Negative serial CUS studies reliably
exclude DVT. If doubt about the diagnosis persists, limited
venography with abdominal shielding may be performed.
This exposes the fetus to 0.05 mSv (1 mSv 0.1 rad) of
radiation. In patients with suspected pulmonary embolism
(PE), a negative ventilation/perfusion (V/Q) scan excludes
PE, whereas a positive scan conrms the diagnosis. All other
abnormal V/Q scans should be considered nondiagnostic,
and helical computed tomographic (CT) scanning or, alter-
natively, shielded pulmonary angiography should be pursued
to assess the pregnant patient further. There is insufcient
information at this time to dene accurately the utility of
D-dimer testing in diagnosing DVT during pregnancy.
Estimates for the positive predictive value of several
thrombophilic disorders in pregnancy-associated thrombo-
embolism have been reported. Assuming an underlying rate
of 0.66 cases of VTE per 1000 pregnancies in Western popu-
lations, the risk for thrombosis was estimated at approxi-
mately 1:500 in patients with factor V Leiden, 1:200 in
patients with heterozygous prothrombin G20210A, 1:113 in
patients with protein C deciency, and 1:2.8 in patients with
type 1 antithrombin deciency. Patients homozygous for
factor V Leiden have an absolute risk of thrombosis during
pregnancy of approximately 916%, whereas patients doubly
heterozygous for factor V Leiden and prothrombin G20210
have an estimated absolute risk of thrombosis of 4%. aPL
antibodies, in addition to their association with poor preg-
nancy outcomes (see discussion later in this section), are
associated with an increased risk of thrombosis in pregnancy,
although the exact magnitude of the risk is uncertain. This
risk is likely to be greatest in patients with a history of prior
thrombosis because these antibodies are found in approxi-
mately 2% of the general low-risk obstetric population.
Estimates of the risk of recurrent VTE during a subsequent
pregnancy in a patient with VTE in a prior pregnancy have
varied widely. One prospective study examined 125 pregnant
women with prior VTE. No antepartum heparin was admin-
istered, and patients received coumadin for 46 weeks after
delivery. Afterward, 95 patients underwent thrombophilia
testing. The rate of antepartum VTE recurrence was 2.4%,
but there were no recurrences among the 44 patients who did
not have thrombophilia. The recurrence rate in patients with
thrombophilia or an idiopathic initial thrombosis was 5.9%.
Several important issues unique to the use of anticoagu-
lants during pregnancy bear consideration. First, the oral
vitamin K antagonist (VKA) warfarin is teratogenic, causing
an embryopathy consisting of nasal hypoplasia and/or stip-
pled epiphyses. The teratogenic effects occur during weeks
612, whereas warfarin is probably safe during the rst 6
weeks of gestation. However, more recently an increased risk
of neurodevelopmental abnormalities have been described in
the offspring of women exposed to warfarin during the sec-
ond and third trimesters. Warfarin may also cause an antico-
agulant effect in the fetus. Second, UFH is associated with a
substantial risk of osteoporosis and a 23% incidence of ver-
tebral fractures when administered throughout pregnancy.
Recent reports suggest substantially less osteoporosis when
patients receive LMWH. LMWH also displays better pharma-
cokinetic proles and a 510-fold reduction in the incidence
Taken together, these risk estimates and other consider-
ations have led to the development of comprehensive guide-
lines prepared by the American College of Chest Physicians.
These are depicted in Table 18-8. Of note, postpartum anti-
coagulation is recommended for many situations. Warfarin
of HIT. As pregnancy progresses and the LMWH volume of
distribution increases, dose adjustments of LMWH may be
required. Moreover, some have suggested that LMWH be
changed to UFH at the 36th week of gestation, due to the shorter
half-life of the former in the setting of incipient parturition.
Obstetrics and gynecology
| 485
Table 18-8 American College of Chest Physicians Guidelines for Anticoagulant Treatment of VTE/Thrombophilia in Pregnancy.*
Scenario Recommended treatment Comments
Acute VTE
Acute VTE Adjusted-dose LMWH throughout pregnancy, or
IV UFH (bolus followed by continuous infusion)
for at least 5 days followed by adjusted-dose
UFH or LMWH for the remainder of the pregnancy
Continue anticoagulants at
least 6 weeks postpartum
Prior episode of VTE
Prior VTE associated with transient risk factor Close clinical surveillance and postpartum
anticoagulation
If initial event was estrogen
or pregnancy related, use
antenatal prophylaxis also
Single VTE, idiopathic Prophylactic LMWH, or
Mini- or moderate-dose UFH, or
Clinical surveillance and postpartum anticoagulants
Single episode of VTE and thrombophilia or strong
family history of thrombosis, not receiving
anticoagulants
Prophylactic- or intermediate-dose LMWH, or
Mini-dose or moderate-dose UFH
Postpartum anticoagulants
Antithrombin deciency, compound heterozygous
prothrombin G20210A and factor V Leiden, or
homozygotes for these conditions, with history
of VTE
Intermediate-dose LMWH prophylaxis, or
Moderate-dose UFH
Postpartum anticoagulants
Multiple episodes of VTE and/or receiving long-term
anticoagulants (ie, single VTE, idiopathic or
associated with thrombophilia)
Adjusted-dose UFH, or
Adjusted-dose LMWH
Resume long-term
anticoagulants postpartum
All women with prior DVT Antenatal and postpartum compression stockings
No prior VTE
Antithrombin deciency, compound heterozygous
prothrombin G20210A and factor V Leiden, or
homozygotes for these conditions, with history
of VTE
Intermediate-dose LMWH prophylaxis, or
Moderate-dose UFH
Postpartum anticoagulants
No prior VTE and thrombophilia Surveillance, or
Prophylactic LMWH, or
Mini-dose UFH
Postpartum anticoagulants
Denitions of anticoagulant regimens:
Mini-dose UFH: UFH 5000 U SC every 12 hours.
Moderate-dose UFH: UFH SC every 12 hours in doses sufcient to achieve an anti-Xa level of 0.10.3.
Adjusted-dose UFH: UFH SC every 12 hours in doses sufcient to achieve a midinterval aPTT into the therapeutic range.
Prophylactic- or intermediate-dose LMWH: eg, enoxaparin 40 mg SC every 24 hours, dalteparin 5000 U SC every 12 hours.
Adjusted-dose LMWH: weight-adjusted full-treatment doses of LMWH once or twice daily. Twice-daily dosing is preferred, due to shortened
half-life in pregnancy.
Postpartum anticoagulants: warfarin for 46 weeks, with target INR of 2.03.0, with initial UFH or LMWH overlap until the INR is >2.0.
Surveillance: clinical vigilance and aggressive investigation of women with symptoms suspicious of VTE.
*Adapted from Bates SM, Greer IA, Hirsh J, et al. Use of antithrombotic agents during pregnancy. The Seventh ACCP Conference on
Antithrombotic and Thrombolytic Therapy. Chest. 2004;126(suppl):S62744.
aPTT activated partial thromboplastin time; INR international normalized ratio; IV intravenous; LMWH low-molecular-weight
heparin; SC subcutaneous; UFH unfractionated heparin; VTE venous thromboembolic event.
|
Consultative hematology 486
is not present in high concentrations in breast milk, so
mothers on warfarin may breast-feed safely.
Thrombophilia, fetal loss, and poor pregnancy outcomes
Most studies addressing the relationship between thrombo-
philia and fetal loss have been small, retrospective, and
uncontrolled case-control or cohort studies, making conclu-
sions difcult. However, recent metaanalyses have helped to
assimilate this large body of data.
In a recent metaanalysis, the contribution of thrombo-
philias to pregnancy outcomes was examined in detail.
Briey, this analysis found that both hetero- and homozy-
gous factor V Leiden, heterozygous prothrombin G20210A,
anticardiolipin antibodies, lupus anticoagulants, acquired
activated protein C resistance (APCR), and hyperhomo-
cysteinemia, but not antithrombin deciency, homozygous
MTHFR C677T, or protein C and S deciencies were associ-
ated with early pregnancy loss (dened as rst or second tri-
mester). For rst trimester pregnancy loss alone, signicant
associations were seen for factor V Leiden (homo- and het-
erozygous grouped together), heterozygous prothrombin
G20210A, anticardiolipin antibodies, acquired APCR, and
homocysteinemia, but not homozygous MTHFR. For nonre-
current second trimester loss, factor V Leiden (homo- and
heterozygous), heterozygous prothrombin G20210A, and
lupus anticoagulants were identied as major risk factors.
Both factor V Leiden and prothrombin G20210A were more
strongly associated with second trimester loss than rst. For
late pregnancy loss (third trimester), protein S deciency was
most strongly associated, but signicant associations were
also observed for factor V Leiden, heterozygous prothrombin
G20210A, and anticardiolipin antibodies. In interpreting all
of this data, however, the limitations of metaanalyses must
be considered. Moreover, in most cases the odds ratios are
low and the condence intervals wide.
Associations of thrombophilias with pregnancy complica-
tions, in particular preeclampsia, as well as intrauterine
growth retardation (IUGR) have also been demonstrated,
although less well established.
These associations have led to widespread screening of
patients with poor pregnancy outcomes for congenital and
acquired thrombophilias and have led to the increasing use of
anticoagulant therapy in attempts to optimize pregnancy
outcomes. Indeed, current guidelines developed by the Amer-
ican College of Chest Physicians suggest that women with
recurrent pregnancy loss (3 or more miscarriages) and women
with prior severe or recurrent preeclampsia, abruptions, or
otherwise unexplained intrauterine death should be screened
for congenital thrombophilia and aPL antibodies. However,
despite the association of thrombophilias with poor preg-
nancy outcomes, the trials assessing the efcacy of anticoagu-
lation therapy have yielded conicting results. Most such
studies have not been randomized or placebo controlled, and
outcomes often compared with historical controls. The larg-
est body of evidence for efcacy of anticoagulant therapy for
recurrent fetal loss has been in patients with aPL antibodies.
In these patients, current guidelines suggest the use of low-
dose aspirin (5081 mg/d) and heparin to improve pregnancy
outcome. However, even in this disorder, a recent randomized
study failed to demonstrate improved outcomes in patients
treated with LMWH and aspirin compared to that in patients
treated with aspirin alone. In other trials, aspirin alone did
not improve outcome compared with placebo.
Current recommendations from the American College
Chest Physicians for the use of anticoagulants in the treat-
ment of fetal loss and pregnancy complications due to
thrombophilias are summarized in Table 18-9.
Articial heart valves
Without anticoagulant therapy, patients with mechanical
heart valves have a high risk of arterial thromboembolism.
Warfarin appears to be more effective than heparin at pre-
venting valvular thrombosis in these patients. However,
debate continues as to whether the benet to the mother in
prevention of valvular thrombosis offsets the risk of warfarin-
induced embryopathy and neurodevelopmental abnormali-
ties. A recent systematic review assessed the merits of 3
approaches: (i) the continued use of VKAs throughout preg-
nancy, (ii) replacement of VKAs with heparin from 612
weeks, or (iii) use of UFH throughout pregnancy. This analy-
sis concluded that the use of VKAs throughout pregnancy
was associated with a 6.4% incidence of warfarin embryo-
pathy; this was eliminated when heparin was substituted at
or before 6 weeks. The incidence of fetal wastage, and of bleed-
ing, was similar in all groups. The use of VKAs throughout
pregnancy led to the lowest incidence of valve thrombosis
and systemic embolism (3.9%); substitution of heparin from
weeks 612 increased this risk to 9.2%. This analysis is
consistent with other retrospective studies in that VKAs were
associated with a lower risk of valve thrombosis and systemic
embolism, though estimates of the risk of warfarin embryo-
pathy in smaller studies have ranged from 025%. LMWH
has also been used in pregnant patients with mechanical
heart valves; however, valvular thrombosis has been reported
at an apparently higher incidence than might be expected.
Hence, at present, there remains insufcient data to make
denitive recommendations concerning the use of VKAs
versus heparin in pregnant patients with mechanical heart
valves. It is possible that some of the apparently increased
risk of heart valve thrombosis observed in patients on UFH
or LMWH may have resulted from inadequate dosing. Cur-
rent guidelines developed by the American College of Chest
Physicians are depicted in Table 18-10 and take this possibil-
ity into account by recommending stringent monitoring of
both of these agents using plasma anti-Xa levels.
Hematologic malignancies
Hematologic malignancies account for 25% of cancers diag-
nosed during pregnancy, similar to the rates of gestational
breast and cervical cancers. The most common diagnoses are
Hodgkin disease (HD; 50% of cases; incidence: 0.060.10% of
pregnancies) and acute leukemia (incidence: 0.001%). Among
the leukemias, acute myeloid leukemia (AML) and acute lym-
phoblastic leukemia (ALL) account for roughly 67% and 25%
of cases, with the remainder consisting of chronic leukemias.
The appropriate evaluation and therapeutic strategies for
hematologic malignancies during pregnancy depend heavily
on the aggressiveness of the disease, the stage of pregnancy,
the patients willingness to bear personal risk to avoid ther-
apy to protect her fetus, and the option of accepting fetal risk
to receive optimal management for the malignancy. If aggres-
sive malignancy strikes in the rst trimester, when chemo-
therapy incurs a high risk of spontaneous abortion, fetal
death, and major malformations, it is appropriate to explore
the patients attitudes toward abortion. Accordingly, a plan
should be made in full partnership with the patient and her
trusted personal advisors or with the help of the hospitals
ethics committee.
HD during pregnancy usually presents with the same fea-
tures that are typical among nonpregnant patients (ie, supra-
diaphragmatic lymphadenopathy, pruritis, fever, sweats, and
weight loss). Diagnosis should be based on histopathologic
evaluation of excisional lymph node biopsy material or other
involved tissue. Because CT scans deliver a signicant radia-
tion exposure to the fetus, magnetic resonance imaging
(MRI) should be used to stage disease in the chest, abdomen,
and pelvis. Ultrasound is a less sensitive alternative. Bone
marrow biopsies from either posterior iliac crest are safe.
Roughly 70% of pregnant women with HD are diagnosed
with stage I or II disease. In general, concurrent HD does not
adversely affect pregnancy and fetal development. In turn,
the response rate of HD to treatment and the overall long-
term survival are not usually compromised by pregnancy.
Whenever possible, chemotherapy should be delayed until
the second trimester or until after delivery in a woman with
Table 18-9 American College of Chest Physicians Guidelines for Anticoagulant Therapy of Thrombophilia-Related Fetal Loss.*
Scenario Recommended treatment Comments
aPL antibodies and a history of multiple (2 or
more) late-pregnancy losses, preeclampsia,
IUGR, or abruption
Antepartum aspirin plus mini-dose or
moderate-dose UFH or prophylactic
LMWH
Homozygous for thermolabile variant (C677T)
of MTHFR
Folic acid supplementation prior to
conception, or if already pregnant
as soon as possible, and throughout
pregnancy
Congenital thrombophilic decit and recurrent
miscarriages, a second trimester or later loss,
severe or recurrent preeclampsia, or abruption
Low-dose aspirin therapy plus either
mini-dose heparin or prophylactic
LMWH therapy
Postpartum anticoagulation should
also be administered
aPL antibodies and a history of venous thrombosis,
usually receiving long-term oral anticoagulation
Adjusted-dose LMWH or UFH, plus
low-dose aspirin
Resume long-term anticoagulation
therapy postpartum
aPL antibodies and no prior VTE or pregnancy loss Surveillance, or
Mini-dose heparin, or
Prophylactic heparin, or
Low-dose aspirin, 75162 mg/d
For prevention of VTE, as well
as possible pregnancy loss or
complications
Denitions of anticoagulant regimens:
Mini-dose UFH: UFH 5000 U SC every 12 hours.
Moderate-dose UFH: UFH SC every 12 hours in doses sufcient to achieve an anti-Xa level of 0.10.3.
Adjusted-dose UFH: UFH SC every 12 hours in doses sufcient to achieve a midinterval aPTT into the therapeutic range.
Prophylactic- or intermediate-dose LMWH: eg, enoxaparin 40 mg SC every 24 hours, dalteparin 5000 U SC every 12 hours.
Adjusted-dose LMWH: weight-adjusted full-treatment doses of LMWH once or twice daily. Twice-daily dosing is preferred, due to shortened
half-life in pregnancy.
Postpartum anticoagulants: warfarin for 46 weeks, with target INR of 2.03.0, with initial UFH or LMWH overlap until the INR is >2.0.
*Adapted from Bates SM, Greer IA, Hirsh J, et al. Use of antithrombotic agents during pregnancy. The Seventh ACCP Conference on
Antithrombotic and Thrombolytic Therapy. Chest. 2004;126(suppl):S62744.
aPL antiphospholipid; aPTT activated partial thromboplastin time; IUGR intrauterine growth retardation; LMWH low-molecular-
weight heparin; SC subcutaneous; UFH unfractionated heparin; VTE venous thromboembolic event.
Obstetrics and gynecology
| 487
|
Consultative hematology 488
stable low-stage disease. A patient with symptomatic, local-
ized supradiaphragmatic disease (ie, diagnostic studies sug-
gest a low risk of intraabdominal involvement) can be
adequately treated with radiation to axillary, cervical, and
upper mediastinal nodes without undue risk to the fetus. Of
note, however, preliminary studies suggest that breast cancer
may occur at a higher rate and earlier among women who
received radiotherapy for HD around the time of pregnancy,
compared with nonpregnant women who received similar
radiotherapy.
Patients with aggressive HD during pregnancy after the
rst trimester should receive multiagent chemotherapy with
curative intent (eg, mechlorethamine, vincristine, prednisone,
and procarbazine or doxorubicin, bleomycin, vinblastine,
and dacarbazine). Retrospective and long-term follow-up
studies of children exposed to chemotherapy in utero have
shown relatively few congenital malformations and low
incidence rates of hematologic malignancies and cardiac
dysfunction (among those exposed to anthracyclines). Neo-
natal transmission of HD has not been reported; however,
placental metastases have been noted with HD and NHL,
and rare cases of infant NHL identical to the maternal dis-
ease have occurred. Therefore, the placenta should be
inspected closely after birth, and consideration should be
given to storage of umbilical cord blood cells.
Pregnancy-associated AML or ALL demands immediate
attention. Untreated, the disease may be rapidly fatal to the
mother and, consequently, to the fetus. Management of a
pregnant woman with acute leukemia is the same as manage-
ment of a nonpregnant woman, including the use of aggres-
sive induction chemotherapy and supportive care. Because of
the high incidence of fetal loss when induction chemotherapy
is given during the rst trimester, treatment should be tempo-
rarily delayed, if possible, or termination should be strongly
considered. Cytarabine, commonly used for AML, appears to
be particularly teratogenic and should be avoided in the rst
trimester. In selected cases, the patient may be transiently
managed with aggressive supportive care and leukapheresis.
When induction chemotherapy is given later in pregnancy, the
fetal prognosis depends on the treatment outcome for the
mother. The chances of achieving complete remission during
pregnancy are similar to those in nonpregnant patients. Drugs
are dosed according to the maternal weight, although phar-
macokinetics may be affected by pregnancy-associated changes
in plasma volume, renal function, and hepatic function. All-
trans retinoic acid has been anecdotally successful in treating
acute promyelocytic leukemia during pregnancy. In women
with ALL, L-asparaginase should be withheld because of the
added prothrombotic risks of pregnancy. Fetal transmission
of maternal acute leukemia has been reported, so the neonate
should be screened for this complication.
A number of reports have recently described pregnancies
occurring in women with a complete hematologic remission
of chronic myeloid leukemia (CML) while on imatinib
therapy. In many cases, the imatinib therapy was discontin-
ued to retain the pregnancy, and CML recurred, requiring
interferon-alfa or leukapheresis. Healthy babies resulted in
80% of those pregnancies that were not electively termi-
nated. Minor complications (hypospadius and intestinal
rotation) have been noted in a couple of neonates. Anec-
dotal reports have also described babies with no congenital
defects born to mothers who continued imatinib through-
out pregnancy. Despite the reassurance of these reports,
imatinib is teratogenic in rats when administered at doses
equivalent to a human adult dose of 800 mg/d. Therefore,
women on imatinib should continue to practice adequate
contraception, and the drug should be withheld during
pregnancy or lactation.
Table 18-10 American College of Chest Physicians Guidelines for Anticoagulant Treatment of Pregnant Patients With Articial Heart Valves.*
Scenario Recommended treatment Comments
Pregnant women with prosthetic
heart valves
Adjusted-dose, twice-daily LMWH throughout pregnancy, in doses adjusted
either to keep a 4-hour postinjection anti-Xa heparin level at approximately
1.01.2 U/mL (preferable) or according to weight, or
Resume long-term
anticoagulants postpartum
with all regimens
Aggressive adjusted-dose UFH throughout pregnancy, ie, administered SC
every 12 hours in doses adjusted to keep the midinterval APTT at least twice
control or to attain an anti-Xa heparin level of 0.35 to 0.70 U/mL, or
UFH or LMWH (as above) until the 13th week, change to warfarin until the
middle of the third trimester, and then restart UFH of LMWH
Pregnant women with prosthetic
heart valves at high risk
One of the above regimens, plus low-dose aspirin, 75162 mg/d
*Adapted from Bates SM, Greer IA, Hirsh J, et al. Use of antithrombotic agents during pregnancy. The Seventh ACCP Conference on
Antithrombotic and Thrombolytic Therapy. Chest. 2004;126(suppl):S62744.
APTT activated partial thromboplastin time; LMWH low-molecular-weight heparin; SC subcutaneously; UFH unfractionated heparin.
Selected outpatient hematology topics
Mild thrombocytopenia
Although the denition of mild is arbitrary, patients with
platelet counts in the range of 80,000/L to 120,000/L are
commonly referred for consultation. Mild thrombocyto-
penia may result from less robust presentations of many of
the same disorders that cause severe thrombocytopenia, as
well as several additional disorders.
The differential diagnosis of mild thrombocytopenia may
be narrowed by determining whether it is new in onset or
chronic; thus, the results of prior blood counts from as far in
the past as possible should be vigorously sought. New onset
thrombocytopenia generally results from development of a
new disease process, with thrombocytopenia as either a
primary (eg, ITP) or secondary (eg, HIV or other infectious
process, bone marrow inltration, myelodysplasia) manifes-
tation. Chronic thrombocytopenia may suggest an inherited
process (eg, variant of an MYH9-related disorder, Mediter-
ranean thrombocytopenia).
Mild thrombocytopenia that is generally not clinically
important in terms of increasing bleeding risk may be sig-
nicant as a marker of an underlying disorder. Thus, the
patient should be questioned carefully for signs or symptoms
of infection, autoimmune disease, or malignancy, and the
physical examination should pay particular attention to the
assessment of lymphadenopathy, hepatosplenomegaly, skin
rashes, and musculoskeletal abnormalities.
Drug-induced thrombocytopenia has been discussed
previously in reference to acutely ill patients but should be
considered in patients with mild thrombocytopenia as well.
Although drugs such as sulfa-containing antibiotics often
cause severe thrombocytopenia, othersfor example, psy-
chotropic agentsmay cause a number of blood dyscra sias,
including mild thrombocytopenia. A listing of drugs asso-
ciated with thrombocytopenia and the strength of the
apparent association may be found at http://moon.ouhsc.
edu/jgeorge. However, when assessing a patient, it is often
reasonable to assume that most drugs may cause thrombo-
cytopenia; even those such as H2 blockers, which do so
only rarely, are so commonly used that their overall contri-
bution to cases of isolated thrombocytopenia may be
substantial.
As with any hematologic disorder, examination of the
peripheral blood lm is an essential part of the evaluation.
Clumped platelets may suggest pseudothrombocytopenia; if
this is observed, the platelet count should be repeated in a
citrated tube. Large platelets may be consistent with an
MYH9-related disorder, such as the MayHegglin anomaly;
neutrophil inclusions would support this diagnosis. Hyper-
segmented neutrophils and macrocytosis may suggest vita-
min B
12
deciency. Lymphocytosis may suggest underlying
Iron deciency is the most common cause of anemia during
pregnancy, and routine iron supplementation should be offered. In
pregnant patients with iron deciency who do not tolerate oral iron,
iron sucrose or ferric gluconate are the preferred parenteral agents.
Thrombocytopenia in a pregnant patient may result from a
number of different etiologies. The most common cause is
incidental thrombocytopenia of pregnancy, in which the
platelet count does not fall below 70,000/L. This disorder is
not associated with increased maternal or fetal morbidity.
ITP is the most common cause of isolated thrombocytopenia
during the rst trimester of pregnancy.
TTPHUS, preeclampsia, and HELLP may be very difcult to
distinguish from one another in some patients. If TTPHUS is suspected,
a trial of plasma exchange should be initiated immediately.
Levels of vWF and factor VIII generally increase during
normal pregnancy in patients with type I vWD, the most
common subtype. Levels should be monitored throughout
pregnancy to conrm the expected increase. However, levels of
vWF may decrease precipitously after delivery in some patients,
and supplementation with vWFFVIII containing products may
be required for up to 2 weeks.
Thrombophilia is associated with an increased risk of
thrombosis during pregnancy and in the puerperium. The
greatest risk is in the latter. The relative risks for heterozygous
thrombophilic conditions are small, but those for homozygous
or doubly heterozygous conditions are substantial. Guidelines
are available (Table 18-7) for management of pregnant
individuals with thrombophilia.
Thrombophilia is also associated with an increased incidence
of fetal loss and probably with pregnancy complications.
However, the role of anticoagulation in this setting has not been
well dened in large prospective studies.
VKAs are associated with embryopathy when administered
between weeks 6 and 12 of pregnancy. Use of these agents
during the second and third trimester may be associated with
neurodevelopmental abnormalities as well. Warfarin is not
present in high concentrations in breast milk, and mothers on
warfarin may breast-feed safely.
The optimal anticoagulation regimen for pregnant patients
with articial heart valves is uncertain, but guidelines favor the
use of adjusted-dose LMWH.
The most common hematologic malignancies diagnosed
during pregnancy are HD (50% of cases; incidence: 0.06
0.10% of pregnancies) and acute and chronic leukemias
(incidence: 0.001%). Among gestational leukemias, AML and
ALL account for 67% and 25% of cases, respectively.
The evaluation and therapeutic strategies for women with a
hematologic malignancy diagnosed during pregnancy depend
heavily on the aggressiveness of the disease, the stage of
pregnancy, the patients willingness to bear personal risk to
avoid therapy to protect her fetus, and the option of accepting
fetal risk to receive optimal management for the malignancy. In
all cases, delaying therapy until at least the second trimester
minimizes the risk of fetal complications and pregnancy loss.
Key points
Selected outpatient hematology topics
| 489
|
Consultative hematology 490
chronic lymphocytic leukemia, and circulating plasma cells
may be consistent with myeloma. Dysmorphic red cells,
hypogranulated neutrophils, or PelgerHuet cells may sug-
gest underlying myelodysplasia, which may occasionally pres-
ent with isolated thrombocytopenia.
There are no guidelines as to when, or whether, the bone
marrow should be examined. Although the incidence of pri-
mary marrow disorders such as myelodysplastic syndromes
increases with age, recent epidemiologic studies demonstrate
that ITP is also common in elderly patients. Many experts
advise that if the thrombocytopenia is isolated and there are
no other abnormalities on the peripheral blood lm or
detected on the physical examination, bone marrow exami-
nation is not immediately necessary, and the patient may be
followed with serial platelet counts. Should unexplained
symptoms arise, other hematologic abnormalities appear, or
the thrombocytopenia worsens, further evaluation should be
considered.
In a recent study, the outcomes of 191 otherwise healthy
individuals with incidentally discovered borderline throm-
bocytopenia (platelet count 100,000150,000/L) who had
been stable for 6 months were determined. With a median
observation time of 64 months, 64% of the cases of throm-
bocytopenia either resolved spontaneously or persisted with
no other disorders becoming apparent. The most common
pathologic event developing during follow-up was autoim-
mune disease. The 10-year probability of developing ITP was
6.9% and of developing autoimmune diseases other than
ITP was 12.0%.
Leukocytosis
Hematology consultation is occasionally requested for
asymptomatic leukocytosis. Classication of the complaint
depends on the absolute number and phenotype of the cells.
The consultant should begin with examination of freshly
performed blood counts and peripheral smear cell morphology.
In some cases, an immunophenotypic analysis by ow
cytometry may be necessary at the outset to characterize an
abnormal leukocyte (especially lymphocyte) population.
The history and clinical records are of key importance.
Progressive leukocytosis may accompany disorders associ-
ated with inammation, infection, infestation, malignancy,
trauma, or exposure to a foreign substance or drug.
Symptoms may not be evident. During the evaluation of a
reactive process, it may be helpful for the patient to keep a
diary to record chills, sweats, and periodic temperature
measurements.
Physical examination of patients with leukocytosis should
include a search for manifestations of acute or chronic
inammation, rash, lymphadenopathy, sinusitus, thyromegaly,
pulmonary disease, heart murmur, hepatomegaly, or evidence
of cirrhosis, splenomegaly, and arthropathy. Pulmonary
function tests and echocardiograpy should be considered to
assess for end-organ complications of hypereosinophilic
syndrome in the setting of unexplained eosinophilia. The
next step is to exclude occult diseases such as miliary tuber-
culosis, an abscess, parasitic infection, hypercortisolism,
occult malignancy, or a primary hematologic disorder. A
diagnostic bone marrow aspiration and biopsy with ow
cytometry and chromosomal analysis or specic molecular
studies may also be helpful.
Disorders that may prompt hematology consultation from
the medical, surgical, or obstetric clinics are listed in Table
18-11. Neutrophilia and monocytosis are further discussed
in Chapter 7, and hypereosinophilic syndrome is discussed
in Chapter 8.
Leukopenia
Consultation for leukopenia is usually obtained when the
peripheral blood white count is 3700/L. It is useful to
remember that there are racial and gender differences in the
normal distribution of leukocyte counts. Women have a
lower median and minimum than men. African Americans
and Yemenite Jews have a lower median and minimum neu-
trophil count than Caucasians. Asymptomatic individuals
from populations with a lower normal range and a benign
history do not need to be further evaluated.
The signicance of leukopenia relates to the clinical con-
text, the acuity, and the affected cell line. Isolated mild neu-
tropenia is of less clinical signicance than the combination
of the same degree of leukopenia with mild anemia and mild
thrombocytopenia. Acute severe neutropenia in the context
of an acute but otherwise mild viral or gram-negative bacte-
rial infection may be of no clinical importance and generally
resolves rapidly. In contrast, acute severe neutropenia in a
debilitated elderly individual who is septic or toxic is evi-
dence of a grim prognosis. The causes and management of
neutropenia in children and adults are reviewed extensively
in Chapter 7.
Lymphopenia in an infant with recurrent infections
should prompt investigations for a congenital immunode-
ciency syndrome. Common variable immunodeciency
may also present with recurrent infections and lymphope-
nia later in life. Acquired lymphopenia in a child or adult
is usually related to decreases in the circulating T cells sec-
ondary to an infection, inammatory disorder, advanced
age, or drug (Table 18-12). In some patients, persistent mild
lymphopenia is related to an idiopathic decrease in the
CD4