1. health and fitness
2. disorders

Seminars in Fetal & Neonatal Medicine (2008) 13, 256e264
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/siny
Neonatal thrombocytopenia
Irene Roberts a,*, Neil A. Murray b
a
Department of Haematology, Imperial College London and Imperial Healthcare NHS Trust,
Du Cane Road, London W12 0NN, UK
b
National Blood Service, John Radcliffe Hospital, Oxford, UK
KEYWORDS
Fetal
thrombocytopenia;
Neonatal alloimmune
thrombocytopenia
(NAIT);
Neonatal
thrombocytopenia;
Platelet transfusion
Summary Thrombocytopenia (platelets <150 109/L) is one of the most common haematological problems in neonates, particularly those who are preterm and sick. In those preterm
neonates with early thrombocytopenia who present within 72 h of birth, the most common
cause is reduced platelet production secondary to intrauterine growth restriction and/or maternal hypertension. By contrast, the most common causes of thrombocytopenia arising after
the first 72 h of life, both in preterm and term infants, are sepsis and necrotizing enterocolitis.
The most important cause of severe thrombocytopenia (platelets <50 109/L) is neonatal
alloimmune thrombocytopenia (NAIT), as diagnosis can be delayed and death or long-term disability due to intracranial haemorrhage may occur. Platelet transfusion is the mainstay of
treatment for severe thrombocytopenia. However, the correlation between thrombocytopenia
and bleeding is unclear and no studies have yet shown clinical benefit for platelet transfusion
in neonates. Studies to identify optimal platelet transfusion practice for neonatal thrombocytopenia are urgently required.
ª 2008 Elsevier Ltd. All rights reserved.
Introduction
Sick neonates often develop thrombocytopenia. Of neonates admitted to neonatal intensive care units (NICUs), the
platelet count drops below 150 109/L in one in four babies
and to below 50 109/L in one in twenty. Many clinical conditions are associated with neonatal thrombocytopenia (NT)
but, until recently, almost half of all cases were classified as
idiopathic. However, recent studies detailing the natural
history of NT have identified reduced platelet production
as the main underlying mechanism of many idiopathic cases.
This has led to newer classifications of NT based on the
timing of onset of thrombocytopenia (i.e. early versus late
* Corresponding author. Tel.: þ44 208 383 2163; fax: þ44 208 742
9335.
E-mail address: irene.roberts@imperial.ac.uk (I. Roberts).
NT; Table 1). These are more useful for neonatal clinicians
than previous classifications, with their focus on associations and rare disorders, and will help to facilitate systematic studies to improve the management of NT.
The need for systematic studies is emphasized by the
variation in platelet transfusion practice between centres,
with reported rates ranging from 2% to 9% for similar NICU
populations.1e3 In addition, it is important to note that no
study has yet shown clinical benefit of platelet transfusion
in NT. Indeed, recent reports suggest possible adverse
effects, including increased mortality in transfused versus
non-transfused neonates with a similar degree of NT,4 and
an increased risk of short bowel syndrome and/or cholestasis in transfused neonates surviving necrotizing enterocolitis (NEC).5 Furthermore, no study has yet defined the
magnitude of increased bleeding risk in sick neonates
developing severe NT.
1744-165X/$ - see front matter ª 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.siny.2008.02.004
Neonatal thrombocytopenia
Table 1
257
Classification of fetal and neonatal thrombocytopenias
Condition
Fetal
Alloimmune
Congenital infection (e.g. CMV, toxoplasma, rubella, HIV)
Aneuploidy (e.g. trisomies 18, 13, 21, or triploidy)
Autoimmune (e.g. ITP, SLE)
Severe Rhesus disease
Congenital/inherited (e.g. WiskotteAldrich syndrome)
Early-onset neonatal (<72 h)
Placental insufficiency (e.g. PET, IUGR, diabetes)
Perinatal asphyxia
Perinatal infection (e.g. Escherichia coli, Haemophilus influenzae, GBS)
DIC
Alloimmune
Autoimmune (e.g. ITP, SLE)
Congenital infection (e.g. CMV, toxoplasma, rubella, HIV)
Thrombosis (e.g. aortic, renal vein)
Bone marrow replacement (e.g. congenital leukaemia)
KasabacheMerritt syndrome
Metabolic disease (e.g. proprionic and methylmalonic acidaemia)
Congenital/inherited (e.g. TAR, CAMT)
Late-onset neonatal (>72 h)
Late-onset sepsis
NEC
Congenital infection (e.g. CMV, toxoplasma, rubella, HIV)
Autoimmune
KasabacheMerritt syndrome
Metabolic disease (e.g. proprionic and methylmalonic acidaemia)
Congenital/inherited (e.g. TAR, CAMT)
Bold type indicates the most common conditions. CAMT, congenital amegakaryocytic thrombocytopenia; CMV, cytomegalovirus; DIC, disseminated intravascular coagulation; GBS, group B Streptococcus; HIV, human immunodeficiency virus; ITP, idiopathic thrombocytopenic
purpura; IUGR, intrauterine growth restriction; NEC, necrotizing enterocolitis; PET, pre-eclampsia; SLE, systemic lupus erythematosus;
TAR, thrombocytopenia with absent radii.
This paper comments briefly on recent progress in
understanding the causes and mechanisms of NT, then
assesses whether a relation between platelet count and
bleeding can be defined in neonates. Finally developments
aimed at rationalizing platelet transfusion therapy for sick
neonates with severe NT are outlined.
Definition and prevalence of neonatal
thrombocytopenia
Fetal studies have shown that the platelet count reaches
150 109/L by the end of the first trimester of pregnancy,6
and is maintained at or above this level to term in healthy
fetuses.7e9 NT can therefore be defined as a platelet count
<150 109/L in any healthy neonate of a viable gestational
age. This definition is supported by large population studies
that show that >98% of term neonates born to mothers with
normal platelet counts have platelets above 150 109/L at
birth.10e13 Although some studies have reported a higher
prevalence of NT, the variation in prevalence largely
depends on the population studied, with higher rates seen
in populations containing high-risk neonates.14 In neonates
admitted to NICUs, for example, NT develops in 22e35% of
all admissions.15e17 The rate increases as gestational age
decreases14 and is higher in intrauterine growth restricted
(IUGR) neonates (see below). Severe thrombocytopenia
(platelets <50 109/L) is seen in w10% of cases of NT
present at birth. It is these neonates above all who need urgent investigation to identify the cause, particularly those
cases due to neonatal alloimmune thrombocytopenia
(NAIT) in which prompt treatment is necessary to prevent
severe long-term disability (see below).
Causes of neonatal thrombocytopenia
NT usually presents in one of two clinical patterns, which
reflect the most common causes: (1) early NT (within 72 h
of birth); or (2) late NT (after 72 h of life; see Table 1). Clinically, the most important cause of severe early NT is NAIT.
However, NAIT accounts for only a small proportion (<5%)
of early NT overall. The most frequent cause, particularly
in preterm infants, is that associated with chronic fetal
hypoxia, such as infants born to mothers with pregnancyinduced hypertension or diabetes and/or those with
IUGR.17,18 NT occurs in a large proportion of such infants,
although the thrombocytopenia is self-limiting. It usually
resolves within 10 days17,18 and is rarely severe (platelets
<50 109/L), except in neonates with severe IUGR. The
most common and clinically important causes of late NT
are sepsis and NEC, which together account for >80% of
cases.1,19 This form of NT usually develops very rapidly
over 1e2 days, is often very severe (platelets <30 109/L)
and takes 1e2 weeks to recover.1 Such babies frequently
require repeated platelet transfusion.1
258
Neonatal thrombocytopenia secondary to
chronic fetal hypoxia
Neonates with IUGR secondary to chronic fetal hypoxia, and
those born to mothers with pregnancy-induced hypertension or diabetes, have a number of distinct haematological
abnormalities that are present at birth. These include
varying degrees of NT, transient neutropenia and increased
erythropoiesis (high numbers of circulating nucleated red
cells with or without associated polycythaemia); many also
have evidence on the blood film of hyposplenism (spherocytes, target cells and HowelleJolly bodies).18 Erythropoietin levels are increased and the severity of the
haematological abnormalities correlates both with serum
erythropoietin levels and with the severity of the placental
dysfunction consistent with fetal hypoxia as a direct cause
of the abnormalities.18,20 We and others have shown that
megakaryocytopoiesis is severely impaired at birth in such
neonates, as shown by a marked reduction in circulating
megakaryocytes and their precursor cells.1,18,21
Neonatal alloimmune thrombocytopenia
In NAIT, the platelet equivalent of haemolytic disease of
the newborn, thrombocytopenia results from transplacental passage of maternal platelet-specific antibodies to
paternal human platelet antigens (HPA) expressed on fetal
platelets that the mother lacks. Sixteen HPAs have been
identified but fetomaternal incompatibility between only
three (HPA-1a, HPA-5b and HPA-15b), singly or in combination, causes 95% of cases in Caucasian populations.22 Other
antibodies, such as anti-HPA-3a, are occasionally involved.23,24 Fetomaternal incompatibility for HPA-1a is
responsible for 75% of cases22e24 and occurs in 1:350 pregnancies, although thrombocytopenia develops in only
1:1000e1500 pregnancies.22 Recent data indicate that the
ability of an HPA-1a-negative woman to form anti-HPA-1a
is controlled by the HLA DRB3*0101 allele such that HLA
DRB3*0101-positive women are 140 times more likely than
HLA DRB3*0101-negative women to make anti-HPA-1a,25
thereby explaining the frequency of the clinical problem.
NAIT occurs in the first pregnancy in almost 50% of cases.
Thrombocytopenia is frequently extremely severe (platelet
count <20 109/L) and commonly results in major bleeding,
particularly intracranial haemorrhage (ICH). The incidence
of ICH is difficult to ascertain precisely, but large series report
its occurrence in 10e20% of untreated pregnancies.22,26,27 As
thrombocytopenia is frequently present in the fetus, often as
early as 20e24 weeks gestation,28,29 ICH might occur during
fetal development,30,31 sometimes resulting in fetal death.
Fetal ICH is most common in untreated pregnancies, in which
it might account for up to 75% of all the fetuses and neonates
developing such bleeds.24,32 The neonatal course in otherwise well neonates is variable, with thrombocytopenia
resolving in most cases within 1 week (with or without platelet transfusion) with no long-term sequelae. However,
thrombocytopenia can last for several weeks before resolving, and in such cases it requires repeated platelet transfusion. In a minority of affected neonates ICH occurs for the
first time following birth, adding to the overall mortality
and morbidity of the condition.
I. Roberts, N.A. Murray
Neonatal ICH is most common in neonates whose mothers
received no antenatal therapy.32 The neurodevelopmental
outcome of neonates with untreated NAIT is poorer than
their siblings where maternal treatment was instituted.33
However, the poorest outcome is seen in those neonates
with ICH; two-thirds of such cases show neurodevelopmental problems, approximately half of which are severe, e.g.
severe cerebral palsy and/or sensory impairment.22
The severe fetal and neonatal consequences of NAIT
mean that this disorder requires expert management with
close collaboration between experienced fetal medicine
specialists, haematologists and neonatologists. The guiding
principle of therapy has been the knowledge that for mothers
with known HPA antibodies, the fetal and neonatal course in
subsequent pregnancies (with an antigen-positive fetus)
closely reflects that in previously affected pregnancies.
Thus, mothers with previous neonates suffering an ICH are
at high risk of future children also having a severe course.
Antenatal management of neonatal alloimmune
thrombocytopenia
Antenatal management of pregnancies at risk of NAIT remains
controversial. Three general approaches are used in different
centres (these are reviewed in refs. [24,34] and [35]). First,
an invasive approach, focused on ‘high-risk’ mothers with
previous severely affected children. This is based on repeated
fetal blood samples (FBS) and intrauterine transfusions (IUT)
of HPA-compatible platelets in thrombocytopenic fetuses,
combined with preterm delivery at 32e34 weeks. Second,
a non-invasive approach based on monitoring by fetal ultrasound scan (USS) and maternal intravenous weekly highdose intravenous immunoglobulin therapy (IVIG) þ steroids;
this is used most often in ‘low-risk’ women. Third, combination therapy based on high-dose maternal IVIG together
with infrequent FBS to monitor the response to therapy during
pregnancy and to decide on the mode of delivery.
It has recently become clear that the rate of fetal loss
and emergency preterm delivery associated with repeated
FBS and IUT approximates to the rate of fetal ICH in
untreated pregnancies and exceeds that in ‘low-risk’
pregnancies treated with IVIG.22,36,37 As a result, therapy
for NAIT, even in high-risk cases, is increasingly moving
towards the non-invasive approach.
Recently, van den Akker et al. reported on their
experience of treating 52 pregnant women (five with
a previous sibling with ICH) with known HPA incompatibility
with IVIG alone at a dose of 1 g/kg per week.38 IVIG was
begun at 16 weeks if the previous sibling had an ICH and
32 weeks if not. All the pregnancies resulted in live births;
there were no ICHs and no neonatal deaths. As a result,
these researchers have adopted this non-invasive regimen
as their standard protocol for antenatal therapy for all
mothers with known HPA incompatibilities.34
Management of affected neonates with neonatal
alloimmune thrombocytopenia
Clinical symptoms range from asymptomatic neonates with
incidental thrombocytopenia, to limited or widespread skin
petechiae or purpura, to symptoms of ICH (e.g. neonatal
Neonatal thrombocytopenia
259
seizures). The diagnosis of NAIT is made by demonstrating
platelet antigen incompatibility between mother and baby
serologically or by PCR, and is carried out in reference
transfusion labs. The most important aspect of management
of new cases is to consider NAIT as a possible diagnosis in any
case of unexpected severe NT presenting at birth, particularly in term neonates. As the vast majority of affected
neonates (w80%) will not have suffered an ICH before birth,
preventing ICH during the period of thrombocytopenia
should be considered a neonatal emergency. Although this
might appear straightforward, the best way of achieving this
is unclear. A ‘safe’ platelet count has not been identified for
neonates with NAIT; it is not clear what threshold should be
used to trigger platelet transfusion and what regimen should
be used to achieve and maintain a ‘safe’ count.39 This situation is further complicated by recent evidence that
suggests that neonates with HPA-5b incompatibility might
be at risk of bleeding at higher platelet counts than that
seen with other HPA incompatibilities.22
Table 2
First presentation of neonatal alloimmune
thrombocytopenia
Where NAIT is suspected and there is no prior family
history, it is prudent to consider neonates with a platelet
count <50 109/L as being at high risk of ICH and to aim to
maintain the platelet count above 50 109/L for the first
2 weeks of life (Table 2). All babies with severe thrombocytopenia due to NAIT should have a cranial USS to look for
evidence of ICH. The treatment of choice for NAIT is transfusion with HPA-compatible platelets (available ‘off the
shelf’ from transfusion centres in the UK and many other
European countries) as these produce the best platelet
increment and longest half-life of transfused platelets.40
Alternatives are random donor platelet transfusion, IVIG,
steroids, an expectant approach monitoring the count until
thrombocytopenia resolves and washed maternal platelets.
Where there is a delay in obtaining HPA-compatible platelets, random donor platelet transfusions can be used in an
emergency as they can produce a significant platelet
Guidelines for platelet transfusion thresholds for neonates
Platelet
count
(109/L)
Non-bleeding
neonate
(first week of life)
Non-bleeding
neonate (week
2 onwards)
Neonate with
major bleeding
Auto-IT
NAIT
(new case
suspected)
NAIT
(known case)
<30
Transfuse
all patients
Transfuse
all patients
Transfuse
Transfuse
if bleeding
present or
IVIG
unavailable
Transfuse
using HPA
compatible
platelets
30e49
Do not transfuse
if clinically stable.
Transfusion
appropriate if:
<1000 g and < 1
week of age;
clinically unstable
(e.g. high ventilation
requirements or
fluctuating
BP/hypovolaemia);
previous major
bleeding tendency
(e.g. grade 3 or 4
IVH); concurrent
coagulopathy
surgery or exchange
transfusion
Do not transfuse
Do not
transfuse
Transfuse
Do not
transfuse
if stable and
not bleeding
Transfuse
using HPA-1a/5b
negative platelets
(random donor
platelets only if
compatible
platelets
unavailable)
Transfuse using
HPA-1a/5b
negative platelets
(random donor
platelets only if
compatible
platelets
unavailable)
Do not
transfuse
Transfuse
Do not
transfuse
50e99
Transfuse using
HPA-1a/5b
negative platelets
(major bleeding)
Transfuse
using HPA
compatible
platelets
(minor
bleeding)
Transfuse
using HPA
compatible
platelets
(major
bleeding)
Auto-IT, autoimmune thgrombocytopenia; BP, blood pressure; HPA, human platelet antigen; IVH, intravascular haemorrhage; IVIG,
intravenous immunoglobulin; NAIT, neonatal alloimmune thrombocytopenia.
260
increment in NAIT.41 The role of IVIG is less clear. The available literature suggests that any platelet increment following IVIG is likely to be delayed for 12e36 h,26,29 making it
unacceptable as a monotherapy. In addition, a recent study
has suggested that increases in platelet count following
IVIG therapy in some neonates with NAIT might be confused
with natural platelet count recovery questioning the true
effectiveness of neonatal IVIG therapy.32 In new cases without major bleeding and with an initial count >50 109/L,
an expectant approach is appropriate. However, thrombocytopenia often worsens over the first few days, making
close monitoring of the platelet count essential. In affected
neonates with active bleeding (e.g. new or worsening ICH,
gastrointestinal, frank haematuria) it seems reasonable to
maintain the platelet count above 100 109/L (Table 2).
I. Roberts, N.A. Murray
Known cases of neonatal alloimmune thrombocytopenia
Neonates born after treatment with antenatal maternal
IVIG appear to have a reduced risk of major haemorrhage,22,32 although some will still be born with platelet
counts <30 109/L. These neonates are also likely to be
born in specialist centres, the HPA-incompatibility will be
known and delivery can be planned with HPA-compatible
platelets immediately available. It is possible that these
neonates can be maintained with lower platelet counts
and a platelet threshold of 30 109/L rather than
50 109/L, as they appear to have a reduced risk of ICH
(see Table 2). However, this is not yet clear and awaits
evaluation by carefully designed clinical trials.
thrombocytopenia. Neonates with NAIT, for example, have
a high risk of bleeding, possibly because some anti-HPA
antibodies interfere with platelet function, although this is
unproven as yet.24 Similarly, clinical experience suggests
that thrombocytopenic neonates with IUGR have a low
risk of major haemorrhage and those with sepsis or NEC
an intermediate risk, perhaps reflecting differences in
platelet function or co-existing coagulopathy. Although
these observations influence our clinical practice, there is
very little evidence that allows us to measure the risk of
bleeding accurately, at a particular platelet count, in an
individual neonate or group of neonates (e.g. those with
IUGR). Indeed, no neonatal studies have addressed this
question and none has compared different groups of thrombocytopenic neonates by underlying mechanism of thrombocytopenia. We therefore recently carried out a detailed
observational study in 169 neonates with severe NT in
whom the prevalence and nature of major and minor haemorrhage was recorded according to the cause of the
thrombocytopenia (Fig. 1). In this study we found that neonates with NT secondary to IUGR or maternal pregnancyinduced hypertension formed the most common diagnosis
in the group of babies with no major haemorrhage (38%) despite severe NT, whereas they formed a much smaller proportion of the group of babies who did develop major
haemorrhage. Severe sepsis and NEC were the most common diagnoses in those neonates who did bleed, suggesting
that these disorders do affect both platelet number and
function.
Neonatal autoimmune thrombocytopenia
Neonatal platelet transfusion
NT secondary to transplacental passage of maternal platelet autoantibodies occurs in babies born to mothers with
idiopathic thrombocytopenic purpura (ITP) or systemic
lupus erythematosus (SLE). Around 10% of infants of
affected mothers develop thrombocytopenia. It is a less
common cause of NT than NAITP, affecting 1e5 in 10,000
pregnancies.42,43 Thrombocytopenia is usually mild and ICH
is rare (<1% of at-risk babies). In affected babies with
severe thrombocytopenia treatment with IVIG 1 g/kg for
2 days is usually effective (see Table 2).44
Platelet transfusion is the only specific therapy available to
treat NT. However, no neonatal trials have yet evaluated
whether platelet transfusion reduces haemorrhage or
improves outcome in NT. In the only randomized controlled
trial addressing this issue, Andrew et al.46 found no benefit
in terms of haemorrhage of maintaining a normal platelet
count (platelets >150 109/L) by platelet transfusion in
preterm neonates compared to controls with moderate
thrombocytopenia (platelets 50e150 109/L). This trial
was reported almost 15 years ago and neonates with platelet counts <50 109/L were transfused and excluded from
analysis. The relevance of this study to modern neonatal
practice is therefore limited, both because of considerable
changes in practice since the early 1990s and because the
vast majority of platelet transfusions are now given to
neonates with platelet counts <50 109/L.
The lack of clinical trial data in neonates has led to
uncertainties regarding optimum transfusion therapy, with
the result that NICUs often have quite different platelet
transfusion protocols. Three publications1e3 retrospectively
document platelet transfusion practice in three separate
NICUs in the USA, UK and Mexico. Despite the geographical
differences, several consistent observations are apparent in
these reports:
Inherited thrombocytopenia
A number of rare inherited disorders present with thrombocytopenia in the fetus or neonate (see Table 1). In most
cases, the thrombocytopenia is due to reduced platelet
production secondary to abnormal haemopoietic stem cell
development and there are often associated congenital
anomalies, which are useful in guiding investigations and
establishing the diagnosis. Many of the recent advances in
identifying the molecular basis of a number of these disorders have informed both the diagnosis and the management
of neonates with unexplained, persistent NT (reviewed in
ref. [45]).
Neonatal thrombocytopenia and bleeding risk
From clinical experience it is clear that different neonates
have very different risks of bleeding for the same degree of
Different platelet count triggers for transfusion are used
by each unit. As a result, platelet transfusion rates vary
considerably between units: from 2% to 9.4% of admissions (some of this variation is explained because more
transfusions are given in NICUs with a high percentage
Neonatal thrombocytopenia
261
Neonates admitted to NICUs
n = 3498
Platelets < 60 109/L during
admission
Entered into study
n = 169
Major intracranial haemorrhage
before thrombocytopenia started
n = 26
No major intracranial haemorrhage
before thrombocytopenia started
n = 143
Received platelet Tx: 20/26 (77 )
Mean Tx per patient: 3
Mean Tx per patient transfused: 3.9
Major haemorrhage during
thrombocytopenic episodes
n = 19
Received platelet Tx: 79/124 (64 )
Mean Tx per patient: 1.8
Mean Tx per patient transfused: 2.8
Received platelet Tx: 17/19 (89 )
Mean Tx per patient: 6.3
Mean Tx per patient transfused: 7
On study mortality
6/19 (32 )
On study mortality
9/124 (7 )
On study mortality
10/26 (38 )
Figure 1 Major haemorrhage, platelet transfusions and mortality in 169 neonates with platelet count <60 109/L. Tx,
transfusions.
of intensive care patients and in those units practising
extracorporeal membrane oxygenation (ECMO)).
The majority of platelet transfusions are given prophylactically to non-bleeding neonates.
Most transfused neonates receive only one platelet
transfusion.
Thrombocytopenic neonates who receive platelets are
up to 10 times more likely to die than neonates who
do not receive platelet transfusion.
Evidence of possible harmful effects of
neonatal platelet transfusion
The excess mortality seen in thrombocytopenic neonates
remains an area of considerable interest. Although the
quality of current evidence is low, it is reasonable to
believe that thrombocytopenic neonates are at increased
risk of haemorrhage. It therefore seems counterintuitive
that platelet transfusions might harm these neonates.
However, evidence is accumulating to suggest that
repeated platelet transfusion can be harmful, particularly
in sick neonates. First, neonates who receive platelet
transfusions have a higher mortality than non-transfused
neonates.1e3 Although this might be due to the severity of
the conditions causing severe thrombocytopenia (e.g. sepsis and NEC) rather than a direct effect of platelet transfusion, the specific contribution of platelet transfusion has
not been properly evaluated. In a retrospective study of
46 neonates (mean gestational age 28 weeks) with Bell’s
stage 2 or stage 3 NEC, Kenton et al.5 found no improvement in either mortality or morbidity with increasing number or volume of platelet transfusions. However, study
neonates who developed short bowel syndrome and cholestasis had been given a significantly higher number and volume of platelet transfusions than those who did not
develop these morbidities. Recently, Bonifacio et al.4
looked more closely at the relationship between illness
severity, thrombocytopenia and platelet transfusion in preterm neonates <28 weeks gestational age with platelet
counts <100 109/L and found a considerably higher mortality in neonates who had received platelet transfusions
compared to non-transfused neonates.
The mechanism of a detrimental effect of platelet
transfusion in neonates has not been investigated. For
neonates with sepsis or NEC, it seems reasonable to suggest
that repeated transfusion of platelets might perpetuate the
activation of inflammatory and/or coagulation cascades
seen in these conditions, contributing to a poorer outcome
after transfusion. More specifically, repeated platelet
262
transfusion might perpetuate the coagulative necrosis seen
within the bowel of neonates with NEC, increasing the risk
of death or the development of short bowel syndrome.
To make informed clinical decisions about the role and
regimen of platelet transfusion, particularly for sick neonates, who might suffer deleterious consequences of such
transfusions, detailed studies of NT and platelet transfusion
are required. It is only by critically evaluating the relationship between NT and haemorrhage (and the effects of
platelet transfusion on both of these indices) that controlled clinical trials of neonatal platelet transfusion can be
safely carried out. Our group therefore recently completed
a large, prospective, observational study of NICU patients
developing platelet counts <60 109/L (unpublished data)
by documenting practice in seven different NICUs over a period of 18 months between 2005 and 2006. During this
period, the units admitted 3498 neonates, of whom 194
developed a qualifying platelet count; 169 of these were
enrolled in the study (see Fig. 1). Each had daily assessments of bleeding and all platelet counts and platelet
transfusions given were recorded, including the reason for
transfusion. This study therefore provides the first precise
information on severe NT, associated major bleeding and
its therapy by platelet transfusion. The major findings of
the study were:
I. Roberts, N.A. Murray
thrombocytopenia who receive multiple transfusions (typically those with uncontrolled sepsis or NEC) have a poor
outcome and a high mortality. This divides affected
neonates into two major groups with at least two separate
clinical questions. Does the single transfusion given to the
neonates with transient thrombocytopenia improve outcome or is it unnecessary therapy? Do the multiple transfusions given to neonates with prolonged thrombocytopenia
improve outcome or is this therapy actually detrimental to
outcome? We are currently designing randomized controlled trials to answer these important clinical questions.
Developing guidelines for platelet transfusion
in neonates
A number of platelet transfusion guidelines for the newborn
have been proposed. However, given the uncertainties
discussed above, it remains a difficult task to construct
practical guidelines that encompass the clinical variations
in neonates with NT. We have used the three recent clinical
reports,1e3 plus our own recent survey, to develop guidelines for platelet transfusion in different groups of sick
neonates (Table 2).
Haemorrhage in sick neonates
NT with counts <60 109/L occurred in 5.5% of NICU
admissions, the incidence rising with decreasing gestational age to 37% in neonates <26 weeks gestation.
The first period of thrombocytopenia typically occurred
in the first week of life, at a median of 4 days of age,
and was of short duration (median 4 days).
Major haemorrhage occurred in 13% of affected neonates [intravascular (IVH) 53%, pulmonary 26%, renal
11%, other 10%] and of these 84% were <30 weeks gestational age and 42% were <26 weeks.
Platelet transfusions were given to 69% of affected
neonates with a median of three transfusions per neonate, although 66% of those transfused received only
one transfusion.
The platelet thresholds triggering prophylactic platelet
transfusion in non-bleeding neonates varied in the
participating units from 20 to 40 109/L. The most
common reason for platelet transfusion (88%) was
a count below the unit protocol threshold, whereas
clinical bleeding triggered only 6% of transfusions.
Mortality during NT varied both with incidence of major
haemorrhage (32% in those with major haemorrhage
versus 9% in those without) and with the number of
platelet transfusions (35% in those receiving >4 transfusions versus 6% in those receiving <1 transfusion).
These prospectively collected data clearly demonstrated for the first time many of the characteristics of
NT that need to be considered in the design of safe, useful
randomized controlled trials of platelet transfusion in
neonates. In particular, it is pertinent that the majority
of neonates who develop severe NT have a short-lived
course, receive no or a single platelet transfusion, have
a low incidence of major haemorrhage and have a low
mortality. However, those neonates with prolonged
As shown in our recent study, most neonates who bleed
(particularly those with IVH) do so in the first days of life.
However, with the exception of perinatal asphyxia, the
conditions precipitating the majority of episodes of severe
NT (late-onset sepsis and NEC) usually develop after the first
few days of life. However, major haemorrhage outside the
period where IVH is common (first week) is relatively rare
even in severely thrombocytopenic neonates. Thus, we
believe that prophylactic platelet transfusions are not
required outside the first week of life for thrombocytopenic
neonates of any gestational age until the platelet count falls
below 30 109/L (see Table 2). (Indeed, it is likely that
many neonates might not be at significant risk of haemorrhage until the platelet count falls much lower.) Within
the first week of life, firm recommendations are more difficult. The only previous randomized controlled trial46
suggested that a platelet count of 50 109/L and above
was a safe level in preterm neonates and many neonatologists have adopted this cut-off in practice. Until high-quality
trial evidence is available, we recommend that for patients
in the first week of life with the greatest risk of haemorrhage (e.g. unstable extremely preterm neonates), prophylactic platelet transfusions using trigger thresholds up to
50 109/L represent acceptable and safe clinical practice.
When such neonates have progressed to the second week of
life and have achieved greater clinical stability, then their
platelet transfusion trigger threshold can be reduced to
30 109/L (see Table 2). Platelet transfusions in neonates
with platelet counts >50 109/L should be reserved for
patients with active major bleeding (i.e. IVH, new or extension, pulmonary, gastrointestinal, renal) as there is no
evidence that higher platelet counts are of any benefit in
non-bleeding neonates. Neonates with suspected or proven
NAIT require individualized protocols as above (see Table 2).
Neonatal thrombocytopenia
Platelet dose and product
No trial evidence is currently available regarding the
optimal volume (dose) of platelets to administer or
when to administer further transfusions. Larger volumes
(20 mL/kg) appear to result in larger and more sustained
rises in platelet count than smaller volumes (10 mL/kg),
and are generally well tolerated (personal observations).
As most neonates receive only one platelet transfusion it
seems prudent to ensure that this results in a good platelet
increment by using a large volume strategy to minimize donor exposure. As highlighted in our study, a small number of
neonates whose thrombocytopenia is the result of marked
platelet consumption (e.g. NEC) might show no measurable
response to repeated platelet transfusions. It is possible
that such transfusions are having a positive effect at a microvascular level not reflected by the circulating platelet
count. However, the accumulating evidence that repeated
platelet transfusion might be detrimental to outcome in
such neonates only adds to the difficulty for neonatologists
in prescribing logical and effective therapy given the
current paucity of trial evidence.
For guidance on appropriate platelet products for transfusion in neonates, the reader is referred to the British
Committee for Standard in Haematology website.47
263
uncommon even in severely thrombocytopenic
neonates.
Although the mainstay of treatment of neonatal
thrombocytopenia is platelet transfusion, the correlation between thrombocytopenia and bleeding
is unclear and no studies have yet shown clinical
benefit of platelet transfusion in neonates; studies
to identify optimal neonatal platelet transfusion
practice are an urgent priority.
Research agenda
To define more clearly the safe lower limit for
platelet counts within the different groups of neonates in tandem with defining which neonates will
benefit from platelet transfusion for their
thrombocytopenia.
To ensure accurate diagnosis and to determine the
most effective fetal and neonatal therapy for
NAIT, the most common cause of unexpected neonatal mortality and morbidity associated with NT.
Conclusions
Neonatal thrombocytopenia remains a common clinical
problem. Fortunately, most episodes of thrombocytopenia
are mild or moderate and resolve spontaneously without
apparent clinical sequelae. For more severe episodes, the
recent demonstration of impaired megakaryocytopoiesis
and platelet production as a major contributor to NT is an
important advance both for our understanding of the
underlying disease processes and the potential for improved
therapies. Clinical studies are now critically exploring how
changes in platelet count relate to the different causes of
NT and to the role of platelet transfusion in the management
of the different thrombocytopenic syndromes.
Practice points
Neonatal thrombocytopenia is defined as a platelet
count <150 109/L in neonates of any viable gestational age.
The commonest cause of thrombocytopenia developing within 72 h of birth is intrauterine growth restriction and/or maternal hypertension; this form
of thrombocytopenia is self-limiting and rarely
severe.
The commonest cause of severe thrombocytopenia
developing within 72 h of life is NAIT; the major
complication is intracranial haemorrhage which is
associated with long-term neurodevelopmental
problems in two-thirds of cases.
Our recent study indicates that most neonates who
bleed (particularly those with IVH) do so in the first
days of life; major haemorrhage after this is
References
1. Murray NA, Howarth LJ, McCloy MP, et al. Platelet transfusion
in the management of severe thrombocytopenia in neonatal
intensive care unit patients. Transfus Med 2002;12:35e41.
2. Garcia MG, Duenas E, Sola MC, et al. Epidemiologic and
outcome studies of patients who received platelet transfusions
in the neonatal intensive care unit. J Perinatol 2001;21:
415e20.
3. Del Vecchio A, Sola MC, Theriaque DW, et al. Platelet transfusions in the neonatal intensive care unit: factors predicting
which patients will require multiple transfusions. Transfusion
2001;41:803e8.
4. Bonifacio L, Petrova A, Nanjundaswamy S, Mehta R. Thrombocytopenia related neonatal outcome in preterms. Indian
J Pediatr 2007;74:269e74.
5. Kenton AB, Hegemier S, Smith EO, et al. Platelet transfusions
in infants with necrotizing enterocolitis do not lower mortality
but may increase morbidity. J Perinatol 2005;25:173e7.
6. Pahal G, Jauniaux E, Kinnon C, Thrasher AJ, Rodeck CH. Normal
development of human fetal hematopoiesis between eight and
seventeen weeks’ gestation. Am J Obstet Gynecol 2000;183:
1029e34.
7. Holmberg L, Gustavii B, Jonsson A. A prenatal study of fetal
platelet count and size with application to the fetus at risk
of Wiskott Aldrich syndrome. J Pediatr 1983;102:773e81.
8. Forestier F, Daffos F, Galacteros F. Haematological values of
163 normal fetuses between 18 and 30 weeks of gestation.
Pediatr Res 1986;20:342e6.
9. Forestier F, Daffos F, Catherine N, Renard M, Andreux JP.
Developmental hematopoiesis in normal human fetal blood.
Blood 1991;77:2360e3.
10. Burrows RF, Kelton JG. Incidentally detected thrombocytopenia in healthy mothers and their infants. N Engl J Med 1988;
319:142e5.
264
11. Burrows RF, Kelton JG. Thrombocytopenia at delivery: a prospective survey of 6715 deliveries. Am J Obstet Gynecol
1990;162:731e4.
12. Burrows RF, Kelton JG. Fetal thrombocytopenia and its relation
to maternal thrombocytopenia. N Engl J Med 1993;329:
1463e6.
13. Sainio S, Jarvenpaa A-S, Renlund M, et al. Thrombocytopenia in
term infants: a population-based study. Obstet Gynecol 2000;
95:441e6.
14. Christensen RD, Henry E, Wiedmeier SE, et al. Thrombocytopenia among extremely low birth weight neonates: data from
a multihospital healthcare system. J Perinatol 2006;26:
348e53.
15. Castle V, Andrew M, Kelton J, et al. Frequency and mechanism
of neonatal thrombocytopenia. J Pediatr 1986;108:749e55.
16. Mehta P, Rohitkumar V, Neumann L, Karpatkin M. Thrombocytopenia in the high risk infant. J Pediatr 1980;97:791e4.
17. Murray NA, Roberts IAG. Circulating megakaryocytes and their
progenitors in early thrombocytopenia in preterm neonates.
Pediatr Res 1996;40:112e9.
18. Watts TL, Roberts IAG. Haematological abnormalities in the
growth-restricted infant. Semin Neonatol 1999;4:41e54.
19. Roberts IAG, Murray NA. Neonatal thrombocytopenia: new
insights into pathogenesis and implications for clinical management. Curr Opin Pediatr 2001;13:16e21.
20. Salvesen DR, Brudenell JM, Snijders RJ, Ireland RM,
Nicolaides KH. Fetal plasma erythropoietin in pregnancies
complicated by maternal diabetes mellitus. Am J Obstet Gynecol 1993;168:88e94.
21. Watts TL, Murray NA, Roberts IAG. Thrombopoietin has a primary role in the regulation of platelet production in preterm
babies. Pediatr Res 1999;46:28e32.
22. Ghevaert C, Campbell K, Walton J, et al. Management and
outcome of 200 cases of fetomaternal alloimmune thrombocytopenia. Transfusion 2007;47:901e10.
23. Ouwehand WH, Smith G, Ranasinghe E. Management of severe
alloimmune thrombocytopenia in the newborn. Arch Dis Child
Fetal Neonatal 2000;82:F173e5.
24. Bussel JB, Primiani A. Fetal and neonatal alloimmune thrombocytopenia: progress and ongoing debates. Blood Rev 2008;22:
33e52.
25. Williamson LM, Hackett G, Rennie J, et al. The natural history
of fetomaternal alloimmunization to the platelet-specific antigen HPA-1a (PlA1, Zwa) as determined by antenatal screening.
Blood 1998;92:2280e7.
26. Mueller-Eckhardt C, Kiefel V, Grubert A, et al. 348 cases of suspected neonatal alloimmune thrombocytopenia. Lancet 1989;
1:363e6.
27. Bussel JB, Zacharoulis S, Kramer K, et al. Clinical and diagnostic comparison of neonatal alloimmune thrombocytopenia to
non-immune cases of thrombocytopenia. Pediatr Blood Cancer
2005;45:176e83.
28. Giovangrandi Y, Daffos F, Kaplan C, et al. Very early intracranial haemorrhage in alloimmune fetal thrombocytopenia. Lancet 1990;336:310.
I. Roberts, N.A. Murray
29. Bussel JB, Zabusky MR, Berkowitz RL, McFarland JG. Fetal alloimmune thrombocytopenia. N Engl J Med 1997;337:22e6.
30. Singh SA, Pollard J, Singhal N. Fetomaternal alloimmune
thrombocytopenia presenting as intracerebral bleeding in
utero. Indian J Pediatr 2005;72:269.
31. Sharif U, Kuban K. Prenatal intracranial hemorrhage and neurologic complications in alloimmune thrombocytopenia. J Child
Neurol 2001;16:838e42.
32. te Pas AB, Lopriore E, van den Akker ES, et al. Postnatal management of fetal and neonatal alloimmune thrombocytopenia:
the role of matched platelet transfusion and IVIG. Eur J Pediatr
2007;166:1057e63.
33. Ward MJ, Pauliny J, Lipper EG, Bussel JB. Long-term effects of
fetal and neonatal alloimmune thrombocytopenia and its antenatal treatment on the medical and developmental outcomes
of affected children. Am J Perinatol 2006;23:487e92.
34. van den Akker ES, Oepkes D. Fetal and neonatal alloimmune
thrombocytopenia. Best Pract Res Clin Obstet Gynaecol 2008;
22:3e14.
35. Murphy MF, Bussel JB. Advances in the management of alloimmune thrombocytopenia. Br J Haematol 2007;136:366e78.
36. Overton TG, Duncan KR, Jolly M, Letsky E, Fisk NM. Serial aggressive platelet transfusion for fetal alloimmune thrombocytopenia: platelet dynamics and perinatal outcome. Am J
Obstet Gynecol 2002;186:826e31.
37. Berkowitz RL, Kolb EA, McFarland JG, et al. Parallel randomized trials of risk-based therapy for fetal alloimmune thrombocytopenia. Obstet Gynecol 2006;107:91e6.
38. van den Akker ES, Oepkes D, Lopriore E, Brand A, Kanhai HH.
Noninvasive antenatal management of fetal and neonatal alloimmune thrombocytopenia: safe and effective. Br J Obstet
Gynaecol 2007;114:469e73.
39. Bassler D, Greinacher A, Okascharoen C, et al. A systematic
review and survey of the management of unexpected neonatal
alloimmune thrombocytopenia. Transfusion 2008;48:92e8.
40. Allen D, Verjee S, Rees S, Murphy MF, Roberts DJ. Platelet
transfusion in neonatal alloimmune thrombocytopenia. Blood
2007;109:388e9.
41. Kiefel V, Bassler D, Kroll H, et al. Antigen-positive platelet
transfusion in neonatal alloimmune thrombocytopenia (NAIT).
Blood 2006;107:3761e3.
42. Kelton JG. Idiopathic thrombocytopenic purpura complicating
pregnancy. Blood Rev 2002;16:43e6.
43. Bussel JB. Immune thrombocytopenia in pregnancy: autoimmune and alloimmune. J Reprod Immunol 1997;37:35e61.
44. Gill KK, Kelton JG. Management of idiopathic thrombocytopenic purpura in pregnancy. Semin Hematol 2000;37:275e89.
45. Alter BP. Diagnosis, genetics, and management of inherited
bone marrow failure syndromes. Hematol Am Soc Hematol
Educ Prog; 2007:29e39.
46. Andrew M, Vegh P, Caco C. A randomized, controlled trial of
platelet transfusions in thrombocytopenic premature infants.
J Pediatr 1993;123:285e91.
47. British Committee for Standard in Haematology, Available
online at: http://www.bcshguidelines.com.