Advances in Pediatrics 55 (2008) 255269

ADVANCES IN PEDIATRICS
Advances and Controversies
in Neonatal ICU Platelet
Transfusion Practice
Robert D. Christensen, MDa,b,*
a

Neonatology, Intermountain Healthcare, 36 S. State Street, Salt Lake City, UT, USA
Intermountain Healthcare, Urban North Region, McKay-Dee Hospital Center, NICU, 4300
Harrison Boulevard, Ogden, UT 84403, USA

hrombocytopenia is one of the most common hematopathologic findings

among patients in neonatal ICUs (NICUs) [14]. Up to 30% of NICU
patients have a low platelet count detected at some point during their hospital stay [58]. Among extremely low birth weight (ELBW) neonates (<1000 g),
thrombocytopenia is even more common, with approximately 75% having
thrombocytopenia recognized at some time during their NICU course [4]. As
shown in Fig. 1, the smallest ELBW neonates, those weighing less than 750 g
at birth, have an 85% to 90% incidence of developing thrombocytopenia.
For certain varieties of neonatal thrombocytopenia, intravenous immunoglobulin or corticosteroids are recommended as treatments [9,10]. Platelet
transfusions, however, remain the principal means of treating thrombocytopenia in NICUs [11]. Platelet transfusions generally are used in two situations: (1)
treatment of thrombocytopenic hemorrhagethrombocytopenic neonates who
have active bleeding and (2) prophylactic administrationthrombocytopenic
neonates who do not have active bleeding but where the risk for bleeding is
judged to significantly outweigh the risks and costs of platelet transfusion.
More than 95% percent of the platelet transfusions given to NICU patients
are in the second categorythey are given prophylactically, with the supposition that this reduces the risk for a serious hemorrhage. Unfortunately, this
common use of platelet transfusions in NICUs lacks a substantial evidence
base. This review focuses on the issue of prophylactic platelet transfusion practice in NICUs and evaluates the rationale, benefits, risks, and the studies
needed to improve this important aspect of neonatology practice.

*Corresponding author. Intermountain Healthcare, McKay-Dee Hospital Center, NICU,

4300 Harrison Boulevard, Ogden, UT 84403. E-mail address: rdchris4@ihc.com
0065-3101/08/$ see front matter
doi:10.1016/j.yapd.2008.07.003

2008 Elsevier Inc. All rights reserved.

CHRISTENSEN

256

Percent With Thrombocytopenia

100%

80%

60%

40%

20%

0%
<500g

500-600g 601-700g 701-800g 801-900g 901-1000g

Birth Weight
Fig. 1. Thrombocytopenia as related to birth weight. The incidence of thrombocytopenia
among 284 neonates weighing less than 1000 g at birth, as related to birth weight. (From Christensen RD, Henry E, Wiedmeier SE, et al. Thrombocytopenia among extremely-low birth-weight
neonates: data from a multihospital healthcare system. J Perinatol 2006;26:34853; with
permission.)

DEFINING NEONATAL THROMBOCYTOPENIA

Following conflicting reports, it now seems that the anatomic site of blood
drawing does not affect a neonates platelet count [12]. Thurlbeck and McIntosh obtained blood samples from various sources, including umbilical lines,
venipunctures, and capillary sticks, and found no differences in platelet counts.
This is in contrast to the significant differences in leukocyte concentrations
(lower in arterial sources) and in erythrocyte concentrations (higher in capillary
sources). Largely on the basis of the Thrulbeck study, it now is common practice to obtain platelet counts from neonates using the most convenient sampling
site, without concern regarding sample-site differences [13].
An expected range of platelet counts, spanning from a low of 150,000/lL to
a high of 450,000/lL, has been reported among various groups of neonates,
including healthy neonates, very low birth weight (VLBW) neonates, and
NICU patients. No reports can be cited, however, that sampled large numbers
of subjects (more than 1000), who were all tested using the same drawing
methods and all run on the same types of laboratory equipment. Arad and coworkers [14] reported platelet counts of 323,336  106,918 (mean  SD)
among 119 healthy term neonates. These values were slightly higher than
those reported by Hathaway and Bonnar [15] and Tan and colleagues [16],
but all reports concluded that the lower limit of normal is approximately
150,000/lL [13]. Thus, a platelet count below 150,000/lL generally is considered abnormally low, thus defining neonatal thrombocytopenia. The incidence
of neonatal thrombocytopenia, reported among healthy term neonates, NICU
patients, and ELBW neonates, is reviewed in Table 1.

PLATELET TRANSFUSION PRACTICE

257

Table 1
Reported incidence of neonatal thrombocytopenia
Author

Year

Definition of
thrombocytopenia

Surveys of patients in well baby nurseries

Uhrynowska
1997
<100,000/lL
Dreyfus
1997
<150,000/lL
Patients in a neonatal ICU
Mehta
1980
<150,000/lL
Castle
1986
<100,000/lL
Oren
1994
<100,000/lL
Christensen
2006
<150,000/lL

Subjects

Incidence

Healthy, term
Healthy, term

0.7%
0.9%

NICU
NICU
NICU
ELBW

35%
22%
18%
73%

The mean platelet volume (MPV) can be important to consider when evaluating neonates who have thrombocytopenia. The mean corpuscular volume,
an electronic estimate of the average size of platelets, expressed in femtoliters
(fL), is analogous to the mean corpuscular volume used in erythrocyte evaluation. From 143 healthy newborn infants, approximately half of whom were
preterm, Patrick and Lazarchick [17] reported a MPV of 8.7  1.0 fL. This
is similar to that reported by Arad and colleagues [14]. One way to view a neonates capacity to generate an effective platelet plug is to consider the platelet
count and the MPV together. Specifically, by multiplying the number of platelets/lL of blood by the MPV (in fL), the circulating platelet mass is calculated,
using the units, fL of platelets/lL of blood (Table 2). These comparisons have
not been reported for VLBW or ELBW neonates.
THE DEGREE OF THROMBOCYTOPENIA
Categorizing patients into groups according to their platelet count is artificial
but perhaps helpful in assigning risk preliminarily (Table 3). For example,
neonates who have platelet counts in the range of 100,000 to 149,999/lL are
categorized as having mild thrombocytopenia, those who have counts in
the range of 50,000 to 99,999/lL as having moderately severe thrombocytopenia, and those who have counts below 50,000/lL as having severe thrombocytopenia. This classification system leads to inaccuracies, because counts
Table 2
Normal platelet values of newborn infants
Population

Platelets/lL

MPV (fL)

Circulating platelet mass

(fL platelets/lL blood)

Term
LBW
VLBW

310,000  68,000
290,000  70,000
275,000  60,000

8.7  1.0
8.9  1.1
NR

2.70  106
2.58  106
NR

Abbreviations: LBW, low birth weight, <2500 g; NR, not reported; VLBW, <1500 g.

CHRISTENSEN

258

Table 3
Classification of neonatal thrombocytopenia according to degree
Platelet count

Classification

>450,000/lL
150,000450,000/lL
100,000149,999/lL
50,00099,999/lL
<50,000/lL

Thrombocytosis
Normal
Mild thrombocytopenia
Moderately severe thrombocytopenia
Severe thrombocytopenia

of 5000/lL and 45,000/lL are categorized together as severe, but the two
counts likely have very different bleeding risks. This classification system
also ignores the influence of the MPV on the capacity to generate an effective
platelet plug. Better methods should be developed for classifying the degree of
neonatal thrombocytopenia, which are more helpful for clinicians as they seek
to estimate the risks and benefits of platelet transfusions for specific patients.
Nevertheless, for this review, lacking a better classification system, the bleeding
risks are discussed for mild, moderate, and severe neonatal thrombocytopenia
as defined previously.
THE BLEEDING RISK ASCRIBABLE TO MILD NEONATAL
THROMBOCYTOPENIA
Assigning a bleeding risk on the basis of the platelet count is imprecise and
overly simplistic, yet such an assignment is necessary if prophylactic platelet
transfusions are used. Much evidence links thrombocytopenia in neonates to
pathologic bleeding [18]. For neonates who have mild thrombocytopenia, however, no increase in bleeding risk has been proved. Possible ways to evaluate
bleeding risk in neonates include laboratory tests and clinical outcomes studies.
Two laboratory tests that have been used to try to estimate the bleeding risk
are the template bleeding time and the PFA-100 platelet function analyzer closure time.
The template bleeding time assesses hemostatic effectiveness of platelet-endothelial interactions in vivo [19,20]. In 1980, Feusner [21] reported modifying the
standard adult template bleeding time method for use in infants by decreasing
the incision size. He reported that the bleeding time for healthy term neonates
(3.4  0.9 minutes) and for preterm neonates (3.6  1.0 minutes) did not differ
from that of healthy children (3.4  1.3 minutes). Almost a decade later,
Andrew and colleagues [22,23] modified the technique by using an automated
device with a retractable blade. They found slightly shorter bleeding times with
this new method. Using the Andrew method, the authors group reported that
during the first week of life, a hematocrit less than 28% causes a prolongation of
the bleeding time and that indomethacin administration prolongs the bleeding
time significantly [24,25].
Del Vecchio and colleagues [26] performed template bleeding times on 240
neonates, who had a range of platelet counts from 110,000 to more than

PLATELET TRANSFUSION PRACTICE

259

700,000/lL. They found only a weak relationship between the bleeding time
and the platelet count (Fig. 2). Because the bleeding time was not lengthened
by platelet counts within the mild thrombocytopenic range, they concluded
that transfusing platelets provides no benefit to neonates who have mild thrombocytopenia. In adult subjects, Harker and Slicter [27] showed that once the
platelet count falls below 50,000 to 60,000/lL, the bleeding time lengthens,
approximately in proportion to the platelet count. Thus, patients who have
the lowest platelet counts have the longest bleeding times. Studies, such as
the Harker bleeding time study, are needed in neonates, to obtain information,
albeit indirect, regarding how low a platelet count must fall before any adverse
effect is observed on platelet plug generation.
The PFA-100 is an in vitro method for assessing platelet plug formation
[2830]. It serves as a rapid and reproducible measurement, reducing the interoperator variation that is inherent in the template bleeding time test. The
PFA-100 test requires 1.6 mL of anticoagulated blood. The blood specimen
is aspirated through an aperture that has been cut into a membrane coated
with collagen and ADP or epinephrine. Platelets adhere to the membrane
and aggregate until a stable platelet plug occludes the flow of blood through
the aperture. The number of seconds required to close the aperture with a platelet plug is termed, closure time, and this measurement approximates the template
bleeding time. Generally, each test includes two closure time measurements,
one obtained with ADP and the other obtained with epinephrine.
Patients taking nonsteroidal anti-inflammatory medications typically have
a prolonged closure time with epinephrine but a normal closure time with
ADP, whereas most qualitative platelet disorders cause prolongation of both.
New studies are needed of PFA-100 closure times in neonates who have various platelet counts. Such studies may give indirect evidence regarding the risk

Bleeding Time (Seconds)

480
420
360
300
240
180
120
r = -0.10

60
0
0

100,000 200,000 300,000 400,000 500,000 600,000 700,000

Platelets/uL blood
Fig. 2. Relationship of platelet count (platelets/lL) and template bleeding time (seconds) in 240
neonates. (From Del Vecchio A, Latini G, Henry E, et al. Template bleeding times of 240 neonates
born at 24 to 41weeks gestation. J Perinatol 2008;28:42731; with permission.)

260

CHRISTENSEN

for various platelet counts relative to platelet plug formation and, by extrapolation, to the risk for mucocutaneous hemorrhage [2527].
Considering all discussed previously, the author finds no evidence to support
the use of prophylactic platelet transfusions for nonbleeding neonates who have
mild thrombocytopenia. With more accurate means of assessing platelet function, certain neonates who have mild thrombocytopenia but very poor platelet
function could benefit from prophylactic platelet transfusions. For the present
time, however, it is difficult to recommend a prophylactic platelet transfusion
on the basis of mild thrombocytopenia. Not all neonatologists agree with this interpretation. In a recent survey of neonatologists in the United States and Canada, many declared that they would order a platelet transfusion for a nonbleeding
patient on the basis of a platelet count in the range of 100,000 to 149,999/lL [31].
THE BLEEDING RISK ASCRIBABLE TO MODERATELY SEVERE
THROMBOCYTOPENIA
The decision to use any treatment in a NICU should be based on its anticipated benefits and risks. The benefits of platelet transfusions in nonbleeding
neonates who have moderately severe thrombocytopenia are highly questionable. The intent of prophylactic platelet transfusions to such patients is based
on the supposition that such transfusions prevent hemorrhage. Andrew and
colleagues [32] tested that hypothesis by randomizing preterm neonates to
a platelet transfusion group (receiving up to three platelet transfusions in the
first week to keep platelet counts 150,000/lL) or a no transfusion group (letting the counts fall to <50,000/lL or until there was evidence of clinical bleeding before a platelet transfusion was given). They observed no difference in
bleeding outcomes between the two groups and concluded that prophylactic
platelet transfusions were not helpful in preventing bleeding in neonates who
have platelet counts above 50,000/lL.
Despite Andrew and colleagues [32] findings and conclusions, survey results
indicate that many or most neonatologists advocate platelet transfusions to
nonbleeding neonates who have moderately severe thrombocytopenia, particularly if patients are VLBW or ELBW and within the first week of life.
A study that might be helpful in sorting out these issues would be one where
the template bleeding time and the PFA-100 closure time were measured in
neonates who had moderately severe thrombocytopenia, with values repeated
before and after those who receive a platelet transfusions. It also would be useful to see the effect of various medications used in NICUs on the template
bleeding time and the PFA-100 closure time and the effect of gestational age
and postnatal age on these measurements. Clinical trials seeking to validate
Andrew and colleagues finding also would be valuable.
THE BLEEDING RISK ASCRIBABLE TO SEVERE
THROMBOCYTOPENIA
Life-threatening or fatal hemorrhage certainly can occur as a consequence of
severe thrombocytopenia. This relationship has been established among adults,

PLATELET TRANSFUSION PRACTICE

261

children, and neonates [33]. Otherwise healthy, term neonates who have
extremely low platelet counts can be found to have significant intracranial hemorrhage with no recognized antecedent trauma. Intracranial hemorrhage in
utero can occur when a fetus has severe thrombocytopenia and no other
recognized risk factors for hemorrhage [34].
The level of platelet count that conveys an increased risk for hemorrhage, in
NICU patients, is not clear. Factors other than the platelet count undoubtedly
are important in determining the hemorrhagic risk. For instance, the MPV, gestational age, postnatal age, medications, stability of blood pressure, and the
presence of other conditions, such as patent ductus arteriosus, sepsis, shock,
or pulmonary hypertension, all likely weigh heavily into the hemorrhagic risk.
Adults who develop thrombocytopenia secondary to cancer or chemotherapy generally are not given prophylactic platelet transfusions unless their
platelet counts falls below 10,000/lL [3537]. Multicentered, prospective, cluster-randomization trials have contrasted prophylactic transfusion triggers of
20,000 versus 10,000/lL and found no difference, thus 10,000/lL is the transfusion trigger generally selected.
No evidence base is available for assigning a platelet transfusion trigger to
NICU patients. Few if any neonatologists would participate in a clinical trial
where the platelet count would be allowed to fall to10,000/lL before platelet
transfusions were offered. Murray suggests, in nonbleeding neonates, permitting
the platelet count to drop to 30,000/lL before considering a prophylactic platelet
transfusion [38]. The Intermountain Healthcare platelet transfusion guidelines
(discussed later) recommend permitting the platelet count to drop to 50,000/lL
in unstable NICU patients but letting the count drop to 20,000/lL in stable
patient before ordering a prophylactic platelet transfusion.
RISKS OF PLATELET TRANSFUSIONS
Several studies have illustrated that the number of platelet transfusions administered predicts the mortality rate (Fig. 3) [2,4,5,39]. Some of this correlation
between number of platelet transfusions and mortality rate is ascribable to
unknown and unmeasured factors, such as increasing level of illness. Growing
evidence indicates, however, that some of this correlation is the result of harmful effects of multiple platelet transfusions in this group of patients [3842].
Much evidence indicates that platelet transfusions are not benign. Risks
include transmission of bacterial and nonbacterial infections, alloimmunizaiton,
febrile reactions, hemolytic reactions, allergic reactions, and transfusion-related
lung injury [3945]. The most common complication of platelet transfusion is
bacterial infection [42,43]. Because of the nature of their preparative and storage processes, such as the need to store platelet concentrates at room temperature, platelet products have a much higher risk for bacterial contamination
than do erythrocytes or plasma products. Up to 10% of platelet preparations
are contaminated with small concentrations of bacteria.
Another risk of platelet transfusions relates to the concentrations of various
bioactive factors, including platelet-activating factor (PAF), which might lead to

CHRISTENSEN

262

80%

Mortality rate

70%
60%
50%
40%
30%
20%
10%
0%
0

7-8

9-10 11-29 >30

Number of platelet transfusions

Fig. 3. The mortality rate as a function of the number of platelet transfusions received. The
figure depicts an increasing mortality rate associated with an increasing number of platelet
transfusions administered to 1600 thrombocytopenic NICU patients. (From Baer VL, Lambert
DK, Henry E, et al. Do platelet transfusions in the NICU adversely effect survival? Analysis
of 1600 thrombocytopenic neonates in a multihospital healthcare system. J Perinatol
2007;27:7906; with permission.)

or exacerbate inflammatory injury [41,46]. Although prophylactic platelet transfusions are given to prevent bleeding, Kahn and colleagues [47] reported that
neonates receiving multiple platelet transfusions actually had a higher incidence
of intraventricular hemorrhage. Additionally, unlike the case with erythrocyte
transfusions, where a neonate can receive multiple transfusions with only one
donor exposure, each platelet transfusion involves an additional donor exposure.

# of Transfusions

Mortality

g
Unknownunmeasured
factors

Fig. 4. Sensitivity analysis. Relationships are shown schematically between three factors: (1)
the number of platelet transfusions received, (2) the mortality rate, and (3) unknown-unmeasured factors (such as sickness level of the infant) that could potentially influence the other
two factors. The solid arrow (r) is the correlation between the unmeasured factors and the number of platelet transfusions given. The solid arrow (g) is the correlation between the unmeasured factors and the mortality rate. The dashed arrow with the question mark asks whether
or not there is any direct relationship between the number of platelet transfusions received
and the mortality rate. (From Baer VL, Lambert DK, Henry E, et al. Do platelet transfusions
in the NICU adversely effect survival? Analysis of 1600 thrombocytopenic neonates in a multihospital healthcare system. J Perinatol 2007;27:7906; with permission.)

r\g
0
0.5
0.75
0.85
0.9
0.95
0.98
0.99

0
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)
1.14
(1.10,1.18)

0.2

0.4

0.6

0.8

1.14
(1.10,1.18)
1.11
(1.08,1.15)
1.10
(1.07,1.14)
1.10
(1.07,1.14)
1.10
(1.06,1.13)
1.09
(1.06,1.13)
1.09
(1.06,1.13)
1.09
(1.06,1.13)

1.14
(1.10,1.18)
1.09
(1.06,1.13)
1.07
(1.04,1.11)
1.06
(1.03,1.10)
1.06
(1.02,1.09)
1.05
(1.02,1.09)
1.05
(1.02,1.08)
1.05
(1.02,1.08)

1.14
(1.10,1.18)
1.07
(1.04,1.11)
1.04
(1.01,1.07)
1.02
(0.99,1.06)
1.02
(0.99,1.05)
1.01
(0.98,1.05)
1.01
(0.98,1.04)
1.01
(0.98,1.04)

1.14
(1.10,1.18)
1.05
(1.02,1.08)
1.00
(0.97,1.04)
0.99
(0.96,1.02)
0.98
(0.95,1.01)
0.97
(0.94,1.01)
0.97
(0.94,1.00)
0.97
(0.94,1.00)

1.14
(1.10,1.18)
1.03
(0.99,1.06)
0.97
(0.94,1.01)
0.95
(0.93,0.99)
0.94
(0.92,0.98)
0.93
(0.91,0.97)
0.93
(0.90,0.96)
0.93
(0.90,0.96)

PLATELET TRANSFUSION PRACTICE

Table 4
Sensitivity analysis

This table shows the odds-ratios relating the increase in mortality rate for each additional platelet transfusion given, along with the 95% CI, for various values of r (the correlation
between the unmeasured variable and the number of transfusions, in the rows) and g (the log-odds ratio of mortality and the unmeasured variable, in the columns). Odds ratios in
red represent conditions where platelet transfusions cause an increase in the mortality rate. Those in black represent conditions where platelet transfusions neither significantly
increase nor decrease the mortality rate. Values in green represent conditions where platelet transfusions decrease the mortality rate.
From Baer VL, Lambert DK, Henry E, et al. Do platelet transfusions in the NICU adversely effect survival? Analysis of 1600 thrombocytopenic neonates in a multihospital healthcare system. J Perinatol 2007;27:7906; with permission.

263

CHRISTENSEN

264

Thus, when multiple platelet transfusions are given, all of the known plus the
unknown risks inherent in blood donor exposure are compounded [39].
In thrombocytopenic NICU patients, at any level of platelet count, some
neonatologists order a transfusion whereas others do not. Taking advantage
of this variance in practice, the author examined outcomes of NICU patients,
stratified by platelet count, and found that at all levels of platelet count, the
BASIC PREMISE: A NICU patient with clinically significant hemorrhage might urgently need
packed red blood cells (PRBC), platelets, and frozen plasma (FP). A NICU patient who has no signs
of hemorrhage should receive PRBC according to guidelines, platelet transfusions according to
guidelines, cryoprecipitate according to guidelines, and generally should not receive FP.
A PATIENT WITH CLINCALLY SIGNIFICANT BLEEDING: For a NICU patient with
hemorrhage, consideration should be given for an urgent PRBC transfusion (20 mL/k or more) and a
STAT CBC and DIC screen should be performed. If platelets are <100,000/L a platelet transfusion
(at least 15 mL/k) should be considered. If coagulopathic*, FP (at least 15 ml/k) should be
considered.
HEMORRHAGE

PRBC STAT
> = 20mL/k

CBC/DIC SCREEN STAT

PL <100k, PL Tx
> = 15mL/k

Coagulopathic, FFP
> = 15mL/k

A PATIENT WITH NO HEMORRHAGE: Transfuse PRBC according to guidelines. Transfuse

platelets according to guidelines. Give Cryoprecipitate (0.5 U/k) if the fibrinogen is <100 mg/dL. In
general, FP is not used for a non-bleeding NICU patient.
NO HEMORRHAGE

PRBC by guidelines

PL Tx by Guidelines

Cryo if Fibrinogen < 100 mg/dL

PLATELET TRANSFUSION GUIDELINE SUMMARY

Transfuse if:
PLTS <100,000/uL and ECMO, or going to or just had surgery.
PLTS <50,000/uL and unstable.
PLTS <20,000/uL and stable.

Fig. 5. Intermountain Healthcare NICU platelet transfusion guidelines. Coagulopathic is

defined by abnormal clotting studies (in particular PT and fibrinogen) that are outside the normal
range (From Christensen RD, editor. Hematologic problems of the neonate. W.B. Saunders;
2000. p. 2445 [Tables 12-1 and Table 12-2]); with permission. Hemorrhage is defined as
abnormal bleeding, including visible blood in the endotracheal tube, stools, urine, mucous
membranes, skin, or bleeding into any body cavity or potential space such as with an active
intraventricular hemorrhage.

PLATELET TRANSFUSION PRACTICE

265

patients who received platelet transfusions had a higher mortality rate than
those who did not [39]. Furthermore, neonates who were not given any platelet
transfusions had a mortality rate of 2%; those who had one or two transfusions
had a mortality rate of 11% (P < .001); those who had greater than 10 had
a mortality rate of 35% (P < .001); and those who had 20 or more had a mortality rate of 50% (P < .001). A sensitivity analysis (Fig. 4 and Table 4) suggests
that the platelet transfusions themselves likely were responsible for some fraction of the increasing mortality rate.

% transfusions outside
guidelines

PLATELET TRANSFUSION GUIDELINES

Establishing NICU transfusion guidelines has been advocated as a way of
improving practice and reducing unnecessary transfusions [4754]. In 2002, Intermountain Healthcare, a not-for-profit healthcare organization in the Western
United States, established and implemented a written set of NICU transfusion
guidelines. The guidelines that pertain to platelet transfusions are shown as
Fig. 5. Intermountain Healthcare recently conducted a study to examine compliance with these guidelines. Specifically, every blood component transfusion
administered to every NICU patient who had a birthdate of January 1, 2006,
through December 31, 2006, and was cared for in one of three perinatal center
associated level III NICUs of Intermountain Healthcare was critically
reviewed. For every transfusion given, two or more reviewers examined all
available information in the medical record to judge whether or not the transfusion was administered in compliance with or not in compliance with the
guidelines.
During the year 2006, 4% of all Intermountain Healthcare NICU admissions
received one or more platelet transfusions [52]. Each neonate who received
a platelet transfusion actually received an average of three platelet transfusions.
100
90
80
70
60
50
40
30
20
10
0

112
32

0-9

10-19

37

13

21

20-29

30-39

=>40

Day of life transfused

Fig. 6. The percent of platelet transfusions that were given outside of guidelines is shown in
association with the day of life the transfusion was given. The numbers above the bars refer to
the total number of transfusions given on that day of life. (From Baer VL, Lambert DK, Schmutz
N, et al. Adherence to NICU transfusion guidelines: data from a multihospital healthcare
system. J Perinatol 2008;28:4927; with permission.)

266

CHRISTENSEN

Extrapolating this to the 400,000 NICU admissions annually in the United

States, approximately 16,000 neonates were estimated to have received approximately 48,000 platelet transfusions each year. Of the platelet transfusions
given in 2006, 69% were compliant with the platelet transfusion guidelines
(ranging from 63% to 73% in various NICUs). As shown in Fig. 6, the platelet
transfusions administered in violation of the guidelines tended to be during the
first week of life. These out-of-guideline transfusions involved nonbleeding
neonates who had platelet counts in the range of 50,000/lL to 100,000/lL.
The guidelines call for platelet transfusions to nonbleeding neonates only if
the platelet count falls below 50,000/lL, but in actual practice, platelet transfusions often were given with platelet counts in the 50,000 to 99,999/lL range.
This was the most common transfusion guidelines violation detected [50]
and illustrates the widespread belief, also shown in surveys, that transfusing
platelets to nonbleeding neonates who have moderately severe thrombocytopenia provides benefits that outweigh risks and costs [31].
SUMMARY
Some of the platelet transfusions currently given to NICU patients are
unnecessary and convey no benefits. Although ordered with good intentions,
unnecessary platelet transfusions carry known and unknown risks. Identifying
and eliminating any unnecessary platelet transfusions in NICUs would be
a step toward better care, lower costs, and more careful preservation of blood
component resources.
A renewed interest in platelet transfusion studies is needed, if essential data is
to be gathered to improve NICU platelet transfusion practice. Retrospective
studies can be of value: for instance, seeking associations between bleeding
events and platelet counts can suggest the possibility of cause and effect relationships. Such studies might identify approximate platelet count levels that
convey high hemorrhagic risk and might help focus future prospective trials.
Prospective indirect studies also can be of value, for instance, measuring the
template bleeding time and the PFA-100 closure time as a function of platelet
count and perhaps as a function of circulating platelet mass, and would provide
new information with relevance to platelet transfusion benefits. Such studies
might give a better awareness of how low the platelet count can fall before
platelet plug formation is impaired. It seems inescapable, however, that new,
multicentered, randomized, prospective studies are needed, where NICU
patients are assigned different platelet transfusion triggers and then carefully
tracked for bleeding events and long-term neurodevelopmental outcomes.
Only that type of study is likely to generate the evidence base needed for widespread implementation of improvements in NICU platelet transfusion practice.
References
[1] Murray NA, Roberts IAG. Neonatal transfusion practice. Archives of Diseases in Childhood. Fetal and Neonatal Edition 2004;89:1017.

PLATELET TRANSFUSION PRACTICE

267

[2] Del Vecchio A, Sola MC, Teriaque DW, et al. Platelet transfusions in the neonatal intensive
care unit: factors predicting patients who will require multiple transfusions. Transfusion
2001;41:8038.
[3] Chakravorty S, Murray N, Roberts I. Neonatal thrombocytopenia. Early Hum Dev
2005;81:3541.
[4] 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:
34853.
[5] Garcia MG, Duemas E, Sola MC, et al. Epidemiologic and outcome studies of patients who
received platelet transfusions in the neonatal intensive care unit. J Perinatol 2001;21:41520.
[6] Roberts IAG, Murray NA. Thrombocytopenia in the newborn. Current Opinion in Pediatrics
2003;15:1723.
[7] Sola MC. Evaluation and treatment of severe and prolonged thrombocytopenia in neonates.
Clin Perinatol 2004;31:114.
[8] Roberts IAG, Murray NA. Neonatal thrombocytopenia. Current Hematology Reports
2006;5:5563.
[9] Dunsmore KP. Intravenous immunoglobulin g therapy in fetal and neonatal alloimmune
thrombocytopenia. BioDrugs 1997;8:26572.
[10] Murphy MF, Bussel JB. Advances in the management of alloimmune thrombocytopenia.
Br J Haematol 2007;136:36678.
[11] Luban NLC. Transfusion of blood and blood products. In: MacDonald MG, Ramasethu J, editors. Atlas of procedures in neonatology. Philadelphia: Lippincott/Williams & Wilkins;
2007. p. 31728.
[12] Thurlbeck SM, McIntosh N. Preterm blood counts vary with sampling site. Arch Dis Child
1987;62:745.
[13] Christensen RD. Expected hematologic values for term and preterm neonates. In:
Christensen RD, editor. Hematologic problems of the neonate. Philadelphia: WB Saunders
Co; 2000. p. 1301.
[14] Arad ID, Alpan G, Sznajderman DS, et al. The mean platelet volume in the neonatal period.
Am J Perinatol 1986;3:14.
[15] Hathaway WE, Bonnar J. Perinatol coagulation. New York: Grune and Scratton; 1978.
[16] Tan KL, Yip YY, Low SH. Haematological status of full-term healthy infants in the first week of
life. Ann Acad Med Singapore 1985;14:5728.
[17] Patrick CH, Lazarchick J. The effect of bacteremia on automated platelet measurements in
neonates. Am J Clin Pathol 1990;93:391401.
[18] Andrew M, Castle V, Saigal S, et al. Clinical impact of neonatal thrombocytopenia. J Pediatr
1987;110:45762.
[19] Del Vecchio A, Sola MC. Performing and interpreting the bleeding time in the neonatal
intensive care unit. Clin Perinatol 2000;27:64354.
[20] Del Vecchio A. Use of the bleeding time in the neonatal intensive care unit. Acta Paediatr
2002;91(Suppl):826.
[21] Feusner JH. Normal and abnormal bleeding times in neonates and young children utilizing
a fully standardized template technic. Am J Clin Pathol 1980;74:737.
[22] Andrew M, Castle V, Mitchell L, et al. Modified bleeding time in the infant. Am J Hematol
1989;30:190219.
[23] Andrew M, Paes B, Bowker J, et al. Evaluation of an automated bleeding time device in the
newborn. Am J Hematol 1990;35:2757.
[24] Sola MC, Del Vecchio A, Edwards TJ, et al. The relationship between hematocrit and bleeding
time in very low birth weight infants during the first week of life. J Perinatol 2001;21:36871.
[25] Del Vecchio A, Sullivan SE, Christensen RD, et al. Indomethacin prolongs the bleeding time
in neonates significantly more than ibuprofen. Pediatr Res 2002;51:446A.
[26] Del Vecchio A, Latini G, Henry E, et al. Template bleeding times of 240 neonates born at 24
to 41weeks gestation. J Perinatol 2008;28:42731.

268

CHRISTENSEN

[27] Harker LA, Slicter SJ. The bleeding time as a screening test for evaluation of platelet function.
N Engl J Med 1972;287:1559.
[28] Mammen EF, Comp PC, Gosselin R, et al. PFA-100 system: a new method for assessment of
platelet dysfunction. Semin Thromb Hemost 1998;24:195202.
[29] Ridefelt P, Egberg N, Hillarp A, et al. New in vitro analysis tested: bleeding time is still the
best method for evaluation of primary hemostasis. Lakartidningen 2001;98:39224.
[30] Saxonhouse MA, Sola MC. Platelet function in term and preterm neonates. Clin Perinatol
2004;31:1528.
[31] Josephson C, Su L, Christensen RD, et al. Platelet transfusion practice among neonatologists
in the U.S. and Canada: results of a survey. Pediatr Res 2007;61:6315.
[32] Andrew M, Vegh P, Caco VC, et al. A randomized, controlled trial of platelet transfusions in
thrombocytopenic premature infants. J Pediatr 1993;123:28591.
[33] Corazza MS, Davis RF, Merritt TA, et al. Prolonged bleeding time in preterm infants
receiving indomethacin for patent ductus arteriosus. J Pediatr 1984;105:2926.
[34] Fass RJ, Copelan EA, Brandt JT, et al. Platelet-mediated bleeding caused by broad-spectrum
penicillins. J Infect Dis 1987;155:12428.
[35] Kantarjian H, Giles F, List A, et al. The incidence and impact of thrombocytopenia in
myelodysplastic syndromes. Cancer 2007;109:170514.
[36] Slichter SJ. Evidence-based platelet transfusion guidelines. Hematology Am Soc Hematol
Educ Program 2007;2007:1728.
[37] Greeno E, McCullough J, Weisdorf D. Platelet utilization and the transfusion trigger: a prospective analysis. Transfusion 2007;47:2015.
[38] Murray NA. Evaluation and treatment of thrombocytopenia in the neonatal intensive care
unit. Acta Paediatr Suppl 2002;91:7481.
[39] Baer VL, Lambert DK, Henry E, et al. Do platelet transfusions in the NICU adversely effect
survival? Analysis of 1600 thrombocytopenic neonates in a multihospital healthcare system.
J Perinatol 2007;27:7906.
[40] Fernandes CJ, ODonovan DJ. Platelet transfusions in infants with necrotizing entercolitis.
Curr Hematol Rep 2006;5:7681.
[41] Kenton AB, Hegemier S, Smith EO, et al. Platelet transfusions in infants with necrotizing
entercolitis do not lower mortality but may increase morbidity. J Perinatol 2005;25:1737.
[42] Khashu M, Osiovich H, Henry D, et al. Persistent bacteremia and severe thrombocytopenia
caused by coagulase-negative staphylococcus in a neonatal intensive care unit. Pediatrics
2006;117:3418.
[43] Kuehnert MJ, Roth VR, Haley NR, et al. Transfusion-transmitted bacterial infection in the
United States, 1998 through 2000. Transfusion 2001;41:14939.
[44] Kopko PM, Holland PV. Mechanisms of severe transfusion reactions. Elsevier Science. Paris.
Transfus Clin Biol 2001;8:27881.
[45] Kleinman S, Chan P, Robillard P. Risks associated with transfusion of cellular blood components in Canada. Transfus Med Rev 2003;17:12062.
[46] Murray NA, Hawarth LJ, McCloy MP, et al. Platelet transfusion in the management of severe
thrombocytopenia in neonatal intensive unit patients. Transfus Med 2002;12:3541.
[47] Kahn DJ, Richardson DK, Billett HH. Inter-NICU variation in rates and management of
thrombocytopenia among very low birth-weight infants. J Perinatol 2003;23:3126.
[48] Miyashiro AM, dosSantos N, Guinsburg BI, et al. Strict red blood cell transfusion guideline
reduces the need for transfusions in very-low-birthweight infants in the first 4 weeks of life:
a multicentre trial. Vox Sang 2005;88:10713.
[49] Calhoun DA, Christensen RD, Edstrom CS, et al. Consistent approaches to procedures and
practices in neonatal hematology. Clin Perinatol 2000;27:73353.
ncio JP, Santos AM, Guinsburg R, et al. Strict guideline reduces the need for RBC trans[50] Vena
fusions in premature infants. J Trop Pediatr 2007;53:7882.
[51] Ringer SA, Richardson DK, Sacher RA, et al. Variations in transfusion practice in neonatal
intensive care. Pediatrics 1998;101:194200.

PLATELET TRANSFUSION PRACTICE

269

[52] Baer VL, Lambert DK, Schmutz N, et al. Adherence to NICU transfusion guidelines: data
from a multihospital healthcare system. J Perinatol 2008;28:4927.
[53] Banchette VS, Rand ML. Platelet disorders in newborn infants: diagnosis and management.
Semin Perinatol 1997;21:5362.
[54] Petaja J, Andersson S, Syrjala M. A simple automatized audit system for following and
managing practices of platelet and plasma transfusions in a neonatal intensive care unit.
Transfus Med 2004;4:2818.