Professional Documents
Culture Documents
Thrombocytopenia in Late Preterm and Term Neonates After Perinatal Asphyxia
Thrombocytopenia in Late Preterm and Term Neonates After Perinatal Asphyxia
Thrombocytopenia in Late Preterm and Term Neonates After Perinatal Asphyxia
P
latelet (PLT) transfusions can be lifesaving for
BACKGROUND: A recent NHLBI conference concluded certain thrombocytopenic neonates.1 However,
that platelet (PLT) transfusions of neonates must a recent multidisciplinary NHLBI think tank
become more evidence based. One neonatal disorder expressed the concern that some neonates may
for which transfusions are given is a poorly defined be receiving PLT transfusions unnecessarily, thereby con-
entity, the thrombocytopenia of perinatal asphyxia. To veying little or no benefit with an unfavorable benefit-to-
expand the evidence base for this entity, we performed risk ratio.2
a multicentered, retrospective analysis of neonates with One neonatal condition for which PLT transfusions
perinatal asphyxia. are sometimes given, yet the supporting evidence is
STUDY DESIGN AND METHODS: We analyzed meager, is an entity termed the thrombocytopenia of
records of term and late preterm neonates with perina- perinatal asphyxia. Although this condition has been the
tal asphyxia defined by a cord blood pH of not more subject of previous reports, many aspects are unclear.
than 6.99 and/or base deficit of at least 16 mmol/L. Sola-Visner and Saxonhouse3 recently commented that
From these we identified neonates with at least two for this condition the causative mechanism, the usual
PLT counts of fewer than 150 109/L in the first week range of severity, the typical duration, and the role of PLT
of life and described the severity, nadir, and duration of transfusions are all relatively unknown.
the thrombocytopenia. As a step toward enhancing the knowledge base of the
RESULTS: Thrombocytopenia occurred in 31% (117/ thrombocytopenia of perinatal asphyxia, we conducted a
375) of neonates with asphyxia versus 5% of matched retrospective analysis of all neonates in the past 9 years
nonasphyxiated controls admitted to a neonatal inten- who met the definition of perinatal asphyxia. This was
sive care unit (p < 0.0001). Twenty-one of the 117 done to find the incidence, severity, and usual duration of
asphyxiated neonates were excluded from the remain- this variety of thrombocytopenia. Next we sought to
ing analysis due to disseminated intravascular coagula- understand the value of PLT transfusions in these patients
tion or extracorporeal membrane oxygenation. Nadir by examining, among those where a PLT transfusion was
PLT counts of the remaining 96 were on Day 3
(75 109/L; 90% confidence interval, 35.7 109-
128.6 109/L) and normalized by Days 19 to 21. PLT
ABBREVIATIONS: DIC = disseminated intravascular
counts after asphyxia roughly correlated inversely with
coagulation; ECMO = extracorporeal membrane oxygenation;
elevated nucleated red blood cell count (NRBC) counts
MPV = mean platelet volume; NRBC = nucleated red blood cell
at birth. Thirty of the 96 received at least one PLT
count; TPO = thrombopoietin.
transfusion, all given prophylactically, none for bleeding.
CONCLUSIONS: We maintain that the thrombocytope- From the 1Women and Newborns Clinical Program,
nia of perinatal asphyxia is an authentic entity. Its asso- Intermountain Healthcare, and the 2Department of Pediatrics,
ciation with elevated NRBC counts suggests that Division of Hematology/Oncology, University of Utah School of
hypoxia is involved in the pathogenesis. Because PLT Medicine, Salt Lake City, Utah.
counts are only moderately low, the condition is tran- Address reprint requests to: Robert D. Christensen, MD,
sient, and bleeding problems seem rare, we speculate Women and Newborns Clinical Program, Intermountain
that PLT transfusions should not be needed for most Healthcare, 4403 Harrison Boulevard, Ogden, UT 84403; e-mail:
neonates with this condition. robert.christensen@imail.org.
Received for publication March 10, 2014; revision received
May 28, 2014, and accepted May 30, 2014.
doi: 10.1111/trf.12777
2014 AABB
TRANSFUSION **;**:**-**.
given, the reasons for and response to transfusion. We blood cells (NRBCs) were quantified using either auto-
then attempted to identify the underlying cause of this mated cell counts, performed in accordance with the
variety of thrombocytopenia and to assess whether thera- hematology analyzer manufacturers instructions, or
peutic hypothermia (cooling) used in the treatment of manual enumeration, performed by certified medical
neonates with perinatal asphyxia played a role in the technologists on Wright-stained blood smears enumerat-
pathogenesis of the thrombocytopenia. Last, we sought to ing a minimum of 100 white blood cells per test. The ref-
determine the implication of finding thrombocytopenia erence intervals for complete blood cell count variables
after perinatal asphyxia on the mortality rate. are those we previously published from Intermountain
Healthcare data bases.8
Guidelines for administering PLT transfusions in the
MATERIALS AND METHODS
Intermountain Healthcare neonatal intensive care units
Patient information (NICUs) during this period have been previously pub-
Data were collected retrospectively as deidentified limited lished.9 All PLT transfusions were type specific or AB (D+
data sets from archived Intermountain Healthcare or D), and were derived from apheresis. They were not
records. Intermountain Healthcare is a not-for-profit pooled or volume-reduced but were all subjected to
health care system that owns and operates 19 hospitals leukofiltration and irradiation, and administered in a
with labor and delivery units in Utah and Idaho. The infor- volume of 10 to 15 mL/kg body weight. Gestational age
mation collected was limited to the complete blood was determined by obstetric assignment unless this was
counts and the information displayed in tables and figures changed by the neonatologist on the basis of gestational
of this report. Patient records were accessed if the neonate age assessment (physical examination and neurological-
had a date of birth from June 1, 2004, through May 31, neurodevelopmental findings).
2013. The Intermountain Healthcare Institutional Review Neonates meeting the definition of perinatal
Board approved this as a deidentified data-only study, not asphyxia, and found to have thrombocytopenia (2 PLT
requiring the consent of individual subjects. counts <150 109/L), were matched 1:1 with neonates
A standard operating procedure was in place at all from the same hospitals born during the same period of
Intermountain Healthcare hospitals during all study time who had cord gasses that did not meet either defi-
years, guiding the delivering physician (obstetrician or nition (pH or BE) of perinatal asphyxia. Matching was
family physician) on collecting umbilical cord blood for performed on the basis of gestational age (1 week), sex,
blood gas determination of any birth for which the meta- and birth weight (250 g). Patients whose data would
bolic status is in question.4 After the neonate is delivered otherwise qualify for inclusion in the study were
and the umbilical cord clamped and cut, the umbilical excluded if they were treated with extracorporeal mem-
artery (and sometimes the umbilical vein also), on the side brane oxygenation (ECMO) or had a confirmed diagnosis
of the cord still attached to the placenta, is punctured with of disseminated intravascular coagulation (DIC). This
a needle attached to a 1-mL syringe and 0.2 mL of blood was done because of the confounding effect of PLT con-
is withdrawn for immediate blood gas determination. The sumption caused by ECMO or DIC on the issues relating
cord blood gas findings defining perinatal asphyxia were to thrombocytopenia. DIC was diagnosed by the combi-
pH of not more than 6.99 and/or base deficit of at least nation of hypofibrinogenemia for age,10 schistocytosis,11
16.0 mm/L, because these were suggested in the Ameri- prolongation of the prothrombin time and activated
can College of Obstetricians and Gynecologists Commit- partial thromboplastin time for age,12 and elevated
tee Opinion on cord blood gas analysis4 and were used by D-dimers.
two large studies of therapeutic hypothermia5,6 and are
consistent with the 2003 report of the American College of
Obstetricians and Gynecologists Task Force on Neonatal Data collection and statistical analysis
Encephalopathy and Cerebral Palsy.7 The program used for data collection was a modified sub-
system of clinical workstation. The 3 M Company (Minne-
apolis, MN) approved the structure and definitions of all
Blood cell counting and PLT transfusions data points for use within the program. Data were
PLT counts and mean PLT volumes (MPVs) were deter- managed and accessed by authorized data analysts.
mined in all hospitals with a hematology analyzer (Model Means and standard deviations (SDs) were used to express
LH750, Beckman Coulter, Fullerton, CA) from 2004 values in groups that were normally distributed and
through mid-2012. After mid-2012 the PLT counts and medians and ranges to express values in groups that were
MPVs were determined using counters (Sysmex America, not. Differences in categorical variables were assessed
Inc., Lincolnshire, IL). All blood tests were performed in using the Fisher exact test or chi square. A nonpaired t test
accordance with Intermountain Healthcare Laboratory was used to assess continuous variables. Significance was
Services standard operating procedures. Nucleated red set as a p value of less than 0.05.
Did not have a blood gas definition of perinatal asphyxia Had a blood gas definition of perinatal asphyxia
(n = 23,912) (n = 439)
Controls matched 1:1 with the 375 with perinatal Had 2 platelet counts obtained during the first
asphyxia who had 2 PLT counts in the first week week (n = 375)
Fig. 1. Study flow diagram to identify neonates with the thrombocytopenia of perinatal asphyxia.
Fig. 2. All PLT counts of 96 neonates with the thrombocytopenia of perinatal asphyxia during their first 3 weeks are included in the
figure. Patients were at least 35 weeks gestation at birth, met a cord blood gas definition of perinatal asphyxia, and in the first week
had at least two PLT counts of less than 150 109/L. The median values, first and third interquartile ranges (box), and 90% CIs
(whiskers) are shown. The shaded area represents the normal reference interval for PLT counts of neonates of at least 35 weeks
gestation during the first 21 days after birth (modified from Wiedmeier et al., J Perinatol 2009;29:130-136).13
Fig. 3. PLT counts after 56 apheresis PLT transfusions to 30 neonates are included in the figure. All values are normalized to per-
centage of the pretransfusion PLT count (all pretransfusion counts are shown as 100%). The median values, first and third
interquartile ranges (box), and 90% CIs (whiskers) are shown. The peak and characteristics of the subsequent fall are used to infer
the kinetic mechanism responsible for the thrombocytopenia (reduced PLT production vs. accelerated PLT destruction).
period after transfusion, we conducted a subgroup analy- relationship (R2 = 0.107) was seen between high NRBCs at
sis including only the PLT counts performed in the 6-hour birth and low PLT counts.
period after completing the transfusions. This showed an Table 2 shows data from the neonates with perinatal
average 6-hour increase to 287 109 143 109 PLTs/L. asphyxia divided into two groups: 1) acute hypoxia,
We next compared PLT counts after apheresis PLT defined as having a normal NRBC count on the day of
transfusion of six nonasphyxiated neonates during this birth (50th percentile reference range, n = 172), and 2)
time period that had a diagnosis of alloimmune thrombo- chronic hypoxia, defined as having an elevated NRBC
cytopenia (immune-mediated PLT destruction). Their count on the day of birth (>95th percentile reference range
counts increased by only 5.7 109 4.5 109 PLTs/L. limit, 3.0 109 NRBCs/L, n = 167).15 Many characteristics
MPV measurements accompanying the PLT counts of of neonates in the two groups were similar; however,
the 96 thrombocytopenic neonates over their first week thrombocytopenia was much more common (52% vs.
after birth are displayed in Figure 4. Overlying the values 15%, p < 0.0001) in the chronic hypoxia group.
in the shaded area is the MPV reference interval (5th to Therapeutic hypothermia was administered to 91 of
95th percentile limits) that we published previously.12 the 375 neonates with perinatal asphyxia. All of these were
during the period 2008 to 2013, because the Intermoun-
tain Healthcare NICUs began offering this treatment in
Relationships with hypoxia and 2008.13 The 91 with therapeutic hypothermia had slightly
therapeutic hypothermia lower PLT counts at birth (168.7 109 75.9 109/L,
An elevated NRBC count at birth has been used as a mean SD) than did those not treated with hypothermia
marker for prolonged (exceeding 24 hr) hypoxia in (195.5 109 68.6 109/L, p < 0.0001). However, both
utero.14,15 Figure 5 includes data of all 375 neonates with groups had a decrease in PLT count over the first 3 to
perinatal asphyxia and shows their NRBC count at birth 4 days (nadir counts, 108.9 109 65.2 109/L in those on
and their lowest PLT count during the first 3 days. A weak hypothermia vs. 142.5 109 78.2 109/L in those not
Fig. 4. MPV measured in fL. Values are shown from 96 neonates who met a definition of thrombocytopenia of perinatal asphyxia.
The median values, first and third interquartile ranges (box), and 90% CIs (whiskers) are shown. The shaded area represents the
normal reference interval for MPV (modified from Wiedmeier et al., J Perinatol 2009;29:130-136).13
TABLE 2. Characteristics of neonates at least 35 weeks gestation meeting a definition of perinatal asphyxia and
grouped according to whether their NRBC count on the day of birth was normal (Group 1) or elevated
(Group 2)*
Thrombocytopenia
Apgar Initial PLT count diagnosed
Gestational Birthweight score base deficit at birth during the
Group age (weeks) (g) (5 min) Initial pH (mmol/L) (109/L) first week
1: normal NRBC count at birth 39 1 3172 535 5.5 2.2 6.89 0.09 18 3 220 67 26/172 (15%)
(acute hypoxia) (n = 172)
2: elevated NRBC count at birth 38 1 3228 645 4.8 2.5 6.87 0.12 18 3 183 75 87/167 (52%)
(chronic hypoxia) (n = 167)
p value 0.229 0.379 0.009 0.223 0.762 <0.0001 <0.0001
* A normal NRBC count at birth was used as a marker of acute perinatal hypoxia. An elevated NRBC count at birth (>95th% upper reference
interval) was used as a marker of hypoxia with a duration of more than 24 to 36 hours before birth.14,15
Two or more PLT counts of fewer than 150 109/L.
TABLE 3. Causes of, and associations with, death among neonates with thrombocytopenia after
perinatal asphyxia
Death caused by or associated with Withdrawal of respiratory support because
Category of thrombocytopenia Death bleeding problems of futility
Thrombocytopenia of perinatal 11 (11.5%) 0 Eleven all had severe HIE. None were
asphyxia (n = 96) recorded to have died because of or in
association with bleeding problems.
ECMO after perinatal asphyxia 2 (12.5%) One developed a large subdural The patient with a large subdural
(n = 16) hemorrhage on ECMO with a marked hemorrhage had multiple other
midline brain shift unresolvable medical problems and
support was withdrawn.
One had CDH and developed an
occlusive SVC thrombus on ECMO, in
addition to multiple other problems, and
support was withdrawn.
DIC after perinatal asphyxia 4 (80%) Three had bleeding and oozing requiring One patient had support withdrawn for
(n = 5) dressings on multiple body parts up to severe HIE. At the time of death
the time of death abnormal bleeding/oozing was no longer
present.
CDH = congenital diaphragmatic hernia; HIE = hypoxemic-ischemic encephalopathy; SVC = superior vena cava.
highlighted by Josephson and colleagues2 in a recent sym- with this variety of thrombocytopenia, with PLT counts in
posium, PLT transfusion strategies for neonates currently the 40 109 to 60 109/L and free of other complications,
lack a strong evidence base, yet this is needed as a high should do well without PLT transfusions. Thirty of the 96
priority. From our present findings we conclude that most neonates in this analysis with this variety of thrombocy-
neonates with this variety of thrombocytopenia will have topenia received one or more PLT transfusions. In retro-
nadir PLT counts of higher than 40 109/L (the mean spect, we question how many of these, if any, were
nadir count was 75 109/L; 90% CI, 35.7 109-128.6 109/ beneficial. None of the 30 had a PLT count of fewer than
L). Surveys of neonatologists in Central Europe25 and 20 109/L, almost half had a pretransfusion PLT count of
North America26 agree that PLT transfusions are recom- more than 50 109/L, and all 30 were transfused prophy-
mended when PLT counts fall below 20 109/L, but lactically with no bleeding.
counts that low should be quite rare in the thrombocyto- Our data suggest that therapeutic hypothermia does
penia of perinatal asphyxia. Whether transfusions should not independently cause neonatal thrombocytopenia.
be recommended when the counts are in the 40 109 to Hypothermia for 72 hours has recently become a common
60 109/L range is not clear.2,27 Stanworth and coworkers28 and effective means of improving neurodevelopment of
reported a prospective observational study of 194 neo- term neonates after perinatal asphyxia.29,30 Hypothermia
nates with thrombocytopenia, where a subgroup of 47 had does adversely affect PLT function, as manifested by pro-
their lowest count in the 40 109 to 60 109/L range. Only longation of the bleeding time and lengthening of the
two of those who were at least 35 weeks gestation had a PFA100 closure time.13 Once rewarming has occurred,
major hemorrhage and both of those had DIC. Thus we PLT function rapidly normalizes.13 Additional study is
judge it is very likely that late preterm and term neonates needed to determine whether PLT dysfunction during
therapeutic hypothermia warrants giving PLT transfu- 5. Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head
sions at higher PLT counts than for neonates not on thera- cooling with mild systemic hypothermia after neonatal
peutic hypothermia. encepathology; multicenter randomized trial. Lancet 2005;
Taken together our study results suggest that 365:663-70.
approximately 30% of term and late preterm neonates 6. Shankaran S, Laptook AR, Ehrenkranz RA, et al.; National
with a cord pH of not more than 6.99 and/or a base deficit Institute of Child Health and Human Development Neona-
of at least 16 mmol/L will have a condition that can be tal Research Network. Whole body hypothermia for neo-
called the thrombocytopenia of perinatal asphyxia. A nates with hypoxic ischemic encephalopathy. N Engl J Med
minority of these neonates will have coexisting severe 2005;353:1573-84.
DIC. Likely those will be the most ill patients, with 7. American Academy of Pediatrics, American College of
elevated coagulation times and D-dimers,31 very low and Obstetricians and Gynecologists. Neonatal encephalopathy
rapidly decreasing PLT counts, persistent hypotension and cerebral palsy: defining the pathogenesis and patho-
and acidosis, and a poor prognosis for survival. However, physiology. Elk Grove Village, IL, AAP; Washington, DC,
we speculate that the great majority of neonates with ACOG, 2003.
thrombocytopenia after perinatal asphyxia will have a 8. Christensen RD. Reference ranges in neonatal hematology.
much more benign condition, with PLT counts that do not In: de Alarcn PA, Werner EJ, Christensen RD, editors. 2nd
fall below 40 109/L. We maintain that these should not ed. Neonatal hematology. Cambridge: Cambridge Univer-
need treatment with either PLT transfusions or TPO ago- sity Press; 2013. p. 385-408.
nists and that their PLT counts will gradually increase 9. Baer VL, Henry E, Lambert DK, et al. Implementing a
from a nadir on approximately Day 3 to the normal range program to improve compliance with neonatal intensive
over about 3 weeks without any specific intervention. care unit transfusion guidelines was accompanied by a
reduction in transfusion rate: a pre-post analysis within a
multihospital health care system. Transfusion 2011;51:
ACKNOWLEDGMENT
264-9.
The authors thank Martha C. Sola-Visner, Harvard University and 10. Christensen RD, Baer VL, Lambert DK, et al. Reference
Boston Childrens Hospital, for helpful discussions about this intervals for common coagulation tests of preterm infants.
topic and for suggestions in manuscript editing. We also thank Transfusion 2014;54:627-32.
Erick Henry, MPH, Intermountain Healthcare, for assistance with 11. Christensen RD, Yaish HM, Lemons RS. Neonatal hemo-
the graphics. lytic jaundice: morphologic features of erythrocytes that
will help you diagnose the underlying condition. Neonatol-
ogy 2014;105:243-9.
CONFLICT OF INTEREST
12. Wiedmeier SE, Henry E, Sola-Visner MC, et al. Platelet ref-
The authors have disclosed no conflicts of interest. erence ranges for neonates, defined using data from over
47,000 patients in a multihospital healthcare system.
J Perinatol 2009;29:130-6.
REFERENCES 13. Christensen RD, Sheffield MJ, Lambert DK, et al. Effect of
1. Ferrer-Marin F, Stanworth S, Josephson C, et al. Distinct therapeutic hypothermia in neonates with hypoxic-
differences in platelet production and function between ischemic encephalopathy on platelet function. Neonatol-
neonates and adults: implications for platelet transfusion ogy 2012;10:91-4.
practice. Transfusion 2013;53:2814-21. 14. Christensen RD, Henry E, Andres RL, et al. Reference
2. Josephson CD, Glynn SA, Kleinman SH, et al.; State-of-the ranges for blood concentrations of nucleated red blood
Science Symposium Transfusion Medicine Committee. A cells in neonates. Neonatology 2011;99:289-94.
multidisciplinary think tank: the top 10 clinical trial 15. Christensen RD, Lambert DK, Richards DS. Estimating the
opportunities in transfusion medicine from the National nucleated red blood cell emergence time in neonates.
Heart, Lung, and Blood Institute-sponsored 2009 state-of- J Perinatol 2014;34:116-9.
the-science symposium. Transfusion 2011;51:828-41. 16. Roberts I, Stanworth S, Murray NA. Thrombocytopenia in
3. Sola-Visner MC, Saxonhouse MA. Acquired thrombocyto- the neonate. Blood Rev 2008;22:173-86.
penia. In: de Alarcn PA, Werner EJ, Christensen RD, 17. Sola-Visner M. Platelets in the neonatal period: develop-
editors. 2nd ed. Neonatal hematology. Cambridge: Cam- mental differences in platelet production, function, and
bridge University Press; 2013. p. 162-4. hemostasis and the potential impact of therapies.
4. Task Force on Neonatal Encephalopathy. Acid-base param- Hematology Am Soc Hematol Educ Program 2012;2012:
eters of umbilical cord blood. In: DAlton ME, editor. Neo- 506-11.
natal encephalopathy and neurologic outcome. 2nd ed. 18. Castle V, Coates G, Mitchell GL, et al. The effect of hypoxia
Washington, DC: American College of Obstetricians and on platelet survival and site of sequestration in the
Gynecologists; 2014. p. 94-7. newborn rabbit. Thromb Haemost 1988;59:45-8.
19. Gerday E, Baer VL, Lambert DK, et al. Testing platelet mass 26. Josephson CD, Su LL, Christensen RD, et al. Platelet trans-
versus platelet count to guide platelet transfusions in the fusion practices among neonatologists in the United States
neonatal intensive care unit. Transfusion 2009;49:2034-9. and Canada: results of a survey. Pediatrics 2009;123:278-
20. McDonald TP, Cottrell MB, Clift RE, et al. Effects of 85.
hypoxia on megakaryocyte size and number of C3H and 27. Venkatesh V, Curley AE, Clarke P, et al. Do we know when
BALB/c mice. Proc Soc Exp Biol Med 1992;199:287-90. to treat neonatal thrombocytopaenia? Arch Dis Child Fetal
21. Saxonhouse MA, Rimsza LM, Christensen RD, et al. Effects Neonatal Ed 2013;98:F380-2.
of anoxia on megakaryocyte progenitors derived from cord 28. Stanworth SJ, Clarke P, Watts T, et al.; Platelets and Neona-
blood CD34pos cells. Eur J Haematol 2003;71:359-65. tal Transfusion Study Group. Prospective, observational
22. Saxonhouse MA, Rimsza LM, Stevens G, et al. Effects of study of outcomes in neonates with severe thrombocyto-
hypoxia on megakaryocyte progenitors obtained from the penia. Pediatrics 2009;124:e826-34.
umbilical cord blood of term and preterm neonates. Biol 29. Oncel MY, Erdeve O, Calisici E, et al. The effect of whole-
Neonate 2006;89:104-8. body cooling on hematological and coagulation param-
23. LaIuppa JA, Papoutsakis ET, Miller WM. Oxygen tension eters in asphyxic newborns. Pediatr Hematol Oncol 2013;
alters the effects of cytokines on the megakaryocyte, eryth- 30:246-52.
rocyte, and granulocyte lineages. Exp Hematol 1998;26: 30. Groenendaal F, Brouwer AJ. Clinical aspects of induced
835-43. hypothermia in full term neonates with perinatal asphyxia.
24. Sallmon H, Gutti RK, Ferrer-Marin F, et al. Increasing Early Hum Develop 2009;85:73-6.
platelets without transfusion: is it time to introduce novel 31. Rogers GM, Lehman CM. Diagnostic approach to the
thrombopoietic agents in neonatal care? J Perinatol 2010; bleeding disorders. In: Greer JP, Arber DA, Glader B, et al.,
30:765-9. editors. Philadelphia (PA): Wolters Kluwer; 2014.
25. Cremer M, Sola-Visner M, Roll S, et al. Platelet transfusions Wintrobes clinical hematology. 13th ed. p. 1050.
in neonates: practices in the United States vary signifi-
cantly from those in Austria, Germany, and Switzerland.
Transfusion 2011;51:2634-41.