How I diagnose and treat neonatal thrombocytopenia

Thrombocytopenia is a frequent observation in babies admitted to neonatal intensive care units (NICUs) but also in otherwise healthy infants. Mild and self-limiting thrombocytopenia in infants with intrauterine growth restriction or in those born to mothers with preeclampsia has even been described as physiological responses to stress in neonates.¹ However, in other settings, thrombocytopenia can indicate a broad array of hematological or nonhematological disorders with significant diagnostic and management challenges.^2-5 The etiologies of neonatal thrombocytopenia range from rare inherited disorders of platelet production to more common acquired causes, such as immune-mediated thrombocytopenia and sepsis.^1,6 However, despite this diverse etiology, the main treatment for thrombocytopenia in clinical practice remains platelet transfusion.^7,8 

Prematurity and its complications are the most common reasons for hospital admission in newborns. All preterm babies under 32 weeks require intensive care, and many have to remain in the hospital for extended periods. Advances in care mean that neonates born as early as 22 and 23 weeks of gestation are now surviving.^9,10 The key management challenge for premature infants is to minimize the long-term consequences of very early life care, especially neurodevelopmental impairment.¹¹ A mainstay of medical treatment in an NICU is supportive care, including respiratory support, thermoregulation, nutritional management, infection management, and blood transfusion. Although intensive care given to preterm neonates is essential for their survival, each intervention also carries risk. The entire hospital course for preterm neonates presents significant challenges for families.

Thrombocytopenia (platelet count <150 × 10⁹/L) is a prevalent laboratory abnormality observed in ∼1% to 4% of all neonates at birth but up to ∼10% to 25% of those admitted to a (high-dependency) NICU and 50% of those with a birth weight <1000 g.^1,12-17 Unlike thrombocytopenia in patients receiving myeloablative chemotherapy, the trajectory of neonatal thrombocytopenia is highly variable, reflecting the diverse etiologies. Thrombocytopenia in some neonates is transient, but in others, it may persist and fluctuate over prolonged periods, even months or years.^14,18 Bleeding may be a major clinical sequela of thrombocytopenia yet correlates poorly with platelet count, as will be discussed later in this article.

The different temporal patterns of thrombocytopenia enable classification of thrombocytopenia in neonates into the following categories, based on the timing of presentation with respect to birth⁴ (Table 1; Figure 1):

1. Fetal thrombocytopenia: thrombocytopenia before delivery is usually discovered via fetal blood sampling performed because of ultrasound findings suggestive of congenital infection, aneuploidy, or genetic thrombocytopenia or because of maternal autoimmune disease, such as immune thrombocytopenic purpura or alloimmune disease.

2. Early thrombocytopenia (<72 hours from birth): disorders that develop before birth but are discovered early in life, or disorders that develop early after delivery. This accounts for 75% of all episodes of thrombocytopenia in patients admitted to an NICU¹⁵ and commonly affects preterm neonates from pregnancies complicated because of placental insufficiency, perinatal asphyxia, or maternal illness, such as pre-eclampsia or diabetes.^19-21 Thrombocytopenia arising from these causes is typically preceded by healthy or mildly reduced platelet counts at birth, falling to nadir counts of >50 × 10⁹/L at 4 or 5 days of life before spontaneous recovery by day 7 or 14.²² By contrast, more severe thrombocytopenia (<50 × 10⁹/L), which cannot be completely explained by these causes or shows a different trajectory, such as persistence for >7 or 10 days, may be associated with fetal onset disorders, such as fetal and neonatal alloimmune thrombocytopenia (FNAIT), congenital infection, aneuploidy, or genetic thrombocytopenia or with early perinatal infection, including group B Streptococcus infections.

3. Late thrombocytopenia (≥72 hours from birth): most commonly explained by late-onset sepsis and necrotizing enterocolitis (NEC). For these and other causes, including spontaneous intestinal perforation, there are typically additional clinical and laboratory abnormalities pointing toward such diagnoses, but the thrombocytopenia may be severe (<50 × 10⁹/L) and rapidly progressive and/or prolonged.^4,14,23,24 

In this review, we present a series of case scenarios with thrombocytopenia of different etiologies that illustrate the importance of correct diagnosis and intervention. We consider the following questions:

1. What are the causes of thrombocytopenia and how are they diagnosed?

2. How does thrombocytopenia affect clinical outcomes, including bleeding?

3. How should thrombocytopenia be managed with platelet transfusion or other interventions?

4. How can new genetic technologies assist in the diagnosis and treatment of neonatal thrombocytopenia?

A baby born at 25 weeks of gestation is now postnatal age day 11 and is being ventilated in the NICU. The baby is clinically stable, although the clinical team is concerned about abdominal distention and a rising serum C-reactive protein level. The platelet count measured as part of routine daily testing has fallen gradually from 150 × 10⁹/L to 30 × 10⁹/L. There is no evidence of bleeding.

This case illustrates the importance of clinical context and supporting laboratory findings in resolving the differential diagnoses of thrombocytopenia. The initial diagnostic approach for a neonate who is thrombocytopenic is based on gestation at birth, time of the first diagnosis of thrombocytopenia, antenatal and obstetric history, family history, and current clinical status, including features of comorbidities and bleeding. Thrombocytopenia that is late (>72 hours) in a neonate who is critically ill is much more likely to be caused by sepsis, NEC, or other acquired causes such as spontaneous intestinal perforation, in which there is a single perforation of the bowel, typically in the terminal ileum, rather than alternative explanations such as FNAIT, in which neonates usually appear generally well. Sepsis may be associated with increased mean platelet volume (MPV) and immature platelet fraction (IPF) and the appearance of large platelets on the blood film, reflecting increased platelet turnover and production.^25-27 However, these parameters may be significantly influenced by variables such as the transport time of blood samples before analysis and the type of hematology analyzer. Increased MPV and IPF are also hallmarks of certain types of genetic thrombocytopenia (see “Case 2”). The history and rising C-reactive protein level in case 1 point toward thrombocytopenia arising from causes such as infection or NEC; further investigation and management of these significant sources of neonatal morbidity are outside the scope of this review.

Platelets contribute significantly to innate immunity via direct interactions with bacterial pathogens and effector cells, such as neutrophils.²⁸ The association between thrombocytopenia and infection has been recognized for many years,²⁹ especially in neonates.²⁴ Multiple pathophysiological pathways have been identified, including abnormal platelet interactions with microbial pathogens, immune cells, or activated endothelial cells or hepatocytes, leading to increased platelet clearance.³⁰ Infection is also associated with increased platelet apoptosis.³¹ NEC is a life-threatening illness in which there is inflammation of the intestine leading to bacterial translocation into the circulation and risk of intestinal perforation; it is largely restricted to neonates.³² Thrombocytopenia commonly accompanies NEC because of the profound inflammatory response leading to complement and coagulation activation and platelet consumption. This may evolve into the coagulopathy of neonatal disseminated intravascular coagulation, an indicator of clinical deterioration. Evidence for disseminated intravascular coagulation management is limited to small trials and comprises proactive blood component support.^33-35 Although this thrombocytopenia is associated with increased neonatal mortality, this likely reflects an indirect effect from underlying sepsis rather than a direct consequence of thrombocytopenia. Thrombocytopenia in neonates at early postnatal ages may have other infective causes, such as intrauterine infections and congenital cytomegalovirus, which is, however, beyond the scope of this article. A general clinical decision-making aid for neonatal thrombocytopenia is shown in Figure 1.

A term male with abnormal scalp and shoulder bruising noted 3 days after a forceps delivery had a platelet count of 84 × 10⁹/L. A blood smear confirmed thrombocytopenia but otherwise showed healthy blood cell morphology. The mother had a platelet count of 105 × 10⁹/L in early pregnancy, which remained stable and was attributed to gestational thrombocytopenia. Delivery was complicated because of a 900 mL postpartum hemorrhage. The mother had a history of heavy menstrual bleeding, frequent epistaxis, and prolonged bleeding after a dental extraction but had no prior hematological diagnoses. Both her father and a paternal uncle had died of acute myeloid leukemia.

Genetic thrombocytopenia is an important differential diagnosis and may sometimes be accompanied by syndromic features in neonates with thrombocytopenia. Many forms of genetic thrombocytopenia have long-term health implications for the neonate, family members, and furture children for the parents. Case 2 illustrates that although the clinical characteristics alone of the neonate were insufficient to indicate genetic thrombocytopenia, there were important clues in the family history to support this diagnosis. These clues included the maternal thrombocytopenia detected at the antenatal booking visit. Although thrombocytopenia has multiple causes in pregnancy, genetic thrombocytopenia is often revealed for the first time when a complete blood count is performed at early pregnancy assessment. As in this case, adults and children with genetic thrombocytopenia frequently receive incorrect diagnoses of more common acquired thrombocytopenias and sometimes undergo unnecessary treatments, such as immunosuppression. The mother’s abnormal bleeding history is a key pointer toward a hemostatic defect greater than what was expected, given her platelet count, and suggestive of certain forms of genetic thrombocytopenia in which thrombocytopenia is accompanied by defective platelet function.

In this example, there was sufficient suspicion of genetic thrombocytopenia for the mother to receive genetic counseling and then testing of both herself and the neonate for bleeding and platelet disorders using a 90-gene next-generation sequencing panel.³⁶ Both mother and neonate harbored a monoallelic missense variant in RUNX1, reported by the genomics multidisciplinary team as likely pathogenic and indicative of familial platelet disorder with associated myeloid malignancies (FPDMM).³⁷ This diagnosis accounts for the inheritance pattern of thrombocytopenia in this pedigree and the maternal history of disproportionate bleeding, explained by the platelet storage pool defect, which is a feature of FPDMM.³⁸ The wider family history of hematological malignancy further supports this diagnosis because FPDMM confers a lifetime risk of ∼20% to 50% for myelodysplasia or acute myeloid leukemia.³⁹ 

Genetic thrombocytopenia is an increasingly well-characterized group of Mendelian disorders that are genetically and phenotypically heterogeneous.^40,41 The prevalence of individual genetic thrombocytopenia disorders in adult populations ranges from 1:5000 to <1:1 000 000; however, there are no reliable estimates of the composite prevalence in neonates presenting with thrombocytopenia. For reference, the more common glycoprotein deficiencies are probably the best characterized group of genetic platelet disorders in the United Kingdom, the data for which have been submitted via the United Kingdom Haemophilia Centre Doctors’ Organisation registry. Bernard-Soulier syndrome may present with mild thrombocytopenia and, based on a UK population of 67.5 million, has an estimated prevalence of 1 in 650 000 (102 cases).⁴² However, when considering different genetic thrombocytopenias, we are likely to underrecognize milder phenotypes.

More frequently encountered forms of genetic thrombocytopenia among older children and adults are caused by monoallelic pathogenic variants in MYH9, TUBB1, ACTN1, and GP1BB, which cause mild reductions in platelet count but seldom abnormal bleeding. These seldom present clinically in neonates but may be identified as incidental discoveries.^43-46 By contrast, rarer disorders such as thrombocytopenia absent radius syndrome, congenital amegakaryocytic thrombocytopenia, and amegakaryocytic thrombocytopenia with radioulnar synostosis caused by biallelic variants in RBM8A, MPL, and HOXA1, respectively, often cause more severe thrombocytopenia, presenting in neonates alongside nonhematological manifestations.^47,48 The spectrum of genetic thrombocytopenia also includes X-linked thrombocytopenia with dyserythropoiesis or thalassemia due to GATA1 variants, in which there may be other hematological features in neonates or which develop later in childhood.⁴⁹ The association of thrombocytopenia with reduced platelet size, eczema, and immunodeficiency may indicate the X-linked disorder Wiskott-Aldrich syndrome.⁵⁰ Thrombocytopenia in neonates may be associated with the chromosomal aneuploidy trisomy 18, 13, or 21, likely to be the most common form of genetic thrombocytopenia among patients in NICUs. Thrombocytopenia is part of the disease phenotype of neonates with triploidy, Turner syndrome,⁵¹ or chromosomal deletions causing Paris-Trousseau/Jacobsen (11q) or DiGeorge/velocardiofacial (22q11.2) syndrome, in which thrombocytopenia is accompanied by syndromic features that may be detected antenatally.^52,53 

Incompletely penetrant late manifestations, such as hematological malignancy, are a feature of genetic thrombocytopenia caused by pathogenic RUNX1, ANKRD26, and ETV6 variants.^54,55 Detection of pathogenic variants in these genes requires careful consideration of family counseling, cascade screening of family members, and the performance of long-term surveillance complete blood count or bone marrow tests in affected individuals.^56,57 An updated list of genes associated with monogenic genetic thrombocytopenia is presented in Table 2.

Detection and molecular classification of genetic thrombocytopenia are essential for ongoing management of neonates with thrombocytopenia and, as appropriate, for family screening. However, the wide differential diagnosis of genetic thrombocytopenia necessitates careful assessment of other hematological and nonhematological features in the affected neonate and, sometimes, in family members. We recommend early consultation between neonatologists and either hematological or clinical geneticists for neonates with persistent thrombocytopenia of unknown cause and, particularly, if there are syndromic features or potentially relevant family history.

Multiple specialist laboratory tests, such as flow cytometric measurement of platelet surface proteins, platelet ultrastructure, functional testing, and plasma biomarker assays may give diagnostically useful information by suggesting individual forms of genetic thrombocytopenia.^61,62 The presence of large or small platelets on the blood smear or deviating values for MPV in the complete blood count have been proposed as a broader tool to distinguish genetic thrombocytopenia from immune or other acquired thrombocytopenias in older children and adults.^63,64 However, the diagnostic reliability of using platelet size in neonates remains untested and may be confounded by observations that MPV is increased in other causes of neonatal thrombocytopenia, including sepsis.⁶⁵ IPF is usually not elevated in genetic thrombocytopenia, with the exception of some macrothrombocytopenias, such as MYH9-related disorder, thereby potentially assisting differentiation from immune-mediated thrombocytopenias, such as FNAIT.

Definitive diagnosis of genetic thrombocytopenia requires genomic testing, now enabled by next-generation technologies that allow high-throughput analysis of large gene panels, exomes, or genomes. Precision diagnosis using these technologies is reported to achieve diagnosis in up to 48% of adults with potential Mendelian thrombocytopenic disorders.^36,66 These tests may also be accessible as a rapid diagnostic test in some health care services for neonates who are seriously ill.^67,68 Effective genomic diagnosis requires standardized selection of candidate gene panels and rigorous variant calling and reporting strategies, often adapted on a gene-by-gene basis.⁶⁹ Accurate consideration of all the phenotypic characteristics of neonates with suspected genetic thrombocytopenia, phenotype of family members, and inferred pattern of inheritance are essential prerequisites for accurate genomic diagnosis, which is best supported through the structures of multidisciplinary teams. Despite these advances, there remain practical issues in the administration of genomic testing, particularly in distinguishing pathogenic from nonpathogenic variants for many genes and a lack of clarity about the position of genomic testing in the wider diagnostic pathway for neonatal thrombocytopenia.

With the exception of the small subset of genetic thrombocytopenia presenting with severe thrombocytopenia, platelet transfusion or other prohemostatic interventions are seldom indicated in neonates with genetic thrombocytopenia. Potentially related and cryptic syndromic features in different genetic thrombocytopenias, such as hearing loss and cardiac or urogenital abnormalities and wider bone marrow failure disorders, which may accompany some genetic thrombocytopenia, must be considered carefully. The long-term management and follow-up of genetic thrombocytopenia is important and has recently been reviewed comprehensively.^70,71 

An otherwise healthy term baby is noted to have a widespread mild purpuric rash. The platelet count is 40 × 10⁹/L but drops to 27 × 10⁹/L the following day. This is the parents’ first child. The mother has a healthy platelet count and no history of abnormal bleeding.

This case of severe thrombocytopenia (platelet count <50 × 10⁹/L) in an otherwise healthy term baby is a classical presentation of FNAIT, which has an estimated incidence of ∼1 in 1500 pregnancies.⁷² The clinical course of presentation is often mild, although, as in this example, the platelet count may drop to a level that raises clinical concerns about bleeding. The thrombocytopenia is usually accompanied by increased IPF in the complete blood count test, reflecting platelet consumption and increased platelet production. The major clinical concern in fetuses and neonates with FNAIT is intracranial hemorrhage and, sometimes, other types of major bleeding.^73-75 

FNAIT arises from differences in common variants between the parental genes encoding immunogenic platelet surface proteins. These can result in the development of maternal antibodies directed against paternal antigens on fetal platelets, which, if transferred across the placenta, can result in fetal or neonatal thrombocytopenia. Although >30 different platelet-specific antigens have been implicated in the pathogenesis of FNAIT,⁷² common platelet antigens implicated in European-ancestry mothers are human platelet antigen 1a (HPA-1a) and HPA-5b, but there are different serotypes prevalent in other ancestries. Susceptibility to FNAIT is modified by other genetic loci, such as HLA-DRB3∗01:01, which increases the risk of HPA-1a immunization in pregnant women with HPA-1a.⁷⁶ Thrombocytopenia in FNAIT typically lasts from 1 to 6 weeks after birth. Although the severity of thrombocytopenia is generally a poor predictor of major bleeding, in a mouse model of immune neonatal thrombocytopenia, a platelet count <10% of normal conferred increased bleeding risk at specific developmental periods such as the first week of life.⁷⁷ 

Management is often needed before the results of diagnostic serological tests, which usually takes several days. International consensus guidelines for transfusion currently recommend that neonates who are severely thrombocytopenic and not bleeding should receive prophylactic platelet transfusions, with the aim of preventing the development of bleeding.⁷⁸ Platelet count thresholds for prophylactic transfusion are usually recommended between 25 × 10⁹/L and 50 × 10⁹/L. Although this is not supported by trial data specific to FNAIT, these thresholds appear reasonable but should be modifiable after consideration of factors such as age of the neonate, comorbidities, and history of bleeding in the neonate or other affected siblings. For example, higher platelet count thresholds for transfusion may be considered in neonates with active or prior bleeding, especially intraventricular hemorrhage. UK Blood Services stock platelets that are tested negative for the common HPA-1a/5b antigens, but if not available, standard HPA-unselected, random donor platelets may be given, with pending results of serological investigations.⁷⁹ Platelet count increments after transfusion should be recorded, and repeated transfusions considered if increments are poor. IV immunoglobulin may be considered in circumstances in which platelet transfusion is unavailable and there is life-threatening bleeding and/or a persisting platelet count of <30 ×10⁹/L with evidence of poor increments.^79,80 IV immunoglobulin may be associated with adverse events such as hemolysis and is unlikely to improve platelet counts for at least 24 hours.⁸¹ Outpatient follow-up of the neonate is essential to document complete resolution of thrombocytopenia. There should be a maternal referral to a high-risk obstetric service for future pregnancies.

There is a need for collaborative work to address FNAIT research uncertainties and advance care.⁸² An international web-based registry recently reported extreme thrombocytopenia (platelet count < 10 × 10⁹/L) in 24% of neonates with FNAIT but also identified considerable variation in practices for postnatal FNAIT treatments.⁸³ This registry provided evidence that platelet count increments were higher after HPA-matched transfusions than those after HPA-unmatched transfusions, although the effects of higher platelet increments on the risk of bleeding were less clear.⁸³ 

Standardization of scoring tools to reliably record and grade bleeding events has revealed insights about bleeding risk in neonates who are thrombocytopenic. When ascertained using a validated bleeding assessment tool specific to neonates, major or minor bleeding was found to occur in 37 of 146 (25%) infants but was more common in preterm infants.⁸⁴ In many studies, bleeding in neonates who are thrombocytopenic occurred across a wide range of platelet counts,^14,22,85-87 raising uncertainties about the predictive value of thrombocytopenia alone (rather than clinical factors) for neonatal bleeding,^88,89 similar to findings in adults⁹⁰ and older children.⁹¹ Caution should also be exercised in ensuring that low platelet counts generated by automated analyzers are not spurious. Assessment should always include inspection of the blood film to exclude platelet clumping (pseudothrombocytopenia), in which case, more reliable platelet count can often be obtained by analyzing citrate- rather than EDTA-anticoagulated blood samples.⁹² Bleeding arising from platelet dysfunction, which is a feature of some genetic thrombocytopenias, correlates poorly with the results of platelet function tests. Some more relevant risk factors for bleeding have been reported to include gestational age and NEC.⁹³ The absence of a clear relationship between platelet count and bleeding in neonates could reflect that the pathogenesis of major bleeding in this group is less dependent on the classical role of platelets in primary hemostasis. Instead, bleeding such as intraventricular hemorrhage may depend more on hemodynamic disturbances and the integrity of vascular tissues such as the endothelium.^94,95 Although platelets may contribute to endothelial cell integrity, this may require circulating platelet numbers lower than required for primary hemostasis.^96,97 

Platelet transfusions are a common prohemostatic intervention used in neonatal care and have been reported to be administered to ∼10% of preterm neonates at some stage during their inpatient stay.^1,14,17 Thresholds in platelet count at which transfusion is administered have been highly variable between neonatal centers over many years,^7,14,87,98 although recent national guidelines^79,99 have now provided more clarity. These guidelines include a stronger recommendation for a platelet count transfusion threshold of 25 × 10⁹/L for preterm neonates who are not bleeding, unless there are additional exceptional bleeding risk factors. This recommendation was informed by the PlaNeT-2/MATISSE study, a multicenter randomized controlled trial that compared liberal vs restrictive prophylactic platelet transfusions in preterm neonates.¹⁰⁰ The study completed a target recruitment of 660 neonates and achieved follow-up in 98% for the composite primary outcome of death or major bleeding. Neonates in the liberal arm (platelet counts maintained at >50 × 10⁹/L) had a significantly higher rate of death or major bleeding within 28 days of randomization compared with neonates in the restrictive arm (platelet counts maintained at >25 × 10⁹/L), and of neurodevelopmental impairment at a corrected age of 2 years.¹⁰¹ The findings of the PlaNeT-2/MATISSE study are supported by similar findings in smaller randomized trials of platelet count thresholds in neonates^86,102,103 and together inform stronger evidence that liberal transfusion thresholds do not improve clinical outcomes and indeed may be associated with harm.

Given the diverse etiologies of thrombocytopenia in neonates, a discussion point of the PlaNeT-2/MATISSE study was whether the findings were generalizable to all preterm neonates. Limitations of the PlaNeT-2/MATISSE study, as acknowledged by the trial authors, included the subgroups of enrolled neonates who had received platelet transfusions before randomization or who were recruited in the first few days of life, possibly reflecting concerns by clinicians about randomizing very sick or extremely premature neonates to lower platelet thresholds for transfusion. This was partly explored in a modeling analysis of PlaNet-2/MATISSE data that considered outcomes in neonates substratified into different degrees of baseline risk based on other risk factors, such as gestational age, postnatal age, diagnosis, and prior major bleeding.¹⁰⁴ This showed that the 25 × 10⁹/L threshold was associated with absolute-risk reduction for death or major bleeding in all subgroups, thereby supporting the general recommendation for the restrictive threshold of 25 × 10⁹/L in all preterm neonates.⁹⁹ The focus should now be on strategies for effective dissemination,^3,105 given the evidence of ongoing poor uptake of research findings.¹⁰⁶ The safety of platelet transfusion in NEC warrants additional scrutiny because platelet activation triggered by thrombin may contribute to the pathogenesis of NEC.¹⁰⁷ Consistent with this, some clinical studies have suggested a detrimental effect of platelet transfusion in this setting.^108-110 

It should be noted that the PlaNeT-2/MATISSE study tested different platelet count thresholds in preterm neonates who are not bleeding and not in neonates before major surgery, in which case, higher platelet count thresholds are often suggested for platelet transfusion.⁷⁹ The timings of platelet counts measured in the PlaNeT-2/MATISSE trial were those undertaken in routine clinical practice, typically once daily (morning), and there was no recommendation for more frequent monitoring of count increments, which would inevitably add to volumes for blood draws in small neonates. The majority of enrolled neonates received only 1 (or 2) platelet transfusions. Some babies do receive multiple repeated platelet transfusions, often in the context of refectory persisting thrombocytopenia associated with sepsis and NEC. Although a restrictive threshold (25 × 10⁹/L) should be adopted in these neonates, the benefits for repeated transfusions, often associated with minimal sustained platelet count increments, are sometimes questioned in clinical practice, and it may be prudent to avoid targeting a specific daily platelet count in these neonates who are not bleeding and to follow an individualized transfusion policy with clinical judgment.

Although the specification of platelet components has important implications for the efficacy and safety of transfusion in neonates, there is considerable variation between institutions and blood service operators. The heterogeneity between platelet components that arises from differences between blood donors/donations and changes in processing steps, such as pathogen reduction and storage times, is often not recognized by clinicians.^111,112 One contemporary illustration is the recent change (from 2020) in the recommendations for neonatal platelet processing, instituted by the National Health Service Blood and Transplant (England) after suboptimal results of neonatal platelet pH quality monitoring. To improve the pH at expiry and thereby the platelet quality, all neonatal platelet packs are now supplemented with a small volume of platelet additive solution (∼20% of the platelet pack volume).¹¹³ Such changes may have clinical implications and need clinical consultation and communication, although, in this case, it was considered that any dilution of the platelet concentrate would have minimal clinical impact for neonatal prophylactic transfusions.

Platelet concentrates may sometimes be issued with minor serotype incompatibilities to recipients, particularly if there is need to manage shortages in the platelet supply chain,¹¹⁴ but the full impact of this in preterm neonates is unclear.⁷ Most neonatologists prescribe from 10 to 20 mL/kg of platelet concentrates for neonates, which is a high volume compared with a standard therapeutic equivalent dose of 2 to 4 mL/kg in adults.¹¹⁵ Circulatory overload consequent of this relatively large volume has been suggested as an explanation for the poorer outcomes associated with liberal transfusion in the PlaNeT-2/MATISSE trial, although the delayed effect of the reported events relative to the time period when most platelet transfusions were given might argue against this hypothesis.¹⁰⁰ However, further recognition of this complication is currently hampered because of the absence of clear definitions of transfusion-associated circulatory overload in neonates. The Serious Hazards of Transfusion (an independent, professionally led hemovigilance system in the United Kingdom) annual report includes a chapter on transfusion reactions and adverse events in children and babies.¹¹⁶ An older analysis based on Serious Hazards of Transfusion reporting has suggested that such events in children and neonates are proportionately more common than in adults.¹¹⁷ Although some studies of basic neonatal platelet transfusion practice have been completed,¹¹⁸ further research is required to better define many common practices of blood transfusion. New genetic technologies may also offer significant advances in blood transfusion services by enabling selection of optimally matched product for transfusion based on genotype rather than serological test results. Potential applications may include the provision of platelets for neonates with alloimmune causes of thrombocytopenia.^119-121 

All platelet donations for neonates are collected from adult donors. Yet, the neonatal hemostatic system has important differences to that among older children and adults, and the full functional consequences of introducing adult donor platelets are not completely understood.^106,122,123 Neonatal platelets have distinct in vitro patterns of reactivity when compared with adult platelets but are generally hyporeactive to many activating agonists in vitro.^{72-75,124-127} Although this might suggest that neonates are at greater bleeding risks compared with adults, reduced platelet reactivity is offset in vivo by compensatory hemostatic mechanisms, such as higher hematocrit and circulating von Willebrand factor levels, resulting in a net effect of enhanced primary hemostasis compared with adults. It is unclear whether there is an association between adult platelet transfusions in neonates and thrombosis because there are currently no robust datasets available to assess this. However, it is noteworthy that the addition of adult platelets to neonatal blood in an ex vivo model of platelet transfusion in neonates resulted in enhanced platelet reactivity and greater clot viscoelastic strength, illustrating the potential for a prothrombotic effect in vivo.¹²³ 

This discussion has focused on platelet transfusion in preterm neonates, which is the most common clinical setting for transfusion. Causes of thrombocytopenia in term neonates are usually different (Table 1), and extrapolation of the results from trials in preterm neonates should be considered with caution. Given the emerging data that indicate adverse nonhemostatic effects of transfused platelets in children and adults,¹¹² it is reasonable to follow restrictive policies for prophylactic platelet transfusions in term infants with thresholds of ≤25 × 10⁹/L except in a setting of higher bleeding risk such as surgery or extracorporeal membrane oxygenation. There are limited alternatives to platelet transfusion in preterm neonates. Antifibrinolytics like tranexamic acid have been far less studied for neonates compared with that for adults. A recent trial of tranexamic acid in adult patients with hematological cancers and thrombocytopenia did not report any benefit for clinical bleeding.¹²⁸ Tranexamic acid is recommended in infant cardiac surgery with bypass,¹²⁹ but, currently, we would not recommend its widespread use in preterm neonates.

Neonatal thrombocytopenia is common and may require input from hematologists for diagnosis and management. Research has advanced our approaches to neonatal thrombocytopenia, including the application of rapid and comprehensive genetic testing for genetic thrombocytopenia and randomized trials of platelet transfusion thresholds. Despite this, our understanding remains incomplete in key areas, including the prediction of bleeding risk.^130,131 The impact of different platelet donation characteristics on outcomes may be greater in preterm neonates, given their biological immaturity, compared with older transfusion recipients. The nonhemostatic effects of platelets transfused into neonates are unclear, as are the reasons for harm as demonstrated in the PlaNeT2/MATISSE trial. It was a potentially informative observation in this trial that more surviving neonates in the liberal arm at 36 weeks of corrected age experienced bronchopulmonary dysplasia, a form of chronic lung disease in which developmental and immunological abnormalities lead to reduced alveoralization and a need for oxygen or respiratory support.⁸¹ This could represent a clearer signal suggesting that adverse immune effects can result from liberal platelet transfusion.^132,133 In addition, angiogenic signals, if provided through adult platelets, may precipitate bleeding during a vulnerable period of neonatal brain development.^134,135 

The authors thank Helen New and Martha Sola-Visner for their critical review and Kim Lacey for administration support.

Contribution: S.J.S. and A.D.M. wrote the manuscript.

Conflict-of-interest disclosure: S.J.S. reports receiving funds from government sources (National Institutes of Health Research and NHS Blood and Transplant) for research in the field of platelet transfusion. A.D.M. declares no competing financial interests.

Correspondence: Simon J. Stanworth, Radcliffe Department of Medicine, University of Oxford, Headley Way, Headington, Oxford OX3 9BQ, United Kingdom; e-mail: simon.stanworth@nhsbt.nhs.uk.