Platelet (PLT) transfusions can be lifesaving for certain thrombocytopenic neonates. However, a recent multidisciplinary NHLBI “think tank” expressed the concern that some neonates may be receiving PLT transfusions unnecessarily, thereby conveying little or no benefit with an unfavorable benefit-to-risk ratio.
One neonatal condition for which PLT transfusions are sometimes given, yet the supporting evidence is meager, is an entity termed the “thrombocytopenia of perinatal asphyxia.” Although this condition has been the subject of previous reports, many aspects are unclear. Sola-Visner and Saxonhouse recently commented that for this condition the causative mechanism, the usual range of severity, the typical duration, and the role of PLT transfusions are all relatively unknown.
As a step toward enhancing the knowledge base of the thrombocytopenia of perinatal asphyxia, we conducted a retrospective analysis of all neonates in the past 9 years who met the definition of “perinatal asphyxia.” This was done to find the incidence, severity, and usual duration of this variety of thrombocytopenia. Next we sought to understand the value of PLT transfusions in these patients by examining, among those where a PLT transfusion was given, the reasons for and response to transfusion. We then attempted to identify the underlying cause of this variety of thrombocytopenia and to assess whether therapeutic hypothermia (cooling) used in the treatment of neonates with perinatal asphyxia played a role in the pathogenesis of the thrombocytopenia. Last, we sought to determine the implication of finding thrombocytopenia after perinatal asphyxia on the mortality rate.
- Top of page
- Materials and Methods
- Conflict of Interest
Thrombocytopenia is a relatively common problem among patients in NICUs.[16, 17] The majority of neonates with thrombocytopenia are small and ill and have an acquired variety of consumptive thrombocytopenia accompanying bacterial or fungal sepsis or necrotizing enterocolitis.[3, 16] A different variety of thrombocytopenia has been described among neonates after perinatal asphyxia.[3, 18] Although DIC can occur in such patients, and can cause thrombocytopenia from PLT consumption, the majority of neonates reported with thrombocytopenia after birth asphyxia do not have evidence of consumptive coagulopathy. This study was an attempt to better define the thrombocytopenia of perinatal asphyxia as a needed step toward evidence-based NICU PLT transfusion practices.
We began our study by identifying all patients in the past 9 years at least 35 weeks' gestation who met a cord blood gas definition of perinatal asphyxia and had at least two PLT counts drawn during their first week. Among 375 such neonates we sought the incidence of thrombocytopenia, and among those we sought information on the severity and duration of the thrombocytopenia and its underlying mechanism. Our findings led us to speculate that the thrombocytopenia of perinatal asphyxia is most likely: 1) an authentic entity, although the condition can cooccur with DIC; 2) probably a hyporegenerative condition, as opposed to being due entirely to accelerated PLT consumption and/or destruction; 3) associated with intrauterine hypoxemia; 4) only moderately severe (mean nadir count, 75 × 109/L); and 5) self-limited, typically with a nadir count on Day 3 and a duration of about 3 weeks.
This was a retrospective study, and we acknowledge significant problems with that approach. For instance, PLT counts after PLT transfusions were not obtained on a standard schedule, PLT transfusions were given to some and not to others with no clear explanation, surely some relevant data were sometimes missing from the charts or electronic databases, and bias may have been unintentionally introduced. Also, changes in obstetric and neonatology practices occurred during the 9 years of study. Examples of practice changes we recognize include the marked overall reduction in NICU transfusion during this period and the specific reduction in PLT transfusions on the basis of a change from PLT count–based guidelines to PLT mass–based guidelines. Also, therapeutic hypothermia was introduced in 2008. These flaws limit the confidence in our conclusions, but the data do provide, in our opinion, reasonable hypotheses for future observational and prospective testing.
The molecular mechanisms resulting in the thrombocytopenia of perinatal asphyxia are obscure, but several experiments have generated potential explanations. For instance, McDonald and colleagues subjected adult mice to hypoxia and found a decrease in the size of individual megakaryocytes in the marrow and an overall reduction in marrow megakaryocyte mass, suggesting that the primary kinetic mechanism was reduced PLT production. Saxonhouse and coworkers evaluated whether megakaryocyte progenitors are damaged directly by hypoxia. He found no evidence for this by culturing CD34(+) hematopoietic progenitor cells obtained from umbilical cord blood in 0% versus 20% oxygen environments. The anoxic environment did not result in CD34(+) cell death or diminished megakaryocyte clonogenic potential. In a follow-up study Saxonhouse and coworkers cocultured CD34(+) progenitors with or without mononuclear (accessory) cells under various oxygen environments (1, 5, and 20%). Although the progenitors themselves were unaffected by the ambient oxygen concentration, the cocultures with accessory cells showed a progressive reduction in megakaryocytic clones with decreasing O2 concentrations. The conclusion was that nonprogenitor cells in the hematopoietic microenvironment were likely damaged by hypoxia resulting in impaired megakaryocytopoiesis. This conclusion is consistent with the in vitro experiments of LaIuppa and colleagues showing that hypoxia alters the effects of cytokines on megakaryocytic progenitors.
We suspect that the new thrombopoietin (TPO) receptor agonists romiplostim and eltrombopag will not be of value in treating the thrombocytopenia of perinatal asphyxia. Three reasons form the basis of our assumption: 1) These medications require approximately 14 days between commencement of dosing and a significant increase in PLT count, and the duration of this variety of thrombocytopenia is generally only 3 weeks. 2) Ninety percent of neonates with this disorder will not decrease their PLT count below approximately 40 × 109/L and thus may not need any specific treatment. 3) We suspect that impaired TPO production is not a pathogenic component of this disorder, because McDonald and coworkers reported that, in mice, hypoxia does not reduce TPO production.
One treatment option for neonates with the thrombocytopenia of perinatal asphyxia is PLT transfusion, but much investigation is needed to determine the risks and benefits of transfusions for neonates with this condition and to establish transfusion guidelines for this group. As highlighted by Josephson and colleagues in a recent symposium, PLT transfusion strategies for neonates currently lack a strong evidence base, yet this is needed as a high priority. From our present findings we conclude that most neonates with this variety of thrombocytopenia will have nadir PLT counts of higher than 40 × 109/L (the mean nadir count was 75 × 109/L; 90% CI, 35.7 × 109-128.6 × 109/L). Surveys of neonatologists in Central Europe and North America agree that PLT transfusions are recommended when PLT counts fall below 20 × 109/L, but counts that low should be quite rare in the thrombocytopenia of perinatal asphyxia. Whether transfusions should be recommended when the counts are in the 40 × 109 to 60 × 109/L range is not clear.[2, 27] Stanworth and coworkers reported a prospective observational study of 194 neonates with thrombocytopenia, where a subgroup of 47 had their lowest count in the 40 × 109 to 60 × 109/L range. Only two of those who were at least 35 weeks' gestation had a major hemorrhage and both of those had DIC. Thus we judge it is very likely that late preterm and term neonates with this variety of thrombocytopenia, with PLT counts in the 40 × 109 to 60 × 109/L and free of other complications, should do well without PLT transfusions. Thirty of the 96 neonates in this analysis with this variety of thrombocytopenia received one or more PLT transfusions. In retrospect, we question how many of these, if any, were beneficial. None of the 30 had a PLT count of fewer than 20 × 109/L, almost half had a pretransfusion PLT count of more than 50 × 109/L, and all 30 were transfused prophylactically with no bleeding.
Our data suggest that therapeutic hypothermia does not independently cause neonatal thrombocytopenia. Hypothermia for 72 hours has recently become a common and effective means of improving neurodevelopment of term neonates after perinatal asphyxia.[29, 30] Hypothermia does adversely affect PLT function, as manifested by prolongation of the bleeding time and lengthening of the PFA100 closure time. Once rewarming has occurred, PLT function rapidly normalizes. Additional study is needed to determine whether PLT dysfunction during therapeutic hypothermia warrants giving PLT transfusions at higher PLT counts than for neonates not on therapeutic hypothermia.
Taken together our study results suggest that approximately 30% of term and late preterm neonates with a cord pH of not more than 6.99 and/or a base deficit of at least 16 mmol/L will have a condition that can be called the thrombocytopenia of perinatal asphyxia. A minority of these neonates will have coexisting severe DIC. Likely those will be the most ill patients, with elevated coagulation times and D-dimers, very low and rapidly decreasing PLT counts, persistent hypotension and acidosis, and a poor prognosis for survival. However, we speculate that the great majority of neonates with thrombocytopenia after perinatal asphyxia will have a much more benign condition, with PLT counts that do not fall below 40 × 109/L. We maintain that these should not need treatment with either PLT transfusions or TPO agonists and that their PLT counts will gradually increase from a nadir on approximately Day 3 to the normal range over about 3 weeks without any specific intervention.