Neonatal thrombocytopenia : from bleeding risk to new therapies on the horizon

Abstract

Background: Thrombocytopenia is common in the Neonatal Intensive Care Unit and if severe, it is treated with platelet transfusions to prevent bleeding. Previous studies found a weak correlation between thrombocytopenia and incidence of bleeding. The poor predictive value of the platelet count has highlighted the need for better tests to assess bleeding risk and to guide transfusion decisions. Romiplostim (ROM) and eltrombopag (ELT) – two thrombopoietin (TPO) mimetics approved for treatment of thrombocytopenia in adults – may benefit thrombocytopenic neonates by potentially reducing platelet transfusions.

Objectives: First, the relationship between platelet count and in vitro bleeding time (PFA-100 closure times, CTs) in thrombocytopenic neonates was evaluated. Second, other risk factors associated with CTs were determined. Third, the relationship between the PFA-100 CTs and clinical bleeding was assessed. Fourth, in vitro responses of human neonatal vs. adult megakaryocyte (MK) progenitors to TPO, ROM and ELT, were compared.

Methods: We conducted a single institution cross-sectional study (Study 1), and a multicenter prospective longitudinal study (Study 2). Blood samples from thrombocytopenic neonates were tested with PFA-100 to measure in vitro bleeding time. In addition to platelet counts, other variables including severity of illness were obtained. Bleedings were quantified using the Neonatal Bleeding Assessment Tool (NeoBAT). In a pre-clinical study (Study 3), hematopoietic progenitor (CD34+) cells isolated from umbilical cord blood (CB) and adult peripheral blood (PB) were cultured in escalating concentrations of TPO, ROM, or ELT. After 14 days, the number of MKs and their maturation (CD42b expression) and ploidy were evaluated by flow cytometry.

Results: In Study 1, we found that all infants with platelet counts >90x109/L had normal CT-ADPs. At platelet counts <90x109/L, 74 % (14/19) of the infants had normal or minimally prolonged CT-ADPs, while 16 % (3/19) had markedly prolonged values. Study 2 showed that in preterm infants with gestational age <27 weeks and platelet counts <100×109/L, CT-ADP but not platelet counts was associated with NeoBAT scores. This association was robust also after adjustment for demographic and clinical variables. Study 3 found that with escalating concentrations of TPO, ROM, or ELT, CB progenitors generated 10 times more MKs than PB-MKs. At low concentrations, ELT stimulated megakaryopoiesis, but at high concentrations, ELT suppressed MK differentiation and proliferation via its intracellular iron chelating properties.

Conclusions: The CT-ADP test (an in vitro test of whole blood primary hemostasis) is a significantly better marker of bleeding risk than platelet counts among thrombocytopenic preterm neonates. TPO mimetics increased the numbers of cord blood MKs but did not influence their maturation or ploidy level. ELT generated fewer MKs than TPO or ROM. At low concentrations ELT stimulated megakaryopoiesis, but higher concentrations resulted in reduced MK differentiation and proliferation. These findings may have implications for the management of thrombocytopenic infants.