Thrombocytosis and thrombocytopenia in the NICU: incidence, mechanisms and treatments

Objective: Quantitative and qualitative platelet abnormalities of neonates must be defined using evidence-based reference ranges, constructed according to gestational and postnatal age. Methods: Platelet counts, and demographic and outcome data, were obtained from neonates in the Intermountain Healthcare system in the western USA and template bleeding times were determined from neonates in Italy. Results: Reference ranges were constructed by excluding values from neonates with diagnoses associated with abnormal platelet counts (small for gestational age (SGA), pregnancy-induced hypertension (PIH), infection and necrotizing enterocolitis (NEC)). Values remaining in the database after excluding these diagnoses were organized into 5th to 95th percentile ranges. At 23–25 weeks gestation, thrombocytopenia (<5th percentile) was defined by a platelet count <100,000/µl. Severe thrombocytopenia (platelet count <50,000/µl) occurred in 2.4% of neonatal intensive care unit (NICU) admissions and was largely due to acquired consumptive causes (bacterial and fungal sepsis, NEC and extracorporeal membrane oxygenation). No correlation was found between platelet count and subsequent central nervous system (CNS), pulmonary or gastrointestinal (GI) bleeding. The mortality rate did not correlate with the lowest platelet count but was proportionate to the number of platelet transfusions received. Platelet transfusions, administered according to guidelines, were given to 7% of NICU admissions, but a change in the guidelines from “count-based” to “mass-based” was associated with a reduction to 4%, with no increase in CNS, pulmonary, GI or cutaneous haemorrhage. Bleeding times were twice as long in neonates <33 weeks gestation as in term neonates, and shortened to term values by day of life ten. Conclusions: When reference ranges for platelets, appropriate to gestational and postnatal ages, are used, more uniformity occurs in definitions. This uniformity will foster consistency in diagnosis, treatment and outcomes-reporting.