Redefining neonatal thrombocytopenia: an evidence-based approach, improved by the refineR algorithm.

OBJECTIVE: A common definition of neonatal thrombocytopenia is a platelet count <150,000/µL. The international hematologic consensus designation of cytopenia is a blood count <2.5th percentile lower reference interval for an appropriate comparison population. We hypothesized that a new reference interval modeling algorithm, refineR, would generate more accurate platelet reference intervals and, therefore, more accurate and evidence-based definitions of neonatal thrombocytopenia. STUDY DESIGN: We utilized refineR to analyze platelet counts from large clinical databases to generate new neonatal platelet reference intervals. RESULTS: With 155,062 neonatal platelet counts, we created reference intervals for the day of birth and over the first 30 days. These included a lower (2.5th percentile) reference interval at birth for neonates ≥34 weeks of 119,600 platelets/µL (95% CI, 117,700 to 135,300/µL) and for neonates <34 weeks of 103,000 platelets/µL (98,700 to 105,600). CONCLUSION: Platelet counts of neonates at birth in the range of 120,000-150,000/µL are normal.