The role of HLA antibodies in neonatal alloimmune thrombocytopenia is controversial. We prospectively studied the sera of obstetric patients at delivery for HLA antibodies and correlated their presence with umbilical cord blood platelet counts. We studied 493 births at The Johns Hopkins Hospital comprising of 357 African American, 115 Caucasian, and 21 babies of other racial groups. One hundred and thirty nine mothers had HLA antibodies. Of these HLA alloimmunized mothers, only ten infants had platelet counts of 150,000/ microL or less. Three hundred and eight mothers with no detectable antibodies gave birth to 27 infants with platelet counts of 150,000/microL or less. Yates corrected Chi square analysis showed no significant relationship between maternal HLA alloimmunization and baby platelet count (p = 0.709). Only 8 of sixty cord sera from babies of HLA alloimmunized mothers were positive for HLA antibodies. The HLA cord blood antibody results were then correlated with the neonatal platelet counts. The Fisher's exact test showed no significant relationship between the presence of HLA antibodies in cord blood samples and neonatal platelet counts (p = 0.232). Although one third (31%) of mothers have HLA antibodies, neonatal thrombocytopenia is rarely associated with this finding. However, HLA antibodies can cross the placenta, and in these unusual cases, may be associated with a higher risk of neonatal thrombocytopenia.