Of infants born to women with idiopathic thrombocytopenic purpura (ITP), about 10–15% have transient neonatal immune thrombocytopenic purpura (NITP). Of concern is the lack of a reliable predictor for NITP. We conducted a retrospective study of all pregnancies with ITP at Osaka University Hospital over the past 16 years analyzing a total of 127 pregnancies in 88 women with ITP to assess the predictive value of various clinical factors regarding neonatal platelet count in the current pregnancy. We also reviewed the literature concerning ITP in pregnancy and NITP prediction. Neonatal platelet counts were less than 100 × 109/L in 20 of 130 neonates (15.4%), less than 50 × 109/L in 11 neonates (8.5%), and less than 20 × 109/L in three neonates (2.3%). There was a strong correlation between the first and second sibling regarding the occurrence and the severity of NITP with Spearman correlation coefficient of 0.55 (P= 0.001) at birth and 0.63 (P< 0.0001) at nadir after birth. A maternal platelet count less than 100 × 109/L at delivery showed a statistical trend for an association with the occurrence of NITP (P= 0.043). Moreover, maternal ITP refractory to splenectomy correlated with a higher risk for fetal or neonatal ICH according to the review of the literature. In conclusion, pregnant women who have had a previous offspring with NITP or who have ITP refractory to splenectomy may be at particular risk of delivering an offspring with significant NITP. Management decisions, including mode of delivery, may be altered by the degree of risk for potentially severe NITP.