Valproic acid and thrombocytopenia in children: a case-controlled retrospective study.

Thrombocytopenia in association with valproic acid (VPA) therapy has been reported in patients of various ages with incidences ranging from 1 to 32%. To evaluate this association in a pediatric population, we retrospectively studied 167 children treated with VPA at our institution between 1989 and 1993. Ninety-one patients on VPA monotherapy and 76 on VPA polytherapy (VPA plus 1 to 3 other antiepileptic drugs) were compared with 92 age- and sex-matched control patients treated with antiepileptic drugs other than VPA. Study variables included patient age, most recent platelet count, VPA dose, dose/kg, and serum VPA level. Thrombocytopenia, defined as a platelet count < 200 x 10(3)/mm3, was present in 21.6% of the children treated with VPA (26.4% in those on VPA monotherapy and 15.8% in those on VPA polytherapy) but in only 5.4% of controls untreated with VPA. Serum VPA level was the highest risk factor for the development of thrombocytopenia in these patients (P = .0001) and greater age also independently predicted thrombocytopenia. The dose of VPA or dose/kg were not independent predictors of thrombocytopenia (P = .0025). The degree of thrombocytopenia was mild among these patients, reflected in the absence of significant difference in mean platelet count between the three groups and the absence of any bleeding complication or excessive bruising. We conclude from these findings that the risk for children developing severe thrombocytopenia with bleeding complications while taking VPA is low and that mild thrombocytopenia does not necessarily mandate discontinuing the drug. Because higher serum VPA levels do predict thrombocytopenia, platelet counts should be closely monitored after dose escalation in such patients.