To evaluate sodium valproate-induced hemostatic side effects in children. A variety of both pro- and anticoagulatory parameters were longitudinally investigated in 80 children before therapy and up to 720 days after initiation of sodium valproate (VPA) therapy. VPA caused a significant reduction in platelet count (309 000/µl ± 122 000 before treatment to 261 000/µl ± 150 000 under VPA therapy, p = 0.007). However platelet function was not impaired. While vWF antigen was reduced during VPA therapy (1.05 U/ml ± 0.4 U/ml before therapy, 0.95 ± 0.4 U/ml under VPA therapy), the in vivo activity of vWF (ratio between function and antigen concentration) increased significantly (1.06 ± 0.2 before therapy, 1.36 ± 0.3 under VPA therapy, p = 0.01). Both procoagulatory and anticoagulatory factors were significantly reduced (fibrinogen: 264.5 ± 64.5 mg/dl before therapy, 221.4 ± 47.5 mg/dl under therapy, p = 0.001; protein C: 81.3 % ± 18 before therapy, 65.6 % ± 21.4 under VPA therapy, p = 0.005, antithrombin: 122.7 % ± 23.7 before therapy, 101.7 % ± 18 under VPA therapy, p = 0.04). With the exception of fibrinogen, these effects were identical in children treated either with monotherapy or with polytherapy. Besides already known alterations of a variety of procoagulatory parameters, a relevant influence of VPA on the anticoagulatory system is demonstrated. We hypothesize that this additional alteration of anticoagulatory parameters might reduce the absolute bleeding risk of children treated with VPA.