Valproate (VPA) is associated with a variety of idiosyncratic hematological effects including thrombocytopenia, inhibition of platelet aggregation, and fibrinogen depletion. This has led some investigators to recommend discontinuation of VPA therapy prior to elective surgical procedures. However, administration of VPA therapy is not altered prior to surgery at the authors' center and no VPA-associated bleeding complications have been observed. Therefore, a retrospective chart review was conducted to verify the clinical observations in patients who had undergone cortical resection while receiving antiepileptic drugs (AEDs). Baseline, surgical, and postoperative laboratory data were available for a total of 313 patients, 111 of whom were receiving VPA and 202 of whom were receiving AEDs without VPA (control patients). Eighty-seven percent of the patients receiving VPA were also being treated with at least one other AED. The control group was approximately equally divided between monotherapy (55%) and polytherapy (45%) treatments. Platelet counts were significantly lower in the control polytherapy (284 +/- 74 x 10(9)/L) and both VPA groups (279 +/- 113 x 10(9)/L) as compared with the control monotherapy group (314 +/- 85 x 10(9)/L). Baseline fibrinogen levels were significantly lower in the VPA than in the control groups (223 +/- 91 g/dl vs. 278 +/- 95 g/dl). Both pre- and postoperatively, the VPA group had lower red blood cells counts, hematocrit, and hemoglobin levels. There was no significant difference between groups in estimated blood loss during surgery or qualitative wound discharge postsurgery. There was only one case of a bleeding complication, which occurred 14 days postoperatively in a patient receiving carbamazepine monotherapy. The results of this study confirm the clinical observations of an absence of bleeding complications in patients receiving VPA therapy at the time of surgery, despite differences in laboratory parameters.