Is There a Need for Extensive Haemostatic Screening in Neurosurgical Patients Using Valproic Acid?

Various haemostatic abnormalities have been reported in association with use of the anti-epileptic drug valproic acid (VPA). When neurosurgical procedures were performed in patients on VPA, our center performed extensive haemostatic screening preoperatively. However, the clinical importance of abnormal findings is unclear. In the current report we evaluate the need for elaborate haemostatic investigations and make recommendations for clinical practice.

We reviewed medical records of all neurosurgical patients on VPA in our university clinic from February 2006 through May 2011 and recorded outcome of the extensive haemostatic screening, any perioperative measures to prevent extensive blood loss and the occurrence of bleeding complications during or after surgery.

A total of 65 consecutive patients using VPA had a neurosurgical procedure performed and underwent preoperative haemostatic screening. In 34 patients persistent laboratory abnormalities were found and the proportion was similar in the 13 patients who reported a bleeding tendency as in the 52 who did not. However, 5/6 patients who reported a clear time relation between start of VPA use and their bleeding tendency were found to have lab abnormalities. Five of the 65 patients had a mild thrombocytopenia (range 100–150 *10E9/L). 17/63 patients had a prolonged bleeding time of which 5 had normalized on repeat testing. 9/65 had a decreased von Willebrand cofactor activity. Only 2 of these had a prolonged bleeding time. Twelve patients had abnormal platelet function tests but in 3 these normalized on repeat testing. CT and PT were normal in all patients and fibrinogen was marginally lowered in 14/65 (lowest 1.2 g/L). Factor XIII was tested in 57 patients and no clinically relevant lowering was perceived (lowest 0.47 IU/mL). Clinical consequences of the laboratory results were left to the treating haematologist. Additional perioperative measures were taken in 15 patients (stopping VPA in 3, administering DDAVP in 12 and tranexaminic monotherapy in one). Two of 65 patients had a bleeding complication: in one no bleeding time was performed but all other tests were normal. Despite tranexaminic acid perioperatively she developed a large subdural hematoma and seizures. None of the other patients with normal lab values had a bleeding complication. The second patient had had uncomplicated non-neurological surgery during VPA use and was found to have a mildly abnormal ATP/ADP ratio. Because of his uncomplicated history no additional measures were taken and he developed a large epidural hematoma.

In our cohort of 65 neurosurgical patients using VPA, none had clinically relevant abnormalities in PT, aPTT, thrombocyte count, fibrinogen or fXIII level. Therefore, these tests appear unnecessary in the preoperative screening of patients on VPA. However, even in the presence of a normal bleeding time, clinically relevant lowering of von Willebrand factor, and abnormalities in platelet function should actively be sought after. In these cases, we recommend treatment with DDAVP perioperatively. Unlike others (Kreuz et al, Epilepsia 1992), we found DDAVP to be safe and effective in our neurosurgical population. Whether an isolated prolonged bleeding time in absence of demonstrated platelet abnormalities necessitates additional perioperative measures should be the subject of further study. In our cohort the majority of patients with an unexplained prolonged bleeding time received DDAVP perioperatively.

No relevant conflicts of interest to declare.