Bone flap explantation, steroid use, and rates of infection in patients with epilepsy undergoing craniotomy for implantation of subdural electrodes

Abstract

Subdural implantation of electrodes is commonly performed to localize an epileptic focus. Whether to temporarily explant the bone plate and whether to treat patients with perioperative steroid agents is unclear. The authors' aim was to evaluate the utility and risk of bone plate explantation and perioperative steroid use. The authors reviewed the records of all patients who underwent unilateral craniotomy for electrode implantation performed between November 2001 and June 2011 at their institution. Patients were divided into 3 groups: Group 1 (n=24), bone explanted, no perioperative steroid use; Group 2 (n=42), bone left in place, no perioperative steroid use; Group 3 (n=25), bone left in place, steroid agents administered perioperatively. Complications, mass effect, and seizure rates were examined by means of statistical analysis. Of 324 cranial epilepsy surgeries, 91 were unilateral subdural electrode implants that met our inclusion criteria. A total of 11 infections were reported, and there was a significantly higher rate of infection when the bone was explanted (8 cases [33.3%]) than when the bone was left in place (3 cases [4.5%], p<0.01). Leaving the bone in place also increased the rate of asymptomatic subdural hematomas and frequency of seizures, although there was no increase in midline shift, severity of headache, or rate of emergency reoperation. The use of steroid agents did not appear to have an effect on any of the outcome measures. Temporary bone flap explantation during craniotomy for implantation of subdural electrodes can result in high rates of infection, possibly due to the frequent change of hands in transferring the bone to the bone bank. Leaving the bone in place may increase the frequency of seizures and appearance of asymptomatic subdural hematomas but does not increase the rate of complications. These results may be institution dependent.