Researchers in the Department of Paediatric Neurosurgery and Epilepsy Unit, Great Ormond Street Hospital, London, UK analyzed the hospital charts of 95 children operated on between 1995 and 2009 for medication-resistant focal epilepsy and requiring presurgical intracranial monitoring of the EEG. Patients who underwent implantation of subdural grids and/or depth electrodes had received prior noninvasive assessment including video-EEG monitoring. Indications for invasive EEG recording included an MRI-negative focal epilepsy, discordant imaging and electroclinical data, suspected multifocal epilepsy, and seizures arising adjacent to eloquent cortex. Mean age at surgery was 10.8 years (range, 0.7-18.5 years). Mean age at the time of the first seizure was 3.4 years (range, 3 days-12 years). Mean number of seizures per month was 239.5 (range, 1-1000; SD 239). The mean number of grids or strips implanted per patient was 3 (range, 1-6). A depth electrode was placed in 31 patients.
Subdural grid recording was uneventful in 51.1% cases. Adverse events in 49% included subdural hemorrhage in 17% patients, wound infection in 14.9%, CSF leak in 10.6%, and symptomatic brain swelling in 6.4%. Adverse events were grade 1 in 19.1% (18 patients with no lengthening of hospital stay); grade 2 in 13.8% (13 with prolonged hospital stay); grade 3 in 16% (15 patients with a reduction in the Glasgow Coma Score); none was grade 4, an adverse event that would result in death. The frequency of adverse events was 20% in children <2 years of age and 50% in children >2 years of age. Of the 46 patients with adverse events, unplanned surgery had to be performed in 17 cases. No permanent neurologic deficits incurred as a result of any adverse event. Predictors of adverse events included age (brain swelling occurred in older patients >5 years); and length of recording (shorter with a complication such as hemorrhage). Functional cortical mapping with stimulation in 68 (71.6%) cases allowed identification of localized seizure onset zone in 68.9% patients. It was inconclusive in 18.9% (17 patients). The outcome of surgery was significantly related to the localizing accuracy of the invasive recording. 
COMMENT. Two US neurosurgeons, Drs JG Ojemann and HL Weiner, each compliments the authors on this important contribution to the surgical management of medically refractory epilepsy. The cerebral swelling and other temporary complications of invasive monitoring are a concern. Advances in intraoperative electrocorticography in the future may lessen the need for preoperative invasive recordings.