Seizure incidence in the acute postneurosurgical period diagnosed using continuous electroencephalography

Abstract

OBJECTIVE Delay in diagnosis and subsequent treatment of nonconvulsive seizures can lead to worsened outcomes. The gold standard in detecting nonconvulsive seizures is continuous video-electroencephalography (cEEG). Compared to routine, 30-minute EEG, the use of cEEG increases the likelihood of capturing intermittent nonconvulsive seizures. Studies of critically ill patients in intensive care units demonstrate a particularly high rate of nonconvulsive seizures. Some of these studies included postneurosurgical patients, but often subanalyses of specific populations were not done. In particular, few studies have specifically evaluated postneurosurgical patients by using cEEG in the acute postoperative setting. Therefore, the incidence and predictors of acute postneurosurgical seizures are unclear. METHODS In this study, the authors focused on patients who were admitted to the neurological critical care unit following neurosurgery and who underwent cEEG monitoring within 72 hours of surgery. RESULTS A total of 105 cEEG studies were performed in 102 patients. Twenty-nine patients demonstrated electrographic (subclinical) seizures, of whom 10 had clinical seizures clearly documented either before or during cEEG monitoring. Twenty-two patients had subclinical seizures only detected on cEEG, 19 of whom did not have clinical seizure activity at any point during hospitalization. Those with seizures were more likely to have had a history of epilepsy (p = 0.006). The EEG studies of patients with seizures were more likely to show lateralized periodic discharges (p = 0.012) and lateralized rhythmic delta activity (p = 0.012). The underlying neuropathological disorders most associated with seizure risk were lobar tumor on presentation (p = 0.048), subdural hematoma (SDH) requiring craniotomy for evacuation (p = 0.002), subarachnoid hemorrhage (SAH) (p = 0.026), and perioperative SAH (p = 0.019). In those undergoing craniotomy, the presence of SDH (p = 0.032), particularly if requiring evacuation (p = 0.003), increased the risk of seizures. In those without preoperative intracranial bleeding, perioperative SAH after craniotomy was associated with a higher incidence of seizures (p = 0.014). There was an additive effect on seizure incidence when perioperative SAH as well as concomitant intraparenchymal hemorrhage and/or stroke were present. The clinical examination of the patient, including the presence or absence of altered mental status and the presence or absence of repetitive movements, was not predictive of subclinical seizures. CONCLUSIONS In postneurosurgical patients referred for cEEG monitoring, there is a high rate of both clinical and subclinical seizures in the early postoperative period. Seizures are particularly common in patients with SDH or lobar tumor and perioperative SAH. There was an additive effect on seizure incidence when more extensive brain injury was present. As expected, those with a history of epilepsy also demonstrated higher seizure rates. Further studies are needed to evaluate the time period of maximum seizure incidence after surgery, and the effects acute postneurosurgical seizures have on long-term outcomes.