Abstracts

The incidence and predictors of nonconvulsive seizures (NCSzs) on continuous EEG monitoring (cEEG) in patients with intracerebral hemorrhage (ICH) are largely unknown. The objective of this study was to determine the frequency and risk factors for NCSzs in patients with acute spontaneous, nontraumatic, nonaneurysmal ICH on cEEG. The Columbia Comprehensive Epilepsy Center cEEG database was reviewed to identify patients [gt]18 y.o. with ICH who underwent cEEG over a 4 year period. Video-EEG recordings and medical charts were reviewed to identify demographic, clinical and EEG findings. Multivariate logistic regression was carried out to identify factors associated with NCSzs in this patient population. Patients with purely cerebellar hemorrhages and those with a past history of epilepsy were excluded. 48 patients (mean age 61 [plusmn] 18, 60% male) were monitored for a median of 2 days (range 1-18 days). Hypertension was the main cause of hemorrhage, noted in 42% of patients, followed by amyloid angiopathy (15%) and vascular malformation (10%). Location of the ICH was lobar (superficial) in 65% patients and deep in 35%. ICH volumes were [lt]30 cc in 44% and [gt]30 cc in 54%. 45% of patients developed shift, 25% had hydrocephalus, 30% required an external ventricular drain, and another 15% required a ventriculoperitoneal shunt. 11 patients (23%) had seizures on cEEG; 10/11 (91%) had purely nonconvulsive seizures. In multivariate analysis, NCSzs were independently associated with in-hospital clinical seizures prior to cEEG hookup (p=0.03, OR = 7.1; 40% of those with clinical seizures while in hospital had NCSzs on cEEG vs 12% of those without) and the presence of PLEDs (p=0.008, OR 14.6; 63% vs 13%). Additional variables found to be associated with NCSzs in the univariate analysis included the presence of BiPLEDs, SIRPIDs, associated ischemic infarct, and lobar location. Of the 11 patients with seizures, seizures occurred upon hookup in 3 (27%), within one hour in 7 (64%), within 24 hours in 8 (73%) and within 48 hours in 9 (82%). Seizures are common during cEEG in patients with acute ICH. The majority of patients with seizures have purely NCSzs that can only be detected with cEEG. Over half are detected in the first hour of recording, while 18% are not detected until [gt] 48 hours of cEEG recording. NCSzs are most strongly associated with prior in-hospital clinical seizures and PLEDs, but may also be associated with BiPLEDs, SIRPIDs, associated infarct and lobar location. This study confirms the importance of prolonged cEEG for patients with acute ICH.