Can Continuous EEG Influence Outcome in Patients with Intracerebral Hemorrhage (ICH)?

Abstract & Commentary

By Padmaja Kandula, MD, Assistant Professor of Neurology and Neuroscience, Comprehensive Epilepsy Center, Weill Medical College of Cornell University. Dr. Kandula reports no financial relationships relevant to this field of study.

Synopsis: This retrospective study demonstrated that 31% of patients with ICH had an electrographic or clinical seizure while in the hospital.

Source: Claassen J, Jetté N, Chum F, et al. Electrographic seizures and periodic discharges after intracerebral hemorrhage. Neurology 2007; 69:1356-1365.

Since the advent of continuous eeg monitoring (cEEG), medical awareness of electrographic seizures has increased. In a previous retrospective study by Claassen and colleagues, nearly 20% of critically ill patients monitored for unexplained mental status had subclinical seizures.1 A separate observational study by Vespa and coworkers suggested that seizures after intraparenchymal hemorrhage, particularly lobar hemorrhage, are not uncommon and are associated with poor outcome.2 Thus, clinical evidence that electrographic seizures do occur with regular frequency in this subpopulation has emerged. This recent retrospective study by Claassen et al aims to determine the frequency and electrographic and radiologic variables associated with subclinical seizures and periodic discharges in patients with intracerebral hemorrhage (ICH).

Over a six-year period, 102 patients with nontraumatic spontaneous ICH who underwent cEEG were retrospectively identified. Patients with ICH with unexplained mental status or suspicion of seizures underwent cEEG at the clinical discretion of the treating physician. Inclusion criteria included age older than 17 and cEEG duration of at least 12 hours. Patients with ICH secondary to trauma or aneurysmal bleed were excluded. Patients with a decreased level of consciousness and lobar hemorrhage or signs of increased intracranial pressure were loaded with 20 mg/kg of fosphenytoin, followed by maintenance phenytoin. Clinical, radiographic, and electrographic data were retrospectively obtained on each of the 102 study patients. Clinical data recorded by a study neurologist included: vitals signs/laboratory testing (admission blood pressure, serum glucose, toxicology screen), admission neurologic status, and hemorrhage etiology (hypertension, vascular malformation, amyloid angiopathy, anticoagulation, unknown, other).

Four radiologic features were assessed: change in ICH volume, midline shift between baseline admission CT scan and follow up CT (at 24 hours, 48-72 hours, and between days 3-7), ICH location (deep versus lobar), and closest distance in mm of the ICH from the cortical surface.

The presence of the following were determined: convulsive and electrographic seizures, PEDs (periodic epileptiform discharges), GPDs (generalized periodic discharges bilaterally synchronous with no consistent laterality), BIPLEDs (bilateral independent PLEDs), triphasic waves, FIRDA (frontal intermittent rhythmic delta activity), burst-suppression activity, reactivity to external stimuli, stage II sleep transients (K-complexes, sleep spindles), and SIRPIDs (stimulus induced rhythmic, periodic, or ictal discharges).

A total of 31% of patients with ICH experienced a seizure (either clinical or electrographic) between hemorrhage onset and hospital discharge. Of the patients, 18% had electrographic seizures, of which 56% were detected within one hour of cEEG initiation. Nonconvulsive status epilepticus (NCSE) occurred in 39% of patients with electrographic seizures. Electrographic seizures were more frequent in those with PEDs (59%) versus those without PEDs (9%), in those with PLEDs than those without (77% versus 9%, respectively), and in those with focal SIRPIDs than those without (77% versus 14%, respectively).

An increase in ICH volume of 30% or more between admission and 24-hour follow-up CT scan was associated with electrographic seizures. PEDs were significantly less frequently seen in patients with ICH located 1 mm or deeper from the cortical surface (8% of hemorrhages 1 mm or deeper versus 29% of hemorrhages within 1 mm of cortex).

The only electrographic variable associated with poor outcome (a rating of 1-2 on the Glasgow outcome scale) was the presence of any periodic epileptiform discharge. In addition, ICH volumes > 60 mL and lower systolic blood pressure on admission also were associated with poor outcome.


The inherent retrospective nature of this paper raises the obvious concern of possible selection bias. Nonetheless, this study, along with a prior prospective study by Vespa et al,2 do support the idea that electrographic seizures are not an uncommon entity and that they can be missed by visual inspection alone. In both studies, seizures occurred in 28-31% of patients and more than half of all seizures were purely electrographic. Among the patients who had seizures, nearly all seizures were detected within 48 hours. In addition, only 13% of all ICH admissions were monitored via cEEG; thus, the above numbers may not represent the true incidence of electrographic seizures in this population of patients. However, despite methodological limitations, this paper does raise two important clinical questions. Can radiologic criteria be used as an adjunct to clinical history when deciding which ICH patients undergo cEEG monitoring?

The finding that expanding ICH volume is associated with electrographic seizures could potentially identify patients at higher risk of electrographic seizures within 24 hours. This identification of "high-risk" patients seems an important clinical point since nonconvulsive status, a known entity responsible for cell hippocampal cell death,3 occurred in nearly 40% of patients with electrographic seizures. However, the clinical implications of discrete electrographic seizures in critically ill patients remains unanswered. To further complicate matters, the authors found that PEDs and not seizures were associated with poor outcome as measured by the Glasgow outcome scale (originally designed to assess outcome after brain injury). Although this scale is based on a fixed numerical rating performed at discharge and may not accurately reflect lasting disability and quality of life, this further begs the question, how aggressively should electrographic seizures, and more importantly PEDs, be treated? The evidence is compelling that cEEG is indeed an important diagnostic tool, but can aggressively suppressing electrographic seizures and PEDs improve patient outcome? A definitive prospective treatment trial of subclinical seizures and PEDs and long-term outcome is needed to further shed light on the significance of these electroencephalographic entities.


1. Claassen J, et al. Neurology 2004; 62:1743-1748.

2. Vespa PM, et al. Neurology 2003; 60:1441-1446.

3. DeGiorgio CM, et al. Epilepsia 1992; 33:23-27.