Neuroimaging seems warranted in children with new-onset seizures presenting as status epilepticus; MRI preferable to CT when available, data suggest

In children with new-onset seizures presenting as status epilepticus (SE), neuroimaging offers potentially useful information related to diagnosis, treatment, and long-term prognosis, according to an analysis of patients treated at the Children's National Medical Center in Washington, DC. In addition, findings from the study suggest that magnetic resonance imaging (MRI) is preferable to computed tomography (CT) because of its superior yield. The purpose of the analysis was to better...

In children with new-onset seizures presenting as status epilepticus (SE), neuroimaging offers potentially useful information related to diagnosis, treatment, and long-term prognosis, according to an analysis of patients treated at the Children's National Medical Center in Washington, DC. In addition, findings from the study suggest that magnetic resonance imaging (MRI) is preferable to computed tomography (CT) because of its superior yield.

The purpose of the analysis was to better characterize children with new-onset seizures presenting as SE. To do this, researchers evaluated prospectively collected data for 1,382 children (median age, 2 years) who presented with new-onset seizures between January 2001 and December 2007. Of these children, 144 had SE (defined as tonic, clonic, or tonic-clonic unremitting seizures lasting >20 minutes), with SE lasting >1 hour in 28%. Among the patients with SE, a majority (62%) had no history of significant medical problems; 25% had a family history of seizures or epilepsy.

Of the 139 patients (97% of the cohort) who were evaluated with electroencephalography (EEG), 115 (83%) received a routine 30-minute EEG, 6 (4%) received multiple routine EEGs, and 18 (13%) underwent prolonged digital video EEG monitoring. Electrographic seizures were observed in 5 patients (4%) and were captured only with prolonged monitoring, the authors noted. Nonconvulsive SE was identified in 4 patients, all of whom underwent prolonged monitoring.

The most common seizure etiology was febrile convulsions, followed by cryptogenic etiology. The most common acute symptomatic cause was a central nervous system infection, whereas the most common remote symptomatic cause was cerebral dysgenesis.

Almost all of the patients (99%) underwent a head CT scan, and 31% (n=45) underwent an MRI. Of the CT scans, 115 (80%) were normal, 14 (10%) showed acute abnormalities, and 14 showed relevant chronic abnormalities.

Among the 44 patients who underwent both CT and MRI, MRI detected abnormalities in 14 of 30 patients (47%) with normal CT scans. Two patients had focal signal abnormalities on CT scans that were later deemed normal after reviewing MRI data. Although CT identified acute vascular lesions and acute edema, the investigators noted, MRI was superior in identifying subtle abnormalities and remote symptomatic causes such as dysplasia and mesial temporal sclerosis.

Combined, the use of CT and MRI led to a diagnosis in 30% of the patients and directed acute management of SE in 24%.

The study authors suggested that these data support the use of neuroimaging--particularly MRI--in patients with new-onset seizures presenting as SE that is not a prolonged febrile convulsion. They added that the findings also indicate a benefit of prolonged video EEG monitoring, particularly in children who present with SE that does not return to baseline (ie, potential cases of nonconvulsive SE). (Singh RK, et al. Neurology 2010;74:636-642.)

In an accompanying editorial, Dr. Shlomo Shinnar with the Albert Einstein College of Medicine and Dale Hesdorffer with Columbia University suggested that the current study offers "convincing evidence that MRI can provide valuable diagnostic information regarding etiology in a significant minority of patients. Although MRI may not alter acute care in most cases, it will provide important information to inform and potentially alter decisions about subsequent treatment and prognosis."

They added that these data "substantially strengthen" recommendations for imaging and are likely to affect clinical practice, particularly for the treatment of SE with an unclear etiology. (Neurology 2010;74:624-625.)