Abstracts

Rationale: Refractory status epilepticus (RSE) is associated with significant mortality. Neuronal loss can occur within 20 minutes of sustained seizure activity, to which the hippocampus is particularly vulnerable. In humans, neuronal damage is most often inferred via non-invasive methods such as MRI. However, few studies have investigated the neuropathological changes that accompany these MRI changes. We sought to characterize hippocampal changes at autopsy in cases of RSE. Methods: We performed a retrospective review of fatal cases of RSE who underwent autopsy. The cause of RSE, EEG findings, medications used to treat RSE, MRI findings and principal cause of death were collected. Sections of the hippocampi were stained with H&E, GFAP, NeuN and HLA-DR and studied microscopically. Results: Five cases of RSE undergoing autopsy were identified. Three cases were new onset RSE (NORSE), and 2 were related to hepatic encephalopathy. The average patient age was 41.2 years, and 3 were female. Duration of RSE ranged from 1-59 days (average 16.2). Prominent astrogliosis was found in CA4 and granule layer in 3/5 cases. Two other cases showed prominent astrogliosis in CA1: one with early hippocampal sclerosis, the other acute neuronal necrosis. Clasmatodendrosis was evident in all cases. There was increased staining with HLA-DR in areas most prominently affected with astrogliosis. Conclusions: This study demonstrates prominent involvement of CA4 and granule layer of the hippocampus. This differs from previous findings in which seizure-related damage occurred primarily in CA1. Clasmatodendrosis, an irreversible astroglial change thought to represent autophagy, protein degradation and cellular death, was found in all our cases and occurred primarily in the CA1 region andadjacent white matter. In rat models, clasmatodendrosis was found in the CA1 region of the hippocampus, and represents chronic necrotic as the result of status epilepticus, indicating selective astrocyte vulnerability to the effects of status epilepticus. This finding has not previously been reported in humans. This suggests hippocampal damage in RSE may be due to factors other than ischemia. The increased staining of HLA-DR may represent an inflammatory response, but whether this is primary and potentially etiological, or secondary is unknown. Status epilepticus was poorly controlled in all but one case, and interestingly, this case had the most prominent pathological findings. However, the duration of RSE in this patient was significantly longer than any other patient (59 days). This suggests that the duration of status epilepticus may have more influence on neuropathological changes than does seizure control; however, this observation is limited to a single patient.