Mapping the Cerebral Burden of Status Epilepticus - Results from a Longitudinal MRI Study

Status epilepticus (SE) is a neurological emergency with a mortality of up to 39% in population-based studies. Animal studies suggest that, beyond a critical timepoint (t2), SE induces neuronal injury exceeding what can be expected from the underlying aetiology, but in vivo evidence in humans is scarce. Thirty-six prospectively recruited individuals with SE with a mean age of 59 years underwent serial high-resolution T1-weighted brain MRI over a mean follow-up period of 5 months and were compared with 34 individuals with drug-resistant focal epilepsy and 36 propensity-score matched healthy controls, which were retrospectively included. Cortical thickness and volume of subcortical structures were estimated and harmonised for scanner-related effects. Longitudinal change was assessed using linear mixed-effects models correcting for age at baseline and interscan interval and compared between groups. We further investigated the independent effects of SE duration, semiology and level of consciousness after mutual adjustment for each variable as well as aetiology and age at baseline, and investigated longitudinal structural change associated with key locations of peri-ictal MRI abnormalities (PMA). SE was associated with pronounced bilateral hippocampal atrophy in comparison to normal aging and drug-resistant epilepsy, alongside volume increase in several deep grey matter nuclei as well as a trend for cortical thinning of medial brain structures. SE duration was the strongest independent driver of brain change, producing widespread cortical thinning and bilateral hippocampal atrophy. A convulsive semiology was independently associated with accelerated medial temporal cortical thinning and bilateral hippocampal volume loss when compared with non-convulsive (NCSE) and other prominent-motor SE, while reduced consciousness predicted faster thinning of medial frontoparietal cortex. PMA was associated with distinct longitudinal trajectories of subcortical volume, with pulvinar and hippocampus involvement predicting thalamic and hippocampal atrophy patterns. According to our findings, a single episode of SE was associated with a measurable structural imprint in the brain that evolved over months beyond what would be expected for aetiology alone, with atrophy trajectories highlighting a vulnerability of the hippocampus and other limbic structures. A long SE duration, together with, to a lesser extent, convulsive semiology and impaired consciousness, independently amplified this damage. These findings corroborate the timepoint-based (t2) concept of SE and reinforce the clinical imperative of rapid seizure termination to contain long-term structural brain injury.