Seizure-induced reversible magnetic resonance imaging abnormalities: A retrospective cohort study

Abstract

Purpose: Seizure-induced reversible magnetic resonance imaging (MRI) abnormalities (SRMA) present challenges in seizure management. We sought to investigate the frequency, risk factors, evolution and prognostic value of SRMA.

Methods: A retrospective observational cohort study of consecutive seizure patients investigated with an MRI of the brain was conducted. Clinical and MRI data were reviewed to determine the clinical characteristics and imaging findings of SRMA. Outcomes (seizure freedom versus uncontrolled seizures and deaths) were assessed upon the last clinic follow-up. Mann-Whitney U test and chi-square test for independence with Bonferroni correction were used to explore the statistical significance of predictive factors.

Results: The study included 483 consecutive seizure patients with 7.6% developing SRMA. Patients with SRMA were older (median age 57 years, interquartile range-IQR 52-66, p < 0.001) and experienced longer seizures (median 5 minutes, IQR 2-15, p = 0.002) compared with seizure patients with normal MRI. Seizure type (provoked versus unprovoked), recurrence (first versus recurrent) and epileptiform EEG changes did not demonstrate a significant association. Diffusion restriction and ADC reduction observed in SRMA resolved earlier, while T2, FLAIR hyperintensities and temporal lobes changes persisted longer on follow-up scans. The median time interval from seizure to complete resolution of SRMA was 87 days (IQR 45-225). No statistical difference in outcomes was seen between patients with SRMA and normal MRIs (p = 0.19).

Conclusions: SRMA is an uncommon finding following seizures. It is not associated with poor seizure control or mortality. Risk factors associated with SRMA include older age and longer seizure duration including status epilepticus.

Keywords: Imaging; Laminar necrosis; Prognosis; Seizure; Status epilepticus.