Abstracts

Rationale: Currently there are available anecdotal case reports showing that chronic exposure to valproate (VPA) may lead to reversible brain pseudoatrophy in patients with epilepsy. Recent morphometric MRI studies suggested that VPA use could be associated with parietal and occipital cortical thinning in epilepsy patients. However, their results might be confounded by the small sample size and clinical heterogeneity. We examined the effect of VPA on the regional grey matter (GM) structural changes using a large number of homogenous patients with idiopathic generalized epilepsy (IGE).

Methods: Volumetric MRI were acquired using a 3 Tesla scanner in 81 IGE patients not currently taking valproate (VPA– group; 54 females), 112 IGE patients currently taking valproate (VPA+ group; 44 females), and 120 healthy subjects (control group; 53 females). Voxel-based morphometry and surface-based morphometry were used to evaluate GM volume and cortical thickness alterations between the groups, respectively.

Results: Three groups did not differ in age and education years, but differed in sex. There were no differences in age of seizure onset, disease duration, and seizure frequency between VPA– and VPA+ groups. In voxel-based morphometry analysis, one-way analysis of variance showed a significant main effect of group in bilateral cerebellum, hippocampus, insula/basal ganglia, motor/premotor cortex, anterior cingulate cortex, and thalamus. Compared with control group, VPA+ group revealed GM volume reductions in the above-mentioned regions. Compared with control group, VPA– group revealed GM volume reductions in bilateral hippocampus, temporal lobe, and thalamus. Compared with VPA– group, VPA+ group exhibited GM volume reductions in the anterior cingulate cortex, bilateral cerebellum and motor/premotor cortex. In surface-based morphometry analysis, one-way analysis of variance showed a significant main effect of group in the mesiofrontal cortex and bilateral motor/premotor and prefrontal cortices. Compared with control group, VPA+ group revealed cortical thinning in the above-mentioned regions. No difference in cortical thickness was found between control and VPA– groups. Compared with VPA– group, VPA+ group exhibited cortical thinning in the mesiofrontal cortex and bilateral motor/premotor and prefrontal cortices.

Conclusions: We have identified that chronic exposure to VPA may lead to cortical GM volume and thickness reductions in multiple brain regions including mesiofrontal and motor/premotor cortices. The differences were independent of demographics and clinical characteristics as well as total brain volume, suggesting that VPA treatment per se may be associated with structural brain changes, particularly in the frontal cortex. The clinical significance of frontal structural changes observed in our study remains unknown but may provide a mechanistic understanding of reported frontal cognitive dysfunctions that are associated with VPA use. Our results indicate that VPA use should be taken into account as a potential confounding factor in future morphometric MRI studies in which subjects taking VPA are included.

Funding: None