Retinal neuronal layers in children with idiopathic generalized epilepsy: effects of epilepsy and long-term valproate use.

OBJECTIVE: Idiopathic generalized epilepsies (IGEs) and long-term antiseizure treatment may differentially affect retinal neuronal structures. This study aimed to evaluate optical coherence tomography (OCT)-derived retinal and optic nerve head (ONH) parameters in adolescents with IGEs, in order to explore retinal alterations associated with epilepsy and valproic acid (VPA) exposure. METHODS: This cross-sectional study included three groups: patients with IGEs receiving VPA monotherapy for at least one year (PwE+VPA, n = 35), newly diagnosed patients with IGEs who had not yet initiated antiseizure medication (PwE-VPA, n = 22), and healthy controls (HC, n = 30). Retinal and ONH structures were evaluated using OCT. Segmental and global measurements of the Bruch's membrane opening (BMO), peripapillary retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), inner plexiform layer (IPL), and inner nuclear layer (INL) were obtained. RESULTS: A total of 87 individuals aged 10-18 years (47 males and 40 females) were included in the study. In PwE+VPA, the mean duration of epilepsy/VPA exposure was 3.9 ± 2.3 years, and the mean serum VPA concentration was 69.34 ± 18.03 mg/L. There were no statistically significant differences among the groups in age or sex distribution. Compared with HC, PwE+VPA showed significant thinning of the global RNFL, RNFL temporal-inferior, nasal IPL, and global INL. In the comparison between PwE+VPA and PwE-VPA, BMO temporal-inferior was significantly lower in PwE+VPA, whereas the difference in global RNFL thickness did not reach statistical significance. In PwE-VPA, RNFL temporal-inferior, nasal IPL, and global INL were significantly thinner than in HC. In multivariable analyses, nasal IPL and global INL were independently associated with the presence of epilepsy, while RNFL temporal-inferior showed a borderline association. BMO temporal-inferior was independently associated with VPA exposure status, whereas global RNFL was not. Within PwE+VPA, longer epilepsy/VPA treatment duration was independently associated with lower global RNFL thickness but was not associated with BMO temporal-inferior. CONCLUSION: Our findings suggest that retinal involvement may be detectable from the early stages of epilepsy and may predominantly affect the peripapillary RNFL and inner retinal layers. Within the PwE+VPA group, longer epilepsy/VPA exposure duration was associated with lower global RNFL thickness. Further longitudinal studies are needed to clarify the tmporal relationship and underlying mechanisms.