Abstracts

Rationale: Vigabatrin (VGB) is commonly used as first- or second-line therapy for the treatment of infantile spasms. Although numerous studies have linked VGB to retinotoxicity and irreversible peripheral vision loss, minimal attention has been focused on potential extraretinal toxicities. In particular, there is inadequate characterization of Vigabatrin-Associated Brain Abnormalities on MRI (VABAM), namely reversible hyperintense T2 signal changes in the thalami, basal ganglia, brainstem tegmentum, and cerebellar dentate nuclei. There are rare reports of VABAM associated with reversible hyperkinetic movement disorders, severe encephalopathy, and respiratory compromise. Employing a large infantile spasms cohort, we set out to estimate the risk of VABAM, with and without clinical symptoms, among VGB-treated children with infantile spasms. Methods: Children with video-EEG confirmed infantile spasms evaluated at our center between February 2007 and February 2014 were retrospectively identified. For each patient, brain MRI reports were serially reviewed to ascertain the presence of VABAM in the brainstem, basal ganglia, cerebellum, and/or thalami. Separately, each neurology progress note was sequentially reviewed to identify and quantify VGB exposure. When VABAM was identified, progress notes (from any provider) were systematically reviewed to ascertain the presence of symptoms reasonably attributable to VGB. Results: We identified 257 patients with video-EEG confirmed infantile spasms, who underwent a total of 507 brain MRI studies. 143 children were treated with VGB, with detailed dates of exposure available for 104 patients, of whom 45 underwent at least one MRI study during vigabatrin treatment. 5 children underwent MRI for evaluation symptoms suspicious for VGB toxicity (i.e. hyperkinetic movements, severe encephalopathy, and/or respiratory compromise), and VABAM was seen in 4 cases. Among the remaining 40 children who underwent MRI for other reasons (i.e. etiologic or surgical evaluation), we encountered 6 cases in which asymptomatic VABAM was discovered. Based on the cohort of 40 patients among whom we could reasonably encounter asymptomatic MRI changes, we estimate the risk of asymptomatic VABAM is 6/40, or 15.0% (95%CI 5.7 ?" 29.8). Based on the larger cohort of 104 patients with detailed exposure data, we estimate the cumulative risk of symptomatic VABAM is 4/104, or 3.8% (95%CI 1.1 ?" 9.6). See Figures 1 and 2. Conclusions: This is the first study to quantify the risk of symptomatic VABAM. We have identified a small but significant risk of symptomatic VABAM among children with infantile spasms, which should be emphasized in risk-benefit discussions. Further study is needed to determine whether VABAM is dose-dependent or idiosyncratic, whether it is truly reversible, and to identify other predictors of risk. Funding: The work was accomplished with the support of the Elsie and Isaac Fogelman Endowment, the Hughes family Foundation, and the UCLA Children's Discovery and Innovation Institute.