Risk of vigabatrin-associated brain abnormalities on MRI in the treatment of infantile spasms is dose-dependent



Although the link between vigabatrin (VGB) and retinotoxicity is well known, little attention has been focused on the risk of VGB-associated brain abnormalities on magnetic resonance imaging (MRI) (VABAM), namely reversible—and largely asymptomatic—signal changes in the thalami, basal ganglia, brainstem tegmentum, and cerebellar nuclei. Using a large infantile spasms cohort, we set out to identify predictors of these phenomena.


Children with infantile spasms were retrospectively identified. Brain MRI reports were serially reviewed without knowledge of VGB exposure. Upon VABAM discovery, records were systematically reviewed to ascertain presence of symptoms attributable to VGB. Separately, progress notes were sequentially reviewed to identify and quantify VGB exposure.


We identified 507 brain MRI studies among 257 patients with infantile spasms. VGB treatment was documented in 143 children, with detailed exposure data available for 104, of whom 45 had at least one MRI study during VGB treatment. Among the limited subset of asymptomatic children who underwent MRI (n = 40), 6 exhibited VABAM. Risk of asymptomatic VABAM was dose-dependent, as peak (but not cumulative) VGB dosage was strongly associated with asymptomatic VABAM (p = 0.0028). In an exploratory analysis, we encountered 4 children with symptomatic VABAM among 104 patients with detailed VGB exposure data. Risk of symptomatic VABAM was seemingly dose-independent, and potentially associated with concomitant hormonal therapy (i.e., prednisolone and adrenocorticotropic hormone [ACTH]) (p = 0.039).


We have demonstrated dose-dependent risk of asymptomatic VABAM and uncovered a possible association between symptomatic VABAM and concomitant hormonal therapy. Caution should be exercised in the use of high VGB dosage (i.e., >175 mg/kg/day), and further study is warranted to confirm the potential impact of hormonal therapy.