Abstracts

Rationale: Vigabatrin (VGB) is an anti-epileptic drug designed to increase the levels of γ-aminobutyric acid (GABA) through selective, irreversible inhibition of GABA transaminase. VGB has been shown to be effective in decreasing or eliminating seizures in multiple, well-controlled studies, with particular efficacy in Infantile Spasms (IS), an age-specific epilepsy syndrome that occurs during infancy. Experimental VGB neurotoxicity in certain animal studies has raised an additional concern. Neuronal damage characterized by intramyelinic edema (IME) of myelinated tracts of the cerebellum in mice, rats and dogs after long-term administration of high doses of VGB was present and appeared reversible after discontinuation of VGB treatment. Despite extensive toxicity studies, VGB-induced IME has yet to be observed in humans. Even though VGB has proven to be an effective antiepileptic medication, demonstrating an adequate risk-to-benefit ratio, past concerns of VGB-induced IME in non-human animal species has resulted in part to the disapproval of VGB as an antiepileptic in the United States. The incidence of magnetic resonance (MR)/diffusion-weighted imaging (DWI) abnormalities in patients with IS undergoing VGB therapy has not been previously documented. The objective of this study was to report the findings of DWI and apparent diffusion coefficient (ADC) mapping of 73 pediatric patients diagnosed with IS on VGB therapy. Methods: Seventy-three patients with IS receiving VGB (40-200mg/kg/day) for more than three months between 1999 and 2007 were investigated. Patients were grouped into three arms: (1) diagnosed with IS but not on VGB (n=10); (2) diagnosed with IS on VGB (n=58); and (3) diagnosed with IS and taken off VGB (n=5). A pediatric neuroradiologist who was blinded to whether the patients were receiving VGB treatment, retrospectively reviewed the MR imaging including DWI. Special attention was given to the globi pallidi, thalami and brainstem. Pediatric patients diagnosed with meningitis, hypoxic ischemic encephalopathy, non-accidental injury, metabolic disorder or any type of disorder that involves destruction of the brain parenchyma were excluded from the study. Results: In patients with IS but not on VGB (n=10) (group 1), one patient had a non-restricted DWI. In patients diagnosed with IS on VGB (n=59) (group 2), 16 patients had baseline MR that where within normal limits; but with initiation of drug, 21% showed an abnormal MR and/or restricted DWI. Forty percent of the remainder of the patients diagnosed with IS on VGB who did not have a baseline MR (n=42) had an abnormal MR and/or restricted DWI on VGB. Yet, the MR and diffusion abnormalities reversed with withdrawal of VGB in all but one patient. Interestingly, one patient diagnosed with IS who was on VGB therapy for ten years showed no MR or DWI signal changes. Conclusions: VGB caused reversible MR signal intensity changes and also reversal of the diffusion restriction in the deep nuclei. Prospective study evaluating the brain parenchyma on MR before, during and after VGB treatment is required for the continuing human evaluation of the safety of VGB treatment.