The GABA-transaminase inhibitor, vigabatrin, has been shown to have a rather low degree of acute toxicity in several animal species. Oral administration of the drug at 1,000 mg/kg/day for 2-4 weeks caused decreased food consumption and weight loss with resultant prostration and death in both rats and dogs. Dosages of 200 mg/kg/day were tolerated for a year without clinical signs in dogs, although rats suffered reduced weight gains and convulsions after 3-4 months when given the drug in the diet. The convulsions continued to occur frequently throughout the one-yr study, but abated 3-4 months after cessation of treatment. The only consistent histopathologic evidence of toxicity in rats and dogs has been the finding of intramyelinic edema (microvacuolation) in the brain, most notably in certain areas of white matter (cerebellum, reticular formation and optic tract in rats and columns of fornix and optic tract in dogs). No lesions were found in the spinal cord or peripheral nervous system. It took several weeks for the microvacuolation to develop, even at high dosages, but it did not continue to progress thereafter, even though a slight effect was noted at dosages as low as 30-50 mg/kg/day after one yr of treatment. The intramyelinic edemia disappeared within a few weeks after treatment was withdrawn. No residual effects were observed in dogs, whereas rats exhibited swollen axons and microscopic mineralized bodies in the cerebellum. Monkeys exhibited no adverse clinical effects except for occasional loose stools at 300 mg/kg/day. After 16 months of oral treatment at 300 mg/kg/day any suggestion of intramyelinic edema was considered to be equivocal, and there was no evidence of any effect in the 50 or 100 mg/kg/day monkeys after 6 yr of treatment. Higher doses caused chronic diarrhea, thus limiting the dosage in this species. Vigabatrin was shown to be well absorbed in rat, dog and man, whereas dose-limited absorption occurred in the monkey. Metabolism is practically nil in all 4 species and the primary elimination pathway is by glomerular filtration. Because vigabatrin is an irreversible inhibitor of GABA-transaminase and the enzyme has a slow turnover rate, plasma levels of the drug are not indicative of its pharmacologic activity. For this reason cerebrospinal fluid levels of GABA and vigabatrin were evaluated, with considerable species differences being noted. The significance of these differences in relation to the differences in toxic response is discussed.