Summary:Purpose: To evaluate effects of vigabatrin (VGB) by using [1231)iomazenil single-photon emission computed tomography (SPECT) to estimate central y-aminobutyric acid (GABA,)/benzodiazepine receptors (BZRs), and magnetic resonance spectroscopy (MRS) to assess tissue GABA levels.Methods: Six patients with partial seizures had both SPECT and MRS before and 25-84 days after starting VGB (3 g P.o., q.d.). SPECT was acquired by using the constant-infusion method and, after nonuniform attenuation correction, coregistered with T,-weighted MR Imaging (MRI) A volume of interest (VOI) of 3 x 2 x 2 cc over the occipital cortex, used for MRS acquisition, was positioned on both MRI and coregistered SPECT. Occipital activity was divided by either total plasma activity or plasma [1231]iomazenil concentration to estimate BZR distribution volume (V,-p and Vk, respectively). Wilcoxon's test was used for VOI differences in GABA levels, BZR V,-p or V;. SPM96 (either no global normalization or proportional scaling) was used to compare BZR V,-p changes in the patients with and without VGB with test-retest data in eight healthy age-matched controls.Results: Occipital GABA levels were increased threefold (without VGB, 1.1 -t 0.1 pmol/g; with VGB, 2.9 & 0.5 pmol/g; p = 0.027). BZR distribution volumes showed no change, when estimated by either V,-p (without VGB, 6.00 2 0.91 mllg; with VGB, 5.86 k 0.44 ml/g; p = 0.92) or V; (without VGB, 41.1 2 11.2 ml/g; with VGB, 41.2 _t 9.9 ml/g; p = 0.75). No significant changes were detected by SPM96.Conclusions: A clinically effective dose of VGB caused a threefold increase in tissue GABA levels but was not associated with a substantial BZR downregulation. Key Words: Vigabatrin-Benzodiazepine receptor-SPECT-GAB A-MRS .y-Aminobutyric acid (GABA) is the major inhibitory neurotransmitter in human cerebral cortex. Animal models show that GABA plays a pivotal role in suppressing the origin and spread of seizure activity. The antiepileptic drug (AED) vigabatrin (VGB; y-vinyl GABA, or 4-amino-hex-5-enoic acid) irreversibly inhibits GABA transaminase, preferentially in neurons but also in astrocytes, markedly increasing GABA concentrations in vitro and in vivo (1). VGB may in addition compete with GABA for the neuronal GABA transporter (2). A previous study using 'H nuclear magnetic resonance spectroscopy (MRS) showed a twofold elevation of occipital cor- (3). The vast majority of tissue GABA measured by MRS is intracellular. Nevertheless, the increased tissue GABA levels induced by VGB probably also reflect the extracellular pool of GABA, which interacts with and may modulate GABA*ibenzodiazepine receptors (BZR) levels. That is, lumbar cerebrospinal fluid (CSF) GABA concentrations also are increased threefold with daily VGB therapy, and the increased CSF GABA concentrations correlate with improved seizure control (4)(5)(6)(7)(8). Therefore the increase in GABA levels appears to be critical to the efficacy of VGB .Alterations in BZR may modulate the effect of increased GABA levels. For example, incr...