Vigabatrin is an antiepileptic drug (AED) that acts as a selective irreversible inhibitor of gamma-aminobutyric acid (GABA) transaminase. In 1997, 3 cases of severe symptomatic and persistent visual field constriction associated with vigabatrin treatment were described. During 1997 to 1998, similar concentric visual field constrictions were described in patients with drug-resistant epilepsy who were receiving vigabatrin concurrently with other AEDs. However, a study of patients treated with vigabatrin monotherapy alone showed that there was a causal relationship between vigabatrin treatment and the specific bilateral concentric visual field constriction. The Marketing Authorisation Holders survey (involving 335 vigabatrin recipients aged >14 years) indicated that 31% of patients [95% confidence interval (CI) 26 to 36%] had a visual field defect attributable to vigabatrin, compared with a 0% incidence of visual field defects (upper 95% CI 3%) in an unexposed control group. Other studies in adults have given similar overall prevalences, with a total of 169 of 528 patients diagnosed with vigabatrin-associated field defects (32%, 95% CI 28 to 36%). Male gender seems to be associated with an increase in the relative risk of visual field loss of approximately 2-fold. The pattern of defect is typically a bilateral, absolute concentric constriction of the visual field, the severity of which varies from mild to severe. Data gathered so far suggest that the cumulative incidence increases rapidly during the first 2 years of treatment and within the first 2 kg of vigabatrin intake, stabilising at 3 years and after a total vigabatrin dose of 3 kg. The prevalence of vigabatrin-associated field defects seems to be lower in children, but there are also methodological problems and greater variability in the assessment of visual fields in children. There is particular concern that the increased risk of the visual field defects will outweigh the benefit of the drug in patients who could be controlled with other AEDs. Vigabatrin should currently be used only in combination with other AEDs for patients with resistant partial epilepsy when all other appropriate drug combinations have proved inadequate or have not been tolerated. Regular visual field testing should be performed before the start of treatment and at regular intervals during treatment. Patients with pre-existent visual field defects due to other causes should not be treated with vigabatrin. Currently, the benefits of treating infantile spasms with vigabatrin monotherapy seem to outweigh the risks, but further prospective studies and follow-up of children receiving treatment are needed to evaluate the place of vigabatrin in this indication.