Tiagabine is a relatively new anticonvulsive agent. Data concerning safety and efficacy come from randomised controlled trials whose relation to everyday clinical practice is poorly defined. We analysed retrospectively the data of 56 patients to whom tiagabine was routinely prescribed in a special clinic. Effect and adverse events were registered according to documentation of routine visits in the outpatient clinic. After a median of 89 weeks, 22 patients (39%) still received tiagabine. All of them noted a reduction in seizure frequency, and eight (14%) became seizure-free. Reasons for stopping the medication were: an increase in seizure frequency, lack of efficacy, tiagabine-associated non-convulsive status epilepticus and sudden and short episodes of mental chang. However, tiagabine seems to be an effective anticonvulsant in clinical practice but should remain in the hands of experienced prescribers until further clarification of possible risk factors for proconvulsive effects.