During the last decade, intermittent electrical stimulation of the left cervical vagus nerve was established as a new add-on treatment of drug-resistant seizures. Particularly in Europe, the acceptance of vagus nerve stimulation (VNS) was tentative in the beginning because of unknown mechanisms of action. We report the outcome in a sample of 95 adult patients with drug-resistant seizures who have received implants since 1998. The last available follow-up data are included. Unavoidable medication changes (e.g., intoxication) were accepted to examine VNS under usual clinical conditions. Median percentage of reduction in seizure frequency as compared to baseline was 30%. The seizure responder rate (> or =50% reduction) was 45%. Four patients experienced total release from seizures. Adverse effects were mild in general. Seizure outcome was positively correlated with VNS duration. No potential clinical factor (e.g., syndrome, cause, or lesion) could be identified as an indicator of favorable outcome. Patients with on stimulation-on periods of 30 seconds (standard cycle) had a better outcome than patients with stimulation-on periods of 7 seconds (rapid cycle). During an embedded, randomized, controlled trial, no evidence was found for a differential outcome of initial standard cycle versus initial rapid cycle stimulation conditions. Taking into account the good cost-benefit ratio as well as positive effects on well-being, VNS has to be considered an appropriate strategy for the add-on treatment of drug-resistant seizures, particularly in cases not suitable for epilepsy surgery.