Although VNS can lessen the number of seizures in some patients, there is no evidence to indicate that this would reflect a reduction in associated morbidity and mortality or significant quality of life improvement. The evidence to date must be cautiously considered, since, in general, it comes from studies of poor methodological quality and with no control group. Special care must be exercised in less studied subpopulations such as patients with generalized epilepsy. At present, no criteria have been identified to select those patients with higher chances of benefiting from VNS. As general guideline, it is recommended to confirm epilepsy diagnosis first as well as adequate compliance to pharmacological treatment, duly documenting failure. When treating a patient with refractory epilepsy, different therapeutic alternatives can be considered, including surgical treatment (which can be curative up to 60-70% of patients with temporal epilepsy) or pharmacological treatment with new drugs. In those patients in whom surgery is not a viable option, the decision to use VNS should be based on frequency and type of seizures, severity of the episodes, toxicity to antiepileptic drugs and the expected overall impact of these factors on the patient s quality of life.