Vagus nerve stimulation (VNS) is an established option of adjunctive treatment for patients with drug-resistant epilepsy, however, evidence for long-term efficacy is still limited. Studies on clinical outcomes of VNS in Asia are also limited. We report the overall outcome of a national, prospective registry that included all patients implanted in Japan. The registry included patients of all ages with all seizure types who underwent VNS implantation for drug-resistant epilepsy in the first three years after approval of VNS in 2010. The registry excluded patients who were expected to benefit from resective surgery. Efficacy analysis was assessed based on the change in frequency of all seizure types and the rate of responders. Changes in cognitive, behavioural and social status, quality of life (QOL), antiepileptic drug (AED) use, and overall AED burden were analysed as other efficacy indices. A total of 385 patients were initially registered. Efficacy analyses included data from 362 patients. Age range at the time of VNS implantation was 12 months to 72 years; 21.5% of patients were under 12 years of age and 49.7% had prior epilepsy surgery. Follow-up rate was >90%, even at 36 months. Seizure control improved over time with median seizure reduction of 25.0%, 40.9%, 53.3%, 60.0%, and 66.2%, and responder rates of 38.9%, 46.8%, 55.8%, 57.7%, and 58.8% at three, six, 12, 24, and 36 months of VNS therapy, respectively. There were no substantial changes in other indices throughout the three years of the study, except for self/family-accessed QOL which improved over time. No new safety issues were identified. Although this was not a controlled comparative study, this prospective national registry of Japanese patients with drug-resistant epilepsy, with >90% follow-up rate, indicates long-term efficacy of VNS therapy which increased over time, over a period of up to three years. The limits of such trials, in terms of AED modifications and during follow-up and difficulties in seizure counting are also discussed.