Abstract: Background：The indication of vagus nerve stimulation (VNS) is approved for children with drugresistant epilepsy (DRE) older than three years old in China. Nowadays, VNS treatment is also used offlabel to treat infants with DRE in clinical practice, but there is still a lack of relevant research on its efficacy and safety. Objective：To evaluate the longterm efficacy, safety and efficacyrelated factors of VNS treatment for infants with DRE. Design：Case series report. Methods：This study included infants with DRE after VNS treatment in the children epilepsy center of Peking University First Hospital between March 2015 and March 2021. All the patients were regularly programmed and followed up after VNS implantation for at least 1 year with current output over 1mA. Patients who received other newadded antiseizure therapy including epileptic surgery, ketogenic diet, glucocorticoid and adrenocorticotropic hormone, or turned off VNS due to any reason were excluded. The frequency of all seizure types recorded in the epilepsy diary of parents was calculated to evaluate followup responder rate defined as seizure frequency decreasing by ≥ 50% at the last followup compared with 3 months before VNS implantation，and seizure frequency reduction rate so as to investigate the effect of clinical information and programmed parameters on the efficacy and safety. Main outcome measures：Responder rate and seizure frequency reduction rate after VNS treatment. Results：Twentyfive infants with DRE after VNS treatment met the inclusion and exclusion criteria, including 16 males (64%) and 9 females. The median age of seizure onset was 4.9 months. The median duration of epilepsy was 1.7 years. The median seizure frequency was 555 times per month, and the median age at VNS implantation was 2.3 years. The number of historical antiseizure medications (ASMs) used and the number of ASMs used 3 months before VNS implantation ranged 310 and 14, respectively. Twelve cases (48%) had specific DRE etiology, and eighteen (72%) were diagnosed with the epileptic syndrome. Four, five, six and ten cases had 4, 3, 2 and 1 seizure types, respectively. All the twentyfive cases of DRE infants implanted with VNS had developmental delay. The median postoperative followup duration was 2.4 years. The change rate of seizure frequency ranged from +1367% to -100% at 1year followup, with a median change rate of -52.8%. The change rate of seizure frequency ranged from +1712% to -100% at 2year followup, with a median change rate of -59.5%. At the last followup, the responder rate was 64%, the seizurefree rate was 20% (5/25) , and the longest seizurefree duration during followup was 2 years and 5 months. The responder rate for 1 year and 2 years after VNS were both 60% (15/25, 9/25). One of fifteen responders at 1 year after VNS became a nonresponder at 2 years after VNS. Two of ten nonresponders at 1 year after VNS became responders at 2 years after VNS. No perioperative and devicerelated adverse reactions were observed, and 4 patients developed stimulationrelated adverse reactions during programming, including 2 cases of transient cough and 2 cases of pharyngeal discomfort, which could be tolerated in a short time. There were no significant differences in clinical information, parameters at the last followup and postoperative ASMs adjustment between responders (16 cases) and nonresponders (9 cases) during VNS efficacyrelated factors analysis. Conclusion：The responder rate of VNS treatment for infants with DRE could reach 60% at 1 and 2 years of followup, and the safety of VNS treatment was good. The lower age limit of VNS still needs to be further evaluated.

Key words: Vagus nerve stimulation, Drug-resistant epilepsy, Infants, Efficacy