Some authors recommend a routine "protective" antireflux procedure (ARP) in neurologically impaired children undergoing feeding gastrostomy (FG). Over 4 years, we performed FG in 107 neurologically impaired children aged 1 month to 16 years. Ninety-eight had preoperative radiological assessment for gastroesophageal reflux (GER), which was documented in 44, of whom 33 had FG plus ARP and 11 had FG alone. Seven of the 11 subsequently developed symptomatic GER and 5 of them had a secondary ARP. Of the 54 children with no demonstrable GER preoperatively, 3 children underwent FG plus ARP. The remaining 51 had FG alone. There was one postoperative death. Of the 50 surviving patients, 22 developed symptomatic GER and 17 of these had a subsequent ARP. Mean follow-up of 20.0 months showed that the risk of developing GER after FG alone was not influenced by age, sex, indication for FG, underlying diagnosis, or method of gastrostomy. There was no significant difference in mortality and early morbidity between patients undergoing FG alone and those having FG with simultaneous or subsequent ARP. Only 44% of our patients in whom GER was not demonstrated initially developed symptomatic GER followed FG alone. This incidence does not justify a routine "protective" ARP.