BACKGROUND AND AIM Although percutaneous endoscopic gastrostomy (PEG) has become established as a useful enteral nutrition technique, the associated risks must always be kept in mind. Recently, we experienced several patients who could orally ingest after PEG. To avoid unnecessary PEG, we investigated patients who could orally ingest after PEG, and analyzed predictive factors of postoperative oral feeding. METHODS We retrospectively analyzed data of 302 patients who underwent PEG at our hospital. After all patients were divided according to postoperative oral feeding status, we assessed factors of patients' backgrounds. In patients who could orally ingest after PEG, we investigated the course of oral feeding status. We attempted to identify predictive factors for postoperative oral feeding using logistic regression analysis. RESULTS Mean age was high in both groups, and overall condition was markedly poor. Forty-four patients (15%) were able to ingest orally after PEG. Enteral nutrition could be avoided during our observation period in 15 cases, because sufficient oral intake was achieved. Conversely, oral feeding was reduced or discontinued in 14 cases. Multivariate analysis identified the following independent predictive factors for postoperative oral feeding: (i) absence of dysphagia or aphagia; (ii) younger age; (iii) favorable performance status; (iv) presence of post-traumatic encephalopathy; and (v) preoperative swallowing training. CONCLUSIONS A total of 15% of PEG cases were able to ingest orally after PEG. In patients showing positive predictive factors, indications for PEG should be carefully considered.