Artificial feeding by a percutaneous endoscopic gastrostomy (PEG) tube in patients with dementia has increased since the introduction of the endoscopic method of tube placement. Few studies have documented survival benefit from this intervention. This report reviews our experience with PEG tube placement for feeding patients with dementia. All consultations for PEG tube placement were evaluated by a certified nutrition support nurse (L.M.M.) in consultation with a member of the gastroenterology physician staff (T.O.L.) for 24 months. Evaluation included the attainment of a brief medical history, a physical examination, and a review of comorbid conditions, laboratory variables for nutrition status, and bleeding risk. Interviews with patients or surrogates were conducted, including an explanation of the risks and benefits of PEG tube placement. A Kaplan-Meier survival curve was used to compare the median survival between patients with dementia who received a PEG tube and patients with dementia in whom PEG tube placement was refused. We received 41 consultations for PEG tube placement in patients with dementia. Percutaneous endoscopic gastrostomy was performed in 23 patients; 18 patients met the medical criteria for PEG tube placement, but surrogates refused placement. The median survival for the 23 patients who underwent PEG was 59 days; the median survival for the 18 patients who did not undergo PEG was 60 days. There seems to be no survival benefit in patients with dementia who receive artificial feeding by a PEG tube.