Abstract

The indications for percutaneous endoscopic gastrostomy (PEG) and patient outcome, were examined prospectively in the setting of a general hospital. In the course of 26 months, 76 patients underwent PEG (median age 62 years (range 18-99)) and were followed up for 6887 patient days. The median (range) duration of PEG feeding was 93 (3-785) days. The procedure was carried out for neurological indications in 76% of cases (stroke 51%) and 53% of patients were severely malnourished (body mass index < 17 kg/m2) at the time of referral. In 12 (16%) patients swallowing recovered and the PEG was removed after a median (range) of 55 days (20-150). Three (4%) deaths were related to PEG (one oesophageal perforation, one haemorrhage, and one aspiration pneumonia). One patient developed peritonism and ileus, which resolved with conservative treatment. Minor complications included local sepsis 3%, tube blockage 12%, and tube connector leak 5%. During seven days of observation, demands on nursing time for routine care of the PEG were the same as for nasogastric tube feeding, median (range) 21 (4-42) v 16 (4-40) min/day respectively, but in about half the latter cases the tube had to be replaced at least once. Over 15 months, 29 patients were randomised to receive a 1.9 mm inner, 2.9 mm (9F) outer diameter Fresenius and 27 a 3.0 mm inner, 4.0 mm (12F) outer diameter Bower polyurethane tube and were followed for 2920 and 2388 patient days respectively. There was no difference in the insertion time (median (range) 20 (10-45) v 24 (10-45) min respectively) or number of patients with complications (three v eight patients NS), although there were more minor mechanical problems (three v 12, p < 0.01) with the 12F tube. The internal anchoring device of the 12F tube allowed its non-endoscopic removal, a method applicable too 16% of cases. No tubes were removed because of blockage.