Risk factors for mortality in patients undergoing percutaneous endoscopic gastrostomy

Background and Study Aims: Percutaneous endoscopic gastrostomy (PEG) is a method used for feeding patients who are unable to eat. High early mortality rates among hospitalized patients have been reported. The aim of this study was to shed light on the risk factors for early mortality after PEG tube insertion. Patients and Methods: Outpatients from nursing homes and hospitalized patients who underwent PEG between July 1995 and July 2001 were compared. Survival analysis was used to assess mortality after PEG. In a logistic regression analysis, mortality within 30 days among hospitalized patients was chosen as the outcome variable and the predictor variables were demographic characteristics, co-morbid conditions, and indication for PEG. Results: A total of 502 PEG tubes were inserted in 419 hospitalized and 83 nursing-home patients. The prevalence of co-morbid conditions was similar in the two groups. Both the 30- and 60-day mortality rates were around six times higher in the hospitalized patient group than in the nursing-home patient group (30-day mortality rate 8% vs. 1.2%, P = 0.034; 60-day mortality rate 12% vs. 2.4%, P = 0.016). Risk factors for 30-day mortality among hospitalized patients were: serum albumin <3 g/dl (odds ratio 2.82, 95% CI 1.34-5.96), chronic obstructive pulmonary disease (odds ratio 2.79, 95% CI 1.26-6.14), and diabetes mellitus (odds ratio 2.44, 95% CI 1.20-4.97). Conclusions: Compared with nursing-home patients, hospitalized patients are at higher risk for early mortality after PEG. The presence of diabetes, chronic obstructive pulmonary disease, and a low serum albumin level each increase the 30-day mortality risk among hospitalized patients threefold.