Survival analysis in percutaneous endoscopic gastrostomy feeding: a worse outcome in patients with dementia

Am J Gastroenterol. 2000 Jun;95(6):1472-5. doi: 10.1111/j.1572-0241.2000.02079.x.


Objective: Percutaneous endoscopic gastrostomy (PEG) feeding has been validated in specific clinical situations such as acute stroke with dysphagia and oropharyngeal malignancy. The perception that gastrostomy insertion is safe and technically simple has led to an increase in the demands for PEG insertion, encompassing clinical applications such as in patients with dementia, in whom its role has not been justified. The purpose of this study was to compare the mortality of patients with dementia who were fed by PEG to that of other subgroups of patients requiring gastrostomy feeding.

Methods: The study focused on a cohort of 361 consecutive patients requiring PEG feeding between August 1992 and July 1997 from two District General Hospitals (Rotherham District General Hospital and Doncaster Royal Infirmary) in South Yorkshire. A retrospective cohort survival analysis was performed using the Kaplan-Meier survival method and Cox proportional hazards analysis.

Results: In all patients requiring gastrostomy feeding there is a high initial mortality of 28% at 30 days. However, patients with dementia have a worse prognosis compared to other subgroups, with 54% having died at 1 month and 90% at 1 yr (log rank test p < 0.0001). This difference remained significant (log rank p < 0.0001) after adjusting for age at the time of PEG insertion.

Conclusions: This is the first demonstration in the United Kingdom that the mortality of patients with dementia who are fed by gastrostomy is considerable. Consequently, we may wish to advise against gastrostomy feeding in selected patients within this clinical setting.

MeSH terms

  • Aged
  • Cohort Studies
  • Contraindications
  • Dementia / complications*
  • Dementia / mortality
  • Endoscopy*
  • Enteral Nutrition*
  • Gastrostomy / methods*
  • Humans
  • Middle Aged
  • Proportional Hazards Models
  • Retrospective Studies
  • Survival Analysis
  • Treatment Outcome
  • United Kingdom