Outcome after percutaneous endoscopic gastrostomy in children and young adults

Abstract

Background and objectives: Factors predicting outcome after percutaneous endoscopic gastrostomy (PEG) in large pediatric cohorts are not well defined. We hypothesized that definable preoperative clinical factors predict the need for further intervention to provide enteral access after PEG. Our aim was to identify factors associated with PEG outcome.

Materials and methods: A retrospective review of 760 (407 boys and 353 girls) patients was performed after PEG at the Johns Hopkins Children's Center from 1994 to 2005. Logistic or multiple linear regression was used to analyze indication; diagnosis; age; prematurity; neurological impairment; weight-for-age z scores; modified barium swallow; postoperative complications; need for fundoplication (FP), gastrojejunal tube, or jejunostomy; and length of hospital stay.

Results: The median age was 1 year (range 0-26 years). The most common indications given for PEG were failure to thrive (n = 373) and dysphagia (n = 27). Postoperative FP, gastrojejunal tube, or jejunostomy were performed in 66 (10%), 24 (4%), and 9 (1%) patients, respectively. Preoperative report indicated that dysphagia and direct aspiration on modified barium swallow was strongly associated with patients undergoing FP after PEG, 10.6% of patients (P = 0.008, odds ratio 2.4) and 11.2% of patients (P = 0.013, odds ratio 2.8), respectively. Younger preoperative age was also associated with the need for FP (P = 0.0006; median age of 5.8 vs 14 months). Patients with preoperative dysphagia had a longer median length of hospital stay: 8 versus 3 days (P < 0.00001). Patients with neurological impairment demonstrated greater weight gain than neurologically normal patients after PEG (P = 0.04). Minor postoperative complications (most commonly wound infection) were observed in 4% (27/747) of children before hospital discharge from PEG and in 20% of children (138/682) after discharge. There were only 2 major complications (gastric separation and gastrocolonic fistula.). There were no fatalities.

Conclusions: Preoperative diagnosis, indication, prematurity, and neurological impairment did not influence postoperative complications.