Background: The appropriate duration of fasting therapy after endoscopic hemostasis for the prevention of recurrent bleeding from non-variceal upper gastrointestinal bleeding (NVUGIB) remains unclear.
Objective: To determine the appropriate duration of fasting therapy using a national administrative database.
Design: Retrospective cohort study. Setting: 475 hospitals between April and December in 2008.
Patients and interventions: 3254 patients with endoscopic hemostasis for NVUGIB.
Main outcome measurements: Patients were divided into 5 groups according to duration of fasting therapy, namely 1, 2, 3, 4, and 5 d or more. We compared the risk-adjusted rate of recurrent bleeding in relation to the duration of fasting therapy. Odds ratio (OR) and their 95% confidence intervals (CI) for recurrent bleeding after fasting therapy were estimated using a multiple logistic regression model.
Results: Recurrent bleeding rate after fasting therapy significantly differed by fasting duration, at 21.7% of patients with 1 d of fasting therapy, 12.3% with 2 d, 4.8% with 3 d, 4.6% with 4 d, and 3.0% with 5 d or more (p<0.001). Multiple logistic regression analysis showed no significant difference in the risk of recurrent bleeding between fasting for 3 or 4 d and 5 d or more (OR of fasting for 3 d 1.635, 95% CI 0.793–3.366, p=0.182; and OR of fasting for 4 d 1.548, 95% CI 0.705–3.402, p=0.276).
Limitations: Use of claims data lacking clinical detail.
Conclusions: 3 d of fasting therapy seems to be suitable for the prevention of recurrent bleeding from NVUGIB.