Prospective evaluation of immediate versus delayed refeeding and prognostic value of endoscopy in patients with upper...

The effects of immediate vs. delayed refeeding and the prognostic value of endoscopic findings in patients with major upper gastrointestinal hemorrhage were assessed in a prospective randomized study. Entry criteria were clinical evidence of major hemorrhage and endoscopic evidence of a Mallory-Weiss tear or an ulcer with a clean base, flat spot, or clot. Two hundred fifty-eight patients were randomly assigned to groups receiving a regular diet immediately or nothing by mouth for 36 hours, then clear liquids for 12 hours, and a regular diet thereafter. Outcomes in the immediate and delayed refeeding groups were comparable: rebleeding occurred in 4% vs. 5%; urgent intervention, 2% vs. 2%; and deaths, 1% vs. 1%, respectively. Rebleeding occurred in 2 (2%) of 96 patients with cleanbased ulcers, 5 (8%) of 65 with ulcers with spots, 3 (14%) of 21 with ulcers with clots (P = 0.05, 3 x 2 chi2 test), and 1 (2%) of 66 with Mallory-Weiss tears. It is concluded that the time of refeeding does not influence the hospital course of patients with a low risk of recurrent bleeding. Patients with clean-based ulcers or nonbleeding Mallory-Weiss tears may be refed and discharged home immediately after stabilization.