Background and study aims: Rebleeding after initial endoscopic hemostasis remains an important determinant of poor prognosis in patients with ulcer hemorrhage. Prospective identification of patients who are at high risk for rebleeding, and directing further therapeutic measures in these patients, would be expected to improve the prognosis.
Patients and methods: We used our previously described scoring system (the Baylor Bleeding Score) to identify patients at increased risk for rebleeding after initially successful endoscopic hemostasis, and randomized them into retreatment and no-retreatment groups. Patients in the retreatment group received additional endoscopic heat probe therapy at 24 hours, the follow-up was otherwise similar. Forty men with major ulcer hemorrhage were randomized, 19 to the retreatment group and 21 to the no-retreatment group. Patients in the two groups had similar Baylor Bleeding Scores, blood transfusion needs (median three vs. two units), ulcer sites (duodenal 57% vs. 43%), stigmata of hemorrhage (actively bleeding 68% vs 67%), and received similar initial endoscopic therapy (heat probe 240 J vs. 300 J). Patients in the retreatment group received additional heat probe therapy (median 165 J) at 24 hours.
Results: Rebleeding, requiring significantly more blood transfusion, occurred in 24% of the no-retreatment group patients compared to none of the retreatment group patients (p < 0.05).
Conclusions: Endoscopic retreatment after successful initial endoscopic hemostasis prevents rebleeding in highrisk patients with ulcer bleeding.