Is endoscopic paravascular injection of sclerosing agents reasonable in the control of GI bleeding?

Background: The pharmacologic response and microvascular effects associated with the endoscopic injection of sclerosing agents around vessels (paravas cular injection) to stop bleeding from the digestive tract remain to be cla rified. Methods: Using in vivo microscopy, we directly visualized submucosal microv essels of the rat stomach and intestine. We studied differences among scler osing agents in thrombus formation and vascular diameter change that occur through a pharmacologic response and/or local compression after topical app lication or paravascular injection of the agents. Results: Except for absolute ethanol, topical application of the agents did not cause constriction or thrombi in either arterioles or venules. Polidoc anol topical application and paravascular injection significantly dilated a rterioles. Injecting ethanolamine oleate near venules constricted them the longest and most effectively, but vasoconstriction in arterioles was transi ent. Injecting absolute ethanol formed long-lasting thrombi and caused vaso constriction in venules, but arteriole thrombi persisted no more than 3 min utes. The vascular response to thrombin did not significantly differ from t hat to physiologic saline. Conclusion: The paravascular injection of ethanolamine oleate, because of i ts long-lasting vasoconstriction, or of absolute ethanol, because of its th rombogenic effect, is a valid therapeutic approach to treating venous bleed ing. The efficacy of paravascular injection of sclerosing agents for treati ng acute arterial bleeding, however, is not supported in this experimental model.