ABSTRACT Objective The objective of this practice guideline is to offer evidenced-based recommendations regarding the optimal management of patients with gastrointestinal bleeding. This guideline is meant to apply to adult patients who present to medical attention with subacute or acute gastrointestinal bleeding thought to arise from either the upper or lower gastrointestinal tract. While some of the recommendations may apply to patients who develop gastrointestinal bleeding during the course of hospitalization for another disorder, such patients are excluded from formal consideration. The guideline is intended to improve the overall care of patients with gastrointestinal bleeding, to reduce unnecessary practice variations, and to lower costs without reducing quality of care. Options The options for diagnosis and treatment include a variety of laboratory and diagnostic tests and risk stratification techniques intended to guide 1) initial resuscitation and stabilization efforts 2) triage from the emergency department and 3) triage and length of stay following endoscopy. Other treatment options considered include the use and timing of endoscopy, the use of blood products and antisecretory agents, the use of medications for the treatment of esophageal disorders, and the role of surgical consultation. Outcomes The outcomes considered in reviewing the literature relevant to the development of this guideline included rebleeding rates, transfusion requirements, mortality, length of stay and cost of treatment. Evidence Guideline recommendations are derived from multiple sources ranging from well designed randomized controlled trials to expert opinion. These sources were identified by computerized literature search, reviews of bibliographies of relevant articles and discussions with local experts in the field. Articles were included for consideration as late as July 1997. Risk stratification techniques were derived and adapted from several large prospective studies of prognosis in gastrointestinal bleeding. Recommendations concerning the use of therapeutic endoscopy, octreotide, and H2 receptor antagonists were derived from well designed blinded randomized controlled trials. All evidence was examined by committee members and graded according to a UCSF Practice Guideline Review Committee approved format. Values  The assignment of values to health outcomes was made jointly by a committee which included subspecialty (GI and Anesthesiology) and generalist (General Internal Medicine and General Surgery) physicians, nurses (Emergency Department and Medical Unit) and quality analysts. Valuation of outcomes was made by informal consensus. Where data was insufficient for the expert panel to make evidenced based recommendations, a high value was placed on patient safety. For example, there are no clinical trials to guide length of stay for patients with moderate risk endoscopic findings. In these situations, the committee chose to err on the side of caution and suggest a length of stay similar to that of patients with high risk endoscopic findings. Benefits, Harms and Costs The application of this guideline is intended to improve the overall care of patients with gastrointestinal bleeding by identifying and codifying current best practices. We are not able to estimate the potential effectiveness of such recommendation in quantitative terms. Similar guidelines developed at other institutions have reported reductions in overall length of stay and costs without adversely effecting outcome.   Key Recommendations Rapid evaluation and triage of patients from the Emergency Department is the critical first step towards ensuring successful outcomes. Resuscitation efforts should be guided primarily by the hemodynamic consequences of the bleeding episode. History and physical examination should focus on those factors which have been shown to predict etiology and determine outcome from gastrointestinal bleeding. Laboratory testing should be used to evaluate the extent of blood loss and to corroborate the presence of co-morbidities. Nasogastric lavage and stool examination are important bedside diagnostic tests which provide valuable information for risk stratification. Triage from the Emergency Department should be guided by an assessment of the patients risk for further bleeding or death. Patients judged to be at low risk should be transferred to the Endoscopy Unit prior to admission since many will be found to have lesions amenable to outpatient management. Patients at moderate or high risk should be admitted for further stabilization and treatment of co-morbidities with the goal of performing endoscopy as soon as possible. Non endoscopic therapies of proven benefit include Octreotide for the control of variceal hemorrhage and various measures used to reverse coagulopathy due to clotting factor deficiency or qualitative or quantitative platelet disorders . H-2 Blockers have not been shown to improve outcome of gastrointestinal bleeding, although they do promote ulcer healing. Recommendations regarding the need for continued hospitalization following endoscopy are based on the combination of clinical and endoscopic predictors of outcome. Sponsors Key sponsors of this guideline included the UCSF Department of Medicine and the UCSF Office of Clinical Resources Management. Copyright 1998 Regents of the University of California. All Rights Reserved. Last updated . Please direct comments and questions to the OCRM Practice Guideline Program at pgp@email.his.ucsf.edu For technical support service please contact: webmaster@email.his.ucsf.edu