The main purposes of this study are to compare clinical outcomes of two groups of patients with similar medical conditions (one with non-variceal upper gastrointestinal (UGI) lesions such as ulcers and another group with varices or portal hypertensive lesions) who are treated either with current standard visually guided endoscopic treatment according to stigmata of hemorrhage or with endoscopic Doppler endoscopic ultrasound probe (DEP) monitoring of blood flow in the lesion.

Background: For patients with severe upper gastrointestinal (UGI) hemorrhage, risk stratification for rebleeding or endoscopic hemostasis has been based on visually defined endoscopic stigmata of hemorrhage for more than 40 years. These are imperfect, because there is significant interobserver variation and the investigators were previously unable to detect underlying vessel blood flow, which is directly related to the risk of rebleeding. Doppler ultrasound endoscopic probes (DEP) are relatively new in the US and can detect underlying lesion blood flow. However, few patients with severe UGI hemorrhage have been studied with endoscopic doppler ultrasound probe in the US and none in VA's. The Center for Ulcer Research and Education (CURE)/VA GI Hemostasis Research Group will perform prospective randomized studies at the West Los Angeles VA and University of California Los Angeles (UCLA) Ronald Reagan Medical Centers to define the role of DEP in management of severe UGI hemorrhage.

Objectives: The specific aims (SA) of this research are: #1) In a randomized, blinded prospective controlled (RCT) study of patients with severe UGI hemorrhage (from varices/portal hypertensive lesions as one separate group vs. ulcers, & other benign non-variceal sources as another group) to compare 30 day outcomes of patients in these 2 major disease groups, managed by current standards (according to endoscopic visualization & stigmata of hemorrhage) with similar patients assessed, risk stratified, & treated with DEP monitoring. #2). For patients randomized to the DEP group, to determine the initial prevalence, type (arterial or venous), & location & course of blood flow underlying stigmata for the different lesion groups. #3). For DEP patients with different lesion types, to determine the rates of persistent blood flow after endoscopic hemostasis treatments & whether blood flow after under stigmata can be eliminated by further endoscopic hemostasis with different techniques. #4). To determine the proportion of patients whose risk stratification (for rebleeding) &/or endoscopic treatment are changed by utilizing DEP for detection of blood flow under stigmata of hemorrhage or lesions before as a guide for endoscopic treatment & absence of flow after treatment as a treatment endpoint rather than visual guidelines (stigmata) alone for endoscopic treatment. #5). To compare the outcomes for a large cohort of historical controls previously treated for hemostasis of the two lesion types in the UGI tract by the CURE/VA Hemostasis Research Group according to visual guidelines & stigmata alone with patients in the RCT managed with DEP findings & stigmata, contrasting demographics, hemostasis rates, rebleeding rates & other outcomes up to 30 days for major diagnoses (ulcers and other non-variceal lesions vs varices & other lesions related to portal hypertension). #6). For patients who are treated with surgery or angiography for continued bleeding or rebleeding of UGI lesions, to correlate & compare their vessel depth & location (relative to stigmata), type of vessel, & lumen patency at surgery or angiography vs. Doppler endoscopic probe findings recorded previously. Research Plan and Methods: All studies will be performed over 5 years. The investigators will utilize a large RCT (blinded), a very large cohort study, & prospective observational studies to complete the specific aims of the study. Statistical Analysis System (SAS) will be utilized for data management. For SA #1, about 240 new patients (150 with non-variceal lesions and 90 with variceal-portal hypertensive lesions) admitted to West Los Angeles (WLA) VA or UCLA Hospitals with severe UGI hemorrhage will be randomized in RCT of Doppler assisted management versus standard endoscopic/medical diagnosis, risk stratification, & treatment. During urgent endoscopy, patients with clean varices (as the source of bleeding) or other UGI lesions with stigmata of recent hemorrhage (from the ulcers, Mallory Weiss tears, esophageal or gastric varices, and Dieulafoy's lesions) will be randomized. Routine clinical outcomes will be assessed prospectively & compared by major diagnostic groups (ulcers-non-variceal lesions or varices-portal hypertensive lesions). For SA #5 (UGI cohort study), demographics, outcomes, & risk factors will be compared for about 150 patients with non-variceal lesions & about 90 other variceal-portal hypertensive lesion matched historical control patients (from CURE Hemostasis Research Databases) treated previously only based upon stigmata vs. 75 new patients with non-variceal UGI lesions or 50 variceal-portal hypertension lesions in this study treated according to stigmata of hemorrhage & DEP as a guide to risk stratification & endoscopic hemostasis. SA #2-4 & 6 will be prospectively performed according to the methods in the proposal & all analysis will be performed with the collaboration of an experienced biostatistician.

Potential Impact on Veterans and Non- VA Healthcare: These studies will increase our knowledge about UGI bleeding, UGI lesion vasculature, blood flow, and effects of endoscopic hemostasis. Also, DEP may improve risk stratification of Veteran patients, medical and endoscopic management & outcomes of patients with severe UGI hemorrhage from different etiologies. This is particularly relevant to patient care of Veterans with severe UGI hemorrhage, since this is a common clinical condition which requires considerable health care resources in every VA hospital. These results will also be generalizable to non-VA hospitals and the US population who is hospitalized with severe UGI bleeding.

Inclusion Criteria:

• Patients must have evidence of severe UGI bleeding by laboratory tests (Hgb less than or equal to 9 gms; with red blood cell (RBC) transfusions; or documented decrease in Hgb of greater than or equal to 2 gms relative to baseline) and by clinical parameters (melena; hematemesis or hematochezia; nasogastric tube (NG) evidence of UGI bleeding- fresh blood, clots, or old blood).
• The following non-variceal UGI lesions will be included if stigmata of hemorrhage are found on emergency endoscopy (active arterial bleeding, oozing, non-bleeding visible vessel (NBVV), adherent clot, or flat spot or a combination of these) for peptic ulcers (gastric, duodenal, esophageal, or anastomatic), Mallory-Weiss (MW) tears without portal hypertension (PHTN), or Dieulafoy's lesions.
• For other types of severe UGI bleeding related to PHTN, the investigators will include patients with esophageal or gastric varices (with or without stigmata, if no other UGI lesion is the source of the bleed); post-rubber band ligation (RBL) ulcers, and MW tears associated with PHTN and having some stigmata of recent hemorrhage.
• Life expectancy of at least 60 days based on lack of very severe or terminal comorbidity, as judged by the generalists or specialists caring for the patient.
• Written informed consent by patient or surrogate.

Exclusion Criteria:

• Patients who are uncooperative, unable to give written informed consent, who cannot return for 30 day follow-up, or refuse informed consent.
• Patients with UGI known malignancies or malignant appearing ulcers; diffuse bleeding from mucosal lesions or esophagitis; infectious UGI lesions; or other bleeding lesions (polyps, post-endoscopic mucosal resection (EMR), or post-sphincterotomy).
• End-stage, very severe, recurrent or ongoing co-morbid illness, e.g. severe liver, renal, cardiac, respiratory failure; peritonitis; or sepsis that preclude emergency procedures or clinical follow-up and limit survival.
• Persistent shock or hypotension (e.g. systolic blood pressure less than or equal to 99 mm mercury) that is unresponsive to less than or equal to 6 units of packed red blood cell (RBC) transfusions or requires continuous intravenous infusions of vasoactive drugs for blood pressure elevation.
• Severe coagulopathy unresponsive to blood transfusions e.g. international normalized ratio (INR) > 2.0, platelet count < 20,000, activated partial thromboplastin time (APTT) greater than 2.0 x normal, or bleeding time > 10 minutes.
• Contraindication to urgent endoscopy or follow-up procedures.