Continuing Education Activity

Upper gastrointestinal bleeding describes blood loss from a gastrointestinal source localized to the esophagus, stomach, or duodenum, proximal to the ligament of Treitz. In the United States, gastrointestinal (GI) bleeding is the most common GI diagnosis requiring hospitalization; upper GI bleeding is estimated to occur in 80 to 150 out of 100,000 people each year, with an estimated mortality rate of 2% to 10%. Upper GI bleeding may be acute or chronic, slow or brisk, obscure or overt, depending on the underlying etiology, bleeding rate, and chronicity of blood loss. Gastrointestinal manifestations of upper GI bleeding include hematemesis, coffee-ground emesis, hematochezia, or melena. Patients may also experience systemic symptoms such as syncope, fatigue, palpitations, exertional dyspnea, or weakness. Timely triage, early resuscitation, and endoscopic evaluation are critical to improving outcomes for patients with upper GI bleeding. This activity for healthcare professionals is designed to enhance the learner's competence in recognizing upper gastrointestinal bleeding risk factors and clinical features, performing the recommended evaluation, and implementing an appropriate interprofessional management approach to improve patient outcomes.

Objectives:

• Identify patients with upper gastrointestinal bleeding based on their clinical history and presentation.

• Determine the etiology of upper gastrointestinal bleeding using diagnostic tools.

• Apply best practices and evidence-based guidelines when evaluating and treating a patient with upper gastrointestinal bleeding.

• Develop and implement interprofessional team strategies to improve care coordination and outcomes in patients with upper gastrointestinal bleeding.

Introduction

Upper gastrointestinal bleeding (UGIB) is a common problem with an annual incidence of approximately 80 to 150 per 100,000 population, with an estimated mortality rate of 2% to 10%.[1] UGIB is classified as any blood loss from a gastrointestinal source in the esophagus, stomach, or duodenum; historically defined as above the ligament of Treitz, a fibromuscular band extending from the upper surface of the duodenojejunal junction. UGIB can manifest as hematemesis, be it bright red or coffee-ground emesis, hematochezia, or melena. Patients may also present with systemic symptoms secondary to blood loss, such as orthostasis, syncope, fatigue, and weakness.[2][3][4] Prior data from the United States show that UGIB accounts for nearly 40% of hospital admissions for gastrointestinal bleeding, with an estimated annual cost of over 2 billion dollars.[5]

Etiology

Peptic ulcer disease (PUD), esophagitis, gastroduodenitis, and varices are the most common etiologies of UGIB. Older data estimate PUD accounting for up to 50% of cases; more recent studies show a decrease of PUD-related UGIB to 32% to 36%. Esophagitis accounts for 24%, gastritis 18% to 22%, duodenitis 13%, and varices 11%.[1][6][7][8] Other etiologies of UGIB include Mallory-Weiss tears (5% to 15%), vascular ectasia (5%), neoplasms, and portal hypertensive gastropathy. Less common causes include Dieulafoy lesions, gastric antral vascular ectasia, aortoenteric fistulae, hemobilia, and hemosuccus pancreaticus. 

Epidemiology

UGIB is more common in men than women, and the incidence increases with advancing age, irrespective of genotype.[3] UGIB accounts for 75% of all acute gastrointestinal (GI) bleeding cases. The annual incidence of UGIB is approximately 80 to 150 per 100,000 population.[1] UGIB hospitalizations decreased by 21% in the United States from 2002 to 2012.[9] The in-hospital upper GI bleed mortality rate in the United States declined between 1989 and 2009. This shifting epidemiology has been attributed to increased awareness and treatment of Helicobacter pylori infections and antisecretory medications.[10]

Patients on long-term, low-dose aspirin have a higher risk of overt UGIB compared to placebo. When aspirin is combined with P2Y12 inhibitors, such as clopidogrel, the number of UGIB cases increases by 2- to 3-fold. When a patient requires triple therapy, such as a regimen comprising aspirin, a P2Y12 inhibitor, and a vitamin K antagonist, the risk of UGIB is even higher.[11] Patients who are 75 years of age or older on combination antiplatelet and anticoagulant therapies have the highest GI bleeding risk of up to 17% per year.[12]

Pathophysiology

The most common cause of UGIB is peptic ulcer disease (PUD). Ulceration occurs due to a breakdown in the mucosal defense mechanism from endogenous (eg, acid, pepsin, bile) and exogenous (eg, infection, drugs, smoking) factors. H pylori infection and the use of nonsteroidal antiinflammatory drugs (NSAIDs) are the 2 most common causes of PUD. H pylori causes mucosal inflammation, leading to epithelial cell degeneration and injury. NSAIDs cause mucosal injury by blocking the cyclooxygenase-1 (COX-1) pathway, leading to inhibition of prostaglandin synthesis. Prostaglandins are necessary molecules that maintain the mucosal barrier.[13]

Variceal bleeding can occur in patients with an elevated portal vein pressure, defined as >12 mm Hg. Locations of increased resistance of portal blood flow include prehepatic (eg, portal vein thrombosis), intrahepatic (eg, cirrhosis), and posthepatic (eg, hepatic vein or inferior vena cava thrombosis). Increased portal vein pressure leads to the formation of portosystemic collaterals, a complex process of gradual dilation and hypertrophy of preexisting vascular channels called varices.[14][15][16]

History and Physical

During history-taking, attention should be given to comorbidities, prior endoscopic interventions, and abdominal surgeries. A detailed review of current medications should be performed, and patients should be directly asked about the use of NSAIDs, antiplatelet drugs, aspirin, anticoagulants, and herbs. Various medications and herbs can interfere with the metabolism of anticoagulants, leading to increased bleeding risk. A detailed medical and social history regarding alcohol and illicit substance use can further assess the risk for underlying liver disease. Clinicians should also assess for a family history of gastrointestinal malignancies and bleeding disorders.

The clinical presentation of a patient with UGIB can vary but should be well-characterized by asking the patient to describe the bloody output and performing a rectal exam. Hematemesis is vomiting of fresh blood or clots. The term "coffee-ground" describes gastric aspirate or vomitus that contains dark specks of old blood. Melena refers to black, sticky, and tarry-appearing stools with a distinctive odor. Hematochezia is the passage of bright red blood or clots per rectum. The latter usually reflects lower gastrointestinal bleeding (LGIB) but may be seen in patients with brisk UGIB. Patients may also present with syncope or orthostatic hypotension if bleeding is severe enough to cause hemodynamic instability.

As part of a thorough physical examination, clinicians should evaluate for tenderness to abdominal palpation and associated rebound or guarding, as these findings suggest possible gastrointestinal perforation. Clinicians should also search for evidence of chronic liver diseases, including palmar erythema, spider angiomas, gynecomastia, jaundice, and ascites, which may suggest the underlying etiology of the bleeding (eg, variceal bleeding).

Evaluation

Laboratory Evaluations

Initial laboratory assessments should include a complete blood cell count (CBC) to evaluate current hemoglobin, hematocrit, and platelet levels. A low mean corpuscular volume can indicate chronic blood loss and iron deficiency anemia. Blood chemistries should also be evaluated. Specifically, an elevated blood urea nitrogen to creatinine ratio >30 suggests an upper GI source of bleeding.[17] A coagulation panel, including the international normalized ratio (INR), should also be performed, particularly in patients with liver dysfunction, taking anticoagulant therapy, or with underlying bleeding diatheses.[18][19][20]

Risk Stratification

Scoring systems help predict which patients need intervention and are at higher risk for rebleeding and mortality. Risk stratification tools guide resuscitation, endoscopy timing, and discharge planning. 

• The Rockall score was designed to predict rebleeding and mortality and includes age, comorbidities, the presence of shock, and endoscopic stigmata. A pre-endoscopic Rockall scoring system is also available and can be used to stratify a patient's risk for rebleeding and mortality before the endoscopic evaluation. When the Rockall score is used, patients with ≤2 points are considered low risk and have a 4.3% probability of rebleeding and 0.1% mortality. In contrast, patients with a score of ≥6 have a rebleeding rate of 15% and mortality of 39%.[21][22]
• The AIMS65 score predicts in-hospital mortality, length of stay, and cost in patients with UGIB. The individual parameters of the score are albumin, INR, mental status change, systolic blood pressure, and age 65 years or older.[23]
• The Glasgow-Blatchford Score was designed to predict the need for endoscopic therapy and includes hemoglobin levels, blood pressure, presentation of syncope, melena, liver disease, and heart failure. A score of ≥6 is associated with a >50% risk of needing an intervention. Patients with a low Glasgow-Blatchford Score score of 0 or 1 can be discharged from the emergency department with outpatient follow-up.[24] In an international, multicenter, prospective study comparing various scoring systems, the Glasgow-Blatchford Score best predicted the need for intervention or death; hence, the Glasgow-Blatchford Score has become the recommended risk stratification tool in both the American College of Gastroenterology (ACG) and the European Society of Gastrointestinal Endoscopy (ESGE) UGIB guidelines.[25][26][27]

Endoscopy

Upper endoscopy or esophagogastroduodenoscopy (EGD) remains the gold standard for identifying the etiology and potentially treating upper GI bleeding. A recent randomized controlled trial addressed the optimal timing of EGD in UGIB, comparing urgent endoscopy (<6 hours) to early endoscopy (6 to 24 hours) after initial gastroenterology consultation in patients presenting with Glasgow-Blatchford Score score of ≥12. The study excluded patients with refractory hypotension despite resuscitation. The mean time from presentation to gastroenterology consult was approximately 8 hours for both groups, and the mean time from presentation to endoscopy was 9.9 +/- 6.1 hours in the urgent group and 24.7+/- 9.0 hours in the early group. The primary endpoint was death from any cause within 30 days after randomization. Results showed early endoscopy was not associated with lower 30-day mortality and no significant difference in rebleeding rates was observed. However, more endoscopic hemostatic treatment was required in the urgent-endoscopy group.[28] Observational studies suggest endoscopy performed within 1 day of hospital admission leads to shorter hospitalization.[29][30] 

Based on these data, the ACG and ESGE continue to recommend that all patients with UGIB undergo endoscopy within 24 hours of admission, following resuscitative efforts to optimize hemodynamic parameters and other medical problems. The American Association for the Study of Liver Diseases (AASLD) recommends timely EGD performed within 12 hours of presentation in patients suspected to have an acute variceal hemorrhage.[16] This recommendation is based on studies showing increased rebleeding and mortality in patients with cirrhosis and acute variceal hemorrhage who received delayed endoscopy.[31][32]

Imaging

Computed tomography angiography (CTA) can be employed to localize the source of bleeding when upper endoscopy fails to identify an actively bleeding etiology, a gastroenterologist is unavailable, the patient is unsuitable to proceed to endoscopy, or postsurgical anatomy limits endoscopic access. The data for CTA performance in UGIB is relatively small, as prior studies mainly evaluated lower GI bleeding. CTA detects bleeding rates of around 0.1 mL/min and is a diagnostic procedure lacking therapeutic capabilities.[33] Thus, a positive CTA warrants follow-up interventions such as endoscopy or transcatheter angiography with interventional radiology to treat active bleeding.[34] 

Treatment / Management

In 2021, the ACG and ESGE published guidelines on UGIB management.[26][27] The American Society of Gastrointestinal Endoscopy (ASGE) and the ACG published guidelines on anticoagulant and antiplatelet management during acute gastrointestinal bleeding in 2016 and 2022, respectively.[35][36] In 2024, the American Association for the Study of Liver Diseases (AASLD) published guidelines on managing varices and bleeding. [16] These guidelines state:

Resuscitation

Early intensive resuscitation reduces mortality in patients presenting with UGIB.[37] Treatment of underlying medical comorbidities is of prime importance as the vast majority of patients presenting with peptic ulcer bleeding die of other causes.[38] Patients should have a minimum of 2 large-bore (at least 18-gauge) peripheral access catheters. Intravenous fluids should be administered to maintain adequate blood pressure and hemodynamic stability. If patients are unable to protect their airway or have ongoing severe hematemesis, elective endotracheal intubation is advised.

The recommended threshold for initiating blood transfusion in a patient presenting with acute UGIB is hemoglobin <7 g/dL. Data initially from a randomized controlled trial of 921 patients from Villenueva et al and a meta-analysis of 5 randomized controlled trials totaling 1965 patients comparing the restrictive transfusion strategy (<7 g/dL) to liberal transfusion (<9 g/dL) showed reduced mortality and rebleeding in patients who were treated with the restrictive transfusion strategy.[39][40] In patients with existing cardiovascular disease or experiencing acute coronary syndrome, initiating transfusion when hemoglobin falls below 8 g/dL is advised due to the paucity of data in this patient population. Prior trials either excluded patients with recent cardiovascular events or patients presenting with acute coronary syndrome. 

High-quality data is lacking for platelet transfusion threshold in acute gastrointestinal bleeding for patients presenting with thrombocytopenia. A retrospective study of 144 patients who underwent endoscopy within 24 hours of overt GI bleeding with platelets 20,000/μL to <50,000/μL showed 94% initial hemostasis.[41] The median platelet count before endoscopy was 41,000/μL, and recurrent bleeding occurred in 22% of patients at 1 month. Sixty-one percent of the study cohort had cirrhosis. Results from this study suggest endoscopy can be performed in actively bleeding patients with platelet counts <50,000/μL with successful hemostasis; higher quality data is still needed. 

The American Association of Blood Banks (AABB) advises prophylactic platelet transfusion for counts >50,000/μL for major, non-neuraxial surgery but is not specific to gastrointestinal bleeding.[42] The ASGE recommends a target platelet count of >20,000/μL for diagnostic EGD and >50,000/μL for biopsies in their guideline on adverse outcomes in upper endoscopy; this recommendation is also not specific to patients presenting with bleeding.[43] Based on available data, for patients presenting with acute GI bleeding, particularly life-threatening bleeds, aim for a platelet count >50,000/μL if possible and consider consulting transfusion medicine in cases where platelet sequestration or destruction is problematic.

Anticoagulant and Antiplatelet Management

In patients with preexisting cardiovascular disease on aspirin for secondary prophylaxis, holding aspirin in the setting of acute UGIB is not advised due to prior RCT data showing increased 30-day mortality with aspirin cessation.[44] If aspirin is held, it should be promptly resumed as soon as possible once hemostasis is achieved, preferably within 3 to 5 days.[36][27] For patients on dual antiplatelet therapy, the ESGE advises continuing cardiac aspirin and holding the second antiplatelet therapy with the goal of resuming it as soon as possible, preferably within 5 days. In patients with significant cardiovascular history, cardiology consultation is advised. Platelet transfusion is not routinely advised due to the possibility of thrombotic events in patients requiring antiplatelet therapy.

For patients on anticoagulant therapy, be it a vitamin K antagonist or direct oral anticoagulants, the ACG suggests risk-stratifying patients based on the severity of gastrointestinal bleeding to determine whether holding anticoagulant therapy and reversal therapy is warranted. For patients presenting with a life-threatening bleed defined as hypovolemic shock or severe hypotension requiring vasopressors or surgery, or associated with a decrease in hemoglobin of >5 g/dL, or requiring ≥5 units of blood transfusion, or likely causing death, anticoagulant therapy should be held, and a reversal agent is advised. For patients with hemodynamic instability on vitamin K antagonists (VKA), the ACG and ESGE strongly recommend infusing prothrombin complex concentrate (PCC) due to rapid onset of action. Fresh frozen plasma (FFP) should not be routinely used due to increased infusion volume and risk of transfusion-related lung injury. FFP can be considered if PCC is not readily available.

The ACG does not advise routine use of Vitamin K in acute GI bleeding due to its relatively slow onset of action with effect seen around 24 hours and subsequent delays in reaching a therapeutic target once the vitamin K antagonist is restarted. Vitamin K can be used in cases where there are no plans for VKA reinitiation. For patients taking direct oral anticoagulants (DOAC), DOAC-specific reversal agents can be considered for patients with life-threatening bleeding. However, studies show marginal benefits, and the high cost and availability limit use.[45] Alternatively, PCC can be considered in patients with life-threatening bleeding who had taken a DOAC within the past 24 hours of presentation.

For patients on anticoagulant therapy presenting without life-threatening bleeding, VKA and DOAC can be held often without the need for a reversal agent. DOACs have a relatively short half-life (<12 to 15 hours) and are excreted mainly by the kidneys. Maximizing renal perfusion is essential to allow for natural clearance. In patients on VKA with supratherapeutic INR, PCC is preferred over FFP for rapid, reliable INR correction.[36]

Cohort studies examining patients presenting with nonvariceal UGIB and the impact of anticoagulation on rebleeding showed INR at the time of endoscopy was not a predictor of rebleeding after an endoscopic intervention. The percentage of patients with INR >2.5 in 2 studies was 7.6% and 22.9%.[46][47][48] Based on these results, the ESGE recommends against using an INR cutoff level to define or guide the timing of endoscopy in UGIB. While the ACG guideline does not specifically comment on the INR threshold for endoscopy in acute GI bleeding, the ASGE advises a reasonable INR <2.5 for urgent or emergent endoscopies.[35]

Medications

Proton pump inhibitors (PPI) block gastric acid production and stabilize clot formation, leading to mucosal healing. A 2007 randomized controlled trial by Lau et al compared giving high-dose PPI versus placebo before endoscopy in patients presenting with UGIB. Results showed that preendoscopy PPI reduced high-risk features in peptic ulcers and reduced the need for endoscopic therapy. However, there was no significant difference in transfusion requirement, rebleeding rate, and 30-day mortality.[49] Patients with significant bleeding should be treated with an 80mg bolus of PPI followed by continuous 8 mg/hr infusion or intermittent dosing of 40 mg 2 to 4 times daily. If endoscopy findings show peptic ulcers with low-risk features, then an oral daily dose can be continued for ulcer treatment. In patients with endoscopically identified peptic ulcers with high-risk features for rebleeding, high-dose PPI for 72 hours given intermittently or continuously after achieving endoscopic hemostasis has been found in multiple randomized controlled trials to lower risks of rebleeding, mortality, and surgery.[50][51][52][53] After 72 hours of high-dose therapy, patients can be transitioned to twice-daily oral PPI for 14 days and then weaned to a daily dose to complete ulcer treatment.[26][54] 

Vasoactive therapy (e.g., octreotide, somatostatin, or terlipressin) should be initiated if variceal bleeding is suspected and continued for 2 to 5 days.[55][56] These medications reduce portal pressure and collateral blood flow. Octreotide is given as an intravenous bolus of 50 µg, followed by a continuous infusion of 25 to 50 µg/hour. Somatostatin is given as an initial intravenous bolus of 250 µg, followed by a continuous infusion of 250 to 500 µg/hour. Terlipressin is given 2 mg intravenously every 4 to 6 hours for the initial 24 to 48 hours and then 1 mg intravenously every 4 to 6 hours. Antibiotic prophylaxis for 2 to 5 days is advised in patients with cirrhosis presenting with acute GI bleeding due to the risks of bacterial translocation, aspiration pneumonia, and spontaneous bacterial peritonitis. Ceftriaxone 1 g intravenously per 24 hours is commonly used. Antibiotics can be discontinued once bleeding is controlled and no active infection is evident. 

Prokinetic therapy with erythromycin before anticipated endoscopy has been shown through a meta-analysis of randomized controlled trials to improve gastric mucosal visualization during endoscopy and decrease the need for second-look endoscopy.[57] Erythromycin binds motilin receptors in the stomach and facilitates gastric motility. The ACG and ESGE recommend infusing erythromycin 250 mg IV 20 to 90 minutes before endoscopy to propel clots distally and optimize intraprocedural visualization.[26][58]

Endoscopy 

Endoscopic intervention may be warranted depending on endoscopy findings. The Forrest classification describes peptic ulcer appearance and guides endoscopic management. Peptic ulcers with high-risk features have active spurting or oozing, a Forrest classification of Ia and Ib, respectively, or nonbleeding visible vessel, a Forrest classification of IIa. Endoscopic treatment for Forrest classification Ia, Ib, and IIa ulcers is advised due to the high risk for persistent or recurrent bleeding without endoscopic intervention. Low-risk features such as flat pigmented spot, Forrest classification IIc, and clean base, Forrest classification III, do not warrant endoscopic therapy due to the low risk of rebleeding. Ulcers with overlying clots, Forrest classification IIb, could be considered for clot removal and further treatment based on Forrest classification after clot removal. 

Standard-of-care Therapies 

Standard-of-care therapies include epinephrine injection, contact and noncontact thermal coagulation, mechanical through-the-scope and over-the-scope clips, and variceal ligation. Epinephrine injection reduces or stops bleeding through vasoconstriction and local tamponade. Epinephrine monotherapy is not advised due to its temporizing effect. Bipolar electrocoagulation is a contact thermal treatment that uses a probe to compress the vessel, coagulates directly, and seals the vessel. A meta-analysis of 15 RCTs showed that bipolar electrocoagulation reduces further rebleeding and mortality compared to no treatment.[59] 

Argon plasma coagulation (APC) is a noncontact thermal therapy commonly used for treating vascular ectasias. A probe delivers inert argon gas to the probe tip, where an electrode converts the gas to an ionized form, leading to tissue coagulation. The APC probe is placed 2 to 10 mm from the bleeding lesion. Through-the-scope clips are available with varying diameters of jaw-opening. Through-the-scope clips are best used for closing mucosal defects ≤10mm in size. Data from 4 RCTs on bleeding treatment do not show a significant difference in further bleeding or mortality in patients receiving through-the-scope clips versus thermal contact therapies.[59]

Over-the-scope clips (OTSC) have a circumferential closure design and can close mucosal defects ≤20 mm in size. A systematic review and meta-analysis of 10 studies (4 RCTs), 914 patients, for OTSC versus standard therapy in high-risk nonvariceal UGIB showed OTSC had an overall lower risk of 7-day (RR 0.41) and 30-day (RR 0.46) rebleeding. Mortality or length of stay demonstrated did not have significant differences.[60] Routine use of OTSC over thermal therapies and through-the-scope clips is limited by the potential need for more extensive, multi-channel therapeutic endoscopes, which have less flexibility and are more difficult to navigate within the lumen. 

Endoscopic variceal ligation is performed for esophageal varices and type 1 gastric varices (GOV1) with active bleeding or stigmata of recent bleeding. In cases where immediate hemostasis is not achieved with endoscopic variceal ligation with ongoing hemorrhage, proceeding to balloon tamponade with a Sengstaken-Blakemore or Minnesota tube or esophageal stenting as temporizing measures is advised, followed by interventional radiology consultation for transjugular intrahepatic portosystemic shunt (TIPS). Balloon tamponade is limited to 24 hours, whereas esophageal stents can be in place for up to 1 week. In patients with bleeding type 2 gastric varices (GOV2), isolated gastric varices (IGV), and ectopic varices, endoscopic cyanoacrylate injection or endoscopic coiling can be done depending on available expertise. Alternatively, interventional radiology should be consulted for obliteration therapy or TIPS. 

Hemostatic Powders and Gels

Hemostatic powders and gels are newer endoscopic therapies for nonvariceal bleeding. Hemostatic powders adsorb water, exert mechanical tamponade, and activate the clotting cascade. The powder is temporizing and sloughs off after 24 hours. A meta-analysis of 2 RCTs and 18 observational studies with 1280 patients presenting with UGIB showed a 28% rebleeding rate within 72 hours.[61] A recent noninferiority RCT compared hemostatic powder and standard therapies for primary treatment of UGIB and showed no significant difference in rebleeding within 30 days between the 2 groups. However, this noninferiority trial only had a small number of spurting bleeds (8%).[62] Thus, the widespread use of hemostatic powders, particularly for primary therapy in UGIB, is limited due to concerns of rebleeding and cost.

However, hemostatic powder shows promise as a primary treatment for malignant bleeding. A multicenter RCT of 106 patients randomized to hemostatic powder or standard therapies for active tumor bleeding involving the gastrointestinal tract showed increased initial hemostasis and low 30-day rebleeding in the group receiving hemostatic powder. No difference was seen in 6-month mortality. A potential explanation for these results may be reduced mechanical disruption (eg, coagulation, injection, clip) to the friable tumor tissue with hemostatic powder.[63]

Hemostatic gel matrix is a viscous, transparent peptide that self-assembles after contact with tissue or blood, creating a sealant. Use in GI bleeding comes primarily from endoscopic submucosal dissection data showing the reduced need for coagulation forceps to stop oozing postendoscopic submucosal dissection when the gel is applied to the defect base.[64] A prior observational pilot study of 111 patients presenting with GI bleeding, the majority peptic ulcer or post-polypectomy bleeds, showed 94% hemostasis when used as a primary agent. Spurting bleeding was seen in only 6% of patients, and rebleeding was noted in 16% of cases within 30 days.[65]

Transcatheter Angiography and Embolization 

Transcatheter angiography and embolization with interventional radiology have diagnostic and therapeutic capabilities. In UGIB, this intervention is generally reserved for refractory bleeding despite endoscopic therapy or in cases of life-threatening bleeds where the patient is deemed not an endoscopy candidate. Brisk bleeding of at least 0.5 to 1 mL/min is needed for extravasation to be seen on transcatheter angiography. Complications can include bowel infarction, renal failure, hematomas, thromboses, and dissection.[34]

Transjugular Intrahepatic Portosystemic Shunt

Transjugular intrahepatic portosystemic shunt (TIPS) is a therapeutic procedure for variceal bleeding performed by interventional radiology, involving the placement of a stent between a hepatic vein and a branch of the portal vein to decompress portal pressures. Salvage TIPS is performed in patients who undergo upper endoscopy with evidence of acute variceal hemorrhage but fail to achieve hemostasis. In patients who achieve hemostasis after initial endoscopy, high-risk patients should undergo evaluation for preemptive or "early" TIPS. Data show that a preemptive TIPS performed within 24 to 72 hours after the initial endoscopy may improve bleeding control and survival.[66][67][68][69] High-risk patients are defined as those with Child-Turcotte-Pugh class B scores of >7 with active bleeding on endoscopy and Child-Turcotte-Pugh class C scores of 10 to 13. 

Surgery

Surgery is generally reserved for refractory UGIB despite attempts to stop bleeding with endoscopy and interventional radiology. Localization of the bleeding source is vital before undergoing surgical evaluation.

Postendoscopy Management

Treatment of ulcers with PPI therapy is discussed above. To minimize the recurrence of PUD, check H pylori status, treat if testing is positive, and confirm eradication posttreatment. Clinicians should also counsel patients on NSAID cessation and minimize medication and herbal interactions with anticoagulants that could affect drug levels of warfarin and DOACs. Resuming anticoagulants and antiplatelets requires balancing the patient's thromboembolic risk and recurrent bleeding risk. These discussions are multidisciplinary and may involve a patient's cardiologist, primary care practitioner, neurologist, and vascular surgeon. In patients with a high risk for recurrent bleeds (eg, elderly requiring multiple antiplatelet/anticoagulants, on chronic steroids, chronic NSAIDs, or persistent H pylori gastritis), consider GI prophylaxis with PPI therapy.

In patients with cirrhosis and treatment of variceal bleeding with endoscopic variceal ligation, endoscopy should be repeated after discharge every 2 to 4 weeks for potential endoscopic variceal ligation until variceal obliteration. For patients who do not undergo TIPS after treatment of acute variceal hemorrhage, a nonselective beta blocker should be initiated to decrease portal pressures after vasoactive therapy is discontinued unless the patient has an absolute contraindication to nonselective beta-blocker therapy (eg, asthma, second- and third-degree atrioventricular block without a pacemaker, sick sinus syndrome, and extreme bradycardia <50 bpm).

Differential Diagnosis

Aortoenteric fistula is a rare but life-threatening cause of UGIB that needs to be excluded. Aortoenteric fistulae may present with melena as a sentinel bleed before the development of brisk, hemodynamically significant bleeding. Patients typically have a history of endovascular intervention of the aorta, where graft material erodes through the gastrointestinal lumen. Other inflammatory conditions involving the aorta, such as chronic infectious processes, can also lead to aortoenteric fistula formation and subsequent bleeding.

Hemosuccus pancreaticus and hemobilia are rare causes of UGIB that are difficult to detect and require a high index of clinical suspicion. Hemosuccus pancreaticus results from the rupture of a visceral aneurysm into the main pancreatic duct, while hemobilia is bleeding from the biliary tree. Both result in blood loss through the ampulla of Vater. 

Depending on medical history, occult malignancy of the upper gastrointestinal tract and metastases from other organs should also be excluded.

Prognosis

Endoscopic therapy effectively achieves permanent control of UGIB in 80% to 90% of patients. However, 10% to 20% will rebleed. In a prospective randomized study of patients with rebleeding following initial endoscopic intervention of peptic ulcers, 73% of these patients achieved hemostasis with repeat endoscopy and avoided surgery.[70] The majority of patients with ulcer bleeding die from other diseases rather than from bleeding. In a cohort study of 10,428 patients presenting with peptic ulcer disease, 80% of the deaths were not due to bleeding but to underlying medical comorbidities.[38] Despite therapeutic advancements, patients presenting with acute variceal hemorrhage still have a 6-week mortality risk of 10% to 15%.[71]

Complications

If patients presenting with acute UGIB do not receive adequate initial resuscitation, complications resulting from acute blood loss can occur, including:

• Myocardial infarction
• Respiratory distress
• Syncope
• Shock
• Infection
• Death.

Consultations

Treatment of UGIB often requires an interprofessional team of gastroenterologists, intensivists, interventional radiologists, and surgeons in refractory bleeding. If patients have underlying cardiovascular risks requiring antiplatelet or anticoagulant therapy, the involvement of cardiologists and vascular surgeons is critical. If patients have underlying clotting disorders or cytopenias, transfusion medicine consultants help guide transfusion management. 

Deterrence and Patient Education

H pylori and NSAIDs are the 2 most common causes of peptic ulcers and UGIB. Patients who have upper abdominal pain or discomfort should discuss with their primary care clinicians to check H pylori status (particularly if they live in or come from an endemic part of the world) and refrain from NSAID use. In patients without red-flag symptoms such as anemia, overt GI bleeding, nausea, vomiting, dysphagia, or weight loss, their primary care clinicians may trial a course of proton pump inhibitors to empirically treat common causes of esophagitis, gastroduodenitis, or ulcers. Upper endoscopy is the initial test of choice for UGIB after adequate resuscitation. 

In patients with overt GI blood loss, expedited evaluation through the emergency room is recommended for risk stratification. In patients with cirrhosis, GI bleeding should be taken seriously due to the increased risk of morbidity and mortality. These patients should be advised to undergo expedited evaluation through the emergency room.

Pearls and Other Issues

Management of patients presenting with UGIB should always follow a step-wise approach. The first step is to assess the hemodynamic status and initiate resuscitative efforts as needed, including fluids and blood transfusions. Patients should be risk stratified based on their initial presentation, hemodynamic status, comorbidities, age, and initial laboratory tests. A low Glasgow Blatchford Score of 0 to 1 suggests an unlikely need for endoscopic intervention and a low risk of death. These patients can be discharged from the emergency room with outpatient follow-up. For patients at higher risk or with persistent bleeding, upper endoscopy should be offered within 24 hours, or in cases of suspected variceal bleeding within 12 hours, to help diagnose the source of bleeding and help further guide management. 

If a bleeding ulcer is found to be the culprit lesion, efforts should be taken to prevent the recurrence of bleeding. If the patient is found to have H pylori, eradication should be a target. If NSAIDs were the likely cause of the bleeding, they should be stopped, and if NSAIDs cannot be stopped, concurrent PPI should be used for GI prophylaxis. Patients with established cardiovascular disease who require aspirin or other antiplatelet agents should be on PPI therapy and generally can have antiplatelet therapy reinstituted after bleeding ceases. In patients with high-risk lesions treated during endoscopy and high thromboembolic risk, consider a trial of unfractionated heparin on the day hemostasis is achieved.

For variceal bleeding that is refractory to endoscopic intervention, proceed to balloon tamponade with a Sengstaken-Blakemore or Minnesota tube or consider esophageal stenting, depending on available expertise. Consult interventional radiology for salvage TIPS or variceal obliteration. In variceal bleeding that stops with endoscopic intervention, patients with a higher risk for rebleeding should be considered for preemptive TIPS. 

Enhancing Healthcare Team Outcomes

The diagnosis and management of UGIB involves an interprofessional team that includes staff from the emergency department, a gastroenterologist, and an internist. Other specialists, including interventional radiologists and surgeons, may be needed depending on the response to endoscopic therapy. The initial steps in resuscitation should follow the Advanced Trauma Life Support protocol. Risk assessment should be performed on presentation to the emergency department to identify high-risk patients requiring expedited resuscitation, monitoring, and a higher level of care. Early resuscitation improves patient outcomes. In cases where endoscopic intervention does not achieve hemostasis, clear and timely communication with the primary treatment team is needed to proceed with other therapeutic modalities (eg, angiography with embolization or surgery). In patients taking antiplatelets and anticoagulants, consultation with cardiovascular specialties should be done to weigh the risks of thromboembolic disease and rebleeding. Balancing risks helps patients make informed decisions on whether or not to resume these therapies.