Medical School · Year 3 · Internal Medicine · includes a quiz and discussion video

Seminar 10: Gastrointestinal Bleeding

Internal Medicine Clerkship

Learning Objectives

By the end of this seminar, students will be able to:

1. Differentiate upper from lower gastrointestinal bleeding based on clinical features, anatomical landmarks, and presenting symptoms
2. Describe the major etiologies of upper and lower gastrointestinal bleeding, including their relative prevalence and pathophysiological mechanisms
3. Apply validated risk stratification tools such as the Glasgow-Blatchford Score and Rockall Score to guide clinical decision-making and disposition
4. Outline the principles of initial resuscitation, including fluid management, restrictive transfusion strategies, and pharmacological interventions
5. Describe the role of upper and lower endoscopy in the diagnosis and therapeutic management of gastrointestinal hemorrhage
6. Recognize life-threatening gastrointestinal bleeding scenarios requiring urgent surgical, endoscopic, or interventional radiological intervention

Seminar Outline

Section 1: Overview and Classification

Gastrointestinal bleeding is broadly classified by its anatomical origin relative to the ligament of Treitz, a peritoneal fold that suspends the duodenojejunal junction. Upper gastrointestinal bleeding originates proximal to this landmark and includes hemorrhage from the esophagus, stomach, and duodenum. Lower gastrointestinal bleeding arises distal to the ligament of Treitz and encompasses the jejunum, ileum, colon, rectum, and anus. When standard upper endoscopy and colonoscopy fail to identify a source, the bleeding is termed obscure, which may be further divided into overt bleeding with visible blood loss and occult bleeding detected only by fecal occult blood testing or the presence of iron deficiency anemia.

The clinical presentation of gastrointestinal bleeding provides essential clues to localize the source. Hematemesis, defined as vomiting fresh blood or blood clots, strongly suggests an upper gastrointestinal origin from the esophagus, stomach, or duodenum. Coffee-ground emesis indicates upper gastrointestinal bleeding in which blood has been exposed to gastric acid, converting hemoglobin to acid hematin. Melena, the passage of black, tarry, malodorous stools, originates from the upper gastrointestinal tract in approximately ninety percent of cases but may also arise from the right colon. Hematochezia, the passage of bright red or maroon blood per rectum, typically indicates a lower gastrointestinal source, although massive upper gastrointestinal bleeding with rapid transit can occasionally present in this manner. Maroon-colored stool generally suggests bleeding from the right colon or small bowel.

Severity assessment of gastrointestinal bleeding is critical for guiding the intensity of resuscitation and the urgency of intervention. Mild bleeding is characterized by hemodynamic stability with a normal hemoglobin level. Moderate bleeding manifests with tachycardia and a hemoglobin between eight and ten grams per deciliter. Severe bleeding presents with hypotension, a hemoglobin below eight grams per deciliter, and evidence of active ongoing hemorrhage. Massive gastrointestinal bleeding is defined by the requirement for transfusion of four or more units of packed red blood cells and the presence of hemodynamic shock, necessitating immediate aggressive resuscitation and emergent intervention.

The epidemiology of gastrointestinal bleeding underscores its clinical importance. Upper gastrointestinal bleeding is more common than lower gastrointestinal bleeding, with hospital admission rates of approximately one hundred per one hundred thousand population per year. The in-hospital mortality for upper gastrointestinal bleeding ranges from five to ten percent, reflecting the severity of common etiologies such as variceal hemorrhage and peptic ulcer disease. Lower gastrointestinal bleeding carries a somewhat lower mortality of two to four percent, in part because many causes, such as diverticular bleeding, are self-limited. Despite advances in endoscopic and pharmacological therapy, gastrointestinal bleeding remains a significant cause of morbidity, hospitalization, and healthcare expenditure.

<image>Panel A: Anatomical diagram of the gastrointestinal tract showing the ligament of Treitz dividing upper from lower GI sources, with labeled sites of bleeding including esophagus, stomach, duodenum, small bowel, and colon. Panel B: Clinical photographs comparing hematemesis (fresh red blood vomitus), coffee-ground emesis (dark granular vomitus), melena (black tarry stool), and hematochezia (bright red rectal blood). Panel C: Severity classification chart showing vital signs, hemoglobin levels, and clinical features for mild, moderate, severe, and massive GI bleeding with corresponding color-coded triage categories. Panel D: Bar graph displaying epidemiological data comparing upper versus lower GI bleeding incidence, hospitalization rates, and mortality percentages.</image>

Section 2: Upper Gastrointestinal Bleeding Causes

The most common cause of upper gastrointestinal bleeding is peptic ulcer disease, accounting for thirty to forty percent of cases. Erosive gastritis or gastropathy represents fifteen to twenty-five percent of cases and is often related to nonsteroidal anti-inflammatory drug use or physiologic stress. Esophageal varices, which develop in the setting of portal hypertension from cirrhosis, account for ten to twenty percent of upper gastrointestinal hemorrhage and carry a particularly high mortality. Mallory-Weiss tears, resulting from forceful retching or vomiting that causes a mucosal laceration at the gastroesophageal junction, account for five to ten percent of cases. Erosive esophagitis contributes another five to ten percent, while malignancy accounts for one to two percent and Dieulafoy lesions, which involve a submucosal artery that erodes through the mucosa, account for less than one percent.

Peptic ulcer disease is the leading cause of upper gastrointestinal bleeding and warrants detailed understanding. Duodenal ulcers are more common than gastric ulcers as a cause of hemorrhage. The two predominant etiologies are Helicobacter pylori infection and chronic nonsteroidal anti-inflammatory drug use. The most feared complication occurs when a posterior duodenal ulcer erodes into the gastroduodenal artery, causing massive hemorrhage. Risk factors for ulcer-related bleeding include concurrent use of NSAIDs and anticoagulants, a prior history of peptic ulcer disease, advanced age, and comorbid illness. Endoscopic stigmata that predict rebleeding include a visible vessel in the ulcer base and an adherent clot that resists washing.

Variceal bleeding represents a distinct and particularly dangerous form of upper gastrointestinal hemorrhage. It occurs in the setting of portal hypertension, most commonly due to cirrhosis. Each episode of variceal hemorrhage carries a mortality of fifteen to twenty percent, making it one of the most lethal causes of gastrointestinal bleeding. Clinical stigmata of underlying liver disease, including spider angiomata, palmar erythema, ascites, jaundice, and gynecomastia, provide important diagnostic clues. Varices may be esophageal or gastric in location, and their management requires a specific therapeutic approach distinct from non-variceal bleeding, including vasoactive medications, prophylactic antibiotics, and endoscopic band ligation.

Several less common but clinically important causes of upper gastrointestinal bleeding deserve recognition. Mallory-Weiss tears are strongly suggested by a history of retching or vomiting preceding hematemesis and typically heal spontaneously. Dieulafoy lesions involve an abnormally large submucosal artery, usually in the gastric fundus, that can cause sudden massive hemorrhage without a surrounding ulcer. Cameron lesions are linear erosions found in large hiatal hernias, often causing chronic occult blood loss. Aortoenteric fistulas should be considered in any patient with prior aortic surgery who presents with gastrointestinal bleeding, as this condition represents a surgical emergency with catastrophic hemorrhage if not promptly recognized and treated.

<image>Panel A: Pie chart showing relative prevalence of upper GI bleeding causes including peptic ulcer disease (35%), erosive gastritis (20%), esophageal varices (15%), Mallory-Weiss tear (8%), erosive esophagitis (8%), malignancy (2%), and Dieulafoy lesion (1%). Panel B: Endoscopic photograph of a duodenal ulcer with a visible vessel in the base, with labeled anatomy showing the gastroduodenal artery and posterior duodenal wall. Panel C: Endoscopic image of esophageal varices with blue-tinged, tortuous, dilated submucosal veins in the distal esophagus, alongside clinical stigmata of cirrhosis including spider angiomata and ascites. Panel D: Endoscopic images comparing a Mallory-Weiss tear at the gastroesophageal junction, a Dieulafoy lesion with protruding vessel, Cameron lesions in a hiatal hernia, and a diagram of an aortoenteric fistula.</image>

Section 3: Lower Gastrointestinal Bleeding Causes

Lower gastrointestinal bleeding arises from sources distal to the ligament of Treitz and has a distinct spectrum of etiologies compared to upper gastrointestinal hemorrhage. Diverticulosis is the most common cause, accounting for twenty to forty percent of cases, followed by hemorrhoids at five to twenty percent, ischemic colitis at five to twenty percent, angiodysplasia at five to ten percent, colorectal cancer at five to ten percent, and inflammatory bowel disease at five to ten percent. Post-polypectomy bleeding occurs at a variable rate depending on the size and location of the polyp removed and the technique used. Unlike upper gastrointestinal bleeding, the majority of lower gastrointestinal hemorrhage episodes are self-limited, with spontaneous cessation occurring in approximately eighty percent of patients.

Diverticular bleeding is the most frequent cause of significant lower gastrointestinal hemorrhage. Colonic diverticula develop at points of weakness in the bowel wall where vasa recta penetrate, and bleeding occurs when these arterial vessels are disrupted by the diverticulum. The classic presentation is painless, large-volume hematochezia that often resolves spontaneously. Although diverticula are more common in the left colon, diverticular bleeding more frequently originates from the right colon. The recurrence rate is twenty to forty percent, and patients who experience recurrent episodes may require elective surgical resection. Management is typically conservative with supportive care; however, persistent bleeding may necessitate angiographic embolization or surgical intervention.

Angiodysplasia, also known as arteriovenous malformations, represents an important cause of both acute and chronic lower gastrointestinal bleeding. These lesions consist of dilated, tortuous, thin-walled vascular structures located most commonly in the cecum and right colon. Angiodysplasia has well-established associations with aortic stenosis (Heyde syndrome), chronic kidney disease, and von Willebrand disease. The bleeding pattern is often slow and chronic, leading to iron deficiency anemia rather than acute hemodynamic compromise. Diagnosis may be made by colonoscopy, where the lesions appear as flat, red, fern-like vascular patterns, or by angiography, which can demonstrate a characteristic early-filling vein, vascular tuft, or late-draining vein.

Ischemic colitis is a common cause of acute lower gastrointestinal bleeding, particularly in elderly patients with cardiovascular disease. It results from transient reduction in mesenteric blood flow, most commonly affecting the watershed areas of the colon, particularly the splenic flexure and rectosigmoid junction. The classic presentation is the acute onset of crampy left-sided abdominal pain followed by bloody diarrhea within twenty-four hours. Risk factors include advanced age, atherosclerotic cardiovascular disease, atrial fibrillation, hypotension from any cause, and vasoactive medications. Management is generally supportive with bowel rest, intravenous fluids, and broad-spectrum antibiotics if transmural involvement is suspected. Surgical intervention is reserved for patients who develop peritonitis, perforation, or persistent hemorrhage.

<image>Panel A: Anatomical diagram of the colon with labeled sites of common lower GI bleeding sources including diverticula predominantly in the right colon, angiodysplasia in the cecum, ischemic colitis at the splenic flexure, and colorectal cancer in the rectosigmoid region. Panel B: Colonoscopic image of a diverticulum with active arterial bleeding, accompanied by a cross-sectional diagram showing the vasa recta penetrating the colonic wall at the diverticular neck. Panel C: Colonoscopic image of angiodysplasia showing a flat, red, fern-like vascular lesion in the cecal mucosa with surrounding normal colonic tissue. Panel D: Colonoscopic image of ischemic colitis showing edematous, hemorrhagic, and ulcerated colonic mucosa in a segmental distribution at the splenic flexure watershed area.</image>

Section 4: Initial Assessment

The initial assessment of a patient presenting with gastrointestinal bleeding begins with a focused history that helps localize the source and identify contributing factors. The character of the blood loss is paramount: hematemesis and coffee-ground emesis point to an upper source, while hematochezia typically suggests a lower origin. Associated symptoms including abdominal pain, vomiting, diarrhea, and weight loss provide additional diagnostic information. A thorough medication history is essential, with particular attention to nonsteroidal anti-inflammatory drugs, anticoagulants such as warfarin and direct oral anticoagulants, and antiplatelet agents including aspirin and clopidogrel. Prior episodes of gastrointestinal bleeding, including the identified source and interventions performed, should be documented. A history of liver disease, alcohol use, or portal hypertension raises concern for variceal hemorrhage, while recent endoscopic procedures such as polypectomy or percutaneous endoscopic gastrostomy placement may identify iatrogenic causes.

The physical examination in gastrointestinal bleeding serves to assess hemodynamic stability, identify the likely source, and detect underlying conditions. Vital signs are critical: tachycardia, hypotension, and orthostatic changes indicate significant volume depletion and hemodynamic compromise. The abdominal examination should evaluate for tenderness, distension, masses, and signs of peritonitis that might suggest perforation or ischemia. A digital rectal examination is mandatory to assess stool color (black, maroon, or bright red), to identify hemorrhoids or rectal masses, and to obtain a sample for fecal occult blood testing. Stigmata of chronic liver disease, including spider angiomata, palmar erythema, caput medusae, ascites, jaundice, and gynecomastia, strongly suggest variceal bleeding as the cause. In cases where the distinction between upper and lower gastrointestinal bleeding is unclear, nasogastric aspirate may be helpful, although a clear aspirate does not exclude an upper source.

Laboratory evaluation provides essential data for assessing severity, guiding resuscitation, and preparing for potential transfusion. A complete blood count reveals the hemoglobin level, which may initially be normal in acute bleeding before equilibration with extravascular fluid. The platelet count is important for assessing coagulation capacity and detecting thrombocytopenia associated with liver disease or disseminated intravascular coagulation. A basic metabolic panel evaluates renal function and electrolytes, and coagulation studies including the prothrombin time, international normalized ratio, and partial thromboplastin time identify coagulopathies that may require correction. A type and screen should be sent early to prepare for potential transfusion. Liver function tests help assess for underlying hepatic disease that may contribute to coagulopathy and suggest variceal bleeding.

The blood urea nitrogen to creatinine ratio is a useful adjunct in distinguishing upper from lower gastrointestinal bleeding. A BUN-to-creatinine ratio greater than thirty-six to one suggests an upper gastrointestinal source, as digested blood in the upper tract results in significant absorption of urea from the small intestine, disproportionately elevating the BUN relative to creatinine. A ratio less than thirty-six to one is less specific and does not reliably exclude an upper source. This mechanism reflects the fact that blood acts as a high-protein meal when it enters the small bowel, where bacterial urease converts the protein-derived urea precursors into ammonia, which is then absorbed and converted to urea in the liver. While this laboratory finding is a helpful screening tool, it must be interpreted in the clinical context alongside the presenting symptoms, physical examination findings, and other diagnostic data.

<image>Panel A: Flowchart diagram showing the systematic approach to history-taking in GI bleeding, with branching pathways for character of blood loss, associated symptoms, medication history, prior bleeding episodes, and liver disease risk factors. Panel B: Illustration of a focused physical examination for GI bleeding showing vital sign assessment, abdominal palpation, digital rectal examination with stool color assessment, and inspection for stigmata of liver disease including spider angiomata and ascites. Panel C: Laboratory panel display showing normal and abnormal ranges for CBC, BMP, coagulation studies, type and screen, and liver function tests in the context of GI bleeding evaluation. Panel D: Diagram illustrating the mechanism of elevated BUN-to-creatinine ratio in upper GI bleeding, showing digested blood in the small intestine with urea absorption and hepatic conversion, with a comparison bar chart of ratios in upper versus lower GI bleeding.</image>

Section 5: Risk Stratification

The Glasgow-Blatchford Score is a validated pre-endoscopic risk stratification tool used to identify patients at low risk who may be safely managed as outpatients and those at high risk who require urgent intervention. The score incorporates blood urea nitrogen level (zero to six points), hemoglobin level (zero to six points), systolic blood pressure (zero to three points), and additional markers including heart rate, presence of melena, syncope, hepatic disease, and heart failure (zero to two points each), yielding a total possible score of zero to twenty-three. A Glasgow-Blatchford Score of zero identifies patients at very low risk who may be candidates for outpatient endoscopy without hospitalization, as they have an extremely low probability of requiring intervention. Scores of six or higher indicate high-risk patients who are likely to require endoscopic intervention, blood transfusion, or surgical management and should be admitted and undergo early endoscopy.

The Rockall Score is a complementary risk assessment tool that incorporates both pre-endoscopic clinical features and post-endoscopic findings to predict rebleeding and mortality. The pre-endoscopic components include age (zero to two points), shock status based on heart rate and blood pressure (zero to two points), and comorbid conditions including heart failure, ischemic heart disease, renal failure, liver failure, and metastatic malignancy (zero to three points). After endoscopy, additional points are assigned based on the diagnosis (zero to two points) and the presence of stigmata of recent hemorrhage such as active bleeding, visible vessel, or adherent clot (zero to two points). The complete Rockall Score helps predict the risk of rebleeding and mortality, and it is particularly valuable for post-endoscopic risk assessment and guiding the intensity of post-procedural monitoring.

Several high-risk features in gastrointestinal bleeding warrant identification because they predict adverse outcomes and guide the urgency of intervention. Hemodynamic instability, manifested by persistent tachycardia, hypotension, or orthostatic changes despite initial fluid resuscitation, indicates the need for intensive care unit admission and urgent blood product transfusion. Ongoing active bleeding, evidenced by continued hematemesis, hematochezia, or inability to stabilize hemoglobin despite transfusion, requires emergent intervention. Concurrent anticoagulation increases the risk of both initial bleeding severity and rebleeding, and decisions regarding reversal must balance bleeding risk against thromboembolic risk. Significant comorbidities, particularly cardiovascular, hepatic, and renal disease, independently increase mortality and should factor into disposition decisions. Endoscopic stigmata of high-risk bleeding, including a spurting vessel or a non-bleeding visible vessel, predict high rebleeding rates and mandate endoscopic therapy.

Disposition decisions are guided by the integration of clinical assessment and validated risk stratification scores. Patients with a Glasgow-Blatchford Score of zero are considered very low risk and may be discharged for outpatient endoscopy within days, provided they have reliable follow-up and no social barriers to safe discharge. Low-risk patients should be admitted to a general medical ward for monitoring, intravenous fluid resuscitation, and endoscopy within twenty-four hours. High-risk patients, defined by hemodynamic instability, active ongoing bleeding, significant comorbidities, or high-risk scoring, should be admitted to an intensive care unit with continuous monitoring and undergo urgent endoscopy after appropriate resuscitation. The decision to admit, the level of care, and the timing of endoscopy should be individualized based on the totality of clinical data, and clinical judgment should always complement standardized scoring tools.

<image>Panel A: Detailed scoring table for the Glasgow-Blatchford Score showing point assignments for BUN, hemoglobin, systolic blood pressure, heart rate, melena, syncope, hepatic disease, and heart failure, with risk category thresholds highlighted. Panel B: Diagram of the complete Rockall Score with pre-endoscopic and post-endoscopic components, showing point assignments for age, shock, comorbidities, endoscopic diagnosis, and stigmata of hemorrhage. Panel C: Infographic highlighting high-risk features in GI bleeding including hemodynamic instability, ongoing hemorrhage, anticoagulant use, major comorbidities, and high-risk endoscopic stigmata, each with associated outcome data. Panel D: Flowchart showing disposition pathways based on risk stratification from outpatient management for very low risk, ward admission for low risk, and ICU admission with urgent endoscopy for high-risk patients.</image>

Section 6: Initial Management

Initial resuscitation in gastrointestinal bleeding follows a systematic approach prioritizing hemodynamic stabilization and preparation for definitive intervention. Two large-bore (eighteen-gauge or larger) intravenous catheters should be placed immediately to allow rapid administration of fluids and blood products. Fluid resuscitation begins with crystalloid solution, typically normal saline or lactated Ringer solution, to restore intravascular volume. Blood transfusion is initiated when hemoglobin falls below the transfusion threshold, with a goal hemoglobin of seven to eight grams per deciliter for most patients and a higher target for patients with coronary artery disease. Coagulopathy should be corrected with vitamin K, prothrombin complex concentrate, fresh frozen plasma, or platelet transfusion as indicated. Patients with hemodynamic instability or ongoing active bleeding should be admitted to an intensive care unit for continuous monitoring and aggressive resuscitation.

Pharmacological management plays a critical role in the initial treatment of gastrointestinal bleeding. Intravenous proton pump inhibitor therapy is initiated with an eighty-milligram bolus followed by a continuous infusion of eight milligrams per hour, which raises gastric pH to promote clot stability in peptic ulcer disease. When variceal bleeding is suspected based on clinical features or a history of liver disease, octreotide should be started with a fifty-microgram intravenous bolus followed by a continuous infusion of fifty micrograms per hour to reduce splanchnic blood flow and portal pressure. Prophylactic antibiotics, specifically ceftriaxone one gram intravenously daily, are recommended for all cirrhotic patients with gastrointestinal bleeding to reduce the risk of bacterial infections, which are associated with increased rebleeding and mortality. Anticoagulants and antiplatelet agents should be held unless there is a critical indication for their continuation, and intravenous erythromycin may be administered as a prokinetic agent to clear blood from the stomach prior to endoscopy.

The role of nasogastric tube placement in gastrointestinal bleeding has evolved and remains somewhat controversial. The primary purpose of nasogastric aspiration is to assess for the presence of blood or coffee-ground material in the stomach, which would confirm an upper gastrointestinal source. The finding of bloody or coffee-ground aspirate confirms upper gastrointestinal bleeding and should prompt early upper endoscopy. However, a clear nasogastric aspirate does not reliably rule out an upper source, as twelve to fifteen percent of patients with upper gastrointestinal bleeding have a clear aspirate due to a duodenal source with a competent pylorus or intermittent bleeding. Nasogastric lavage to clear the stomach of blood has not been shown to improve clinical outcomes, and the tube is relatively contraindicated in patients with suspected esophageal varices due to the risk of precipitating bleeding, though this remains a relative rather than absolute contraindication.

A restrictive transfusion strategy has become the standard of care for most patients with gastrointestinal bleeding, supported by evidence demonstrating improved outcomes compared to liberal transfusion. The recommended transfusion threshold is a hemoglobin below seven grams per deciliter for the majority of patients, with transfusion targeting a hemoglobin of seven to eight grams per deciliter. Patients with coronary artery disease or acute coronary syndrome may benefit from a slightly higher threshold of eight grams per deciliter to ensure adequate myocardial oxygen delivery. Studies have demonstrated that a restrictive strategy is associated with lower mortality, reduced rebleeding rates, and fewer complications compared to liberal transfusion, likely because over-transfusion in variceal bleeding can increase portal pressure and perpetuate hemorrhage. The primary exception to restrictive transfusion is massive hemorrhage with hemodynamic instability, where aggressive resuscitation with blood products is necessary to maintain organ perfusion and prevent cardiovascular collapse.

<image>Panel A: Step-by-step illustration of initial resuscitation showing placement of two large-bore IV lines, crystalloid fluid bags, blood product transfusion setup, and ICU monitoring equipment with vital sign displays. Panel B: Pharmacological management diagram showing intravenous PPI infusion pump, octreotide syringe for variceal bleeding, ceftriaxone antibiotic vial for cirrhotic patients, and erythromycin prokinetic agent, each with dosing annotations. Panel C: Diagram illustrating nasogastric tube placement and aspiration findings, comparing bloody aspirate (confirms upper GI source), coffee-ground aspirate (confirms upper GI source), and clear aspirate (does not exclude upper source), with percentage annotations. Panel D: Graph comparing outcomes of restrictive versus liberal transfusion strategies in GI bleeding, showing mortality rates, rebleeding rates, and transfusion-related complications for hemoglobin thresholds of 7 versus 9 g/dL.</image>

Section 7: Endoscopic Management

The timing of endoscopy in gastrointestinal bleeding is determined by the clinical scenario and risk stratification. For hemodynamically stable patients with low-risk features, endoscopy should be performed within twenty-four hours of presentation, as earlier timing in this group has not demonstrated improved outcomes. High-risk patients with non-variceal upper gastrointestinal bleeding should also undergo endoscopy within twenty-four hours, which is considered early endoscopy and allows timely diagnosis and therapeutic intervention. Suspected variceal hemorrhage warrants more urgent endoscopy within twelve hours of presentation, given the high mortality and the need for specific endoscopic therapy such as band ligation. Patients who are hemodynamically unstable should first undergo adequate resuscitation, including airway protection if needed, before proceeding to endoscopy, as performing endoscopy in an unstable patient increases procedural risk without clear benefit.

Endoscopic findings in peptic ulcer bleeding are classified by the Forrest classification, which predicts the risk of rebleeding and guides the need for endoscopic therapy. A spurting arterial vessel, the highest-risk stigma, carries a rebleeding rate of approximately ninety percent without intervention. A non-bleeding visible vessel, appearing as a pigmented protuberance in the ulcer base, has a rebleeding rate of approximately fifty percent. An adherent clot that resists vigorous washing carries a thirty percent rebleeding risk, while a flat pigmented spot in the ulcer base carries only a ten percent risk. A clean-based ulcer has the lowest rebleeding risk at approximately five percent and does not require endoscopic therapy. Understanding these stigmata is essential for determining which patients require endoscopic treatment and which can be managed conservatively.

Multiple endoscopic therapeutic modalities are available for the treatment of gastrointestinal bleeding. Epinephrine injection, typically using a one-to-ten-thousand dilution in one to two milliliter aliquots around the bleeding site, provides initial hemostasis through local vasoconstriction and a tamponade effect, but should not be used as monotherapy due to high rebleeding rates. Thermal coagulation, including heater probe and bipolar electrocoagulation, achieves hemostasis through tissue coaptation and sealing of the bleeding vessel. Mechanical hemostasis with through-the-scope hemostatic clips directly approximates the edges of the bleeding vessel. Band ligation is the preferred endoscopic treatment for esophageal varices. Hemostatic spray, such as TC-325 powder, provides a topical hemostatic effect and is useful as an adjunctive measure or as a bridge in diffuse or difficult-to-access bleeding. Combination therapy, using epinephrine injection followed by a second modality such as thermal coagulation or clips, has been shown to be more effective than any single modality alone.

Post-endoscopy management is guided by the endoscopic findings and the risk of rebleeding. Patients with low-risk stigmata, such as a clean-based ulcer or a flat pigmented spot, can resume an oral diet and be transitioned to oral proton pump inhibitor therapy, with early discharge considered. Patients with high-risk stigmata, including active bleeding, visible vessels, or adherent clots treated endoscopically, should receive intravenous proton pump inhibitor infusion for seventy-two hours and be monitored in an intensive care unit setting for signs of rebleeding. Following endoscopic variceal band ligation, patients should continue octreotide infusion for three to five days and complete a course of prophylactic antibiotics. In the event of rebleeding, options include repeat endoscopy with therapeutic intervention, interventional radiology with angiographic embolization, or surgical consultation for definitive management.

<image>Panel A: Timeline diagram showing recommended endoscopy timing for different clinical scenarios: stable low-risk patients within 24 hours, high-risk non-variceal within 24 hours, suspected variceal within 12 hours, and unstable patients after resuscitation. Panel B: Endoscopic photographs illustrating the Forrest classification of ulcer bleeding stigmata including spurting vessel (Ia), oozing vessel (Ib), non-bleeding visible vessel (IIa), adherent clot (IIb), flat pigmented spot (IIc), and clean base (III), with corresponding rebleeding percentages. Panel C: Illustration of endoscopic therapeutic modalities including epinephrine injection, bipolar thermal coagulation probe, through-the-scope hemostatic clip application, variceal band ligation, and hemostatic spray deployment. Panel D: Post-endoscopy management algorithm showing pathways for low-risk stigmata (oral PPI, early discharge) versus high-risk stigmata (IV PPI infusion, ICU monitoring, 72-hour observation).</image>

Section 8: Variceal Bleeding Management

The initial management of variceal hemorrhage requires a coordinated, multimodal approach that differs significantly from non-variceal bleeding. Resuscitation follows a restrictive transfusion strategy with a hemoglobin target of seven to eight grams per deciliter, as over-transfusion can increase portal pressure and exacerbate variceal bleeding. Vasoactive drug therapy should be initiated immediately upon suspicion of variceal hemorrhage, with octreotide administered as a fifty-microgram intravenous bolus followed by a continuous infusion of fifty micrograms per hour for three to five days. Prophylactic antibiotics, specifically ceftriaxone one gram intravenously daily for seven days, are administered to all cirrhotic patients with gastrointestinal bleeding and have been shown to reduce mortality, bacterial infections, and rebleeding. Airway protection through endotracheal intubation should be strongly considered in patients with massive hematemesis or altered mental status from hepatic encephalopathy to prevent aspiration during endoscopy.

Endoscopic therapy is the cornerstone of definitive management for variceal bleeding. Endoscopic variceal band ligation is the preferred first-line treatment for esophageal varices, as it has been shown to have superior efficacy and fewer complications compared to injection sclerotherapy. Sclerotherapy, which involves injection of a sclerosant agent directly into the varix, is reserved for situations where band ligation is not technically feasible, such as active massive hemorrhage obscuring the endoscopic view. Cyanoacrylate glue injection is the preferred treatment for gastric varices, which are not amenable to band ligation due to their submucosal depth and larger caliber. Endoscopy should be performed within twelve hours of presentation to maximize the likelihood of successful hemostasis and improve clinical outcomes.

Rescue therapy is required when endoscopic treatment fails to achieve hemostasis or when early rebleeding occurs despite endoscopic therapy. Balloon tamponade using a Sengstaken-Blakemore tube or Minnesota tube provides temporary mechanical compression of esophageal and gastric varices and serves as a bridge to definitive therapy, although it carries significant risks including esophageal necrosis and aspiration. Transjugular intrahepatic portosystemic shunt placement is the preferred rescue intervention for refractory variceal bleeding, as it reduces portal pressure by creating a low-resistance channel between the hepatic vein and portal vein within the liver. Surgical portosystemic shunt creation or devascularization procedures are reserved for cases where TIPS is unavailable or technically not feasible. Current guidelines recommend considering TIPS after a maximum of two failed endoscopic treatment attempts, and early TIPS within seventy-two hours may be considered for high-risk patients with Child-Pugh class C cirrhosis or those with persistent active bleeding.

Secondary prophylaxis following an episode of variceal hemorrhage is essential to prevent rebleeding, which occurs in approximately sixty percent of patients within one to two years without prophylaxis. Repeat endoscopic band ligation sessions are performed every two to four weeks until complete variceal eradication is achieved, typically requiring three to five sessions. Non-selective beta-blockers, specifically propranolol or nadolol, are initiated after the acute bleeding episode resolves to reduce portal pressure by decreasing cardiac output (beta-1 blockade) and causing splanchnic vasoconstriction (beta-2 blockade). The combination of endoscopic band ligation with non-selective beta-blocker therapy provides superior protection against rebleeding compared to either modality alone and represents the current standard of care for secondary prophylaxis. Surveillance endoscopy should continue after variceal eradication to monitor for recurrence.

<image>Panel A: Comprehensive treatment algorithm for acute variceal bleeding showing simultaneous initiation of resuscitation (restrictive transfusion), octreotide infusion, antibiotics (ceftriaxone), airway assessment, and urgent endoscopy within 12 hours. Panel B: Endoscopic image sequence showing the steps of variceal band ligation: identification of varix, suction of varix into ligator cap, deployment of rubber band at base, and post-ligation appearance with banded varix. Panel C: Diagram of rescue therapies including balloon tamponade with Sengstaken-Blakemore tube placement, TIPS procedure showing transjugular catheter creating an intrahepatic shunt between portal and hepatic veins, and post-TIPS portal pressure reduction. Panel D: Secondary prophylaxis timeline showing repeat band ligation sessions every 2-4 weeks until eradication, concurrent initiation of non-selective beta-blocker therapy, and surveillance endoscopy schedule.</image>

Section 9: Lower Gastrointestinal Bleeding Management

The initial approach to lower gastrointestinal bleeding follows the same resuscitation principles as upper gastrointestinal hemorrhage, with prompt establishment of intravenous access, fluid resuscitation, and hemodynamic stabilization. It is important to exclude an upper gastrointestinal source in patients with significant lower gastrointestinal bleeding, as approximately ten to fifteen percent of patients presenting with hematochezia have an upper gastrointestinal source, particularly when bleeding is massive and transit is rapid. An anorectal examination should be performed early to identify hemorrhoids, fissures, or rectal masses as potential sources. Colonoscopy is the primary diagnostic and therapeutic modality for lower gastrointestinal bleeding, offering direct visualization of the colonic mucosa, the ability to identify the bleeding source, and the option for endoscopic therapeutic intervention.

The timing of colonoscopy in lower gastrointestinal bleeding depends on the clinical severity and hemodynamic status. For hemodynamically stable patients, colonoscopy should be performed within twenty-four hours of presentation after adequate bowel preparation, which typically requires a rapid purge with four to six liters of polyethylene glycol solution administered over three to four hours via nasogastric tube or oral ingestion. Severe bleeding requires initial hemodynamic stabilization with resuscitation before bowel preparation and colonoscopy can be safely performed. In patients with very severe or massive lower gastrointestinal bleeding who cannot be adequately stabilized, computed tomographic angiography or conventional angiography may be required as the initial diagnostic modality to localize the bleeding source before endoscopic or interventional management. The diagnostic yield of colonoscopy is significantly improved with adequate bowel preparation, even in the setting of acute bleeding.

Alternative diagnostic modalities are employed when colonoscopy is not feasible, fails to identify a source, or when bleeding is too massive for endoscopic evaluation. Computed tomographic angiography is a rapidly available, non-invasive study that can localize active bleeding and identify underlying pathology such as tumors, diverticular disease, or vascular malformations, making it useful when colonoscopy is not immediately feasible. Conventional mesenteric angiography provides both diagnostic and therapeutic capability, allowing identification and embolization of actively bleeding vessels, but requires an active bleeding rate of at least 0.5 to 1.0 milliliter per minute for detection. Tagged red blood cell scanning using technetium-99m-labeled erythrocytes can detect bleeding at a lower rate of 0.1 to 0.4 milliliter per minute and is useful for intermittent bleeding, providing localization information to guide subsequent angiography or surgery. Capsule endoscopy is reserved for patients with obscure gastrointestinal bleeding when standard endoscopy and colonoscopy have failed to identify a source, allowing visualization of the small bowel.

Interventional radiology plays an important role in the management of lower gastrointestinal bleeding when endoscopic therapy is not feasible or has failed. Angiographic embolization involves selective catheterization of the mesenteric artery supplying the bleeding site, followed by deployment of embolic agents such as microcoils or polyvinyl alcohol particles to achieve hemostasis. Successful embolization requires the demonstration of active bleeding on angiography, with an active bleeding rate of at least 0.5 to 1.0 milliliter per minute. The primary risk of embolization is bowel ischemia, which occurs in approximately five to ten percent of cases due to interruption of the arterial blood supply to the affected segment. Embolization is particularly valuable as an alternative to surgery in patients who are at high surgical risk due to advanced age, multiple comorbidities, or hemodynamic instability, and it may serve as a bridge to definitive surgical management.

<image>Panel A: Flowchart showing the initial approach to lower GI bleeding including resuscitation, exclusion of upper source (with nasogastric aspirate or upper endoscopy), anorectal examination, and colonoscopy as the primary diagnostic modality. Panel B: Step-by-step diagram of colonoscopy timing and preparation, showing rapid bowel preparation with polyethylene glycol, hemodynamic stabilization, and colonoscopy within 24 hours with endoscopic therapy options. Panel C: Comparison images of alternative diagnostic modalities including CT angiography showing active contrast extravasation in the right colon, conventional mesenteric angiography with contrast blush, tagged red blood cell scan with focal radiotracer uptake, and capsule endoscopy image of a small bowel lesion. Panel D: Interventional radiology diagram showing selective mesenteric artery catheterization, angiographic identification of bleeding site with contrast extravasation, and coil embolization of the bleeding vessel with post-embolization hemostasis.</image>

Section 10: Special Situations

Obscure gastrointestinal bleeding is defined as bleeding that persists or recurs without an identified source after standard upper endoscopy and colonoscopy. The most common causes are small bowel lesions, including angiodysplasia, small bowel tumors (adenocarcinoma, carcinoid, gastrointestinal stromal tumors, and lymphoma), Crohn disease, and Meckel diverticulum. Capsule endoscopy is the initial diagnostic modality of choice, as it allows non-invasive visualization of the entire small bowel with a diagnostic yield of approximately sixty to seventy percent. Deep enteroscopy, including push enteroscopy, single-balloon enteroscopy, and double-balloon enteroscopy, provides both diagnostic and therapeutic capability for small bowel lesions identified by capsule endoscopy. Computed tomographic enterography is an alternative modality that provides cross-sectional imaging of the small bowel wall and mesentery and is particularly useful for detecting tumors and structural abnormalities.

Management of gastrointestinal bleeding in anticoagulated patients requires careful balancing of the bleeding risk against the thrombotic risk. An initial assessment of bleeding severity guides the decision regarding anticoagulation reversal: life-threatening hemorrhage necessitates immediate reversal using appropriate agents such as vitamin K, prothrombin complex concentrate for warfarin, idarucizumab for dabigatran, or andexanet alfa for factor Xa inhibitors. Mild bleeding may be observed without reversal if the patient has a high thrombotic risk, such as a mechanical heart valve or recent venous thromboembolism. The timing of anticoagulation resumption after hemostasis is achieved must balance the risk of rebleeding against the risk of thromboembolic events, with most guidelines recommending resumption within seven to fourteen days if there are no high-risk features for rebleeding. This decision should involve multidisciplinary discussion among gastroenterology, hematology, and the referring service.

Rebleeding following initial hemostasis occurs in ten to twenty percent of patients with peptic ulcer bleeding and is associated with increased morbidity and mortality. High-risk endoscopic stigmata, including a visible vessel or active bleeding at the index endoscopy, are the strongest predictors of rebleeding. When rebleeding occurs, a repeat endoscopy with therapeutic intervention is the recommended first-line approach, and successful hemostasis is achieved in approximately seventy to seventy-five percent of cases with a second endoscopic attempt. If endoscopic therapy fails to achieve durable hemostasis, the next step is interventional radiologic embolization, which can target the bleeding vessel under angiographic guidance. If both endoscopic and interventional radiologic approaches fail, surgical intervention is required, with the specific procedure determined by the location and nature of the bleeding source.

Surgical intervention for gastrointestinal bleeding is reserved for patients who have failed endoscopic and interventional radiologic management or who present with indications that require operative treatment. Specific indications include failure of two or more endoscopic treatment attempts, failure of interventional radiologic embolization, massive transfusion requirements exceeding six to eight units of packed red blood cells within twenty-four hours, persistent hemodynamic instability despite aggressive resuscitation, and gastrointestinal perforation as a surgical emergency. The choice of surgical procedure depends on the identified source and may include oversewing of a bleeding ulcer, partial gastrectomy, segmental colectomy, or subtotal colectomy when the bleeding source cannot be localized. Preoperative localization of the bleeding source is highly desirable to guide a targeted surgical approach and minimize the extent of resection, as blind subtotal colectomy carries significantly higher morbidity and mortality.

<image>Panel A: Diagnostic algorithm for obscure GI bleeding showing initial negative EGD and colonoscopy, followed by capsule endoscopy, deep enteroscopy (single-balloon and double-balloon), and CT enterography, with common small bowel findings including angiodysplasia, tumors, and Meckel diverticulum. Panel B: Decision-making flowchart for anticoagulated patients with GI bleeding, showing assessment of bleeding severity, reversal agents for different anticoagulants (vitamin K and PCC for warfarin, idarucizumab for dabigatran, andexanet alfa for factor Xa inhibitors), and timelines for anticoagulation resumption. Panel C: Stepwise management of rebleeding showing repeat endoscopy as first-line, interventional radiologic embolization as second-line, and surgical intervention as third-line, with success rates annotated at each step. Panel D: Surgical indications infographic showing failed endoscopy and IR, massive transfusion requirements, persistent hemodynamic instability, and perforation, with corresponding surgical procedures including ulcer oversewing, partial gastrectomy, and segmental colectomy.</image>

Summary

• Upper GI bleeding originates proximal to the ligament of Treitz and presents with hematemesis or melena; lower GI bleeding originates distally and presents with hematochezia
• Common upper GI causes include peptic ulcer disease, esophageal varices, erosive gastropathy, and Mallory-Weiss tears
• Common lower GI causes include diverticulosis, hemorrhoids, angiodysplasia, ischemic colitis, and colorectal cancer
• Initial resuscitation requires large-bore IV access, crystalloid fluids, and a restrictive transfusion strategy targeting hemoglobin of 7-8 g/dL
• Upper GI bleeding management includes IV PPI infusion; suspected variceal bleeding requires octreotide and prophylactic antibiotics
• Risk stratification with Glasgow-Blatchford Score (pre-endoscopy) and Rockall Score (post-endoscopy) guides disposition and intensity of care
• Endoscopy timing: within 24 hours for most cases, within 12 hours for suspected variceal hemorrhage
• Endoscopic therapy combines injection with thermal coagulation or clips for ulcer bleeding; band ligation is preferred for esophageal varices
• Lower GI bleeding is primarily managed with colonoscopy after bowel preparation; angiography is used for massive or ongoing bleeding
• Surgical intervention is reserved for failure of endoscopic and interventional radiologic therapy, massive transfusion requirements, or perforation

Key Terms

This content is subject to the MIT License. © 2024–2026 Hibbert School of Medicine.

Read this lecture as Markdown