ACP medicine, 3rd Edition


Gastrointestinal Bleeding

Elizabeth Rajan MD1

Associate Professor of Medicine

1Mayo Foundation

The author has no commercial relationships with manufacturers of products or providers of services discussed in this chapter.

February 2007

Gastrointestinal (GI) bleeding occurs commonly, has many causes, and ranges from trivial to torrential and life-threatening in severity. Practical classification of GI bleeding—on the basis of presentation, site, and mechanism of the bleed—aids the clinician in selecting an appropriate management algorithm.

GI bleeding is defined as overt when visible red or altered blood is noted in emesis or feces. Overt bleeding is considered major when accompanied by hemodynamic instability and considered minor when not. Occult bleeding is visibly inapparent but is detected directly by stool testing or suggested indirectly by iron deficiency anemia.

GI bleeding occurs when a pathologic process such as ulceration, inflammation, or neoplasia causes erosion of a blood vessel. The size of the eroded artery is an important determinant of the rate of bleeding, the risk of rebleeding, and the clinical outcome. The blood flow and, thus, the rate of blood loss vary directly with the diameter of the vessel; small changes in vessel diameter have dramatic effects on bleeding rates. Most GI bleeds result from erosion of pathologic processes into small vessels and are trivial and self-limited. Erosion into larger vessels can produce lesions that exceed the capacity of normal hemostasis and result in overt major bleeding. A study of the external diameter of arteries in gastric ulcers that bled recurrently showed a range of 0.1 to 1.8 mm, with a mean of 0.7 mm.1 Deep, large ulcers are more likely to erode into large blood vessels. Recurrent or persistent bleeding may result from inadequate vasoconstriction because of large vessel size or inflammatory necrosis of the vessel wall, from pseudoaneurysm formation at the bleeding site, or from systemic coagulopathies.

Overt Bleeding


The reported incidence of GI bleeding varies widely, in part because of varying definitions. Overt minor bleeding, such as from anorectal hemorrhoids, is exceedingly common. Most major bleeding arises from upper GI lesions; the estimated annual incidence ranges from 40 to 150 episodes per 100,000 population.2,3 Mortality from upper GI bleeding has remained at 8% to 10% over the past 50 years.4,5 The fact that, over this period, mortality has failed to decrease substantially despite advances in patient care and technology may reflect the increasing number of elderly patients with complicated comorbidities. Cases in individuals older than 60 years account for 35% to 45% of all cases of acute upper GI bleeding but nearly all of the associated mortality.6 Lower GI sources account for an estimated 15% to 20% of all major GI bleeds. The incidence of lower GI bleeds increases with age.7,8


The causes of GI bleeding are protean [see Table 1]. The most common etiologies are briefly elaborated in this subsection.

Table 1 Major Causes of Gastrointestinal Bleeding

Inflammatory    Peptic ulcer disease
   Esophagitis or esophageal ulceration
   Diaphragmatic hernia (Cameron erosions)
   Diverticular disease
   Inflammatory bowel disease
   Meckel diverticulum
Cancers and neoplasms
   Primary lesion at any site
   Metastatic deposits at any site
   Large polyps
   Gastrointestinal stromal tumors
Vascular anomalies
   Gastroesophageal varices
   Dieulafoy lesion
   Watermelon stomach
   Radiation proctopathy
   Nonsteroidal anti-inflammatory drugs
   Mallory-Weiss tear

Upper GI Bleeding

Upper GI bleeding is arbitrarily defined as hemorrhage from a source proximal to the ligament of Treitz (i.e., the esophagus, stomach, or duodenum). Hematemesis essentially always reflects upper GI bleeding, and stools may range from black (melena) to bright red (hematochezia), depending on rates of bleeding and intestinal transit.

Peptic ulcer disease

The most common cause of upper GI bleeding is peptic ulcer disease (PUD), accounting for 60% of cases found at emergency endoscopy.9About 50% of patients have a clean-based ulcer with a low probability of rebleeding; only pharmacologic intervention is required in such cases.10 Adherent clots, visible vessels, or active bleeding [see Figure 1] portend progressively less favorable outcomes unless endoscopic or surgical treatment is applied. The two most important risk factors for bleeding in cases of PUD are the use of nonsteroidal anti-inflammatory drugs (NSAIDs) and Helicobacter pylori infection; heavy alcohol ingestion and smoking are also associated with increased risk.11,12,13


Figure 1. High-risk posterior duodenal bulb ulcer with nonbleeding visible vessel (arrow).


Aspirin and other NSAIDs are responsible for most cases of drug-induced GI bleeding. In the United States, more than 30 billion NSAID tablets are consumed annually. Except for sodium salicylate, all NSAIDs can cause bleeding. Acetaminophen is not associated with GI bleeding.

The elderly are especially susceptible to NSAID-induced GI bleeding.14 NSAIDs may cause bleeding at any level of the GI tract, but they most commonly do so in the stomach or duodenum. Although the bleeding risk increases in proportion to NSAID dose, any amount (including low-dose aspirin taken for prophylaxis against cardiovascular events) may cause bleeding. Use of selective serotonin reuptake inhibitors (SSRIs) is associated with a higher risk of upper GI bleeding, especially in patients who are also taking NSAIDs or low-dose aspirin.15Anticoagulants and nonaspirin antiplatelet drugs do not cause GI bleeding per se, but they can unmask or aggravate hemorrhage from preexisting lesions.

Variceal bleeding

Gastroesophageal variceal bleeding accounts for 10% to 30% of all upper GI hemorrhage. Patients present with overt major bleeding that is sudden in onset. Variceal bleeding is distinctive, with large-volume hematemesis of bright-red blood or clots and associated severe hemodynamic instability [see Figure 2]. Because of the cathartic nature of blood, patients may also present with hematochezia. A prospective review found that 75% of bleeding varices were esophageal and 25% were gastric.16 The most common site of bleeding is the distal 5 cm of the esophagus, because of relatively greater variceal distention and thinner supporting tissue surrounding the veins in this region, compared with the upper and the middle esophagus. Varices are present in 40% to 60% of patients with cirrhosis, and hemorrhage occurs in 25% to 35% of them.17,18,19 Approximately one third of first variceal bleeds are fatal.19 Physicians should bear in mind that in up to half of patients with portal hypertension who experience bleeding, the bleeding has a nonvariceal cause.16


Figure 2. High-risk esophageal varices with red wale marking.

Mallory-Weiss tear

Mallory-Weiss tear is a longitudinal mucosal laceration at the gastroesophageal junction or gastric cardia caused by forceful retching or vomiting. Most tears occur within 2 cm of the gastroesophageal junction on the lesser curvature aspect of the cardia. Mallory-Weiss tears account for 5% to 11% of all major upper GI hemorrhages.20 Most patients present with hematemesis, often associated with alcohol use. Typically, overt bleeding is minor and bleeding ceases spontaneously. Mallory-Weiss tears can also occur with upper GI endoscopy when a patient struggles or retches during the procedure.

Dieulafoy lesions

Dieulafoy lesions account for approximately 5% of cases of major upper GI bleeding.21 Their characteristic feature is the presence of a large-caliber, tortuous artery in the submucosa close to the mucosal surface, which bleeds upon erosion of the overlying mucosa and artery wall [see Figure 3]. They can be extremely difficult to detect endoscopically unless they are actively bleeding at the time of the procedure. Dieulafoy lesions are usually single lesions located in the proximal stomach. However, these lesions can occur anywhere throughout the GI tract. In a review of 90 Dieulafoy lesions, 34% of the lesions were extragastric.22


Figure 3. Dieulafoy lesion in gastric fundus (arrow).

Lower GI Bleeding


Diverticular disease is one of the most common causes of lower GI bleeding, particularly in the elderly. Diverticulosis is uncommon in persons younger than 40 years, but it affects roughly two thirds of persons older than 80 years.23,24 The mean age period for diverticular hemorrhage is the sixth decade of life. The true incidence of diverticular bleeding is difficult to ascertain, given the different definitions and evaluations used in various studies. Bleeding occurs from an arteriole at either the dome or neck of a diverticulum. Typically, there is no associated diverticulitis. Diverticula are most commonly found in the left colon, but many bleeds arise from diverticula in the right colon. Patients typically present with painless, large-volume hematochezia. Because diverticular bleeding tends to stop spontaneously, the diagnosis is often presumptive and is based on exclusion of other sources of bleeding in a patient with diverticulosis.25


Angiodysplasia is an acquired vascular ectasia that is considered to be degenerative in origin, given its propensity to occur in the elderly. Typically, patients are between 60 and 80 years of age. The pathogenesis of angiodysplasia remains unclear, but a proposed cause is chronic, intermittent, low-grade obstruction of submucosal veins, leading to dilatation of mucosal capillaries [see Figure 4]. The lesions of angiodysplasia are usually small (2 to 5 mm in diameter) and can be single or multiple. These lesions can occur anywhere along the GI tract but are most commonly found in the proximal colon (approximately 80%), particularly the cecum.26 Angiodysplasia is an incidental finding at colonoscopy in 2% of nonbleeding patients older than 65 years.27,28 Fewer than 10% of cases of angiodysplasia are associated with bleeding.27 Bleeding stops spontaneously in the majority of patients, but rebleeding is common.


Figure 4. The lesions of angiodysplasia are usually small (2 to 5 mm in diameter) and can be single or multiple. These lesions can occur anywhere along the GI tract but are most commonly found in the proximal colon, particularly the cecum.


Colonoscopic polypectomy is generally considered a safe procedure, but hemorrhage is reported to occur in 0.3% to 6.0% of cases.29 In a retrospective study of 83 patients who underwent a total of 274 polypectomies, bleeding occurred at a median of 5 days (range, 0 to 17 days) after the procedure.30 Bleeding was associated with advanced age, polyps greater than 1 cm in diameter, sessile polyps, and polyps in the cecum. The prognosis for these patients is favorable. Most cases are managed with observation or endoscopic hemostasis.


Emergent Evaluation and Management

Management of GI bleeding is determined by the severity of the bleed; algorithms differ with regard to major bleeding [see Figure 5] and minor bleeding [see Figure 6]. Patients with overt major bleeding require immediate hospitalization with intensive monitoring. Patients are initially stabilized with fluid and blood component replacement and with correction of any coagulopathy or electrolyte imbalances. Endotracheal intubation may be necessary. Stabilization is followed by immediate endoscopic evaluation and therapy as indicated. If hemorrhage control is ineffective and the patient continues to be hemodynamically unstable, radiologic or surgical interventions are considered.


Figure 5. Evaluation and management of overt major gastrointestinal bleeding


Figure 6. Evaluation and management of overt minor gastrointestinal bleeding. (EGD—esophagogastroduodenoscopy)

Clinical and Laboratory Assessment

The history and physical examination provide vital information on the location, severity, and duration of bleeding and can help identify patients at increased risk for exsanguination and rebleeding [see Table 2]. It is important to remember that patients with overt major bleeding from an upper GI source can present with hematochezia. These patients can experience visceral discomfort and orthostatic symptoms shortly after the onset of bleeding. Abdominal pain—especially periumbilical cramping and gaseous distention—usually indicates rapid intestinal transit of blood and suggests a major bleed.

Table 2 Clinical High-Risk Criteria for Rebleeding and Mortality

Advanced age (≥ 70 yr)
Major organ comorbidities
In-hospital bleed
Bright-red hematemesis in patient with liver cirrhosis
Hypotension (systolic blood pressure < 100 mg Hg)
Tachycardia (heart rate > 100 bpm)
Orthostasis (BP drop > 20 mm Hg; HR rise > 20 bpm)
Hemoglobin < 10 g/dl or drop of ≥ 2 g/dl
≥ 4 units of blood transfused in 24 hr

The physician should look for evidence of liver disease, PUD, coagulopathy, previous abdominal aortic aneurysm repair, and significant comorbidities such as heart disease and diabetes mellitus. A history of drug or alcohol ingestion may suggest a diagnosis.

After cessation of active upper GI bleeding, patients may experience melena for 2 to 3 days. In itself, such melena is not necessarily an indication of rebleeding, especially if the patient's hemoglobin level does not decrease.

Serial recording of vital signs is crucial in determining whether an overt major bleed has occurred. Significant volume loss is indicated by hypotension (i.e., systolic blood pressure less than 100 mm Hg), orthostasis (i.e., a decrease in systolic pressure of more than 20 mm Hg or an increase in heart rate of more than 20 beats/min), tachycardia (i.e., heart rate greater than 100 beats/min), or a drop in hemoglobin of more than 2 g/dl. Further assessment of skin pallor, features of liver disease or portal hypertension, and stool color as part of the rectal examination can also help with diagnosis or management.

A nasogastric tube can be placed if there is uncertainty about the location of the bleed in a patient with hematochezia or if bleeding persists in a patient with hematemesis. Aspiration of blood indicates a recent upper GI bleed, but absence of blood in the aspirate does not exclude a recent bleed.

The most important laboratory measurement to assess severity of the initial bleed and to monitor rebleeding is the hemoglobin level. An abrupt drop of more than 2 g/dl indicates a significant bleeding episode. An increase in the ratio of blood urea nitrogen to creatinine to more than 25:1 strongly suggests an upper GI source. Measurement of serum electrolyte concentrations, coagulation indices, platelet count, and liver enzyme levels aids in the diagnosis and guides management.


In most patients, the location of the bleed is identified by upper GI endoscopy or colonoscopy. Endoscopy also provides therapeutic options and essential information on the risk of rebleeding [see Table 3]. There are established visual criteria, based on stigmata of recent hemorrhage, that the endoscopist can use to identify patients at high or low risk for rebleeding.

Table 3 Endoscopic High-Risk Stigmata for Rebleeding and Indications for Endoscopic Therapy

Nonvariceal bleeding

Arterial spurting
Oozing bleed
Nonbleeding visible vessel
Adherent clot

Variceal bleeding

Large varices (> 5 mm esophageal, > 1 cm gastric)
Red wale marks (longitudinal dilated venules resembling whip marks)
Cherry-red spots (< 2 mm diameter)
Hematocystic spots (> 4 mm diameter)

During upper GI endoscopy, if massive active bleeding is encountered, it is prudent to discontinue the procedure and protect the airway by endotracheal intubation before proceeding. If visualization is impaired, use of large-bore orogastric lavage or a jumbo-channel (6 mm) therapeutic endoscope to evacuate blood and clots may be effective. Erythromycin lactobionate (125 mg intravenously) can also be used to promote quick intestinal transit of blood when active bleeding has stopped.

Before colonoscopy, whenever possible, patients should receive a rapid colonic lavage with 2 to 3 L of a nonabsorbable polyethylene glycol solution administered through a nasogastric tube over 2 hours to cleanse the colon and facilitate adequate visualization.


Selective visceral angiography is considered when endoscopic therapy for an established lesion has failed and surgery is not an option or when the site of an active bleed remains obscure after endoscopy. An optimal examination with a high positive yield is best obtained when there is active bleeding at rates exceeding 0.5 to 1 ml/min. Significant complications—including contrast reaction, acute renal failure, and femoral artery thrombosis—have been reported in approximately 9% of cases.31,32 The reported sensitivity of angiography varies from 22% to 87%. The specificity approaches 100%.33

Radionuclide Technetium Scan

A technetium-99m-labeled red cell scan should be considered when active bleeding is suspected but the results of endoscopy are negative. Nuclear scans can detect bleeding at rates that exceed 0.1 ml/min. On scans, however, pooled blood may sometimes be mistaken for active bleeding, which contributes to a reported false positive rate of about 22%.34 Upper GI bleeding may be misdiagnosed as lower GI bleeding because of pooling in the distal ileum or right colon. A positive result is more reliable when the scan is done early rather than delayed for several hours.

Other Measures

Endoscopic techniques for examination of the small bowel that are currently available include push enteroscopy, wireless capsule endoscopy, double-balloon enteroscopy, and intraoperative enteroscopy. Push enteroscopy typically reaches into the proximal jejunum only, whereas wireless capsule enteroscopy, double-balloon enteroscopy, and intraoperative enteroscopy reach the entire small bowel.

Push enteroscopy is currently performed using a pediatric colonoscope with or without an overtube. In one study, the diagnostic yield of enteroscopy in overt GI bleeding was 46%; the most common lesions seen were angiodysplasia and ulcers.35

Wireless capsule endoscopy represents a new technology involving an easily swallowed 11 by 26 mm capsule. No sedation is required. The capsule, which is disposable, contains a color video chip, light source, and transmitter. The patient wears an antenna array on a belt (data recorder). While transiting through the intestine by peristalsis, the capsule takes color photos and sends them to the data recorder. These images are then downloaded onto a computer after the examination [see Figure 7]. There is a total of 6 to 8 hours of recording time. This noninvasive technology can provide images of the entire small bowel; however, it is a purely diagnostic modality. This technique may be beneficial in patients with recurrent or occult GI bleeding of obscure origin, but it is not appropriate in hemodynamically unstable patients with major active bleeding. The diagnostic yield of capsule endoscopy in patients with obscure GI bleeding ranges from 45% to 66%.36,37


Figure 7. Capsule endoscopy photograph showing a small-bowel ulcer (arrow) induced by a nonsteroidal anti-inflammatory drug.

Double-balloon enteroscopy is an emerging endoscopic technique that allows visualization of the entire small bowel, as well as tissue acquisition and therapeutic intervention. The procedure is performed using an endoscope and overtube. Two separate latex balloons are attached to the tip of the endoscope and to the overtube; the balloons are inflated and deflated independently, facilitating advancement of the endoscope through the intestine. The endoscope may be passed via an antegrade (oral) or retrograde (anal) approach. The procedure is particularly limited by a long procedure time (approximately 90 minutes), patient discomfort, and need for operator expertise. The diagnostic yield of double balloon enteroscopy in patients with obscure GI bleeding ranges from 60% to 80%.38,39 Its role in the diagnosis and therapy of GI bleeding remains to be defined.

Intraoperative enteroscopy, performed during exploratory laparotomy through single or multiple enterotomy sites, is indicated for the occasional patient with active or recurrent major bleeding of obscure origin. Complications include mucosal laceration, intramural hematoma, mesenteric hemorrhage, and intestinal ischemia.40


Nonvariceal Bleeding

Endoscopic Therapy

A variety of endoscopic modalities are currently available for the management of GI bleeding. These can be categorized as thermal, mechanical, and injection devices. Thermal devices are either contact (e.g., heater probe, multipolar electrocautery) or noncontact (e.g., argon plasma coagulator, laser). These devices generate sufficient heat to create a hemostatic bond through tissue desiccation. The heater probe consists of a Teflon-coated hollow aluminum cylinder with a heating coil. Only heat (no electrical current) is delivered to the tissue. Multipolar or bipolar cautery works by completion of an electrical circuit between two electrodes on the probe tip. The argon plasma coagulator utilizes high-frequency monopolar alternating current delivered to target tissue through ionized argon gas. The conduit of argon gas is called the argon plasma. Electrons flow through a channel of electrically activated, ionized argon gas from the probe electrode to the tissue, causing a thermal effect at the interface. In laser photocoagulation, which is less frequently used, the conversion of light to heat results in coagulation or vaporization of tissue. The neodymium:yttrium-aluminum-garnet (Nd:YAG) laser is the most commonly used laser. Mechanical devices for hemostasis include metallic clips and rubber-band ligators. An injection solution that is generally used to achieve hemostasis consists of saline mixed with epinephrine at a 1:10,000 concentration.

These therapeutic modalities are used alone or in combination. A common practice is to start by injecting epinephrine and saline submucosally in the region of active bleeding so as to stop or slow hemorrhaging and therefore allow for adequate inspection. Thermal or mechanical modalities are then used to achieve definitive hemostasis. Prospective, controlled studies have confirmed the benefit of endoscopic intervention in achieving initial hemostasis and in prevention of rebleeding.41 Combination therapy (i.e., injection plus thermal therapy) has been demonstrated to reduce rebleeding rates more successfully than single therapy.42,43 Currently, combination therapy using injection followed by either a thermal or a mechanical intervention is the most effective approach. Rebleeding after endoscopic therapy occurs in approximately 20% of cases, typically within 48 to 72 hours after treatment. However, rebleeding can occur as late as 7 days after therapy.


Initial drug therapy for major nonvariceal upper GI bleeding is directed at gastric acid suppression. In a randomized, double-blind study comparing high-dose omeprazole with placebo, rebleeding after endoscopic therapy occurred less frequently in the omeprazole group (7% versus 23%).44 In general, proton pump inhibitors are administered in doses that reduce gastric acidity. Blood clot stability depends on intragastric pH, with optimum stability at a pH of 6 or higher.45

In patients with PUD, long-term acid suppression and eradication of Helicobacter pylori infection after endoscopic intervention promote ulcer healing, including ulceration at the treatment site, and reduce rebleeding substantially. GI bleeding from NSAIDs is best prevented by avoiding these drugs.

Radiologic Intervention

Selective arterial embolization and selective vasoconstriction with intra-arterial infusion of vasopressin are the methods currently available for the control of major nonvariceal GI bleeding. The proponents of embolization favor this form of therapy because it reduces the need for observation in the intensive care unit, and it eliminates indwelling arterial lines, the risk of catheter dislodgement, and problematic systemic side effects of intravenous vasopressin [see Pharmacotherapy, below]. Advances in catheter design have allowed for superselective embolization of vasa recta; in experienced units, this modality is probably the treatment of choice. A study of superselective embolization in 48 patients with lower GI bleeding showed that embolization was the definitive treatment in 44% of patients, with a 27% technical failure rate.46 The risks associated with embolization include misplacement of embolic material, inadvertent distal reflux of embolic agent, and excessive devascularization of an organ leading to ischemia and eventual luminal stenosis. Endoscopy can be helpful in determining ischemic injury if suspected. Microcoils (e.g., stainless steel, platinum), gelfoam pledgets, polyvinyl alcohol particles, and collagen suspensions have been used for embolization.

Intra-arterial vasopressin is the drug of choice for selective vasoconstrictive therapy and is generally infused for a minimum of 24 hours. It is associated with a 70% rate of bleeding control and an 18% rate of rebleeding.47,48,49 Vasopressin may be ineffective when bleeding arises from large arteries that do not constrict in response to therapy. In a study comparing embolization with vasopressin, initial hemostasis rates were similar for the two modalities, but a higher rate of rebleeding was seen with vasopressin.2 The use of intra-arterial provocative mesenteric angiography with heparin and tissue plasminogen activator (t-PA) to aid in diagnosis has been described but is still in the experimental stage.


Despite the high overall success rate of endoscopic therapy in the treatment of major GI bleeding, surgery is still indicated when (1) initial hemostatic control cannot be achieved, (2) rebleeding occurs despite repeated endoscopic sessions, (3) a large (> 2 cm) penetrating ulcer is present, (4) a vessel larger than 2 mm in diameter is visible within the culprit lesion, (5) the ulcer is located in the posterior duodenal bulb (this location is associated with the large gastroduodenal artery), (6) the patient requires substantial transfusion (i.e., four or more units of blood over 24 hours, after preadmission losses have been replaced), and (7) radiologic intervention has failed, is unavailable, or is not appropriate for the particular lesion. The choice of surgery depends on the location of the bleed and the presence of comorbidities. Localization of the site of bleeding is critical for surgical planning.

Variceal Bleeding

Endoscopic Therapy

With variceal bleeding, endoscopic treatment is used primarily for esophageal varices; the techniques include band ligation and sclerotherapy. Band ligation is considered the first-line endoscopic therapy for esophageal varices. The band ligator is readily attached to the distal end of the endoscope, which is advanced to the varix; the endoscopist then suctions the varix into the ligator cap and deploys a rubber band around the varix. This results in the plication of the varices and surrounding submucosal tissue, with fibrosis and eventual obliteration of varices. Comparative studies report a better initial control of bleeding (control rates, 91% versus 77%) and lower rebleeding rates (rebleeding, 24% versus 47%) with band ligation than with sclerotherapy.50 Complications of banding include retrosternal chest pain, dysphagia from compromise of the esophageal lumen, band ulceration (usually superficial ulcers that heal within 2 weeks), or esophageal perforation. Complication rates vary from 2% to 19%.51,52

Sclerotherapy utilizes a variety of sclerosants to induce variceal thrombosis, with sodium tetradecyl sulfate and ethanolamine oleate used most frequently. Intravariceal injections are more effective than paraesophageal injections in controlling bleeding. Compared with a sham injection, sclerotherapy is significantly more likely to stop bleeding (91% versus 60%), reduce mortality during hospitalization (mortality, 25% versus 49%), reduce rebleeding rates (rebleeding, 20% versus 51%), and reduce the need for transfusion (four versus eight units).53Complications of sclerotherapy include retrosternal chest pain, fever, ulceration (usually deep ulcers that heal within 3 weeks), dysphagia, delayed perforation (1 to 4 weeks later), and stricture formation. Complication rates vary from 19% to 35%.50,54,55 The popularity of sclerotherapy has diminished as a result of these complications.

Endoscopic glue injection therapy using cyanoacrylate-based compounds has shown efficacy for the control and prevention of bleeding from gastric varices.56,57 However, this therapy is not yet approved by the Food and Drug Administration.

If bleeding continues despite endoscopic therapy or if endoscopic therapy cannot be initiated, then a modified Sengstaken-Blakemore (Minnesota) tube should be inserted. However, this is only a temporary measure until more definitive treatment—endoscopic, radiologic, or surgical—can be undertaken.

Preventive measures may be indicated in patients with esophageal varices. Preventive measures are generally offered to patients who have a history of GI bleeding and to those who have large esophageal varices without a prior bleeding event. Currently, the accepted preventive measures for variceal bleeding include endoscopic band ligation, beta-blocker therapy, or a combination of both. Ligation is performed every 14 to 21 days until varices are completely eradicated, which typically requires three or four sessions.


In acute variceal bleeding, splanchnic blood flow and portal pressure can be reduced by intravenous infusion of vasoconstrictors such as somatostatin, octreotide, vasopressin, and terlipressin. Somatostatin, a naturally occurring peptide, is reported to stop variceal bleeding in 80% of patients.17,58 Side effects are few and include hyperglycemia and abdominal pain. Octreotide is a synthetic analogue of somatostatin that is preferred because of its longer half-life. The combination of pharmacologic treatment (e.g., octreotide for 5 days) and endoscopic therapy appears to offer better control of acute bleeding than either alone.

Vasopressin is a potent vasoconstrictor that has a reported overall success rate of 50% but a high rebleeding rate when treatment is discontinued.59 It has a short half-life and therefore is given as a continuous infusion. Because of the systemic vasoconstrictive side effects associated with vasopressin that may lead to myocardial or mesenteric ischemia, it is not commonly used. To minimize these side effects, it is given in conjunction with nitroglycerin. Terlipressin is a synthetic analogue of vasopressin that has fewer side effects and a longer half-life and is given in bolus injections.

The role of beta blockers is primarily prophylactic. These agents are not used in the acute management of variceal bleeding. The use of isosorbide mononitrate with beta-blocker therapy does not offer a survival advantage and in fact reduces the tolerability of therapy. In cirrhotic patients with GI bleeding, short-term antibiotic prophylaxis decreases the incidence of infections, particularly spontaneous bacterial peritonitis, and improves survival.60

Radiologic Intervention

The radiologic intervention available for variceal bleeding is transjugular intrahepatic portosystemic shunt (TIPS). The accepted indications for TIPS are bleeding or rebleeding that cannot be controlled by either pharmacologic or endoscopic therapy. TIPS is contraindicated in patients with severe hepatic failure, chronic heart failure, hepatic encephalopathy, bile duct obstruction, or cholangitis. TIPS is reported to control bleeding in at least 90% of patients, with rebleeding rates of 12% to 26% at 1 year and 16% to 44% at 2 years.61,62 Patients require close surveillance for shunt dysfunction (evidenced by reduced flow by Doppler ultrasound or the reappearance of varices) because primary shunt patency rates are poor (reported cumulative patency rates of 50% at 1 year and 21% at 3 years), but cumulative secondary shunt patency rates can be as satisfactory as 85% and 55% at 1 and 3 years, respectively.55

TIPS should not be undertaken lightly, because the overall procedure-related mortality can be as high as 1% to 2%,54 largely from intraperitoneal hemorrhage. Other complications include hepatic encephalopathy, portal vein thrombosis, renal failure, sepsis, and stent migration or stenosis.

Surgical Intervention

Surgical intervention is rarely used for variceal bleeding; it is considered when other measures have proved ineffective. Surgical treatments include portosystemic venous shunt operations and esophageal devascularization. A variety of surgical shunts are available. These are generally classified as total, partial, or selective, depending on the intended impact of portal flow diversion. The end-to-side portacaval shunt is a total shunt that diverts all portal blood flow into the inferior vena cava. The side-to-side portacaval shunt diverts only a part of the portal blood flow. Selective shunts decompress variceal flow while preserving portal blood flow. The distal splenorenal shunt is a selective shunt designed to prevent encephalopathy, which is often seen with total shunts. Surgical shunts are used for both esophageal and gastric varices. Encephalopathy, accelerated progression of liver failure, and perioperative morbidity can occur with surgical intervention. Esophageal devascularization may be an effective means of controlling acute variceal bleeding, but bleeding can recur as additional varices develop.

Occult Bleeding

The critical metabolic sequela of occult GI bleeding is iron deficiency.63 Occult GI bleeding causes most cases of iron deficiency in adults, especially in men and postmenopausal women.


Most of the many lesions that cause overt bleeding can also produce occult blood loss. However, variceal and diverticular hemorrhage invariably bleed overtly, whereas lesions such as watermelon stomach (gastric antral vascular ectasia) and diaphragmatic hernia with Cameron erosions tend to bleed occultly. Occult GI bleeding in most patients is suspected only when manifested by fatigue, pallor, or the finding of iron deficiency.


In Western countries, erosive or ulcerative diseases of the esophagus, stomach, and duodenum are the most common GI lesions associated with occult bleeding and iron deficiency anemia. Most peptic ulcer disease is caused by either H. pylori infection or the use of drugs such as aspirin or other NSAIDs. The association between large diaphragmatic hernias and iron deficiency anemia has long been known. A large diaphragmatic hernia is found in about 10% of iron-deficient patients.64 Blood loss in these patients is generally caused by longitudinal mucosal erosions (Cameron erosions) located proximally in the hernia and believed to be secondary to repeated mechanical trauma from respiration.

Cancers and Neoplasms

In adults from Western countries, GI tumors are second only to PUD as a cause of occult bleeding with subsequent iron deficiency anemia.65Colorectal cancer is currently the most common source of occult bleeding from GI tract malignancies.

Vascular Causes

Vascular malformations are found in approximately 6% of adults with iron deficiency anemia.66,67 Vascular malformations may be acquired or hereditary (hereditary hemorrhagic telan-giectasia). An increasingly recognized and endoscopically treatable vascular lesion is watermelon stomach [see Figures 8a and 8b], which typically presents as iron deficiency anemia in older women.


Figure 8. Watermelon stomach with (a) typical spokes of vascular ectasia radiating from the pylorus into the antrum and (bClose-up view.


When a patient is found to have iron deficiency and occult GI bleeding, it is critical to conduct a thorough GI investigation. Such an evaluation may disclose a health-threatening lesion, in which case specific therapy can be given to prevent associated morbidity and further iron loss. Management with iron therapy and monitoring is not appropriate until a specific lesion has been treated or a lesion has been ruled out as the cause of the iron deficiency [see Figure 9].


Figure 9. Evaluation and management of occult gastrointestinal bleeding.

Regardless of the type of lesion causing the bleeding, treatment with iron supplementation is important to correct iron deficiency. With conditions such as Cameron erosions, iron supplementation is the mainstay of treatment. Most patients can be managed as outpatients. Oral iron therapy with ferrous sulfate is preferred because it is inexpensive, effective, and, in most cases, well tolerated [see 5:II Red Blood Cell Function and Disorders of Iron Metabolism]. The maximal adult dose of ferrous sulfate is 325 mg three times a day. Absorption is not appreciably increased with higher doses. Oral iron is as effective as parenteral iron in repleting iron stores, except in patients with a malabsorption syndrome, and is safer.


Figures 1,2,3,4 and 8 © 2003 Mayo Foundation for Medical Education and Research. Used by permission.


  1. Swain CP, Storey DW, Bown SG, et al: Nature of the bleeding vessel in recurrently bleeding gastric ulcers. Gastroenterology 90:595, 1986
  2. Peter DJ, Dougherty JM: Evaluation of the patient with gastrointestinal bleeding: an evidence based approach. Emerg Med Clin North Am 17:239, 1999
  3. Vreeburg EM, Snel P, de Bruijne JW, et al: Acute upper gastrointestinal bleeding in the Amsterdam area: incidence, diagnosis, and clinical outcome. Am J Gastroenterol 92:236, 1997
  4. Friedman LS, Martin P: The problem of gastrointestinal bleeding. Gastroenterol Clin North Am 22:717, 1993
  5. Pianka JD, Affronti J: Management principles of gastrointestinal bleeding. Prim Care 28:557, 2001
  6. Farrell JJ, Friedman LS: Gastrointestinal bleeding in the elderly. Gastroenterol Clin North Am 30:377, 2001
  7. Gostout CJ, Wang KK, Ahlquist DA, et al: Acute gastrointestinal bleeding: experience of a specialized management team. J Clin Gastroenterol 14:260, 1992
  8. Gostout CJ: Acute lower GI bleeding. Current Medicine: Clinical Practice of Gastroenterology. Brandt LJ, Daum F, Eds. Churchill Livingstone, Philadelphia, 1998, p 651
  9. Hay JA, Lyubashevsky E, Elashoff J, et al: Upper gastrointestinal hemorrhage: clinical guideline determining the optimal hospital length of stay. Am J Med 100:313, 1996
  10. Rockall TA, Logan RF, Devlin HB, et al: Selection of patients for early discharge or outpatient care after acute upper gastrointestinal haemorrhage. National Audit of Acute Upper Gastrointestinal Haemorrhage. Lancet 347:1138, 1996
  11. Andersen IB, Jorgensen T, Bonnevie O, et al: Smoking and alcohol intake as risk factors for bleeding and perforated peptic ulcers: a population-based cohort study. Epidemiology 11:434, 2000
  12. Bardhan KD, Graham DY, Hunt RH, et al: Effects of smoking on cure of Helicobacter pyloriinfection and duodenal ulcer recurrence in patients treated with clarithromycin and omeprazole. Helicobacter 2:27, 1997
  13. Cohen H: Peptic ulcer and Helicobacter pylori. Gastroenterol Clin North Am 29:775, 2000
  14. Solomon DH, Gurwitz JH: Toxicity of nonsteroidal anti-inflammatory drugs in the elderly: is advanced age a risk factor? Am J Med 102:208, 1997
  15. Dalton SO, Johansen C, Mellemkjaer L, et al: Use of selective serotonin reuptake inhibitors and risk of upper gastrointestinal tract bleeding: a population-based cohort study. Arch Intern Med 163:59, 2003
  16. Gostout CJ: Patient assessment and resuscitation. Gastrointest Endosc Clin N Am 9:175, 1999
  17. Sharara AI, Rockey DC: Gastroesophageal variceal hemorrhage. N Engl J Med 345:669, 2001
  18. Cales P, Zabotto B, Meskens C, et al: Gastroesophageal endoscopic features in cirrhosis: observer variability, interassociations, and relationship to hepatic dysfunction. Gastroenterology 98:156, 1990
  19. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices: a prospective multicenter study. The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. N Engl J Med 319:983, 1988
  20. Sugawa C, Steffes CP, Nakamura R, et al: Upper GI bleeding in an urban hospital: etiology, recurrence, and prognosis. Ann Surg 212:521, 1990
  21. Kasapidis P, Georgopoulos P, Delis V: Endoscopic management and long-term follow-up of Dieulafoy's lesions in the upper GI tract. Gastrointest Endosc 55:527, 2002
  22. Norton ID, Petersen BT, Sorbi D: Management and long-term prognosis of Dieulafoy lesion. Gastrointest Endosc 50:762, 1999
  23. Buttenschoen K, Buttenschoen DC, Odermath R, et al: Diverticular disease-associated hemorrhage in the elderly. Langenbecks Arch Surg 386:8, 2001
  24. Freeman SR, McNally PR: Diverticulitis. Med Clin North Am 77:1149, 1993
  25. McGuire HH: Urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med 342:1609, 2000
  26. Hodgson H: Hormonal therapy for gastrointestinal angiodysplasia. Lancet 359:1630, 2002
  27. Imdahl A: Genesis and pathophysiology of lower gastrointestinal bleeding. Langenbecks Arch Surg 386:1, 2001
  28. Hochter W, Weingart J, Kuhner W, et al: Angiodysplasia in the colon and rectum: endoscopic morphology, localisation and frequency. Endoscopy 17:182, 1985
  29. Rosen L, Bub DS, Reed JF 3rd, et al: Hemorrhage following colonoscopic polypectomy. Dis Colon Rectum 36:1126, 1993
  30. Sorbi D, Norton I, Conio M, et al: Postpolypectomy lower GI bleeding: descriptive analysis. Gastrointest Endosc 51:690, 2000
  31. Lingenfelser T, Ell C: Gastrointestinal bleeding in the elderly. Best Pract Res Clin Gastroenterol 15:963, 2001
  32. Egglin TK, O'Moore PV, Feinstein AR, et al: Complications of peripheral arteriography: a new system to identify patients at increased risk. J Vasc Surg 22:787, 1995
  33. Koval G, Benner KG, Rosch J, et al: Aggressive angiographic diagnosis in acute lower gastrointestinal hemorrhage. Dig Dis Sci 32:248, 1987
  34. Ng DA, Opelka FG, Beck DE, et al: Predictive value of technetium Tc 99m-labeled red blood cell scintigraphy for positive angiogram in massive lower gastrointestinal hemorrhage. Dis Colon Rectum 40:471, 1997
  35. Feitoza A, Rajan E, Gostout CJ: Overt gastrointestinal bleeding: diagnostic and therapeutic yield of push-enteroscopy. Am J Gastroenterol 96:S104, 2001
  36. Ell C, Remke S, May A, et al: The first prospective controlled trial comparing wireless capsule endoscopy with push enteroscopy in chronic gastrointestinal bleeding. Endoscopy S34:685, 2002
  37. Triester SL, Leighton JA, Leontiadis GI, et al: A meta-analysis of the yield of capsule endoscopy compared to other diagnostic modalities in patients with obscure gastrointestinal bleeding. Am J Gastroenterol 100:2407, 2005
  38. Sun B, Rajan E, Cheng S, et al: Diagnostic yield and therapeutic impact of double-balloon enteroscopy in a large cohort of patients with obscure gastrointestinal bleeding. Am J Gastroenterol 101:2011, 2006
  39. Yamamoto H, Kita H, Sunada K, et al: Clinical outcomes of double-balloon endoscopy for the diagnosis and treatment of small-intestinal diseases. Clin Gastroenterol Hepatol 2:1010, 2004
  40. Lopez MJ, Cooley JS, Petros JG, et al: Complete intraoperative small-bowel endoscopy in the evaluation of occult gastrointestinal bleeding using the Sonde enteroscope. Arch Surg 131:272, 1996
  41. Fullarton GM, Birnie GG, Macdonald A, et al: Controlled trial of heater probe treatment in bleeding peptic ulcers. Br J Surg 76:541, 1989
  42. Lin HJ, Tseng GY, Perng CL, et al: Comparison of adrenaline injection and bipolar electrocoagulation for the arrest of peptic ulcer bleeding. Gut 44:715, 1999
  43. Bleau BL, Gostout CJ, Sherman KE, et al: Recurrent bleeding from peptic ulcer associated with adherent clot: a randomized study comparing endoscopic treatment with medical therapy. Gastrointest Endosc 56:1, 2002
  44. Palmer KR: Intravenous omeprazole after endoscopic treatment of bleeding peptic ulcers. Gut 49:610, 2001
  45. Green FW Jr, Kaplan MM, Curtis LE, et al: Effect of acid and pepsin on blood coagulation and platelet aggregation: a possible contributor to prolonged gastroduodenal mucosal hemorrhage. Gastroenterology 74:38, 1978
  46. Bandi R, Shetty PC, Sharma RP, et al: Superselective arterial embolization for the treatment of lower gastrointestinal hemorrhage. J Vasc Interv Radiol 12:1399, 2001
  47. Rosen RJ, Sanchez G: Angiographic diagnosis and management of gastrointestinal hemorrhage: current concepts. Radiol Clin North Am 32:951, 1994
  48. Gostout CJ: The evaluation and management of acute upper gastrointestinal bleeding. Clinical Practice of Gastroenterology. Brandt LJ, Daum F, Eds. Churchill Livingstone, Philadelphia, 1998, p 1514
  49. Zuccaro G Jr: Management of the adult patient with acute lower gastrointestinal bleeding. American College of Gastroenterology. Practice Parameters Committee. Am J Gastroenterol 93:1202, 1998
  50. Qureshi W, Adler DG, Davila R, et al: ASGE Guideline: the role of endoscopy in the management of variceal hemorrhage, updated 2005. Gastrointest Endosc 62:651, 2005
  51. Sarin SK, Govil A, Jain AK, et al: Prospective randomized trial of endoscopic sclerotherapy versus variceal band ligation for esophageal varices: influence on gastropathy, gastric varices and variceal recurrence. J Hepatol 26:826, 1997
  52. Laine L, el-Newihi HM, Migikovsky B, et al: Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 119:1, 1993
  53. Sclerotherapy for actively bleeding esophageal varices in male alcoholics with cirrhosis. Veterans Affairs Cooperative Variceal Sclerotherapy Group. Gastrointest Endosc 46:1, 1997
  54. Stiegmann GV, Goff JS, Michaletz-Onody PA, et al: Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices. N Engl J Med 326:1527, 1992
  55. Lo GH, Lai KH, Cheng JS, et al: Emergency banding ligation versus sclerotherapy for the control of active bleeding from esophageal varices. Hepatology 25:1101, 1997
  56. Sarin SK, Jain AK, Jain M, et al: A randomized controlled trial of cyanoacrylate versus alcohol injection in patients with isolated fundic varices. Am J Gastroenterol 97:1010, 2002
  57. Dhiman RK, Chawla Y, Taneja S, et al: Endoscopic sclerotherapy of gastric variceal bleeding with N-butyl-2-cyanoacrylate. J Clin Gastroenterol 35:222, 2002
  58. Double-blind randomized controlled trial comparing terlipressin and somatostatin for acute variceal hemorrhage. Variceal Bleeding Study Group. Gastroenterology 111:1291, 1996
  59. Nader A, Grace ND: Pharmacological intervention during the acute bleeding episode. Gastrointest Endosc Clin N Am 9:301, 1999
  60. Bernard B, Grange JD, Khac EN, et al: Antibiotic prophylaxis for the prevention of bacterial infections in cirrhotic patients with gastrointestinal bleeding: a meta-analysis. Hepatology 29:1655, 1999
  61. McKusick MA: Interventional radiology for the control and prevention of bleeding. Gastrointest Endosc Clin N Am 9:311, 1999
  62. Zhuang ZW, Teng GJ, Jeffery RF, et al: Long-term results and quality of life in patients treated with transjugular intrahepatic portosystemic shunts. AJR Am J Roentgenol 179:1597, 2002
  63. Ahlquist DA: Approach to the patient with occult gastrointestinal bleeding. Textbook of Gastroenterology, 4th ed. Yamada T, Alpers DH, Kaplowitz N, et al, Eds. Lippincott Williams & Wilkins, Philadelphia, 2003
  64. Descamps C, Schmit A, Van Gossum A: “Missed” upper gastrointestinal tract lesions may explain “occult” bleeding. Endoscopy 31:452, 1999
  65. Rockey DC, Celo JP: Evaluation of the gastrointestinal tract in patients with iron deficiency anemia. N Engl J Med 329:1691, 1993
  66. Kendrick ML, Buttar NS, Anderson MA: Contribution of intraoperative enteroscopy in the management of obscure gastrointestinal bleeding. J Gastrointest Surg 5:162, 2001
  67. Gostout CJ, Ahlquist DA, Radford CM, et al: Endoscopic laser therapy for watermelon stomach. Gastroenterology 96:1462, 1989

Editors: Dale, David C.; Federman, Daniel D.