Angiodysplasia of the Colon 

Updated: May 08, 2019
Author: Hussein Al-Hamid, MD; Chief Editor: BS Anand, MD 



Angiodysplasia is the most common vascular lesion of the gastrointestinal tract, and this condition may be asymptomatic, or it may cause gastrointestinal (GI) bleeding.[1] The vessel walls are thin, with little or no smooth muscle, and the vessels are ectatic and thin (see image below).

Angiodysplasia identified on the cecum wall during Angiodysplasia identified on the cecum wall during colonoscopy.

Phillips first described a vascular abnormality that caused bleeding from the large bowel in a letter to the London Medical Gazette in 1839. During the 1920s, neoplasms were considered the major source of GI hemorrhage. However, in the 1940s and 1950s, diverticular disease was recognized as an important source of bleeding. In 1951, Smith described active bleeding from a diverticulum visualized through a sigmoidoscope. An association between colonic angiodysplasia and aortic stenosis was described by Heyde in 1958.[2]

Vascular abnormalities as a source of active bleeding were once considered controversial. In 1960, Margulis and colleagues identified a vascular malformation in the cecum of a 69-year-old woman who presented with massive bleeding.[3] This diagnosis was accomplished with operative mesenteric arteriography.

Galdabini first used the name angiodysplasia in 1974; however, confusion about the exact nature of these lesions resulted in a multitude of terms that included arteriovenous malformation, hemangioma, telangiectasia, and vascular ectasia. These terms have varying pathophysiologies, with a common presentation of GI bleeding.

Angiodysplasia is a degenerative lesion of previously healthy blood vessels found most commonly in the cecum and proximal ascending colon. Seventy-seven percent of angiodysplasias are located in the cecum and ascending colon, 15% are located in the jejunum and ileum, and the remainder is distributed throughout the alimentary tract. These lesions typically are nonpalpable and small (< 5 mm).

Angiodysplasia is the most common vascular abnormality of the GI tract. After diverticulosis, it is the second leading cause of lower GI bleeding in patients older than 60 years. Angiodysplasia may account for approximately 6% of cases of lower GI bleeding. It may be observed incidentally at colonoscopy in as many as 0.8% of patients older than 50 years. The prevalence for upper GI lesions is approximately 1%-2%.

Small bowel angiodysplasia may account for 30%-40% of cases of GI bleeding of obscure origin. In a recent retrospective colonoscopic analyses, it was shown that 12.1% of 642 persons without symptoms of irritable bowel syndrome (IBS), and 11.9% of those with IBS had colonic angiodysplasia.[4]

Colonic arteriovenous malformation (AVM) is one of the causes of lower GI bleeding. Unlike small vascular ectasia or angiodysplasia, colonic AVM tends to be solitary, large in size, and identified endoscopically as flat or as an elevated bright red lesion.[5]

Angiodysplasia may present as an isolated lesion or as multiple vascular lesions. Unlike congenital or neoplastic vascular lesions of the GI tract, this lesion is not associated with angiomatous lesions of the skin or other viscera.

Clinical presentation in patients with angiodysplasia is usually characterized by maroon-colored stool, melena, or hematochezia. Bleeding is usually low grade, but it can be massive in approximately 15% of patients. In 20%-25% of bleeding episodes, only tarry stools are passed. Iron deficiency anemia and stools that are intermittently positive for occult blood can be the only manifestations of angiodysplasia in 10%-15% of patients. Bleeding stops spontaneously in greater than 90% of cases but is often recurrent.


The exact mechanism of development of angiodysplasia is not known, but chronic venous obstruction may play a role.[6, 7] This hypothesis accounts for the high prevalence of these lesions in the right colon and is based on the Laplace law. The Laplace law relates wall tension to luminal size and transmural pressure difference in a cylinder, whereby the wall tension is equal to the pressure difference multiplied by the radius of the cylinder. In the case of the colon, wall tension refers to intramural tension, the pressure difference is that between the bowel lumen and the peritoneal cavity, and cylinder radius is the radius of the right colon. Wall tension is highest in bowel segments with the greatest diameter, such as the right colon.

This theory involving chronic venous obstruction suggests that repeated episodes of colonic distention are associated with transient increases in lumen pressure and size., which results in multiple episodes of increasing wall tension with obstruction of submucosal venous outflow, especially where these vessels pierce the smooth muscle layers of the colon. Over many years, this process causes gradual dilation of the submucosal veins and, eventually, dilation of the venules and arteriolar capillary units feeding them. Ultimately, the capillary rings dilate, the precapillary sphincters lose their competency, and a small arteriovenous communication forms. This accounts for the characteristic early-filling vein observed during mesenteric angiography.

The developmental theory of angiodysplasia accounts for several clinical and pathologic features, including the occurrence in older individuals, location in the cecum and proximal right colon, and prominent submucosal veins that dilate after traversing the muscularis propria. In addition, it also accounts for the lack of pathologic changes in arterioles supplying vascular ectasias and the absence of any mucosal lesion associated with them. Previous studies demonstrating that colonic motility, increased tension in the bowel wall, and increased intraluminal pressure can diminish venous flow lend further support to this theory. Dilated submucosal veins have been one of the most consistent histologic findings and may represent the earliest abnormality in colonic angiodysplasia. This histologic feature supports the theory of chronic venous obstruction in the genesis of angiodysplasia.

Of note, the aforementioned pathophysiologic mechanisms responsible for the development of cecal lesions are unlikely to apply to lesions in the upper GI tract, despite being morphologically identical.

Increased expression of angiogenic factors, like basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF), is also believed to play a role in the pathogenesis of colonic angiodysplasia.[8]

A link between the mechanical disruption of high molecular-weight multimers of von Willebrand factor, due to the turbulent blood flow through a narrowed valve in patients with aortic stenosis, and colonic angiodysplasia has been proposed.

Colonic angiodysplasia and true diverticula may be associated.[9] Portal hypertension colopathy, a form of colonic angiodysplasia, has been described.[10] Vascular ectasia of the entire GI tract has been reported in a patient receiving high-dose chemotherapy and autologous stem cell transplantation for relapsing Hodgkin disease.[11]


Angiodysplasia, a vascular malformation, is the most common cause of recurrent lower intestinal hemorrhage in patients with renal failure. Lesions are multiple in 40%-75% of cases and are often located in the stomach and duodenum but can also affect the colon and the jejunum. Diagnosis is improved by endoscopy, which has a much higher sensitivity compared with angiography. Capsular endoscopy may reveal the hemorrhage site in the small intestine when regular endoscopy fails, and therapeutic intervention usually includes argon plasma coagulation.[12]

The exact cause of angiodysplasia is unknown, but theories include degenerative changes of the small blood vessels associated with aging (most widely accepted theory) and long-term local hypo-oxygenation of the microcirculation from cardiac, vascular, or pulmonary disease

Angiodysplasia has been reported to be associated with aortic stenosis. Heyde first reported this association in 1958, describing Heyde syndrome as the combination of calcific aortic stenosis and GI bleeding due to angiodysplasia of the colon.[2] He reported on 10 patients with GI bleeding of unknown origin who had clinical signs of aortic stenosis and speculated that these patients bled from sclerotic GI vessels.[2] One month later, Schwartz et al suggested a similar association.[13]

Catell was quoted in a clinicopathologic conference on such a case in 1965. He suggested that these patients bled from a vascular lesion in the ascending colon that the pathologists could not demonstrate. Catell recommended a blind right hemicolectomy, which, in his experience, had resulted in cessation of bleeding in these patients.

Mucosal hypoperfusion from cardiac disease was later postulated to be the underlying cause for the development of angiodysplasia. Studies using echocardiograms indicated that only a few patients with angiodysplastic lesions had significant valvular heart disease, such as aortic stenosis. More patients had aortic sclerosis rather than aortic stenosis. Aortic valve replacement or colectomy may be effective in the cessation of recurrent bleeding or after correction of heart failure in hypertrophic subaortic stenosis.[14]

In most persons with angiodysplasia, cardiac findings have no importance in the development of angiodysplasia, although in Japan the most prevalent underlying condition in patients with colonic angiodysplasia was cardiovascular disease (56%).[15] Critiques of the literature by Imperiale and Ransohoff found a lack of conclusive evidence to support the association of aortic stenosis, angiodysplasia, and GI bleeding[16]

Hypoperfusion or hypo-oxygenation from cardiac or pulmonary disease possibly results in ischemic necrosis of an existing angiodysplastic lesion. The observation that low cardiac output usually is a late occurrence in the course of aortic valve disease has not supported this possibility. In addition, the low cardiac output associated with mitral stenosis is not associated with a propensity for bleeding in angiodysplastic lesions.

Cessation of angiodysplastic bleeding after aortic valve replacement has been reported in patients with severe aortic stenosis.

Pate et al suggested that the Heyde syndrome consists of bleeding from presumably latent angiodysplasia as a result of a hematologic defect, such as a lack of high molecular weight von Willebrand factor multimers.[14]

Bleeding angiodysplastic lesions in the upper GI tract have been found with a high prevalence in patients with chronic renal failure requiring dialysis.[17] However, this has not been a consistent finding. Patients with chronic renal failure are more likely to have coagulopathies that are related to quantitative and qualitative platelet defects and abnormal function and the structure of von Willebrand factor.

Bleeding from angiodysplastic lesions in the upper and lower GI tract has been reported in patients with von Willebrand disease. Because factor VIII complex is synthesized partly in vascular endothelial cells, patients with von Willebrand disease and angiodysplasia have been proposed to have an underlying endothelial defect that may be related to the subsequent development of the two disorders. However, as with renal failure, the coagulopathy is more likely responsible for bleeding than for the development of the lesions.

Moreover, degenerative aortic stenosis is associated with increased destruction of high molecular weight multimers of von Willebrand factor which can promote bleeding from intestinal angiodysplasias. Aortic valve replacement is the first line therapy for advanced stage of valve disease but can also be an effective treatment for coexistent bleeding angiodysplasias and acquired von Willebrand disease.[18]

Roskell et al demonstrated a relative deficiency of collagen type IV in the mucosal vessels in angiodysplasia compared to controls.[19] The authors proposed that this deficiency may be related to the patients' susceptibility to ectasia and hemorrhage.

In a small study, Junquera et al observed an increased expression of angiogenic factors in human colonic angiodysplasia.[8] This study noted that vascular immunoreactivity for basic fibroblast growth factor was observed in 7 (39%) specimens from patients with colonic angiodysplasia, whereas either very limited or no immunostaining was found in sections from specimens of patients with colonic cancer and healthy margins.

Patients with scleroderma may also have a higher incidence of angiodysplasia throughout the GI tract, including the colon.

The incidence of GI bleeding after implantation of the continuous-flow left-ventricular assist devices (CF-LVAD) varies between 18% and 40% in many studies, and it is believed to be higher compared to the older generation pulsatile LVAD. GI bleeding due to angiodysplasia and arteriovenous malformations (AVMs) is more common and appears to be related to the blood-flow rheology of these devices.[20, 21]

A study from the University of Minnesota showed that 51 of 233 (22%) patients who underwent CF-LVAD (Heartmate 2) implantation between 2005 and 2013 had GI bleeding.[21] Acquired von Willebrand disease is believed to be the proposed mechanism in some patients due to the reduction in high molecular weight (HMW) multimers of von Willebrand factor (vWF) from the shear stress affecting blood flow in the CF-LVAD .[21] In addition to acquired von Willebrand disease, activation of the fibrinolytic system and a loss of platelet numbers and function during CF-LVAD support is a possibility.[21] The low pulse pressure and the secondary intestinal hypoperfusion is another proposed mechanism for GI bleeding after CF-LVAD. Screening of patients for angiodysplasia and von Willebrand disease before CF-LVAD implant may allow for effective preemptive treatment.[20, 21]

In Heyde syndrome, the association of aortic stenosis and bleeding from angiodysplasia appears to be related to subtle alterations in the plasma coagulation factors. vWF is the strongest possible link between aortic stenosis and bleeding associated with GI angiodysplasia. Aortic valve replacement appears to offer the best hope of long-term resolution of the bleeding.[22]

Aortic valve replacement corrects the vWF abnormalities with long-term resolution of GI bleeding.[23] Resolution of anemia usually follows aortic valve replacement.[24]

Hereditary hemorrhagic telangiectasia (HHT), also known as Osler-Weber-Rendu disease, is an autosomal dominant disorder of the fibrovascular tissue. It is characterized by the classic triad of mucocutaneous telangiectasias, recurrent hemorrhages, and familial occurrence. Gastric angiodysplasia of the fundus and body of the stomach is observed. Histopathologic study shows dilated capillaries lined by flat endothelial cells in the papillary dermis.[25]

GI angiodysplasia is a very common cause of digestive hemorrhage among patients with chronic renal insufficiency.[26]

About two thirds of patients with systemic sclerosis who have gastric antral vascular ectasia (GAVE) have diffuse cutaneous systemic sclerosis and the remainder have limited cutaneous systemic sclerosis. The mean disease duration at diagnosis with GAVE was 22 months for diffuse systemic sclerosis and 84 months for limited cutaneous systemic sclerosis (p = 0.025); diffuse cutaneous systemic sclerosis is associated with earlier development of GAVE, as well as more severe anemia requiring more therapeutic interventions.[27]

GAVE appears to be related to autoimmune disorders or to hepatic cirrhosis, whereas radiation proctitis is the result of pelvic irradiation, most commonly used for the treatment of pelvic malignancies. Argon plasma photocoagulation (APC) is the most commonly used endoscopic modality in the treatment of GAVE and radiation proctitis.[28] GAVE is often associated with systemic illnesses, such as cirrhosis of the liver, autoimmune connective tissue disorders, bone marrow transplantation, and chronic renal failure.[29]


United States statistics

The incidence per 100,000 person-years of hospitalizations due to upper GI ulcer bleeding and perforation has decreased over time, from 54.6 and 3.9 in 1996 (R² = 0.944) to 25.8 and 2.9 in 2005 (R² = 0.410), respectively. In contrast, the incidence of colonic diverticular and angiodysplasia bleeding per 100 000 person-years has increased over time from 3.3 and 0.9 in 1996 (R² = 0.443) to 8 and 2.6 in 2005 (R² = 0.715), respectively. Recent recorded drug intake showed an increased frequency of anticoagulants with colonic diverticular and angiodysplasia bleeding, whereas NSAID and low-dose aspirin use were more prevalent in peptic ulcer bleeding and colonic diverticular bleeding, respectively.[30]

The prevalence of angiodysplasia is 0.8% in healthy patients older than 50 years who are undergoing screening colonoscopy.

Foutch et al noted the prevalence of angiodysplasia to be 0.83% from 3 prospective studies in which screening colonoscopies were performed in 964 asymptomatic individuals (mean age, 62 y).[31]

Angiodysplasia is the most common reason (50%) for occult GI bleeding. The pooled completion rate was 84%. The pooled retention rates were approximately 2%.[32]

Angiodysplasia accounts for 20%-30% of GI bleeding episodes in patients with end-stage renal disease[33] and up to 50% of recurrent GI bleeding in this patient population.[34]

Patients with von Willebrand disease may have an increased incidence of GI bleeding from colonic angiodysplasia.[35, 36, 37, 38, 39, 40, 41]

International statistics

No widespread studies to determine the international incidence of angiodysplasia have been conducted, but the incidence probably is similar to that in the United States.

Colonic angiodysplasia in Japanese patients is predominantly located in the left colon, whereas in Western patients it is mainly located in the right colon. The percentage of colonic lesions with a size of more than 5 mm or elevated type detected in Japanese patients was significantly higher than in Western patients.[42] Two cases have been reported in Nigeria.[43]

Race-, sex-, and age-related demographics

No racial predilection exists in cases of angiodysplasia of the colon.

Angiodysplasia of the colon occurs with equal frequency in men and women.

Most patients found to have angiodysplasia are older than 60 years; of these patients, most are older than 70 years. However, case reports exist of occurrence in young people.[44]


The prognosis in patients with angiodysplasia is favorable because most angiodysplasias spontaneously cease bleeding (90% of cases).

Richter et al reviewed the clinical course of 101 patients with colonic angiodysplasia.[45] The cases of 15 asymptomatic individuals who had never bled were followed for as long as 68 months (mean, 23 mo), and no patient experienced bleeding during this observation period.[45] Therefore, one should conservatively manage nonbleeding angiodysplasia that is discovered as an incidental finding. Thirty-one patients with overt bleeding or anemia managed with blood transfusions alone had rebleeding rates at 1 year of 26% and 3 years of 46%. The high rate of rebleeding justifies treatment for angiodysplasia in symptomatic individuals.

Rebleeding after hemicolectomies occurs in 5%-30% of patients, which is much less than that of endoscopic techniques.

Individuals with angiodysplasia lesions longer than 10 mm have higher transfusion requirements, a higher proportion of therapeutic procedures performed after CE, lower hemoglobin concentration, and a lower rebleeding rate. Patients with 10 or more angiodysplasia lesions had also higher transfusion requirements and lower hemoglobin levels, but no differences in the number of therapeutic procedures or rebleeding rate were noted between groups. Angiodysplasia with a size of 10 mm or more suggests a worse clinical impact and a greater possibility of receiving a therapeutic procedure.[46]


Bleeding from angiodysplasia is usually self-limited, but it can be chronic, recurrent, or even acute and life threatening. Approximately 90% of bleeding angiodysplasias spontaneously cease bleeding, presumably because of its venous nature.

Mortality is related to the severity of bleeding, hemodynamic instability, age, and the presence of comorbid medical conditions.


Hemodynamic instability may result from massive bleeding.

Patient Education

If angiodysplasia is identified incidentally, most patients can be reassured because most remain asymptomatic.

Preventive treatment with endoscopic obliteration should be decided on a patient-to-patient basis and should not be done routinely.

For patient education resources, see the Digestive Disorders Center, as well as Gastrointestinal Bleeding (GI Bleeding), Rectal Bleeding, Diverticulosis and Diverticulitis, and Irritable Bowel Syndrome (IBS).




Many patients with angiodysplasia are asymptomatic, and the lesions are found incidentally, such as with screening colonoscopy. Clinical presentation and physical examination are related to GI bleeding or its consequences.

The estimated incidence of active GI bleed in patients with angiodysplasia is less than 10%. However, because these lesions may be located throughout the GI tract and because the rate of bleeding may be variable, the clinical presentation ranges from hematemesis or hematochezia to occult anemia. Bleeding is usually chronic or recurrent and, in most cases, low grade and painless because of the venous source.

Angiodysplastic lesions are often present in more than one location within the GI tract, and the presentation may vary during a patient's clinical course.

GI bleeding from small bowel lesions has occurred in as many as 22% of patients in whom angiodysplasia of the colon was the presumed index source of bleeding.

In 40%-60% of patients with gastric and duodenal angiodysplasia, multiple lesions are observed at endoscopy. Colonic lesions will be associated in 15%-20% of these patients. In addition, angiodysplastic lesions in the colon are more frequently multiple than single. To diagnose and treat patients with suspected angiodysplasia, the diffuse location of lesions and the propensity for multiplicity must be considered. A possible association of true colonic diverticula and angiodysplasia has been proposed and should be kept in mind.

Hematemesis and melena are frequently observed in patients with angiodysplasia of the upper GI tract. Presentation with hemodynamically well-compensated, chronic bleeding is typical and often suggests the diagnosis. Patients with upper tract lesions may have had bleeding from days to years.

Bleeding from colonic lesions is most often chronic and low grade, but as many as 15% of patients present with acute massive hemorrhage. Patients with colonic angiodysplasia may present with hematochezia (0%-60%), melena (0%-26%), hemoccult positive stool (4%-47%), or iron deficiency anemia (0%-51%).

Melena occurs in at least one fourth of patients with colonic bleeding.

Spontaneous cessation of bleeding (occurring in 90% of patients) is the rule for angiodysplastic lesions located in any part of the GI tract.

Physical Examination

Physical examination in a patient suspected of having angiodysplasia should include assessment of their hemodynamic stability and the likely origin of the blood loss. Note the following:

  • Extracolonic angiodysplasias occur in 17% of persons with colonic lesions.[47]

  • Vital signs may demonstrate tachycardia, hypotension, and postural changes based on the amount of blood loss.

  • Stool is typically guaiac positive. Because bleeding may be intermittent, alternating positive and negative guaiac stools can be found.

  • In most cases, bleeding presents as bright red blood, but it can also be maroon in color or melena.

  • A microcytic hypochromic anemia, reflecting iron deficiency, is observed in 10%-15% of cases.

  • Hemodynamic instability may occur if bleeding is massive. This is observed in 15% of cases.





Approach Considerations

The best workup approach depends on many factors, including whether the bleeding is occult or overt, the presence of laboratory or clinical signs of severe bleeding, the patient’s age, and the presence of other comorbidities.

The workup usually starts with endoscopy (colonoscopy, with or without upper endoscopy, and wireless video capsule endoscopy). Angiography (usually preceded by a bleeding scan to guide angiography), deep small bowel enteroscopy, or even intraoperative enteroscopy might be warranted if the bleeding is massive, obscure, and ongoing.

Laboratory Studies

The following laboratory tests are used in the evaluation of patients with suspected angiodysplasia of the colon:

  • Complete blood cell (CBC) count: Approximately 10% of patients who bleed from angiodysplasia present with anemia.

  • Serum iron level: Iron deficiency is found in 10% of patients with bleeding angiodysplasia.

  • Stool for occult blood: As many as 15% of patients with bleeding angiodysplasia will be intermittently positive for occult blood.

  • See the algorithm for acute GI bleeding below.

    Algorithm for acute gastrointestinal (GI) bleeding Algorithm for acute gastrointestinal (GI) bleeding. CBC = complete blood cell count; CXR = chest x-ray; EKG = electrocardiography; IVF = intravenous fluid; NG = nasogastric.

Imaging Studies


Selective mesenteric angiography is a useful diagnostic technique for angiodysplasias, especially in patients with massive bleeding in whom a colonoscopic diagnosis is difficult.

The sensitivity of angiography ranges from 58% to 86%. Detection of bleeding depends on the rate of bleeding (as low as 0.5 mL/min), technique, and timing of angiography in relation to the period of bleeding.

The most frequent and earliest angiographic sign is a densely opacified, dilated, and slowly emptying draining vein within the intestinal wall. This vein is detected during the venous phase of the study and is present in more than 90% of angiodysplastic lesions.

As the lesion progresses, a vascular tuft may become apparent during the arterial phase of the study. This is observed in as many as 70%-80% of patients with angiodysplasia. It represents an extension of the dilation process to the mucosal venules.

The latest sign, an early-filling vein, may be observed in the arterial phase, indicating a more developed arteriovenous communication through the angiodysplastic lesion. It is observed in only 60%-70% of cases of angiodysplasia.

The above three angiographic signs correspond to the development of ectatic changes in the vascular lesions. Their prevalence has been recorded less systematically in other parts of the intestinal tract. However, the interpreter must consider all clinical information, because these angiographic findings may be observed in other disorders, such as malignancy

Demonstration of extravasation of contrast dye in the bowel lumen from angiodysplastic lesions is definite evidence of bleeding; however, this is observed in only 6%-20% of patients. This low percentage is attributed to the episodic nature of the bleeding lesions.

Angiography of resected specimens has been used to confirm appropriate resection when preoperative studies are equivocal or unsatisfactory. An intraluminal formalin fixation technique on the resected specimen, followed by mucosal dissection, has also been used for documenting correct resection. Vasopressin infusion during angiography can arrest bleeding, but potential complications of vasopressin include bowel infarction, arterial vasospasm, and lower extremity ischemia.

Small bowel angiodysplasia accounts for 30%-40% of cases of obscure GI bleeding and is associated with significant morbidity and mortality. Small bowel capsule endoscopy (SBCE) is a significant advance on earlier diagnostic techniques. Lymphangiectasia are seen in about half of those with angiodysplasia, in 19% of those with obscure GI bleeding without angiodysplasia, and in 14% of those without GI bleeding. Lymphangiectasia are strongly associated with the presence of small intestinal angiodysplasia and may represent a useful clinical marker for this condition.[48]

When endoscopy is not available, angiography is a useful diagnostic tool in cases with massive GI bleeding such as angiodysplasia and varicosis. Angiodysplasia and varicosis have distinguishable characteristic features on angiography, such as the presence of a nidus, visible late-draining veins, and the typical vascular tuft. Histological examination may reveal that the lesion is an angiodysplasia and not varicosis.[49]

Nuclear imaging studies

Radionuclide scanning using technetium-99m (99m Tc)–labeled red blood cells or99m Tc sulfur colloid is helpful in detecting and localizing active bleeding from angiodysplasia. Scanning can detect bleeding with rates as low as 0.1 mL/min. The intermittent bleeding nature of angiodysplasia has limited the utility of radionuclide studies in this disorder.

Red blood cells that are labeled with technetium have a long half-life in the intravascular compartment and are especially useful in patients with intermittent hemorrhage. The usefulness of technetium-labeled red blood cells is attributed to the ability to detect bleeding during the 12 or so hours after a single injection of radiolabeled cells.

The reticuloendothelial system rapidly clears99m Tc sulfur colloid.99m Tc sulfur colloid has a half-life of only 3 minutes; hence, it is helpful only in patients with active bleeding. The extravasated labeled sulfur colloid easily demonstrates the site of bleeding in the absence of confusing background activity in the circulation, although uptake into the liver and spleen restricts the image area.

Nuclear scans lack the specificity of an angiogram in differentiating the nature of bleeding lesions, despite the fact that they are noninvasive and relatively easy to perform. Nuclear scans have proven more useful as an adjunct to angiography by localizing and confirming the presence of bleeding, minimizing the number of angiograms that do not yield meaningful diagnostic information, and allowing rapid selection of the artery to be injected by angiography. Confirm positive findings from a radionuclide study by colonoscopy or angiography if surgical resection is contemplated (see image below).

Angiodysplasia identified on the cecum wall during Angiodysplasia identified on the cecum wall during colonoscopy.

Other imaging studies

Helical computed tomography (CT) angiography can detect extravasation from angiodysplasia and is potentially an important noninvasive test in patients with obscure bleeding sites.

Capsule endoscopy has been reported to detect cecal angiodysplasias in selected cases. Capsule endoscopy is particularly useful to demonstrate small intestinal lesions, but its role as a diagnostic test for the colon is still experimental.[50]

Double-balloon endoscopy is useful to detect small bowel lesions, and retrograde double-balloon endoscopy may allow for careful inspection of the cecum and ileocecal valve.[51, 52]

It is recommended not to use air contrast barium enema during acute GI bleeding. Air contrast barium enemas can help exclude other causes of chronic colonic bleeding; however, they are not useful in detecting angiodysplasia, because the lesions are small and usually do not distort the colonic mucosa. In addition, barium enema can obscure other diagnostic studies.



Endoscopy is the most common method of diagnosing angiodysplasia in both the upper and lower GI tract.

Upper endoscopy is used to establish a diagnosis of gastric and duodenal angiodysplasia. Celiac artery and superior mesenteric artery arteriograms frequently fail to demonstrate these lesions, although large lesions with well-formed and enlarged draining veins have been reported in the gastric antrum. Angiography can demonstrate lesions in the more distal small intestine, a region less accessible to endoscopic evaluation. Push enteroscopy has proven to be a successful method of identifying angiodysplasia of the proximal small intestine in patients with obscure bleeding. Other methods of visually evaluating the small bowel include capsule endoscopy and deep or balloon enteroscopy, which are currently used in clinical practice.

The endoscopic appearance of gastric lesions typically has been described as discrete, flat, or slightly raised (2-10 mm) and bright red. These lesions have fernlike margins or stellate configurations. Proximal small intestinal lesions are the size of a pinpoint, with a similar gross appearance. A surrounding pale rim or halo also characterizes upper tract lesions.

Angiography or colonoscopy

Either angiography or colonoscopy may be used to detect colonic lesions. Because colonoscopy is a principal method in the evaluation of GI bleeding, the diagnosis of these lesions often results from colonoscopic examination.

Comparative studies using selective angiography and colonoscopy indicate that the sensitivity of colonoscopy exceeds 80% when the lesions are located in the area examined by colonoscopy. Most angiodysplastic lesions are located in the right colon, thus, the entire colon must be examined. Angiography has the advantage of detecting additional angiodysplastic lesions not depicted by colonoscopy. In a series by Emanuel et al, 17% of subjects were found to have concomitant colonic and extracolonic angiodysplasia when studied by triple-vessel angiography.[47] Angiography and colonoscopy can play important complementary roles.

Skibba et al first described the colonoscopic appearance of angiodysplasia in 1976.[53] Angiodysplastic lesions are often described as discrete and small, with scalloped or frondlike edges and a visible draining vein. They can be flat or slightly raised and can be hidden within the mucosal folds. Although angiodysplasia may be detected anywhere in the colon, a strong propensity exists for the cecum and ascending colon.

Angiodysplastic lesions encountered as incidental findings are generally small lesions with pale coloration compared with lesions with a recent hemorrhage, which are described as extremely bright with elevated centers.

The endoscopic appearance of angiodysplasia can be confused with the ectasias associated with systemic diseases, such as hereditary hemorrhagic telangiectasia (HHT), Turner syndrome, Ehlers-Danlos syndrome, blue rubber web nevus syndrome, the CREST variant syndrome (calcinosis, Raynaud phenomenon, esophageal hypomotility, sclerodactyly, and telangiectasia) of scleroderma, gastric antral vascular ectasia (GAVE), portal hypertensive colopathy, and radiation-induced injury. The lack of systemic manifestations distinguishes angiodysplasia from these syndromes.

In addition, the endoscopic appearance of angiodysplasia may be difficult to discern from spider angiomata, radiation injury, ulcerative colitis (see image below), Crohn disease (see image below), ischemic colitis, and suction artifacts.

Double-contrast barium enema studies in a 44-year- Double-contrast barium enema studies in a 44-year-old man known to have a long history of ulcerative colitis. These images show total colitis and extensive pseudopolyposis.
Double-contrast barium enema examination in a pati Double-contrast barium enema examination in a patient with Crohn colitis demonstrates numerous aphthous ulcers.

Evaluation of a patient who may have angiodysplasia requires that the colonoscope be inserted carefully, with minimal use of suction, and that the mucosa should be examined more carefully as the instrument is advanced than during withdrawal. Obtaining a biopsy of suspicious lesions may be necessary if classic features are not present.

Because blood pressure and volume influence the colonoscopic appearance of vascular lesions, angiodysplasia may not be evident in patients who have bled recently. An accurate evaluation of the colon may not be possible until after effective fluid resuscitation and blood transfusions.

Some studies have noted that administering meperidine may diminish the prominence of vascular abnormalities, and some researchers have advocated reversal by naloxone to aid in the accurate detection of these lesions.

In persons with obscure GI bleeding, the overall diagnostic yield of video capsule endoscopy was 61%. Angiodysplasia was diagnosed in (54%) patients older than 65 years, whereas mucosal breaks, (such as erosions or ulcers), accounted for over a quarter of finding in persons aged 65 years or younger.[54]

The underlying etiologies for lower GI bleeding included diverticular disease (59%), angiodysplasia (19%), ulcers (19%), and malignancy (3%). The use of superselective mesenteric embolization for the treatment of lower GI bleeding is technically successful in 97%, whereas 3% develop postembolization ischemia.[55]

Other procedures

Lesions outside the small bowel as possible sources of obscure GI bleeding are found in about half of patients; a definite source of bleeding outside the small bowel (SB) was detected in about a quarter. Therefore, repeat esophagogastroduodenoscopy (EGD) and ileocolonoscopy should be taken into consideration before video capsule endoscopy or deep enteroscopy.[56]

Despite standard endoscopic diagnostic evaluations, as many as 30% of patients with iron deficiency anemia (IDA) remain without a definitive diagnosis.[57]

In one study, capsule endoscopy (CE) in persons with unexplained IDA revealed angiodysplasias in 51 patients; jejunal and/or ileal micro-ulcerations in 12 patients; tumors/polyps in 9 patients; erosive gastritis in 4 patients; Crohn disease in 5 patients; jejunal villous atrophy in 5 patients; a solitary ileal ulcer in 1 patient, and active bleeding in the last 4 patients. Follow-up data were available for 80 of 91 patients (87.9%). In a third of patients with angiodysplasias, IDA spontaneously resolved without any treatment; about one fifth of the patients required iron supplementation, and one fifth healed after lanreotide (somatostatin analog) administration. APC was performed in one fifth of patients, and the remainder required regular blood transfusions.[58]

Angiodysplasia is the most common reason (50%) for obscure GI bleeding, as shown on CE. The pooled completion rate of CE was 84%. The pooled retention rates were approximately 2%.[59]

Angiodysplasia can be detected infrequently by visual inspection of the serosal side of the bowel during laparotomy. Richardson et al made a correct diagnosis of angiodysplasia during surgical exploration in only 1 of 39 cases (2.6%).[60] The remaining patients were diagnosed by angiography or colonoscopy. Fourteen of these individuals had undergone 21 nondiagnostic surgeries before their evaluation.

Intraoperative enteroscopy can help in the localization of distal small bowel lesions. In addition, an angiographic catheter can be placed before surgery into the appropriate feeding vessel supplying the angiodysplastic lesion. The surgeon then can identify the catheter during surgery and explore and resect the appropriate small bowel segment.

Angiodysplasia (9 mm), GI stromal tumor (15 mm), and AVM may be clearly detected by virtual enteroscopy.

CT enteroclysis is a noninvasive method and useful approach in the diagnosis of small intestinal diseases.

Wireless capsule endoscopy (CE) was first FDA approved to evaluate small intestinal disease in 2001. It can be performed in an ambulatory setting. The most common indication is evaluating obscure GI bleed (OGIB), whether overt or occult.[61]

Deep enteroscopy procedures, whether double-balloon enteroscopy (DBE), single-balloon enteroscopy (SBE), or spiral enteroscopy (SE), are newer, advanced endoscopic procedures that have the ability to access the entire small bowel via both the anterograde and retrograde approach. They have the advantage of being diagnostic and therapeutic in many small intestinal diseases, including angiodysplasia.

A multicenter prospective trial that compared DBE to SBE in 100 patients revealed that DBE had a higher therapeutic yield and total enteroscopy yield with a 3-fold higher rate of complete enteroscopy compared to SBE.[62, 63] A prospective randomized trial of 241 patients that compared DBE and SE showed that both were comparable in terms of diagnostic yield, therapeutic yield, procedure time, and maximal insertion depth.[64] Another study that compared SE and SBE in 92 patients showed that maximal insertion depth was significantly higher for SE than for SBE; however, diagnostic yield and procedure time were comparable.[65]

A likely cause for obscure GI bleeding was found in the small bowel during balloon-assisted enteroscopy (BAE); angiodysplasia or AVM were found in 70% of patients, ulcerative lesions was found in 12% of patients, tumors were found in 5% of patients, and other findings were found in 13% of patients.[66] In 76% of patients with findings at BAE, endoscopic treatment was administered in 71%, but anemia also improved spontaneously in 63% with normal findings during BAE.[66]

To evaluate the diagnostic value of DBE for obscure GI bleeding, one study showed that bleeding was most commonly associated with small bowel tumor (28%), angiodysplasia (19%), and Crohn disease (11%).[67]

The overall complication rate DBE is about 0.9, with no significant difference between age groups. Elderly patients were more likely to have angiectasias (39% vs 23%; P = .01), and were more likely to require endoscopic therapy during DBE (46.8% vs 29.2%; P = .01).[68]

One of the main indications for DBE is suspected small-bowel bleeding (SSBB). The most common finding may be angiodysplasia. The rate of rebleeding in patients who had undergone therapeutic interventions (20%) was similar to that in patients who did not (18%).[69]

The combination of double-balloon enteroscopy (DBE) and laparoscopically assisted bowel surgery (LABS) represents a useful therapeutic method for obscure GI bleeding caused by small bleeding neoplasms or vascular lesions.[70]

No consensus has been reached on the relative accuracy of CE versus DBE to investigate obscure GI bleeding. The positive findings for each type of lesion were similar for angiodysplasia (CE, 65% and DBE, 61%), ulcers (CE, 20% and DBE, 19%), and mass (CE, 10% and DBE, 9%). CE was superior to DBE in detecting bleeding lesions (CE, 43% and DBE, 15%, P=0.0004). DBE provides the advantage of therapeutic intervention by argon plasma coagulation, clipping, polypectomy, tattooing, and biopsies. CE and DBE are complementary, and both evaluate obscure GI bleeding more fully than either modality alone.[71] Three meta-analyses that compared the diagnostic yield of DBE to CE in patients with OGIB showed comparable results. CE is also superior to DBE in being noninvasive and having higher rates of achieving total small intestinal examination.[72]

CE and magnetic resonance enterography (MRE) are complementary methods which, when used in conjunction, may better characterize suspected small bowel disease.[73]

Histologic Findings

Endoscopic forceps biopsy has revealed characteristic histopathologic features of angiodysplasia in only 31%-60% of specimens. Endoscopic mucosal biopsies for the purposes of diagnosis are generally not recommended because of the low diagnostic yield and the risk of provoking hemorrhage.

However, the histologic diagnosis of angiodysplasia is difficult. Acquired lesions such as angiodysplasias must be differentiated from vascular tumors, lesions associated with congenital or systemic disease, or radiation damage.

Angiodysplasias typically are irregularly shaped clusters of ectatic small arteries, small veins, and their capillary connections. They are more often multiple than single. Microscopically, angiodysplastic lesions are dilated, distorted, thin-walled vessels. The amount of smooth muscle in the vessel wall is variable. The vessel wall can become so thinned that it appears to be composed only of endothelium.

Markedly dilated submucosal vasculature is the most consistent abnormality and the earliest change identified. More advanced lesions involve the mucosa. Because the major portion of the lesion is often submucosal, endoscopic mucosal biopsies are often not diagnostic. Characteristic histopathologic findings of angiodysplasia are identified only in 31%-60% of endoscopically obtained biopsies.

In addition, routine pathologic examination usually discovers less than one third of lesions. Injecting the colonic vasculature with silicone rubber and clearing the specimen can be used to identify almost all lesions. In this process, the rubber compound is injected through a catheter placed in one or more of the arteries supplying the colon, after which the specimen is refrigerated for 24 hours to allow the silicone to polymerize. Specimens are then dehydrated in increasing concentrations of ethyl alcohol and cleared with methyl salicylate. The result is a transparent specimen with a filled vascular bed, which is studied through a dissecting microscope using direct light as well as transillumination.

Overlapping features and poor endoscopic-histologic correlation make the distinction challenging between portal hypertensive gastropathy (PHG) and GAVE. Thrombi were revealed in 100% of histologically confirmed cases of GAVE and in 60% of cases suspected of GAVE on endoscopy alone; control biopsies were negative. Mucosal microvessel density (MVD) was significantly higher in GAVE than PHG.[74]

Comparative histological and electron microscopic study of arteriovenous and venous angiodysplasias revealed specific features of their structure, presumably reflecting differences in their morphogenesis. Specific ultrastructural characteristics of angiodysplasias are modified shape of endotheliocytes, impaired structure of the basal membrane, and reduced count of pericytes.[75]



Medical Care

Medically manage each patient with angiodysplasia in accordance with the severity of bleeding, hemodynamic stability, and recurrence of symptoms. A conservative approach to patients who are hemodynamically stable is recommended, because most bleeding angiodysplasias will cease spontaneously. Treatment is usually not advocated for asymptomatic patients when angiodysplasias are found incidentally.

Initially, hemodynamically stabilize all patients with active bleeding with intravenous fluids and packed red blood cells as needed. In addition, correct coagulopathies.

Admit the patient with colonic angiodysplasia to the intensive care unit (ICU) if the patient is hemodynamically unstable. Monitor for recurrent bleeding and stabilization of the hematocrit. Transfuse as needed.

When intervention is warranted, institute steps to control hemorrhage. Endoscopic techniques have been employed most frequently.

Gastric and duodenal angiodysplastic lesions have been managed most commonly with endoscopic obliteration techniques. Rebleeding after these techniques has been attributed to other areas of bleeding angiodysplasia rather than failure of obliteration. These techniques include monopolar electrocautery, heater probe, sclerotherapy, band ligation, and argon and neodymium:yttrium-aluminum-garnet (Nd:YAG) lasers.

Monopolar electrocautery has been used to obliterate angiodysplasia; however, bleeding recurs in approximately 50% of subjects. A reduction in the posttherapy transfusion requirement was not reported to be statistically superior to no therapy. Heater probe or multipolar coagulation devices have more favorable results. Monopolar electrocautery has a higher risk of perforation.

Sclerotherapy using 0.5-1 mL of 1.5% sodium tetradecyl sulfate has been used to obliterate upper tract angiodysplastic lesions; however, bleeding recurs in half of the subjects. In each case, bleeding arose from another area of angiodysplasia. A significant rate of complications has been reported with sclerotherapy, including perforation, embolism, bacteremia, and stricture. The authors do not recommend the use of sclerotherapy for obliteration of colonic angiodysplasia.

Argon plasma coagulation (APC) and Nd:YAG lasers are the most successful endoscopic obliterative techniques for upper tract lesions. APC is a no-touch electrocoagulation technique in which high-frequency alternating current is delivered to the tissue through ionized argon gas. A reduction in both the bleeding rate and transfusion requirement has been demonstrated for at least 12 months after laser therapy. However, active bleeding decreases the ablative efficacy of APC by dissipation of the energy, and APC has been associated with colonic perforation.

GI angiodysplasia (GIAD) may be either asymptomatic, or may induce overt or obscure bleeding with a high risk of recurrence. Endoscopic destruction, preferably using noncontact endoscopic techniques, is most often proposed as a first-line treatment for GIAD. In addition, APC is preferred over Nd:Yag laser due to the lower risk of perforation. Octreotide and estroprogestative treatments are the best evaluated drugs; however, no appropriate comparison on cost-effectiveness and tolerance has been performed.[76]

Submucosal injection of a saline epinephrine solution followed by the application of APC has been reported.[77] Effectiveness appeared to be reduced in patients with more numerous lesions, those with coagulation disorders, and those who are older. Rebleeding commonly occurred over time.[77]

New endoscopic techniques such as the Olympus EVIS LUCERA variable indices of hemoglobin chart function have been developed to assess completeness of vascular mucosal ablation.[78] However, their clinical use is still experimental. Argon plasma coagulation appears the best endoscopic option at the moment to control bleeding in these patients with a low rate of adverse effects and complications and relatively lower costs.

Fifty percent of patients with distal small bowel lesions and no other defined GI bleeding sites have benefited from enteroscopy and lesion obliteration. In one report, blood replacement requirements for a group of 13 patients decreased by more than 50%, comparing the years before and after endoscopic treatment, and 31% required no further transfusion.[79] This group of patients had small bowel angiodysplastic lesions and unexplained bleeding. New endoscopic techniques to examine the small bowel, such as double-balloon enteroscopy, have been developed but are time intensive and operator dependent.

Angiodysplasia of colonic origin has been managed by endoscopic obliteration. Heater probe and multipolar electrocoagulation probe have been more successful than monopolar electrocoagulation. Rebleeding rates for monopolar electrocoagulation are approximately 50%, with the transfusion requirement resembling that of patients receiving no therapy.

Super selective embolization of visceral arterial branches is central to the management of patients with lower GI bleeding, including bleeding from colonic angiodysplasia.[80] Immediate cessation of bleeding was achieved in 97% of patients with injection of microcoils, polyvinyl alcohol particles, gel form, or by selective vasopressin infusion. Postembolization ischemia occurs in 3%, and overall mortality in high-risk patients is 9%.[80, 81] Selective infusion of vasopressin is less effective than embolization as a definitive therapy because of high rebleeding rates associated with its use. Despite the fact that intra-arterial vasopressin can achieve hemostasis for massive lower GI bleeding in 70%-91% of patients, bleeding recurs after discontinuation of vasopressin in 22%-71% of patients.

Endoscopic laser photocoagulation has been successful in controlling bleeding from colonic angiodysplasia, especially right-sided lesions.[77, 82] However, complications occur in as many as 15% of patients and are more common when the Nd:YAG laser is used in the right colon. Complications may be attributed to the deeper coagulation of the vascular abnormalities from laser sources, which incidentally has been responsible for more effective bleeding cessation. Patients with colonic angiodysplasia generally have a 60% chance of remaining free of bleeding at 24 months after laser obliteration.

Emergency embolization with a liquid polyvinyl alcohol copolymer in acute arterial bleeding of the GI tract may be clinically useful.

Endoclips have been used in anecdotal case reports for bleeding angiodysplasia of the cecum and right colon.[80]

Angiodysplasia that presents with acute hemorrhage can be controlled effectively with angiography, although it is seldom needed. Angiography is appropriate in severely ill patients who are not candidates for surgical intervention. In these patients, transcatheter embolization of selected mesenteric arteries has been quite effective. However, the rate of complications is sufficiently high and must be balanced against the risk of surgical resection.

Angiography plays a more important role in the preoperative localization of small bowel lesions immediately before surgical resection, because intraoperative palpation, endoscopy, and visual inspection through multiple enterotomies are of little value with angiodysplasia.

Injection of dyes, such as methylene blue, indigo carmine, and fluorescein, has been used to assist in the localization of angiodysplasia before surgical resection.

Patients with systemic sclerosis with severe GAVE refractory to laser ablation may show clinical and endoscopic improvement following intravenous (IV) pulse cyclophosphamide (CYC) treatment.[83] APC is an effective and safe endoscopic treatment for GAVE in patients with liver cirrhosis.[84]

Traditionally, gastroenterologists prefer to use endoscopic modalities like argon plasma coagulation and electrocoagulation to treat accessible vascular lesions. Prospective studies involving 10 or more patients were included in one analysis that showed a clinical response to treatment of 0.76 (95% CI, 0.64-0.85). The weighted mean difference in transfusion requirements before starting therapy (control group) and after treatment initiation (treatment group) was -2.2 (95% CI, -3.9 to -0.5).[85]

Recurrence of acute hemorrhage from GI angiodysplasia after hospital discharge occurred in 30% of patients after a mean follow-up of 33 (±40) months. In a multivariate analysis, earlier history of bleeding with a high bleeding rate, over anticoagulation, and the presence of multiple lesions were predictive factors of recurrence. Surprisingly endoscopic APC therapy was not associated with lower rates of recurrent bleeding.[86]

Different types of gastric vascular ectasia include focal vascular ectasia, portal hypertensive gastropathy, and GAVE. Endoscopic thermal ablation with APC is effective in managing upper GI bleeding.[87]

Endoscopic treatment by APC was successful in about 80% of patients with bleeding gastric vascular ectasia with or without cirrhosis. Noncirrhotic patients required significantly more APC sessions to achieve complete treatment.[88]

Treatment of GAVE, characterized by mucosal and submucosal vascular ectasia causing recurrent GI hemorrhage, with endoscopic thermal therapy (ETT) requires multiple sessions for destruction of vascular ectasia and control of bleeding. Compared with ETT, EBL has a significantly higher rate of bleeding cessation (67% vs 23%, P = .04), fewer treatment sessions required for cessation of bleeding (1.9 vs 4.7, P = .05), a greater increase in hemoglobin values (2.8 g/dL vs 0.9 g/dL, P = .05), a greater decrease in transfusion requirements (-12.7 vs -5.2, P = .02), and a greater decrease in hospital admissions (-2.6 vs -0.5, P = .02) during the follow-up period.[89]

The role of somatostatin analogues for refractory bleeding GI angiodysplasias has been systematically reviewed by Brown et al.[90]

Treatment with continuous octreotide LAR 20mg once a month, reduces transfusion requirements in persons with diffuse small bowel angiodysplasia.[91]

The use of short-acting SMS analogues is recognized in acute variceal hemorrhage secondary to portal hypertension in cirrhosis. Long-acting SMS analogue therapy has been successfully used in obscure GI bleeding thought to be secondary to angiodysplasia.[92]

Rarely, thalidomide may be used in persons with GI angiodysplasia, which has shown no response to standard treatment.[93]

Thalidomide should be considered as a therapeutic option in patients who are resistant to conventional therapy for chronic angiodysplasia bleeding requiring ongoing transfusions. It has a high discontinuation rate because of its side-effects.[94] Thalidomide, with its antiangiogenic mechanism of action, is a promising drug in bleeding angiodysplasias as a treatment option for patients unable to benefit from other available modalities of treatment.[95]


Consultations should be arranged with a gastroenterologist, an interventional radiologist, and a surgeon.

Diet and activity

Withhold oral intake until the diagnosis has been made and treatment has been initiated.

Restrict activity until hemodynamic stability can be maintained.


No preventive methods for angiodysplasia have been definitely identified at this time. Avoidance of nonsteroidal anti-inflammatory drugs (NSAIDs) is recommended in patients with chronic bleeding.

Surgical Care

Surgical resection is the definitive treatment for angiodysplasia.

Partial or complete gastrectomy for the management of gastric angiodysplasia has been reported to be followed by bleeding in as many as 50% of patients. Rebleeding was attributed to other angiodysplastic lesions.

Right hemicolectomy for angiodysplasia is a second-line therapy after endoscopic ablation, if repeated endoscopic coagulation has failed, if endoscopic therapies are not available, and for life-threatening hemorrhage.

The mortality rate associated with surgical resection ranges from 10% to 50%. This is based on the view that surgery carries a much higher risk in elderly patients, who often have multiple coexisting medical problems, including coronary artery disease, coagulopathy, and renal and pulmonary dysfunction.

In a study by Meyer et al, right hemicolectomy resulted in 63% of the subjects remaining free of intestinal bleeding (mean follow-up, 3.6 y), and 37% had some degree of recurrent bleeding.[96]

Trends toward reduced transfusion requirements have been observed after surgical resection, as well as after electrocoagulation as the only mode of therapy, and in patients who received no specific intervention.

Surgical resection is preferred for the acute management of severe hemorrhage or for the management of recurrent hemorrhage over a relatively short period accompanied by a large transfusion requirement.

Long-Term Monitoring

The exact time frame for follow-up colonoscopy in patients with angiodysplasia is controversial. If the patient is asymptomatic, a repeat colonoscopy is not recommended. Outpatient monitoring of hemoglobin and repeated tests for occult blood can be performed. Patients with chronic GI bleeding may need repeated colonoscopies.



Medication Summary

Medical treatment has been used in active and recurrent bleeding from colonic angiodysplasia with controversial results. Hormonal treatment with estrogen and progesterone has been evaluated by randomized trials but remains controversial and is probably not effective. Octreotide, both short and long acting, has been described as effective in a few case reports and case series only. Other agents, such as thalidomide, remain experimental. Desmopressin (DDAVP) has also been used in specific subsets of patients. At the moment, no medical therapy has been proven to effectively treat bleeding from angiodysplasia. A recent meta-analysis suggested that hormonal therapy was not effective in bleeding cessation; however, somatostatin analogs could be effective therapy for gastrointestinal angiodysplasias.[97]

Oral Contraceptives

Class Summary

Only use hormonal therapy for the small subset of patients who are transfusion-dependent from bleeding angiodysplasia refractory to conservative and endoscopic therapy and who are poor surgical candidates. This therapy is not for routine management of bleeding angiodysplasia. No large-scale, randomized, double-blinded studies have demonstrated its effectiveness.

Estrogen-progesterone therapy, previously used to treat bleeding associated with hereditary hemorrhagic telangiectasia (HHT), also has been tried in patients with GI bleeding from angiodysplasia.

The proposed mechanisms by which hormonal therapy might affect bleeding include improvement in coagulation, alterations in microvascular circulation, and improvements in endothelial integrity.

Data from a double-blinded, crossover trial using 0.05 mg ethinyl estradiol and 1 mg norethisterone administered daily to 10 elderly patients with GI ectasia (6 of the patients had HHT) indicated that the combination significantly reduced bleeding and transfusion requirements. Several other small series with anecdotal success have been described, but one must be skeptical.

A retrospective cohort study of 64 patients by Lewis et al refutes the benefits of hormonal therapy in angiodysplasia.[98] Thirty patients were administered 5-10 mg of norethynodrel with mestranol (0.075-0.15 mg) or with conjugated estrogens (0.625 mg); the bleeding rates did not differ before and after therapy, and they did not differ from bleeding rates of historical controls or from patients who refused therapy.[98] Treatment adverse effects in this study included vaginal bleeding, fluid retention, and stroke (23% of the treated patients).

Overall, the current data do not support the use of hormonal therapy in patients with colonic angiodysplasia.

Ethinyl estradiol and norethindrone (Ovcon 50)

Suggested mechanisms by which hormonal therapy might affect bleeding include improvement in coagulation, alterations in the microvascular circulation, and improvements in endothelial integrity. One tab contains ethinyl estradiol 0.05 mg and norethindrone 1 mg.

Somatostatin Analogues

Class Summary

Somatostatin analogues have been reported to decrease the rate of bleeding from intestinal angiodysplasia. In our experience, these agents are usually well tolerated and may decrease the rate of chronic bleeding. Octreotide should be the first choice in patients with portal hypertension.

Octreotide (Sandostatin)

Mechanism of action in this setting is not fully understood. Used in acute variceal bleeding and for recurrent bleeding after endoscopic therapy.

May reduce the transfusion requirement.