Obscure gastrointestinal (GI) bleeding accounts for approximately 5% of all GI bleeding and is defined as bleeding from an unknown source that persists or recurs after negative endoscopic diagnostic evaluation.¹ A negative diagnostic evaluation is commonly agreed on as consisting of negative upper endoscopy (esophagogastroduodenoscopy [EGD]) and colonoscopy with careful evaluation of the terminal ileum. Obscure GI bleeding is subcategorized as overt or occult. Obscure-overt GI bleeding is persistent or recurrent visible evidence of bleeding (hematemesis, hematochezia, or melena), whereas obscure-occult GI bleeding is persistent or recurrent positive fecal occult blood, iron deficiency anemia, or both, when there is no evidence of visible blood loss to the patient or health care provider.²

Etiology of Obscure Gastrointestinal Bleeding

Obscure GI bleeding has numerous possible causes and may originate from the upper GI tract (proximal to the ligament of Treitz), the mid-GI tract (ligament of Treitz to the terminal ileum), or the colon. Most causes (approximately 75%) of obscure GI bleeding are found between the ligament of Treitz and the ileocecal valve—in other words, in the small bowel.³ Arteriovenous malformations (AVMs) of the small bowel account for 30% to 40% of obscure GI bleeding and are the most common source of obscure GI bleeding in older patients (Fig. 15.1).⁴ In persons 30 to 50 years of age, tumors such as GI stromal tumors, leiomyomas, leiomyosarcomas, schwannomas, carcinoids, lymphomas, and adenocarcinomas predominate (Fig. 15.2). Younger individuals most commonly have obscure bleeding from Crohn’s disease or Meckel’s diverticula–associated ulceration (Fig. 15.3).⁵ Nonsteroidal antiinflammatory drug (NSAID) enteropathy has been associated with erosions, ulcers, and strictures of the small bowel and can also be a potential cause of obscure GI bleeding.⁶ Less common causes of obscure GI bleeding include Dieulafoy’s lesions, hemosuccus pancreaticus, Strongyloides stercoralis infection, radiation-induced enteritis, and pseudoxanthoma elasticum (Table 15.1).⁷^–¹¹

Fig. 15.1 Arteriovenous malformation of the small bowel as seen by capsule endoscopy.

Fig. 15.2 Gastrointestinal stromal tumor as seen by double-balloon enteroscopy.

(Courtesy of Professor Roberto De Franchis.)

Fig. 15.3 Small bowel ulceration in Crohn’s disease as seen by capsule endoscopy.

Table 15.1 Etiology of Obscure Gastrointestinal (GI) Bleeding

UPPER GI LESIONS

Gastric antral vascular ectasia

MID–GI TRACT LESIONS Younger than 40 Years of Age

Meckel’s diverticulum

Older than 40 Years of Age

Hemosuccus pancreaticus

Aortoenteric fistula

LOWER GI LESIONS

Angiectasis

Neoplasms

NSAID, Nonsteroidal antiinflammatory drug.

Investigating Obscure Gastrointestinal Bleeding

Medical history, patient age, symptoms, physical examination findings, and laboratory data all may provide clues and help guide diagnostic investigations. Several elements of the medical history and physical examination can provide important information about the etiology of obscure bleeding and help define the intensity with which a bleeding site should be sought. Recurrent hematemesis indicates bleeding proximal to the ligament of Treitz, whereas recurrent passage of hematochezia suggests a colonic source. Melena can originate from bleeding anywhere from proximal to the ligament of Treitz to the right colon. As a result, a history of melena provides only limited value in terms of localization of obscure GI bleeding.

A review of medications (including over-the-counter medications) may reveal inadvertent use of NSAIDs or products containing acetylsalicylic acid (ASA) and NSAIDs. A family history of cancer occurring at an early age, particularly colorectal or endometrial, may suggest the presence of hereditary nonpolyposis colorectal cancer. Skin, nail, and oral mucosal changes may suggest the presence of several disorders associated with obscure GI bleeding or iron deficiency anemia, including telangiectasias, which may reflect hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu syndrome); dermatitis herpetiformis, which may reflect celiac disease; or other conditions with cutaneous and GI manifestations (e.g., Kaposi’s sarcoma, Peutz-Jeghers syndrome, tylosis, pseudoxanthoma elasticum, Ehlers-Danlos syndrome, blue rubber bleb nevus syndrome, Henoch-Schönlein purpura, neurofibromatosis, malignant atrophic papulosis, and Klippel-Trenaunay-Weber syndrome).²

Diagnostic Investigations

Numerous diagnostic investigations may be performed to identify the etiology of obscure GI bleeding. Endoscopic investigations include repeat EGD and colonoscopy, push enteroscopy, capsule endoscopy (CE), deep enteroscopy using balloon-assisted or non–balloon-assisted techniques, and intraoperative enteroscopy (IOE). Radiologic testing includes barium studies such as small bowel follow-through and enteroclysis, nuclear medicine studies such as tagged red blood cell (RBC) scan and Meckel’s scan, computed tomography (CT) and magnetic resonance (MR) imaging in combination with enteroclysis (CT and MR enterography), and angiography (Fig. 15.4).

Fig. 15.4 Proposed algorithm for diagnosis and management of obscure gastrointestinal bleeding. CE, Capsule endoscopy; PE, push enteroscopy.

(Adapted from Pennazio M, Eisen G, Goldfarb N: ICCE consensus for obscure gastrointestinal bleeding. Endoscopy 37:1046–1050, 2005.)

Endoscopic Investigations

Repeat Upper Endoscopy (Esophagogastroduodenoscopy) and Colonoscopy

Second-look upper endoscopy may be helpful in identifying lesions potentially overlooked or unrecognized at the time of the initial endoscopic evaluation.¹² Data suggest that 20% of lesions leading to obscure GI bleeding may be overlooked and are within reach of a standard upper endoscope.¹³ Commonly overlooked lesions in the upper GI tract include Cameron’s erosions or ulcerations in large hiatal hernias, isolated gastric fundal varices, peptic ulcers, AVMs including gastric antral vascular ectasia (watermelon stomach), and Dieulafoy’s lesion (Fig. 15.5).¹⁴^–¹⁶ An aortoenteric fistula should be considered in patients with prior abdominal aortic aneurysm repair. Although the reported yield of repeat colonoscopy is low (3% to 6%), it may be helpful in selected patients when the original colonoscopy examination was documented to have had a mediocre or poor preparation, the extent of the examination was not to the cecum, or the terminal ileum was not evaluated.¹⁷ Lesions missed or unrecognized during colonoscopy may include AVMs, polyps, solitary rectal ulcers, rectal varices, and neoplasms.

Fig. 15.5 Gastric antral vascular ectasia, also referred to as watermelon stomach.

Push Enteroscopy

Push enteroscopy permits evaluation of the proximal small intestine to a distance that is approximately 50 to 100 cm beyond the ligament of Treitz. Dedicated videoenteroscopes (160 to 250 cm in length) are commercially available, but if these instruments are not available at the endoscopy site, a pediatric or standard adult colonoscope can be used instead.¹⁸ The use of an overtube, back-loaded onto the endoscope insertion tube, may help limit looping of the enteroscope within the stomach and facilitate deeper small bowel intubation.¹⁹ The diagnostic yield of push enteroscopy is reported to increase with greater depth of scope insertion. When using a pediatric or adult colonoscope, the reported diagnostic yield in the evaluation of obscure GI bleeding ranges from 13% to 38%. With the use of a dedicated videoenteroscope, reported diagnostic yield rates increase to 26% to 80%.^13,^20,²¹

With the development of endoscopic devices for dedicated videoenteroscopes, such as biopsy forceps, snares, thermal probes (contact and noncontact), and injection needles, push enteroscopy is preferred over radiologic diagnostic modalities because of the ability to obtain tissue, perform polypectomy or hemostasis if necessary, and mark lesion sites with India ink tattoo.¹³ However, push enteroscopy does not allow for the visualization of the entire small bowel, and complications, including perforation and mucosal laceration, have been reported with the use of an overtube. With the advent of CE and “deep” enteroscopy, diagnostic push enteroscopy is less commonly used.

Capsule Endoscopy

CE is an endoscopic technology that is capable of obtaining endoscopic images from the entire small bowel.^22,²³ CE is safe, easy, minimally invasive, and patient-friendly and has become a first-line tool in imaging small bowel pathologies. With this realization, there has been rapid uptake and wide acceptance of this revolutionary endoscopic technology for detecting small bowel abnormalities.²⁴

The Pillcam SB video capsule endoscope (Given Imaging Ltd, Yoqneam, Israel) is a wireless capsule (11 mm × 26 mm) composed of a light source, lens, complementary metal oxide semiconductor imager, battery, and wireless transmitter. The Pillcam SB has a battery life of approximately 7 to 8 hours in which time the capsule captures two images per second (approximately 60,000 total images per examination) in a 140-degree field of view and 8 : 1 magnification.²⁵ The smooth outer coating of the capsule allows easy ingestion and prevents adhesion of intestinal contents, whereas the capsule moves via natural peristalsis from the mouth to the anus. Endoscopic images are transmitted via sensor arrays to a recording device worn as a belt by the patient. The recorded images are downloaded into a Reporting and Processing of Images and Data (RAPID) computer workstation and reviewed as a continuous video by the physician. The PillCam SB 2 video capsule offers advanced optics and a wider field of view for imaging the small bowel. It has the same dimensions as the PillCam SB and captures nearly twice the mucosal area per image compared with the PillCam SB. The PillCam SB 2 also provides Automatic Light Control for optimal illumination of each image.

Olympus (Olympus Corporation, Tokyo, Japan) has introduced a wireless capsule endoscope (EndoCapsule) with similar features. The Pillcam SB, Pillcam SB 2, and EndoCapsule have been approved by the U.S. Food and Drug Administration (FDA) (Fig. 15.6). Other small bowel capsule endoscopes available in the marketplace but not yet approved by the FDA are the OMOM pill (Jinshan Science & Technology, Chongqing, China) and the MiroCam (Intromedic, Seoul, Korea).

Fig. 15.6 A, Capsule endoscope (Pillcam SB 2). B, Capsule endoscope (Olympus EndoCapsule).

(A, Courtesy of Given Imaging Ltd, Yoqneam, Israel; B, courtesy of Olympus Corporation, Tokyo, Japan.)

Compared with other diagnostic modalities for evaluating obscure GI bleeding (e.g., small bowel series and enteroclysis, push enteroscopy, IOE), CE has been shown to have very good test characteristics and diagnostic yields in the range of 63% to 74%.² Patient selection is important and seems to influence the diagnostic yield of CE. CE performed close to the time of the bleeding event, hemoglobin less than 10 g/dL, recurrent episodes of bleeding, or persistent bleeding (>6 months) may increase CE diagnostic yield.² Limitations of CE include that at the present time biopsy, therapy, and endoscopic marking (e.g., India ink tattooing) are impossible. In addition, not all CE examinations reach the cecum (complete small bowel examination occurs approximately 80% to 85% of the time), and in some patients, luminal debris and bubbles interfere with viewing. Some physicians use polyethylene glycol–based preparations, prokinetic agents, simethicone, or a combination of these products before capsule ingestion.²⁶ Other limitations are that the imaging cannot be controlled or localized, and images are not viewed in real time.

Some patients may be unsuitable candidates for CE, including patients with cardiac pacemakers, patients with defibrillators, and patients with suspected small bowel obstruction. However, small case series suggest that CE, when performed with careful patient monitoring, is safe in patients with pacemakers and cardiac defibrillators.² Patients with swallowing disorders may have the capsule placed endoscopically into the duodenum using a capsule delivery device. Capsule retention is the major, and for all practical purposes, the only complication of CE. The reported incidence of capsule retention ranges from 0% in healthy volunteers, to 1.4% in patients with obscure GI hemorrhage, to 21% in patients with suspected small bowel obstruction.²⁷

Double-Balloon Enteroscopy

Double-balloon enteroscopy (DBE), initially described and reported by Yamamoto and colleagues,²⁸ is a novel endoscopic insertion technique that attempts to improve on currently available endoscopic insertion methods to evaluate the entire length of the small bowel. The DBE system (Fujinon Inc, Saitama, Japan), uses a high-resolution, dedicated videoendoscope that has a working length of 200 cm and two soft, latex balloons: One balloon is attached to the tip of the endoscope, and the other is attached to the distal end of a soft, flexible overtube (Fig. 15.7). The balloons can be inflated and deflated using an air pump that is controlled by the endoscopist while monitoring air pressure.²⁹ The balloons grip the wall of the bowel, allowing the endoscope to be advanced without looping.

Fig. 15.7 Double-balloon enteroscope.

(Courtesy of Fujinon Inc, Saitama, Japan.)

The procedure can be performed via an oral or anal approach with or without fluoroscopic guidance. Choice of oral or transanal approach may be dictated by suspicion for the location of a possible lesion as determined by preceding small bowel CE or other nonendoscopic small bowel imaging technique.^30,³¹ In a peroral approach, when the two balloons reach the duodenum, the overtube balloon is inflated to fix the overtube to the small bowel wall. The overtube is held in place as the endoscope is inserted further. Once the tip of the endoscope is maximally inserted, the balloon on the tip of the endoscope is inflated, the balloon on the overtube is deflated, and the overtube is advanced over the shaft of the endoscope. When the distal end of the overtube reaches the tip of the endoscope, the overtube balloon is reinflated, again fixing the overtube to a second point on the small bowel wall. This sequence is repeated until the entire small bowel is evaluated or further advancement of the endoscope is difficult.³²

Two types of double-balloon enteroscopes are available: one for general diagnostic use (EN-450P5) and one for therapeutic use (EN-450T5). The EN-450P5 is a thinner endoscope with an external diameter of 8.5 mm and maximum working channel diameter of 2.2 mm. The EN450T5 has an external diameter of 9.4 mm and working channel of 2.8 mm.²⁹ DBE has been shown to be able to visualize the entire length of the small bowel and allows for biopsy, marking of lesions for subsequent surgical resection (e.g., India ink tattoo), and therapeutics.

Published studies evaluating DBE for obscure GI bleeding have reported diagnostic yields ranging from 40% to 80% and diagnostic or treatment success in 43% to 76% of cases.² Limitations of DBE include concerns about the endoscopic learning curve, common need for endoscopy on two separate days (peroral followed by transanal approach), limitations in visualization of the entire small bowel, miss rates for submucosal lesions owing to insufflation issues, time-consuming procedure that also requires a high level of ancillary staffing, increased sedation requirements (including possible anesthesiology support), and patient tolerance and preferences. Although initial reports of complete evaluation of the small bowel using DBE were encouraging, subsequent studies have failed to show similar results.² In addition, although uncommon, the reported incidence of severe complications associated with DBE has ranged from 0% to 2.5% and has included pancreatitis, perforations, bleeding, abdominal pain, and fever.²⁹

Single-Balloon Enteroscopy

A novel balloon enteroscope system has been developed more recently using only a single balloon (single-balloon enteroscopy [SBE]) (Fig. 15.8). The endoscopist needs to manipulate only one balloon, and time and complexity for preparation of the system and for the examination itself may be reduced.³³ Data for the SBE system to compare its performance and safety profile with DBE are sparse.³⁴ In the limited literature available, SBE seems to provide similar diagnostic and therapeutic yield compared with DBE.^35,³⁶ SBE is performed using the SBE endoscope system (SIF-Q180, Olympus Optical Co, Ltd, Tokyo, Japan). The SBE endoscope consists of a 200-cm long videoendoscope with an outer diameter of 9.2 mm and a flexible overtube with a length of 140 cm and an outer diameter of 13.2 mm. One single silicone balloon is attached to the tip of the overtube. The insertion process follows the method used for DBE, but instead of inflation of the endoscope balloon, the tip of the endoscope may be optionally angulated during the pullback or straightening of the enteroscope and balloon-inflated overtube.³⁷ Similar to DBE, SBE can be performed via an oral or transanal approach, according to suspicion for the location of a possible small bowel lesion.

Fig. 15.8 Single-balloon enteroscope.

(Courtesy of Olympus Optical Co, Ltd, Tokyo, Japan.)

Spiral Enteroscopy

Spiral enteroscopy is an alternative technique for accomplishing deep small bowel intubation that uses a special overtube (Endo-Ease, Discovery Small Bowel (DSB), Spirus Medical) to pleat the small bowel onto itself (Fig. 15.9).³⁸ The DSB overtube can be paired with existing enteroscopes or colonoscopes. Rotation of the DSB beyond the ligament of Treitz allows the manually rotating overtube with an outer fixed spiral coil to pleat the small bowel onto the overtube. Preliminary data suggest that use of the Endo-Ease DSB overtube for enteroscopy is a safe and effective technique for visualization of the small bowel.³⁹ Studies of spiral enteroscopy have shown diagnostic yield, total time of procedure, and depth of small bowel insertion to compare favorably with DBE and SBE.⁴⁰ The strengths of spiral enteroscopy are rapid advancement in the small bowel and controlled, stable withdrawal that facilitates endoscopic therapy.

Fig. 15.9 Spiral enteroscope.

(Courtesy of Spirus Medical.)

Intraoperative Enteroscopy

Exploratory laparotomy with IOE has been used since the 1980s and is a diagnostic and potentially therapeutic endoscopic modality in suspected small bowel disease.^41,⁴² IOE is considered to be the ultimate endoscopic evaluation of the small bowel. In addition to being able to diagnose small bowel sources of obscure bleeding, IOE allows for identification of lesions for definitive surgical resection. Using an orally passed colonoscope or, more optimally, a dedicated small bowel videoenteroscope, the endoscope is either advanced beyond the ligament of Treitz into the proximal jejunum or advanced into a mid–small bowel enterotomy. The latter technique expedites the examination, especially of the distal small bowel, and minimizes the potential for iatrogenic trauma to the small bowel arising from excessive manipulation that can arise from peroral passage. At that point, the surgeon gently telescopes the small bowel over the shaft of the endoscope, allowing for the careful inspection of the mucosa.

Inspection of the small bowel mucosa should be performed in an anterograde manner. When using a mid–small bowel enterotomy, the endoscope is first advanced toward the cecum and then redirected toward the stomach. Dimming the operating room lights facilitates endoscopic visualization and extrinsic inspection of the bowel by the surgeons. Lesions identified endoscopically can be marked for resection by the surgeon with a suture placed on the serosal aspect of the bowel. After completion of the enteroscopy and withdrawal of the endoscope, the marked sites can be surgically resected. An alternative approach to IOE involves insertion of a sterilized endoscope through multiple, surgically created enterotomies, rectal insertion, or laparoscopy-assisted enteroscopy.

Reported complications associated with IOE include mucosal lacerations, perforations, prolonged ileus, abdominal abscess, and bowel ischemia.^41,⁴² Because of its invasive nature and potential for complications, the decision to perform IOE must not be made lightly. All risks and benefits need to be fully considered and explained well to the patient; only experienced endoscopists and surgeons should perform this procedure. Currently, this technique is reserved as a last option if deep enteroscopy using DBE, SBE, or spiral enteroscopy cannot be performed successfully because of the presence of abdominal adhesions, obesity, or other technical factors.

Small Bowel Contrast Radiography

A small bowel series (upper GI series with small bowel follow-through) involves the oral ingestion of a dilute barium solution with serial abdominal images obtained of the small bowel. Enteroclysis is a double-contrast study performed by passing a tube into the proximal small bowel and injecting barium, methylcellulose, and air.⁴³ This technique is considered to be superior to standard small bowel follow-through.⁴⁴ An upper GI series with small bowel follow-through is sometimes used to evaluate obscure GI bleeding either before performing push enteroscopy or after a negative push enteroscopy examination.

The role of small bowel series and enteroclysis in the evaluation of obscure GI bleeding has declined substantially because of low diagnostic yield (0% to 6% for small bowel follow-through and 10% to 25% for enteroclysis) and the advent of improved endoscopic diagnostic techniques for the small bowel.^45,⁴⁶ In addition to limited diagnostic sensitivity in obscure GI bleeding, enteroclysis can be uncomfortable for the patient and involves more radiation exposure than small bowel follow-through. Generally, the role of enteroclysis in the evaluation of obscure GI bleeding is limited to settings in which CE and enteroscopy are unavailable or contraindicated, such as clinical situations that suggest small bowel obstruction secondary to malignancy, Crohn’s disease, or prior significant NSAID use. Otherwise, there is little or no role today for either small bowel series or enteroclysis in the evaluation of obscure GI bleeding.

Cross-Sectional Imaging

Novel cross-sectional imaging techniques for evaluation of the small bowel include helical CT enteroclysis, helical CT angiography, and MR enteroclysis.² Helical CT enteroclysis combines standard enteroclysis to distend the small bowel followed by helical CT.⁴⁷ More recent reports suggest the utility of helical CT in identifying small and large intestinal bleeding angiectasias.^48,⁴⁹ Further research and clinical experience should define the precise role of CT enteroclysis in the investigation of small bowel disease.

CT angiography involves catheterization of the abdominal aorta followed by helical CT angiography before and after intraarterial injections of a contrast medium. The site of hemorrhage is recognized as a hyperdense area owing to the extravasation of contrast medium into the intestinal lumen. In a prospective study of 18 patients with bleeding colonic angiectasis, the sensitivity, specificity, and positive predictive value of helical CT angiography were 70%, 100%, and 100% compared with a “gold standard” of colonoscopy and mesenteric angiography.⁵⁰ In a report of 22 patients with obscure GI bleeding, CT enteroclysis was found to be inferior to CE in the detection of potential bleeding lesions such as angiectasis in the small bowel.⁵¹

Nuclear Medicine Scans

Tagged RBC scans are noninvasive, safe, and readily available. Most commonly performed with technetium-labeled RBCs, a radionuclide bleeding scan detects bleeding that is occurring at a rate of 0.1 to 0.5 mL/min. It is of little or no value in patients with obscure-occult bleeding who appear to have a low rate of blood loss because the rate of bleeding is too slow to be detected. Bleeding scans may be more sensitive than angiography but less specific than either a positive endoscopic or angiographic examination.^52,⁵³ Early scans (up to 4 hours after initial injection) may be helpful in gross localization of bleeding when the rate of blood loss exceeds 0.1 to 0.5 mL/min. Tagged RBC scans can identify only a general area of bleeding, however, and assessment with other diagnostic modalities such as angiography is often necessary after nuclear scans. The role of nuclear scans continues to be limited in patients with obscure GI bleeding, and the ability to localize the source of bleeding accurately is poor.⁵⁴

Angiography

The data on the clinical utility of angiography in the setting of obscure GI bleeding are very limited.⁵⁵ Compared with tagged RBC scans, mesenteric angiography is more likely to document the specific site of bleeding, yet the rate of bleeding must be greater than 0.5 mL/min. Angiography can also identify lesions that are not actively bleeding because of demonstration of typical vascular features seen in vascular ectasia (e.g., slow-emptying veins, vascular tuft, and early-filling veins) and tumors. It is possible to administer embolization therapy if an amenable lesion is detected.

Provocation angiography using anticoagulation, vasodilators, or thrombolytic agents may increase the likelihood that a source of bleeding can be identified and has been advocated by some investigators.^56,⁵⁷ However, the risk of inducing uncontrolled bleeding limits the use of this technique. As noted with tagged RBC scans, the utility of mesenteric angiography in the evaluation of patients with obscure GI bleeding is limited, and it is not commonly recommended.

Proposed Diagnostic Strategy

In a patient with obscure GI bleeding, defined as careful, quality endoscopic examinations (EGD plus ileocolonoscopy) being negative, the small bowel should be assumed to be the source of blood loss. Unless contraindicated, CE should be the next method of small bowel evaluation (see Fig. 15.4).^2,⁵⁸ In patients with obscure overt bleeding, data support the role of expedited CE.^12,⁵⁹ CE performed closer to the episode of obscure overt bleeding has a higher likelihood of finding the site of small bowel bleeding.^12,⁵⁹ If CE defines a small bowel source of obscure bleeding, appropriate specific therapy may be instituted (e.g., endoscopic, surgical, or pharmacologic). If no small bowel source is identified, a decision must be made as to whether additional diagnostic evaluations are immediately indicated (e.g., repeat EGD with or without ileocolonoscopy, repeat CE, Meckel’s scan in pediatric and young adult patients) or a period of “watchful waiting” may be warranted.

Pharmacologic Therapy

Pharmacotherapy should be considered whenever endoscopic therapy, surgical intervention, or angiographic therapy is impractical or ineffective, such as in patients in whom the source and the etiology of bleeding remains undefined or the underlying bleeding pathology is too diffuse to be amenable to ablative therapies. The role of hormonal therapy continues to be most debated.^60,⁶¹ Somatostatin and its analogue octreotide have been anecdotally reported to be beneficial in patients with obscure GI bleeding from angiectasis and blue rubber bleb nevus syndrome, possibly secondary to their inhibitory effect on angiogenesis and splanchnic blood flow.⁶²^–⁶⁴ A beneficial effect of thalidomide and its antiangiogenic effects in patients with obscure GI bleeding secondary to angiectasis and hereditary hemorrhagic telangiectasia has also been reported.^65,⁶⁶ In a case report, erythropoietin was believed to stop chronic diffuse hemorrhage from the GI mucosa. The exact mechanism of erythropoietin is unknown but may be related to the complex effect of erythropoietin on the platelet-subendothelial interactions and on protein C, protein S, and anti–thrombin III levels.⁶⁷