Diagnosis



The first step in the diagnosis of the cause of gastrointestinal bleeding is to differentiate be­tween upper and lower bleeding. The patient’s history may be of considerable value in this dif­ferentiation. Hematemesis, for example, suggests bleeding proximal to the ligament of Treitz, and recent ingestion of anti-inflammatory drugs or al­cohol suggests gastritis. A typical history of peptic ulcer pain may be helpful, as will a prior or family history of GI bleeding. Recent retching followed by hematemesis points to the possibility of a MalloryWeiss syndrome. A change in bowel habits may suggest colonic cancer, and acute abdominal pain with bleeding may indicate ischemic colitis. The history may also reveal the presence of an aortic bypass graft.

Physical examination may reveal stigmata of chronic liver disease, suggesting a variceal source of bleeding, whereas enlargement of lymph nodes or an abdominal mass may reflect an intra-abdom­inal malignancy. Examination of the skin may show the telangiectasia characteristic of Osler-Weber-Rendu disease. Rectal examination will help to rule out a mass.

A simple and expedient diagnostic procedure is the passage of a nasogastric tube. A clear as­pirate rules out active bleeding proximal to the pylorus. Active bleeding from a duodenal ulcer may be missed but is ruled out if clear bile is as­pirated. If bleeding from an upper GI source is intermittent, the aspirate may be negative. An el­evated BUN also suggests upper, as opposed to lower, GI bleeding.

Further diagnostic procedures are selected ac­cording to the clinical suspicion of upper versus lower GI bleeding and the rate of blood loss. In acute bleeding, barium studies are generally in­appropriate, and if very rapid bleeding is present, particularly if the source is colonic, endoscopic examination may fail to visualize the site of bleed­ing. Under such conditions, angiography is often the procedure of first choice and may be preceded by a radiolabeled red-cell scan to localize further the region of blood loss. If bleeding is moderate, angiography often fails to show extravasation of contrast at the site of bleeding, and endoscopic procedures are preferred.

In most cases of suspected upper GI bleeding, upper GI endoscopy will provide $ rapid, safe, definitive means for diagnosis of the site of blood loss. If bleeding has stopped, a double-contrast barium upper GI series is an inexpensive proce­dure that will detect most malignancies and pep­tic ulcers, although gastritis is often missed. Endoscopy is a more specific and sensitive diagnostic tool that has occasional therapeutic ad­vantages as well as the advantage of allowing bi­opsy of lesions suspicious of malignancy. Thus endoscopy is preferred by many physicians for upper GI hemorrhage even in the patient who has stopped bleeding.

Bleeding from the lower GI tract is first evalu­ated by anoscopy and proctosigmoidoscopy, which will detect lesions in the rectosigmoid re­gion, including hemorrhoids, polyps, carcinoma, and inflammatory bowel disease. More proximal colonic lesions may be detected by colonoscopy after colonic cleansing, but if bleeding is brisk, colonoscopy may prove futile and recourse to an­giography may be necessary. If bleeding has stopped, double-contrast barium enema studies may be considered, but they are often inconclu­sive and make colonoscopy and angiography dif­ficult to perform in the event of early recurrent bleeding. Colonoscopy is thus often performed even after lower GI bleeding has stopped. If eval­uation of presumed lower GI bleeding is unre­warding, an upper GI source should always be ex­cluded. Small bowel lesions below the ligament of Treitz are difficult to detect, but may be seen with a small bowel barium series.

For patients with chronic GI blood loss, the ap­proach is usually more leisurely and the initial search for a responsible lesion is undertaken with barium studies followed, as indicated, by endo­scopic evaluation and biopsy of lesions suspicious of malignancy. Angiography is occasionally em­ployed, particularly to search for vascular lesions of the bowel.

In spite of the use of sophisticated diagnostic techniques, the site of bleeding is not established in manypatients with GI bleeding. Sometimes this reflects a subtle or inaccessible lesion, at other times a multitude of lesions without iden­tification of the one responsible for the bleeding. This does not necessarily compromise the patient, although repeated hospitalization for bleeding from an undiagnosed site is the lot of an unfor­tunate few and represents a difficult challenge to the physician.