Gastrointestinal Hemorrhage

From WiserWiki

[edit] Gastrointestinal Hemorrhage

David Lichtenstein

Umar Beejay

The management of patients with gastrointestinal (GI) hemorrhage constitutes a major problem that mandates close cooperation among primary care providers, gastroenterologists, surgeons, and interventional radiologists. Pragmatically, GI bleeding is divided into upper and lower GI hemorrhage, with the ligament of Treitz being the anatomic dividing line. Bleeding can be acute, with hematemesis, melena, and hematochezia, or chronic, presenting as iron deficiency anemia or guaiac-positive stool. The clinical approach depends not only on the acuity and localization of the bleeding, but also on patient characteristics and the medical skills and resources available. A consistent, planned approach to GI hemorrhage optimizes patient management, reduces hospital length of stay and expenditures, and improves patient survival.

[edit] INITIAL MANAGEMENT

The initial approach to the patient with acute GI hemorrhage includes assessing the severity of hemorrhage, initiating resuscitative efforts, and performing a brief history and limited physical examination (Fig. 109-1). The key to the successful management is the rapid institution of fluid resuscitation to maintain organ perfusion and tissue oxygenation. Depending on the acuity of illness, patients can be triaged to the intensive care unit (ICU) or a less intensively monitored setting. In those with severe bleeding, a multidisciplinary approach should be instituted early.

Figure 109-1 Approach to patient with acute gastrointestinal hemorrhage. EGD, Esophagogastroduodenoscopy.

[edit] Clinical Assessment

The primary initial objective of the clinical evaluation is to assess the patient's severity of hemorrhage by determining hemodynamic status. This assessment takes precedence over measures aimed at localizing and directing treatment of the bleeding site. Vital signs are monitored frequently, including blood pressure, pulse, and postural changes. The patient's ability to manifest compensatory hemodynamics is often influenced by other factors, such as medication intake (e.g., β-adrenergic blocker), age, vascular integrity, and the intactness of the autonomic nervous system. Initial hematocrit (Hct) determination can be misleading at presentation because vascular fluid redistribution requires 24 to 72 hours to equilibrate (Fig. 109-2). The patient's general appearance is also important; skin turgor and jugular venous pressure supplement vital signs in determining the level of volume depletion. Whether the patient is pale or jaundiced should also be noted. Cachexia, lymphadenopathy, and abdominal masses suggest malignancy. The abdominal examination should evaluate for tenderness, masses, hepatosplenomegaly, ascites, and bowel sounds. A rectal examination provides valuable information on the color and consistency of the stool (e.g., maroon vs. bright red, melena vs. occult blood). Cutaneous stigmata may suggest the etiology of GI hemorrhage. Jaundice, spider nevi, palmar erythema, and ecchymoses suggest liver disease, whereas orofacial telangiectasias suggest Osler-Weber-Rendu syndrome (hereditary hemorrhagic telangiectasia). Café au lait spots suggest neurofibromatosis, and eythema nodosum is associated with inflammatory bowel disease (IBD).

Figure 109-2 Reliability of hematocrit (Hct) as indicator of severity of gastrointestinal bleeding. In acute period after hemorrhage, plasma volume and red blood cell volume have decreased in parallel so that Hct remains essentially unchanged. Over the ensuing 72 hours, however, equilibration of extravascular and intravascular fluid spaces results in Hct that accurately reflects total blood losses.

Clinical assessment is poor at predicting the source of bleeding, but the history helps direct therapy because the distribution of etiologies varies with patient age, particularly in lower GI bleeding. Thus in young patients the diagnosis frequently includes IBD, Meckel's diverticulum, and juvenile polyps; in middle-aged patients, polyps, hemorrhoids, and diverticula; and in elderly patients, angiodysplasia, malignancy, ischemia, diverticulosis, and polyps. With lower GI hemorrhage the presence or absence of abdominal pain is also important (see Chapter 107 ). Diffuse abdominal pain can suggest inflammatory bowel disease or ischemic bowel, whereas painless bleeding is common with diverticula, angiodysplasia, malignancy, polyps, and Meckel's diverticulum. Bloody diarrhea may indicate infectious (e.g., Clostridium difficile) colitis or IBD, or recent antibiotic use. Recent constipation suggests anorectal sources, such as hemorrhoidal bleeding, anal fissure, or solitary rectal ulcer syndrome. The physician should inquire about symptoms of extraintestinal manifestations of IBD (see Chapter 108 ). The history should include prior episodes of bleeding, family history of GI diseases (e.g., hereditary hemorrhagic telangiectasia), known illnesses (e.g., ulcers, cirrhosis, cancer, bleeding diathesis), and previous abdominal surgery. A prior aortic aneurysm repair suggests aortoenteric fistula; undiagnosed bleeding from an aortoenteric fistula is ultimately fatal. The medication profile is important, especially use of nonsteroidal antiinflammatory drugs (NSAIDs) and anticoagulants. Within the spectrum of upper GI hemorrhage, classic features of epigastric pain relieved by food suggests peptic ulcer disease, whereas chronic renal failure should suggest arteriovenous malformations. Similarly, a classic history of retching followed by hematemesis should suggest a Mallory-Weiss tear, although this classic history, which was initially considered to be a prerequisite for the syndrome, is now known to occur in only 30% to 50% of patients. Many patients experience hematemesis without previous retching or present infrequently with melena alone.

[edit] Resuscitation and Transfusion of Blood Products

Resuscitation of the patient is initiated concurrently with the clinical assessment. Blood is sent to the laboratory for determination of the complete blood count (CBC), coagulation profile, and serum chemistries, including blood urea nitrogen (BUN), creatinine, and liver function tests. Blood is also sent to the blood bank for typing and crossmatching. While obtaining blood specimens, at least two large-bore (14-to 18-gauge) intravenous (IV) catheters should be placed for the administration of fluids and blood products. More aggressive invasive monitoring may be needed in those with significant cardiorespiratory disease. Strict charting of fluid intake and output is helpful to assess tissue perfusion. Patients in shock unresponsive to fluid resuscitation require IV pressors. The type and quantity of fluid replacement must be tailored to the patient's specific needs.

The use of rigid transfusion guidelines, such as maintaining Hct above 30%, should be avoided. A hemoglobin (Hb) level greater than 7 to 8 g/dl is generally acceptable if the patient is young and not actively bleeding and if fluid resuscitation has achieved a near steady-state value. In elderly patients or those with known cardiovascular disease, an Hb level of approximately 10 g/dl should be maintained. Empiric transfusion of fresh-frozen plasma (FFP) or platelets after packed red blood cells (PRBCs) is unnecessary provided rapid access to the coagulation profile and platelet count is available. FFP may be used to reverse the effect of warfarin; correct deficiencies of coagulation factors (II, V, VII, VIII, IX, X, XI, XIII), antithrombin III, heparin cofactor II, protein C, and protein S; and correct deficiencies of multiple factors, as in severe liver disease and vitamin K depletion. Early factor replacement should be considered in patients with known cirrhosis, with an average requirement of 1 unit of FFP for every 4 units of PRBCs. Factor VII biologic half-life is only 6 hours, and additional infusions may need to be repeated every 6 to 12 hours. The usual adult dosage of FFP is 3 to 5 units (12 to 15 ml/kg) given over 1 hour to several hours, depending on the patient's cardiac status. Platelet counts less than 80,000/dl are associated with increased bleeding and, in the setting of active GI bleeding, should be corrected with transfused platelets. If existing platelets are dysfunctional, as with recent aspirin or NSAID use, platelet replacement should be considered for the actively bleeding patient even when platelets are quantitatively normal.

The majority of patients who develop GI bleeding while taking anticoagulants have underlying mucosal pathology. When bleeding occurs in the setting of anticoagulation, the degree of bleeding and the reason for anticoagulation need to be reassessed for possible reversal of therapy. Patients with GI bleeding who are taking warfarin may require only discontinuation of their anticoagulation therapy if bleeding is not severe and if the coagulation profile is within the therapeutic range. When warfarin is discontinued, the international normalized ratio (INR) does not normalize for 3 to 4 days. Alternatively, a low dose of vitamin K (1 mg) or FFP may be given intravenously. Heparin therapy can simply be stopped, unless bleeding is severe, in which case the anticoagulant effects of heparin should be reversed with protamine sulfate.

Patients with renal failure often have an underlying bleeding diathesis reflected in a prolonged bleeding time. The hemostatic defects in patients with chronic renal failure are caused by abnormalities in platelet function, vessel wall–coagulation interaction, and plasma coagulation factor deficiencies. Desmopressin (DDAVP) is a synthetic analog of arginine vasopressin that possesses hemostatic properties because of the release from vascular endothelium of preformed high-molecular-weight multimers of von Willebrand's factor and factor VIII:C. DDAVP has a rapid onset of effect but is limited by its short duration of action of 6 hours. The IV dose is 0.3 μg/kg diluted in 50 ml of normal saline given over 15 to 30 minutes. Repeat dosing in 12 to 24 hours reproduces an increase in coagulant factors, although a decreased physiologic response is then seen because of depletion of preformed factors from vascular endothelium. Although DDAVP is frequently used to correct the bleeding diathesis in uremic patients with acute bleeding, improved survival has not been shown. Uncontrolled trials also report improved bleeding times and reduced rebleeding rates and transfusion requirements after administration of cryoprecipitate, conjugated estrogens, or combined estrogen and progesterone.

[edit] Localization of Gastrointestinal Hemorrhage

The initial challenge after stabilization of the patient is to determine whether the hemorrhage is emanating from above (upper intestinal) or below (lower intestinal) the ligament of Treitz. The presence of hematemesis is virtually always indicative of an upper gastrointestinal (UGI) hemorrhage; however, epistaxis or oropharyngeal bleeding may infrequently simulate a UGI bleed. Melena usually indicates UGI bleeding and occurs as a result of bacterial degradation of Hb to hematin and other hemochromes as blood traverses the small bowel and colon. Melena can result from as little as 50 to 100 ml of blood introduced into the UGI tract. Rarely,melena will occur during a lower GI bleed, but only when bleeding is slow and is localized to the right colon or small bowel. Other clinical findings suggestive of UGI hemorrhage include hyperactive bowel sounds and elevated BUN out of proportion to serum creatinine.

The passage of red blood from the rectum usually indicates a lower GI bleed. A brisk upper GI bleed can present similarly, however, and therefore nasogastric (NG) tube aspiration should be performed in all patients with suspected lower GI bleeding. Patients with hematochezia from UGI bleeding usually have evidence of volume depletion from the history and physical examination. NG tube placement and lavage are not recommended for suspected UGI hemorrhage (provided endoscopy is planned within several hours) because of (1) delays in definitive therapy caused by prolonged efforts to clear the stomach of its contents; (2) risks of pulmonary aspiration, intestinal perforation, incorrect tube placement into the respiratory tree, and mucosal suction artifacts, which can be indistinguishable from true mucosal vascular lesions; and (3) lack of evidence that gastric lavage is capable of stopping active bleeding or preventing recurrence. NG aspiration should be considered in patients with suspected UGI hemorrhage when esophagogastroduodenoscopy (EGD) will be delayed in order to (1) confirm the upper tract nature of bleeding when melena or hematochezia is present rather than hematemesis, (2) provide prognostic information, and (3) cleanse the stomach to facilitate endoscopic examination. The presence of gross blood or “coffee grounds” on gastric aspirate confirms an UGI source. A clear aspirate does not exclude an upper bleed, however, since bleeding may be episodic or the lesion may be distal to the stomach, with a competent pylorus preventing duodenogastric reflux of blood. Bile in a nonbloody aspirate makes an upper source less likely, although this can be misleading because the visual appearance of the aspirate does not reliably discriminate between the presence and absence of bile. When NG aspiration is performed, the use of chemical tests (e.g., guaiac reaction) are of little clinical value and should be avoided.

[edit] NONVARICEAL UPPER GASTROINTESTINAL HEMORRHAGE

Nonvariceal UGI bleeding continues to rank among the most common emergencies encountered by physicians, with more than 300,000 hospital admissions annually in the United States. With only supportive therapy, approximately 80% to 85% of patients with UGI hemorrhage will stop bleeding (Fig. 109-3). The remaining 15% to 20% who continue to bleed or have recurrent hemorrhage account for the majority of complications and deaths. It is these high-risk patients who require early identification and in whom endoscopic, surgical, and angiographic therapies have improved patient outcome. Severe UGI hemorrhage is most often caused by peptic ulcer disease, Mallory-Weiss tears (Fig. 109-4), esophageal varices, vascular malformations, and erosive gastroduodenal lesions (Table 109-1).

Figure 109-3 Outcome of acute upper gastrointestinal (UGI) bleeding.

Figure 109-4 Endoscopic view of Mallory-Weiss tear, with linear mucosal laceration at gastroesophageal junction.

Table 109-1 Common Causes of Upper Gastrointestinal (UGI) Hemorrhage

[edit] Prognostic Factors

Clinical factors that predict a poor outcome include advanced patient age, comorbid medical illnesses, onset of bleeding while already hospitalized, recurrent hemorrhage, and specific etiology and severity of the bleeding episode (Box 109-1). Clinical indicators of severe bleeding include hemodynamic instability on presentation, persistent hypotension, transfusion requirements in excess of 4 to 6 units within 24 hours of hospital admission, or bleeding manifested as hematemesis, hematochezia, or a bloody gastric aspirate that does not clear with lavage.^[1]

Although clinical criteria are useful predictors for determining outcome, the endoscopic information indicating the cause of bleeding (and in the case of ulcers the appearance of the ulcer base) provides equally important prognostic information (Fig. 109-5). The bleeding ulcer may have no stigmata of recent hemorrhage (SRH), that is, a clean base. Major SRH include active bleeding (oozing or spurting) or visible vessel, and minor SRH include an adherent clot or flat pigmented spots (red, black, purple, or brown) (Table 109-2).^[1] An actively bleeding ulcer predicts continued or recurrent bleeding in more than 85% of cases (Fig. 109-6). The term nonbleeding visible vessel describes a 2-mm to 3-mm protuberance in the floor of an ulcer crater that is typically the vessel itself or more often represents an adherent clot plugging the eroded artery (Fig. 109-7). The visible vessel is found in 17% of bleeding ulcers; 43% of patients have rebleeding, 34% require emergency surgery, and 11% die when endoscopic therapy is withheld. A clean ulcer base (i.e., without endoscopic SRH) is found in 20% to 50% of bleeding ulcer patients and reliably predicts a benign course with a negligible risk for rebleeding (0% to 5%) and virtually no mortality. Ulcers with minor SRH of flat spots or adherent clots account for approximately one third of the ulcers found at endoscopy (Fig. 109-8). These patients have a slightly increased incidence of rebleeding, surgery, and mortality, with mean rates of 10% to 20%, 5% to 10%, and 3% to 7%, respectively. Other endoscopic prognostic factors shown to influence outcome include ulcer location and size. Ulcers greater than 1 to 2 cm diameter more often have SRH and indicate an increased likelihood of rebleeding and death. Ulcers located in the posteroinferior duodenal bulb and to a lesser extent the high lesser gastric curve may adversely influence outcome because of their proximity to large vessels, the gastroduodenal artery and left gastric artery, respectively.

Figure 109-5 Influence of diagnosis on outcome of UGI hemorrhage.

Figure 109-6 Endoscopic view of a spurting ulcer.

Figure 109-7 Gastric ulcer containing nonbleeding visible vessel, manifested by raised red spot within ulcer crater.

Figure 109-8 Duodenal ulcer with minor stigmata of recent hemorrhage (SRH), manifested by nonraised black spot within ulcer crater.

Table 109-2 Endoscopic Stigmata of Recent Hemorrhage (SRH) in Peptic Ulcer Bleeding

[edit] Patient Triage

A treatment and triage algorithm incorporates clinical criteria with endoscopic findings (Fig. 109-9). Patients can be divided on clinical grounds into two groups: those with hemodynamically significant bleeding at high risk for rebleeding and death and those at low risk with limited blood loss. The high-risk group requires admission to a monitored unit or ICU. The low-risk group can be triaged in the emergency room with diagnostic endoscopy or admitted to the ward before EGD. Endoscopy in both groups can predict the likelihood of further bleeding and guides further decisions regarding therapy, feeding, and length of hospitalization.

Figure 109-9 Proposed algorithm to triage and manage patients with UGI hemorrhage from peptic ulcer disease. EGD, Esophagogastroduodenoscopy;ER, emergency room;ICU, intensive care unit.

Patients with major SRH of active bleeding or nonbleeding visible vessels are treated endoscopically and observed for 24 hours in an ICU and then transferred to the ward if there are no signs of rebleeding. The patients with an uncomplicated hospital course can safely be discharged after 3 to 4 days of observation, since the vast majority of ulcer rebleeding occurs within 72 hours of bleeding onset. Patients with endoscopic SRH predicting a high likelihood of rebleeding should receive nothing by mouth or a clear-liquid diet for 24 hours after endoscopic therapy so that food in thestomach does not interfere with an urgent endoscopic or surgical procedure, which may be necessary if rebleeding occurs. Patients with clean ulcer bases can be fed and discharged on the first day of admission immediately after volume resuscitation and stabilization. Factors favoring outpatient care include absence of severe anemia, serious comorbid medical illness, or liver disease; lack of concomitant anticoagulation therapy or coagulopathy; no signs of active bleeding; and adequate volume resuscitation and home support.^[2][3] Patients with intermediate risk of rebleeding, characterized by minor SRH, should be refed early and observed on a medical ward for up to 3 days before discharge.

[edit] Diagnostic Studies

EGD is the most useful study for determining the location and type of lesion responsible for UGI hemorrhage. The skilled endoscopist can identify the bleeding source in more than 95% of cases and can offer a number of therapeutic options for the endoscopic control of hemorrhage and prevention of rebleeding. Radiographic contrast studies should be avoided in the setting of acute hemorrhage. Disadvantages of the early use of barium include (1) interference with diagnostic and therapeutic angiographic and endoscopic procedures; (2) lack of sensitivity in identifying superficial mucosal sources of bleeding (e.g., Mallory-Weiss tear, gastritis, vascular malformation); and (3) inability to distinguish the source of hemorrhage in patients with more than one mucosal lesion.

EGD should be performed urgently for the high-risk bleeder and within 12 to 24 hours for all others with an acute, self-limited episode of UGI bleeding. Included in the high-risk category are patients with chronic liver disease, suspected aortoenteric fistula, large-volume blood loss, and active hemorrhage, as well as patients experiencing or unlikely to tolerate a recurrent episode of hemorrhage, including those who have comorbid illness or who object to blood transfusions on religious grounds. Uncorrectable coagulopathy is a relative contraindication to endoscopic hemostasis in the patient who is not actively bleeding; this should not preclude attempts to control active bleeding, however, despite that effective hemostasis may be only temporary. Airway protection with endotracheal intubation is recommended in the pediatric, obtunded, or massively bleeding patient. EGD in the bleeding patient is technically more demanding than an elective endoscopic procedure. Although complications occur slightly more often in the actively bleeding patient, the procedure is usually well tolerated and safe, with a reported rate of major complications (e.g., perforation, pulmonary aspiration, induced bleeding) of only 0.5% and a mortality of less than 0.1%. Recent myocardial infarction is not an absolute contraindication to EGD, and a risk-benefit assessment needs to be individualized for each patient.

[edit] Management

Management of UGI hemorrhage may be divided into acute treatment of the bleeding lesion and strategies for prevention of rebleeding. It is important to develop both a short-term and a long-term strategy for preventing recurrent ulceration and rebleeding, since recurrent hemorrhage may develop acutely during the initial hospitalization (typically within 24 to 48 hours) or may be delayed for months to years after peptic ulcer bleeding.

[edit] Acute Treatment

[edit] Medical Therapy.

Theoretically, by slowing or halting bleeding, pharmacologic agents may be useful in the resuscitative phase of management and in enhancing the visualization of lesions during endoscopy. Iced saline gastric lavage has been a time-honored treatment considered essential for controlling UGI hemorrhage; however, the therapeutic value of lavage has since been refuted. Pharmacologic agents that act as antifibrinolytics (e.g., tranexamic acid), reduce gastric acid (e.g., histamine receptor antagonists [H₂RAs], proton pump inhibitor [PPIs], and prostaglandin analogs), act as mesenteric vasoconstrictors (e.g., vasopressin), or both reduce acid and act as vasoconstrictors (e.g., somatostatin) have been studied in an attempt to control actively bleeding ulcers. Although a few studies have reported benefit, most have found these agents to be ineffective, and at present, no medical therapy can be strongly recommended for the acute treatment of nonvariceal UGI bleeding. Data from in vitro studies suggest that clotting occurs more effectively and that lysis of clots by proteolytic enzymes occurs more slowly at high intraluminal pH levels.^[4] This forms the basis for regimens designed to reduce gastric acidity in patients with intestinal bleeding. No single trial has convincingly demonstrated an overall benefit of H₂RAs in stopping active bleeding or preventing rebleeding. The availability of the more potent PPIs to control peptic ulcer bleeding has renewed interest in antisecretory therapy. The available information is promising but inconclusive.^[5][6] Therefore it is reasonable to initiate treatment with a PPI for the potential benefit and as an alternative therapy when endoscopy is unsuccessful, contraindicated, or unavailable. PPIs also may be used once the diagnosis of an acid-peptic disorder is confirmed, acutely to promote healing of the lesion and as maintenance therapy to prevent recurrence.

[edit] Endoscopic Therapy.

A National Institutes of Health (NIH) Consensus Conference^[7] on therapeutic endoscopy and bleeding ulcers concluded that endoscopic therapy should be limited to patients at high risk for persistent or recurrent bleeding and death. This group includes those with clinical evidence of significant blood loss and endoscopic evidence of active bleeding or a nonbleeding visible vessel. Meta-analysis of studies evaluating endoscopic hemostasis for nonvariceal UGI bleeding found a significant reduction in mortality (relative risk reduction of 30%) and a decrease in both continued or rebleeding rates (69%) and need for emergency surgery (62%).^[8] Ulcers with minor SRH or a clean base have a much lower rate of rebleeding and do not benefit from endoscopic therapy. Other nonulcer lesions amenable to endoscopic therapy include select vascular malformations, Dieulafoy's lesion, and Mallory-Weiss tears. The two major complications of endoscopic therapy are the induction of bleeding (0.5%) and intestinal perforation (0.3%).

[edit] Prevention of Ulcer Rebleeding.

Effective endoscopic hemostasis decreases the risk of acute rebleeding to approximately 10% to 20% in patients presenting with peptic ulcer hemorrhage. Repeat upper endoscopy and retreatment are reserved for patients with recurrent bleeding; permanent hemostasis is achieved in about 50% of these patients, for an overall endoscopic success rate of approximately 90%. Although no known pharmacologic intervention has been shown to reduce reliably the incidence of acute rebleeding in patients with peptic ulcer hemorrhage after endoscopic hemostasis, acid-suppressive medication, eradication of Helicobacter pylori,^[9] and cessation of NSAIDs greatly reduce delayed rebleeding. The presence of H. pylori is determined at the initial EGD, with gastric biopsy for rapid urease testing or histologic examination, or subsequently with noninvasive breath testing or serology. Follow-up endoscopy is recommended to document healing, exclude malignancy, and assess for H. pylori eradication in patients with bleeding from a gastric ulcer. The less expensive and noninvasive breath testing alone is recommended 4 to 6 weeks after completion of antimicrobial therapy for bleeding duodenal ulcer disease. Antisecretory therapy is continued until H. pylori eradication is documented. Although previously recommended for all peptic ulcer–related bleeds, maintenance antisecretory therapy should be restricted to select patients, including ulcer patients not infected with H. pylori, those with failed H. pylori eradication, and successfully treated patients who are likely to tolerate poorly recurrent hemorrhage because of concurrent illness. Patients with H. pylori infection and a history of NSAID use should have both problems corrected; when therapy is successful, maintenance ulcer therapy generally is not recommended. Although eradication of H. pylori infection would also seem to decrease the risk of delayed recurrent hemorrhage in patients continuing to use NSAIDs, the extent of this reduction has not been evaluated. Patients still requiring NSAIDs should receive long-term prophylactic therapy with misoprostol or preferably a PPI and then a selective cyclooxygenase II (COX-II) inhibitor.

[edit] Angiographic Intervention.

Angiographic therapy is used infrequently to treat patients with UGI bleeding and should be considered only for severe, persistent bleeding if surgery confers a high risk and endoscopic therapy is not available or has been unsuccessful. Uncontrolled studies suggest that arterial embolization with absorbable gelatin sponge (Gelfoam) and autologous clot or with nonabsorbable polyvinyl alcohol and sponge-wire coils may control bleeding identified angiographically in 75% to 80% of patients, although recurrent bleeding may occur in more than half. Alternatively, selective intraarterial vasopressin infusion may temporarily stop bleeding in up to 50% of patients. Complications of angiographic therapy include allergic reactions, renal insufficiency, ischemia, and perforation in target and nontarget organs.

[edit] Emergency Surgery.

All patients with UGI bleeding severe enough to cause significant hemodynamic instability that requires ICU admission warrant evaluation by a surgeon. Criteria for emergency surgery include (1) failure to control bleeding with nonoperative means; (2) severe rebleeding despite two attempts at endoscopic hemostasis; (3) a lesion inaccessible to endoscopy because of prior surgery, anatomic anomaly, or pyloric stenosis; (4) severe shock when emergent surgery may prevent exsanguination; and (5) a severe complication of endoscopic therapy (e.g., perforation, worsening bleeding lesion).

[edit] ACUTE LOWER GI BLEEDING^[10]

It is helpful to divide the common causes of lower GI bleeding by age and clinical presentation, since knowledge of the etiology helps plan optimal management (Tables 109-3 and 109-4). After initial evaluation and volume resuscitation, further management depends on the results of an NG tube aspirate. About 1000 ml or more of blood is required to cause hematochezia from an upper source, and hemodynamic compromise is typically an accompanying feature. If copious nonbloody bile is seen on NG aspiration, the physician should proceed directly to a colonoscopy. In all other cases, however, the colonoscopy should be preceded by an EGD because as many as 10% to 15% of patients with suspected lower GI bleeding have a UGI source. The diagnostic yield from colonoscopy ranges from 60% to 80%. Timing of endoscopy has not been systematically studied but should be performed as soon as possible in patients with continuous hematochezia. Patients who have stopped bleeding can undergo examination on a semielective basis.

Table 109-3 Age and Clinical Presentation of Lower Gastrointestinal Bleeding

Table 109-4 Common Causes of Lower Gastrointestinal Bleeding

The three outcomes to colonoscopy are source identification, negative examination, or incomplete colonoscopy secondary to severe bleeding. If identified, the source of bleeding should be treated appropriately. If the examination is unsuccessful in identifying the source because the bleeding is sudden and intermittent or is severe, arteriography is indicated. In some patients a bleeding scan is an appropriate initial step before arteriography. Abdominal scintigraphy plays a limited diagnostic role, however, because it requires active bleeding, is often obscured by liver and spleen tracer uptake, and does not provide information on the nature of the bleeding site. Angiography may be helpful when bleeding is active and has not been localized endoscopically. The bleeding must be arterial and at least 0.5 ml per minute. A negative colonoscopy and arteriographic study may require further evaluation with a Meckel's scan, small bowel enteroscopy, or a repeat colonoscopy if bleeding recurs. A Meckel's scan with technetium-99m pertechnetate should be considered in young patients to exclude Meckel's diverticulum. If bleeding has ceased, contrast radiography to evaluate for small bowel pathology with enteroclysis is preferred to standard barium small bowel series because of greater sensitivity. The technique requires intubation of the small bowel to the duodenojejunal junction, with controlled infusion of barium, methylcellulose, and water to achieve a double-contrast effect. If the results of small bowel studies are negative and the blood loss is self-limited or maintained by oral iron supplementation, further evaluation may not be necessary. Surgery is reserved for treatment of a defined site of hemorrhage or for diagnostic purposes when combined with intraoperative endoscopy.

[edit] CHRONIC GASTROINTESTINAL HEMORRHAGE

[edit] Occult Bleeding^[11]

The most common presentation of occult GI bleeding includes a positive fecal occult blood test with or without concomitant iron deficiency anemia. Although fecal occult blood tests were designed as a screening test for colorectal carcinoma, they are used to obtain evidence of blood loss in any acute or chronic bleeding lesion. The standard workup for occult GI blood loss in the adult population is to perform a colonoscopy. In the presence of UGI symptoms or iron deficiency anemia, a negative colonoscopy should be followed by an EGD.

[edit] Obscure Bleeding

Obscure bleeding is defined as active or recurrent hemorrhage that evades diagnosis on routine endoscopic and radiologic evaluation. It has an estimated prevalence of 5% to 10% in patients with GI bleeding. A wide range of diagnostic studies are available for evaluating bleeding of obscure origin, and the choice depends on severity of bleeding and availability of endoscopic, radiologic, and surgical expertise. Initially, standard UGI endoscopy should be repeated, particularly if blood prevented adequate mucosal inspection, since missed lesions account for a significant proportion of identified lesions on subsequent enteroscopy evaluation. Patients with massive or acute recurrent bleeding require a more invasive approach, including small bowel enteroscopy and angiography. Small bowel enteroscopy can be accomplished with push, sonde, or intraoperative techniques.^[12][13]Push enteroscopy permits inspection of the entire duodenum and proximal jejunum to a level 40 to 50 cm beyond the ligament of Treitz. Sonde enteroscopy or intraoperative techniques examine the length of small intestine to the level of the terminal ileum. An etiology of bleeding can be determined in up to 35% of cases evaluated. Most patients with a small bowel source have bleeding from vascular malformations, but other lesions occasionally can be identified, including celiac disease, tumor, and occult IBD.

[edit] REFERENCES