Learning Objectives

1. Identify the most common causes of upper gastrointestinal (UGI) hemorrhage.
2. Describe the role of medical and surgical management for acute nonvariceal UGI bleeding.
3. Describe which patients with ulcer hemorrhage should be treated endoscopically.
4. Describe long-term medical therapies to prevent recurrence of ulcer hemorrhage.
5. Discuss the most common causes and the current diagnosis and treatment of patients with severe hematochezia and presumed lower (L) GI hemorrhage.

[TOP]

Gastrointestinal bleeding is a common clinical problem. Upper gastrointestinal (UGI) hemorrhage accounts for about 400 000 hospitalizations per year. Despite major advances in patient care, the mortality rate for acute UGI hemorrhage has remained relatively constant between 5 and 10% over the past 50 years. One possible explanation is that increases in both the age and the prevalence of comorbid conditions among those presenting with UGI hemorrhage have offset the diagnostic and therapeutic advances of the past 20 years.

Acute gastrointestinal bleeding is a common worldwide clinical problem and continues to be associated with significant morbidity and mortality. The annual rate of hospitalization for UGI bleeding has been estimated at 30 to 100 patients per 100 000 with about 400 000 hospitalizations per year for acute nonvariceal upper GI bleeding in the United States. Lower (L) GI bleeding occurs less frequently, with an incidence of 6 to 20 per 100 000.

Mortality rates from UGI hemorrhage are high, varying from 3.5 to 7% in the United States. A large United Kingdom study reported a mortality rate of 14%.1 Mortality rates for LGI bleeding range from 2 to 5%.

Despite decades of overall improvement in emergency and intensive unit care, blood banking, pharmacological therapy, and endoscopic hemostasis, mortality from UGI bleeding has not changed significantly over the past 20 to 30 years. The most likely reason for this is the increasing proportion of elderly patients comprising the population of UGI bleeding patients. Currently 44% of UGI hemorrhage hospitalizations occur in patients older than 60 years of age. In one study, the mortality rate in patients over 80 years old with UGI hemorrhage was 11.2%, compared to 0.4% for patients under the age of 60. Most deaths from bleeding are not directly caused by bleeding but are related to worsening of underlying coexistent medical illness or to complications of hospitalization and surgery.

After initial stabilization of the patient, the management of GI hemorrhage includes determining the source of bleeding, stopping active bleeding or endoscopically treating major stigmata of hemorrhage, treating the underlying abnormalities or causes, and preventing recurrent bleeding.

[TOP]

UGI Bleed Etiology

[ZOOM]

Peptic ulcer disease is the most common cause of acute UGI hemorrhage, accounting for about 50% of cases.2 Conversely, UGI hemorrhage is the most common complication of peptic ulcer disease (Table 5.1). About 25% of ulcer patients bleed at some time during the course of their disease with a greater proportion of duodenal ulcer than gastric ulcer patients experiencing hemorrhage. In about 75 to 80% of cases, hemorrhage from ulcers stops spontaneously, and subsequent mortality is negligible. In a subgroup of patients, however, bleeding persists or recurs and may be associated with mortality rates of 25 to 30% in patients with severe comorbid conditions. Other conditions, such as Mallory-Weiss tear, angiodysplasia, watermelon stomach, and Dieulafoy's lesion, occur less frequently than peptic ulcers but are important causes of UGI hemorrhage and contribute substantially to the morbidity and cost of care associated with UGI bleeding (Table 5.1).3

[TOP]

Risk Factors

Independent risk factors for ulcer hemorrhage are older age (over 65 years), use of nonsteroidal anti-inflammatory drugs (NSAIDs) or aspirin, and a history of prior peptic ulcer. After a single episode of ulcer hemorrhage and ulcer healing, the relative risk for recurrent ulcer hemorrhage during long-term follow-up is 10 to 20 times that of a control population, depending on the number of risk factors present. These risk factors appear to be additive. For example, NSAID use and a history of prior peptic ulcer disease result in a 17-times-higher relative risk for ulcer hemorrhage than in a control population.

Ingestion of NSAIDs may cause either gastric ulcers (GUs) or duodenal ulcers (DUs). Complications may develop shortly after starting NSAIDs and are more frequent during the first month of treatment. The risk of NSAID-related ulcer bleeding is dose dependant. In a placebo-controlled trial comparing 300 mg or 1200 mg daily of aspirin for prophylaxis of transient ischemic attacks, the 1200 mg dose was found to have twice the relative risk of hemorrhage as the 300 mg dose. Nevertheless, the lower aspirin dose had a statistically significant higher relative risk of bleeding (7.7 times higher) than the group randomized to placebo. Corticosteroid use alone is not associated with an increased risk for complications such as ulceration or bleeding. However, concomitant steroid and NSAID use doubles the risk of ulcer complications compared with NSAID use alone and increases by the chance of UGI bleeding by 10-fold. Although Helicobacter pylori is a risk factor for the development and recurrence of peptic ulcer disease, it has not been shown to be an independent risk factor for ulcer hemorrhage.

In large trials (for example, over 8000 patients) of COX2-inhibitors compared to nonselective NSAIDs, other independent risk factors for development of ulcers or ulcer complications included cardiovascular disease, coumadin ingestion, and corticosteroids.

[TOP]

Presentation

[ZOOM]

Patients with UGI bleeding present with hematemesis (30%), melena (20%), or both (50%). Hematochezia may be the only manifestation in about 5 to 10% of patients with bleeding ulcers. In one study, about 18% of all patients presenting with hematochezia were found to have an UGI source (Table 5.2).4

Various symptoms and signs have been used to predict severity of the bleeding episode (Table 5.3). These are shock, red hematemesis (as opposed to coffee-ground emesis), hematochezia, or inability to clear the nasogastric (NG) lavage. Severe coagulopathy is another indicator of poor prognosis. Mortality after a bleeding episode has also been correlated with age greater than 60 years, inpatient hemorrhage, and presence of clinically significant coexistent disease, especially if cardiovascular, respiratory, hepatic, or malignant in nature.1

[TOP]

Initial Resuscitation and Medical Management

The initial approach to the patient with UGI hemorrhage should combine an evaluation of the severity of the bleed, a brief history, and physical examination with early and vigorous resuscitation. If the patient is not hypotensive, orthostatic changes in pulse and blood pressure should be determined. An early decision must be made whether hospitalization is required and whether ICU or monitored bed admission is indicated. All patients with severe, acute UGI bleeding should be considered for admission to an ICU or telemetry bed, because in most hospitals the major causes of mortality, such as bleeding, associated illness, or postoperative complications, are all better managed in an ICU or telemetry unit than on a regular hospital ward.

[ZOOM]

Resuscitation should be started immediately and concurrently with the initial evaluation. Important determinants of resuscitation include adequate intravenous (IV) access, accurate assessment of blood loss, and appropriate fluid and blood product infusion.

Patients with UGI hemorrhage may present with hematochezia in the absence of hematemesis. A grossly bloody NG aspirate in such patients identifies a UGI location of bleeding. However, a clear NG aspirate does not rule out an upper source. Even the presence of bile without blood may be misleading because bleeding from a duodenal lesion may be intermittent.

In patients with a bloody NG aspirate, gastric lavage is usually done to remove blood from the stomach before endoscopy to improve visualization. There is no therapeutic value to iced saline lavage. If lavage is done, lukewarm tap water should be used because it is as safe as, and cheaper than, saline. Large-volume gastric lavage may be uncomfortable for the patient and may increase the risk of aspiration. Intravenous metoclopramide (10 mg) or erythromycin (250 mg) before endoscopy has also been reported to clear the stomach of clots and improve emergency endoscopic examinations in patients with severe UGI hemorrhage.

Airway protection is critical during severe UGI hemorrhage. For example, in one study of severe UGI hemorrhage, respiratory complications occurred in 22% of patients. Patients with respiratory complications had a much higher mortality than patients without these complications (70 versus 4%). Endotracheal intubation should be strongly considered to prevent aspiration in patients with ongoing hematemesis, altered mental status, altered respiratory status, or severe neuromuscular disorders (where mechanically assisted ventilation is indicated for sedated procedures).

[TOP]

Medical Therapy

Several agents have been evaluated for initial control of active bleeding or prevention of rebleeding from peptic ulcers. Since in vitro data show that coagulation parameters and platelet function improve at neutral pH in comparison with an acid milieu, a long-standing therapeutic goal in the treatment of ulcer hemorrhage has been to inhibit gastric acid secretion to promote clot formation and prevent clot dissolution. A meta-analysis of randomized trials of H2-receptor blockers for acute UGI bleeds showed no significant difference in persistent or recurrent bleeding. This may have been due to inability to sustain acid suppression.

Proton-pump inhibitors (PPI) such as omeprazole (40 mg orally twice daily), which produce more potent acid suppression than H2 receptor antagonists, have been reported to significantly reduce the rates of rebleeding and surgery in patients with ulcer hemorrhage.5 High-dose intravenous omeprazole (80 mg bolus followed by8 mg/h for 72 hours) following successful endoscopic therapy was also reported to reduce rebleeding and need for emergent surgery.6 Other studies have also suggested that, after successful endoscopic treatment of bleeding peptic ulcers, both oral omeprazole and lower IV doses improve outcomes. A meta-analysis of proton-pump inhibition confirmed that high-dose intravenous PPI use significantly reduces ulcer rebleeding, surgery, and mortality.7 This meta-analysis also suggested that both high-dose oral and non-high-dose intravenous proton-pump inhibition improved outcomes, although there were fewer trials to evaluate. PPIs are now the medical therapy of choice for patients with ulcer hemorrhage after successful endoscopic hemostasis, but the optimal dose or route administration is not well defined, and further studies are indicated in high-risk ulcer bleeding patients. Other medications that reduce gastric acid secretion may be useful in the medical management of peptic ulcer bleeding. Somatostatin has theoretical advantages over other antisecretory drugs because it inhibits both acid and pepsin secretion and reduces gastroduodenal mucosal blood flow. The inhibition of pepsin secretion may provide additional benefits by preventing the dissolution of freshly formed clots at the bleeding site. Although conflicting results have been reported, a meta-analysis of 14 trials showed that somatostatin decreased the risk of persistent or recurrent hemorrhage in comparison to controls. Octreotide, a long-acting somatostatin analogue had similar effects, although there were fewer studies. Medications such as prostaglandins and tranexamic acid (an antifibrinolytic agent) have not been shown to benefit patients with ulcer hemorrhage.

[TOP]

Endoscopic Stigmata of Ulcer Hemorrhage

After patient stabilization, endoscopy is the preferred procedure for diagnosis and treatment of UGI hemorrhage because of its high accuracy and low complication rate. Endoscopy is diagnostic in almost 95% of patients with severe UGI hemorrhage. Endoscopy also may show stigmata of recent hemorrhage on ulcers that have important prognostic value.8

[ZOOM]

By consensus, stigmata have been divided into those that reflect either active bleeding, manifested as (a) arterial spurting, (b) oozing, or (c) oozing beneath an overlying clot; or recent hemorrhage, such as (a) nonbleeding visible vessel, (b) overlying clot without oozing, or (c) flat spots or dark slough. The terms visible vessel and pigmented protuberance usually refer to an elevated red, blue, purple, or white plug or mound, protruding from an ulcer base, that resists washing.9 The visible vessel is usually visualized as a small (less than 4 mm in diameter), smooth, organized structure distinguishable from a clot. The visible vessel is usually single and often is found in the center of the ulcer crater (Figure 5.1). The color, shape, and size of a nonbleeding visible vessel can be quite variable. In a deep duodenal ulcer or in a posterior gastric ulcer, a visible vessel can be pulsatile.

Pathologically, the visible vessels in resected gastric ulcers correlate with a small organized clot projecting above an ulcer base, rather than a protruding vessel. This is sometimes referred to as a sentinel or pigmented clot.

The implications of these stigmata of recent hemorrhage have been confirmed in several controlled clinical trials of endoscopic hemostasis. In one study, active spurting ulcer bleeding (Figure 5.2) seen at endoscopy (but not treated endoscopically) was associated with a need for emergency surgery in 69% of patients with a 23% mortality rate. In contrast, a rebleeding rate of 30 to 33% was reported for ulcers oozing slowly at endoscopy and treated medically. In another study, an ulcer with a nonbleeding visible vessel had a 50% chance of rebleeding on medical therapy during hospitalization (Table 5.4). Other endoscopic stigmata of recent hemorrhage, such as an overlying clot without oozing (Figure 5.3) or flat spots or dark slough, had a less than 10% chance of rebleeding on medical therapy (Figure 5.4).

[ZOOM]	[ZOOM]
[ZOOM]	[ZOOM]

[ZOOM]

Recent analyses of medically treated patients in randomized, controlled trials by the Center for Ulcer Research and Education (CURE) Hemostasis Research Group have yielded somewhat different rebleeding rates, according to stigmata of ulcer hemorrhage, in comparison to other studies.2 These are summarized in Figure 5.5. In prospective CURE studies, patients with active arterial hemorrhage had a greater than 90% rate of continued bleeding or rebleeding when treated medically over the last 3 decades, no matter what the medical therapy. In the Hong Kong series, for less severe active bleeding described as oozing bleeding without a clot or visible vessel, patients had a 27% rate of continued bleeding, whereas in CURE studies, continued bleeding occurred in only 10% of patients with oozing bleeding and no other ulcer stigmata. For nonbleeding adherent clots, CURE patients on medical management had a 33% rebleeding rate in comparison to a 10 to 15% rate for other clinical trials. The prevalence of stigmata of ulcer hemorrhage and their related outcomes in ICU patients with severe ulcer bleeding are summarized on Table 5.4.

Recent experience has shown that approxi-mately 20% of patients with severe ulcer bleeding who require ICU admission have a nonbleeding visible vessel. Approximately 50% of these patients rebleed during hospitalization when treated medically, and 40% with severe hemorrhage, who were not treated with endoscopic hemostasis, required ulcer surgery for hemostasis. Only about 15% of ulcer patients treated medically without endoscopic hemostasis have active bleeding (ranging from arterial spurt to oozing) at endoscopy. Of those treated medically with arterial bleeding, 80% had continued bleeding, and 70% needed surgery. The data on patients who have oozing bleeding without other stigmata are variable, with less than 10% in CURE studies having further bleeding, compared to 27% of oozers with further bleeding in Hong Kong studies. Most investigators distinguish arterial bleeders from oozers. Less than 5% of patients presented with shock or had ulcers that are inaccessible to the endoscope. Most of these patients continued to bleed or rebleed and needed surgery.

In the patient with severe UGI bleeding, resuscitative measures should be started immediately, and the patient should be hemodynamically stable before initiating emergency upper endoscopy. Hypotension or shock, hematemesis of red blood, and red blood per rectum are indicators of significant blood loss. Endoscopy in such patients should be performed within 4 to 8 hours after admission to an ICU and/or when there is an adequate resuscitation with fluid and blood products. Endoscopy is done sooner in patients who continue to bleed despite resuscitative measures, who rebleed in hospital, or who have cirrhosis. This approach allows the majority of patients with UGI hemorrhage to stop bleeding and results in the evaluation of a subgroup of patients with more severe UGI bleeding. Table 5.4 summarizes recent experience with endoscopic appearance of bleeding peptic ulcers and subsequent clinical outcomes based on this approach. Carefully performed endoscopy will provide an accurate diagnosis of the source of the UGI bleed and reliably identify those high-risk subgroups that will benefit most from endoscopic hemostasis.

Both endoscopic and clinical criteria have been used to determine which patients with acute ulcer hemorrhage could be managed as outpatients or considered for early discharge after endoscopy. Several scoring systems have been reported that are based on factors associated with recurrent bleeding or death. Shock, advanced age, severe comorbid illnesses, and endoscopic findings of active arterial bleeding, nonbleeding visible vessel, and large ulcer (over 2 cm) suggest a higher probability of rebleeding and death. The absence of these high-risk factors has been incorporated into guidelines to identify patients who can be safely managed as outpatients or be discharged early from the hospital. Prompt endoscopy (within 24 hours) is essential if such guidelines are to be successful. Additional valuable criteria include the availability of adequate home support for patients judged to be at low risk for ulcer rebleeding. Using these scoring systems, several authors have shown that a significant proportion of patients with acute ulcer hemorrhage can be safely managed as outpatients. Similar guidelines have been used to predict who could be safely discharged early from the hospital, thus reducing the overall length of stay for patients with ulcer bleeding.10

[TOP]

Endoscopic Therapies for Ulcer Hemostasis

[ZOOM]

An ideal endoscopic hemostatic technique should possess the following features: (a) reproducible effectiveness; (b) easy and rapid application; (c) low complication rate; (d) relatively low expense; (e) portability to the bedside; and (f) widespread availability. Endoscopic techniques have been grouped into 3 general types of methods and are categorized (Table 5.5) according to whether or not tissue contact is required to achieve hemostasis.8 A combination group is considered separately for epinephrine injection plus thermal or mechanical treatment.

Thermally Active Methods

The most widely tested modalities of endoscopic hemostasis are the thermal devices. Heating can be produced by tissue absorption of laser light energy, passage of electrical current through tissue, or heat diffusion from another source. Heat causes edema, coagulation of tissue proteins, and contraction of arteries. An important added effect with a thermally active contact probe, such as the bipolar circumactive probe or heater probe, is its ability to compress the target artery before heat delivery. Vessel compression reduces the loss of heat associated with blood flow (heat sink effect). Initial compression at the bleeding point or nonbleeding visible vessel followed by delivery of a small amount of heat welds the vessel walls together (coaptation). Under laboratory conditions, mesenteric arteries up to 2.5 mm in diameter could be coagulated with a bipolar or heater probe if arterial blood flow was decreased effectively by tamponade before coagulation.

Noncontact thermal techniques include Nd: YAG (neodymium-yttrium aluminum garnet) and argon laser, argon plasma coagulation, and microwave. Lasers have been extensively investigated for endoscopic hemostasis, but currently their use for ulcer hemostasis is mainly of historical interest. Other techniques have surpassed lasers because of their lack of portability to the bedside, high cost, and risk of perforation.

Argon plasma coagulation (APC) is a noncontact form of electrocoagulation in which argon gas forms a plasma and conducts monopolar current to the treatment site. One small study suggested that APC was comparable to heater probe for hemostasis of bleeding ulcers. In view of the superficial coagulation (1 to 2 mm) of APC and the fact that tamponade of arteries is not possible with this device, it is unlikely that APC will be extensively used for ulcer hemorrhage. APC contacting the tissue acts as a monopolar probe, which can coagulate deeper but also increases the risk of complications of radiofrequency monopolar coagulation.

Heater Probe

Mechanism of Action The heater probe is a hollow aluminum cylinder with an inner coil. The aluminum cylinder effectively transfers heat from its end or sides to tissue. This allows heat transfer whether the probe is applied perpendicularly or tangentially and facilitates probe placement in areas where endoscopic access is difficult. The probe can be passed through the biopsy channels of therapeutic endoscopes and placed directly on the bleeding lesion to produce both a tamponade effect and heat. Clots or overlying blood can be irrigated with an axial water spray. The tip is Teflon coated to diminish sticking. The probe is available in both small (2.4 mm) and large (3.2 mm) diameter sizes.

Technique Laboratory studies at CURE revealed that effective coagulation of gut arteries up to 2.5 mm in diameter could be obtained with firm tamponade followed by coagulation with the heater probe. It was realized that high bond strength of the arteries could be obtained when a total of 100 to 120 joules (J) was delivered in a succession of 4 pulses of 30 J each. The present recommendations for heater probe coagulation of patients with clinically severe ulcer hemorrhage and major stigmata are to use (a) the large heater probe (3.2 mm), (b) double-channel therapeutic endoscope, (c) firm tamponade directly on the bleeding point or the visible vessel, and (d) coagulation with at least 120 J (4 pulses of 30 J each) before changing the position of the probe.

Electrocoagulation

High-frequency electrical current generates heat that can coagulate or cut tissue. Thermal electrocoagulation is the classic treatment for bleeding during surgery. This has been used for several years to treat GI bleeding via endoscopy. Two types of endoscopic electrodes are currently available

Monopolar Probes

The monopolar probe uses an electrical current that passes from the electrode tip through the patient's body to a grounding plate. Although monopolar electrocoagulation generators are readily available, and endoscopic coagulation by experts with monopolar probes appears to be relatively effective, these devices have not become widely used by gastroenterologists for GI hemostasis. Reasons include the availability of newer and improved methods, such as the heater probe or bipolar electrocoagulation, and difficulties with the device, such as (a) sticking, (b) inconsistent coagulation, (c) deep coagulation pattern, (d) difficulty with standardization, (e) vessel erosion, and (f) perforation or transmural injury.

Multipolar (or Bipolar) Probes

Mechanism of Action In bipolar (or multipolar) electrocoagulation, the current flows between 2 or more electrodes separated by 1 to 2 mm at the probe tip. The advantage of this probe is that the current is concentrated much closer to the tip than in the monopolar probe, and therefore a distant patient plate is not required. This results in less depth of tissue injury and lower perforation potential. The commercially available multipolar probes have several electrode plates in an array at the tip and a central irrigation channel.

Technique The technique of multipolar electrocoagulation has not been standardized outside of randomized prospective trials, and there are many differences among the reported clinical trials. Laine described 4 major variables in technique: size of the probe, power setting, duration of electrocoagulation, and appositional pressure applied to the lesion. In canine arteries, a large probe (3.2 mm) produced better hemostasis than did a smaller (2.3 mm) probe. The greatest depth of coagulation in these studies was achieved with low BICAP (ConMed Corporation) settings of 3 to 5 (on a BICAP I generator) or 12 to 16 watts (on a -BICAPII generator). It is postulated that at the lower settings more energy is provided to the tissue before desiccation occurs, which then disengages the electrical flow between the electrodes in the tip of the BICAP. Higher hemostatic bond strength is obtained by increasing the length or number of BICAP pulses. The fourth major variable is the appositional pressure applied to the bleeding lesion. Increased pressure on the bleeding site results in higher hemostatic bond strength and depth of coagulation. Laine therefore recommends using the large periods of coagulation such as 14 seconds, or 7 pulses of 2 seconds. Jensen and colleagues use a dial setting of 3 to 4 with 10 second pulses on a 50 watt BICAP generator or 20 watts and 10 second pulses on a BICAP II generator. A newer bipolar device, the injector-Gold probe (Microvasive, Boston Scientific), which is different in tip design and stiffer than the BICAP probe, has been shown to be effective at low power settings and longer pulse duration under both laboratory conditions and comparative randomized trials.

Injection Treatments

Injection therapy for nonvariceal causes for UGI bleeding has been advocated for reasons of simplicity, -relative inexpense, and wide clinical experience with variceal hemorrhage. There have been several recent trials showing the efficacy of injection treatments for endoscopic hemostasis of bleeding ulcers. Most have used initial injection of major stigmata and repeat endoscopy at 24 hours, with further treatment of endoscopic bleeding or other stigmata of ulcer hemorrhage.

Mechanism of Action Laboratory experiments have shown that epinephrine at a concentration of 1:10 000 or 1:20 000 in saline is effective and safe. It provides local tamponade, vasoconstriction, and possibly improved platelet aggregation. Because saline injection is also an effective therapy in treating bleeding ulcers, the predominant mechanism may be local vessel compression or tamponade.

Other solutions such as ethanolamine, polidocanol, sodium tetradecyl sulfate, and alcohol have been used as sclerosants. These agents can cause tissue necrosis and may result in complications. Ethanol, in particular, causes dehydration and tissue fixation and predisposes to ulceration and possible perforation.

Technique The technique of injection varies for nonvariceal UGI lesions. A sclerotherapy catheter with a small (25 gauge) retractable needle is passed through the biopsy channel of an endoscope. To treat a lesion such as a nonbleeding visible vessel, a sclerosing -solution, usually ethanol, is injected in increments of 0.1 to 0.2 ml per injection at 3 or 4 sites around the visible vessel, aiming 1 to 3 mm from the vessel. A maximum of 1.5 to 2.0 ml of ethanol is injected, because greater volumes of ethanol are associated with a high rate of extensive tissue injury. Treatment of an actively bleeding ulcer is may not be as effective with injection therapy. One approach uses a dilute epinephrinesaline solution (1:10 000) injected in 1.0 ml increments in all 4 quadrants around the bleeding point, up to a total of 20 to 25 ml or until hemostasis is achieved. This is then followed by injections of ethanol as described above. Another injection technique uses single agents such as ethanol, polidocanol, or adrenalinesaline solution injected directly into or around the bleeding point. Most non-U.S. studies have advocated initial therapy (usually with epinephrine alone, such as in various Hong Kong studies) and surveillance endoscopy in 24 hours with retreatment of endoscopic stigmata of ulcer hemorrhage. This may not be feasible in many countries because of the expense of repeating endoscopies in all patients, including those who do not have clinically significant rebleeding. Injection alone results in substantially higher ulcer rebleeding rates than does combination therapy.

Combination Injection and Thermal Coagulation

Combination treatment with epinephrine injection and thermal therapy (multipolar or heater probe) has theoretical advantages because epinephrine and thermal coagulation have different mechanisms of action for arterial hemostasis. Epinephrine injection produces vasoconstriction and activates platelet coagulation to reduce blood flow, which potentiates the coaptive coagulation produced by thermal energy. The introduction of an injector Gold probe, which combines a retractable sclerotherapy needle with a stiff multipolar probe, facilitates combination treatment. In laboratory studies of bleeding ulcers, hemostasis with the injector Gold probe is faster and more convenient than either separate injectors or thermal probes, and its efficiency is excellent. This has been confirmed in clinical applications of ulcer hemostasis, such as for treatment of active arterial bleeding or adherent clots, where combination therapy is particularly recommended (Table 5.6).

[ZOOM]

Technique The technique is 4-quadrant injection of epinephrine (1:10 000 concentration) in 0.5 to 1.0 ml aliquots, then coagulating with large-diameter(10 French) multipolar probe directly on the bleeding point (or major stigmata), applying firm tamponade pressure until the bleeding stops, then coagulating with long pulses (7 to 10 seconds in duration) at low power settings (16 to 20 watts). The probe is then slowly removed from the ulcer with gentle irrigation to reduce adherence of coagulated tissue, and coagulation is repeated as necessary to stop hemorrhage and flatten any underlying visible vessel.

The injector Gold probe with a small retractable sclerotherapy needle within a multipolar probe permits sequential target irrigation, needle extrusion and epinephrine injection, needle retraction, further target irrigation, tamponade, and coagulation without removing the probe from the endoscope.

Mechanical Methods

Several devices, including metallic clips, sewing devices, rubber band ligation, and endoloops, have been described for the mechanical endoscopic treatment of bleeding ulcers. Hemoclips have been the most extensively studied of the mechanical methods and provide another option in the management of patients with UGI hemorrhage. The hemoclips produce hemostasis similar to surgical ligation. Results from clinical -trials have been conflicting. A trial comparing hemoclips alone, injection therapy or combination treatment, found similar initial failure and recurrent bleeding in ulcer patients among the 3 treatment groups. However, hemoclips had a significantly higher overall failure rate. In another study comparing hemoclips to heater probe, hemoclip therapy significantly decreased rebleeding, transfusions, and hospital stay. A recent trial showed similar ulcer-bleeding related outcomes (primary hemostasis, rebleeding, transfusion requirement, and duration of hospitalization, surgery rates, and mortality) for both hemoclips and combination therapy with epinephrine injection and multipolar electrocoagulation.11

Several factors may be responsible for these divergent results. In laboratory experiments, some commercially available clips were unsuccessful in stopping hemorrhage from 1 to 2 mm arteries due to an inability to provide adequate compression. Hemoclips may be more difficult to place along the high lesser curve or posterior duodenal bulb or over a scarred chronic ulcer base. Other sources of UGI hemorrhage, such as Dieulafoy lesions and Mallory-Weiss tears, have been successfully treated with hemoclips.3 Overall, although experience is limited, hemoclip use for the treatment of ulcer bleeding appears promising (Figure 5.6).

[ZOOM]

Band ligation and endoloops have not been extensively evaluated for treating ulcer bleeding. Although well described and effective for the treatment of variceal bleeding, these mechanical devices are not as well suited for ulcer application because ulcers often have firm bases that preclude suctioning of the ulcer and successful band deployment.

[TOP]

Which Hemostatic Techniques Should Be Used for Ulcer Hemorrhage?

Active Arterial Bleeding

[ZOOM]

We recommend combination therapy with epinephrine injection and a thermal probe, particularly for active arterial bleeding and adherent clots8 because this therapy combines different mechanisms of arterial -hemostasis such as vasoconstriction, activation of platelet coagulation, and coaptive coagulation. Epinephrine injection may be with a standard sclerotherapy catheter or a combination device. The injector Gold probe reduces the treatment time compared to 2 separate catheters in laboratory and clinical settings.

Two groups of investigators have reported significantly better results with combination therapy (epinephrine and thermal probe) than monotherapy for active arterial bleeding. In contrast, combination therapy compared to epinephrine injection alone did not improve outcomes of patients with oozing ulcer bleeding in one large study.

Successful endoscopic hemostasis is achieved in nearly 100% of patients with combination therapy for treatment of chronic-appearing ulcers with active arterial bleeding (Figure 5.7). However, the rebleeding rate for high-risk elderly patients after combination therapy (injector Gold probe) in a CURE randomized prospective study was approximately 20%. On medical therapy alone, similar patients with spurting ulcer bleeding in earlier trials had continued bleeding or rebleeding 85 to 95% of the time, and ulcer surgery was required for hemostasis in 75% of patients.

Nonbleeding Visible Vessel in an Ulcer

For the nonbleeding visible vessel, monotherapy with either a multipolar or heater probe directly on the visible vessel with firm tamponade (for chronic ulcers) coagulates the underlying vessel and significantly reduces rebleeding rates compared with medical therapy alone. Optimum results are achieved by applying firm tamponade, a large-diameter probe, slow coagulation with a low power setting, and treatment until the visible vessel is flat (Figure 5.8). In comparison to a 50% rebleeding rate with medical therapy alone, endoscopic coagulation with large thermal (heater or multipolar) probes reduced rebleeding to less than 15%. In most recent studies, combination endoscopic therapy (with epinephrine plus thermal probe) (Figure 5.9) was not better than thermal -therapy alone for nonbleeding visible vessels.

[ZOOM]

[ZOOM]

Nonbleeding Adherent Clot on an Ulcer

Elderly patients with chronic ulcers and comorbidity have a high (about 33 to 40%) risk of rebleeding from a nonbleeding adherent clot while receiving medical therapy alone. Most adherent clots are hooked onto an underlying vessel, and the clot obscures an underlying stigma or clean ulcer base. Most often, the underlying stigma is a nonbleeding visible vessel (Figure 5.10) rather than a minor stigmata (flat spot or clean ulcer base). Therefore, a safe method of clot removal could theoretically help identify the underlying stigmata and facilitate application of an endoscopic treatment. However, forceful clot removal may precipitate bleeding. Thus, the optimal therapeutic approach to bleeding peptic ulcers with an adherent clot has been controversial and has included medical treatment alone or endoscopic management. CURE studies indicate that combination therapy is superior to monotherapy (either heater probe coagulation or injection therapy) in treating patients with nonbleeding adherent clots. Monotherapies have rebleeding rates similar to control: 33 to 38%. In contrast, combination therapy had very low rebleeding rates compared with medical therapy in randomized prospective blinded studies. After injection of epinephrine (1:10 000 dilution) in 4 quadrants around the base of the clot in 0.5 to 1.0 ml aliquots, a clot that cannot be removed by irrigation or suctioning after the initial epinephrine injection is snared with a polypectomy snare. It is cold guillotined off in sections without pulling off the attachment until it is shaved down to a 2 to 4 mm clot at its attachment. This remnant is then treated with thermal coagulation (Figure 5.11).

[ZOOM]

[ZOOM]

This combined technique to remove adherent clots does not induce severe bleeding. In a CURE multicenter trial, combination therapy significantly reduced the rebleeding rate to zero in comparison to 31.4% with medical therapy alone.12 Combination therapy represents a considerable improvement over monotherapies that were studied previously in a CURE multicenter study. The rebleeding rates after heater probe (33%) or sclerotherapy (38%) were similar to that for medical therapy alone (33%) in that study. A recent meta-analysis of 6 studies of adherent clots on ulcers, comparing endoscopic therapy including clot removal and hemostasis with injection therapy and/or thermal or electrocoagulation to no endoscopic intervention, concluded that endoscopic management was superior in preventing recurrent bleeding.13

Ulcer Oozing without Other Stigmata of Hemorrhage

Minor bleeding from an ulcer base or edge is often classified as "active ulcer hemorrhage." In most reports, this type of bleeding accounts for 75 to 80% of patients treated as active ulcer bleeders. Usually this type of ulcer bleeding is self-limited and markedly different than spurting ulcer hemorrhage or active bleeding associated with a clot or visible vessel. If oozing from an ulcer base persists despite targeted water irrigation and observation, monotherapy with either thermal probes or epinephrine injection is recommended for definitive hemostasis.

Rebleeding rates of ulcers with persistent oozing during irrigation treated medically varied from less than 10% (CURE Hemostasis Research Group) to as high as 27% (Hong Kong Group). Epinephrine injection or thermal probes reduce rebleeding rates to less than 5%. Combination epinephrine and thermal therapy was reported to be no better than injection alone for hemostasis of oozing bleeding.

Flat Spots or Clean-Based Ulcers

[ZOOM]

Patients with flat spots treated medically have a rebleeding rate less than 7%. They do not benefit from endoscopic hemostasis. Patients with UGI bleeding who have clean-based (Figure 5.12) ulcers without stigmata of hemorrhage have a rebleeding rate of less than 3%. Early feeding does not adversely affect patient outcomes in comparison to a several-day delay in resuming oral intake. The findings of clean-based ulcers and other minor stigmata of hemorrhage (such as flat spot or gray slough) have also been used to triage patients to early discharge. In selected patients presenting with UGI hemorrhage without significant hypovolemia, severe coagulopathy, liver disease, or severe comorbidity and who are reliable and agree to return for outpatient follow-up, hospitalization may be safely avoided with additional cost savings.

Safety of Endoscopic Hemostasis

In laboratory studies, thermal probes and epinephrine injection are very safe. A study that compared monotherapy (multipolar or heater probe or epinephrine injection) with combination treatment (injection and coagulation with a heater or multipolar probe) for bleeding gastric ulcers showed no complications such as bleeding or perforation.

Potential complications include perforation or precipitation of bleeding from a nonbleeding visible vessel. In a meta-analysis of injection or thermal probe coagulation, hemorrhage was induced in 0.4% of patients, and perforation was induced in 0.7%. Perforations are more commonly reported with heater probe than multipolar probe. Perforations are also more frequent after endoscopic retreatments.

Second-Look Endoscopy

Some studies suggest that repeat (second-look) endoscopy for repeat endoscopic hemostasis 24 hours after the initial endoscopic therapy helps reduce rebleeding. Others have questioned this practice because outcomes for rebleeding rate, transfusion requirement, hospital stay, surgical intervention, and mortality in patients receiving frequent repeat endoscopy and endoscopic retreatment compared with conservative management with repeat endoscopy only for recurrent hemorrhage were similar. The Baylor group has recommended selective second-look endoscopy only in high-risk patients, based on a scoring system.

Routine second-look endoscopy for surveillance is expensive and unjustified in all patients, particularly these at low risk for recurrent hemorrhage. Clinical judgment or a scoring system such as the Baylor score should guide the decision about repeat endoscopy and hemostasis in selected high-risk patients.

Retreatment

Rebleeding after endoscopic therapy occurs in 10 to 25% of patients and represents a challenging management problem.14 Two separate studies reported that ulcers greater than 2 cm in diameter and active bleeding at endoscopy are independent predictors of failure of hemostasis therapy. A prospective, randomized study compared endoscopic treatment with emergency surgery in patients who rebled after initial successful ulcer hemostasis. After initial successful combination therapy of bleeding ulcers in 92 patients with epinephrine injection and heater probe, 48 patients who rebled were randomized to repeat endoscopic treatment, and 44 other patients were assigned to emergency ulcer surgery. In comparison to surgery, repeat endoscopic therapy produced lower definitive hemostasis rates (73% versus 93%), but had fewer severe complications (14.6% versus 36.4%). The transfusion requirements and hospital stay were similar in the 2 groups. These results, suggest that repeat endoscopic therapy is warranted for rebleeding after initial hemostasis for ulcer hemorrhage.

[TOP]

Surgical Therapy

If endoscopic hemostasis fails or is unavailable for patients with continued severe ulcer bleeding or recurrent hemorrhage, surgery becomes a therapeutic option. The main purpose is to prevent exsanguination, with a secondary aim being to prevent recurrent hemorrhage and ulceration. Emergency surgery has a higher mortality rate than elective procedures, and procedures involving gastric resection have a higher mortality rate than oversewing the ulcer and highly selective vagotomy or vagotomy and pyloroplasty.

In recent CURE multicenter studies, fewer than 5% needed emergency surgery if endoscopic hemostasis was used to treat high-risk patients with active bleeding, visible vessels, or adherent clots. A prospective, randomized study showed that in patients with peptic ulcers and recurrent hemorrhage after initial endoscopic coagulation, endoscopic retreatment reduced the need for surgery and was associated with fewer complications than surgery. Endoscopic therapy is also considerably less expensive than emergency surgery for patients with bleeding ulcers.

Specific guidelines have been created to help in the decision making regarding when to consider surgical treatment. Criteria that are based solely on number of units transfused, and do not consider complete clinical conditions, are inappropriate. Studies comparing early versus later surgical intervention suggested a small benefit for early surgery, especially for patients older than 60 years. A study in an indigent patient population described the beneficial outcomes of a unified team approach, including early endoscopy (within 24 hours) and early surgery in the management of acute GI bleeding associated with peptic disease. Both the death rate and transfusion requirements were lowered.

The choice of operation is related to the surgeon's experience, location of the ulcer, and overall condition of the patient. Historically, truncal vagotomy and antrectomy provided high cure and low recurrence rates (under 3% in some series). Recurrence rates after vagotomy and pyloroplasty or highly selective vagotomy are higher (2.0 to 2.5% per year).

With the application of laparoscopic techniques to ulcer surgery, the laparoscopic approach could become even more appealing to a small subset of ulcer patients likely to benefit from surgery. One study reported that laparoscopic, highly selective vagotomy was a safe and effective alternative to conventional surgery. There was decreased morbidity, shorter hospitalization, and less recovery time. Peptic ulcer bleeding can be treated effectively with surgery and is safer than some other alternative treatments such as angiography.

[TOP]

Angiographic Therapy

Angiography is potentially useful as a diagnostic and therapeutic modality. Its main role is in patients with severe bleeding in whom endoscopy has failed to reveal the bleeding site or in whom endoscopic hemostasis is unsuccessful. It should also be considered in patients who are at high risk for surgical intervention. Angiography can identify the site of bleeding if blood loss is occurring at a rate of 0.5 ml or more per minute. Extravasation will appear as a blush, which with DUs or GUs may remain localized or spread to a more dependent position.

Angiographic therapy of GI bleeding includes 2 different techniques, vasopressin infusion and embolization. Intra-arterial infusion of vasopressin causes vasoconstriction, producing cessation of ulcer hemorrhage in about 50% of cases. Embolic materials, such as -absorbable gelatin sponge (Gelfoam), tissue adhesives, or other occlusion devices, can be selectively injected through a catheter into the bleeding artery. Potential complications of embolization therapy include ischemia and perforation. In one study of angiographic treatment of patients with severe UGI bleeding -(average -transfusion requirement nine units per patient), extravasation occurred in 40% of cases. Selective vasopressin infusion and embolization were similarly successful.

[TOP]

Therapies for Long-Term Prevention of Ulcer Hemorrhage

In a patient with at least 1 prior documented ulcer hemorrhage, there was 36% incidence of rebleeding during a follow-up period which averaged 61 weeks if the patient did not take maintenance medical therapy. Three separate approaches have been shown to decrease the incidence of recurrent ulcer hemorrhage: long-term maintenance acid suppression, eradication of H. pylori, and ulcer surgery. Patients should also be educated about and instructed to stop all forms and doses of aspirin and NSAIDs if not medically indicated. These are independent risk factors for both ulcer indication and bleeding.

For patients with GUs complicated by GI bleeding, a follow-up endoscopy to document healing and to exclude gastric cancer is appropriate. Follow-up endoscopy after a bleeding DU is controversial but may be warranted in high-risk patients (Table 5.2) to document ulcer healing prior to discontinuing therapy or changing to a maintenance regimen.

[TOP]

Medical Therapies

Medical therapy with continuous H2-receptor antagonists (H2RA) effectively decreases the long-term rate of recurrent ulcer bleeding after an index hemorrhage and documented ulcer healing. CURE prospective studies have evaluated such patients during long-term follow-up for up to 7 years. With maintenance dosing of H2RAs (half the therapeutic dose, that is, ranitidine 150 mg or famotidine 20 mg hs), recurrent GU or DU bleeding rates were 10 to 15% during 3-year follow-up, and symptomatic ulcer recurrence rates without bleeding were 20 to 25%. In a placebo-controlled randomized study, 36% of untreated patients had recurrent DU hemorrhage in comparison to 9% of ranitidine-treated patients. Full-dose H2-blocker maintenance (that is, ranitidine 300 mg or famotidine 40 mg) was even more effective-with rates for recurrent hemorrhage of only 4.5% for DUs and 3.8% for GUs during follow-up over 24 months. These results were similar to those observed following surgery for the index hemorrhage. About 25% of ulcers recur on low-dose acid suppression; we now suggest full-dose maintenance therapy with a PPI or full-dose H2RAs.

[TOP]

Helicobacter Pylori Eradication and NSAID Discontinuation

Recent studies have shown that the prevalence of H. pylori infection is less frequent in patients with ulcer hemorrhage (DUs 60 to 70%, GUs 50 to 60%) than in patients with uncomplicated ulcers. There may be a component of false-negative results because rapid urease tests appear to be less sensitive in the presence of UGI bleeding. We recommend that H. pylori status be evaluated with other methods, such as serology or histology of the gastric mucosa, in acute UGI hemorrhage.15 Empiric H. pylori treatment will not prevent recurrent ulcer hemorrhage in many ulcer patients. However, studies have shown that when H. pylori infection was documented, eradication was associated with a marked decrease in ulcer recurrence. Several trials have reported the effects of H. pylori eradication on preventing recurrent ulcer hemorrhage. Although small sample size, short follow-up, absence of double-blind or controlled protocol, and continuation of NSAID ingestion limited most studies, the trends are very encouraging. After H. pylori eradication, both symptomatic ulcer recurrences and recurrent ulcer hemorrhage decreased. The studies, however, show a wide range of risk reduction for bleeding rates varying from 2.5 to 33%. Good results have been recently confirmed by a large multicenter study with long-term follow-up. In a recent large randomized, double-blind CURE multicenter study of H. pylori eradication alone compared to eradication plus daily H2RA, there was a very low recurrence of symptomatic or bleeding GU or DU in both treatment groups. Therefore, if NSAIDs are not ingested, rebleeding can be prevented by H. pylori eradication alone in patients with an index bleed and ulcers etiologically related to H. pylori infection.

As previously described, NSAIDs are an important cause of hemorrhage associated with peptic ulcer. CURE studies have shown that about 60% of GU or DU patients with hemorrhage consumed aspirin or other NSAIDs prior to their bleed, compared with about 20% of similar patients with nonulcer UGI hemorrhage. Discontinuing the aspirin or NSAIDs after ulcer bleeding would be ideal, but many patients require chronic administration of aspirin for cardiovascular protection or NSAIDs for inflammatory conditions. Two different strategies are used to reduce the risk of rebleeding in patients with prior ulcer bleeding and need for continued NSAID ingestion.

Proton-pump inhibitors protect against ulcer formation in patients taking NSAIDs. Omeprazole 20 mg or 40 mg daily has been shown to be as effective as misoprostol (800 µg daily) at healing ulcers and was superior in maintaining remission than misoprostol (400 µg daily) in patients on long-term NSAID therapy. Misoprostol was associated with more adverse events than omeprazole. Another trial showed that omeprazole (20 or 40 mg daily) was more effective than ranitidine (150 mg b.i.d.) in healingulcers and maintaining remission, with similar adverse event rates. Randomized trials have shown that PPIs reduced recurrent ulcer hemorrhage in patients taking nonselective NSAIDs.16

Another option is use of COX-2-specific inhibitors, which decrease upper GI complications in comparison with standard NSAIDs. A study comparing a COX-2 inhibitor (colecoxib 200 mg b.i.d.) with a nonselective NSAID (diclofenac 75 mg b.i.d.) plus omeprazole (20 mg daily) in reducing recurrent -ulcer bleeding in arthritis patients showed that the 2 interventions were similarly effective.17 However, 5 to 6% of the study population had rebleeding after 6 months, suggesting that neither treatment option was completely effective. In patients with prior ulcer hemorrhage and continued NSAID or aspirin use who also require concomitant anticoagulation or corticosteroids, long-term PPI and COX-2 inhibitors are beneficial.

NSAIDs and H. pylori infection may coexist but do not, necessarily, interact. However, H. pylori eradication may protect patients who consume NSAIDs. For patients with both risk factors, the primary cause of ulceration and hemorrhage cannot be ascertained; therefore, eradication of H. pylori should be undertaken and NSAIDs discontinued. Maintenance acid suppression may be of benefit even after H. pylori eradication in patients with multiple risk factors and in whom recurrent NSAID or aspirin use is medically indicated. Daily PPIs appear to be the cotherapy of choice for secondary prevention of ulcer hemorrhage in high-risk patients.

Patients with ulcer bleeding and no history of NSAID consumption should be tested for H. pylori infection, and eradication is recommended if positive. Although not usually necessary for uncomplicatedulcers, patients with ulcer bleeding should have eradication of H. pylori confirmed 4 to 6 weeks after therapy. If the organism is still present, patients should be retreated.

[TOP]

Severe Hematochezia (Lower Gastrointestinal Bleeding)

Introduction

Acute lower gastrointestinal (LGI) bleeding, defined as bleeding from a site distal to the duodenum, most commonly from the colon, has an annual hospitalization rate of about 20 per 100 000 adults. For the clinician, while severe hematochezia is the actual presentation, the site of the lesion cannot be clinically determined as foregut, midgut, or colon. Even without a history of, or signs of, UGI lesions, approximately 15 to 20% of patients hospitalized with severe hematochezia have a foregut (UGI or proximal jejunum) source of the severe hematochezia. Based on these data, the term severe hematochezia is preferred. In most ambulatory patients with hematochezia, the bleeding stops spontaneously, permitting elective diagnostic evaluation. However, some patients with severe hematochezia require urgent attention to minimize further bleeding and complications. Mortality rates still range between 3 and 5% because the incidence of LGI bleeding increases markedly in the elderly, and these patients frequently have significant comorbidity.

For patients who present with severe hematochezia, the diagnostic and therapeutic approach is not standardized in most medical centers. However, we have evaluated and found a standardized approach that is effective, safe, and cost effective. During patient resuscitation, we recommend nasogastric aspiration to exclude a potential UGI source and, if negative (without blood or coffee grounds), rapid oral lavage to cleanse the colon, followed by urgent colonoscopy. Urgent colonoscopy provides an accurate diagnosis and, if required, an opportunity for hemostasis during the same examination. If urgent colonoscopy and anoscopy are not diagnostic for a bleeding site, push enteroscopy is recommended to rule out a potential small bowel site of hemorrhage.18 Although uncommon (3 to 5% of patients with GI bleeding), small bowel lesions such as vascular ectasia, tumors, ulcerative diseases, and Meckel's diverticulum should be suspected in the setting of GI hemorrhage with both negative upper endoscopy and colonoscopy with retrograde ileoscopy. This approach improves the diagnostic and therapeutic efficacy while reducing direct costs of patient care.19

Our primary criterion for proving (that is, classifying as "definitive source") that a lesion caused the bleeding is to identify stigmata of recent hemorrhage (such as active bleeding, nonbleeding visible vessel, or an adherent clot for colonic lesions or also flat spots for ulcers) on the lesion. A lesion is classified as the "presumptive cause" of the bleeding when fresh blood is in that location (such as the colon) or a lesion is found there without stigmata and no other likely bleeding sites are identified on colonoscopy, anoscopy, and enteroscopy. A lesion (such as diverticulosis found during the colonoscopy) is classified as "incidental" when more than one type of lesion is found but another lesion is the bleeding site based upon stigmata of recent hemorrhage or other evidence.

[ZOOM]

These diagnostic criteria were prospectively applied to study 300 consecutive patients hospitalized because of severe hematochezia. Although all patients were presumed to have lower GI hemorrhage at presentation, about 19% did not. Of the group, 15.3% had an UGI source (such as an ulcer, varices, or angiomas), 1.3% had a small bowel source, and 2.4% had no identified source.4 Table 5.7 lists the colonic sources of hemorrhage in this study.

The most common cause was diverticulosis (either presumed or definitive), and other common causes were internal hemorrhoids, ischemic colitis, rectal ulcers, delayed bleeding from postpolypectomy ulcers, colon polyps or cancer, and colon angiomas or radiation telangiectasia. The findings at urgent colonoscopy also permitted triage of patients. Low-risk patients (without stigmata of hemorrhage and/or severe comorbidity) were allocated to less intensive and less expensive care, and this often facilitated early discharge.

[TOP]

Severe Acute LGI Bleeding: General Measures

The first concern in patients with persistent, severe hematochezia should be to start aggressive resuscitative measures in a monitored case setting. A consultation with a general or GI surgeon should be obtained at an early stage to consider emergency surgery. Subsequently, a nasogastric (or orogastric) tube is recommended to determine whether UGI bleeding (coffee grounds, blood clots) is present. In the patient with persistent hematochezia, significant anemia, and hypotension, if the hemorrhage originates from the UGI tract, nasogastric (NG) evidence of bleeding is usually present. If there is bile without blood or coffee grounds in the nasogastric aspirate, a lesion proximal to the ligament of Treitz is unlikely when ongoing hematochezia is documented. In patients with hematochezia, return of clear fluid without bile should not be considered a negative NG aspirate. Because continuity has not been established between the NG tube in the stomach and the duodenum, the patient may have a duodenal ulcer or other duodenal lesions. In addition, 1 to 3% of patients who present with severe hematochezia have a small bowel source of hemorrhage.

Before preparation for emergency colonoscopy, tap water enemas are recommended to clear the distal colon and permit examination of the rectosigmoid colon and anal canal with anoscopy followed by flexible sigmoidoscopy with retroflexion in the rectum. This is particularly indicated in patients with a history of bleeding internal hemorrhoids, anorectal disease, or distal colitis. A rigid sigmoidoscopy is not adequate because there may be lesions in the blind area of the rectum that cannot be seen with a rigid instrument, and the extent of the examination is limited. However, with flexible sigmoidoscopy, an examination of the sigmoid and descending colon and retroflexed view of the rectum are both feasible.

If no evidence of UGI bleeding is found (for example, there is bilious return on NG lavage) and the flexible sigmoidoscopy and anoscopy do not reveal a rectosigmoid source of hemorrhage, cleansing the colon with an oral purge is recommended, followed by emergency colonoscopy in the ICU or monitored bed area when the colon is clear of stool and clots. Should the UGI endoscopy and colonoscopy not be diagnostic, then scintigraphy and angiography are warranted. For those patients who stop bleeding or present with less severe bleeding, colonoscopy within 24 hours of presentation should still be considered the initial diagnostic and therapeutic procedure of choice.

[TOP]

Scintigraphy

The threshold rate of GI bleeding for localization with radioisotope scanning is about 0.1 ml per min or more. Scintigraphy may be particularly useful for -identification of small bowel or colonic bleeding sites that are actively bleeding and at least moderately severe. Two different types of scintigrams are available: (a) sulfur colloid with technetium and (b) autologous red blood cells (RBCs) tagged with technetium. Sulfur colloid is rapidly cleared from the circulation after IV injection but may extravasate into the gut lumen during active bleeding and is not commonly used now. Repeat IV injection may be performed. In contrast, tagged RBCs stay in the vascular space for about 24 hours. Technetium-tagged RBC scans are more commonly used than sulfur colloid scans for patients with severe LGI hemorrhage suspected of having a small bowel source or whose site has not been found by emergency endoscopy and colonoscopy. In many institutions, scintigraphy has replaced emergency visceral angiography as an adjunct to colonoscopy because scintigraphy is more sensitive, is less expensive, and has significantly lower morbidity than angiography. Injection of labeled RBCs and early scanning (at least 30 min, 60 min, and 4 hrs) are recommended to identify potential bleeding sites. Because specific localization and etiologic diagnosis are not possible with RBC scanning, confirmatory examinations such as angiography and/or endoscopy are recommended prior to surgical exploration. Delayed scans (12 to24 hrs) are not reliable for localization in the gut, particularly for surgical exploration.

[TOP]

Angiography, MRI, and CT

If the rate of ongoing arterial bleeding is at least 0.5 ml per min, selective visceral angiography may show extravasation of contrast into the lumen to identify a bleeding site. Emergency visceral angiography can be useful for diagnosis and treatment of colonic, small bowel, or UGI lesions.

Abdominal computed tomography (CT) or -magnetic resonance imaging (MRI) may be helpful for diagnosis of an aortoenteric fistula in selected patients presenting with severe hematochezia. For patients with a previous diagnosis of severe peripheral vascular disease or abdominal aneurysm with or without surgery, the physician should consider performing one of these tests for diagnosis of severe hematochezia, if colonoscopy and endoscopy do not identify a bleeding site. Most patients with severe hematochezia do not require such diagnostic testing because they do not have large abdominal aneurysms or a past surgery for this diagnosis. Barium studies have no role in the assessment of severe hematochezia because they cannot show active bleeding and may obscure subsequent evaluation by colonoscopy or angiography.

[TOP]

Emergency Surgery

There are specific indications for emergency surgical intervention in patients with persistent or recurrent lower GI bleeding. Emergency surgery should be considered for patients with: (1) hypotension or shock, despite resuscitative efforts; and (2) continued bleeding with transfusion of 6 or more units of blood and no diagnosis by emergency endoscopy.

[TOP]

Specific Diagnoses

Diverticulosis

Diverticular bleeding is the most common cause for patients hospitalized with severe hematochezia. Although diverticulosis most often involves the sigmoid and descending colon, diverticular bleeding originates most frequently from the right colon (about 70%). Actively bleeding colonic diverticula have been treated with epinephrine injection, multipolar probe coagulation, and metallic clips. We have carefully evaluated endoscopic therapy to treat definitive diverticular hemorrhage,20 defined by the presence of active bleeding, a visible vessel, or an adherent clot associated with a specific diverticulum. Of 54 patients with severe LGI bleeding and diverticulosis, 20% had stigmata of diverticular hemorrhage. Patients with stigmata received epinephrine injections and/or bipolar coagulation during colonoscopy and had neither rebleeding nor complications. Their median discharge was 2 days after colonoscopy. All patients were placed on medical therapy, which consisted of fiber and stool softeners to control constipation; abstinence from aspirin, NSAIDs, anticoagulants, and gingko; and avoidance of small hard seeds, popcorn, and nutshells. Only about 3% of the patients rebled during a 30-month follow-up period.

We treat active bleeding or adherent clots (Figure 5.13) with a 1:10 000 epinephrine/saline solution in 1 to 2 ml aliquots for shallow, broad-based diverticula to tamponade.20 Adherent clots can be guillotined off with a snare, similar to peptic ulcers with adherent clots. After the bleeding stops and for nonbleeding visible vessels, multipolar coagulation with 10 to 14 watts of power is then applied using moderate pressure and a 1 second pulse duration until adequate coagulation is achieved. We also advocate india ink labeling of the diverticulum with the major stigmata (active bleed, visible vessel, or adherent clot) after successful endoscopic hemostasis (Figure 5.14). This will facilitate endoscopic treatment, surgery in case of early rebleeding, and histopathologic correlation.

[ZOOM]

[ZOOM]

Colonic Angiodysplasia

[ZOOM]

Bleeding colonic angiodysplasias most often occur in the right colon and are usually multiple (Figure 5.15). They may be associated with advanced age and medical conditions such as chronic renal insufficiency, cirrhosis, valvular heart disease, and collagen vascular disorders. Bleeding from angiodysplasia is usually mild and self-limited. The bleeding is usually intermittent and presents with chronic iron deficiency anemia. The CURE group performed randomized, prospective studies comparing patient outcomes for angiodysplasia treated with heater probe or multipolar coagulation. Patients usually required more than 1 session of endoscopic hemostasis to obliterate multiple colonic angiomas. The main risk of endoscopic coagulation of angiodysplasia is severe delayed bleeding and postcoagulation syndrome (which occurred in 3% and 1.7% of patients, respectively) treated with multipolar probe. In contrast, heater probe has been associated with an approximately 8% complication rate in patients with severe hemorrhage from right colon angiomas. No perforations occurred with these devices, although some have been reported after hot biopsy forceps, monopolar electrocoagulation, Nd-YAG, or APC hemostasis of right colon angiodysplasia.

[TOP]

Focal Ulcers or Colitis

Focal ulcers proximal to the sigmoid colon are an uncommon cause of severe lower gastrointestinal hemorrhage. In one large series, these accounted for the bleeding site in 9% of patients (Table 5.7). Bleeding colonic ulcers were caused by recent polypectomy with ulceration, inflammatory bowel disease (IBD), ischemic ulcers, or infectious colitis (such as pseudomembranous colitis or cytomegalovirus). The most common cause was delayed bleeding from an induced ulcer, 3 to 10 days after piecemeal polypectomy of a large sessile polyp. The majority of these patients resumed over-the-counter (OTC) aspirin, nonsteroidal anti-inflammatory agents, anticoagulants, or gingko after piecemeal polypectomy of large sessile lesions. For patients with severe hematochezia after recent polypectomy, we recommend an oral purge before colonoscopy. Colonoscopy usually reveals an ulceration at the site of a recent polypectomy with either active bleeding, adherent clot, or a nonbleeding visible vessel (Figure 5.16). We treat such stigmata with endoscopic hemostasis similar to peptic ulcer hemostasis treatment. Using this approach in patients with severe hematochezia, we found that 12 of 23 patients with rectal ulcers had major stigmata of recent hemorrhage such as acute bleeding, visible vessel, or adherent clot (Figure 5.17). The finding of major stigmata predicted severe bleeding and -increased bleeding recurrence and mortality. Despite endoscopic hemostasis with epinephrine injection and bipolar coagulation, bleeding recurred at a high rate, requiring repeat endoscopy and/or surgery.21

[ZOOM]

[ZOOM]

Colonic ulcers secondary to infection, ischemia, IBD, or NSAIDs are much less common causes of severe LGI hemorrhage. Urgent colonoscopy usually shows a diffuse mucosal process, and patients do not benefit from endoscopic hemostasis (Figure 5.18).

[ZOOM]

[TOP]

Internal Hemorrhoids

Internal hemorrhoids are the most common cause of LGI bleeding in outpatient ambulatory adults and are the second most common cause of severe hematochezia in patients hospitalized with presumed LGI hemorrhage in a recent CURE study. Most patients with internal hemorrhoids have self-limited, mild bleeding manifested by bright red blood on the toilet tissue. When such bleeding is recurrent and frequent, iron deficiency and microcytic anemia may result. However, internal hemorrhoids may bleed profusely and require hospitalization and emergency hemostasis.22

Prior to assuming that severe hematochezia is from more proximal lesions, the anal canal should always be examined by rigid anoscopy and, if this is not diagnostic, by retroflexed examination of the rectum with a flexible sigmoidoscope to identify internal hemorrhoids. With a slotted anoscope, active bleeding from internal hemorrhoids or stigmata of hemorrhage (such as an adherent clot or visible vessel) will confirm the diagnosis and facilitate bedside treatment. For inpatients, we often use rubberband ligation for emergency hemostasis of bleeding internal hemorrhoids22. Emergency colonoscopy can be obviated in such cases, although an elective colonoscopy might still be considered in patients at risk for concomitant polyps or colorectal cancer.

[TOP]

References

1. Rockall TA, Logan RF, Devlin HB, et al. Risk management after acute upper gastrointestinal hemorrhage. Gut. 1996;38:316-321.
2. Kovacs TOG, Jensen DM. Complications of peptic ulcer disease. In Dimarino AJ Jr, Benjamin SB, eds. Gastrointestinal Disease: An Endoscopic Approach. Thorofare, NJ: Slack; 2002:411-438.
3. Kovacs TOG. Mallory-Weiss tears, angiodysplasia, watermelon stomach, and Dieulafoy's: a potpourri. Techniques in Gastrointestinal Endoscopy. 2005;7:139-147.
4. Kovacs TOG, Jensen DM. Upper or small bowel hemorrhage that presents as hematochezia. Techniques in Gastrointestinal Endoscopy. 2001;3:206-215.
5. Khuroo MS, Yattoo GN, Javid G, et al. A comparison of omeprazole and placebo for bleeding peptic ulcers. N Engl J Med. 1997;336:1054-1058.
6. Lau JYW, Sung JJY, Lee KKC, et al. Effect of intravenous omeprazole and placebo for bleeding peptic ulcer. N Engl J Med. 2000; 343:310-316.
7. Bardou M, Toubouti Y, Benhaberou-Brown D, et al. Meta-analysis: proton pump inhibition in high-risk patients with acute peptic ulcer bleeding. Aliment Pharmacol Ther. 2005;21:677-686.
8. Kovacs TOG, Jensen DM. Recent advances in the endoscopic diagnosis and therapy of upper gastrointestinal, small intestinal and colonic bleeding. Med Clin N Am. 2002, 86:1319-1356.
9. Freeman ML, Cass OW, Peine CJ, et al. The non-bleeding visible vessel versus the sentinel clot: natural History and risk of rebleeding. Gastrointest Endosc. 1993;39:359-366.
10. Cipolletta L, Bianco MA, Rotondano G, et al. Outpatient management for low-risk nonvariceal upper GI bleeding: a randomized controlled trial. Gastrointest Endosc. 2002;55:1-5.
11. Saltzman JR, Strata LL, Di Sena V, et al. Prospective trial of endoscopic clips versus combination therapy in upper GI bleeding (PROTECCT-UGI Bleeding). Am J Gastroenterol. 2005;100:1503-1508.
12. Jensen DM, Kovacs TOG, Jutabha R, et al. Randomized trial of medical or endoscopic therapy to prevent recurrent ulcer hemorrhage in patients with adherent clots. Gastroenterology. 2002;123:407-413.
13. Kahi CJ, Jensen DM, Sung JJ, et al. Endoscopic therapy versus medical therapy for bleeding peptic ulcer with adherent clot: a meta-analysis. Gastroenterology. 2005;129:855-862.
14. Jensen DM. Treatment of patients at high risk for recurrent bleeding from a peptic ulcer. Ann Intern Med. 2003;139:294-295.
15. Kovacs TOG, Kovacs SVB. Role of Helicobacter pylori in gastric disease. In: Ashley SW, ed. Problems in General Surgery: Gastric Surgery. Vol 14. Philadelphia, PA: Lippincott-Raven; 1997:25-36.
16. Lai KC, Lam SK, Chu KM, et al. Lansoprazole for the prevention of recurrences of ulcer complications from long-term low dose aspirin use. N Engl J Med. 2002;246:2033-2038.
17. Chan FK, Hung LC, Sven BY, et al. Colecoxib versus diclofenoc and omeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. N Engl J Med. 2002;347:2104-2110.
18. Kovacs TOG. Small bowel bleeding. Current Treatment Options in Gastroenterology. 2005;8:31-38.
19. Jensen DM. Management of patients with severe hematochezia-with all current evidence available. Am J Gastroenterol. 2005;100:2403-2406.
20. Jensen DM, Machicado GA, Jutabha R, et al. Urgent colonoscopy for diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med. 2000;342:78-82.
21. Kanwal F, Dulai G, Jensen DM, et al. Major stigmata of recent hemorrhage on rectal ulcers in patients with severe hematochezia: endoscopic diagnosis, treatment, and outcomes. Gastrointest Endosc. 2003;57:462-468.
22. Jutabha R, Miura-Jutabha C, Jensen D. Current medical, anoscopic, endoscopic, and surgical treatments for bleeding internal hemorrhoids. Techniques in Gastrointestinal Endoscopy. 2001;3:199-205.

    [TOP]