39 Lower Intestinal Bleeding Disorders
Alexander Meier and Helmut Messmann
Gastrointestinal bleeding from the colon is a common reason for hospitalization, and endoscopic hemostasis is the daily challenge that must be mastered by gastroenterologists. For this, many different endoscopic techniques are available (injection therapy, hemoclips, thermal coagulation, topical hemostatic substances). Depending on the source of bleeding, the suitable and most effective method must be chosen. In this chapter, we discuss the diagnostic approach and definitive treatment of colonic hemorrhage.
The following definitions will guide you while you go through the rest of the chapter.
Lower gastrointestinal (GI) bleeding is defined as acute or chronic abnormal blood loss originating from the colon.
Acute lower intestinal bleeding is arbitrarily defined as bleeding of less than 78 hours duration resulting in instability of vital signs, anemia, and/or a need for blood transfusion.
Chronic lower intestinal bleeding is defined as slow blood loss over a period of several days or longer presenting with symptoms of occult fecal blood, intermittent melena, or scant hematochezia.
Occult gastrointestinal bleeding means that the amounts of blood in the feces are too small to be seen but are detectable by chemical tests.
Obscure gastrointestinal bleeding often presents as lower intestinal bleeding and means bleeding from an unclear site that persists or recurs after a negative initial or primary endoscopy.
39.2 General Aspects
The incidence of lower gastrointestinal bleeding is only one-fifth of that in the upper gastrointestinal tract and is estimated at 21 to 27 cases per 100,000 adults/year. Lower intestinal bleeding is usually chronic and self-limiting. Twenty-one of 100,000 adults/year require hospitalization due to severe bleeding. Among these, men and older patients suffer from more severe lower intestinal bleeding. There is a 200-fold increase in the incidence from the third to the ninth decade, due to diverticulosis and angiodysplasia. 1 In a cross-sectional survey, it was found that 15.5% of the population in the United States suffered from rectal bleeding, but only 13.9% of those affected sought medical care. 2
Some 20% of all gastrointestinal bleeding disorders occur from colonic and anorectal sources. A small bowel source is less common. Studies of lower gastrointestinal bleeding have noted 0.5 to 12% of recurrent bleeding after an initial nondiagnostic colonoscopy result. However, some have estimated that a source in the small bowel is the cause of gastrointestinal bleeding in up to 5% of cases. 3
39.2.2 Clinical Course and Prognosis
There is evidence that upper gastrointestinal bleeding differs in acuteness and severity from lower intestinal bleeding. Patients with lower intestinal bleeding are in shock significantly less often (19 vs. 35%, respectively), require fewer blood transfusions (36 vs. 64%), and have a significantly higher hemoglobin level (84 vs. 61%). 4 Patients with colonic bleeding require fewer blood transfusions in comparison with those who have bleeding from the small intestine. As in upper gastrointestinal bleeding, the majority (80–85%) of cases of bleeding in the lower intestinal tract stop spontaneously. The mortality and morbidity rates increase with age. The overall mortality rate varies between 2 and 3.6%. Patients with bleeding episodes after hospital admission have significantly higher mortality rates (23.1%) in comparison with those who bleed before hospital admission. 5
Several studies have identified clinical features that predict the risk of complications in patients with presumed acute lower GI bleeding. These features can be used to help categorize patients either as low or high risk. 6 , 7 , 8 High-risk features include the following:
Hemodynamic instability (hypotension, tachycardia, orthostasis, syncope)
Significant comorbid illnesses
Bleeding that occurs in a patient who is hospitalized for another reason
A prior history of bleeding from diverticulosis or angiodysplasia
Current anticoagulant or antiplatelet use
Prolonged prothrombin time
A nontender abdomen
An elevated blood urea nitrogen level
An abnormal white blood cell count
Hemodynamic instability (blood pressure < 100 mm Hg, heart rate > 100 beats/min) 1 hour after initial medical evaluation
Active gross bleeding per rectum
Initial hematocrit less than or equal to 35%
39.3 Diagnostic Approach
A focused history helps differentiate the causes of lower intestinal bleeding. Important points include the duration of bleeding, stool color (melena; massive, intermittent, or scant hematochezia; small quantities of blood in the stool), and frequency. Lower intestinal bleeding is usually suspected when hematochezia is present. This means the passage of maroon or bright red blood or blood clots per rectum. This is different from upper gastrointestinal bleeding, which often presents with hematemesis (vomiting blood) and melena. However, massive upper gastrointestinal bleeding can also present with bright red stool; up to 11% of patients with hematochezia may have upper gastrointestinal bleeding. Zuckerman et al reported that melena as a sign of upper gastrointestinal bleeding is correctly described and diagnosed in the acute setting by most physicians, as well as hematochezia for the incidence of a lower gastrointestinal bleeding. 9
Clinical symptoms such as pain, weight loss, changes in bowel habits, or fever are helpful in planning the next diagnostic steps. When the patient’s medical history is being investigated, note should be made of previous bleeding episodes, abdominal and vascular operations, radiotherapy of the pelvic organs, a history of peptic ulcer disease or inflammatory bowel disease (IBD), medication (especially acetylsalicylic acid, nonsteroidal anti-inflammatory drugs [NSAIDs], and anticoagulation treatment), a family history of malignant disease, and comorbidity.
39.3.2 Physical Examination
The physical examination helps differentiate acute from chronic bleeding and includes assessment of circulatory stability. Blood loss of less than 250 mL has no influence either on heart rate or blood pressure. Blood loss of more than 800 mL induces a fall in blood pressure of about 10 mm Hg and a heart rate increase of 10 beats/min. Paleness, weakness, and dizziness are frequent symptoms. Extensive` blood loss of more than 1,500 mL usually presents with shock symptoms, tachypnea, and depressed mental status. Digital rectal examination in combination with a test for occult blood helps confirm the patient’s description of stool color. A digital rectal examination can also detect 40% of rectal carcinomas; in 2% of patients with massive rectal bleeding, the digital rectal examination detected a rectal cancer. 10
39.3.3 Laboratory Studies
The initial laboratory work-up should include the following:
Complete blood count (including hemoglobin, hematocrit, and thrombocytes)
Serum chemistry (electrolytes and creatinine)
Sample for type and cross-match
Flexible endoscopy is considered the mainstay for evaluation of lower intestinal bleeding. The incidence of serious complications is approximately one in 1,000 procedures. Elderly patients and those with cardiovascular or pulmonary diseases are at special risk for cardiopulmonary complications. Aspiration (in upper endoscopy), oversedation, hypoventilation, and vasovagal events are the major problems. Perforation rarely occurs, even in urgent colonoscopy. Patients should be continuously monitored during urgent endoscopy using electrocardiography and noninvasive measurement of oxygen saturation. If there are unstable vital signs, patients must receive resuscitation before endoscopy. Recently guidelines for sedation in endoscopy were published. 11
Esophagogastroduodenoscopy. In patients with hematochezia and hemodynamic instability, esophagogastroduodenoscopy should be undertaken first to exclude an upper gastrointestinal source (▶Fig. 39.1). Particularly in patients with a history of peptic ulcer and portal hypertension, this should be considered in any case.
Colonoscopy. It has been demonstrated that in experienced hands, colonoscopy plays the same role in acute lower intestinal bleeding as esophagogastroduodenoscopy does in acute upper gastrointestinal bleeding. Advantages of colonoscopy compared with other tests for lower GI bleeding include its potential to precisely localize the site of the bleeding regardless of the etiology or rate of bleeding, the ability to collect pathologic specimens, and the potential for therapeutic intervention. As in upper gastrointestinal bleeding, there are three main principles underlying early or urgent colonoscopy:
Determination of the location and type of the bleeding source
Identification of patients with ongoing hemorrhage and those who are at high risk for rebleeding
Assessment of the potential for endoscopic intervention
All patients with acute lower intestinal bleeding must be stabilized. Contraindications for colonoscopy are severe active inflammation and also inadequate visibility conditions. The colonoscopy should be aborted if the patient becomes unstable, the bleeding is so severe that identification of a bleeding source is impossible, or the risk of perforation is too high. The diagnostic yield for urgent (within 12 hours after admission) colonoscopy in acute lower intestinal bleeding is in the range of 48 to 90%. 10 Other publications have reported diagnostic yields of 89 to 97%, 12 which is perhaps a reflection of more consistent use of urgent colonoscopy.
Studies demonstrated that early colonoscopy (examination conducted within 24 hours of admission) is significantly associated with a shorter hospital stay. 5 , 13 Navaneethan et al reported that there was no difference in mortality in patients with lower GI bleeding who had early versus delayed colonoscopy (0.3 vs. 0.4%, p = 0.24), however, patients who underwent early colonoscopy had a shorter length of hospital stay (2.9 vs. 4.6 days, p < 0.001), a decreased need for blood transfusion (44.6 vs. 53.8%, p < 0.001), and lower hospitalization costs (U.S. $22,142 vs. U.S.$28,749, p < 0.001). 5 Some studies suggested that the risk of rebleeding and surgery in patients with severe diverticular hemorrhage can be reduced by early colonoscopy with endoscopic treatment compared with patients treated conservatively. 14 Endoscopic therapy was also more likely to be performed during early examinations. 5 , 15
Urgent colonoscopy is defined as within 12 hours after hospitalization. A randomized trial of patients found that a strategy of urgent colonoscopy improved detection of the source of bleeding compared with expectant/elective colonoscopy alone or with radiographic interventions. But mortality, hospital stay, transfusion requirements, or the need for surgery were not significantly reduced. 16 A concern in urgent colonoscopy is bowel preparation. Chaudhry et al 17 showed that in patients with acute lower intestinal bleeding, a high diagnostic yield (97%) and effective hemostasis was possible even without bowel preparation. They were able to control active bleeding in 17 of 27 patients (63%) by endoscopic intervention. However, other recommendations instead advise cleansing the colon as thoroughly as possible in acute lower gastrointestinal bleeding, which improves evaluation of the mucosa and in turn enhances recognition of smaller lesions and minimizes the risk of complications resulting from poor visualization. However, the role of urgent colonoscopy in lower gastrointestinal bleeding remains controversial. 18 In our institution, bowel cleansing is performed with a polyethylene glycol electrolyte solution. For optimal colon preparation, the patient must consume 4 to 6 L of the solution. 19 In some cases it may be helpful to administer a prokinetic antiemetic drug such as metoclopramide or to administer the solution via a nasogastric tube if the patient has problems with swallowing the fluid.
The segmental location of fresh blood, or the level above which no blood is present, should be carefully documented by the endoscopist. An attempt should be made to reach the cecum whenever possible. This is important as a substantial proportion of bleeding sites are located in the right hemicolon. In addition, the endoscopist should try to intubate the terminal ileum. Blood flowing from above is a clear sign of a more proximal bleeding site. Ohyama et al 12 reported that even in conditions of urgent colonoscopy, it was possible to inspect the cecum in 56% of patients and that advancement as far as the terminal ileum was achieved in 27%.
A case series of colonoscopy with unprepared conditions located the bleeding lesion in 39% of patients and the cecum was reached in 69%. 20 Immediate colonoscopy was performed after tap water enema without oral bowel preparation, aided by waterjet pumps and mechanical suction devices.
In patients with nondiagnostic upper endoscopy and colonoscopy, a repeated endoscopy is helpful in identifying lesions missed at the time of the initial endoscopic evaluation. Studies have reported that lesions within the range of a conventional endoscope were detected by small bowel endoscopy or by a second-look endoscopy in 6 to 20% of cases. 21 , 22 , 23 Repeated capsule endoscopy is acceptable in patients with obscure gastrointestinal bleeding and negative capsule endoscopy findings. Jones et al 24 reported a high yield (75%) with repeated capsule endoscopy in patients with obscure gastrointestinal bleeding when the first examination was negative. In addition, these findings led to a change in management in 62.5% of repeated studies. Limited visualization due to blood and debris during the initial capsule endoscopy appears to be a common reason for repeated studies. Recurrent bleeding is another reason.
39.3.5 Nonendoscopic Methods
Computed Tomographic Angiography
Several reports have described computed tomographic angiography (CTA) for the localization of active hemorrhage. 25 , 26 As CTA is widely available, fast, and minimally invasive, it is an appealing diagnostic modality. In addition, anatomical details that may be helpful for subsequent interventions such as angiography are delivered.
Bleeding at a rate of 0.3 to 0.5 mL/min can be detected with CTA, typically using multidetector-row helical computed tomography (CT). Compared with single-detector–row helical CT, multidetector-row helical CT permits markedly increased resolution and shortens scanning time.
Several studies have examined CTA for the detection of GI bleeding:
A meta-analysis of 22 studies with 672 patients found that CTA had a sensitivity of 85% and a specificity of 92% for detecting active GI bleeding. 27
• In a study of 161 patients, CTA was similar to radionuclide imaging for detecting bleeding on subsequent angiography (sensitivity of 90%, specificity of 20%), but with more precision when it came to localizing the site of the bleeding. 28
Potential drawbacks of CTA include the high-contrast volume that is administered to the patient if angiography is performed after CT.
It is estimated that visceral angiography can only detect active bleeding of at least 0.5 to 1 mL/min. The specificity of this procedure is 100%, but the sensitivity varies with the pattern of bleeding, ranging in one study from 47% with acute to 30% with recurrent bleeding. 29 Data on the clinical utility of angiography in obscure gastrointestinal bleeding are very limited. Advantages of angiography include the lack of a need for bowel preparation, the ability to localize the bleeding source exactly (if identified), and the potential for therapy. Angiography should be reserved for patients who have massive bleeding that precludes colonoscopy or in whom endoscopy has failed to identify the bleeding source. A study group from Hong Kong compared immediate capsule endoscopy (CE) and mesenteric angiography in patients with acute, overt, obscure gastrointestinal bleeding. The diagnostic yield of immediate CE was significantly higher than angiography (53.3 vs. 20.0%, p = 0.016). The cumulative risk of rebleeding in the angiography and CE group was 33.3 and 16.7%, respectively (p = 0.10, log-rank test), and there was no significant difference in the long-term outcomes between the two groups including further transfusion, hospitalization for rebleeding, and mortality. 30
Transcatheter embolization is a more definitive means of controlling hemorrhage and has largely replaced vasopressin infusion. The risk of bowel infarction was decreased with the use of superselective embolization of distal vessels using coaxial catheters. In patients found to have active bleeding, superselective embolization is feasible in 80%, and bleeding is successfully controlled in 97%. 31 However, the embolization procedure is associated with a risk of intestinal infarction of up to 20%, as well as other serious complications including arterial injury, thrombus formation, and renal failure.
A retrospective study analyzed 53 cases of lower intestinal bleeding (20 patients with colonoscopy and 20 patients with visceral angiography and nuclear scintigraphy). Performance of colonoscopy showed a broad diagnostic yield in the urgent setting, with consecutive shorter hospital stays and a less transfusion rate. There was no difference relating to necessity of surgical intervention. 32
Nuclear scintigraphy is a sensitive method of detecting gastrointestinal bleeding at a rate of 0.1 mL/min, but has largely been replaced by CT scan that is more practical and the diagnostic method of choice. The role of nuclear scans, and in particular of technetium-99m (99mTc)-labeled red blood cells, is limited in patients with obscure gastrointestinal bleeding and has substantially declined with the advent of complete endoscopic imaging of the small bowel. A major disadvantage of nuclear imaging is that it localizes bleeding only to an area of the abdomen. For example, bleeding from a redundant sigmoid may appear in the right lower quadrant, suggesting bleeding in the right colon. Another problem is colonic motility, which can move blood in either a peristaltic or antiperistaltic direction. When scans are positive within 2 hours, localization is correct in 95 to 100% of cases, but with positive scans after 2 hours, the accuracy decreases to 57 to 67%. 10 Scintigraphy may be a useful tool for intermittent gastrointestinal bleeding, when endoscopic methods have failed. It is strongly recommended that every positive radionuclide imaging examination should be confirmed by endoscopy or angiography before definitive therapy is considered—such as surgery, for example. Two types of nuclear scans have been used: 99mTc sulfur colloid and 99mTc pertechnetate-labeled autologous red blood cells. Both techniques are noninvasive and sensitive for GI bleeding.
Currently, exploratory laparotomy is seldom performed without intraoperative enteroscopy. Lesions have to be identified by simple palpation and transillumination. In two reports, diagnosis was possible at surgery in 64 and 65% of cases, respectively. 33
39.4 Differential Diagnosis
The colon accounts for one-third of cases of gastrointestinal bleeding. The frequency of colonic bleeding sources reported varies among publications. One reason for this could be that studies often fail to differentiate between probable and definite sources of bleeding. In addition, the definitions of acute lower intestinal bleeding used are far from uniform. A source of lower intestinal bleeding cannot be definitively identified in up to 25% of patients. 34 Age can provide a clue to the cause of acute lower gastrointestinal bleeding as younger patients tend to bleed from hemorrhoids, vascular malformations, and rectal ulcers, while older patients tend to bleed from diverticula, vascular malformations, and neoplasms.
An analysis of 1,159 patients with lower gastrointestinal bleeding identified the following sources of bleeding 32 :
Colonic ischemia (6–18%)
Anorectal diseases (6–16%)
Following polypectomy (0–13%)
▶Table 39.1 provides an overview of the frequencies of bleeding sources in patients presenting with hematochezia.
Diverticula (▶Fig. 39.2) are the reported source of gastrointestinal bleeding in 5 to 42% of patients with lower gastrointestinal bleeding. Although most diverticula are located in the left hemicolon, especially in the sigmoid colon, diverticula in the right hemicolon appear to have a greater bleeding tendency. However, the correlation may not always be causal, as diverticula are often cited as the bleeding source in the colon due to lack of evidence of another source.
Among patients with diverticulosis, the risk of bleeding is approximately 0.5 per 1,000 person-years. 35 In a study of 1,514 asymptomatic patients with diverticulosis, the cumulative incidence of bleeding was 0.2% at 12 months, 2.2% at 60 months, and 9.5% at 120 months. Risk factors for bleeding included age greater than or equal to 70 years (adjusted hazard ratio [aHR] 3.7) and bilateral diverticulosis (aHR 2.4). Interestingly, obesity also appears to increase the risk of diverticulitis and colonic diverticular bleeding. 36