Acute severe lower GI bleeding occurs with an annual hospitalization rate of 22 per 100,000 adult population; this is based on a retrospective study of middle-class Americans who were members of Kaiser Permanente Health Care system in San Diego, California.¹ Assuming that an average full-time clinical gastroenterologist is responsible for 50,000 adult lives, he or she would see more than 10 cases per year. Most cases occur in elderly patients, given the increased frequency and risk for diverticulosis, vascular disease, and colonic malignancy.¹ Risk of lower GI bleeding is also associated with the use of aspirin and nonsteroidal antiinflammatory drugs (NSAIDs).^2,³

Initial Approach to a Patient with Severe Hematochezia

Initial patient assessment includes history, vital signs with orthostatic blood pressure determination, and physical and rectal examinations. Patients should be asked about whether they saw red blood or dark maroon blood, duration of symptoms, abdominal pain, prior history of lower GI bleeding, prior pelvic radiation, history of diverticulosis, and prior colon imaging studies. Patients should also be asked about use of medications associated with GI bleeding (e.g., aspirin, NSAIDs, anticoagulants, and Ginkgo biloba). Weight loss or a change in bowel habits suggests possible colon cancer. Abdominal pain usually suggests ischemic colitis, although abdominal pain can be present in other colitides and malignancy.

The most important parts of the physical examination are the vital signs and the stool examination. The presence of bright red blood on rectal examination strongly suggests the possibility of colonic bleeding. Bright red blood per rectum is always a colonic source, unless it is accompanied by hypotension, which can occur during a severe upper GI or small bowel bleed with rapid transit of blood.⁴ In the setting of hematochezia without hypotension, placement of a diagnostic nasogastric (NG) tube is usually unnecessary because it is unlikely that there is a severe upper GI bleed without hypotension. If there is hypotension and hematochezia, a severe upper GI bleed is possible, and an NG tube should be placed. A clear NG tube lavage does not always imply a lower GI source because 16% of patients with duodenal ulcer bleeds have negative NG lavage.⁵ If bile is seen in the NG tube lavage, it is unlikely to be an upper GI bleed. Physical examination should also focus on abdominal tenderness, surgical scars, and stigmata of liver disease. Most patients with severe hematochezia do not need placement of an NG tube for diagnostic lavage, unless there is a strong suspicion for an upper GI source. At least one large-bore (14-gauge or 16-gauge) intravenous catheter should be placed, with two placed in the setting of ongoing bleeding.

Blood should be sent for hematocrit, platelets, prothrombin time, partial thromboplastin time, chemistry panel, and type and crossmatch for packed red blood cells. Resuscitation should be initiated simultaneously with assessment. Normal saline is infused as fast as needed to keep systolic blood pressure greater than 100 mm Hg and pulse lower than 100 beats/min. Patients are transfused with packed red blood cells, platelets, and fresh frozen plasma as necessary to maintain hematocrit greater than 24%, platelet count greater than 50,000/mm³, and prothrombin time less than 15 seconds. A GI endoscopist should be notified as soon as possible to expedite patient diagnosis and possible therapy; this is especially important in terms of coordinating timing of the bowel purge and procedure.

Most patients with lower GI bleeding can be admitted to internal medicine or gastroenterology services rather than to the general surgery services because these patients rarely require emergency surgical intervention, and given their elderly age, they usually require management of comorbid diseases by an internist. Patients should be admitted to an intensive care unit (ICU) or monitored intermediate care unit. Patients should have automatic blood pressure monitoring every 5 minutes if unstable and hourly if stable. Each patient should receive cardiac rhythm monitoring to observe for arrhythmias and to follow the heart rate as a sign of continued or recurrent bleeding. Laboratory-determined hematocrits (not finger-stick hematocrits, which are less reliable) should be obtained every 4 to 6 hours until the patient has a stable hematocrit. In cases of active bleeding, an indwelling bladder catheter should be placed to help monitor fluid status. Swan-Ganz catheter monitoring is unnecessary except for patients with a history of congestive heart failure or unstable cardiac disease. Patients older than age 60 or with risk factors for coronary artery disease should also have serial electrocardiograms and cardiac enzyme evaluation to determine whether there is any cardiac ischemia.

Endoscopy of any sort should be done only when it can be performed safely and when the information may influence patient care. Patients should be medically resuscitated with fluids and transfusions before endoscopy. Ideally, patients should be hemodynamically stable, with a heart rate of less than 100 beats/min and systolic blood pressure greater than 100 mm Hg. Hematocrit, especially in elderly patients, ideally should be greater than 28%. Severe thrombocytopenia (platelet count <50,000/mm³) should be corrected with transfusions before emergency endoscopy, and prolonged prothrombin time (>15 seconds) should be corrected with fresh frozen plasma. Performance of endoscopy in the middle of the night should be avoided unless well-trained endoscopy nurses, appropriate endoscopy equipment, and surgical backup are available.

Early Predictors of Severity in Acute Lower Gastrointestinal Bleeding

Early predictors (within 4 hours of admission) of severity for continued or recurrent bleeding after 24 hours of hospitalization include heart rate greater than 100 beats/min, systolic blood pressure less than 115 mm Hg, syncope, nontender abdominal examination, observed rectal bleeding during the first 4 hours of hospital evaluation, aspirin ingestion, and the presence of more than two comorbid conditions (Box 14.1).^6,⁷ This prediction model has been prospectively validated; the low-risk group had 0% rebleeding, the moderate-risk group had 45% rebleeding, and the high-risk group had 77% risk of rebleeding.⁷ It is possible that factors such as these can be used to help triage patients to the appropriate level of care, such as ICU, hospital ward, or outpatient evaluation and urgent versus elective endoscopic evaluation.

Box 14.1

Early Predictors of Severity of Continued or Recurrent Lower Gastrointestinal Bleeding

Heart rate >100 beats/min

Systolic blood pressure <115 mm Hg

Syncope

Nontender abdominal examination

Observed rectal bleeding during first 4 hours of hospital evaluation

Aspirin

More than two comorbid conditions

From Strate LL, Orav EJ, Syngal S: Early predictors of severity in acute lower intestinal tract bleeding. Arch Intern Med 163:838–843, 2003; and Strate LL, Saltzman JR, Ookubo R, et al: Validation of a clinical prediction rule for severe acute lower intestinal bleeding. Am J Gastroenterol 100:1821–1827, 2005.

Mortality in Severe Lower Gastrointestinal Bleeding

A large U.S. database study comprising 227,000 patients with discharge diagnoses of lower GI bleed in 2002 reported an overall mortality rate from lower GI bleeding of 3.9%.⁸ Multivariate analysis found that independent predictors of in-hospital mortality were age (>70 years), intestinal ischemia, presence of two or more comorbid illnesses, bleeding while hospitalized for a separate process, coagulopathies, hypovolemia, transfusion of packed red blood cells, and male gender. Colorectal polyps and hemorrhoids were associated with a lower mortality risk. Patients who develop severe lower GI bleeding while hospitalized for other lesions have a much higher mortality rate than patients admitted with lower GI bleeding. In a large Kaiser Permanente San Diego retrospective study, the in-hospital mortality rate for patients with lower GI bleeding who began as outpatients was 2.4% compared with 23% for patients with in-hospital lower GI bleeding (P < .001).¹

Diagnostic Options

Most patients undergo initial evaluation with colonoscopy after bowel preparation, although in selected cases flexible sigmoidoscopy without bowel preparation or with enema preparation may be performed. Other diagnostic tests may be used in selected cases or when colonoscopy is unsuccessful.

Anoscopy

Anoscopy can be useful if actively bleeding hemorrhoids are suspected. However, nearly every patient requires additional visualization of the more proximal colon using colonoscopy or possibly sigmoidoscopy.

Flexible Sigmoidoscopy

Occasionally, flexible sigmoidoscopy may be performed to evaluate the left side of the colon quickly for any bleeding site stigmata rather than waiting for a full colonoscopy bowel preparation, and this results in a diagnosis in approximately 9% of cases.⁹ Flexible sigmoidoscopy may be especially useful in patients with strongly suspected diverticular bleeding or ischemic colitis.

Barium Enema

There is no role for emergency barium enema in a patient with severe lower GI bleeding. This test is rarely diagnostic because it cannot show vascular lesions and may be misleading if only diverticula are present. It also fails to detect 50% of polyps greater than 10 mm in size.¹⁰ Subsequent colonoscopy is needed for any suspicious lesions seen on barium enema, and no therapy can be performed.

Nuclear Medicine Scintigraphy

Nuclear medicine scintigraphy involves injecting a radiolabeled substance in the patient’s bloodstream and then performing serial scintigraphy to detect focal collections of radiolabeled material. It has been reported to detect bleeding at a rate of 0.1 mL/min.¹¹ The overall positive diagnostic rate is approximately 45%, with a 78% accuracy in the localization of the true bleeding site.¹² The most common false-positive result occurs when there is rapid transit of luminal blood such that labeled blood is detected in the colon, although it originated in the upper GI tract.

Angiography

Angiography is positive when the arterial bleeding rate is at least 0.5 mL/min.¹³ The diagnostic yield depends on patient selection, timing of the procedure, and the skill of the angiographer, with positive yields in 12% to 69% of cases. An advantage of angiography is that embolization can be performed to control some bleeding lesions. There is also a 3% rate of major complications, however, including hematoma formation, femoral artery thrombosis, contrast dye reactions, renal failure, and transient ischemic attacks.¹⁴

Computed Tomography Colonography

Computed tomography (CT) visualization of the colon is increasingly used to evaluate the colon for polyps and masses and may be of some benefit in lower GI bleeding. Faster scanners allow CT angiography, CT colonography, and CT evaluation of the small bowel to be performed. Use of CT potentially could allow diagnosis of mass lesions and vascular lesions, which would be an advantage compared with other radiologic imaging studies. One study from France reported that CT accurately diagnosed 17 of 19 lower GI bleeding sites, including diverticula, tumors, angiomas, and varices.¹⁵

Colonoscopy

Urgent colonoscopy using a rapid sulfate purge has been shown to be safe, to provide important diagnostic information, and sometimes to allow therapeutic intervention.¹⁶ Patients usually ingest 4 to 8 L of polyethylene glycol either orally or via NG tube over 3 to 5 hours until the rectal effluent is clear of stool, blood, and clots. Metoclopramide may be given intravenously before the purge and repeated every 3 to 4 hours to facilitate gastric emptying and reduce nausea.

Most “urgent” colonoscopy for lower GI bleeds is performed 6 to 36 hours after the patient is admitted to the hospital. Because most bleeding stops spontaneously, cases are often performed more electively the day after initial hospitalization to allow the patient to receive blood transfusions and to have the bowel preparation during the 1st day of hospitalization.

The overall diagnostic yield of a presumed or definite etiology using colonoscopy in lower GI bleeding ranges from 48% to 90%, with an average of 68%, based on a review of 13 studies.¹² The problem with interpreting these data is that it is often impossible to determine a definite diagnosis of the cause of the bleeding, unless bleeding stigmata are identified such as active bleeding, a visible vessel, an adherent clot, mucosal friability or ulceration, or the presence of fresh blood limited to a specific part of the colon. A presumptive diagnosis often is made, especially in the case of diverticulosis, in which no blood is seen but there is a potential bleeding site present.

The optimal time for performing urgent bowel preparation and colonoscopy is unknown. Theoretically, the sooner the endoscopy is performed, the higher the likelihood of finding a lesion that might be amenable to endoscopic hemostasis, such as a bleeding diverticulum or polyp stalk. However, a retrospective study from the Mayo Clinic suggested that there was no significant association between the time of endoscopy (0 to 12 hours, 12 to 24 hours, >24 hours) and the findings of active bleeding or other stigmata that would prompt colonoscopic hemostasis in patients with diverticular bleeding.¹⁷ Early colonoscopy has been associated with fewer hospitalization days.¹⁸ Active bleeding or lesions at risk for rebleeding can be treated with colonoscopic hemostasis; this mostly applies to postpolypectomy bleeding and diverticula and is discussed later.

Surgery

Surgical management is rarely needed for lower GI bleeding because most bleeding is either self-limited or easily managed with medical or endoscopic therapy. The main indications for surgery are malignant lesions and recurrent bleeding from diverticula. For this reason, it seems prudent that most stable patients be managed by an internist or gastroenterologist rather than a surgeon.

Etiology and Pathogenesis of Severe Lower Gastrointestinal Bleeding

It is not always possible to visualize active bleeding during colonoscopy. The timing of endoscopy may influence visualization: Earlier colonoscopy should increase the chances of detecting an actively bleeding lesion. A definite diagnosis of a bleeding lesion can usually be made if active bleeding is seen or there is an obvious stigma, such as clot or visible vessel. A presumptive diagnosis can be made if there is a suspicious lesion and no other possible sources. Table 14.1 lists the frequency of various presumed or definite sites of acute colonic bleeding.¹⁹ Potential colonic lesions amenable to endoscopic hemostasis include diverticula, postpolypectomy sites, angiomas, hemorrhoids, Dieulafoy’s lesions, tumors, ulcers, and radiation proctitis.

Table 14.1 Etiology of Severe Lower Gastrointestinal Bleeding *

Cause	Cases (%)
Diverticulosis	33
Cancer/polyps	19
Colitis	18
Unknown	16
Angiodysplasia	8
Other	8
Postpolypectomy	6
Anorectal	4

* Summary of 1333 patients in seven published studies.

From Zuckerman GR, Prakash C: Acute lower intestinal bleeding. Part II. Etiology, therapy, and outcomes. Gastrointest Endosc 49:228–238, 1999.

Diverticular Bleeding

Colonic diverticula are herniations of colonic mucosa and submucosa through the muscular layers of the colon. Diverticula in the colon are actually pathologic pseudodiverticula because true diverticula contain all layers of the intestinal wall. Colonic diverticula seem to form when colonic tissue is pushed out by intraluminal pressure. Diverticula occur at the point of entry of the small arteries that supply the colon, the vasa recta, as they penetrate the circular muscle layer of the colonic wall. The entry points of the vasa recta are areas of relative weakness through which the mucosa and submucosa can herniate when under increased intraluminal pressure. They vary in diameter from a few milliliters to several centimeters. The most common location is the left colon.

Most colonic diverticula are asymptomatic and remain uncomplicated. Bleeding may occur from vessels at the neck or base of the diverticulum (Fig 14.1).²⁰ Diverticula are common in Western countries, with a prevalence of 50% in older adults.²¹ In contrast, less than 1% of people living in continental Africa and Asia have diverticula.²² This finding has led to the hypothesis that regional differences in prevalence can be explained by the low amounts of dietary fiber in Western diets. Presumably, a low-fiber diet results in less stool content, longer fecal transit time, increased colonic muscle contraction, and, ultimately, increased intraluminal pressure that results in the formation of propulsion diverticula. In addition, diverticula occur with increasing frequency with advanced age, which could be a result of weakening of the colonic wall and muscle tone.

Fig. 14.1 Vascular anatomy of colonic diverticulum.

It has been estimated that 3% to 5% of patients with diverticulosis develop diverticular bleeding.²³ Although most diverticula are in the left colon, several series suggest that bleeding diverticula occur more often in the right colon.^20,²³^–²⁵ Patients with diverticular bleeding are typically elderly and present with painless hematochezia. They often have been taking aspirin or NSAIDs.² In at least 75% of patients with diverticular bleeding, the bleeding stops spontaneously.²⁴ Patients in whom bleeding stops usually require less than 4 U of blood. In one surgical series, surgical resection was needed in 60% of patients, most of whom had continued bleeding despite transfusion of 4 U of blood.²⁴ Patients with successful resection of a bleeding diverticulum had a rebleed rate of 4%.²⁴ Among patients in whom bleeding stopped spontaneously, the rebleeding rate from colonic diverticulosis was 25% to 38% over the next 4 years.^1,²⁴ Urgent colonoscopy after rapid bowel preparation often reveals that bleeding has stopped by the time of colonoscopy, and only nonbleeding diverticula are detected. These patients are given the diagnosis of “presumptive diverticular bleed” because the diverticula are the only likely source of bleeding, although no stigmata were identified.

Occasionally, urgent colonoscopy reveals stigmata of recent bleeding, such as active bleeding, a visible vessel, clot, or blood limited to one segment of the colon (Fig. 14.2). It seems possible that earlier colonoscopy in lower GI bleeding would result in a greater frequency of finding stigmata of recent diverticular bleeding, although a small case series study from the Mayo Clinic did not find any difference if colonoscopy was performed between 0 and 12 hours, between 12 and 24 hours, or more than 24 hours from the time of hospital admission.¹⁷ There have been attempts to stratify patients with diverticular bleeding at increased risk for rebleeding employing the same endoscopic stigmata used in high-risk peptic ulcer bleeding (active bleeding, visible vessel, and clot), although the natural history for each of these untreated stigmata is unknown. The “pigmented protuberance” found on the edge of some diverticula at histopathology is usually clot at the edge of a ruptured blood vessel.²⁶

Fig. 14.2 Active diverticular bleeding.

The UCLA/CURE group ¹⁷ found that among patients with stigmata of recent diverticular hemorrhage (six active bleeding, four visible vessels, and seven adherent clots), there was a very high rebleed rate of 53% and emergency surgery rate of 35%.²⁷ Colonoscopic hemostasis of actively bleeding diverticula has been reported using bipolar probe coagulation, epinephrine injection, metallic clips, rubber band ligation, and fibrin glue.²⁶^–³³ If fresh red blood is seen in a focal segment of colon, we try to examine this segment of bowel carefully to detect the exact bleeding site. If the bleeding is coming from the edge of the diverticulum or there is a pigmented protuberance on the edge, we initially inject 1 : 10,000 epinephrine in 1-mL aliquots using a sclerotherapy needle into four quadrants around the bleeding site. Then we use either an endoscopic clip or a bipolar probe at a low power setting (10 to 15 W) and light pressure for a 1-second pulse duration to cauterize the diverticular edge and stop bleeding or flatten the visible vessel. If there is a nonbleeding adherent clot, we inject around the clot with 1 : 10,000 epinephrine in four quadrants with 1 mL per quadrant and remove the clot in piecemeal fashion using a cold polyp snare. The clot is shaved down until it is 3 mm above the diverticulum, and then the underlying stigma is treated with either endoscopic clip or bipolar probe coagulation as discussed previously. After performing endoscopic hemostasis of a bleeding diverticulum, a permanent submucosal tattoo and a metal clip (if not previously placed) should be placed in the adjacent mucosa to identify the site in case colonoscopy, angiographic embolization, or surgery is required for recurrent bleeding. For long-term management after colonoscopic hemostasis, patients are told to avoid aspirin (if approved by their cardiologist) and NSAIDs and to take a daily fiber supplement.

In 2000, Jensen and colleagues in the UCLA/CURE group ²⁷ published their results on urgent colonoscopy for diagnosis and treatment of severe diverticular hemorrhage. The investigators found that 20% of patients with severe hematochezia had endoscopic stigmata suggesting a definite diverticular bleed. Compared with a historical control group with high-risk stigmata but no colonoscopic hemostasis, the group receiving colonoscopic hemostasis had a rebleed rate of 0% versus 53% and an emergency hemicolectomy rate of 0% versus 35%. After 3 years of follow-up, there were no rebleeding episodes in the patients who underwent colonoscopic hemostasis. In contrast to the UCLA experience, a smaller, retrospective review of the Duke University Medical Center Endoscopic Database revealed 13 patients with active bleeding or stigmata who received endoscopic treatment with epinephrine or bipolar coagulation or both.³⁴ The 30-day rebleed rate was 38%; four of these patients underwent surgery. The long-term rebleed rate was 23% with a mean follow-up of 3 years.

Angiographic embolization can also be performed in selected cases of diverticular bleeding, but there is a risk of bowel infarction, contrast reactions, and renal failure. Angiography can be helpful before surgical resection. Surgical resection for diverticular bleeding is usually reserved for recurrent bleeding episodes. Resection should be guided by colonoscopic, angiographic, or nuclear medicine studies showing the likely bleeding site. The need for surgery is often guided by certainty regarding the bleeding site and medical comorbidity because diverticular bleeding is often mild, and the risks of surgical complications are increased in elderly patients.

Colon Cancer

Most patients with colon cancer present with occult GI blood loss rather than hematochezia. For adult patients with hematochezia, determining the presence or absence of a colon cancer is imperative because early diagnosis improves survival. Because a cancer must ulcerate for overt bleeding to occur, most bleeding cancers are at a relatively advanced tumor stage (Fig. 14.3). Colon cancer can occur anywhere in the colon or rectum, but there is increased prevalence of right-sided tumors in elderly patients.

Fig. 14.3 Colon cancer as a cause of severe hematochezia.

Colitis

The term colitis refers to any form of inflammation of the colon. With regard to severe lower GI bleeding, this is usually ischemic colitis, inflammatory bowel disease, or possibly infectious colitis. Ischemic colitis generally manifests as hematochezia with mild left-sided abdominal discomfort. It results from mucosal hypoxia and is thought to be caused by hypoperfusion of the intramural vessels of the intestinal wall rather than by large vessel occlusion. Most cases do not have a recognizable cause, but associated conditions include recent aortic or cardiac surgery, vasculitis, and medications.^35,³⁶ Because of collateral circulation, ischemic involvement is usually segmental and primarily affects the mucosal aspect of the intestine. The colon is mostly affected in the watershed areas, such as the splenic flexure or rectosigmoid junction in which there is reduced collateral circulation, although ischemia can occur anywhere.

The diagnosis is usually confirmed by colonoscopy but can be suspected by “thumbprinting” on plain film radiographs or colonic wall thickening on CT scan. The colonoscopic appearance includes erythema, friability, and exudate (Fig. 14.4). Biopsy specimens may be suggestive of ischemic changes but more importantly are used to exclude infectious changes or Crohn’s disease. Ischemic colitis generally resolves in a few days and does not require colonoscopic hemostasis. In a large retrospective series from Kaiser Permanente San Diego, there were no episodes of rebleeding from ischemic colitis over a 4-year period.¹ Inflammatory bowel disease affecting the colon can rarely cause severe acute lower GI bleeding. In a case series from the Mayo Clinic, most patients had Crohn’s disease.³⁷ Most patients were successfully treated medically. Three of the 31 patients in the series received endoscopic therapy with epinephrine injection alone or with bipolar coagulation for adherent clots or oozing ulcers in Crohn’s disease. These patients had no rebleeding. Rebleeding occurred in 23% of patients a median of 3 days after the initial bleed (range 1 to 75 days). Of the patients with Crohn’s disease with severe bleeding, 39% required surgical management.

Fig. 14.4 Ischemic colitis.

Infectious colitis should always be excluded in any patient with severe lower GI bleeding and colitis. Lower GI bleeding can occur with infection by Campylobacter jejuni, Salmonella, Shigella, invasive Escherichia coli, E. coli 0157, or Clostridium difficile (Fig. 14.5). Significant blood loss is rare. Diagnosis is made by stool cultures and flexible sigmoidoscopy.

Fig. 14.5 Clostridium difficile colitis.

Angiodysplasia

Colonic angiomas are also referred to as angiodysplasia, arteriovenous malformations, or vascular ectasias. They are generally uncommon: Less than 1% of asymptomatic patients undergoing screening colonoscopy were found to have angiodysplasia.³⁸ The lesion seems to increase with age and may represent degeneration of previously normal blood vessels in the cecum and proximal ascending colon. Histopathology reveals a large, dilated, submucosal vein and, in advanced cases, dilated mucosal veins with small arteriovenous communications. Proposed explanations for angioma formation include the partial obstruction of submucosal veins passing through the colonic muscle layers, with eventual dilation of the submucosal and mucosal veins, and local mucosal ischemia.

Medical conditions associated with angiomas include chronic renal failure and hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu syndrome). There have been reports suggesting that aortic stenosis is associated with lower GI bleeding, presumably from colonic angiomas.³⁹ The potential biologic explanation is that aortic stenosis causes defects in von Willebrand factor, which causes the patient to have decreased platelet adhesion and increased bleeding tendency, especially if there were preexisting mucosal GI lesions such as angiomas.^40,⁴¹ Clinical studies do not support the association between aortic stenosis and the presence of angiomas, however.^42,⁴³ Bleeding from angiodysplasia is usually painless. The bleeding usually occurs from the right colon or cecum.

In the past 2 decades, it seems that the reported frequency of angiomas as the source of lower GI bleeding has decreased.^16,²⁷ This may be because of better recognition of angiomas with improved endoscope technology and increased attribution to presumed diverticular bleeding as the cause of hematochezia. Endoscopic hemostasis has been successfully reported using thermal modalities (e.g., bipolar probe, heater probe, laser) and injection therapy.⁴⁴ Hormonal therapy has been reported as useful for decreasing bleeding from angiodysplasia, but a more recent randomized controlled trial found no benefit.⁴⁵

Postpolypectomy Bleeding

Postpolypectomy bleeding occurs after 1% to 6% of polypectomies, usually within the first 7 days.⁴⁶ It is generally mild and self-limited. Reported risk factors for postpolypectomy bleeding include large polyps (>2 cm), thick stalks, sessile polyps, and right colon polyps.⁴⁶ Endoscopic management techniques include resnaring the stalk (without cautery), epinephrine injection, thermal coagulation, hemoclips, and endoloops (Fig. 14.6).⁴⁶^–⁴⁹ In a case series from the Mayo Clinic, the median time to bleeding after polypectomy was 5 days (range 0 to 17 days).⁴⁸ Of patients, 65% received aspirin, NSAIDs, warfarin (Coumadin), heparin, or steroids after polypectomy; 76% required transfusions; and 96% were managed endoscopically with coagulation or epinephrine injection or both. The routine use of placing hemoclips after colonic polypectomy or endoscopic mucosal resection does not decrease the subsequent postpolypectomy bleeding rate.⁵⁰ However, in selected patients who have had polypectomies and who were believed to be at increased risk for bleeding, prophylactic hemoclip placement or other endoscopic hemostasis may be considered.

Fig. 14.6 A, Postpolypectomy bleeding. B, Polypectomy site after placement of clips.

Radiation Proctitis

Radiation proctitis usually causes mild chronic hematochezia but occasionally can cause acute severe lower GI bleeding. Ionizing radiation can cause acute and chronic damage to the normal colon and rectum after radiation treatment for gynecologic, prostatic, bladder, or rectal tumors. Approximately 75% of patients who receive 4000 rad develop acute, self-limited diarrhea, tenesmus, abdominal cramping, and rarely bleeding during the first few weeks. Chronic radiation effects occur 6 to 18 months after completion of treatment. Bowel injury resulting from chronic radiation is related to vascular damage, with subsequent mucosal ischemia, thickening, and ulceration. Much of this damage is believed to be due to chronic hypoxic ischemia and oxidative stress. Flexible sigmoidoscopy reveals telangiectasia, friability, and ulceration in the rectum.

Patients should be instructed to avoid all aspirin and NSAIDs and should be put on a high-fiber diet. Medical therapies with topical or oral 5-aminosalicylic acid, sucralfate, or steroids can be tried but are usually unsuccessful.⁵¹ Thermal therapy can be quite successful, including argon plasma coagulation, bipolar probe coagulation, radiofrequency ablation, and cryotherapy (Fig. 14.7).⁵²^–⁵⁴ Topical formalin applied directly to the rectal mucosa can reduce bleeding.⁵⁵ In refractory cases, hyperbaric oxygen can also be used successfully.^56,⁵⁷ A few pilot studies suggest that antioxidant vitamins, such as vitamins A, C, and E, can also decrease bleeding resulting from chronic radiation proctitis.⁵⁸