Lower Gastrointestinal Hemorrhage

Transcription

1 20 Lower Gastrointestinal Hemorrhage Frank G. Opelka, J. Byron Gathright, Jr., and David E. Beck Lower gastrointestinal hemorrhage refers to a spectrum of intestinal bleeding that arises distal to the ligament of Treitz. It may range from occult bleeding or occasional spotting of blood to massive lower intestinal hemorrhage. Massive lower intestinal hemorrhage is difficult to define. Patients often describe massive bleeding into their commode even when a small amount of blood discolors the water. True massive intestinal hemorrhage typically involves hemodynamic compromise or symptomatic anemia. Multiple sources define massive bleeding to include patients with a hematocrit less than 30%, patients with transfusion requirements (up to 3 5 units of blood/blood products), or orthostasis requiring resuscitation. Primarily, lower intestinal hemorrhage arises from within the colon. Billingham 1 described lower gastrointestinal hemorrhage as a conundrum with five key concerns. First, the condition may arise from bleeding throughout the gastrointestinal tract. Second, intermittent bleeding precludes a prompt identification of the site of hemorrhage. Third, patients requiring surgery may undergo a procedure without a specific preplanned site of resection and with considerable morbidity and mortality. Fourth, despite aggressive surgical management, persistent bleeding may occur. And finally, there is no consensus about the precise diagnostic and therapeutic pathways for patients. From the perspective of the emergency room care, massive lower gastrointestinal hemorrhage is a relatively uncommon emergency. Longstreth 2 studied a large health maintenance organization (HMO)-based population in San Diego, California, noting an annual incidence rate of 20.5/100,000 patients with a male predominance. His study reflected a retrospective survey and chart review defining incidence. The incidence of significant bleeding increases with age. The association with aging may suggest senescent changes associated with the small intestine and colon. Certainly aging reflects the surging prevalence of colonic diverticulosis and intestinal angiodysplasia in the elderly. It is of interest to note that the California HMO group had a high incidence of diverticulosis (41.6%) and infrequent angiodysplasia (2.7%). However, Longstreth admits the limitations of the study design may not precisely determine the true etiologies of lower gastrointestinal hemorrhage. Lower gastrointestinal bleeding presents with varying degrees of hemorrhage. Patients may experience minor bleeding when they describe the passage of ml of blood, possibly a few clots, and often mixed with mucous. Other patients experience brisk, copious bleeding with major, selflimited hemorrhage. Finally, certain patients present with massive and continuous hemorrhage associated with hypovolemia. The hemorrhage may present as melena or hematochezia. Melena typically suggests bleeding from a more proximal source in the colon or small intestine. Hematochezia suggests left colonic, rectal, or anal sources. It is wise to note that upper gastrointestinal hemorrhage may present with the rectal bleeding given blood s cathartic effect and rapid intestinal transit. Jensen and Machicado 3 outlined the safety of emergency panendoscopy in patients with hematochezia and negative nasogastric lavage. His 1988 study provided additional insights in demonstrating an upper source for bleeding in 11% of patients. 4 Overall, it is believed that upper sources may present with lower gastrointestinal bleeding symptoms in 10% 15% of cases. Most often the intestinal bleeding resolves spontaneously with supportive hospital care. Once it resolves, investigations should begin to identify the potential sources. Actual bleeding sources are not so frequently identified by the current limitations of our diagnostic tools. In clinical scenarios in which the bleeding resolved spontaneously, the diagnostic evaluation may only unmask potential sources. Without associated attached clot or active bleeding, the true site of hemorrhage may never be elucidated. On occasion, the intestinal hemorrhage does not resolve. It continues, creating hemodynamic compromise. Ongoing hemorrhage demands aggressive medical and surgical management. Oftentimes, patients with massive hemorrhage are plagued with significant comorbidities that complicate their individual resuscitation. Their comorbidities must be considered in the diagnostic and therapeutic phases of the care plan. 299

2 300 F.G. Opelka, J.B. Gathright, Jr., and D.E. Beck History and physical examination fall short of an adequate classification system to ultimately predict patient needs or clinical outcome. A patient may portray a worrisome history of massive hemorrhage and still resolve spontaneously with simple, supportive measures. Other patients may sequester blood in large volume and seem to have stopped bleeding. While under observation their scenario promptly changes with ongoing, massive hemorrhage. They require prompt therapy. Still other patients may bleed aggressively, stop for a few days, and then repeat their massive exsanguinations. In addition, diagnostic studies often are invasive procedures with limited sensitivities and specificities. More and more, physicians witness special patient groups with massive lower gastrointestinal hemorrhage. Current disease managements call for concurrent care with anticoagulants or antiplatelet agents for underlying cardiovascular conditions. Current treatment regimens incorporate long-term anticoagulants and antiplatelet agents. Hemorrhage in these patients proves more life-threatening. Landefeld and Goldman 5 noted a 22% long-term risk of bleeding on anticoagulant therapy with warfarin. Gastrointestinal hemorrhage is one of five independent risk factors. Current increased patient exposure to antiplatelet therapy associated with treatment of cardiovascular conditions may increase the comorbid challenges in patients with lower gastrointestinal massive hemorrhage. Etiologies Common causes for lower gastrointestinal hemorrhage include colonic diverticula, angiodysplasia, ischemic colitis, and inflammatory bowel disease. Hemorrhage also stems from intestinal tumors or malignancies. Unusual causes include nonsteroidal antiinflammatory drug (NSAID)-related nonspecific colitis, Meckel s diverticulum, and anorectal diseases. The reported mortality with varying etiologies is summarized in Table Diverticular Disease Diverticulosis is a common malady in Western civilization. Approximately 50% of the population by age 60 years has evidence of diverituclosis. 10 Most diverticula represent pulsion diverticula or pseudodiverticula that are actually outpouchings of the mucosa and submucosa through defects in the muscular layer of the bowel at sites of penetration of the vasa recta. It is theorized that slow intestinal transit and increased intraluminal pressure within the segmentation process promote the development of the diverticula. The precise mechanism of diverticular hemorrhage is unknown. In the late 1800s, Kebs outlined the vascular anatomy of the vasa recta and the mucosal blood supply. Further, Drummond, 11 in 1916, displayed the relationship between the vasa recta and the neck of the diverticulum. In 1976, Meyers et al. 12 defined the bleeding sites as the ruptured vasa recta in the diverticulum. He noted structural changes located eccentrically in the vasa recta at the site of rupture, intimal thickening with thinning of the media, the absence of any acute or chronic inflammation, and stated that these vascular changes were typically the result of focal injury. It is generally accepted that thinning of the media in the vasa recta predisposes to intraluminal rupture: focal injury may occur from trauma related to a fecalith. It is unclear how frequently diverticula are the true cause of hemorrhage. The incidence spans a range of 15% to 48%. Oftentimes, authors attribute the condition to diverticula after the hemorrhage has ceased despite a lack of proof of actual cause, a presumptive diagnosis. Diagnostic evaluations, such as colonoscopy, do not identify a precise source for the hemorrhage without the presence of witnessed bleeding or an adherent clot. Oftentimes, for lack of a more precise etiology, diverticula are present and, therefore, become the primary culprit. Despite being considered a major source for colonic hemorrhage, bleeding from diverticula is a relatively rare event affecting only 4% 17% of patients with diverticulosis. 13 In most cases, bleeding ceases spontaneously, but in 10% 20% of cases, the bleeding continues unabated in the absence of intervention. 14 Once bleeding has occurred, the natural history and risk of rebleeding are poorly understood. Finne 15 comments that the risk of rebleeding after an episode of bleeding is approximately 25% but increases to 50% among patients who have had two or more prior episodes of diverticular bleeding. Right sided colonic diverticula occur less frequently than left sided or sigmoid diverticula but are thought to be responsible for a disproportionate incidence of diverticular bleeding. This finding is not well established, however, and there is often difficulty distinguishing between bleeding from an arteriovenous malformation or angiodysplasias and bleeding from diverticulosis. The overall high prevalence of diverticulosis in the population at risk for lower gastrointestinal hemorrhage makes the exact diagnosis of many bleeding episodes equivocal. TABLE Mortality of lower gastrointestinal bleeding by etiology Investigator Diverticulosis (%) Angiodysplasia (%) Cancer/polyp (%) Colitis/ulcer (%) Anorectal (%) Other (%) Mortality (%) Jensen and Machicado, NA Longstreth, Bramley et al., Richter et al., Rossini et al., NA Jensen and Machicado, NA

3 20. Lower Gastrointestinal Hemorrhage 301 Operative management of diverticular bleeding is indicated when bleeding continues unabated and is not amenable to angiographic or endoscopic therapy. It also should be considered in patients with recurrent bleeding localized to the same colonic segment. In a stable healthy patient, the operation consists of a segmental bowel resection (usually a right colectomy or sigmoid colectomy) followed by a primary anastomosis. One additional note about diverticular hemorrhage focuses on recurrence for patients who stopped bleeding and required no operative intervention. In Longstreth s San Diego study, 2 the author noted that 9% of patients returned within 1 year with another episode. At 2 years, there was little change, 10%; 19% at 3 years; and 25% at 4 years. Angiodysplasia Angiodysplasia was described by Margolis et al. 16 in 1960 when they noted the radiographic features during an intraoperative angiogram performed for colonic bleeding. Angiodysplasias are thin-walled arteriovenous communications located within the submucosa and mucosa of the intestine. Angiodysplasias may be congenital or, more typically, acquired. They could be isolated or multiple. In the acquired form, distortions of the postcapillary venules may arise as a degenerative lesion associated with increases in intraluminal pressure. The intraluminal pressure occurs from loss of the precapillary sphincter and a resultant increased pressure transmitted through the capillary bed into the venules. As these vessels respond to the arterial flow, it results in thickening and ectasia. The vessels eventually entangle as tufts within the submucosa and erode into the mucosa proper. No one is quite certain precisely why angiodysplasias occur. Current hypotheses suggest a loss of vascular integrity related to loss of transforming growth factor (TGF) β signaling cascade or from a deficiency in mucosal type IV collagen. McAllister et al. 17 suggest that a genetic error in endoglin production alters TGF β and, thus, the integrity of the vascular endothelial cells. Roskell et al. 18 noted the loss of mucosal type IV collagen in pathologic specimens of angiodysplasia. Angiodysplasias are uncommon before age 60, increase with age, and are associated with aortic stenosis (Heyde s syndrome), chronic renal failure, and von Willebrand s disease. Osler-Weber-Rendu (hereditary hemorrhagic telangiectasias) is a hereditary condition with telangiectasias of the lung, nervous system, skin, and intestine. These patients present with multiple lesions. In 1995, Christopher Gostout 19 editorialized in questioning the association of angiodysplasias with aortic stenosis. When angiodysplasias are noted during angiography or colonoscopy, unless a hemorrhagic blush is seen during the angiogram or colonoscopy, it is difficult to accurately accuse this malady as the source of hemorrhage. 20 In the past, angiodysplasia was the diagnosis chosen when no bleeding source could be identified and the abnormal vessels were present. In addition, many drew an association between the angiodysplasias and aortic stenosis. The association between the ecstatic vascular tufts and aortic stenosis was dispelled by Imperiale and Ransohoff 21 in the late 1980s. However, the association persisted in anecdotes until Bhutani and colleagues 20 reviewed 37 patients and found no greater incidence of aortic valvular disease than the control group. Angiography remains the gold standard for the diagnosis of angiodysplasia. After injection of contrast, a series of images are collected in three phases. In the arterial phase, the radiographic findings of angiodysplasia demonstrate early venous filling which normally occurs in later phases. During the next phase, capillary phase, small, tortuous tufts are seen entangled and filled with contrast. Finally, the late phase study demonstrates a persistent of this arteriovenous tuft and a persistent of a slow, emptying vein. 22 When angiography identifies a bleeding angiodysplasia, treatment with embolization therapy or directed infusion of vasopressin will decrease or stop the bleeding. Colonoscopy has increased as a screening agent for colorectal cancers as well as during the investigation for colorectal bleeding. Expectantly, more angiodysplasias are seen during endoscopy than in the past. In contrast to the angiographic findings described by Boley et al., Bhutani et al. 20 highlighted the colonoscopic criteria in describing these lesions. The mucosal surface contains a cherry red lesion that is typically flat. The lesions are greater than 2 mm in size and have a fern-like appearance. A central feeding vessel is not always visible. It is important to identify these lesions during scope insertion. Occasionally, the inexperienced endoscopists may attribute colonoscopic suction trauma to an angiodysplastic area. By searching for the vascular muscular lesions during scope insertion, the endoscopist will avoid misnaming scope mucosal trauma as angiodysplasia. Initial experience in identifying these lesions related to a few angiographic studies. The early evidence suggested the lesions were predominantly right sides. Since colonoscopy has become more available, both left and right sided lesions are thought to occur. Other Causes of Lower Gastrointestinal Hemorrhage Multiple other etiologies cause lower gastrointestinal bleeding and most are not associated with a massive hemorrhage or acute symptomatic anemia. Colonic ischemia, inflammatory bowel disease, and colonic malignancies occur frequently. Each presents in a different manner. Typically, ischemic colitis presents with the abrupt onset of abdominal pain, followed by colic and a mucoid, bloody diarrhea. Inflammatory bowel disease, Crohn s, and ulcerative colitis present with a change in stool patterns. Patients develop diarrhea followed by hematochezia or melena. Localized transmural involvement or colic could add pain-related symptoms. Colorectal carcinomas are associated with exophytic, ulcerative lesions that may bleed insidiously. Only rarely does the malignant process proceed to acute, symptomatic hemorrhage.

4 302 F.G. Opelka, J.B. Gathright, Jr., and D.E. Beck More unusual causes of hemorrhage involve small intestinal tumors, known also as gastrointestinal stromal tumors (GIST). These lesions enlarge and surpass their blood supply. In that event, the ischemia in the tumor will ulcerate and may cause a localized hemorrhage. Meckel s diverticulum represent another atypical cause of bleeding. These lesions occur in the distal ileum. Ectopic gastric mucosa leads to localized acidic contents and resultant ulcerations of the contralateral intestinal wall. Finally, NSAID-associated intestinal hemorrhage occurs most frequently in the terminal ileum and cecum. Long-acting NSAIDS cause a localized mucosal injury. These remnants from the agents have been noted at the site of perforating ulcers. It seems that the terminal ileum and cecum may serve as a reservoir and harbor these agents long enough to establish the mucosal defects. Diaphragm-like strictures are pathognomonic for NSAID injuries and may result from a healing ridge related to repeated injuries from the agents. Occult Hemorrhage Obscure or occult gastrointestinal bleeding is a condition that frustrates the patient and the physicians. The hemorrhage is often massive and intermittent. The traditional tests of nuclear scintigraphy, colonoscopy, and angiography provide no solution. It occurs infrequently. One study noted occult bleeding in no more than 5% of all patients admitted with lower gastrointestinal massive hemorrhage. Frequent recurrences create chronic anemic states in patients and require occasional admissions for transfusions. These patients may harbor angiodysplasias in the small intestine or right colon. Patients in this situation may benefit from small bowel contrast radiography or capsule endoscopy Additionally, elective angiography with cecal magnification may reveal small angiodysplasias. If the occult hemorrhage recurs and investigations fail to reveal the source, a variety of provocative diagnostic angiographic studies have been described. Most studies prefer to incite bleeding using either heparin or thrombolytics. Once the site of bleeding is identified, it may be difficult to control without an operation. In these instances, the surgeon should prepare and hold an operating room. Once the location is identified, a superselective catheter is left in the distal artery. During the conduct of surgery, the surgeon can palpate the catheter within the vessel and direct the surgical resection. Initial Assessment, Resuscitation, and Stabilization Massive lower gastrointestinal hemorrhage requires prompt clinical attention. Patients who present with symptoms secondary to the bleeding have urgent resuscitation needs. These symptoms further define the significance of the hemorrhage. Patients may demonstrate pallor, fatigue, angina, tachypnea, cardiac palpitations, postural hypotension, and syncope. Prompt attention requires placement of vascular access with large bore intravenous fluids. Further hemodynamic monitoring requires cardiac rhythm monitoring and placement of a urinary catheter. A nasogastric tube placed will screen for the presence of upper gastric sources for bleeding. Kovacs and Jensen 27 noted 17.9% of lower gastrointestinal hemorrhage presentations involved an upper gastrointestinal source. The nasogastric tube is effective in detecting prepyloric hemorrhage. The nasogastric decompression need not be continued after an appropriate period of observation to exclude upper intestinal sources. The treatment goals for resuscitation are to restore volume and, replete red blood cell deficiencies and their impact on oxygen delivery. In addition, all coagulopathies require reversal. Patients require laboratory profiles that include a complete blood count, serum electrolytes, a coagulation profile, and a type and crossmatch for packed red blood cells. The initial specific diagnostic evaluation begins with a digital anorectal examination and anoscopy. A rigid proctosigmoidoscopy will allow the examiner to evacuate the rectum of blood and clots. A complete mucosal assessment serves to exclude internal hemorrhoids, anorectal solitary ulcers, neoplasms, and colitis. If nothing is found and subsequent surgery becomes necessary, the evaluation of the rectum and anorectal function greatly aids in surgical decisions. A normal anorectal examination allows the surgeon to consider a primary rectal anastomosis as a treatment possibility. In the event that the physician discovers a source for bleeding during the examination, oftentimes therapy can immediately control the hemorrhage. Once the resuscitation demonstrates a stable patient, the next phase of the diagnostic evaluation ensues. What is the first test to evaluate the cause of bleeding? Currently, three tests are considered for the initial evaluation. These tests include colonoscopy, nuclear scintigraphy, and angiography. Colonoscopy and angiography offer therapeutic intervention whereas nuclear scanning is purely diagnostic. Decisions as to which test to use depend on the clinical judgment, local expertise, severity of the event, and the current activity of the hemorrhage. It may be helpful to subdivide patients into three general clinical categories based on the history, physical, and the initial laboratory data. Is the hemorrhagic event 1) minor and self-limited, 2) major and self-limited, or 3) major and ongoing? Major ongoing hemorrhage requires prompt intervention with angiography or surgery. Minor, self-limited may undergo a colonic lavage and colonoscopy within 24 hours. Major, self-limited may be more difficult to define. Within the spectrum of these three clinical groups, the major, self-limited hemorrhage patients create the current controversy. These patients need a diagnostic test to determine if they require prompt therapy or observation. Should these patients undergo nuclear imaging or colonoscopy? Radionuclide imaging (Figure 20-1) detects the slowest bleeding rates. It is able to detect rates of ml/min.

5 20. Lower Gastrointestinal Hemorrhage 303 and allow for reimaging within 24 hours. Nuclear scintigraphy has variable results, suggesting that scan timing, technical skills, and experience may increase accuracy. Current reports suggest accuracies ranging from 24% to 91%. 29 Ng et al. 30 recommend nuclear imaging for the patients with a major, self-limited hemorrhage. Their data suggest that the timing of the blush predicts the success of angiography. In other words, if the nuclear scan demonstrates an immediately positive blush (within the first 2 minutes of scanning), it is highly predictive of a positive angiogram (60%). The data of Ng et al. seemed predictive for surgery in 24% of patients if the first blush was positive. Just as important, if the initial images in the Ng et al. study did not demonstrate a blush, the study is highly predictive of a negative angiogram (93%) and the need for surgery decreased to 7%. Thus, if the nuclear scan is negative, it provides objective evidence that the patient is not actively bleeding and may be evaluated by colonoscopy. FIGURE Selected images from a 99mTc-labeled RBC gastrointestinal bleeding study in a patient with known diureticulosis. Images acquired at 1 minute (A) and 14 minutes (B). Abnormal increased isotopic activity developed in the proximal transverse colon, which progressed antegrade to the descending colon. Thus, it is a technique that is more sensitive than angiography. Unfortunately, the nuclear scanning cannot reliably localize the site of hemorrhage. The specificity (precise origin) using radionuclide scans of small bowel versus large intestine bleeding does not reliably compare with angiography. 28 Two general techniques are used for nuclear imaging, technetium sulfur colloid scans and 99mTc pertechnetate-tagged red blood cells (RBCs). Sulfur colloid scans have a short half-life and detect very low rates of hemorrhage (0.1 ml/min). It is effective to detect brisk hemorrhage but cannot detect sporadic bleeding. The more frequently preferred agent for lower gastrointestinal hemorrhage radionuclide scanning is the pertechnetate-tagged RBC scans. The tagged RBC scans may cover a period of hours Colonoscopy Many authors believe that colonoscopy has clearly demonstrated the highest efficacy and should be the first study in patients with major bleeding that appears self-limited. 31 In general, this may be true if efficacy of the study includes a broad array of the common etiologies for properly defined massive hemorrhage. Controversy abounds with colonoscopy as the preferred first study if the etiologies for hemorrhage are unlikely sources for major hemorrhage. Whether colonoscopy should be undertaken emergently depends on the general ability to maintain a stable patient. If the hemodynamic profile continues to drift toward hypotension and the massive hemorrhage continues unabated during the resuscitation process, the rate of hemorrhage may require more prompt attention. Patients with extremely brisk hemorrhage require a prompt angiogram. Colonoscopy in such patients proves difficult to prep with lavage and the acute exsanguinations may limit intraluminal visualization to deploy all the therapeutic options except for only the most experienced endoscopists. If the patient appears stable with self-limited hemorrhage, colonoscopy is the preferred diagnostic study. Jensen et al. 3,4,9 have long been proponents of emergency colonoscopy. This group and others have demonstrated high cecal intubation rates (95%) and a diagnostic accuracy of 72% and 86%. On a cautious note, the Jensen diagnostic studies demonstrated atypical etiologies for massive hemorrhage including ischemic colitis, inflammatory bowel disease, and cancer. The rate of bleeding in these conditions may be more amenable to urgent colonoscopy (within 24 hours) rather than emergent colonoscopy in patients diverticular or angiodysplastic, hemorrhagic rates. Should the patient undergo a colonic lavage before colonoscopy? Longstreth 2 reported that 80.8% of patients had colonoscopy after electrolyte-polyethylene glycol solution purge, usually within 24 hours of admission. His report reflects the more typical approach to patients. Once the

6 304 F.G. Opelka, J.B. Gathright, Jr., and D.E. Beck patient undergoes observation and stabilization, the need for acute intervention seems avoided. Then the endoscopist may plan for a more controlled, stable, urgent colonoscopy with a lavage which occurs within the first 24 hours. The Longstreth Kaiser Permanente study demonstrated a broad scope of etiologies (see Table 20-2). The major benefit of colonoscopy depends on the ability to provide a definitive localization of ongoing active bleeding and the potential for therapy. Many landmarks for colonoscopy may be obscured during hemorrhage. Because of the inability to appreciate all intraluminal landmarks and locate the segment that is bleeding, once the endoscopist highlights a bleeding source, the region of the intestine requires a tattoo to mark the site with India ink. In such patients, if the hemorrhage continues and fails medical management, the tattoo greatly assists the surgeon in localizing the hemorrhage. The endoscopist has many therapeutic options to control the bleeding. Kovacs and Jensen 27 have described several therapeutic tools to control bleeding in upper and lower gastrointestinal hemorrhage. Therapeutic armamentarium for the colon includes thermal agents such as heater probes, bipolar coagulation, and laser therapy. Injection therapy primarily uses topical and intramucosal epinephrine. Mechanical therapy includes endoscopically applied clips (Figure 20-2). 27 Angiography Angiography is diagnostic and therapeutic in the treatment of intestinal hemorrhage. The clinical judgment for choosing angiography involves three different types of hemorrhage. First, acute, major hemorrhage with ongoing bleeding requires emergency angiography. Second, patients with an early blush during nuclear scintigraphy may benefit from therapeutic angiography. Finally, angiograms may define a potential source for hemorrhage in occult and recurrent gastrointestinal hemorrhage. To appreciate an angiographic blush of contrast, the study requires a hemorrhage rate of at least 1 ml/min. 32 Positive yields with angiography vary greatly. Patient selection will increase yields and avoid overuse of angiograms. Generally, reports demonstrate yields that range from 40% to 78% TABLE Final diagnosis in patients hospitalized for acute lower gastrointestinal hemorrhage 31 n (%) Colonic diverticulosis 91 (41.6) Colorectal malignancy 20 (9.1) Ischemic colitis 19 (8.7) Acute colitis, unknown cause 11 (5.0) Hemorrhoids 10 (4.6) Postpolypectomy hemorrhage 9 (4.1) Colonic angiodysplasia 6 (2.7) Crohn s disease 5 (2.3) Other 22 (10.1) Unknown 26 (11.9) Total 219 (100) FIGURE Clip applied to bleeding diverticular vessel. Angiography provides highly accurate localization of the site of bleeding (Figure 20-3) and the angiographic blush may suggest a specific etiology, but it lacks the accuracy of colonoscopy. Highly accurate localization provides for focused therapy. Hemorrhagic site may receive highly selective, intraarterial vasopressin infusion. The potent arterial contraction may reduce or halt the hemorrhage. Infusion rates of vasopressin are at concentrations of 0.2 U/min and may progress to 0.4 U/min. The systemic effects and cardiac impact of vasopressin may limit maximizing the dosage. Vasopressin controls bleeding in as many as 91% of patients. Bleeding may recur in as many as 50% of patients once the vasopressin is tapered. Angiographic technology also allows for arterial embolization to control hemorrhage. Superselective mesenteric angiography with current microcatheters allows for embolization of the vasa recta of the intestine, vessels as small as 1 mm. In the past, arterial embolization of larger vessels risked intestinal ischemia or infarction. The risk of intestinal infarctions of larger selective vessels may exceed 20%. Arteriography also has complication rates related to angiography, separate from the therapy delivered at the site of bleeding. These include arterial thrombosis, distant arterial emboli, and renal toxicity from the angiographic dye. Embolization therapy provides immediate arrest of the bleeding. Embolization uses a combination of agents to control bleeding including Gelfoam pledgets, coils, and polyvinyl alcohol particles. In 2001, Funaki et al. 37 reported experience with microcoil embolization in 27 patients. They succeeded in 93% and had reasonable long-term

7 20. Lower Gastrointestinal Hemorrhage 305 FIGURE Angiogram demonstrating extravasation (hemorrhage) in cecum. results 81%. Most of his patients had diverticular hemorrhages. His recurrent bleed patients had angiodysplasias. In a similar experience, Peck et al. 38 reported rebleeding in three of four patients with cecal angiodysplasias. The data suggest that angiodysplasias have multiple feeding vessels and may contribute to the recurrence. Operative Therapy Surgical therapy for massive lower intestinal bleeding is rare, often definitive, and associated with significant mortality. Most sources of bleeding spontaneously resolve or are controlled with the current therapeutic interventions. Few patients currently require surgical treatment. If the patient is hemodynamically unresponsive to the initial resuscitation, then radiographic, radionuclide, and endoscopic evaluations are usurped by the need for urgent surgery. Other patients may have the site of hemorrhage localized, yet the available therapeutic interventions fail to control the bleeding. Patient mortality increases with their transfusion requirements, suggesting the severity of the hemorrhage. Bender noted a reduced mortality (7%) for patients requiring less than 10 units of blood. The mortality increased to 27% for patients in excess of 10 units. 39 Therefore, once a patient reaches 6 7 units during the resuscitation and the hemorrhage remains ongoing, surgical intervention becomes eminent. The surgeon tailors the approach to the patient and depends on the diagnostic information gathered before the operation. All patients require an open laparotomy with a thorough examination of the entire intestine. The first objective in surgery focuses on the location of the intraluminal blood with the hope of segmentally isolating the possible sources of bleeding. If the colon visually appears filled with blood and the small intestine remains spared, the surgeon must still examine the entire abdomen and then focus on colonic sources of bleeding. If the small bowel contains blood, then the operative team has a larger area of concern and close inspection. Once the surgeon completes the initial visual inspection, a complete exploration ensues. The exploration begins in the stomach, duodenum, and considers possible missed upper gastrointestinal sources. Next, the small intestine must undergo examination from the ligament of Treitz to the ileocecal valve. Palpation of the intestine may demonstrate such etiologies as a Meckel s diverticulum, ileitis, colitis, or a GIST. Upon completion of the exploration phase, if no source appears obvious, the surgeon may consider intestinal enteroscopy. The enteroscope or colonoscope will expose the luminal surface and transilluminate the intestinal wall for occult lesions. Transillumination may identify vascular anomalies, small ulcers or tumors. Endoscopic access to the intestine may require upper enteroscope, a transgastric approach, a transcolonic approach, or insertion through the anus. Once a hemorrhage site is identified, the surgeon can perform an appropriate segmental resection. Intraoperative endoscopy is a technically difficult endeavor. A team approach with two surgeons or the availability of an experienced endoscopist is important to identify the elusive lesions causing the hemorrhage. If the source of bleeding cannot be found, and it appears to arise from the colon, the surgeon should perform a subtotal or total colectomy. Stable patients will tolerate a primary ileosigmoid or ileorectal anastomosis in this circumstance. Unstable patients require an end ileostomy with closure of the rectal stump or a mucous fistula. Once stable, the patient may return for ileostomy closure. The rectum and sigmoid colon require reexamination endoscopically to assure no bleeding persists. Before the endoscopy, a simple saline washout with a transanal catheter or via the rigid proctosigmoidoscope may provide for safe passage and careful examination of the remaining mucosa. The key concerns with operative management are, first, a delay in the decision to operate until the hemorrhage reaches a critical point beyond 10 units of blood. This seems to contribute to the high mortality rate. Second, mortality rates for patients requiring urgent surgery consistently reach a range hovering between 10% and 35%. 40 Few authors note mortalities less than 10% or greater than 40%. Third, notable recurrence rates of 10% are attributable to the limits of isolating the

8 306 F.G. Opelka, J.B. Gathright, Jr., and D.E. Beck precise cause of the bleeding. The rates of recurrence increase if a surgeon elects to perform a limited right or left colectomy without precise localization of the hemorrhage. Limited segmental colectomies continue to have high mortality rates and excessive persistent bleed rates of 20%. 41 A total colectomy offers the same mortality with a lower chance of recurrent or persistent hemorrhage. New Frontiers Horton and Fishman 42 commented about the advanced imaging within computerized tomography. Current thinly sliced, fast image acquisition combined with three-dimensional software packages has revolutionized the imaging of the vascular tree. Abdominal, and specifically intestinal vascular imaging now details smaller than named vessels. Current use focuses on chronic conditions such as mesenteric ischemia and inflammatory bowel disease. Case reports and animal modeling note it is a feasible study for gastrointestinal hemorrhage. New scanners promise even more with mm slices acquired in 0.4 seconds. Image acquisition synchronized with intravascular contrast may outline a site of contrast extravasation or blush. The detail available may define intestinal hemangiomas, arteriovenous malformations, and angiodysplasias. The sensitivity and specificity of computed tomographic angiography in patients with gastrointestinal hemorrhage are unknown and require further comparison studies to current diagnostic studies. Anderson 43 noted magnetic resonance angiogram creates images using the bright signal from blood. The three-dimensional images are reconstructed using computerized imaging to project a two-dimensional image that mimics a conventional angiogram. Further improvement develops from contrastenhanced magnetic resonance angiography (CEMRA). With current techniques, the resultant images are not as specific or as refined as an angiogram. The technique may detect the extravasation of blood pooling in various segments of the intestine. In addition to localizing the side, the study may distinguish small intestine versus large intestine. These studies may prove an enhancement when compared with nuclear scintigraphy. Wireless capsular endoscopy is an ideal diagnostic adjunct for patients with occult hemorrhage. 24,25 The first generation of capsules are mm. The capsules are easily swallowed and tolerated. The current system captures two images per second and transmits the images to a recording apparatus secured to a belt the patient wears. Transmitted images are later reviewed by the endoscopist. Lewis and Swain 44 reported the results from the first clinical trial. They noted a source of occult hemorrhage in 7 of 11 patients. The sites noted included angiodysplasia, ileal ulcers, and a tumor. Rossini et al. 45 corroborated their work in noting success in 7 of 10 patients. The evaluation and management of lower gastrointestinal hemorrhage remains a challenge for surgeons. An algorithm summarizing the management is provided in Figure FIGURE Algorithm for the management of lower gastrointestinal hemorrhage. References 1. Billingham RP. The conundrum of lower gastrointestinal bleeding. Surg Clin North Am 1997;77(1): Longstreth GF. Epidemiology and outcome of patients hospitalized with acute lower gastrointestinal hemorrhage: a populationbased study. Am J Gastroenterol 1997;92: Jensen DM, Machicado GA. Colonoscopy for diagnosis and treatment of severe lower gastrointestinal bleeding. Routine outcomes and cost analysis. Gastrointest Endosc Clin North Am 1997;7(3): Jensen DM. Diagnosis and treatment of severe hematochezia. The role of urgent colonoscopy after purge. Gastroenterology 1988;95(6): Landefeld CS, Goldman L. Major bleeding in outpatients treated with warfarin: incidence and prediction by factors known at the start of outpatient therapy. Am J Med 1989;87(2): Bramley PN, Masson JW, McKnight G, et al. The role of an open-access bleeding unit in the management of colonic haemorrhage. A 2-year prospective study. Scand J Gastroenterol 1996; 31(8): Richter JM, Christensen MR, Kaplan LM, Nishioka NS. Effectiveness of current technology in the diagnosis and management of lower gastrointestinal hemorrhage. Gastrointest Endosc 1995;41(2): Rossini FP, Ferrari A, Spandre M, et al. Emergency colonoscopy. World J Surg 1989;13(2): Jensen DM, Machicado GA. Diagnosis and treatment of severe hematochezia: the role of urgent colonoscopy after purge. Gastroenterology 1988;95: McGuire HW, Haynes BW. Massive hemorrhage from diverticular disease of the colon: guidelines for therapy based on bleeding pattern in fifty cases. Ann Surg 1972;175: Drummond H. Sacculi of the large intestine. Br J Surg 1916; 4: Meyers MA, Alonso DR, Gray GF, Baer JW. Pathogenesis of bleeding colonic diverticulosis. Gastroenterology 1976;71:

9 20. Lower Gastrointestinal Hemorrhage Ure T, Vernava AM, Longo WE. Diverticular bleeding. Semin Col Rect Surg 1994;5: McGuire HW, Haynes BW. Massive hemorrhage from diverticular disease of the colon: guidelines for therapy based on bleeding pattern in fifty cases. Ann Surg 1972;175: Finne CO III. The aggressive management of serious lower gastrointestinal bleeding. Probl Gen Surg 1992;9: Margolis AR, Heinbecker P, Bernard HR. Operative mesenteric angiography in the search for the site of bleeding in unexplained gastrointestinal hemorrhage. A preliminary report. Surgery 1960;48: McAllister K, Grogg K, Johnson D, et al. Endoglin, a TGF-beta binding protein of endothelial cells, is the gene for hereditary hemorrhagic telangiectasia type 1. Nat Genet 1994;8: Roskell DE, Biddolph SC, Warren BF. Apparent deficiency of mucosal vascular collagen type IV associated with angiodysplasia of the colon. J Clin Pathol 1998;51: Gostout CJ. Angiodysplasia and aortic valve disease: let s close the book on this association. Gastrointest Endosc 1995;42: Bhutani MS, Gupta SC, Markert RJ, Barde CJ, Donese R, Gopalswamy N. A prospective controlled evaluation of endoscopic detection of angiodysplasia and its association with aortic valve disease. Gastrointest Endosc 1995;42: Imperiale TF, Ransohoff DF. Aortic stenosis, idiopathic gastrointestinal bleeding and angiodysplasia: is there an association? Gastroenterology 1988;95: Boley SJ, Sprayregen S, Sammartano RJ, et al. The pathophysiologic basis for the angiographic signs of vascular ectasias of the colon. Radiology 1977;125: Beck DE. Diagnostic imaging. In: Beck DE, ed. Handbook of Colorectal Surgery. 2nd ed. New York: Marcel Dekker; 2003: Iddan G, Meron G, Glukhovsky A, et al. Wireless capsule endoscopy. Nature 2000;405: Appleyard M, Fireman Z, Glukhovsky A, et al. A randomized trial comparing wireless capsule endoscopy with push enteroscopy for the detection of small-bowel lesions. Gastroenterology 2000;119: Appleyard M, Glukhousky A, Swain P. Wireless-capsule diagnostic endoscopy for recurrent small-bowel bleeding. N Engl J Med 2001;344: Kovacs TO, Jensen DM. Recent advances in the endoscopic diagnosis and therapy of upper gastrointestinal, small intestinal, and colonic bleeding. Med Clin North Am 2002;86: Dusold R, Burke K, Carpentier W, et al. The accuracy of technetium-99m-labeled red cell scintigraphy in localizing gastrointestinal bleeding. Am J Gastroenterol 1994;89: Imbembo AL, Bailey RW. Diverticular disease of the colon. In: Sabiston D, ed. Textbook of Surgery. Philadelphia: WB Saunders; 1991: Ng DA, Opelka FG, Beck DE, et al. Predictive value of technetium Tc 99m-labeled red blood cell scintigraphy for positive angiogram in massive lower gastrointestinal hemorrhage. Dis Colon Rectum 1997;40: Bounds BC, Friedman LS. Lower gastrointestinal bleeding. Gastroenterol Clin 2003;32: Zuckerman DA, Bocchini TP, Birnbaum EH. Massive hemorrhage in the lower gastrointestinal tract in adults: diagnostic imaging and intervention. AJR Am J Roentgenol 1993;161: Koval G, Benner KG, Rosch J, et al. Aggressive angiographic diagnosis in acute lower gastrointestinal hemorrhage. Dig Dis Sci 1987;32: Browder W, Cerise EJ, Litwin MS. Impact of emergency angiography in massive lower gastrointestinal bleeding. Ann Surg 1986;204: Britt LG, Warren L, Moore OF III. Selective management of lower gastrointestinal bleeding. Am Surg 1983;49: Gomes AS, Lois JF, McCoy RD. Angiographic treatment of gastrointestinal hemorrhage: comparison of vasopressin infusion and embolization. AJR Am J Roentgenol 1986;146: Funaki B, Kostelic JK, Lorenz J, et al. Superselective microcoil embolization of colonic hemorrhage. AJR Am J Roentgenol 2001;177(4): Peck DJ, McLoughlin RF, Hughson MN, Rankin RN. Percutaneous embolotherapy of lower gastrointestinal hemorrhage. J Vasc Interv Radiol 1998;9: Bender JS, Wiencek RG, Bouwman DL. Morbidity and mortality following total abdominal colectomy for massive lower gastrointestinal bleeding. Am Surg 1991;57: ; discussion Corman ML. Vascular diseases. In: Corman ML, ed. Colon and Rectal Surgery. Philadelphia: JB Lippincott; 1993: Finne CO III. The aggressive management of serious lower gastrointestinal bleeding. Probl Gen Surg 1992;9: Horton KM, Fishman EK. CT angiography of the GI tract. Gastrointest Endosc 2002;55:S Anderson CM. GI magnetic resonance angiography. Gastrointest Endosc 2002;55:S Lewis BS, Swain P. Capsule endoscopy in the evaluation of patients with suspected small intestinal bleeding: the results of the first clinical trial. Gastrointest Endosc 2001;53:AB Rossini FP, Pennazio M, Santui R, et al. Early experience with wireless capsule diagnostic endoscopy in patients with small bowel bleeding. Am J Gastroenterol 2000;96:S

10