Chapter 36: Gastrointestinal Hemorrhage (Upper and Lower)

Gastrointestinal (GI) bleeding accounts for a significant number of hospitalizations every year with a median length of stay of 4 days and with greater than 30% of patients requiring intensive care unit (ICU) care.¹ Early diagnosis of the cause and appropriate intervention can significantly reduce ICU and hospital stays and improve morbidity and mortality. This chapter focuses on a discussion of the differential diagnosis of upper and lower GI bleeding and management principles.

UGIB has an incidence of 48 to 160 cases/100,000 adults per year.² Bleeding from an upper GI source is defined by overt presentation of bleeding represented as melena, hematemesis, or occasionally hematochezia, with localization of the bleeding lesion to the esophagus, stomach, or duodenum. UGIB is 5 to 6 times more common than lower GI bleeding (LGIB). The clinical presentation of UGIB can vary widely from clinically insignificant bleeding to life-threatening bleeding. Mortality from UGIB is generally between 10% and 14%.³ The management principles of UGIB focus on the risk stratification of patients and the early diagnosis and treatment of underlying causes.

UGIB is generally described as the presentation of melena, black tarry stools, or the description of hematemesis, vomiting of blood. The presentation may be insidious in onset or abrupt. Details obtained on history will suggest the duration of the bleeding and whether it is acute or chronic. Patients will often describe the presence of diarrhea associated with UGIB because blood passing through the GI tract often results in multiple loose bowel movements. Rapid UGIB may present as hematochezia (bright red blood or maroon stools from below) in addition to melena. Approximately 10% of patients with hematochezia may have upper GI sources of their bleeding.

Symptoms that accompany UGIB may reflect significant hemodynamic compromise; these include dizziness, light-headedness, chest pain, shortness of breath or dyspnea on exertion, weakness, or fatigue. Symptoms such as chest pain or shortness of breath suggest a more acute hemodynamic compromise and clinically significant blood loss. Patients without accompanying hemodynamic symptoms suggest a slow bleeding rate or a more chronic process. Symptoms such as abdominal pain, abdominal cramping, or postprandial pain may be helpful in narrowing the differential diagnosis of the source of the bleeding.

Essential to the history taking is obtaining information of comorbid conditions the patient may have and any medications the patient may be taking. Comorbid conditions can help in diagnosis, for example, liver disease may suggest a possible diagnosis of varices. Comorbid conditions are also important in treatment; for example underlying cardiac or pulmonary disease may be significant to the management of the patient. Medications that may play a role in UGIB include nonsteroidal anti-inflammatory drugs (NSAIDs), or antithrombotic agents, such as antiplatelet medication or anticoagulants. NSAIDs may cause UGIB through the formation of peptic ulcers, while therapy with antithrombotic agents may exacerbate UGIB from underlying lesions. A recent meta-analysis evaluated 35 randomized controlled trials (RCTs) and demonstrated that antiplatelet therapy with aspirin (ASA) (defined as 75-325 mg/d) increased the risk of UGIB 1.5 times compared to nonusers. The odds ratio (OR) of UGIB in patients taking ASA and clopidogrel was 1.86 when compared to ASA alone and the OR of GI bleeding with anticoagulants was 1.93 when compared to ASA alone.⁴ New anticoagulants, such as thrombin and factor Xa inhibitors, have also been shown to increase risk of GI bleeding; a meta-analysis analyzing 43 RCTs revealed an overall OR of GI bleeding of 1.45 when compared to traditional anticoagulants and an OR as high as 5 when used in the setting of acute coronary syndrome.⁵ Information about the dose, duration, and effects of these medications can help direct further care.

The physical examination focuses initially on hemodynamic evaluation. Assessing a patient's volume status with blood pressure monitoring and heart rate (HR) evaluation is crucial to determining the critical nature of the GI bleed. Patients with hemodynamic instability, suggested by HR greater than 100 beats/min, systolic blood pressure (SBP) less than 100 mm Hb, or the presence of orthostatic hypotension (defined by drop in SBP of > 20 mm Hb or rise in HR of > 20 beats/min from lying to sitting or sitting to standing) or the presence of syncope suggest a loss of blood volume of greater than 10% and indicates an urgency for volume resuscitation.

Other aspects of initial examination include assessment of the abdomen and rectum. Abdominal auscultation and palpation may help to narrow down the differential diagnosis. Patients with physical examination findings of peritonitis or clinical concern for perforation should be triaged for computed tomographic (CT) scan prior to further workup. Rectal examination is essential to confirm the presence of melena and support the etiology of UGIB and to help distinguish UGIB from lower gastrointestinal bleeding (LGIB). Patients in whom UGIB needs to be further confirmed and in whom endoscopy is being considered may benefit from nasogastric (NG) tube placement and gastric lavage. NG lavage provides supporting information of UGIB by witnessing the return of bright red blood or coffee ground material. It allows for the determination of ongoing UGIB: bright red blood that does not clear with lavage suggests active UGIB while clearance of red blood and return of coffee ground material suggest recent UGIB that may have stopped. Finally, NG lavage may allow for clearance of gastric contents facilitating views of the stomach on upper endoscopy. Some physicians advocate for NG lavage in all patients suspected of UGIB; in our practice, we determine its role in diagnosis and management on a case-by-case basis.

Laboratory studies supplement the assessment made on initial physical examination. The patient's Hb should be assessed immediately with a complete blood count (CBC) and compared to previous Hb levels to assess change; however the Hb level can be misleading in acute UGIB if there has not been sufficient time for the cardiovascular system to equilibrate and reflect the blood loss. Other laboratory values that may assist in the assessment are the blood urea nitrogen (BUN) and creatinine, the ratio of which can often be greater than 20:1 in patients with UGIB. All patients with evidence of GI bleeding should have a type-and-cross match sent to establish blood type given the possible need for blood transfusion.

Initial Management

The goals of initial management focus on resuscitation. Additionally, management decisions must be made to minimize ongoing bleeding and to minimize the impact of the UGIB on other organ systems (Figure 36–1).

Resuscitation is initiated and driven by amount of blood loss as assessed on CBC and by hemodynamic compromise. All patients should have 2 large-bore peripheral venous access catheters or a central venous access placed. Patients with altered mental status or respiratory compromise should be intubated for airway protection. Patients with significant hemodynamic compromise should be admitted to the ICU for close monitoring and management. Volume replacement should be initiated immediately with normal saline or lactated Ringer solution. Colloid solutions can be given.

The goal of blood transfusion is to increase the delivery of oxygen to the tissues. The need for blood transfusion is based on initial Hb and assessment of whether UGIB is ongoing and significant. Patients with massive UGIB may benefit from an aggressive transfusion strategy; patients with clinically stable UGIB and low-risk or intermediate-risk UGIB are more likely to benefit from a restrictive transfusion strategy. A recent RCT of patients with acute UGIB comparing a restrictive transfusion strategy (transfusion for Hb < 7) versus a liberal transfusion strategy (transfusion for Hb < 9) found that patients with a restrictive transfusion strategy had higher probability of survival at 6 weeks, decreased rates of rebleeding, and lower number of adverse events. This benefit is most clearly understood in patients with cirrhosis and portal hypertension-related variceal bleeding.⁶ Aggressive reconstitution of blood volume has previously been shown to induce rebound increases in portal pressure, which may precipitate rebleeding of varices; a restrictive transfusion strategy was shown to be beneficial in these patients. Additionally, a restrictive transfusion strategy was shown to be beneficial in patients with nonvariceal UGIB, which may be due to decreased transfusion-related complications, decreased transfusion-related abnormalities in coagulation, or decreased cardiac complications such as pulmonary edema.⁶ Patients should be evaluated on a case-by-case basis for transfusion needs, taking into consideration age, coexisting comorbidities, and hemodynamic status; however, a restrictive transfusion strategy is favored.

Patients should concurrently have prothrombin time (PT) and activated partial thromboplastin time (aPTT) levels checked and corrected as necessary. In patients with ongoing GI bleeding, the goal international normalized ratio (INR) is less than 1.5 and fresh frozen plasma (FFP) can be given to achieve this in the short term. Patients who are hemodynamically stable can proceed to upper endoscopy while transfusion is being given. A recent study demonstrated no difference in risk of rebleeding following endoscopic therapy in patients with mild-to-moderate elevated INR compared to normal INR.⁷ Patients with history of thrombocytopenia should have platelet levels checked and corrected to values greater than 50,000/mm³.

After the initial management, the focus is on minimizing ongoing UGIB. UGIB is often divided between nonvariceal bleeding and variceal bleeding and initial assessment for chronic liver disease helps to differentiate these possibilities. The most common cause of nonvariceal UGIB is peptic ulcer disease (PUD). Thus, to reduce ongoing bleeding from this potential cause, all patients are recommended to receive intravenous (IV) proton pump inhibitor (PPI) therapy until further risk stratification is completed. This is generally given in the form of intravenous pantoprazole starting with a bolus of 80 mg followed by a continuous infusion of 8 mg/h. A RCT of patients receiving PPI infusion versus placebo prior to endoscopy revealed lower need for endoscopic intervention and decreased length of hospital stay in patients treated with pre-endoscopic PPI.⁸ PPI therapy is thought to stabilize blood clot and reduce risk of bleeding by decreasing levels of gastric acid, which can inhibit platelet aggregation and lead to clot lysis.⁹

Patients in whom chronic liver disease is suspected and who have a risk of variceal bleeding are recommended to receive pharmacologic therapy in conjunction with endoscopic therapy as well. Somatostatin and somatostatin analogues such as octreotide and vapreotide have been evaluated in the pre-endoscopic period for patients in whom variceal bleeding is suspected. Octreotide is the only medication of this class currently available in the United States; intravenous octreotide infusion starting with 50 μg bolus followed by 50 μg/h continuous infusion is recommended to be given in all patients in whom variceal bleeding is suspected followed by early endoscopy. A meta-analysis of 8 randomized trials revealed improved endoscopic control of initial bleeding in patients who received pharmacologic therapy in addition to endoscopic therapy.¹⁰ Terlipressin, a synthetic analogue of vasopressin, has also been evaluated and been shown to be effective in acute variceal bleeding; however, this medication is not yet available in the United States. A recent study comparing somatostatin, octreotide, and terlipressin in a randomized fashion showed no difference in the hemostatic effects or safety between these medications.¹¹

Additionally, patients with cirrhosis and suspected UGIB are recommended to receive short-term prophylactic antibiotics. Prophylactic antibiotics are intended to decrease the risk of bacterial infection such as spontaneous bacterial peritonitis and other infections; studies evaluating the role of prophylactic antibiotics have shown decreased rates of infection and improved survival.¹² Norfloxacin 400 mg twice daily given orally for 7 days is the recommended antibiotic, though other fluoroquinolones have also been shown to be effective. Additionally, IV ceftriaxone 1 g/d has been studied and shown to be effective.

An estimated 80% of patients with UGIB will spontaneously stop without intervention other than those described earlier. Patients who have completed initial resuscitation should be evaluated for evidence of ongoing bleeding, risk of rebleeding, and risk of mortality. This assessment should be made first based on clinical and laboratory findings, and then may be modified after endoscopic findings. Several scoring systems have been developed to stratify patients into high risk or low risk for rebleeding and mortality (Table 36–1). In previous studies, clinical predictors of rebleeding and mortality include age greater than 65, shock, comorbid illnesses, and clinically significant UGIB as evidenced by hematemesis, hematochezia, or melena; low initial Hb or high transfusion requirements; and hemodynamic instability or syncope. Additionally, sepsis, elevated urea, creatinine, or serum aminotransferase levels have been predictors of poor clinical outcome.^2,3 Two scoring systems commonly used to risk-stratify patients are the Blatchford score and the Rockall score, which can be initially calculated on clinical and laboratory data and updated based on endoscopic findings. Patients considered high risk should undergo upper endoscopy immediately after resuscitation is complete and the patient is clinically stable to undergo a procedure. These patients often require ICU stay, either for monitoring or to facilitate bedside endoscopy. If patients demonstrate evidence of active UGIB, we will often recommend intubation in the ICU for airway protection. Prior to endoscopy, it is often helpful to consider use of a promotility agent such as metoclopramide or erythromycin to facilitate clearance of the stomach and improve endoscopic visualization. Patients should be monitored closely in the interval between presentation and endoscopy with every 6- to 8-hour CBC depending on the clinical presentation.

For nonvariceal UGIB, endoscopic features can also be used to predict risk of rebleeding and mortality. The presence of high-risk stigmata as defined by the Forest classification has been associated with higher rates of rebleeding and mortality. High-risk stigmata include actively bleeding vessel (Forest class IA), oozing blood (IB), and nonbleeding visible vessel (IIA). Adherent clot (IIB) is felt to be moderate risk. Flat pigmented spots (IIC) and clean-based ulcers (III) are low-risk lesions for rebleeding. The complete Rockwell score (see Table 36–1) can be used when clinical criteria and endoscopic criteria are available to determine risk of rebleeding and mortality.

Varices are present in approximately 50% of patients with cirrhosis and their presence correlates with severity of liver disease. Variceal hemorrhage occurs at a rate of 5% to 15% per year; predictors of variceal hemorrhage include size of varices, decompensated cirrhosis (Child-Pugh class B/C), and the endoscopic presence of red wale signs at the time of screening endoscopy (citation). Patients who have variceal bleeding have a median rebleeding rate of 63% within 1 to 2 years and mortality of at least 20% at 6 weeks.¹³ Predictors of rebleeding and mortality include Child-Pugh class, Model for End-Stage Liver Disease (MELD) score, renal failure, bacterial infection, hypovolemic shock, active bleeding at endoscopy, hepatocellular carcinoma, and hepatic venous pressure gradients (HPVGs) of greater than 20 mm Hb when measured within 24 hours of variceal bleed. A recent study described a MELD-based model to predict risk of mortality among patients with acute variceal bleeding and was demonstrated to be more accurate than other proposed prediction models.¹⁴

The most common causes of UGIB remain similar over time; a recent study reviewing the most common causes of UGIB at a large urban hospital cited PUD as the diagnosis in 34% of cases followed by variceal bleeding in 33% of cases.¹ Ninety percent of all PUD is caused by either infection with the gram-negative bacterium Helicobacter pylori or by the use of NSAIDs. H pylori is found in as many as 90% of patients with duodenal ulcers and 70% of patients with gastric ulcers. The bacteria are transmitted via the fecal-oral route and are most commonly acquired in childhood. Most infected individuals are asymptomatic with no clinical consequence; a subset of patients can develop decreased mucosal barriers and increased mucosal susceptibility to gastric acid damage. Patients with persistent infection and chronic inflammatory changes can develop ulcers.

NSAIDs are a widespread and commonly used class of medications. Chronic use with even low-dose aspirin for cardioprotective measures can lead to mucosal changes throughout the GI system with ulcerations in the stomach, duodenum, and even the distal small bowel and colon. NSAIDs result in prostaglandin inhibition with subsequent decreased mucosal barriers and increased mucosal susceptibility to gastric acid injury. In addition, NSAIDs can cause direct mucosal injury resulting in erosions and ulcerations. PUD can also be seen in patients with severe stress, trauma, burns, sepsis, and prolonged ICU stays.

Variceal formation is caused by collateral vessels in the GI tract as a consequence of portal hypertension in patients with chronic liver disease. The most common site of varices is the esophagus though varices can also form in the gastric fundus and cardia. Portal hypertension can cause other gastric changes in the form of portal hypertensive gastropathy, a diffuse mucosal change in the stomach. Variceal bleeding can present as acute-onset severe UGIB with associated hemodynamic compromise. The decrease in portal pressures by bleeding can lead to self-resolution; however, the risk of rebleeding remains high and can be exacerbated by aggressive volume resuscitation, as described earlier.

Other causes of UGIB include vascular malformations such as arteriovenous malformations (AVMs) or Dieulafoy lesions, traumatic lesions such as Mallory-Weiss tears of the esophagus, or malignant lesions such as gastric cancer or esophageal cancer.

Endoscopic therapy, in combination with pharmacologic therapy, is the main tool used for diagnosis and treatment of UGIB. Early endoscopy, defined by esophagogastroduodenoscopy (EGD) within the first 24 hours after presentation, to localize the site of bleeding and treat bleeding lesions has been shown to improve outcomes and reduce lengths of hospitalization. Most endoscopic findings can be treated within the same setting to stop active bleeding and reduce risks of rebleeding.

Nonvariceal UGIB

In patients with nonvariceal UGIB, high-risk stigmata, such as actively bleeding vessel, oozing vessel, or nonbleeding visible vessel, should be treated endoscopically. It is unclear whether treatment of adherent clot is beneficial; generally accepted practice is to attempt to gently dislodge the clot and treat the lesion based on appearance underneath. Low-risk lesions, such as pigmented spots or clean-based ulcers, do not require endoscopic treatment.

Endoscopic treatment consists of one or more techniques to attempt to stop active bleeding and prevent rebleeding. Tools available to the endoscopist include injection of epinephrine at a concentration of 1:10,000, thermocoagulation of the lesion with bipolar electrocautery probe, or placement of clips. A recent meta-analysis suggested that injection therapy is better than no therapy but that combined therapy (injection plus thermocoagulation or injection plus clips) is better than monotherapy with injection alone. Monotherapy with thermocoagulation or clips is equivalent in efficacy to combined modality therapy.¹⁵ Routine second look endoscopy is not recommended and should be considered on a case-by-case basis. Evidence of rebleeding or inability to identify the bleeding lesion on initial endoscopy is generally considered indication for second endoscopy.

Patients with bleeding that cannot be localized on endoscopic evaluation or patients who have ongoing bleeding despite endoscopic therapy should be considered for surgery or for interventional radiology directed angiography and percutaneous embolization.

Patients with low-risk clinical features and low-risk endoscopic features may be discharged home after endoscopy; RCTs of this highly select group of patients suggest no difference in outcomes with a policy of early discharge versus inpatient observation. These patients need to have close outpatient monitoring and follow-up.

All patients should be discharged with oral PPI therapy to decrease risk of rebleeding with dosage and duration determined by etiology of bleeding; most patients are given once-daily PPI therapy for 8 weeks with close outpatient follow-up. Patients with PUD should be evaluated for H pylori with treatment of any positive results to minimize risk of recurrent ulcers and facilitate ulcer healing. These patients should have eradication confirmed 2 weeks after completion of their PPI therapy. Patients with NSAID use should discontinue use if possible. Patients who require low-dose ASA for cardiovascular protection should resume therapy when the risk of cardiovascular complications outweighs the risk of rebleeding; this appears to occur as early as 5 days after the onset of bleeding and decisions about timing of ASA resumption should be made with a multidisciplinary approach. Patients who require ASA and clopidogrel who have UGIB should continue on PPI therapy. The role of PPI therapy in combination with ASA and clopidogrel was evaluated in a RCT (COGENT trial), which examined the risk of adverse GI events and adverse cardiovascular events in patients taking ASA and clopidogrel randomized to PPI use or placebo. This study found a statistically significant reduction in adverse GI events in patients taking once-daily omeprazole with no difference in cardiovascular events between the 2 groups.¹⁶ This study supports the benefit of PPI use in prevention of primary GI events. The benefit of PPI in preventing rebleeding is also likely to outweigh the hypothesized increase in risk of cardiovascular events attributable to the interaction between PPI therapy and clopidogrel. Recent guidelines on antiplatelet therapy management have been changed to reflect this consensus.

Variceal UGIB

Early endoscopy and endoscopic intervention has also been shown to improve outcomes in patients with variceal bleed. Variceal bleeding is identified when active bleeding is seen from a varix, when a “white nipple” is seen overlying the varix, when clots are seen overlying a varix, or when varices are seen with no other source of UGIB identified. Varices are described as number of columns and are graded into categories of small or large based on their appearance. Endoscopic interventions available for the treatment of variceal bleeding includes endoscopic variceal ligation (EVL) with the use of rubber bands placed at the base of the varix or use of sclerotherapy with injection of agents such as N-butyl-cyanoacrylate or isobutyl-2-cyanoacrylate. Endoscopic therapy in conjunction with pharmacologic therapy is the treatment of choice for bleeding varices; in general, endoscopic treatment with EVL is the preferred form of treatment by consensus. Sclerotherapy is recommended if EVL is not technically feasible.

Patients who fail endoscopic therapy or patients in whom early rebleeding occurs should be evaluated for shunt therapy, either through transjugular intrahepatic portosystemic shunt (TIPS) placement or through surgical shunt placement. One recent study demonstrated survival benefit with early TIPS placement (within 24 hours of bleeding) in patients with acute variceal bleeding and HVPG greater than 20 mm Hb.¹⁷ Patients who undergo shunt placement experience higher rates of hepatic encephalopathy; thus, while TIPS is felt to be equally effective in treatment of variceal bleeding with no difference in overall survival, it is recommended as a rescue therapy in patients who have failed EVL and pharmacologic therapy because of this significant side effect. Balloon tamponade via placement of a Blakemore or Minnesota tube may be used to temporarily control bleeding prior to definitive therapy in patients with acute variceal bleeding; this is generally only recommended in patients with massive variceal bleeding as part of temporizing measures until definitive therapy with endoscopy or TIPS is available.

Patients who survive an episode of acute variceal hemorrhage are at high risk of rebleeding and death. Thus it is important that patients be treated to minimize risk of recurrence. Patients who undergo EVL should be started on a nonselective beta-blocker as soon as they have remained 24 hours without evidence of rebleeding to reduce portal pressures. Patients who have EVL are recommended to undergo repeat EGD at 7- to 14-day intervals until complete obliteration of varices is seen, usually requiring between 2 and 4 sessions. Combination of endoscopic therapy plus pharmacologic therapy with nonselective beta-blocker has been shown to be more effective than EVL alone with decreased rebleeding rates in 2 RCTs. Patients who undergo TIPS for treatment of variceal bleeding do not require any further therapy. All patients with clinically significant variceal bleeding should be referred for liver transplant evaluation if they are appropriate candidates.¹⁸

UGIB should be confirmed quickly by clinical examination with identification of melena or evidence of hematemesis or evidence of positive NG lavage. The patient's hemodynamic status should be assessed and resuscitation initiated rapidly to correct abnormalities. Empiric therapy with intravenous PPI and/or intravenous octreotide can help to reduce active bleeding and/or stabilize clot. Early endoscopic evaluation and therapy of high-risk lesions improve outcomes and reduces length of hospital stay. Patients should be managed endoscopically if high-risk lesions are identified. Patients should be monitored after intervention for rebleeding with continued adjuvant pharmacologic therapy (PPI, octreotide, nonselective beta-blockers, antibiotics) as indicated to decrease rates of rebleeding and mortality.

LGIB is defined as bleeding from a lesion distal to the ligament of Treitz. The most common site of LGIB is the colon; bleeding from the small intestine is much less common. LGIB generally presents clinically as hematochezia or bright red blood from the rectum. LGIB can present in an acute manner or a subacute manner and is managed with many of the same principles used to manage UGIB.

As described, most patients with LGIB will describe the presentation of hematochezia that occurs acutely or subacutely. Hematochezia can present as large quantity or small quantity, can be described in isolation or to be mixed in with brown stool. Patients may have associated symptoms that can help narrow the differential diagnosis such as symptoms of diarrhea, abdominal pain, or tenesmus. Eliciting past medical history and medication use is as important in LGIB as it is in UGIB and may help to narrow the differential diagnosis. Understanding and managing comorbid conditions are similar to that in patients with UGIB and patients should be evaluated for the presence of underlying cardiac or pulmonary diseases.

Similar to UGIB, patients' initial evaluation includes evaluation of vital signs and clinical status. This involves evaluation of heart rate, blood pressure, and orthostatic measurements. Assessing the volume status and proceeding with resuscitation are similar as in patients with UGIB. Patients should have 2 large-bore IVs placed, followed by infusion of crystalloid, colloids, and blood to correct the volume status. Patients with severe volume loss and clinical consequences should be admitted to the ICU for closer monitoring and management. Patients with clinically significant hematochezia should be evaluated for the possibility of UGIB; approximately 10% of patients with UGIB will present with hematochezia. These patients should be evaluated with NG lavage for risk stratification and may require EGD prior to any additional workup to rule out UGIB causes.

Patients who are clinically stable can be further evaluated based on clinical presentation. As with UGIB, checking serum Hb further differentiates the clinical impact of bleeding. Restrictive transfusion strategies are recommended in patients with low or intermediate risk. Additional laboratory studies such as lactic acid may assist in narrowing the differential diagnosis.

The differential diagnosis of overt LGIB is broad. Diverticular bleeding is the most common cause of LGIB and accounts for 40% of all LGIB.¹⁹ Diverticulosis increases with age and the overall prevalence is estimated to be 30% to 50% with as many as 60% of people over 80 years being affected. The incidence of bleeding is estimated to be 5% to 50% in patients with diverticulosis. The clinical presentation of diverticular bleeding is generally consistent with arterial bleeding. Patients may have acute onset of painless hematochezia that ceases spontaneously in most cases with recurrent bleeding in 14% to 38% of cases. Diverticular bleeding can come from either the left colon or the right colon and localization of the site of bleeding can often be challenging. Because most cases resolve spontaneously without rebleeding, diagnosis is often made after exclusion of other potential cases of bleeding.

Ischemic colitis accounts for 1% to 19% of LGIB. This results from sudden loss of blood flow to the mesenteric vessels and most commonly affects the areas of the colon described as “watershed” areas: the splenic flexure and the rectosigmoid junction. Ischemic colitis may occur as a result of an embolic or a thrombotic event or because of transient hypoperfusion or vasospasm of the mesenteric vessels. Patients generally describe sudden onset of abdominal pain accompanied by hematochezia and may have advanced age, or significant cardiovascular disease. Ischemic colitis is diagnosed by a classic appearance of segmental inflamed mucosa on colonoscopy and is generally treated with supportive measures and reversal of the underlying etiology.

Other lesions that can less commonly lead to LGIB include AVMs or Dieulafoy lesions. AVMs are ectatic blood vessels that can be located anywhere in the GI tract but are predominantly found in the cecum and ascending colon. They account for approximately 11% of LGIBs. They may appear as isolated lesions or in clusters of multiple small, flat, red telangiectasias. Dieulafoy lesions are arterial blood vessels that appear close to the mucosal surface that intermittently bleed. Identifying these vascular lesions can be challenging.

Finally, hemorrhoidal bleeding represents 5% to 10% of acute LGIB. Hematochezia tends to be low volume with little to no clinical significance. Episodes can be isolated or sporadic and often is described by the appearance of blood on the toilet paper or on the outside of stool.

Other causes of LGIB represent a wide variety of possible diagnoses including hemorrhagic infectious diarrhea, colitis from infection, inflammatory bowel disease or radiation change, and malignancy, among others.

Similar to UGIB, the gold standard for evaluation of LGIB is endoscopic assessment via colonoscopy, though several other diagnostic tools are available. Patients who present with hematochezia hemodynamic compromise should be evaluated for the possibility of UGIB source. These patients should have NG tube placement with NG lavage to assist in localizing the source of bleeding. Patients with clinical evidence of rapid/ongoing GI blood loss should undergo emergent EGD to rule out UGI etiologies (Figure 36–2).

Patients with hemodynamic stability and hematochezia should be evaluated first by colonoscopy. The patient should be prepared by using a polyethylene glycol-based solution to cleanse the colon prior to procedure. This can be consumed by having the patient drink the preparation or by administering it through the NG tube. Preparation prior to colonoscopy is essential to facilitate visualization, increase the diagnostic yield of the procedure, and minimize the risk of complication such as perforation. The diagnostic yield of colonoscopy varies widely; patients with diverticular bleeding often have low yield at localization of the bleeding site. Alternate etiologies such as ischemic colitis, vascular lesions, infectious/inflammatory colitis, or malignancy can be diagnosed by colonoscopy. Identification of bleeding diverticula or vascular lesions can be treated endoscopically using the same modalities as previously described such as a combination of injection therapy, thermocoagulation, or clip placement.

Patients with nonlocalization of bleeding site and high clinical suspicion for diverticular bleeding can be managed conservatively with close follow-up to assess for ongoing bleeding. If massive bleeding is present, clinical compromise is noted, or ongoing bleeding is suspected, patients may benefit from evaluation by one of several radiologic tests such as CT angiogram, standard angiography, or tagged red blood cell scanning.

The widespread availability of CT angiography has made this a common first test to localize bleeding site. Several studies have demonstrated the high accuracy of CT angiography in identifying and localizing active bleeding either through identification of active extravasation of blood or to localization of intraluminal contents consistent with recent bleeding. The advantage of CT angiography is largely in the ability to perform this evaluation quickly. If localization is successful, patients are able to undergo subsequent management either through IR-directed percutaneous embolization or through surgical resection. IR-directed embolization is highly effective at treating LGIB and often is used with priority over colonoscopy or surgery in patients who have clinically significant GI bleeding. Tagged red blood cell scans are also able to aid in localization of LGIB but are less in favor than CT angiogram because of the time required to perform the evaluation and its inability to localize the precise source of bleeding/lack of structural imaging.

The determination of which localization modality is optimal in LGIB remains an ongoing discussion and should be handled with a multidisciplinary approach between gastroenterologists, radiologists, and surgeons taking into account the clinical stability of the patient and the time required to prepare for colonoscopy.

LGIB follows similar principles for management as UGIB; evaluation begins with confirmation of hematochezia and evaluation for a possible UGIB. Evaluating blood loss and resuscitation is the first step in management. Patients who are clinically stable are recommended to undergo preparation followed by colonoscopy for localization and possible treatment of bleeding lesions; patients with hemodynamic instability or ongoing GI blood loss are recommended to undergo CT angiogram for localization of bleeding site followed by conventional angiography and embolization of bleeding vessel or surgery for definitive treatment.