In the Clinic Acute Gastrointestinal Bleeding In theClinicunmhospitalist.pbworks.com/w/file/fetch/68209086/0000605-201308060... · In the Clinic Acute Gastrointestinal Bleeding Prevention

Inthe

ClinicIn the Clinic

AcuteGastrointestinalBleedingPrevention page ITC2-2

Presentation and Diagnosis page ITC2-3

Treatment page ITC2-7

Practice Improvement page ITC2-13

Tool Kit page ITC2-14

Patient Information page ITC2-15

CME Questions page ITC2-16

Physician WritersMeeta Prasad Kerlin, MD, MSCEJeffrey L. Tokar, MD

Section EditorsDeborah Cotton, MD, MPHDarren Taichman, MD, PhDSankey Williams, MD

The content of In the Clinic is drawn from the clinical information and educationresources of the American College of Physicians (ACP), including PIER (Physicians’Information and Education Resource) and MKSAP (Medical Knowledge and Self-Assessment Program). Annals of Internal Medicine editors develop In the Clinicfrom these primary sources in collaboration with the ACP’s Medical Education andPublishing divisions and with the assistance of science writers and physician writ-ers. Editorial consultants from PIER and MKSAP provide expert review of the con-tent. Readers who are interested in these primary resources for more detail canconsult http://pier.acponline.org, http://www.acponline.org/products_services/mksap/15/?pr31, and other resources referenced in each issue of In the Clinic.

CME Objective: To review current evidence for the prevention, presentation anddiagnosis, treatment, and practice improvement of acute gastrointestinal bleeding.

The information contained herein should never be used as a substitute for clinicaljudgment.

© 2013 American College of Physicians

Downloaded From: http://annals.org/ by a Unviersity of New Mexico User on 08/13/2013

© 2013 American College of Physicians ITC2-2 In the Clinic Annals of Internal Medicine 6 August 2013

1. Hreinsson JP, Kalaitza-kis E, GudmundssonS, Björnsson ES. Up-per gastrointestinalbleeding: incidence,etiology and out-comes in a popula-tion-based setting.Scand J Gastroen-terol. 2013;48:439-47.[PMID: 23356751]

2. Yavorski RT, Wong RK,Maydonovitch C, et al.Analysis of 3,294 casesof upper gastrointesti-nal bleeding in mili-tary medical facilities.Am J Gastroenterol.1995;90:568-73.[PMID: 7717312]

3. Targownik LE, Nabal-amba A. Trends inmanagement andoutcomes of acutenonvariceal uppergastrointestinal bleed-ing: 1993-2003. ClinGastroenterol Hepatol.2006;4:1459-1466.[PMID: 17101296]

4. McColl KE, el-NujumiAM, Chittajallu RS, etal. A study of thepathogenesis of Heli-cobacter pylori nega-tive chronic duode-nal ulceration. Gut.1993;34:762-8.[PMID: 8314508]

5. Kurata JH, NogawaAN. Meta-analysis ofrisk factors for pepticulcer. Nonsteroidalantiinflammatorydrugs, Helicobacterpylori, and smoking. JClin Gastroenterol.1997;24:2-17.[PMID: 9013343]

6. Talamini G, ZamboniG, Cavallini G. Antralmucosal Helicobacterpylori infection densi-ty as a risk factor ofduodenal ulcer. Di-gestion. 1997;58:211-7. [PMID: 9243115]

7. Barkun AN, Bardou M,Pham CQ, Martel M.Proton pump in-hibitors vs. histamine2 receptor antago-nists for stress-relatedmucosal bleedingprophylaxis in critical-ly ill patients: a meta-analysis. Am J Gas-troenterol.2012;107:507-20.[PMID: 22290403]

8. Arisawa T, Tahara T,Shibata T, et al. Asso-ciation between ge-netic polymorphismsin the cyclooxyge-nase-1 gene promot-er and peptic ulcersin Japan. Int J MolMed. 2007;20:373-8.[PMID: 17671743]

Prevention

Acute gastrointestinal (GI) bleeding is common in both the outpatientsetting and the emergency department. Annual U.S. incidence rates overthe past decade are approximately 90–108 per 100 000 persons (1), lead-

ing to approximately 300 000 hospitalizations annually. Most cases are due tononvariceal sources of bleeding (e.g., peptic ulcers) and continue to be associatedwith significant mortality (3–14%) and health economic burden (1–3). The inci-dence of nonvariceal bleeding may be decreasing in the West largely because ofdecreased incidence of Helicobacter pylori infection and increased awareness andimplementation of ulcer-prevention strategies in users of nonsteroidal anti-inflammatory drugs (NSAIDs) (1). However, identifying patients with acute GIbleeding who are in danger of serious adverse events and establishing evidence-based treatment plans are essential in both primary and specialty care.

with lower GI bleeding include in-flammatory bowel disease, infectiouscolitis, neoplasia, angioectasias, andbenign anorectal disease (14).

Approximately 10–20% of patientswith GI bleeding have “obscure”bleeding, defined as an unknowncause despite evaluation with esopha-gogastroduodenoscopy (EGD),colonoscopy, and radiographic smallbowel imaging. Approximately halfof these patients have recurrent orpersistent bleeding and are furthersubclassified as obscure-overt (pas-sage of visible blood with melena orhematochezia) or obscure-occult(iron-deficiency anemia and/or posi-tive for fecal occult blood) (15).Many such patients have bleedingsources in the small intestine, nowsometimes referred to as “mid-GIbleeding” (between the ligament ofTreitz and the ileocecal valve). An-gioectasia is the most common causeof small-bowel bleeding in the West,accounting for 70–80% of cases (16).

Can acute GI bleeding be prevented?Prevention of acute GI bleeding de-pends on the risk factors and causes.For example, reducing use ofNSAIDs and administering antacidtreatment with H

2-inhibitors or pro-

ton-pump inhibitors (PPIs) preventspeptic ulcer bleeding. Prophylacticacid suppression should be consideredin selected hospitalized patients whoare at increased risk for gastroduode-nal ulceration and bleeding, including

Who is at risk for acute GI bleeding?Risk factors for acute GI bleedingvary according to the site and cause.Upper GI bleeding most often re-sults from peptic ulcer disease (ap-proximately one quarter of all cases),the primary risk factors for whichinclude NSAIDs and H. pylori infec-tion and, less commonly, increasedgastric acid production (e.g., theZollinger-Ellison syndrome). Smok-ing (4–6), severe physiologic stress(7), and various host factors (e.g., genetic polymorphisms affecting cy-clooxygenase and prostaglandin pro-duction) (8, 9) may increase risk forpeptic ulcers even in persons withoutconcomitant H. pylori infection orNSAID exposure. Although spicyfoods may cause GI symptoms, thereare no convincing data that they in-crease the risk for peptic ulcers. Oth-er risk factors for acute upper GIbleeding include varices, esophagitis,vascular abnormalities (e.g., angioec-tasias, arteriovenous malformations,Dieulafoy lesions), Mallory–Weisstear from protracted vomiting, andbenign and malignant neoplasms(10–13).

Lower GI bleeding, historically de-fined as bleeding from a source distalto the ligament of Treitz, also resultsfrom several causes with distinct riskfactors. Diverticulosis is the mostcommon cause of hematochezia, ac-counting for up to half of all cases,particularly in patients older than 65years. Other risk factors associated


© 2013 American College of PhysiciansITC2-3In the ClinicAnnals of Internal Medicine6 August 2013

9. Malaty HM, GrahamDY, Isaksson I, En-gstrand L, PedersenNL. Are genetic influ-ences on peptic ulcerdependent or inde-pendent of geneticinfluences for Heli-cobacter pylori infec-tion? Arch InternMed. 2000;160:105-9.[PMID: 10632311]

10. Boonpongmanee S,Fleischer DE, Pezzul-lo JC, et al. The fre-quency of peptic ul-cer as a cause ofupper-GI bleeding isexaggerated. Gas-trointest Endosc.2004;59:788-94.[PMID: 15173790]

11. Enestvedt BK, Gral-nek IM, Mattek N,Lieberman DA, EisenG. An evaluation ofendoscopic indica-tions and findingsrelated to nonva-riceal upper-GI hem-orrhage in a largemulticenter consor-tium. GastrointestEndosc. 2008;67:422-9. [PMID: 18206878]

Prevention... Risk factors for, and therefore prevention of, acute GI bleeding depends on the site and cause of bleeding. In general, minimizing use and appro-priate prescribing of NSAIDs, antiplatelet agents, and anticoagulants, as well asjudicious primary and secondary prophylactic acid suppression in selected pa-tients, are effective measures for ulcer-related upper GI bleeding. Nonselectivebeta-blockers and endoscopic therapy for esophageal and gastric varices are ef-fective for primary and secondary prevention of variceal bleeding. Few measuresare helpful in preventing lower GI bleeding, except for reducing exposure toNSAIDS, anticoagulants, and antiplatelets.

CLINICAL BOTTOM LINE

Presentation andDiagnosis

measures can prevent diverticularbleeding or bleeding related to an-gioectasias. Although high-fiber dietsare often recommended, the role offiber in the pathogenesis or progres-sion of diverticulosis is debatable. Sur-gical intervention for diverticulosis isnot routinely done for prevention, be-cause risk for diverticular bleeding islow relative to the morbidity of pro-phylactic surgical resection (19).

patients who are mechanically venti-lated or who have coagulopathy orthrombocytopenia, traumatic brain orspinal cord injury, or a history ofburns (7, 17). In patients with chronicliver disease, nonselective beta-blockers (such as propanolol ornadolol) to reduce portal hypertensionand endoscopic interventions, such asband ligation (18), can effectively pre-vent variceal bleeding. No specific

it can occur with small-bowel bleed-ing and even slowly bleeding rightcolonic lesions. The source of hema-tochezia is usually the colon. Up to10% of upper GI bleeding episodespresent with hematochezia, which isoften associated with signs and symp-toms of hemodynamic instability.

What are the common causes ofupper and lower GI bleeding?Major causes of GI bleeding areshown in the Box. Upper GI bleed-ing is approximately 5 times as com-mon as lower GI bleeding, withmortality rates of 10%. Upper GIbleeding can be classified as non-variceal or variceal.

Methods allowing visualization andtreatment of lesions deep in thesmall intestine have recently beendeveloped. This has led to increasinguse of the term “mid-GI bleeding”(MGIB) (20). MGIB can presentwith occult bleeding (iron-deficiencyanemia and stools positive for occult

What are the symptoms and signsof acute GI bleeding? Can theyhelp localize the site of bleeding?GI bleeding can present with myriadsigns and symptoms (Table 1). Themost indolent forms may present assevere anemia. Manifestations includefatigue; dizziness; pallor; and rarelyend-organ complications, such as un-stable angina. Brisk bleeding from theupper or lower GI tract can presentwith more specific manifestations,such as visualized blood, syncope, orother symptoms of hypotension.

Distinguishing between upper andlower GI bleeding can help guide ini-tial diagnosis and therapy. Upper GIbleeding is more likely to cause nau-sea and dyspepsia, whereas lower GIbleeding is more likely to result in al-tered bowel habits, lower abdominalpain, or rectal discomfort. Hemate-mesis is exclusive to upper GI bleed-ing. Melena, which can be caused byas little as 50 mL of blood, usually re-sults from upper GI bleeding; however,


12. Longstreth GF. Epi-demiology of hospi-talization for acuteupper gastrointesti-nal hemorrhage: apopulation-basedstudy. Am J Gas-troenterol.1995;90:206-10.[PMID: 7847286]

13. Cook DJ, Fuller HD,Guyatt GH, et al. Riskfactors for gastroin-testinal bleeding incritically ill patients.Canadian CriticalCare Trials Group. NEngl J Med.1994;330:377-81.[PMID: 8284001]

14. Strate LL. Lower GIbleeding: epidemiol-ogy and diagnosis.Gastroenterol ClinNorth Am.2005;34:643-64.[PMID: 16303575]

15. Rondonotti E, Mar-mo R, Petracchini M,de Franchis R, Pen-nazio M. The Ameri-can Society for Gas-trointestinalEndoscopy (ASGE)diagnostic algorithmfor obscure gastroin-testinal bleeding:eight burning ques-tions from everydayclinical practice. DigLiver Dis.2013;45:179-85.[PMID: 22921043]

16. Pennazio M. Tech-niques in Gastrotin-testinal Endoscopy.2012;14:94-99.

17. Alain BB, Wang YJ.Cushing’s ulcer intraumatic brain in-jury. Chin J Trauma-tol. 2008;11:114-9.[PMID: 18377716]


Table 1. History and Physical Examination Elements of Patients with Acute Upper Gastrointestinal BleedingCategory Physical Examination Finding Notes

History Hematemesis Indicates bleeding proximal to the ligament of Treitz

Melena Should not be confused with dark stool of another etiology (test with guaiac to confirm the presence of blood)

Bloody diarrhea 10% of patients will have bleeding sources proximal to the ligament of Treitz

Presyncope and/or syncope Indicates loss of significant blood volume

Physical examination Hypotension (systolic blood pressure Indicates severe intravascular volume loss (≥50%), 90 mm Hg) assuming normal baseline systolic blood pressures

Tachycardia (≥120 beats/min) Indicates severe intravascular volume loss (≥50%)

Orthostatic changes in blood pressure Indicates loss of 20–25% of intravascular volume(≥10 mm Hg) or heart rate (≥30/min)

Nasogastric aspirate shows blood or Indicates an upper gastrointestinal source of bleedingcoffee ground–like material

Pallor Subjective/poor indicator without corroborative evidence

Perioral telangiectasias Suggestive of hereditary hemorrhagic telangiectasia syndrome

Skin abnormalities Stigmata of cirrhosis, pigmented lip lesions, acanthosis nigricans, vascular anomalies

From Bjorkman DJ, Eisen GM. Gastrointestinal bleeding, non-variceal upper. http://pier.acponline.org/physicians/diseases/d184/tables/d184-thp.html. (Dateaccessed 26 May 2009). PIER. Philadelphia: American College of Physicians; 2009.

with acute GI bleeding (e.g., outpa-tient vs. inpatient management) andprovide prognostic information. Pre-dictors of rebleeding risk and mor-tality in patients vary according tothe cause of bleeding; however, sev-eral factors portend a poorer prog-nosis. Adverse outcomes have alsobeen associated with chronic alco-holism and active cancer (22).

The International Consensus UpperGastrointestinal Bleeding Confer-ence Group advocates use of validat-ed risk-stratification tools to facilitatetriage of patients with acute upperGI bleeding by prognostication basedon risk for poor clinical outcomes(e.g., rebleeding, need for emergentsurgery, or mortality) (22). TheRockall scoring system (23) and theGlasgow–Blatchford Scale (24)(Table 2) are 2 commonly used risk-stratification systems that incorporateclinical, laboratory, and/or endoscopicparameters. The Glasgow–Blatchfordand the “Clinical” Rockall stratifica-tion systems are preendoscopic scoring systems and use clinical andlaboratory information to help pre-dict the need for hospitalization orendoscopic intervention. Additionalrisk stratification by incorporating

blood) or overt bleeding (melena,hematochezia). More frequently encountered sources of MGIB arevascular lesions (e.g., angioectasias,Dieulafoy lesions, arteriovenous malformations), ulcerating disorders(e.g., NSAID-enteropathy and in-flammatory bowel disease), and neo-plasms (e.g., neuroendocrine tumors,GI stromal tumors, lymphoma, andprimary or metastatic carcinomas).

The most common source of lowerGI bleeding is diverticulosis, reflectiveof the high prevalence of this condi-tion in the adult population; however,only a few patients with diverticulosishave diverticular bleeding. Othercommon causes of lower GI bleedinginclude colonic polyps, cancer, colitis(due to inflammatory bowel disease,infection, or ischemia), vascular lesions (angioectasia, radiation procti-tis), and anorectal sources (hemor-rhoids, anal fissures) (21).

Can risk for adverse outcomes bepredicted in patients with acuteGI bleeding? Which patients maybe evaluated as outpatients, andwhich require the emergencydepartment or hospitalization?Risk stratification can help deter-mine the disposition of a patient


18. Garcia-Tsao G,Sanyal AJ, Grace ND,Carey W; PracticeGuidelines Commit-tee of the AmericanAssociation for theStudy of Liver Dis-eases. Preventionand management ofgastroesophagealvarices and varicealhemorrhage in cir-rhosis. Hepatology.2007;46:922-38.[PMID: 17879356]

19. McGuire HH Jr. Bleed-ing colonic diverticu-la. A reappraisal ofnatural history andmanagement. AnnSurg. 1994;220:653-6.[PMID: 7979613]

20. Okazaki H, FujiwaraY, Sugimori S, et al.Prevalence of mid-gastrointestinalbleeding in patientswith acute overtgastrointestinalbleeding: multi-cen-ter experience with1,044 consecutivepatients. J Gastroen-terol. 2009;44:550-5.[PMID: 19360374]


Major Causes ofGastrointestinal Bleeding

InflammatoryPeptic ulcer diseaseEsophagitis or esophageal ulcerationDiaphragmatic hernia (Cameron

erosions)Inflammatory bowel disease

Benign and malignant neoplasmsPrimary gastrointestinal tract

neoplasms, at any siteMetastatic deposits in the

gastrointestinal tract, at any site

Vascular anomaliesGastroesophageal varicesAngioectasiasDieulafoy lesionGastric antral vascular ectasia (GAVE,

also known as watermelon stomach)Radiation proctopathy

Drug-inducedAspirinNonsteroidal anti-inflammatory drugs

MiscellaneousColonic diverticulosisPostpolypectomyMallory–Weiss tearMeckel diverticulum

From Rajan E, Ahlquist D. Gastro-intestinal bleeding. In: Dale DC,Federman DD, eds. ACP Medicine, 3rd ed.New York: WebMD; 2007:863.

Table 2. Risk Stratification Tools forUpper Gastrointestinal BleedingScoring System Points

Rockall Risk: Total score = sum of all categoryscores; risk category: high (>5), intermediate(3–4), low (0–2)

Age, y

>80 2

60–79 1

<60 0

Shock

SBP <100 mm Hg 2

SBP >100 mm Hg 1

No shock 0

Comorbid conditions

Renal failure, liver failure, 2widespread cancer

Cardiac failure, ischemic 1heart disease

No other comorbid conditions 0

Diagnosis on EGD

Upper gastrointestinal cancer 2

All other diagnoses 1

No lesion on EGD 0Stigmata of recent hemorrhage

Blood in upper gastrointestinal 2tract, adherent clot, visible vessel

No stigmata 0

Glasgow–Blatchford: Total score = sum of all category scores; a score of zero suggests low riskand the safety of outpatient care

Blood urea nitrogen (mmol/L)

>25 6

10–<25 4

8–<10 3

6.5–<8 2

<6.5 0

Hemoglobin (g/dL)

<10 in men and women 6

10–<12 in men 3

10–<12 in women/12–<13 in men 1

≥12 in women or ≥13 in men 0

SBP (mm Hg)

<90 3

90–99 2

100–109 1

>110 0

Other markers

Cardiac failure 2

Hepatic disease 2

Presentation with syncope 2

Presentation with melena 1

Pulse >100/min 1

EGD = esophagogastroduodenoscopy; SBP =systolic blood pressure.

some endoscopic information to theClinical Rockall scale leads to theFull Rockall scale, which helps pre-dict risk for adverse outcomes fromacute upper GI bleeding. Additionalstratification schemes for upper andlower GI bleeding have been pro-posed (25–28). Outpatient manage-ment may be appropriate for personswith low risk scores (Rockall score of0–2 or Glasgow–Blatchford score of0), whereas those with high riskscores require inpatient care (24,29–34). Admission to an intensivecare unit (ICU) should be consideredfor patients with evidence of brisk,active bleeding, such as hemodynam-ic instability (systolic blood pressure[BP] <100 or pulse >100, orthostatichypotension, evidence of shock), orother clinical parameters associatedwith high risk for rebleeding andmortality. To date, there are no wide-ly accepted validated scoring systemsspecific to acute lower GI bleeding.

What should the initial diagnosticevaluation for possible acute GIbleeding include?The initial evaluation of a patientwith acute GI bleeding should in-clude a focused history and physicalexamination. The history should in-clude associated signs and symptoms;use of NSAIDs, antiplatelet agents,anticoagulants, selective serotonin re-uptake inhibitors, and beta-blockers(because of issues of risk assessmentwith tachycardia and resuscitation im-plications); prior GI bleeding epi-sodes; and comorbid conditions. Vitalsigns on postural changes should beassessed, and stool should be exam-ined. Resting hypotension (systolicBP ≤90 mm Hg) or tachycardia(≥120 bpm) indicates severe intravas-cular volume loss (≥50%). An increaseof ≥30/min in the pulse or severelightheadedness when rising from asupine position also indicates signifi-cant volume loss (35, 36).

Immediate laboratory tests shouldinclude a complete blood count, pro-thrombin and partial thromboplastintimes, platelet count, blood type and


21. Zuckerman GR,Prakash C. Acutelower intestinalbleeding. Part II: eti-ology, therapy, andoutcomes. Gastroin-test Endosc.1999;49:228-38.[PMID: 9925703]

22. Barkun AN, BardouM, Kuipers EJ, et al.International con-sensus recommen-dations on the man-agement of patientswith nonvariceal up-per gastrointestinalbleeding. Ann InternMed. 2010;152:101-13. [PMID: 20083829]

23. Rockall TA, Logan RF,Devlin HB, NorthfieldTC. Selection of pa-tients for early dis-charge or outpatientcare after acute up-per gastrointestinalhaemorrhage. Na-tional Audit of AcuteUpper Gastrointesti-nal Haemorrhage.Lancet.1996;347:1138-40.[PMID: 8609747]

24. Blatchford O, MurrayWR, Blatchford M. Arisk score to predictneed for treatmentfor upper-gastroin-testinal haemor-rhage. Lancet.2000;356:1318-21.[PMID: 11073021]

25. Saltzman JR, TabakYP, Hyett BH, et al. Asimple risk score ac-curately predicts in-hospital mortality,length of stay, andcost in acute upperGI bleeding. Gas-trointest Endosc.2011;74:1215-24.[PMID: 21907980]

26. Strate LL, Orav EJ,Syngal S. Early pre-dictors of severity inacute lower intestin-al tract bleeding.Arch Intern Med.2003;163:838-43.[PMID: 12695275]

27. Strate LL, SaltzmanJR, Ookubo R, Mutin-ga ML, Syngal S. Vali-dation of a clinicalprediction rule forsevere acute lowerintestinal bleeding.Am J Gastroenterol.2005;100:1821-7.[PMID: 16086720]

28. Das A, Wong RC.Prediction of out-come of acute GIhemorrhage: a re-view of risk scoresand predictive mod-els. Gastrointest En-dosc. 2004;60:85-93.[PMID: 15229431]


requirements, rescue surgery, ormortality (41–43), urgent endoscopy(<12 hours) is indicated in patientswith suspected variceal bleeding (39,44, 45). It also provides valuable in-formation for appropriate patienttriage (see “Can risk for adverse out-comes be predicted in patients withacute GI bleeding?”).

What is the role of prokineticmedications before upperendoscopy in patients with acuteGI bleeding?Prokinetic medications (such aserythromycin and, to a lesser extent,metoclopramide) administered intravenously 20 to 120 minutes be-fore upper endoscopy has the poten-tial to facilitate clearance of bloodand clots from the stomach, thusimproving visualization in patientswith upper GI bleeding. However,they do not seem to alter importantclinical outcomes, including diag-nostic rates, need for transfusions orsurgery, or hospital length of stay.Routine use before upper endoscopyis not advocated and should be re-served for patients with red bloodhematemesis or blood in the naso-gastric aspirate (22, 39, 46).

What adjunctive tests can helpevaluate (and/or treat) patientswith acute GI bleeding without anidentified source on EGD orcolonoscopy?Patients without a demonstrablebleeding source on good-quality up-per endoscopy or colonoscopy have“obscure” GI bleeding. Small-bowelbarium radiography, “push” enterog-raphy, technetium-labeled red bloodcell scan, and angiography have his-torically been considered as subse-quent diagnostic tests, however withlow diagnostic yield. Wireless videocapsule endoscopy (VCE) has be-come most gastroenterologists’ test ofchoice for patients with obscure GIbleeding. The diagnostic yield rangesfrom 30–50% (7); however, no VCEsystems can provide hemostatic inter-ventions, and many institutions areunable to perform urgent inpatient

crossmatch in anticipation of bloodtransfusion, and a routine chemistrypanel. An increased ratio of bloodurea nitrogen to creatinine (>25:1 ifin mg/dL) suggests an upper GIsource of bleeding (35, 36).

Nasogastric or orogastric aspirationmay confirm an upper GI source ofbleeding and provide prognostic information on bleeding activity andseverity. However, the procedure isuncomfortable; yields false-negativeresults in approximately 15% of patients with active bleeding; and although it may lead to earlier per-formance of endoscopy, it has notbeen proven to improve clinical out-comes (30, 37). Some expert panelsrecommend it in selected patientsbut do not provide selection criteria(22, 38). Others state that it is notrequired in patients with suspectedupper GI bleeding for diagnosis,prognosis, visualization, or therapeu-tic effect (39). As a result, use of gas-tric aspiration should be dictated byinstitutional preference and practice.

When should a gastroenterologistbe consulted in the evaluation ofacute GI bleeding?A gastroenterologist should be con-sulted early to consider prompt en-doscopy for patients with GI bleedingand to facilitate patient triage. EGDor colonoscopy or both are the initialdiagnostic tests of choice for identifi-cation and treatment of specificbleeding lesions. Clinical presentationdictates which procedures are done.EGD is appropriate for patients withmelena and hematemesis and a subsetof patients with hematochezia result-ing from an upper GI source.

In upper GI bleeding, early en-doscopy (within 24 hours of admis-sion) has a greater impact than delayed endoscopy on some clinicaloutcomes, such as blood transfusionrequirements and hospital length ofstay (40). Although emergent en-doscopy (<6–8 hours) does not seem to further reduce rebleeding,hospital length of stay, transfusion


29. Chen IC, Hung MS,Chiu TF, Chen JC,Hsiao CT. Risk scor-ing systems to pre-dict need for clinicalintervention for pa-tients with nonva-riceal upper gas-trointestinal tractbleeding. Am JEmerg Med.2007;25:774-9.[PMID: 17870480]

30. Hwang JH, FisherDA, Ben-MenachemT, et al. The role ofendoscopy in themanagement ofacute non-varicealupper GI bleeding.Gastrointest Endosc.2012;75:1132-8.[PMID: 22624808]

31. Masaoka T, Suzuki H,Hori S, Aikawa N,Hibi T. Blatchfordscoring system is auseful scoring sys-tem for detectingpatients with uppergastrointestinalbleeding who donot need endoscop-ic intervention. JGastroenterol Hepa-tol. 2007;22:1404-8.[PMID: 17716345]

32. Srirajaskanthan R,Conn R, Bulwer C,Irving P. The Glas-gow Blatchford scor-ing system enablesaccurate risk stratifi-cation of patientswith upper gastroin-testinal haemor-rhage. Int J ClinPract. 2010;64:868-74. [PMID: 20337750]

33. Stanley AJ. Updateon risk scoring sys-tems for patientswith upper gastroin-testinal haemor-rhage [Editorial].World J Gastroen-terol. 2012;18:2739-44. [PMID: 22719181]

34. Stanley AJ, Ashley D,Dalton HR, et al. Out-patient manage-ment of patientswith low-risk upper-gastrointestinalhaemorrhage: multi-centre validationand prospectiveevaluation. Lancet.2009;373:42-7.[PMID: 19091393]

35. Cappell MS, FriedelD. Initial manage-ment of acute uppergastrointestinalbleeding: from initialevaluation up togastrointestinal en-doscopy. Med ClinNorth Am.2008;92:491-509.[PMID: 18387374]


approximately 50–70% and 25–70%of cases, respectively (16). Some cen-ters offer advanced deep enteroscopicprocedures, which are low-risk andenable visualization and therapeuticsdeep within the small intestine, es-sentially eliminating the need forhigh-risk intraoperative enteroscopyfor patients with suspected smallbowel pathology.

VCE. Another option is angiogra-phy, which allows for therapeutic in-tervention (embolization) if a lesionis localized; however, it requires ac-tive bleeding to be diagnostic. Com-puted tomography angiography orcomputed tomography/magnetic resonance enterography are alterna-tive diagnostic options that identifysources of obscure bleeding in

cardiovascular function, with the goalof normal, stable vital signs.

Observational studies and small con-trolled trials have suggested that bloodtransfusion to a target helmoglobinlevel of 9–10 mg/dL might actuallybe detrimental for some patients withhypovolemic anemia, including thosewith GI bleeding.

A recent randomized controlled trial (49)compared a “restrictive” vs. a “liberal” transfu-sion strategy in patients with acute upper GIbleeding. The threshold for transfusion of redblood cells was a hemoglobin value <7 g/dLin the restrictive transfusion group (targetrange, 7–9 g/dL) and 9 g/dL in the liberaltransfusion group (target range, 9–11 g/dL).Overall, the restrictive transfusion strategywas associated with a significantly greaterprobability of survival at 6 weeks, lower inci-dence of further bleeding, fewer adverseevents, fewer total units of blood transfused,and less need for rescue therapy. Subgroupanalysis revealed that the survival advan-tage of the restrictive strategy was limited toChild–Pugh class A and B cirrhosis (not classC). Patients with “massive exsanguinatingbleeding,” the acute coronary syndrome,symptomatic peripheral vasculopathy,stroke, transient ischemic attack, and lowerGI bleeding were excluded from the trial, and

What interventions should bestarted immediately in patientswith acute GI bleeding?Regardless of the source, initial man-agement of a patient with acute GIbleeding involves rapid diagnostic as-sessment and aggressive resuscitationwith isotonic fluids via 2 large-bore(18-gauge or larger) peripheral IVcatheters. Early intensive resuscita-tion has been associated with de-creased mortality from GI bleeding(48). The goal of resuscitation is tomaintain tissue perfusion until bleed-ing resolves spontaneously or is con-trolled by more definitive therapy.Smaller catheters (e.g., 20- or 22-gauge) do not allow rapid fluid ad-ministration. Longer central venouscatheters increase resistance to fluidflow, thereby slowing the rate atwhich fluids may be given. Patientswith emesis who are unable to pro-tect the airway from aspirationshould be intubated. Isotonic IV flu-ids, such as normal saline or lactatedRinger solution, should be given im-mediately, with the goal of replenish-ing intravascular volume as quickly asblood has been lost and as rapidly asis tolerated by the patient’s underlying

Presentation and Diagnosis... GI bleeding can present with myriad signs andsymptoms, ranging from asymptomatic to overt hematemesis or hematochezia,and can be due to a number of causes virtually anywhere along the GI tract. Ini-tial evaluation, including history and physical examination and routine laboratorytests, can help to narrow the differential diagnosis.


Treatment


36. Zuccaro G Jr. Man-agement of the adultpatient with acutelower gastrointestinalbleeding. AmericanCollege of Gastroen-terology. Practice Pa-rameters Committee.Am J Gastroenterol.1998;93:1202-8.[PMID: 9707037]

37. Huang ES, Karsan S,Kanwal F, et al. Im-pact of nasogastriclavage on outcomesin acute GI bleeding.Gastrointest Endosc.2011;74:971-80.[PMID: 21737077]

38. Barkun A, Bardou M,Marshall JK; Nonva-riceal Upper GIBleeding ConsensusConference Group.Consensus recom-mendations formanaging patientswith nonvariceal up-per gastrointestinalbleeding. Ann InternMed. 2003;139:843-57. [PMID: 14623622]

39. Laine L, Jensen DM.Management of pa-tients with ulcerbleeding. Am J Gas-troenterol.2012;107:345-60;quiz 361.[PMID: 22310222]

40. Lin HJ, Wang K,Perng CL, et al. Earlyor delayed en-doscopy for patientswith peptic ulcerbleeding. A prospec-tive randomizedstudy. J Clin Gas-troenterol.1996;22:267-71.[PMID: 8771420]

41. Sarin N, Monga N,Adams PC. Time toendoscopy and out-comes in upper gas-trointestinal bleed-ing. Can JGastroenterol.2009;23:489-93.[PMID: 19623332]

42. Tai CM, Huang SP,Wang HP, et al. High-risk ED patients withnonvariceal uppergastrointestinalhemorrhage under-going emergency orurgent endoscopy: aretrospective analy-sis. Am J EmergMed. 2007;25:273-8.[PMID: 17349900]

43. Targownik LE,Murthy S, Keyvani L,Leeson S. The role ofrapid endoscopy forhigh-risk patientswith acute nonva-riceal upper gas-trointestinal bleed-ing. Can JGastroenterol.2007;21:425-9.[PMID: 17637943]


vessels, or active bleeding are associ-ated with a high-risk for continuedor recurrent bleeding and warrantpharmacologic treatment and suchendoscopic interventions as thermalmethods; injection with vasoconstric-tive agents, such as epinephrine; andmechanical hemostasis with variousclips or sealants; or some combina-tion (53, 54). Strong evidence indicates that monotherapy with epinephrine injection alone is less effective than combination therapy(injection plus thermal therapy, orclips followed by injection) (53, 54).

In patients with peptic ulcer bleed-ing, PPI therapy (bolus followed byinfusion) before endoscopy has beenshown to decrease the likelihood ofhigh-risk stigmata on subsequentendoscopy and to reduce the likeli-hood of requiring an interventionduring endoscopy. It may also be as-sociated with shorter length of stayin the hospital. However, preendo-scopic PPI therapy has not beenshown to reduce mortality or theneed for rescue surgery or rebleedingrates when endoscopic therapy isconsistently delivered (55, 56) (Ap-pendix Figure, available at www.an-nals.org). Thus, preendoscopic PPIsshould be considered but should notdelay early endoscopy (within thefirst 24 h) or replace resuscitationbecause its effect on outcomes is minor at best (57). After endoscopy,hemodynamically stable patientswithout serious comorbid conditionswho are found to have low-risk ul-cers (e.g., clean-based, flat pigment-ed spot) can generally be dischargedearly on once-daily oral PPIs (58).

A meta-analysis of randomized trials ofhigh-dose IV PPI therapy (80 mg bolus then 8mg/h infusion for 72 h after endoscopic ther-apy) after successful endoscopic therapy forulcers in patients with high-risk stigmata (i.e.,active bleeding, nonbleeding visible vessel oradherent clot, or Forrest classification IA, IB,IIa, or IIb) (54, 59) concluded that mortality,need for surgery, and rebleeding is reduced(relative risk 0.41, 0.43, and 0.40, respectively)compared with endoscopy alone. Anothermeta-analysis also concluded that for

a restrictive transfusion strategy may not beapplicable to every patient.

Although coagulopathy has beenshown to be a risk factor for non-variceal upper GI bleeding, data arelimited and conflicting (48, 50, 51).Acknowledging the paucity of pub-lished data on the topic, the recentlypublished International ConsensusRecommendations advise that coagulopathy in patients receivinganticoagulants be treated, but thatmanagement should not delay thera-peutic endoscopy unless the INR issupratherapeutic (>2.5). Furthermore,this approach should not be general-ized to patients with cirrhosis be-cause the INR, at any threshold, cannot predict bleeding risk in thesepatients (22). In addition, a systemat-ic review of published studies report-ed a lack of data on which to recommend optimal platelet countsfor patients with GI bleeding. Basedon expert opinion, they recommend-ed targeting values of 50 000/µL inthe absence of platelet dysfunction,or 100 000/µL if functional plateletdysfunction is suspected (52).

How should acute upper GIbleeding due to peptic ulcerdisease be managed?The initial diagnostic approach tobleeding peptic ulcers is the same asfor other causes of GI bleeding. En-doscopy is nearly 100% specific and>90% sensitive; can be done at thebedside in the emergency depart-ment, endoscopy unit, or ICU; andallows biopsy to assess the cause ofthe ulcer (e.g., H. pylori infection oran ulcer within an adenocarcinoma).Endoscopists use the Forrest classifi-cation to describe peptic ulcers andto predict rebleeding risk associatedwith different ulcer appearances. Approximately 50% of ulcers have anappearance associated with a lowprobability of rebleeding (i.e., cleanulcer base or flat pigmented spot inthe ulcer base). Patients with theseulcers require only pharmacologictreatment. By contrast, ulcers withadherent clots, nonbleeding visible


44. Hearnshaw SA, Lo-gan RF, Lowe D, et al.Use of endoscopy formanagement ofacute upper gastroin-testinal bleeding inthe UK: results of anationwide audit.Gut. 2010;59:1022-9.[PMID: 20357318]

45. Lim CH, Ahmed MM.The optimal timingfor urgent en-doscopy in nonva-riceal upper gas-trointestinalbleeding [Letter]. En-doscopy.2011;43:1018.[PMID: 22057771]

46. Barkun AN, BardouM, Martel M, GralnekIM, Sung JJ. Prokinet-ics in acute upper GIbleeding: a meta-analysis. GastrointestEndosc.2010;72:1138-45.[PMID: 20970794]

47. Pasha SF. Diagnosticyield of deep en-teroscopy tech-niques for small-bowel bleeding andtumors. Tech GastroEndosc. 2012;14:100-105.

48. Baradarian R, Ramd-haney S, Chapala-madugu R, et al. Ear-ly intensiveresuscitation of pa-tients with uppergastrointestinalbleeding decreasesmortality. Am J Gas-troenterol.2004;99:619-22.[PMID: 15089891]

49. Villanueva C, Colo-mo A, Bosch A, et al.Transfusion strate-gies for acute uppergastrointestinalbleeding. N Engl JMed. 2013;368:11-21. [PMID: 23281973]

50. Wolf AT, Wasan SK,Saltzman JR. Impactof anticoagulationon rebleeding fol-lowing endoscopictherapy for nonva-riceal upper gas-trointestinal hemor-rhage. Am JGastroenterol.2007;102:290-6.[PMID: 17100959]

51. Shingina A, BarkunAN, Razzaghi A, Mar-tel M, Bardou M,Gralnek I; RUGBE In-vestigators. System-atic review: the pre-senting internationalnormalised ratio(INR) as a predictorof outcome in pa-tients with uppernonvariceal gastroin-testinal bleeding. Ali-ment PharmacolTher. 2011;33:1010-8.[PMID: 21385193]


monitored closely for adverse effectsof volume replacement, includingpulmonary edema. Targeting a hemoglobin level between 7–8 g/dLmight prevent excessive restitution ofblood volume and subsequent in-creases in portal pressure (62).

There are no definitive guidelines formanagement of coagulopathy andthrombocytopenia in patients withacute variceal hemorrhage. Reversalof coagulopathy requires administra-tion of fresh frozen plasma, althoughthe volume necessary may be prohibi-tive. Recombinant factor VIIa as analternative means of normalizing theprothrombin time has been proposed,but randomized, controlled trials havenot shown a consistent advantage(62). Because prothrombin time–INRis not a reliable indicator of the coagulation status in patients withcirrhosis, recommendations regardingmanagement of coagulopathy/throm-bocytopenia cannot be made on thebasis of available data (63).

Antibiotic prophylaxis reduces infec-tious complications and decreases rebleeding with acute variceal bleed-ing. A short-term quinolone course(e.g., up to 7 days of norfloxacin orciprofloxacin) is recommended. Datafrom a few studies suggest that cef-triaxione may prevent bacterial infec-tions better than norfloxacin, and IVceftriaxone (1 g/day) may be prefer-able in patients with advanced cir-rhosis and/or in regions with highprevalence of quinolone-resistant organisms (62, 64).

Variceal bleeding can often be con-trolled by both medical and endo-scopic means. Medical therapy primarily involves infusion of octreotide, a somatostatin analoguethat causes splanchnic vasoconstric-tion and reduced portal pressure.Numerous trials of octreotide haveyielded equivocal results, but it isrecommended in combination withendoscopic therapies to controlvariceal bleeding and reduce risk forrecurrence (18). Octreotide should

patients with ulcer bleeding, PPIs reduce re-bleeding and need for surgery or repeatedendoscopic procedures and improve mortal-ity in the highest–risk patients (59).

The optimal dose and route of acutePPI administration remain unclear,but patients with high-risk lesionshaving endoscopic therapy should re-ceive in-hospital therapy (IV, high-dose therapy remains the regimenwith the best evidence [22]) for 3days, mirroring the period duringwhich risk for rebleeding is greatest(60, 61). In the absence of compara-tive data, a once-daily oral PPI is rec-ommended after completion of 72hours of IV therapy (22, 30, 39). H-receptor blockers are not as effec-tive and should not replace PPIs foracute management of bleeding ulcers.

How should acute esophagealvariceal bleeding be treated?Acute bleeding from esophagealvarices is frequently life-threateningbecause it can be severe, difficult tocontrol, and rarely resolves sponta-neously. It usually occurs in patientswith end-stage liver disease. Eso-phageal varices are caused by signifi-cant portal hypertension; therefore,bleeding occurs under high pressureand is often brisk. In addition, pa-tients have synthetic liver dysfunc-tion and coagulopathy. Under thesecircumstances, acute managementrequires rapid and aggressive inter-ventions, including fluid resuscita-tion, bleeding control, and effortsaimed at reducing portal pressures.

Intravascular volume replacementshould occur as with any patientwith acute GI hemorrhage. Resusci-tation of patients with end-stage liver disease is often difficult due to hypoalbuminemia and extravasationof fluid from the intravascular space.Some experts advocate preferentialuse of blood products and albuminfor intravascular volume resuscita-tion, because crystalloid fluids delivera sodium load to patients who aretypically already total-body sodiumoverloaded and tend to eventuallyexacerbate ascites. Patients should be


52. Razzaghi A, BarkunAN. Platelet transfu-sion threshold in pa-tients with uppergastrointestinalbleeding: a system-atic review. J ClinGastroenterol.2012;46:482-6.[PMID: 22688143]

53. Barkun AN, MartelM, Toubouti Y,Rahme E, Bardou M.Endoscopic hemo-stasis in peptic ulcerbleeding for patientswith high-risk le-sions: a series ofmeta-analyses. Gas-trointest Endosc.2009;69:786-99.[PMID: 19152905]

54. Laine L, McQuaid KR.Endoscopic therapyfor bleeding ulcers:an evidence-basedapproach based onmeta-analyses ofrandomized con-trolled trials. ClinGastroenterol Hepa-tol. 2009;7:33-47.[PMID: 18986845]

55. Lau JY, Leung WK,Wu JC, et al.Omeprazole beforeendoscopy in pa-tients with gastroin-testinal bleeding. NEngl J Med.2007;356:1631-40.[PMID: 17442905]

56. Sreedharan A, Mar-tin J, Leontiadis GI,et al. Proton pumpinhibitor treatmentinitiated prior to en-doscopic diagnosisin upper gastroin-testinal bleeding.Cochrane DatabaseSyst Rev.2010:CD005415.[PMID: 20614440]


For patients with bleeding gastricvarices, initial clinical and phar-macologic management is similarto that for esophageal varices.However, decision making regard-ing definitive treatment options,particularly for isolated gastricvarices, is more complex. Band lig-ation, injection sclerotherapy, andTIPS have been reported and aresometimes used in clinical prac-tice. More recent therapeutic developments for treatment ofgastric varices include endoscopicinjection of biological glues and aradiologic procedure called bal-loon-occlusion retrograde transve-nous obliteration. Recently, self-expanding esophageal stentshave been used to staunch bleed-ing, particularly in refractory cases; however, further studies areneeded (63).

How should patients with acutelower GI bleeding from colonicdiverticulosis be treated?Initial management of colonic diver-ticular hemorrhage is the same asthat for other causes of GI bleeding,including fluid resuscitation, bloodtransfusion, and diagnostic testing.Colonoscopy should be done to tryto localize the bleeding, althoughwith brisk hemorrhage, localizationcan be difficult due to poor visuali-zation. The timing of the colonos-copy remains controversial; however,it is usually done within 12–24hours of presentation with a rapidcolonic preparation that seems to besafe in this setting (66). The mostimportant contribution of colonos-copy is usually exclusion of othercauses. In rare cases, it can also betherapeutic if a visible vessel or ad-herent clot is noted. More typically,however, bleeding occurs from mul-tiple sites and may be intermittent,so the opportunity for interventionis limited. Other studies that maylocalize the bleeding include nuclearimaging and angiography (19).

Bleeding resolves spontaneously inapproximately 75% of patients with

be given as a bolus dose of 50 µgfollowed by continuous IV infusionof 50 µg/hour for a total of 3 to 5days, although many experts cur-rently prefer a 5-day regimen (63).

Endoscopy should be done within12 hours in patients with known orsuspected varices who present withupper GI bleeding. There are 2 pri-mary endoscopic interventions:sclerotherapy and band ligation.Sclerotherapy involves injection ofa sclerosing agent (e.g., sodiummorrhuate or ethanolamine) intovarices. Band ligation involveswrapping elastic bands over thevarices in the distal esophagus.Clinical trials have shown thatcompared with sclerotherapy, band-ing has more prolonged benefit;lower rates of rebleeding, mortality,and complications (such asesophageal stricture formation, perforation, and mediastinitis); andreduced need for endoscopic treat-ments (65). Thus, band ligation hasbecome the acute treatment ofchoice in most institutions.

For patients with variceal bleedingrefractory to medical and endoscopictherapy, balloon tamponade (e.g.,with a Sengstaken–Blakemore tube)may be used as a temporizing meas-ure. The balloons should be inflatedfor no more than 12 hours. Moredefinitive treatments include trans-jugular intrahepatic portosystemicshunt (TIPS) placement with aPTFE stent and surgical interven-tions. In patients at high risk fortreatment failure (e.g., Child–Pughclass C score < 14 or class B with ac-tive bleeding), TIPS should beplaced within 72 hours (63). Salvage(rescue) TIPS placement is associat-ed with a high mortality rate. Thesetherapies are effective and often render surgical interventions unnec-essary. Surgical options include por-tosystemic shunting, esophagealtransaction, and liver transplantation.Mortality approaches 80% for pa-tients with continued bleeding fromesophageal varices.


57. Barkun AN. Shouldevery patient withsuspected upper GIbleeding receive aproton pump in-hibitor while await-ing endoscopy? [Edi-torial]. GastrointestEndosc.2008;67:1064-6.[PMID: 18513549]

58. Gralnek IM, BarkunAN, Bardou M. Man-agement of acutebleeding from apeptic ulcer. N Engl JMed. 2008;359:928-37. [PMID: 18753649]

59. Leontiadis GI, Shar-ma VK, Howden CW.Proton pump in-hibitor therapy forpeptic ulcer bleed-ing: Cochrane col-laboration meta-analysis ofrandomized con-trolled trials. MayoClin Proc.2007;82:286-96.[PMID: 17352364]

60. Sung JJ, Barkun A,Kuipers EJ, MössnerJ, et al. Intravenousesomeprazole forprevention of recur-rent peptic ulcerbleeding: a random-ized trial. Ann InternMed. 2009;150:455-64. [PMID: 19221370]

61. Lau JY, Chung SC,Leung JW, et al. Theevolution of stigma-ta of hemorrhage inbleeding peptic ul-cers: a sequentialendoscopic study.Endoscopy.1998;30:513-8.[PMID: 9746158]

62. Bari K, Garcia-Tsao G.Treatment of portalhypertension. WorldJ Gastroenterol.2012;18:1166-75.[PMID: 22468079]


How should therapy for acute GIbleeding be monitored?Continued monitoring is required toassess whether bleeding has stopped.Tachycardia may be an early warningof recurrent bleeding, followed soonby hypotension. Hemoglobin levelsshould be checked on a serial basis atleast every several hours initially toassess for stability after blood trans-fusions; levels that do not increase byapproximately 1 g/unit of transfusedpacked red blood cells may indicateongoing blood loss. Additional bloodtransfusions and diagnostic testing(e.g., repeated endoscopy) should beconsidered if the patient has evidenceof ongoing blood loss. Similarly,platelet count and coagulation shouldbe measured serially to assess theneed for repeated transfusions. Pa-tients requiring multiple transfusionsof red blood cells should be moni-tored for hypocalcemia.

When should a surgeon beconsulted for the management ofa patient with acute GI bleeding?Advances in medical and endoscopictherapies have resulted in fewer pa-tients requiring surgery for acute GIbleeding. However, surgical consulta-tion should be considered early in theevaluation and management of pa-tients with severe bleeding. Surgery isindicated when life-threateningbleeding continues, hemodynamiccompromise continues despite initialaggressive resuscitation, or bleedingcannot be stopped by endoscopic orangiographic means. The choice ofsurgery depends on bleeding locationand comorbid conditions. Localiza-tion of the bleeding site is critical forsurgical planning.

What follow-up outpatientevaluations are required inpatients who have experiencedacute GI bleeding?Specific guidelines for managementof patients after discharge other thanthose addressing secondary prophy-laxis are lacking. In general, decisionsregarding timing of postdischargeoutpatient visits and repeated blood

diverticular hemorrhage; recurrentbleeding occurs in 25–35% of pa-tients. If bleeding does not resolvespontaneously or with angiographicintervention, surgical resection isthe remaining therapeutic option.Approximately 20% of patientshospitalized for diverticular bleed-ing require surgery. Segmental/subtotal colectomy may be requiredif bleeding can be localized andendscopic therapy is not feasible. Itmay also be necessary if lower GIbleeding remains nonlocalizable,despite the risk for ongoing bleed-ing (due to failure to resect the ac-tual bleeding site) and may be associated with significant morbidi-ty and mortality (67).

What is the role of angiography?Angiographic interventions includelocal administration of vasopressinand embolization of the source. Va-sopressin is a potent vasoconstrictiveagent, and local instillation can con-trol bleeding in up to 80% of pa-tients; however, rebleeding occurs oncessation of the infusion in up to50% of patients. Therefore, vaso-pressin may be most useful as a tem-porizing measure. Although there areno absolute contraindications, vaso-pressin should be used with cautionin patients with coronary artery dis-ease or peripheral vascular disease be-cause of the risk for vasoconstrictionat sites other than the target lesion.Vasopressin can also cause cardiac ar-rhythmia and hyponatremia.

Embolization is a common inter-vention for control of active hemor-rhaging. It involves injection ofsealant materials, such as gel foam or polyvinyl alcohol, or mechanicaldevices, such as coils and temporaryballoons. Embolization is effective inup to 80% of cases. The primarycontraindication is poor collateralblood supply. The major complica-tions are intestinal ischemia and me-chanical complications of the arterialcannulation (such as localhematoma, arterial dissection, orpseudoaneurysm).


63. de Franchis R;Baveno V Faculty. Re-vising consensus inportal hypertension:report of the BavenoV consensus work-shop on methodolo-gy of diagnosis andtherapy in portal hy-pertension. J Hepa-tol. 2010;53:762-8.[PMID: 20638742]

64. Fernández J, Ruiz delArbol L, Gómez C, etal. Norfloxacin vsceftriaxone in theprophylaxis of infec-tions in patientswith advanced cir-rhosis and hemor-rhage. Gastroen-terology.2006;131:1049-56;quiz 1285.[PMID: 17030175]

65. Laine L, Cook D. En-doscopic ligationcompared with scle-rotherapy for treat-ment of esophagealvariceal bleeding. Ameta-analysis. AnnIntern Med.1995;123:280-7.[PMID: 7611595]

66. Jensen DM, Machi-cado GA, Jutabha R,Kovacs TO. Urgentcolonoscopy for thediagnosis and treat-ment of severe di-verticular hemor-rhage. N Engl J Med.2000;342:78-82.[PMID: 10631275]

67. Stollman N, RaskinJB. Diverticular dis-ease of the colon.Lancet.2004;363:631-9.[PMID: 14987890]

68. Banerjee S, Cash BD,Dominitz JA, et al.The role of en-doscopy in the man-agement of patientswith peptic ulcerdisease. GastrointestEndosc. 2010;71:663-8. [PMID: 20363407]

69. Malfertheiner P,Megraud F, O’MorainCA, et al. Manage-ment of Helicobacterpylori infection—-the Maastricht IV/Florence ConsensusReport. Gut.2012;61:646-64.[PMID: 22491499]


diverges on this topic, reporting in-adequate outcomes and cost-effectiveness to support use of thisstrategy for every patient. Societyguidelines advocate an individualizedapproach (68). Surveillance en-doscopy is most appropriate for pa-tients with suspicious ulcers on initialendoscopy in whom initial biopsieswere negative or were not done, per-sistent symptoms despite a course ofantisecretory medications, absence ofa defined cause, or patients from demographic regions with high inci-dence of gastric carcinoma. If H. py-lori infection has not been assessedvia prior gastric mucosal biopsies,serology, or stool antigen evaluation,gastric mucosal biopsies for H. pyloritesting should be considered. Litera-ture pertaining to detailed manage-ment of patients with H. pylori isavailable (69). Bleeding sources thatmay require additional endoscopictherapy include gastric antral vascularectasia (GAVE, or “watermelonstomach”) and radiation proctopathy.

How long should antisecretorytherapy be continued?Antisecretory therapy for patientswith small (<1 cm), uncomplicatedulcers can usually be discontinued after 4–6 weeks in the absence of on-going ulcer-related symptoms. Main-tenance therapy should be consideredfor higher-risk patients (complicated,recurrent, or large ulcers). In patientswith H. pylori–associated ulcers, ther-apy is often continued until eradica-tion of infection is confirmed, or indefinitely if attempts at H. pylorieradication have failed. Instructionsfor patients with alternative causes ofbleeding are made on a case-by-casebasis.

What instructions do patientsrequire after acute GI bleeding?Specific guidelines for managementof patients after discharge are lackingand should be individualized. Patientsshould be educated regarding thesigns and symptoms of recurrentbleeding and the anticipated benefitand duration of targeted therapies.

work are based on risk-stratification.Early outpatient evaluation is appro-priate for patients with high-risk endoscopic findings, significant comorbid conditions, recent severebleeding episode, and those in needof decision-making pertaining to an-tiplatelet or anticoagulant medicationuse. Conversely, the remaining low-risk patients usually do not requireearly outpatient evaluation. Patientsshould be advised to report symp-toms or signs of recurrent bleeding.With the exception of select clinicalscenarios (see below), empirical re-peated or “surveillance” endoscopy isnot required for most patients partic-ularly if the bleeding source is associ-ated with a low-risk for recurrentbleeding (e.g., Mallory–Weiss tears,postpolypectomy bleeding).

Patients with cirrhosis who survivevariceal hemorrhage but do not havefurther therapy are at significant riskfor rebleeding (60%) and death (ap-proximately 33%) within 1–2 years ofthe initial bleeding episode. Theseoutcomes are improved using com-bined endoscopic variceal band liga-tion and beta-blockers (which are often given once there has been noevidence of hemorrhage for at least24 hours, barring any contraindica-tions). Patients who have shunt sur-gery or TIPS do not require furtherpreventive measures.The AmericanAssociation for the Study of LiverDisease suggests repeating endoscopyat 7- to 14-day intervals until eradica-tion of esophageal varices has beenachieved (usually requires 2–4 ses-sions); and once varices are eradicat-ed, repeating endoscopy every 3–6months initially to evaluate for recur-rence and need for repeated band lig-ation (18). Eventually, the interval canbe increased to 6–12 months.

For patients that recover from gastriculcer bleeding, surveillance en-doscopy 6 to 12 weeks after the lastbleeding episode to exclude a non-healing ulcer (which raises concernfor cancer) is common in clinicalpractice. However, the literature


70. Kanwal F, Barkun A,Gralnek IM, et al.Measuring quality ofcare in patients withnonvariceal uppergastrointestinalhemorrhage: devel-opment of an explic-it quality indicatorset. Am J Gastroen-terol. 2010;105:1710-8. [PMID: 20686458]


recommended (70). Similarly, patientswith a bleeding ulcer while receivinglow-dose aspirin therapy for second-ary prevention of established cardio-vascular disease should resume therapy as soon as possible after ces-sation of bleeding (within 1 week,ideally as early as 3–5 days). Data areless robust for management of aspirintherapy after acute bleeding when as-pirin is being used for primary pre-vention of cardiovascular events, andthe risk and benefits must be weighedon an individual basis.

Patients with bleeding ulcers associat-ed with H. pylori infection should beencouraged to complete a full courseof H. pylori therapy and to have test-ing, such as a urea breath test or H. pylori stool antigen assay, to con-firm successful eradication of the bac-teria. For patients suspected of havingNSAID-related ulcer bleeding, dis-continuation is preferable but some-times not feasible. If a patient mustresume NSAIDs, a cyclooxygenase-2–selective NSAID at the lowest ef-fective dose plus a daily PPI is

Treatment... Treatment of GI bleeding is highly dependent on the cause and severity.Initial evaluation and management in all cases should include history and physical ex-amination, with simultaneous stabilization interventions, such as placement of IV ac-cess and IV fluid resuscitation. Emergent endoscopy (e.g., within 6 hours) is rarely indicated, but urgent endoscopy (e.g., within 12 h) in selected circumstances can beconsidered, and is mandatory if variceal bleeding is suspected. Administering a PPI iswarranted in patients with suspected peptic ulcer disease but should not delay en-doscopy. Transfusion practices should target a hemoglobin threshold of 7–8 g/dL. Out-patient follow-up should be individualized based on the cause of bleeding and the es-timated risk for rebleeding.


colonoscopy have been unrevealing,the American Society of Gastroen-terology recommends early evalua-tion of the small bowel by VCE orangiography, with CT angiography,CT enteroscopy, and deep entero-scopy as secondary considerations,depending on availability and ex-pertise at the institution.

What measures do stakeholdersuse to evaluate the quality of carefor patients with acute GIbleeding?The Agency for Healthcare Re-search and Quality has identifiedmortality rate as the primary qualityindicator for care of patients withGI bleeding. An expert panel inde-pendently developed a list of 26quality indicators for non-varicealupper GI bleeding, categorized aspreendoscopic, endoscopic, andpostendoscopic factors (70).

What do professional organiz-ations recommend with regard to the prevention, diagnosis andtreatment of acute GI bleeding?The International Consensus UpperGastrointestinal Bleeding ConferenceGroup recommends early risk stratifi-cation and early diagnostic endoscopyin most patients with upper GIbleeding.

The American Society of Gastroen-terology recommends earlycolonoscopy for diagnosis of acutelower GI bleeding, with angiogra-phy and tagged red blood cell scan-ning in patients with active bleedingand nondiagnostic colonoscopies. Ifsurgical intervention is contemplat-ed, preoperative localization ofbleeding is desirable.

In patients with obscure acute GIbleeding in whom EGD and

PracticeImprovement


Inthe

C linic

Tool KitIn the Clinic

AcuteGastrointestinalBleeding

PIER Modulehttp://pier.acponline.org/physicians/diseases/d184/d184.htmlPIER module on gastrointestinal (GI) bleeding from the

American College of Physicians.

Patient Informationhttp://pier.acponline.org/physicians/diseases/d184/d184-pi.htmlPatient information that appears on the next page for

duplication and distribution to patients.www.nlm.nih.gov/medlineplus/gastrointestinalbleeding.htmlwww.nlm.nih.gov/medlineplus/tutorials/uppergiendoscopy/

htm/index.htmwww.nlm.nih.gov/medlineplus/spanish/tutorials/ upper

giendoscopy/htm/index.htmResources related to GI bleeding from the National

Institutes of Health’s MedlinePLUS, including aninteractive tutorial on upper GI endoscopy, in Englishand Spanish.

www.gi.org/physician-resources/brochures/Patient brochure from the American College of

Gastroenterology (ACG) on understanding ulcers,common pain medications, and GI Bleeding.

Clinical Guidelineshttp://annals.org/article.aspx?articleid=745521International consensus recommendations for managing

patients with nonvariceal upper gastrointestinal bleeding,published in Annals of Internal Medicine in 2010.

www.sign.ac.uk/guidelines/fulltext/105/index.htmlClinical guideline on the management of acute upper and

lower GI bleeding from the Scottish IntercollegiateGuidelines Network in 2008.

http://circ.ahajournals.org/content/118/18/1894.fullConsensus document on reducing the GI risks of

antiplatelet therapy and NSAID use from the AmericanCollege of Cardiology Foundation (ACCF), ACG, andAmerican Heart Association (AHA) in 2008.

http://content.onlinejacc.org/article.aspx?articleid=1143980Consensus document on concomitant use of proton pump

inhibitors and thienopyridines, an update from theACCF, ACG, and AHA in 2010.

Diagnostic Tests and Criteriahttp://pier.acponline.org/physicians/diseases/d184/tables/

d184-tlab.htmlList of laboratory and other studies for acute upper GI

bleeding from PIER.http://pier.acponline.org/physicians/diseases/d184/tables/

d184-t1.htmlGlasgow–Blatchford screening tool to assess the likelihood

that a patient with an acute upper GI bleeding will needmedical intervention.

http://pier.acponline.org/physicians/diseases/d184/tables/d184-t2.html

Rockall scoring system for identifying patients at risk foradverse outcome after acute upper GI bleeding.

6 August 2013Annals of Internal MedicineIn the ClinicITC2-14© 2013 American College of Physicians


In the ClinicAnnals of Internal Medicine

Pati

ent

Info

rmat

ion

THINGS YOU SHOULDKNOW ABOUTGASTROINTESTINALBLEEDING

What is gastrointestinal (GI) bleeding?• Bleeding in the digestive tract.• The upper digestive tract includes the esophagus,

stomach, and upper portion of the small intestine.• The lower digestive tract includes lower portion of

the small intestine, large intestine (also called thecolon), and anus.

• Most causes of bleeding are curable or controllable,but some may be life-threatening if left untreated.

What causes bleeding in the digestivetract?Causes of bleeding in the upper digestive tract include:

• Peptic ulcers from Helicobacter pylori infections orlong-term use of nonsteroidal anti-inflammatorydrugs, such as aspirin and ibuprofen.

• Varices (enlarged veins) in the lower esophagus thatrupture and bleed.

• Tears in the lining of the esophagus or inflammationin the lining of the stomach (gastritis) or esophagus(esophagitis).

• Cancerous or noncancerous (benign) growths.

Causes of bleeding in the lower digestive tract include:

• Diverticular disease (diverticula are irregular pouch-es that develop in the colon wall).

• Colitis (inflammation of the colon) or angiodysplasia(abnormalities in blood vessels of the intestine).

• Hemorrhoids (ruptured veins in the anus or rectum)or fissures (anal cuts or tears).

• Cancerous or noncancerous (benign) growths.

What are the signs and symptoms?• Vomiting bright-red blood or vomit that looks like

coffee grounds indicates bleeding in upper digestivetract.

• Black or tarry stool or stool that contains dark orbright red blood indicates bleeding in upper or lowerdigestive tract.

• Chronic bleeding (light bleeding that continues for along time or starts and stops) may lead to fatigue,lethargy, and shortness of breath over time.

• Acute bleeding (heavy bleeding) may lead to dizzi-ness or faintness, shortness of breath, abdominalpain, and shock.

How is it treated?• Medical imaging techniques, such as endoscopy or

angiography, may be used to locate the source of thebleeding inside of the digestive tract and to stop thebleeding.

• Surgery may be needed if these interventions do notwork.

• Your doctor will try to prevent future bleeding bytreating the condition that is causing the bleeding.

For More Informationhttp://digestive.niddk.nih.gov/ddiseases/pubs/bleeding/http://digestive.niddk.nih.gov/Spanish/pubs/bleeding/index.aspxInformation on bleeding in the digestive tract from the National

Institute of Diabetes and Digestive and Kidney Diseases(NIDDK), in English and Spanish.

http://digestive.niddk.nih.gov/ddiseases/pubs/lowergi/Lower_GI_Series_T_508.pdfInformation on the lower GI x-rays ordered to help diagnose

problems of the large intestine from the NIDDK.


CME Questions

6 August 2013Annals of Internal MedicineIn the ClinicITC2-16© 2013 American College of Physicians

Questions are largely from the ACP’s Medical Knowledge Self-Assessment Program (MKSAP, accessed at http://www.acponline.org/products_services/mksap/15/?pr31). Go to www.annals.org/intheclinic/

to complete the quiz and earn up to 1.5 CME credits, or to purchase the complete MKSAP program.

1. A 58-year-old man is evaluated in theemergency department for painlessbright-red blood per rectum that began 3 hours ago. The bleeding wasaccompanied by syncope. He has ahistory of rheumatoid arthritis. Hiscurrent medications are adalimumab,methotrexate, and ibuprofen.

On physical examination, temperature is37.2°C (99.0°F), blood pressure is 88/58 mm Hg, pulse rate is 132/min, and respiration rate is 24/min. Abdominal examination is normal. Rectalexamination discloses bright-red blood inthe rectal vault. Nasogastric tubeaspirate shows no evidence of blood orcoffee-ground material.

Laboratory studies reveal a hemoglobinlevel of 7.3 g/dL (73 g/L).

Emergency intravenous fluidresuscitation is begun.

Which of the following is the mostappropriate diagnostic test to performnext?

A. ColonoscopyB. Tagged red blood cell scanC. Upper endoscopyD. Video capsule endoscopy

2. A 46-year-old man is evaluated for a 3-week history of painless occasionalbright-red rectal bleeding. He has nofatigue, lightheadedness, weight loss, orabdominal pain. His stools are frequentlyfirm, occasionally hard, and there is nochange in the frequency or consistencyof bowel movements. He has never beenscreened for colorectal cancer.

On physical examination, temperature is37.2°C (98.9°F), blood pressure is 132/78 mm Hg, and pulse rate is 84/min.Digital rectal examination yields a stoolsample that is positive for occult blood;the examination is otherwise normal.Anoscopy reveals a few internalhemorrhoids without active bleeding.Laboratory studies show a bloodhemoglobin level of 14 g/dL (140 g/L).

Which of the following is the mostappropriate management of this patient?

A. Banding of hemorrhoidsB. ColonoscopyC. Fiber supplementation without

further evaluationD. Home fecal occult blood testing

3. A 60-year-old man hospitalized foradvanced cirrhosis complicated by ascitesand encephalopathy is evaluated formassive hematemesis and hypotension.The patient’s medications arespironolactone, furosemide, andlactulose.

On physical examination, temperature is 35.6°C (96°F), blood pressure is 80/50 mm Hg, pulse rate is 146/min, andrespiration rate is 20/min. The patienthas just vomited red blood and haslarge-volume ascites; the stool is brownand positive for occult blood. Laboratorystudies show hemoglobin of 9 g/dL (90 g/L), platelet count of 60 000/µL (60 × 109/L), and INR of 3.

In addition to rapid volume resuscitation,which of the following is the mostappropriate management of this patient?

A. ArteriographyB. EsophagogastroduodenoscopyC. Intravenous nadololD. Mesocaval shuntE. Transjugular intrahepatic

portosystemic shunt

4. A 78-year-old woman is evaluated in thehospital after being admitted 5 days agofor a 2-week history of abdominal painand nausea. She has also had black, tarrystools for the past 36 hours. On day 1,esophagogastroduodenoscopy showed aclean-based bleeding gastric ulcer thatwas positive for Helicobacter pyloriinfection; the ulcer was treated withinjection therapy and coagulationtherapy with probe cautery, and proton-pump inhibitor therapy was initiated. Thebleeding did not stop, and esophago-gastroduodenoscopy was repeated on day3 with endoclip therapy. The bleedingcontinued, and the patient has receivedeight 8 of packed erythrocytes.

On physical examination on day 5,temperature is 37.2°C (99.0°F), bloodpressure is 95/50 mm Hg, pulse rate is103/min, and respiratory rate is 16/min.Rectal examination reveals melanoticstool. Laboratory studies revealhemoglobin of 10.8 g/dL (108 g/L); allother tests, including coagulationparameters, are normal.

Which of the following is the mostappropriate next step in the managementof this patient?

A. Bleeding scanB. Helicobacter pylori eradication

therapyC. Intravenous octreotideD. Surgery


Documents

In the Clinic Acute Gastrointestinal Bleeding In theClinicunmhospitalist.pbworks.com/w/file/fetch/68209086/0000605-201308060... · In the Clinic Acute Gastrointestinal Bleeding Prevention