II. Definitions
- Upper Gastrointestinal Bleeding
- Bleeding proximal to the ligament of treitz (suspends the distal duodenum) at the duodenojejunal flexure
III. Epidemiology
- Incidence: 48 to 160 cases per 100,000
- Accounts for 300,000 to 400,000 hospitalizations in U.S. yearly
IV. Risk factors
- History of prior Upper Gastrointestinal Bleeding (Relative Risk 13.5)
- Anticoagulant use (Relative Risk 12.7)
-
Helicobacter Pylori infection (present in 64% of cases)
- Upper Gastrointestinal Bleeding Odds Ratio: 1.7
- Adheres to gastric epithelium, predisposing underlying mucosa to injury by toxins
- Together, Helicobacter Pylori and/or NSAID use accounts for 80% of Upper GI Bleeding cases
- Antiplatelet Agents
- Aspirin alone increases risk of GI Bleeding by 37%
- Clopidogrel (Plavix) increases the GI Bleeding risk
- Concurrent Proton Pump Inhibitor lowers the GI Bleeding Odds Ratio
- PPI and Aspirin OR Clopidogrel monotherapy: 0.2 Odds Ratio
- PPI and Aspirin AND Clopidogrel: 0.36 Odds Ratio
-
NSAID use (most common cause)
- See NSAID Gastrointestinal Adverse Effects for Relative Risk of specific NSAIDs
- Upper Gastrointestinal Bleeding Odds Ratio: 4.8 to 5.8
- Upper Gastrointestinal Bleeding Odds Ratio: 6.1 if Helicobacter Pylori positive in addition to NSAID use
- Celecoxib (Celebrex) and NaproxenSodium (Naprosyn) are the lowest risk of Upper Gastrointestinal Bleeding
- Elderly (especially over age 70 years) with Relative Risk 5.6
- Male gender (twice as common as women)
- Acid suppression therapy does not reduce bleeding risk
- End Stage Renal Disease (esp. first year of Dialysis)
- Selective Serotonin Reuptake Inhibitors
V. Causes: Adults
- Adults with acute massive GI Bleeding
- Peptic Ulcer Disease (62%)
- Associated with NSAID use or Helicobacter Pylori infection
- Gastric Ulcer (up to 55%)
- Duodenal Ulcer (30-38%)
- Esophageal Varices (6-10%)
- Gastritis or Duodenitis (5-10%)
- Esophagitis or esophageal ulcer (5-10%)
- Mallory-Weiss tear (3-7%)
- Gastrointestinal malignancy (1-4%)
- Arteriovenous Malformation (10%)
- Dieulafoy's Lesion (1%)
- Artery at gastric fundus may bleed heavily
- Difficult to identify on endoscopy
- Gastric antral vascular ectasia (0.5 to 2%)
- Longitudinal erythematous stripes on gastric mucosa
- Known as Watermelon Stomach
- Aortoenteric Fistula
- Associated with prior aortic surgery or grafting
- Conditions associated with Angiodysplasia of Stomach or duodenum
- Peptic Ulcer Disease (62%)
- Chronic intermittent GI Bleeding
- Gastritis (18 to 35%)
- Esophagitis (18 to 35%)
- Gastric Ulcer (18 to 21%)
- Duodenal Ulcer (3 to 15%)
- Angiodysplasia (5 to 23%)
- Gastric Cancer
- Most commonly missed upper GI sources
- Large Hiatal Hernia erosions
- Arteriovenous Malformation
- Peptic Ulcer Disease
VI. Causes: Children
- Esophagitis
- Gastritis
- Peptic Ulcer Disease
- Esophageal Varices
- Mallory-Weiss Tear
VII. History
VIII. Symptoms
- Lightheadedness, Dizziness or Syncope
- Abdominal Pain
-
Hematemesis with Coffee-ground Emesis
- Distinguish from the frothy, smaller volume blood seen in Hemoptysis
- Hematemesis is associated with higher mortality rates
- Blood per Rectum
- Black tarry stools or Melena
- Bright Red Blood Per Rectum as Maroon Stool or Hematochezia (if Upper GI Bleeding is brisk)
IX. Signs
-
Vital Signs
- Do not be reassured by normal Vital Signs
- In contrast, Abnormal Vital Signs mandate emergent management
- Tachycardia
- Heart Rate may be initially normal despite significant Hemorrhage
- Hypotension or Orthostasis
- Late finding, requires 20% loss of Blood Volume
- Shock Index (SI = HR/SBP)
- Shock Index >0.8 is considered abnormal (hemodynamic instability)
- Do not be reassured by normal Vital Signs
- Cardiovascular Exam
- Signs of poor perfusion (e.g. thready pulses, cool and pale extremities, lethargy)
-
Respiratory Exam
- Exclude Epistaxis source
- Exclude Hemoptysis source
- Abdominal exam
- Observe for signs of Cirrhosis and Esophageal Varices risk (e.g. Jaundice, Ascites, Hepatomegaly, Spider Nevi)
- Hyperactive Bowel Sounds
- Hematemesis
- Large volume, coffee ground or brownish Emesis
- Contrast with Hemoptysis (cough with smaller volume, bright red, frothy Sputum)
- Rebound Tenderness, involuntary guarding (or other signs of peritonitis)
- Consider viscus perforation with Acute Abdomen
- Nasogastric aspirate or lavage bloody
- False Negative Rate is at least 15% and likely higher (negative aspirate does not exclude GI Bleed)
- Fresh blood suggests persistant bleeding and suggests a high risk lesion
- Lavage offers higher Test Sensitivity but increases risk of aspiration and is controversial
- Lavage does not alter mortality, surgery or transfusion requirements
- Lavage positive for coffee ground material or bright red blood confirms upper gastrointestinal source
- However a negative lavage does not exclude Upper Gastrointestinal Bleeding
- Gastric Lavage may also help clear the Stomach of blood prior to endoscopy and improve the evaluation
- Allows for monitoring of ongoing upper gastrointestinal GI Bleeding
X. Labs
-
Complete Blood Count
- Baseline Hemoglobin (trails bleeding by 8 to 24 hours)
- Obtain serial Hemoglobins every 6 to 8 hours after initial Hemoglobin
- Blood Type and Cross-match
-
Coagulation Factors
- Prothrombin Time (INR)
- Partial Thromboplastin Time (PTT)
- Platelet Count
- Consider Fibrinogen level
- Consider Viscoelastic Assay (if available)
- Comprehensive metabolic panel
- Liver Function Tests
- Renal Functions tests
- Blood Urea Nitrogen (BUN)
- Serum Creatinine
- BUN to Creatinine ratio
- Does not reliably distinguish upper GI source
- However some studies show ratio >36 has Test Sensitivity 90%
XI. Imaging
- CT Angiography (CTA)
- See CT Angiography in Gastrointestinal Bleeding
- Consider in patients unable to undergo upper endoscopy with rapid bleeding
- CTA may localize sites of Massive Hemorrhage, allowing for management by Interventional Radiology or surgery
XII. Evaluation
XIII. Management: Stabilization of the Actively Bleeding Patient
- See Acute Gastrointestinal Bleeding Management
- ABC Management
- Initial Supportive care
- Oxygen
- Two large bore IVs (14-18 gauge)
- Larger bore catheters have highest flow rates (250 ml/min for 14G, 100 ml/min for 18G)
- IO Access or Central Line access if unable to obtain timely, adequate large bore peripheral access
- Fluid Resuscitation and Transfusion
- See blood replacement below for additional factors (e.g. Massive Transfusion Protocol)
- May use initial crystalloid fluid (e.g. 1 Liter) to briefly temporize until blood is available
- However, crystalloid further dilutes Coagulation Factors
- Hemorrhagic Shock replacement with Blood Transfusion (pRBC) as soon as available
- Start with 1 to 2 units of unmatched pRBC ( O- Blood) until cross matched blood is available
- Target a mean arterial pressure >65 mmHg
- Increase transfusion rates for Shock Index >1.0, hemodynamic instability, rapid Hemorrhage
- Avoid Vasopressors if possible (aside from peri-intubation as below)
- Limit use to hemodynamic instability refractory to Massive Transfusion Protocol
- Consider Endotracheal Intubation (for airway protection)
- See Rapid Sequence Intubation
- Endotracheal Intubation Preoxygenation
- Have two large bore suctions (one with open tubing, one yanker) on and ready for vomit or blood
- Avoid Peri-Intubation Hypotension (high risk for peri-intubation Cardiac Arrest)
- Start Vasopressors (e.g. peripheral Norepinephrine) or have Push Dose Pressors drawn up and ready
- Monitor Vital Signs including Blood Pressure, Heart Rate and Oxygen Saturation
- Consider Arterial Line
- Avoid Hypothermia (worsens Coagulopathy)
- Correct Acidosis
- Modify mechanical Ventilator settings if intubated (e.g. Respiratory Rates)
- Intensive Care Unit admission for significant bleeding or hemodynamically unstable
- Replace blood and Coagulation Factors as needed
- See Gastrointestinal Bleeding Management
- See Massive Transfusion Protocol
- Unstable Patients: pRBC transfusion
- In acute Hemorrhage, true Hemoglobin level is delayed 8 to 24 hours
- Start Blood Transfusion in hemodynamically Unstable Patients with active Hemorrhage regardless of initial Hemoglobin
- Massive Transfusion Protocol
- See Massive Transfusion Protocol for indications
- Replace pRBC with FFP and Platelets in a balanced, 1:1:1 ratio
- Viscoelastic Assay may be used to direct a more specific protocol for the given patient
- Calcium is also replaced for every 6 units pRBC transfused
- Avoid over-shooting Hemorrhage volume (risk of worsening GI Hemorrhage, esp. Variceal Bleeding)
- Stable Patients: pRBC Transfusion indications
- pRBC transfusion for Hemoglobin <7 g/dl (or <8 g/dl with significant comorbidity)
- Stable Gastrointestinal Bleeding may target a Hemoglobin of 7 g/dl (restrictive transfusion)
- Slow transfusion rates (esp. in CHF, Renal Failure) to prevent Transfusion Associated Circulatory Overload (TACO)
- Platelets
- Platelet Transfusion for Platelet Count <50,000
- Platelets may also be increased with IV Desmopressin in CKD, antiplatelet medications, Thrombocytopenia
- Anticoagulant Reversal
- See Emergent Reversal of Anticoagulation
- Correct INR >2.5 for patients on Warfarin before endoscopy (e.g. PCC4 or FFP and Vitamin K)
- Cirrhosis
- Risk of Hypocalcemia and Hypomagnesemia (associated with citrate in transfusion products)
- Coagulopathy may benefit from Vitamin K Clotting Factor replacement in Massive Hemorrhage
- Consider Vitamin K 10 mg IV
- Consider Cryoprecipitate 10 units/kg IV
- Other agents
- Tranexamic Acid 1g IV has been used
- However, evidence from HALT-IT trial found no benefit and increased adverse events
- (2020) Lancet 395(10241):1927-36 [PubMed]
- Tranexamic Acid 1g IV has been used
- Upper endoscopy urgently to emergently
- Notify GI or surgery on patient presentation to ready for upper endoscopy
- Performed in first 24 hours after fluid Resuscitation and stabilization
- See Upper Endoscopy Evaluation of GI Bleeding
- See Glasgow-Blatchford Bleeding Score
- May forego upper endoscopy if score <1
- See protocols below
- Erythromycin infusion immediately prior to endoscopy
- Prokinetic that increases endoscopy Test Sensitivity for Gastrointestinal Bleeding site
- Medications
- Intravenous Proton Pump Inhibitor (e.g. Pantoprazole IV) - see below
- Variceal Bleeding specific management (see below)
- Prokinetic Agents (e.g. Erythromycin IV before endoscopy)
- Controversial, but may improve endoscopy efficacy
XIV. Management: Specific Interventions
-
Esophageal Varices
- See Bleeding Esophageal Varices
- Upper Endoscopy emergently
- Endoscopic ligation
- Variceal banding
- Sclerotherapy
- Obturation
- Esophageal stenting (refractory cases)
- Vasoactive Agents
- Octreotide 50 mcg IV bolus, followed by 50 mcg/hour IV infusion
- Octreotide is a splanchnic Vasoconstrictor that decreases Variceal Bleeding and decreases transfusion needs
- Non-selective Beta Blocker (e.g. Propranolol, Nadalol, Timolol)
- Prophylactic Antibiotics (e.g. Norfloxacin, Ciprofloxacin, Ceftriaxone)
- Ceftriaxone 2 g IV every 24 hours for 7 days (may be transitioned to oral Ciprofloxacin)
- Consider Esophageal Balloon Tamponade (e.g. Linton Tube, Sengstaken-Blakemore Tube)
- Indicated for stabilization until endoscopy
- Non-variceal persistent bleeding
- Upper endoscopy (see protocol above)
- See Risk stratification below
- Transcatheter Arterial Embolization (via Angiography) Indications for Persistent Bleeding
- Endoscopy not available
- Biliary bleeding
- Pancreatic bleeding
- Gastric Ulcer
- Mallory-Weiss Tear
- General Surgery Indications
- Acute Abdomen (peritonitis)
- Continued massive bleeding without known source
- Failed medical therapy, endoscopy and angiography
- Lower risk of rebleeding compared with transcatheter ablation
- Upper endoscopy (see protocol above)
XV. Management: Acid Reduction
-
Proton Pump Inhibitor (preferred)
- Start Intravenous Proton Pump Inhibitor in all Upper Gastrointestinal Bleeding cases
- Better outcomes with use, although mixed results (see below)
- Dosing
- Typical dosing (even in Massive Hemorrhage)
- Pantoprazole (Protonix) 40 mg IV every 12 hours
- Continue as twice daily dosing (oral or IV) for first 2 weeks after endoscopic therapy
- Massive Hemorrhage
- Pantoprazole (Protonix) 80 mg IV, then 8 mg/hour
- Continuous infusion has largely been replaced in most cases by intermittent bolus
- Typical dosing (even in Massive Hemorrhage)
- Effects
- Full protection (achlorhydria) is delayed for the 5-7 days required to deactivate the proton pump
- Consider H2 Blocker initially (see below)
- Not proven in-vivo to aid clotting
- Omeprazole may heal Peptic Ulcer if near-achlorhydria
- No proven benefit in mortality
- Cochrane study suggested PPI benefit in specific measures when given in endoscopy suite
- May reduce Incidence of re-bleeding (NNT 15)
- May reduce Incidence of Gastrointestinal Bleeding requiring surgery (NNT 30)
- Peptic Ulcer related bleeding appears to benefit from PPI in China and Japan (but possible associated worse outcomes in U.S.)
- Leontiadis (2006) Cochrane Database Syst Rev (1): CD002094 [PubMed]
- No benefit or harm seen when PPI given at Upper GI Bleeding presentation
- Full protection (achlorhydria) is delayed for the 5-7 days required to deactivate the proton pump
- References
- Newman in Herbert (2014) EM:Rap 14(1): 4
- Daneshmend (1992) BMJ 304:143-47 [PubMed]
- Sreedharan (2010) Cochrane Database Syst Rev (2): CD005415 [PubMed]
- Sung (2009) Ann Intern Med 150(7): 455-64 [PubMed]
- Start Intravenous Proton Pump Inhibitor in all Upper Gastrointestinal Bleeding cases
-
H2 Blocker
- Use is controversial in Acute Gastrointestinal Hemorrhage
- Consider starting initially concurrently with Proton Pump Inhibitor to bridge the delay in full PPI activity
-
Ranitidine
- Intermittent: 50 mg IV or IM every 6-8 hours
- Continuous: 6.25 mg/hour IV
XVI. Management: Very low risk patients
- Indications
- Hemodynamically stable with normal lab testing
- No evidence of significant bleeding in last 48 hours
- Nasogastric Tube aspirate without blood
- Upper GI Bleeding Score (Glasgow-Blatchford Bleeding Score) <1
- Protocol
- Home with follow-up within days
- General measures as above
XVII. Management: Low risk patients
- Indications
- Hemodynamically stable within 1 hour of Resuscitation
- Minimal Blood Products required (2 PRBC or less)
- No evidence of active bleeding
- Nasogastric Tube aspirate without blood
- No active comorbid medical conditions
- Protocol
- Consider for rapid protocol
- Immediate Upper Endoscopy Evaluation of GI Bleeding
- Discharge to home if low-risk endoscopy results
- Admit if rapid protocol not available
- Follow moderate risk patient protocol below
- General measures as above
- Consider for rapid protocol
XVIII. Management: Moderate risk patients
- Indications
- Tachycardia persists despite Resuscitation
- Blood Products required >2 PRBC
- Active comorbid condition
- Protocol
- General measures as above
- Admit to regular medical bed
- Upper endoscopy when patient stabilized (<24 hours)
- See Upper Endoscopy Evaluation of GI Bleeding
- Disposition based on Upper Endoscopy results
- Low risk endoscopy: Observe for 24 hours
- Moderate risk endoscopy: Observe for 48-72 hours
- High risk endoscopy
- Initially observe in ICU for at least 24 hours
- Observe in hospital for 72 hours total or more
XIX. Management: High risk patients
- Indications
- Active ongoing bleeding (persistent drop in Hemoglobin despite transfusion)
- Hypotension or other hemodynamic instability persists despite Resuscitation
- Severe active comorbid condition exascerbation
- Liver disease exascerbation
- Endotracheal Intubation for airway protection
- Protocol
- General measures as above
- Admit to Intensive Care unit for first 24 hours
- Observe in hospital for 48 to 72 hours or more
- Urgent upper endoscopy when stabilized
- See Upper Endoscopy Evaluation of GI Bleeding
- Consider repeat routine recheck upper endoscopy in 24 hours for high risk lesions
- Consider arteriography if source not evident
XX. Management: Refractory and Recurrent Bleeding
- Indications
- Persistent or recurrent bleeding despite EGD
- See Upper Endoscopy Evaluation of GI Bleeding
- Procedures
- Consider arteriography with embolization
- First-line study
- Equivalent efficacy to surgical intervention
- Surgical Intervention
- Gastroduodenotomy
- Vagotomy
- Peptic Ulcer resection
- Consider arteriography with embolization
XXI. Management: Disposition
- See prevention measures below
- See Helicobacter Pylori management
-
Proton Pump Inhibitor once daily for 4-8 weeks
- Increase to twice daily for first 2 weeks for high risk lesions
- Repeat Upper Endoscopy
- Perform at 8-12 weeks after initial endoscopy
- Indications
- Gastric Ulcer (Confirm healing and exclude cancer)
- Severe Esophagitis (exclude Barrett's Esophagus)
-
Iron Deficiency Anemia
- Transfused blood will keep Hemoglobin increased only 2-3 weeks
-
Ferrous Sulfate 325 mg daily to three times daily
- Continue until Serum Ferritin and Hemoglobin return to normal
XXII. Management: Cause Specific Approaches
- Erosive Disorders (Esophagitis, Gastritis or Duodenitis)
- Good prognosis and therapeutic endoscopy is not required
- Early discharge on Proton Pump Inhibitor for 8 weeks
-
Mallory Weiss Tear
- Most spontaneously heal
- However, bleeding related mortality approaches that of Peptic Ulcer Bleeding
XXIII. Management: Restarting Anticoagulants and Antiplatelet agents after Upper GI Bleed
-
Warfarin
- Restart 7-15 days after bleeding (7 days if high thromboembolic risk)
- Non-Vitamin K Antagonist Oral Anticoagulant (e.g DOACs, Apixaban )
-
Aspirin
- May restart in 3 days (immediately at 0 days, if low bleeding risk )
- Dual Antiplatelet agents
- First start Aspirin AND
- After 5-7 days, may add back Clopidogrel or other antiplatelet agent in high risk patients (e.g. cardiac stents)
XXIV. Prevention
- See Helicobacter Pylori management
- See GI Prophylaxis
- Avoid NSAIDs
- Consider longterm Proton Pump Inhibitor if Aspirin or Clopidogrel (Plavix) cannot be stopped
- See Clopidogrel (Plavix) for potential Proton Pump Inhibitor interactions
- Proton Pump Inhibitors reduce antiplatelet efficacy (increased cardiovascular events)
- Tobacco Cessation
- Avoid Alcohol
XXV. Prognosis: Outcomes
- Mortality: 2-15% overall (often related to comorbidity)
- In hospital mortality: 13%
- Duodenal Ulcer: 3.7%
- Higher risk of erosions into larger vessels
- Gastric Ulcer: 2.1%
- Course
- Bleeding stops and does not recur: 85% (<2% Mortality)
- Re-bleeding after initially stopped: 15% (10% Mortality)
- Continued active bleed: 5% (30% Mortality)
XXVI. Prognosis: Predictors
- Bleeding characteristic predictors of poor outcome
- See Upper GI Bleeding Score
- Emesis or nasogastric aspirate contains red blood
- Low initial Hematocrit
- Coagulopathy (Low Platelets or high INR)
- Comorbid condition predictors of poor outcome
- Active Coronary Artery Disease
- Congestive Heart Failure
- Active lung disease
- Renal Failure
- Sepsis
- Metastatic cancer
- Advanced liver disease
- Advanced age
XXVII. References
- Azim (2023) Crit Dec Emerg Med 37(4): 23-9
- Spangler, Swadron, Mason and Herbert (2016) EM:Rap C3, p. 1-11
- Gupta (1993) Med Clin North Am 77(5):973-92 [PubMed]
- Fallah (2000) Med Clin North Am 84(5):1183-208 [PubMed]
- Longstreth (1995) Am J Gastroenterol 90(2):206-10 [PubMed]
- Peter (1999) Emerg Med Clin North Am 17(1):239-61 [PubMed]
- Stanley (2019) BMJ 364:1536 +PMID:30910853 [PubMed]
- Terdiman (1998) Postgrad Med 103(6):43-64 [PubMed]
- Wang (2010) Ann Surg 215(1):51-8 [PubMed]
- Wilkins (2012) Am Fam Physician 85(5): 469-76 [PubMed]
- Wilkins (2020) Am Fam Physician 101(5):294-300 [PubMed]
- Zuckerman (2000) Gastroenterology 118:201-21 [PubMed]