Residency · Residency · Internal Medicine

Acute Upper GI Bleeding: Resuscitation and Endoscopic Management

Epidemiology and Etiology

Acute upper GI bleeding occurs with an annual incidence of 50 to 150 per 100,000 and carries a mortality rate of 2-10%, with higher mortality observed in inpatients with comorbidities. Peptic ulcer disease is the most common cause, accounting for 35-50% of cases, with gastric ulcers more common than duodenal; H. pylori infection and NSAID use are the primary risk factors. Variceal bleeding accounts for 10-20% of UGIB with a per-episode mortality of 15-25%; in patients with known cirrhosis, variceal bleeding should be assumed until proven otherwise. Erosive esophagitis and gastritis are often self-limited. Mallory-Weiss tears are mucosal lacerations at the gastroesophageal junction from retching and are usually self-limited. Dieulafoy lesions are large submucosal arteries that cause intermittent massive bleeding and are difficult to identify endoscopically. Gastric and esophageal malignancy are additional causes. Vascular causes include angiodysplasia (more common in CKD and aortic stenosis, known as Heyde syndrome) and aortoenteric fistula, which should always be considered in patients with prior aortic grafts. Cameron erosions occur in large hiatal hernias.

Clinical Presentation

Hematemesis presents as vomiting of bright red blood or coffee-ground emesis. Melena (black, tarry stools) requires as little as 50 to 100 mL of blood in the upper GI tract. Hematochezia, or bright red blood per rectum, may occur with brisk upper GI bleeding exceeding 1 liter. Hemodynamic instability manifests as tachycardia, orthostatic hypotension, or frank shock. An elevated BUN out of proportion to creatinine results from absorption of blood protein in the GI tract.

Initial Resuscitation

Airway Protection

Intubation should be considered for massive hematemesis, altered mental status, or inability to protect the airway, and is particularly important before endoscopy in actively vomiting patients.

Vascular Access and Fluid Resuscitation

Two large-bore (18-gauge or larger) peripheral IVs should be established immediately. Crystalloid resuscitation is initiated, though excessive volume should be avoided as it may worsen variceal bleeding. Type and crossmatch should be sent promptly.

Blood Transfusion -- Restrictive Strategy

A restrictive transfusion threshold (hemoglobin below 7 g/dL) is preferred for most patients. The landmark Villanueva trial (NEJM 2013) demonstrated that a restrictive strategy improved survival, reduced rebleeding, and reduced adverse events compared to a liberal threshold (hemoglobin below 9) in UGIB. Exceptions include patients with symptomatic anemia, active acute coronary syndrome, or hemodynamic instability despite resuscitation, who may warrant transfusion at a higher threshold. In variceal bleeding, over-transfusion increases portal pressure and worsens bleeding. Platelets should be transfused to a target above 50,000/microL for active bleeding. FFP or coagulation factors are used to correct INR if above 1.5 in active bleeding, though endoscopy should not be delayed for correction. Aggressive correction in cirrhosis should be avoided given the balanced coagulopathy.

Medications to Hold/Review

Anticoagulants and antiplatelet agents should be held with reversal assessed based on severity. NSAIDs should be discontinued. Aspirin for secondary cardiovascular prevention should generally be continued in most cases, with restart within 3 to 7 days after endoscopic hemostasis.

Risk Stratification

Glasgow-Blatchford Score (GBS)

The GBS is a pre-endoscopic score predicting the need for intervention. Its components include BUN, hemoglobin, systolic blood pressure, heart rate, melena, syncope, hepatic disease, and cardiac failure. A GBS of zero identifies very low-risk patients who can be considered for outpatient management with early endoscopy. A GBS of one or higher requires inpatient management.

Rockall Score

The Rockall score exists in pre-endoscopic (clinical) and full (post-endoscopic) versions, incorporating age, comorbidities, hemodynamic status, endoscopic findings, and stigmata of recent hemorrhage. It predicts rebleeding and mortality.

AIMS65 Score

The AIMS65 score uses five variables: albumin below 3.0, INR above 1.5, altered mental status, systolic blood pressure below 90, and age above 65. It predicts inpatient mortality.

Pre-Endoscopic Management

Proton Pump Inhibitors

An IV PPI bolus (for example, pantoprazole 80 mg IV followed by 8 mg/hr infusion) before endoscopy reduces high-risk stigmata at endoscopy and the need for endoscopic therapy, though it does not clearly reduce mortality, rebleeding, or need for surgery. After endoscopy, high-dose IV PPI (80 mg bolus plus 8 mg/hr for 72 hours) is continued only for high-risk ulcers that received endoscopic therapy; intermittent dosing (40 mg IV twice daily) is likely equivalent in most cases. Patients are then transitioned to oral PPI (omeprazole 40 mg twice daily for 2 weeks, then daily).

Prokinetic Agents

IV erythromycin (250 mg IV given 30-60 minutes before endoscopy) improves visualization by promoting gastric emptying, reduces the need for repeat endoscopy, and should be considered in patients with significant blood or clot burden. Metoclopramide is an alternative with less supporting evidence.

Variceal-Specific Pre-Endoscopic Treatment

IV octreotide (50 mcg bolus, then 50 mcg/hr infusion for 3-5 days) reduces portal pressure. IV ceftriaxone 1 g daily for 7 days reduces spontaneous bacterial peritonitis, infections, rebleeding, and mortality in cirrhotic patients with UGIB. PPI should not be used for variceal bleeding as there is no evidence of benefit.

Endoscopy

Timing

Endoscopy should be performed within 24 hours of presentation for most patients after resuscitation. Within 12 hours is appropriate for high-risk features including hemodynamic instability, bloody nasogastric tube output, or high GBS. Emergent endoscopy (under 12 hours) is indicated for hemodynamic instability despite resuscitation or suspected variceal bleeding. Very early endoscopy (under 6 hours) has not been clearly shown to be beneficial and may increase mortality if performed before adequate resuscitation.

Endoscopic Findings and Classification

Forrest Classification (Peptic Ulcers)

The Forrest classification categorizes ulcers by bleeding stigmata and rebleeding risk. Class Ia (spurting hemorrhage) and Ib (oozing hemorrhage) carry rebleeding risks of 55-100%. Class IIa (non-bleeding visible vessel) has a 43% rebleeding risk. Class IIb (adherent clot) carries 22% risk. Class IIc (flat pigmented spot) has 10% risk, and Class III (clean base ulcer) has only 5% rebleeding risk.

Endoscopic Hemostasis Techniques

Injection therapy with epinephrine (1:10,000) reduces bleeding temporarily but must always be combined with a second modality, as monotherapy carries a high rebleeding rate. Thermal coagulation includes bipolar electrocautery and heater probe. Mechanical therapy uses hemoclips, including through-the-scope and over-the-scope clips. Combination therapy (injection plus thermal or injection plus clips) is superior to monotherapy for high-risk lesions.

Variceal Bleeding Management

For esophageal varices, endoscopic variceal ligation (banding) is first-line therapy, with sclerotherapy as second-line. For gastric varices, cyanoacrylate (tissue glue) injection is preferred since EVL is less effective. Balloon tamponade (Sengstaken-Blakemore or Minnesota tube) is a temporizing measure for uncontrolled variceal bleeding as a bridge to TIPS or repeat endoscopy; it carries a high complication rate and should not be maintained beyond 24 hours. TIPS (transjugular intrahepatic portosystemic shunt) is indicated for refractory variceal bleeding or as early TIPS (within 72 hours) for high-risk patients, defined as Child-Pugh B with active bleeding at endoscopy or Child-Pugh C with score of 13 or less.

Post-Endoscopic Management

High-Risk Ulcers (Forrest Ia-IIb)

Management includes IV PPI infusion for 72 hours followed by transition to oral twice-daily dosing for 2 weeks and then once daily. Repeat endoscopy is indicated only if rebleeding occurs; routine second-look endoscopy is not recommended. ICU monitoring for 24-48 hours is appropriate for high-risk stigmata.

Low-Risk Ulcers (Forrest IIc-III)

Oral PPI with early feeding is appropriate, and patients can be discharged early, even the same day if GBS equals zero.

H. pylori Testing and Eradication

All patients with peptic ulcer bleeding should be tested for H. pylori. Testing during the index endoscopy (CLO test or biopsy) is standard, though the false-negative rate is higher during acute bleeding. If negative during acute bleeding, retesting with a urea breath test or stool antigen should be performed 4-8 weeks later. Eradication dramatically reduces ulcer recurrence from approximately 60% to less than 5%.

Rebleeding Management

Rebleeding occurs in 10-20% of patients. Repeat endoscopy is first-line for rebleeding. Interventional radiology with angiographic embolization is the next step for endoscopic failure. Surgery is reserved for life-threatening hemorrhage unresponsive to endoscopic and radiologic intervention.

Secondary Prophylaxis for Variceal Bleeding

The combination of a non-selective beta-blocker (nadolol, propranolol, or carvedilol) with EVL is superior to either alone. Heart rate should be reduced to 55-60 bpm or by 25% from baseline. TIPS should be considered if rebleeding occurs despite combined medical and endoscopic therapy. Hepatocellular carcinoma screening and liver transplant evaluation are appropriate in eligible patients.

<image> A clinical algorithm flowchart for acute upper GI bleeding management. Start with "Hematemesis, Melena, or Hematochezia with Suspected UGIB" leading to resuscitation (IV access, fluids, transfusion threshold Hb <7). Branch into risk stratification (GBS = 0 consider outpatient vs. GBS >=1 admit). Show pre-endoscopic interventions (IV PPI, erythromycin, octreotide if variceal suspected). Then endoscopy timing decision and post-endoscopic management based on Forrest classification. </image>

<image> An illustrated guide to the Forrest classification of peptic ulcer bleeding. Show six ulcer cross-sections in a grid format: Ia (spurting vessel), Ib (oozing), IIa (visible vessel), IIb (adherent clot), IIc (pigmented spot), and III (clean base). For each, include the rebleeding risk percentage and whether endoscopic therapy is indicated. Use red color gradients to indicate bleeding severity from active (bright red) to clean base (gray). </image>

<image> A comparison diagram of endoscopic hemostasis techniques. Four panels showing: (1) injection therapy with epinephrine around a bleeding vessel, (2) thermal coagulation with bipolar probe, (3) hemoclip application, and (4) variceal band ligation. Each panel includes a brief description of technique, indication, and key pearl. Highlight that combination therapy is superior to monotherapy for ulcer bleeding. </image>

Clinical Pearls

A restrictive transfusion strategy (hemoglobin below 7 g/dL threshold) improves outcomes in UGIB, and over-transfusion in cirrhotics increases portal pressure and worsens variceal bleeding. A GBS of zero identifies very low-risk patients who may not need inpatient admission or urgent endoscopy, and this score can help avoid unnecessary hospitalization. IV erythromycin before endoscopy improves visualization and is underutilized, particularly in patients with large-volume hematemesis. In any cirrhotic with GI bleeding, IV octreotide and ceftriaxone should be started before endoscopy, as antibiotics reduce mortality in this population. Epinephrine injection alone is not adequate endoscopic therapy and must always be combined with a second modality such as clips or thermal coagulation. H. pylori testing in peptic ulcer bleeding followed by treatment if positive is the single most effective intervention to prevent recurrence. Hematochezia can originate from an upper GI source when bleeding is brisk (exceeding 1 liter), so bright red blood per rectum should not automatically be attributed to a lower GI source, especially when hemodynamic instability is present.

References

• Laine L, et al. ACG Clinical Guideline: Upper Gastrointestinal and Ulcer Bleeding. Am J Gastroenterol. 2021.
• Villanueva C, et al. Transfusion Strategies for Acute Upper GI Bleeding. NEJM. 2013.
• Barkun AN, et al. International Consensus Recommendations on the Management of Patients with Nonvariceal Upper GI Bleeding. Annals of Internal Medicine. 2010 (updated 2019).
• Garcia-Tsao G, et al. Portal Hypertensive Bleeding in Cirrhosis: AASLD Practice Guidance. Hepatology. 2017.
• Forrest JA, et al. Endoscopy in GI Bleeding. Lancet. 1974.
• Stanley AJ, et al. Comparison of Risk Scoring Systems for Patients with Upper GI Bleeding (International GBS Study). BMJ. 2017.
• Blatchford O, et al. A Risk Score to Predict Need for Treatment for Upper GI Hemorrhage. Lancet. 2000.

Read this lecture as Markdown