Management of Lower GI Bleeding (LGIB) - Evidence-Based

Definition and Epidemiology

• Sleisenger and Fordtran's GI and Liver Disease, p. 320

Common Causes

• Sleisenger and Fordtran's, Table 20.8

Step 1 - Resuscitation

• Two large-bore IV lines, IV crystalloid fluid resuscitation
• Blood transfusion: The 2023 ACG Guideline recommends a restrictive transfusion threshold (transfuse at Hb <7 g/dL; Hb <8 g/dL in patients with cardiovascular disease) - identical to the evidence base in upper GI bleeding
• Coagulopathy correction: For patients on vitamin K antagonists with life-threatening LGIB, the ACG 2023 guideline recommends reversal (4-factor PCC preferred over FFP). For patients on direct oral anticoagulants (DOACs) with life-threatening LGIB, specific reversal agents (andexanet alfa for Xa inhibitors, idarucizumab for dabigatran) are recommended
• ICU monitoring for hemodynamically unstable patients

Step 2 - Risk Stratification

Oakland Score (validated tool, ACG 2023 recommended)

• Oakland score ≤8: Minor self-limited bleeding - may be discharged safely from the ED and undergo outpatient colonoscopy within 2 weeks
• Oakland score >8: Requires inpatient management
• Mulholland and Greenfield's Surgery, p. 3213

Strate Clinical Prediction Score (for severe LGI bleeding)

• Aspirin use

• 2 comorbid illnesses

• Heart rate ≥100/min
• Nontender abdomen
• Rectal bleeding within first 4 hours
• Syncope
• Systolic BP ≤115 mmHg

• Sleisenger and Fordtran's, Table 20.9

Step 3 - Diagnostic and Management Algorithm

A) Hemodynamically Unstable - Ongoing Bleeding

1. Urgent CT Angiography (CTA) - first-line investigation
   • Detects bleeding rates as low as 0.3 mL/min
   • Sensitivity 85%, specificity 92% in active bleeders
   • These patients cannot tolerate bowel prep or colonoscopy
2. If CTA positive: Proceed to visceral angiography + superselective embolization (achieves hemostasis in ~80%)
3. If CTA negative: Perform EGD (upper GI source in ~15% of unstable patients with hematochezia)

B) Hemodynamically Stable - Major Bleeding (Oakland >8)

1. Colonoscopy after resuscitation and bowel prep (PEG purge)
   • Bowel prep improves diagnostic yield
   • For cirrhotic patients, recent melena/hematemesis, or PUD history: panendoscopy (EGD + colonoscopy) recommended first
   • Urgent colonoscopy diagnoses the source in 70-95% of cases
2. Timing of colonoscopy - ACG 2023 Update:
   • Two systematic reviews showed colonoscopy within 24 hours improved bleeding localization and reduced hospital stay
   • However, a recent multicenter RCT showed no difference in rebleeding rates between urgent (<24 hrs) and elective colonoscopy
   • ACG 2023 recommends NON-URGENT colonoscopy for most patients (i.e., within the admission but not necessarily within 24 hours), as urgent colonoscopy has not been shown to improve outcomes like rebleeding

   Key evidence: ACG 2023 LGIB Guideline (Sengupta N et al., Am J Gastroenterol 2023; PMID 36735555)

3. If colonoscopy negative: CT angiography for a possible missed bleeding source or mid-GI source

C) Minor, Self-Limited Bleeding (Oakland ≤8)

• Outpatient colonoscopy within 2 weeks
• No need for inpatient admission

Step 4 - Endoscopic Hemostasis

• Rebleeding rate: 0% vs 53% in historical controls without hemostasis
• Emergency hemicolectomy: 0% vs 35% in controls
• In a later series of 63 patients: rebleeding rate 4.8% with endoscopic treatment

• Sleisenger and Fordtran's, p. 323

Step 5 - Angiographic Management

• Active bleeding on CTA in hemodynamically unstable patients
• Failure of endoscopic hemostasis
• Patient cannot tolerate colonoscopy

• First-line angiographic therapy - preferred over vasopressin
• Initial hemostasis success: 85-90%
• Reduced bowel infarction risk with superselective technique
• Detects bleeding at 0.5-1.0 mL/min (requires active extravasation)
• Yamada's Textbook of Gastroenterology, p. 2833

Step 6 - Surgical Management

1. Bleeding unresponsive to endoscopic hemostasis AND angiographic embolization
2. Recurrent bleeding refractory to endoscopic or angiographic therapy
3. Massive hemorrhage unresponsive to resuscitation

• Preoperative localization is essential - allows partial (segmental) colectomy
• Subtotal colectomy reserved for when the bleeding site cannot be localized
  • Recurrent bleeding rate after subtotal colectomy: <4%
  • Perioperative mortality: 2-6%
• Blind subtotal colectomy should be avoided when possible
• Mulholland and Greenfield's Surgery, p. 3231

Step 7 - Specific Cause-Based Management

Diverticular Bleeding

• Stops spontaneously in 75-80% of cases
• Risk factors for recurrence: age >70, NSAID/aspirin/anticoagulant use, pancolonic disease
• Recurrence incidence: 3,626/100,000 person-years; risk after second bleed: 11% annually
• Treatment: hemoclips ± epinephrine injection (preferred in thin-walled right colon)

Hemorrhoids

• Anoscopy should be the initial diagnostic study in ALL LGIB patients
• Mild bleeding: dietary fiber + lifestyle modification (avoids straining)
• Rubber band ligation: effective in >90% of patients
• Severe/grade III-IV: operative hemorrhoidectomy

Ischemic Colitis

• Usually resolves spontaneously in a few days; antibiotic therapy generally not required
• ~10% have a focal ulcer with major stigmata on colonoscopy - treat with epinephrine + hemoclip
• Large-vessel mesenteric ischemia: worse outcomes, higher rates of surgery/perforation/death

Post-Polypectomy Bleeding

• Occurs in ~1% of polypectomies, typically 5-7 days post-procedure
• 50-75% require transfusion; usually self-limited
• Treatment: hemoclip ± epinephrine injection (avoids thermal injury at polypectomy site)

Angioectasia

• APC is the treatment of choice
• If active extravasation: selective arterial embolization

IBD-related Bleeding

• Most cases managed medically
• Surgery (colectomy) for refractory cases

Step 8 - Antithrombotic Medications After Discharge

• In patients on aspirin for secondary prevention: resume aspirin as soon as hemostasis is achieved (do not stop)
• In patients on dual antiplatelet therapy (aspirin + P2Y12 inhibitor):
  • High cardiothrombotic risk: continue at least aspirin; restart P2Y12 inhibitor within 5 days
  • Lower cardiothrombotic risk: continue aspirin; restart P2Y12 within 5 days of hemostasis
• Anticoagulants: Restart after 7-10 days in most cases after hemostasis is confirmed; decision based on thrombotic vs. bleeding risk

Summary Flow

LGIB presentation
│
├─ Oakland ≤8 → Outpatient colonoscopy (2 weeks)
│
├─ Stable + Oakland >8 → Resuscitate → Colonoscopy (non-urgent per ACG 2023)
│   ├─ High-risk stigmata found → Endoscopic hemostasis (clips ± epinephrine)
│   └─ No source on colonoscopy → CTA → angiography/embolization if positive
│
└─ Unstable + ongoing bleeding → Urgent CTA
    ├─ CTA positive → Angiography + superselective embolization
    ├─ CTA negative → EGD (rule out UGI source)
    └─ Failure of all above → Surgery (segmental or subtotal colectomy)

Key Evidence References

Management of Upper Gastrointestinal Bleeding (UGIB) - Evidence-Based

Definition and Causes

• Sabiston Textbook of Surgery, Table 98.1

Management Algorithm

Figure: Algorithm for UGIB - Sabiston Textbook of Surgery

Step 1 - Resuscitation (ABCDE)

Airway

• Patients with hematemesis, altered consciousness, or inability to protect their airway require endotracheal intubation
• Prophylactic intubation is NOT routinely indicated - it is associated with cardiopulmonary complications including pneumonia
• Intubation is mandatory before balloon tamponade (Sengstaken-Blakemore tube)

Circulation - IV Access

• Two large-bore IV catheters (≥18G) in antecubital veins, OR large-bore single-lumen central cordis
• Begin crystalloid resuscitation immediately

Transfusion Strategy

• Restrictive strategy (transfuse at Hb <7 g/dL) - shown by multiple RCTs and guidelines to reduce all-cause mortality and rebleeding vs. liberal strategy
• Higher threshold (Hb <8 g/dL) for patients with ischemic cardiac disease
• For severe/ongoing massive hemorrhage: balanced resuscitation 1:1:1 ratio of packed RBCs : fresh frozen plasma : platelets (as in trauma)
• In active brisk bleeding, do not wait for lab results - transfuse clinically

Acid Suppression - PPI

• High-dose IV PPI is a cornerstone of UGIB management:
  • Active bleeding / hemodynamic instability: Esomeprazole or Pantoprazole 80 mg IV bolus
  • No active bleeding: 40 mg IV
  • If EGD delayed beyond 12 hours: repeat 40 mg IV dose
• IV PPI infusion (80 mg bolus + 8 mg/hr continuous) after endoscopic hemostasis in high-risk ulcers significantly reduces rebleeding, need for surgery, and hospital stay
• H2-receptor antagonists are NOT adequate for acute ulcer bleeding (not shown to reduce rebleeding)
• Pre-endoscopy PPI: A 2022 Cochrane systematic review (PMID 34995368) found that pre-endoscopy PPI reduces the proportion of patients with high-risk stigmata at endoscopy and decreases the need for endoscopic treatment, but does not reduce mortality, rebleeding, or surgery rates

Vasoactive Drugs (for suspected variceal bleeding)

• Octreotide: 50 mcg IV bolus + 50 mcg/hr continuous infusion for up to 5 days
  • Causes selective splanchnic vasoconstriction
  • Meta-analysis: as effective as sclerotherapy for controlling variceal bleeding, with fewer adverse events
  • Preferred over vasopressin due to safety (no cardiac complications)
• Terlipressin: 2 mg IV q4-6h (not available in USA) - shown to reduce mortality in variceal bleeding
• Somatostatin: 250 mcg bolus + 250 mcg/hr infusion

Prokinetics (pre-endoscopy)

• Erythromycin 250 mg IV over 20-30 minutes, 20-90 minutes before endoscopy
  • Motilin receptor agonist - clears stomach of blood, clots, food
  • Improves gastric visualization (77% vs 51%; OR 4.14, 95% CI 2.01-8.53)
  • Reduces need for second-look endoscopy (15% vs 26%; OR 0.51)
  • Shorter hospital stay (mean -1.75 days)
  • Monitor QTc prolongation; check CYP3A4 drug interactions
• Metoclopramide: alternative if erythromycin contraindicated

Antibiotics (cirrhotic patients only)

• Up to 20% of cirrhotic patients admitted with GIB already have bacterial infection; up to 50% develop infection during admission - associated with increased mortality
• Meta-analyses show prophylactic antibiotics reduce both infectious complications and possibly mortality in cirrhotic patients with variceal bleeding
• Regimen: IV ceftriaxone 1 g q24h × 7 days (preferred; superior to fluoroquinolones in areas of high quinolone resistance); alternatives: IV ciprofloxacin 400 mg q12h or IV levofloxacin 500 mg q24h

Nasogastric Tube

• NOT routinely recommended - no benefit on clinical outcomes in RCTs
• May be used to confirm UGI source if unclear, or to clear stomach before endoscopy as an alternative to erythromycin
• Nasogastric aspirate with non-bloody bilious fluid suggests pylorus is open and no active UGI bleeding distal to the pylorus

Step 2 - Risk Stratification

Glasgow-Blatchford Score (GBS) - Pre-endoscopy

• GBS = 0: Very low risk - safe discharge and outpatient management
• GBS ≥7: Predicts need for endoscopic control of bleeding
• GBS is superior to Rockall score for predicting need for intervention - endorsed by ESGE, International Consensus Group, Asia-Pacific Working Group

Rockall Score - Post-endoscopy (predicts mortality and rebleeding)

• Score ≥3: Significant mortality/rebleeding risk
• Score 0-1: Low risk

AIMS65 Score

• Mulholland and Greenfield's Surgery, p. 3204-3205

Step 3 - Endoscopy (EGD)

Timing

• Most patients: EGD within 24 hours of presentation - improves outcomes, allows safe discharge of low-risk patients
• Suspected variceal bleeding or cirrhosis: EGD within 12 hours
• Hemodynamically unstable despite resuscitation: EGD within 12 hours after stabilization
• A 2023 meta-analysis (PMID 36578195) confirmed early colonoscopy improves outcomes, but for non-variceal UGIB, very urgent (<6 hours) endoscopy has NOT been shown to improve mortality
• Upper GI barium studies are contraindicated in acute UGIB (interfere with subsequent endoscopy, angiography, or surgery)
• Routine second-look endoscopy is NOT recommended

Step 4 - Forrest Classification and Endoscopic Hemostasis Decisions

Forrest Classification for Peptic Ulcer Bleeding

• Sabiston Textbook of Surgery, Table 98.2

Endoscopic Hemostasis Techniques

• Endoscopic band ligation (EBL) - preferred over sclerotherapy
  • Initial hemostasis: 80-85%; rebleeding: 25-30%
  • Meta-analysis: EBL reduces rebleeding, overall mortality, and death from bleeding vs sclerotherapy, with fewer complications (especially esophageal strictures)
  • Technique: 2 bands per variceal column - 1 at the GEJ, 1 about 4-6 cm proximally
• Endoscopic variceal sclerotherapy (EVS) - alternative if EBL not feasible
  • Sclerosants: ethanolamine oleate, sodium tetradecyl sulfate, sodium morrhuate
  • Hemostasis in 85-95% of cases; rebleeding 25-30%
  • Complications: esophageal ulcers/perforation, strictures, mediastinitis
• Sleisenger and Fordtran's, p. 318

• Cyanoacrylate (tissue glue) injection - more effective than band ligation for fundal varices (GOV2/IGV1)
• EBL can be used for type 1 gastroesophageal varices (GOV1)

Step 5 - Second-Line Interventions (Failed Endoscopy)

Interventional Radiology

• CT Angiography (CTA): Detects bleeding at ≥0.3 mL/min; sensitivity 85%, specificity 92%
  • Useful when EGD fails to identify source, or patient cannot undergo endoscopy
  • Useful for deep small bowel lesions beyond endoscopic reach
• Visceral angiography + transcatheter arterial embolization (TAE):
  • Detects active bleeding at 0.5-1.0 mL/min
  • Achieves hemostasis in ~80% when bleeding site identified
  • If bleeding has stopped/is intermittent, angiography may be non-diagnostic

• For variceal bleeding refractory to endoscopic + pharmacologic therapy (~10-20% of cases)
• Rescue TIPS (emergency TIPS within 72 hours) is now preferred over surgical shunt in Child-Pugh B/C patients with refractory variceal bleeding
• Preferred over surgery because operative mortality is high in cirrhotic patients

Balloon Tamponade (Sengstaken-Blakemore / Minnesota / Linton-Nachlas tube)

• Used only as a temporizing bridge to definitive therapy in massive variceal hemorrhage
• Achieves initial control in 85-98% of cases
• Rebleeding recurs in 21-60% after deflation
• 30% rate of serious complications: aspiration pneumonia, esophageal rupture, airway obstruction
• Endotracheal intubation is mandatory before placement
• Should NOT be used as definitive therapy
• Sleisenger and Fordtran's, p. 318

Step 6 - Surgical Management

• Ongoing hemorrhage unresponsive to endoscopic + angiographic therapy
• Hemodynamic instability refractory to resuscitation

Bleeding Duodenal Ulcer

• Anterior longitudinal duodenotomy extending across the pylorus
• Figure-of-eight sutures to ligate the gastroduodenal artery at the top and bottom of ulcer crater + U-stitch for transverse pancreatic branches
• Close incision vertically as Heineke-Mikulicz pyloroplasty
• Truncal vagotomy added if patient hemodynamically stable (reduces recurrent ulceration)

Bleeding Gastric Ulcer

• Wedge resection and primary closure for greater curvature, antrum, body lesions
• Distal gastrectomy (Billroth I or II or Roux-en-Y) for lesser curvature ulcers at the incisura
• Anterior gastrotomy with biopsy + oversewing for GEJ ulcers (avoids GEJ compromise)
• Important: 4-5% of benign-appearing gastric ulcers are malignant - always biopsy/excise

Esophageal Varices - Failed TIPS/Endoscopy

• Surgical portosystemic shunts (total or selective):
  • Total/nonselective: end-to-side or side-to-side portocaval shunt (decompresses entire portal system but increases encephalopathy)
  • Selective: distal splenorenal shunt (Warren shunt) - preserves hepatic portal flow while selectively decompressing gastroesophageal varices
• Devascularization procedures (Sugiura procedure): extensive paraesophageal devascularization + esophageal transection + splenectomy + vagotomy + pyloroplasty
  • Rarely used emergently (operative mortality up to 100% in Child C)
  • 5-year survival: 44% (Child A), 22.5% (Child B), 0% (Child C)
• Sabiston Textbook of Surgery, p. 2187-2188

Step 7 - Antithrombotic Drug Management

• Mulholland and Greenfield's Surgery, p. 3210-3211

Summary Algorithm

UGIB (hematemesis / melena / coffee-ground emesis)
│
├─ GBS = 0 → Safe discharge + outpatient EGD
│
├─ Hemodynamically STABLE
│   ├─ Resuscitate (IV access, fluids, blood if Hb <7)
│   ├─ PPI 40-80 mg IV + erythromycin 250 mg IV
│   ├─ Octreotide if suspected varices; ceftriaxone if cirrhotic
│   └─ EGD within 24 hrs (within 12 hrs if suspected varices)
│       ├─ High-risk stigmata (Forrest Ia/Ib/IIa/IIb) → Endoscopic hemostasis
│       │   (combo therapy: injection + thermal/clips) → IV PPI infusion post-EGD
│       └─ Low-risk (Forrest IIc/III) → PPI + H. pylori test and treat → early discharge
│
└─ Hemodynamically UNSTABLE (persistent despite resuscitation)
    ├─ ICU admission + airway protection
    ├─ EGD when stabilized (within 12 hrs)
    ├─ If EGD fails → CTA → angiography + TAE (~80% success)
    ├─ Variceal: rescue TIPS if EBL + octreotide fails
    ├─ Balloon tamponade (SB tube) as bridge to TIPS
    └─ Surgery if all above fail (last resort)

Key Evidence References