GI bleeds are categorized by location in relation to the Ligament of Treitz. Upper GI bleeds (UGIB) occur proximally, in the esophagus, duodenum, or stomach. They classically present with hematemesis, with either frank red blood or coffee-ground emesis, or melena. Less commonly, brisk UGIB can also present as hematochezia which can complicate the identification of the origin of the bleed in some cases. Lower GI bleeds (LGIB) occur distally, in the small bowel, colon and anorectum. In contrast, these usually manifest as hematochezia.

The specific etiologies of GI bleeds are numerous and will not be covered individually in this article, as often the exact origin is unknown at the time of Emergency Department evaluation. Instead, we will focus on areas of interest that may guide or change your management of these patients in the Emergency Department.

Diagnostics in the ED

+ Fecal occult blood testing (FOBT)

Per the American Gastroenterological Association, the only indication for FOBT is in colorectal cancer screening

Intermittent GIB or very acute GIB that has not had sufficient time to travel through the GI tract may not produce positive FOBT and cannot be ruled out based on a negative result (1, 2)

Data suggest FOBT is only 25% sensitive in anemia due to UGIB

There is no strong literature to support its use in trauma, syncope, or hypotension

PPV of FOBT prior to administering anticoagulants for VTE is ~30%

Positive FOBT in normal-colored stool without history of GIB should not preclude patients from receiving thrombolysis in acute MI or stroke if appropriate

Many things can cause false-positives:

Foods: Animal-derived heme, high eroxidase foods (e.g. broccoli, cauliflower, cantaloupe, carrots, squash), food dye additives (red, blue, yellow)
Medications: Activated charcoal, Pepto-Bismol, simethicone, N-acetylcysteine
Iatrogenic: Topical povidone-iodine solution (often used when inserting a Foley catheter)
Extra-intestinal sources: Epistaxis, hemoptysis (3)

If other signs were present, FOBT was not found to change decision in GI consult or endoscopy (4)

Summary: FOBT is not a reliable test in an acute setting and therefore should not be utilized to direct management

+ Utility of CT scans

CT angiography has been shown to have as high as 89% sensitivity and 85% specificity across all GI bleeds (5)

Multidetector CT (MDCT) can identify acute GI bleeds at a similar rate to tagged RBC scans, but can provide additional information on location and underlying etiology (diverticula, tumors, angiodysplasia) (6)

Advantages:

Minimally invasive
There is widespread availability and can be performed immediately in the ED during suspected hemorrhagic episodes
Can detect bleeding in the small bowel distal to Ligament of Treitz, which is not readily accessible to endoscopy
Does not require bowel prep (7)

Disadvantages:

Does not provide direct visualization of the bleed (5)
Only diagnostic, not therapeutic (endoscopy provides both)
Risk of radiation and contrast reactions (allergy, nephropathy)

Summary: CT scans can be useful if they can be done relatively quickly and/or a significant delay to endoscopy/colonoscopy can be expected

There appears to be more benefits in LGIB than UGIB given the better visualization of the small bowel and avoiding delays due to bowel prep

SPECIAL CONSIDERATION IN MANAGEMENT

Reversal of anticoagulation

The decision to reverse is two-fold: (1) Reverse or not reverse and (2) partial or complete reversal. It should be made based on a balance between risk of thromboembolic events vs recurrent bleeding (risk-benefit), as well as cost-benefit. Patients can be stratified into high or low risk for thromboembolism if anticoagulation is reversed:

High risk: DVT or arterial embolism within past 6mo, valvular heart disease with afib, mitral mechanical valve, any mechanical valve with previous thromboembolic event, afib with previous cardioembolic event or risk factor for stroke, recurrent venous or arterial thromboembolism, any hypercoagulable state with at least one thromboembolic event
Low risk: Isolated venous or arterial thromboembolic event older than 6mo, afib without valvular disease, aortic mechanical valve, bioprosthetic valve (8)

Helpful lab tests to assess for anticoagulation based on agent:

PT/INR: Warfarin, can sometimes also be prolonged with rivaroxaban specifically
Anti-factor Xa activity: DOACs, LMWH, fondaparinux
Thrombin time and escarin clotting time: Dabigatran (9)

Vitamin K antagonists or warfarin with no bleed or mild bleed with excessive anticoagulation (INR>6):

Oral Vitamin K: 1-5mg (8)
IV Vitamin K: 1-2.5mg (10)
Can consider FFP or PCC if endoscopy is scheduled within 6-12 hours as it improves rebleeding and mortality rate following endoscopy

Vitamin K antagonists or warfarin with clinically significant bleed:

FFP: IV 15mL/kg

Large volume infusion
Prolonged time needed to match blood group and thaw
Takes several hours for partial reversal and ~9 hours for complete reversal (INR<1.5)

PCC: IV 25-50IU of factor IV/kg depending on baseline INR

Quick reconstitution into small volumes (20mL for 500IU)
Rapid infusion over 20-30 minutes with faster correction
4F-PCC should be preferred over 3F-PCC for acute reversal of oral VKAs

Recombinant factor VIIa:

Not enough large-scale data to support use
Has been correlated with unacceptably high risk of thromboembolism in urgent reversal

IV Vitamin K: 5-10mg infused over 30min, can consider a second dose

Slow infusion helps reduce risk of anaphylaxis
INR decreases within 2-4 hours and reaches normal range within 24 hours (not ideal for urgent reversal)
Should be administered if reversal is to be sustained to prevent rebound coagulopathy

If actively bleeding with persistent or intermittent hemodynamic instability, consider administering coagulation factors even if INR is therapeutic using 4F-PCC

Heparin:

Protamine: 1-1.5mg/100U (if last heparin dose was given immediately prior), 0.5-0.75/100U (30-60 min prior), 0.25-0.375/100U (if given >2 hours prior)
Doses should not exceed 50mg at a time

LMWH:

Protamine can reverse 60-75% of the anti-Xa activity of LMWH
Dalteparin: 1mg protamine/100U

Can give second dose of 0.5mg/100U if bleeding continues or PTT remains prolonged 2-4 hours after first dose

Enoxaparin: 1mg protamine/1mg (if <8 hours after last dose), 0.5mg/1mg (if 8-12 hours), no protamine required for reversal if >12h after last dose

Can give second dose of 0.5mg/1mg if bleeding continues or PTT remains prolonged 2-4 hours after

Fondiparinux:

There is not enough human data to recommend usage of reversal agents

Dabigatran:

Idarucizumab: 5mg (given as 2 2.5g/mL vials) (9)

DOACs:

Andexxa: 400mg bolus + 480mg infusion

High dose administration: 800mg bolus + 960mg infusion (predicted to be rarely used in real-life practice)

Anti-Xa activity decreased by 92%, which was maintained throughout infusion
In Phase 3b-4 studies, 85% of GI bleeds had good hemostasis (defined by drop in Hb <10% from baseline at 12 hours)

Variceal bleeding

Acute variceal bleeding is the most life-threatening complication of liver cirrhosis, with a mortality rate of 20% at 6 weeks. Varices are present in as many as 50% of patients with cirrhosis and acute bleeds can occur in at least 30% of patients who have varices. Acute variceal bleeding is responsible for 70% of all UGIB in patients with cirrhosis (11). Ultimately, patients with variceal bleeding require endoscopic variceal ligation (EVL) but it is important to understand what therapies should be initiated promptly in the ED to improve overall outcomes.

Goals of treatment: Control hemorrhage, reduce risk of early-rebleeding (within 5 days), prevent bleed-related complications (infection, hepatic encephalopathy, AKI) (12)
Standard of care includes: Vasoactive drugs + prophylactic antibiotics + endoscopic techniques
Vasoactive drugs: IV octreotide 50ug bolus + 25 or 50 ug/h infusion

Other options include vasopressin, terlipressin and somatostatin if available
Early use can reduce rate of active bleeding AND make endoscopy easier to perform for diagnostics and therapeutics
Improved hemostasis, reduced 7-day mortality, transfusion requirement, and hospital LOS
When used with EVL, is superior at 5-day success rate than EVL alone (11, 13)

Prophylactic antibiotics: IV ceftriaxone 1g x7 days (13)

Early use in acute bleed has survival benefit, reduced risk of bacterial infections, and reduced risk of re–bleed (12)
Infection is a strong prognostic indicator and complicates the course of 42% of cirrhotic patients after hospitalization (13)

Most frequently encountered infections: SPB, UTI, PNA

Short-course PPI post-EVL can reduce post-banding ligation ulcer size (11)
Strict transfusion criteria should be utilized when the patient is hemodynamically stable

Transfuse to goal Hb 7-8
There is data to suggest that transfusing to 9-11 is associated with higher mortality, due to the higher rates of re-bleeding and increased hepatic venous pressure gradient (12)

Primary prophylaxis with non-selective beta blockers (carvedilol, propranolol, nadolol) to prevent first variceal bleed is not something we typically initiate in the ED

They should not be started during management of an acute bleed
If they are already taking them, they should be discontinued until the acute bleed has resolved (12)

Clinical Decision Making Tools

There are numerous scores that have been studied in relation to GI bleeds, this discussion is limited to three commonly used, well-studied scores that are easily accessible via MDCalc, and can be calculated appropriately with information that is readily available to us in the ED.

Oakland Score

Designed to be evaluated in the ED for patients presenting with LGIB to determine if outpatient management is feasible
Factors included for assessment: Age, sex, previous LGIB admission, DRE findings, HR, SBP, Hb
Score of 8 or less → 95% probability of safe discharge

Safe discharge = absence of rebleeding, blood transfusion, therapeutic intervention, 28-day readmission, death

Can be used in patients taking oral NSAIDs, antiplatelets, or anticoagulants (14)
There is some data to suggest that a modified score that does not include DRE can be safely used AND it is possible that a score of 10 can still maintain significant safety in discharge (15)

Glasgow-Blatchford Bleeding Score

Designed to be used in patients with UGIB being considered for hospital admission to determine patients who are low risk and candidates for outpatient management
Factors included for assessment: Hb, BUN, initial SBP, sex, HR, presence of melena, recent syncope, hepatic disease history, presence of cardiac failure

There is also a modified score that performs similarly and only takes into account HR, BP, BUN, Hb

Score of 0 → low risk, patients do not require medical intervention (blood transfusion, endoscopy, surgery) (16)
There is data to suggest that a score of 1 could also be considered low risk (17)
Data suggests that this score has better sensitivity than the pre-endoscopy and complete Rockall scores (16)
Also data to suggest that it is better at predicting need for intervention than the AIMS65 score (17)

Rockall Score (Pre-Endoscopy)

Designed to determine severity of UGIB in patients who have not yet had endoscopy and can stratify patients into low or high risk for mortality
Factors included for assessment: Age, shock, comorbidities

There is also a complete version of this score that incorporates endoscopy findings, but this is less relevant to our practice

Higher score → greater rate for mortality prior to endoscopy (18)
Does not accurately predict re-bleeding, which has been studied as an independent predictor for mortality (19)

AIMS65 Score

Used to determine risk of in-hospital mortality from acute UGIB
Factors included for assessment: Albumin, INR, presence of altered mental status, SBP, age
Higher score → greater rate of in-hospital mortality (20)

Higher scores also correlate with cost and LOS

Better predictor of in-hospital mortality, ICU admission, and LOS (21)

Post by Gina Qin, MD MPH

Dr. Qin is a PGY-1 in Emergency Medicine at the University of Cincinnati.

Editing by Ryan LaFollette, MD and dr. bronwyn finney, MD

Dr. LaFollette is an Associate Professor in Emergency Medicine at the University of Cincinnati and co-editor of TamingtheSRU.

Dr. Finney is a Chief Resident in Emergency Medicine at the University of Cincinnati.

Taming the SRU

Taming the SRU

Blog

Taming the SRU

Diagnostics: GI Bleeding

Diagnostics in the ED

+ Fecal occult blood testing (FOBT)

+ Utility of CT scans

SPECIAL CONSIDERATION IN MANAGEMENT

Reversal of anticoagulation

Variceal bleeding

Clinical Decision Making Tools

Post by Gina Qin, MD MPH

Editing by Ryan LaFollette, MD and dr. bronwyn finney, MD

References

Taming the SRU

About Us

SRU (pronounced "shrew") = Shock Resuscitation Unit

Stay Connected

Categories

Links

Some Helpful Foam

Academic life in EM

Life in the Fast Lane

EM Lyceum

R.e.b.e.l EM

EMCrit

ERCast

UMEM Educational Pearls

canadiem

P.H.A.R.M.

PEMBlog

Downeast EM

EM Cases

St Emyln's

ECGWaves

PENN TTM Academy Podcasts

Tox in Ten Podcast

Tox Shorts

Tags