Assessment of Upper GI Bleeding.
DR HASHIM UDDIN AZAM
Upper GI Bleeding
• Haemorrhage from the GI Tract and adnexal structures proximal to the Duodenal-Jejunal flexure.
Upper GI Bleeds
v Difficult to assess incidence/prevalence.
v 100 episodes/100 000 people per year US
v (Silverstein et al 1981)
v Represents initial Sx of GI disease in >30% of patients (Fischer et al 1999).
v ~50% of cases develop during hospitalization*.
v Associated with significant Morbidity.
v Mortality 5-12% (mostly cardiovascular)
v Higher in rebleed patients (30-40%)
v (Friedman 1993)
Acute vs Chronic Bleeds
–Present via Acute referral pathway.
Degree of shock due to blood/volume loss.
May have other GI symptomatology.
–Present to GP (ED)
Related to chronic anaemia. May report Bleed.
Other Sx of GI pathology.
No degree of shock.
–Emesis of blood containing fluid.
•Fresh or Altered (Coffee ground)
–Passage of Altered Blood Stool
•Black, Tarry, Offensive Stool
•Indicates Bleed proximal to …
–Passage of fresh blood and clots per rectum.
•Varices (5-10%), Oesophagitis reflux or other, Malignancy, Mallory Weiss Tear, rupture.
•Gastritis, Peptic ulcer(70-80%), Malignancy, Dieulafoy.
•Ulcer, Aorto-enteric fistula, Diverticulum.
Initial Clinical Assessment
–? Symptomatic Shock
–Other features of GI pathology
–Examination of Vomitus/Melaena
–Re-examine Volume status
•Degree of Shock
•FBE, Coags, G+M
•U+E, LFT, Amylase/Lipase
•Aggressive IV Fluids
Oral contrast radiology
–Distinguish Upper from Lower GI Bleeds
•Non-Bloody aspirate seen in up to 16% of UGIB
–Clear stomach for Endoscopy
–Assist haemostatis (lavage)
•No Evidence to support this. (tepid or chilled)
–? Bleeding at presentation
•Initial aspirate 79% sensitivity, If positive 53% specificity.
–Monitor for rebleed (ongoing lavage)
•Most sensitive “online” marker for this.
•Clear aspirate 6% mortality, Bloody 18% mortality
•Low associated morbidity (0.01-0.05%)
•Affords prognostic information.
•Mucosa and its pathology
•Active bleeding sources
•Stigmata of Recent Haemorrhage
–Old blood, adherent clot etc.
•Varices-If not bleeding and no SRH-variceal bleed becomes a diagnosis of exclusion.
•Allows Biopsy, HP testing
–Identifies bleeding site in 70-96% with near 0 false positive rate.
–Differentiates major pathologies in >90%
–SRH still present.
–No delay during which rebleeding (10-30%) may occur.
–(Lieberman 1993, Laine 1998).
•Sooner rather than later
•Spiegel et al 2001 reviewed data from 23 studies
–Most had failings
–Early endoscopy safe, despite earlier reports.
–No decrease in mortality but decrease in transfusion and length of stay.
–Recommended prospective trial.
–Small Bleed (1-2u) now ceased.
–Large bleed (Non-Variceal) now ceased.
•Urgent when stable
–Suspected Variceal Bleed.
–Ongoing or recurrent bleeding.
•AMI, Acute Abdomen
•Patients in extremis
•May require intubation prior to endoscopy.
3--Oral Contrast Radiology
•Limited diagnostic capacity.
–Sensitivity 70-80%, false positive 10-20%
•Can interfere with subsequent testing.
•No prognostic information.
•Poor regional availability.
Endoscopy vs Contrast Radiography
–No benefit of endoscopy over radiography in terms of mortality, rebleeding, admission length or transfusion requirements
–Peterson et al 1981
“the development of therapeutic methods that are better than empiric antacid therapy for specific lesions may demand modofocation of our conclusions”
–endoscopy too successful.
•Localises bleeding points.
–Relies on extravasation not pathology.
•Poor results for varices.
–vasopressin infusion or embolisation.
•Requires bleed rate of >0.5-1.0ml/min for positive result.
–Bleed too fast to safely ‘scope.
•Can localise bleeding point in ~75% of cases.
–Failed endoscopic therapy and need to avoid surgery.
•IA Vasopressin-60-90% successful in stopping gastric bleeds.
•Embolisation (coils, PVA)-~80% success.
Upper GI Bleed prognostics.
–Need for treatment.
–Risk of Re-bleed.
–Risk of Death
–Outpatient vs Inpatient
–Ward vs HDU vs ICU
–Timing of endoscopy
–Length of stay
Need for Treatment
•Blatchford et al 2000
–Developed a scoring system:
•predicts need for intervention Endoscopy, transfusion etc.
•Input data consists of clinical findings and laboratory results.
•Small trial-not yet validated
Risk of Rebleed.
–Dependant on Ulcer appearance (Schaefer, Lieberman)
•Actively bleeding 80% chance on ongoing/recurrent bleed.
•Fibrin clot on visible vessel 40%
•Adherent clot 20%
•Flat pigmented spot in crater 10%
•Clear white base ~5%
•Extensive writings have been published on this point.
–Collated by Hussain et al 2000
–Predictors of mortality
o Age>60 yrs, Signif co-morbidities, raised INR, NSAIDs
Ø Patient course
o Transfusion >5 units, Shock, LFT changes, Renal failure.
Ø Endoscopic factors
o Varices, Active bleeding, SRH, Ulcer site.
The Friendly Version.
–Several described, few validated.
–Rockall et al 1996
•Analysed 4185 cases, prospective data collection.
•Devised numerical scoring system that predicts rebleeding and mortality risk.
•Validated on >1600 cases originally.
•Subsequently validated in US, Holland.
Rockall Scoring System.
–Score 0-2 5%; 3-5 10-25%; 6-8+ 33-45%
–Score 0-2 0%; 3-5 3-10%; 6-8+ 17-40%
•Its trickier than you thought.
•Clinical assessment is important particularly in guiding resuscitation.
•Endoscopy (early) is now the cornerstone of diagnosis.
•Endoscopy also yields prognostic information that may guide clinical decisions.