Acute Upper GI Bleeding
Epidemiology
> 300,000 admissions / year
Cost ~ $2.5 billion
- 10% Overall Mortality (~ 40% if recurrent)
- 10% require OR
- 10% Lower, 85% Upper, 5% SI
- 10% Upper GI Bleeds --> Hematochezia (>1L blood loss)
- 10% Upper GI Bleeds --> Syncope
Presentation
Hematemesis
Coffee-ground emesis
Melena - can be seen with 50cc of blood, can persist for up to 5 days post bleed, result of Hgb --> hematin and SI digestion, occasionally seen with SI or R colon bleeding
Hematochezia - usually indicative of lower GI bleed, seen with rapid transit upper GI bleed >1L
Anemia
- 50% both hematemesis and melana, 30% hematemesis alone, 20% melana alone
Hemorrhagic Shock
Dizziness
Orthostasis (definition HR increases by 20, BP decrease by 10 c/w blood loss ~20%)
BUN increase - out of proportion to their hydration status
Initial Treatment
Hemorrhage Shock Resuscitation
NG
- gastric contents non-bloody output somewhat reassuring but could have duodenal bleeding with pylorospasm
- non-bloody bilious output rules out upper GI bleed
- Room Temp. Lavage removes old blood and lysens that promote ongoing bleeding (iced lavage: coagulopathy of hypothermia more problematic than vasoconstriction is helpful)
Correct any coagulopathy aggressively
IV PPI
+/- IV Octreotide = useful with varicel bleeds, harmless in others but maybe benefitial
EGD - diagnostic and therapeutic test of choice; early 90% sensitivity (don't delay EGD with other theapies), late (>24 hr post bleed) 20% sensitivity, demonstrated decreased transufion need with EGD in 6 vs 24 hrs; get it done in 6 hours
Acidosis causes decrease in platelet aggregation
Etiologies
PUD – 50-60%
Acute Mucosal Lesions – 15-30%: gastritis, duodenitis, esophagitis
Varices – 5-15%: high hospital morality ~30%, overall 1 year mortality 60-70%
Mallory-Weiss Tears – 5%
Esophagitis – 5%
Other – 5% (Dieulafoy’s Lesion, CA, Watermelon Stomach, Arterio-GI fistulas, Angiodysplasia, Crohn’s Dz)
Bleeding Peptic Ulcers
5% - initial manifestation of PUD
20% of peptic ulcers bleed
Most lethal complication - worse than perforation
Due to erosion into submucosal vessel (0.1-1.8 mm, less likely to bleed massively than named vessel)
More likely to represent Neoplasm than DU (10%) - ulcers don't turn into cancer, cancer can present as an ulcer
Bleeding DUs
Posterior DUs can erode into GDA massive bleeding
Endoscopy
EGD Presentation
Stigmata of bleeding prevalance: clean base 49%, dot 23%, clot 13%, nbvv 8%, bleeder 7%
Stigmata of bleeding risk of rebleeding: active ~55%, non-bleeding visible vessel ~45%, clot ~25%, dot ~10%, clean base ~5% (overall risk ~25%)
Question if clean base or dot was really the source of the bleed?
When to restart ASA in heart disease patients? rebleeding increased by 5% if restarted within 24 hours, but overall mortality increased by holding ASA due to cardiac etiologies
EGD Management
Combination therapy is the mainstay of endoscopic treatment
- Epinephrine + heat probe - 3.7% rebleeding
- Epinephrine + thrombin - 4.5% rebleeding
- Epinephrine + fibrin glue - 10% rebleeding
- EGD clipping ()
Gold probe: has heater probe and needle so you can burn then coagulate
Indication for OR = endoscopist says he can't fix it, 2nd attempt at endoscopy is usually recommended, some always repeat endoscopy 24-48 hrs
Limits to endoscopy: the larger the vessel, the less likely the endoscopist is able to stop bleeding,
Reglan and/or erythromycin can help empty stomach of clot pre EGD
Medical Adjuncts
IV-->po PPIs
- Prilosec 80mg IV bolus then 8 mg/hr x 72 hr then transition to PO (80/8x3)
Nutritional Support
H-Pylori testing/eradication
Follow-up EGD - confirm eradication, confirm healing, R/O neoplasm
Pantoprazole 40 mg Oral vs. IV vs. Famotidine 20 mg IV
- H2 blockers (pepcid, etc.) become ineffective very quickly (tachyphylaxis) and therefore didn't seem to make much on an impact on bleeding recurrence, etc.
Octreotide Therapy
- Of signficant benefit with variceal bleeds, no good data for other causes
- Minimal downsides
- Dose: 100 mcg bolus then 50 mcg/hr IV drip
PPI Studies:
H2 blockers provide transient inhibition of acid output with early tachyphylaxis so PPIs are prefered due to lasting inhbition
1998-99 Randomized, Prosp. Controlled Trial Prince of Wales Hosp. in Hong Kong - 240 pts c upper GI bleeds & Ongoing Bleeding or Non-Bleeding Visible Vessel on EGD
Randomized to 80mg IV omeprazole bolus then 8mg/hr x 72hr VS placebo post-EGD
Outcomes:
- Endoscopic retreatment: 5% omeprazole, 20.8% placebo (p<0.001)
- transfusion: 2.7 median units omeprazole, 3.5 median units placebo (p=0.04)
- Surgery: 2.5% omeprazole, 7.5% placebo (p=0.14)
- 30-day mortality: 4.1% omeprazole, 10% placebo (p=0.12)
- median hospital stay: 4d omeprazole, 5d placebo (p=0.006)
- cost per episode of rebleeding prevented: 9000 omeprazole, 12000 placebo
? Benefits of Pre-EGD PPI ?
Same authors/institution/approach, 631 stabilized acute upper GI bleeds, comparing PPI pre- and post- endoscopy vs. just PPI postendoscopy
Randomized to 80/8 dosing of Omeprazole vs placebo with selective EGD within 24 hr.
Saw lower rate of actively bleeding ulcers, nbvv ulcers, and higher rate of ulcers with a clean base
Seems to have modulated the disease process before even getting to endoscopy
PPI recommendations in Acute UGI Bleeds from Ulcers
PPI Pre-EGD if possible
For high / moderate risk findings: 80/8/8: 80 mg IV bolus, 8mg/hr for 72 hrs, PO for 8 months
For low risk (clean base or flat spot): po PPI initially
For all PPI for minimum of 8 weeks
Surgery for Bleeding DU
Important to have some localization - easier than lower GI bleed, usually duodenum, if in stomach, open stomach fix it close stomach
If duodenum: A) Longitudinal opening of pylorus, B) 4 quadrant oversew of bleeder (avoid bile duct, stop bleeding), C) Weinberg modification of Heinke-Mikulicz pyloroplasty (open longitudinally, close transversely) D) Truncal Vagotomy (in some circumstances like failed PPI therapy, Left vagus closely adherent to anterior esophagus, right falls away to the right ~1cm - often send them to path to confirm neural tissue)
Surgical Adjuncts: J-Tube, # 19 – Channel Drains, +/- G - Tube
Other Surgical Options: Vagotomy and antrectomy, Gastric Resection, Selective & Highly Selective Vagotomies, Angioembolization
Angioembolization (cirrhotics, terrible surgical candidates):
- good initial success rate (90-100%)
- high recurrance (~50%): average time to failure = 4 days [range 1-22]
- Possible bail-out option with potential ischemic complications (~5%)
Portal HTN – related UGI Bleeding (See Liver Failure Lecture)
Upper GI Mucosal Lesions
Esophagitis, Gastritis, Duodenitis
Usually not clinically significant bleeding, given PPI
Stress- Related Mucosal Disease (SRMD)
Various terms: stress gastritis, stress ulcers (Curling’s c burns, Cushing’s c CNS Trauma), hemorrhagic gastritis, etc….
Recognized since 1800s, Associated with HCl in 1970s, 1980s – 1990s “stress gastritis” prophylaxis introduced: antiacids, H2 blockers/sulcralfate, PPIs
Extent varies: asymptomatic endoscopic evidence 74-100%, occult bleeding 15-50%, gross bleeding 5-25%, clinical bleeding 2-6%
Decreased rates of clinical bleeding in recent studies (ie since 2000) to 0.1-4% (with or without prophylaxis) - likely due to improved ICU care
Clinical Bleeding (Gross blood/coffee grounds) carries mortality of 50-75% in critical pts - increases ICU mortality by 4x
Risk Factors: respiratory failure requiring vent support >48 hrs (independent risk factor), coagulopathy (INR>1.5, PLT < 50,000, independent risk factor), acute renal failure, acute liver failure, SIRS/Sepsis, hypotension, head/spinal cord injuries >35% TBSA burns, major surgery (>4hr), high dose steroids (>250 mg/day)
Everyone ends up getting PPIs in ICU but keep risk factors, especially independent risk factors above in mind
Pathogenesis of SRMD
- Hypotension -->gastric mucosal ischemia -->loss of mucosal integrity --> SRMD
- SIRS --> hypotension and gastric mucosal ischemia
- Trauma --> hypotension and SNS activation
- SNS activation --> gastric mucosal ischemia
- Gastric mucosal ischemia --> decreased gastric motility --> increased exposure of gastric wall to HCL --> loss of mucosal integrity
Dx of SRMD - endoscopy
Prophylaxis of SRMD
Antacids – some benefit, labor intensive, significant SEs
Sucralfate – labor intensive, ? Benefit of decreased pneumonia rate
H2 Blockers – lose effect, thrombocytopenia
PPIs – current DOC but ? :increased rate of c. diff, some medication interactions, rebound acid hypersecretion once stopped
In some studies ppx reduces incidence of SMRD but not mortality, in others no benefit over placebo
Key – good ICU support
Treatment of SRMD
Endoscopic - difficult, coagulate whole stomach
Angiographic
Operative - total gastrectomy!
Prevention!
Mallory-Weiss Tear
“Non-transmural Boerhaave’s”
Post- emesis/retching
Typically at EG Junction
90% stop spontaneously
Avg ~ 1-2 cm long, 2 mm wide
Is Arterial Bleeding
Treatment: observation, gold probe dual therapy (careful in esophagus, not as thick walled as stomach), surgery with mucosal oversew through high gastrotomy
Dieulafoy’s Lesion
Caliber-persistent submucosal vessel - normally vessels taper as they approach mucosa, these stay inappropriately big and bleed despite large defect
Usually on lesser curve
No erosion --> difficult dx - hardest part is finding them
Treatment: Gold Probe Dual Therapy, Clipping, Surgery with Gastric Wedge Resection
Arterio-enteric Fistula
"The mother of GI bleeds"
Aorto-duodenal (associated with AAA or previous aortic surgery)
Hepatic visceral – enteric (associated with pancreatitis, pancreatectomy)
Both commonly present with “Herald Bleeds” - significant transient bleeding that stops then restarts massively
Aorto-duodenal fistulas
Primary – result of native AAA, Secondary – related to AAA graft infection - complicates 1% of open AAA repairs
Dx: EGD, CT, Angio
Treatment: Angio or Surgery with graft removal & extra-anatomic bypass
Hepatic Arterial – Enteric Fistulas
See 2nd Case Study
Treatment: angio (try to stay out of OR)
Post op hypotension = bleeding, in a whipple patient = viceral pseudoaneurysm or artero-enteric fistula until proven otherwise
High leukocytosis, significant anemia points towards arteroenteric fistula due to enteric contamination
Pro-coagulants help stop bleeding but hurt diagnostic tests to locate bleeding
Visceral Pseudoaneurysms
Rare but significant complication of pancreatitis & pancreatic surgeries
Mechanism: local sepsis &/or pancreatic ENZs causing vessel wall weakening or ligature sloughing +/- contribution from closed-suction drain erosion
? Increased risk with silk suture (use monofilamints - but these slip and cause massive bleeding of another flavor)
Occurs late (usually 3rd week post-op - frequently people are home)
Usually present with “herald bleed” out JPs or GI tract - need high index of suspicion
Dx/Rx: ANGIO - get on both sides to get collateral flow
Propensity for rebleeding if don’t get both sides
Help prevent with Falciform “Patch” over GDA Stump
Hemobilia
Bleeding from biliary system
Described in 1948 by Sandblom Triad: RUQ pain, Jaundice, GI Bleeding
Other Biliary “Triads” - Charcot's triad for cholangitis: RUQ pain, jaundice fever; Choledochal Cyst triad: RUQ pain, jaundice, mass;
Etiologies - Post-procedure (CBDE, ERCP, PTHC), Pseudoaneurysm (hepatic artery), Traumatic, Neoplastic (hepatocellular carcinoma, GB CA, Cholangiocarcinoma, metastatic disease)
Management
Evaluation, resuscitation, exclusion of other sources by endoscopy, coagulation profile, renla functional panel, liver function test
Angiography --> demonstrated lesion --> selective embolization --> if continued bleeding consider repeat angiography vs. operation
Angiography --> no lesion demonstrated --> observation --> if continued bleeding consider repeat angiography vs. operation