Upper GI Bleeding
The purpose of this essay is to help you learn about the causes, workup, treatment and disposition of patients who present to the ED with upper GI bleeding (UGIB).

Diagnosis of UGIB
On a good day the diagnosis is easy - the patient will report vomiting blood and you are able to reasonably confirm it with a rectal exam. That sounds illogical and when you read the foregoing you probably thought "why a rectal exam and not an NG tube (NGT)?" More on that later, but first some facts. These are based on some gruesome studies, probably involving medical student "volunteers". What these intrepid volunteers did was to have varying amounts of their own blood instilled in their stomachs via NGT. Their stools were then monitored. In almost all cases, w/in 20 hours (and in as few a 4 hours) the stool turned black, melenotic, or frankly red. This lasted for up to 5 days. While not stated, I bet their stools remained OBT (+) for quite a bit longer.

So, a history of vomiting blood and bloody stools = UGIB. This is quite reliable. How about "coffee ground emesis (CGE)"? CGE and positive stools also = UGIB. The "coffee grounds" in CGE are not clots but aggregations of hemoglobin from lysed RBC. Clots are unusual in UGIB, but if present will not look like coffee grounds, they will look like clots. If you want to really know what CGE looks likes, do the following demonstration (which I did as an intern).
  • Take some left over blood drawn on a patient
  • Mix with clear gastric juice from a patient with an NGT
  • Observe the results (after a few minutes), which will look something like this:
One thing to be cautious of is that brown emesis does NOT = UGIB. Emesis comes in a lot of colors, including brown. Simply vomiting brown material has no particular significance, AFAIK.
What about the use of the NGT? NG placement was found to be the most painful common ED procedure in one study. It is extremely unpleasant. Use it when you need to, but be sure you need to. The NGT is useful in the context of actual or suspected GIB for the follwing purposes:
  • For diagnosis of UGIB
    • when you suspect UGIB and there is no reliable report of vomiting blood or CGE and a positive stool
    • when you suspect an acute UGIB and not enough time (~24h) has passed for the stool to become bloody
  • To determine if there is active bleeding and estimate the rate
    • the NGT is overused for this purpose
      • someone who vomits bright red blood (BRB) or CGE in front of you is actively bleeding, you don't need an NG to prove it
      • vomiting CGE in front of you argues against active UGIB and implies UGIB that has stopped
      • clear vomitus rules out active GIB
  • To rule out UGIB when you suspect LGIB
    • a trick question that reveals a truth is "what is the commonest cause of lower GIB (LGIB)?
      • the answer is: "UGIB"
        • most bloody or melenotic or black stools are caused by UGIB even w/o any upper GI symptoms
      • true LGIB usually present as bright red blood per rectum (BRBPR)
      • patients may have BRBR with massive UGIB
        • suspect massive UGIB when a patient is tachycardic and perhaps hypotensive and has nausea or upper abdominal pain and large amounts of BRBPR
Caution: ~5% of UGI bleed will not have bloody or CGE or a (+) NG aspirate. Usually this is from bleeding distal the ligament of Treitz, which is relatively rare as a cause of significant bleeding.

Treatment of UGIB in the ED
In one study of ED patients, the following were the causes of UGIB.
  • duodenal ulcer ~30%
  • erosive gastritis/duodentitis ~20%
  • varices ~15%
  • gastric ulcers ~10%
  • Mallory-Weiss tear ~10%
  • Esophagitis ~10%
  • AVM <5%
  • other: rare
While it is unusual for the precise cause of UGIB to be deliniated in the ED, an educated guess is needed to help determine empiric treatment and disposition.

Acid/peptic disease
Suspect acid/peptic causes (the "itisis" and ulcer) when there are symtoms of dyspepsia or GERD and risk factors such as 
    • history of ulcer or (+) testing for H. pylori
    • smoking
    • alcohol
    • NSAID use
    • family history of ulcer disease
    • recent discontinuation of long term PPI use
  • Empiric treatment of acid/peptic disease with UGIB
  • Montoring/resuscitation prn
  • High dose IV PPI (Protonix 80mg IV followed by an 8mg/h drip)
  • Prepare for EGD (see below)
Variceal bleeding should be suspected when a patient has painless UGIB and a history or evidence of advanced chronic liver disease. It is especially likely when there is a known history of varices or variceal bleeding. These bleeds can be massive and made all the worse because advanced liver disease also causes coagulopathy.
  • Other causes of UGIB occur commonly in patients with varices
    • In cirrhotics with UGIB varices are reported as the cause in a wide range of 50-90% of cases.
  • Empiric treat of variceal UGIB
    • standard resuscitation and monitoring
      • even more so than in other causes of UGIB try for two large bore IV's (one for blood products/volume and another for backup and meds)
    • octreotide (somatostatin analog)
      • 250 mcg bolus followed by 250 mcg/hour
    • antibiotics
      • 1 gm ceftriaxone or 400mg cipro
    • FFP if INR prolonged
    • Prepare for EGD (see below)
Mallory-Weiss tear
Suspect a Mallory-Weiss tear when when hemtemesis is preceeded by retaching. There is usually esophageal pain (chest or back) associated.
  • ED treatment of suspected Mallory-Weiss tear
    • standard resuscitation and monitoring
    • prepare for EGD
    • there is no specific treatment
Preparation for EGD
EGD is often both diagnostic and therapeutic in UGIB. Surprisingly, the literature on this subject provides little support for immediate endoscopy versus "early" endsocopy within 24 hours of presentation. The ED role is to do the following.
  • Resuscitate prn
  • Notify the consultant endoscopist
  • Empty the stomach to provide conditions for good visualization
    • traditionally this has meant NG aspiration with lavage until clear
    • recent studies show that erythromycin, 250mg IV is as effective as an NG lavage
      • if contemplating erythro versus NGT discuss with endoscopist
Disposition in UGIB
Traditionally and in our litigious climate all patients with UGIB are admitted, However, patients who have no evidence of active bleeding and are stable can be considered for D/C.
  • if contemplating D/C it's especially reassuring when EGD has revealed the specific cause of bleeding
    • this solidifies prognosis and allows for specific treatment