Upper GI Bleeding

by Bob Stuntz

Coming back from time off, and coming back with an appropriate topic for the occasion…the GI tract.  In particular, upper GI bleeding.  

A 55 year old female presented with complaints of blood in her stools for the last 24 hours.  She said at first it was a dark red color, and seemed to be getting darker.  The presentation was pretty standard: abdomen is soft and nontender and no complaints of pain.  The resident wanted labs including a CBC, chem 7, coags, and a type and screen.  The presentation brought up two interesting topics of conversation.  First, no rectal was performed.  When I asked why not, the answer was, “I don’t see how it’s going to change my management.”  Second, the disposition decision was that if labs were normal, she could likely go home.  We’ll address each issue separately.  

Lets get this out in the open: we don’t want to do certain exams.  No one really wants to do a pelvic, or a rectal, or a male GU exam.  But we have to.  If a patient comes in presenting with any complaints that make you suspicious for GI bleeding, a rectal exam is simply mandatory.  The exam itself can give you a source of the bleeding (fissure, hemorrhoid, etc).  It can also give you a clue as to the location (bright red, likely lower.  Melanotic, likely upper, as we will see below).  Lastly, it actually can change your management.  Again, we will get to this shortly.  These exams are crucial components of your history and physical.  

Now lets talk disposition.  Can a 55 y/o with GI bleeding go home safely?  Possibly.  The real question this comes down to is: will this patient have an adverse event or serious GI bleeding that will require intervention in a time frame that mandates admission?  Problem is we don’t have a crystal ball.  What we do have are our clinical exam and workup to give us the pieces we need to figure this one out, with the help of a score.  

Getting back to that rectal.  The presence of melena on exam gives you a likelihood ratio of 25 for a patient having an acute upper GI bleed.  While the patient reporting a history of melena during the presentation has a likelihood ratio of 5-6, we can’t always take our patient’s word for it.  Other factors will increase the likelihood of an acute upper GI bleed, including NG lavage with blood or coffee grounds, and a BUN:Creatinine ratio > 30.  NG lavage with red blood, tachycardia, or a hemoglobin < 8 g/dL indicate a severe UGIB that may require urgent intervention. (1)  

Once we have our information, how do we risk stratify these patients?  One tool that has come out is the Blatchford score.  While I find many prediction rules to be overly complicated with too many variables, and often too simplified to answer a complex problem, I actually like this score.  Part of the reason is it really is just asking whether or not they have findings that would make me want to admit someone regardless.  A score of 0 indicates that you are low risk and can be safely managed as an outpatient.  The likelihood ratio of someone with a score of 0 having serious upper GI bleed is 1.4.  A score less than or equal to 2 increased that to 1.8.  When you get to scores of 6 or more, > 50% needed emergent intervention.  Based on a pretest probability of 30%, the likelihood of your Blatchford 0 patient having a severe upper GIB needing urgent intervention is < 1%.  I like those odds. (2)

You can calculate a score if you want, but in my mind, this is really more a checklist than a score.  I like to make it dichotomous: if you are a zero, you’re low risk and I can send you home.  If you’re more than zero, you’re not low risk anymore.  So here is my Blatchford checklist (patent pending?):

Hemoglobin > 12.9 g/dL in men or > 11.9 g/dL in women

SBP > 109 mmHg

HR < 100/min

BUN < 18.2 mg/dL

No syncope or melena

No past or present liver disease or heart failure

If the answer is yes to all of the above, you’re good.  Otherwise, they’re not low risk.  

So the final answer for our above patient?  She reported melena (increased risk).  She had melena on exam (Gives you a Blatchford score of at least 1 and makes her automatically higher risk).  She got admitted to the hospital.  

1.  Srygley FD et. al.  Does this patient have a severe upper gastrointestinal bleed?  JAMA. 2012 Mar 14;307(10):1072-9.

2.  Stanley AJ et. al.  Outpatient management of patients with low-risk upper-gastrointestinal haemorrhage: multicentre validation and prospective evaluation. The Lancet, Volume 373, Issue 9657, Pages 42 - 47, 3 January 2009