Saturday, February 16, 2013

You Want Me To Stick That Tube Where? - Utility of the Nasogastric Lavage

The Gist:  Upper gastrointestinal bleeds (UGIB) can be serious and scary, so diagnosis and treatment are essential skills in the ED.  Nasogastric tubes (NGT) are uncomfortable and are not necessary in many UGIB or potential UGIB.  A negative nasogastric lavage (NGL) doesn't mean much, but a positive one is pretty specific.  It may not be worth fighting GI over, but think about the patient and utility of the procedure first and know that the American College of Gastroenterology 2012 Peptic ulcer UGIB guidelines say it's not required for diagnosis, prognosis, or treatment (is airway a problem? patient uncomfortable?). 

The Cases:  
1. 23 y/o female with history of bipolar disorder and recent URI presented after "seeing some blood in my vomit" (further history sounded less like hematemesis and more like a spitting up a small bit of blood after a bout of coughing).  She took 600 mg ibuprofen three times daily for headaches over the past 3 days but had no other symptomatology. Vitals WNL, PE: unremarkable, Hb 13.5 g/dL, chemistry WNL.  

2. 63 y/o male with a history of atrial fibrillation and COPD presented with two episodes of hematemesis and more shortness of breath than his baseline. He was on warfarin, lisinopril, atorvastatin, carvedilol, furosemide, and a myriad of inhalers.  VS: BP 118/78, P 100, R 18, O2 94% RA.  PE: pale, no diaphoresis, abdomen soft, non-tender.  Hb 7.6 g/dL 

The attendings asked the nurses to place an NG tube - in Patient #1 to rule out UGIB and in Patient #2, because it's part of their routine for UGIB.

The outcome: Patient #1 - halted the NGT placement and left AMA.  Patient #2 - difficult passing NGT, NGT abandoned secondary to unsuccessful attempts, increased combativeness, and epistaxis.  His INR then returned at 5.6.  Patient was given vitamin K and admitted to GI for endoscopy.  

NGT placement, cited as one of the most uncomfortable ED procedures, resulted in patient distress in these two cases.  (If you don't believe this, I urge you to have a medical student place one on you...I 'practiced' with a fellow student and began using topical anesthesia after that event).   Many people in the FOAM world likely don't routinely use NGL; however, I found that in keeping with EM practice variation, many emergency physicians do so I opted to explore the utility of this procedure.  

Does NGL confirm UGIB?
  • If positive, yes.  Specificity 91% (95%CI 83-95%), +LR 11 (Witting et al)
    • It's not always clear "how positive" an aspirate is.
  • Negative aspirate cannot rule out an UGIB due to poor sensitivity 42% (95% CI 32-51%) (Witting et al). Recall that many UGIB occur in the duodenum, out of reach of the NGL.
  • Less invasive indicators in the history, physical, and lab work can point clinicians towards localizing the source.  See the Rational Clinical Exam from JAMA on this topic, demonstrating that melena, hematemesis, high BUN/Cr ratio, history of prior UGIB, and medication use are all good predictors.  So, if a patient clearly has hematemesis or melena, they really don't need an NGL to confirm UGIB.  
  • Counterpoint:  if a sick patient has rectal bleeding and signs/symptoms concerning for UGIB without clear stigmata (ex: hematemesis), NGL may point to a source easily and more quickly controllable with EGD.
Any other use for NGL in UGIB?
  • Historically, NGL has been used to predicting the severity of the bleed and need for intervention by determining the rate of bleed.  For example, coffee ground aspirate often indicates a slow bleed whereas bright red aspirate may indicate a more brisk bleed.
  • Most cases of UGIB warrant an endoscopy and the question lies in when this procedure takes place - now or within the next day?
    • NGL may predict high risk lesions, but performance of an NGL doesn't change clinical outcomes (Huang et al
  • The need or anticipated need to claim the airway of a patient with UGIB may warrant an NGT (EMCrit:  airway in the GI bleeder).
There are other ways to risk stratify UGIB..
  • The Glasgow-Blatchford Score (GBS) may have utility in risk stratifying patients prior to endoscopy.  The NICE guidelines allow for consideration of discharge for patients with a score of 0 but this isn't the standard yet in the US.   A simpler, modified GBS has also been developed and needs further validation for use in the ED to discharge patients.  Check out EM Lit of Note's perspective on this.  
Does NGL affect patient oriented outcomes?  No, Huang et al showed NGL made no difference in:
  • Mortality (OR 0.84; 95% CI 0.37-1.92)
  • Length of hospital stay (7.3 vs 8.1 days, P = .57)
  • Surgery  (OR 1.51; 95% CI 0.42-5.43)
  • Transfusions (3.2 vs 3.0 units, P = .94)
What about to clear the view for the endoscopy?
  • Erythromycin is just as good (ref)
If an NGT must be placed... do so humanely
  • Topical anesthesia
  • Whit Fisher's  NGT "Hook" technique (haven't tried this myself)
    • Spray phenylephrine/vasoconstrictor in nose and then viscus lidocaine
    • Clamp nose and have patient lie back.
    • Place NGT tip in oropharyngeal airway and dunk that in ice water for ~ 1 minute. 
The future...Small studies, but some interesting things being done:
Final Note:  So, the FOAM world is moving away from NGL and the literature also supports this.  The following pieces in key journals call for a move away from routine NGL in UGIB. This is an example of the crux of medical interventions - Are we doing something just to do it or because it will give us meaningful information and change the patient's outcome?
  • Pallin DJ, Saltzman JR. Is nasogastric tube lavage in patients with acute upper GI bleeding indicated or antiquated? Gastrointest Endosc. 2011;74:981–984. PubMed
  • Pitera A, Sarko J.  Just say no:  gastric aspiration and lavage rarely provide benefit.  Ann Emerg Med 2010 Apr;55(4):365-6.  PubMed
  • Anderson RS, Witting MD.  Nasogastric aspiration: a useful tool in some patients with gastrointestinal bleeding.  Ann Emerg Med. 2010 Apr;55(4):364-5.  PubMed
Major Studies ReferencedNB:  I haven't found any RCTs on NGL in UGIB and the major studies are in patients admitted to the hospital (high suspicion/obvious bleed or sick).
Witting et al
  • Retrospective records review of patients admitted through the ED with a code for GI bleed
  • n=220
  • Excluded patients with hematemesis
  • Retrospective, cohort of patients admitted to hospital with code consistent with UGIB
  • n=633 
  • Conducted a Propensity analysis (attempt to mitigate selection bias)


  1. Sorry, I kinda feel like I turned two pages at once, and at the risk of asking a blindingly dumb question...

    Is there anywhere in the world where it is actually standard practice to do NGT lavage as a diagnostic tool in patients presenting with upper GI haemorrhage?

  2. I know, rather dated practice. Putting an NGT in to diagnose UGIB is the standard practice of many US physicians. However, I cannot speak as to what fraction of providers still put an NGT in most or all UGIB. Not to say that this is the standard of care, but many many people are still doing it, mostly, from what I understand, outside of the academic setting (and what I've seen). On the EMCast from, November 2011, Dr. Amal Mattu et al discuss this as a "Myth Buster" because, well, there's that huge knowledge translation gap...we know we don't need to do something yet people still do it..