EBM: 4/2/2012

Welcome to the first Evidence Based Medicine blog entry, where we discuss the mounds or molehills of data for what we do in the ED. Are we practicing dogma? Or is there some science behind it?

For the first topic, I bring you…

To Lavage or Not – Are We Asking The Wrong Question?

It’s practically a cliché. A patient arrives in the ER at 1AM complaining of bright red blood and clots in his stools that started earlier today. He denies hematemesis or abdominal discomfort, and has no prior medical history, alcoholism, or similar symptoms previously. His vital signs are stable and he is in no acute distress. After confirming the guaiac positive stool and performing your ABC’s and IV’s, you page the on-call GI fellow to inform them of their latest customer. After establishing a collegial rapport, the fellow asks about the NG aspirate… What to do? Do you go back and sheepishly do an NG lavage? Blame your attending for not letting you do it? Blame the intern for not doing it yet? You think it over and instead decide to dazzle your consultant with your evidence-based knowledge on the topic…

Before we delve into the literature, it’s important to establish that we’re all asking the same question. The general inquiry, “is an NG lavage in GI bleeding useful?” is actually incredibly vague – by ‘useful’ do we mean, helpful in diagnosing? Risk-stratifying? In making a clinical treatment decision? A disposition decision? So, let’s focus. When faced with the issue of whether or not to do an NG lavage, the most important question to ask is if this procedure will help the patient. This means that by performing the procedure, your patient will either directly or indirectly have a better outcome than he would have without it.

If your patient is actively vomiting blood, then we all know what to do. The quandary of the NG lavage comes up when it’s not entirely clear whether the patient is having upper GI bleeding at all, whether it’s active, and when we believe a STAT endoscopy would be helpful and need to convince our consultant to do it in the middle of the night. So, the three main reasons to consider doing an NG lavage are:

  1. Identify upper GI bleeding in patients with melena or hematochezia without hematemesis
  2. Identify ongoing upper GI bleeding in patients who had coffee ground emesis or other prior hematemesis
  3. Convince your gastroenterologist to do an endoscopy faster

Now, let’s look at the literature – is there evidence that NG lavage is helpful for any of these? And, even if there was, does the result of your lavage ultimately lead to better patient outcomes?

(1) Does NG lavage identify upper GI bleeding in patients with melena or hematochezia without hematemesis?

Palmidessi, Sinert, Zehtabchi, et al. (yes, our guys!) published a Cochrane review in 2010 that addressed just this question. Although there were no systematic reviews or meta analyses on the subject, they did ultimately find three retrospective studies with reasonable methodology and reviewed their data on the diagnostic accuracy of NG lavage. All three studies showed a poor sensitivity (42-85%) and a relatively poor negative likelihood ratio (0.65 – 0.2) in ruling out an upper GI source. Given these results, a ‘negative’ lavage can’t be used to definitively rule out upper GI bleeding and all patients with clinically suspected lesions should have an endoscopy. But, what if there’s a positive GI aspirate? The data for that wasn’t great either. The positive likelihood ratios in these studies varied from 1.44 to 4.74 – a wide range that is likely due to varying methodologies as well as broad definitions of what “positive” means (frank blood? Specks of blood? Coffee grounds?). With a poor sensitivity and questionable specificity, it seems that NG lavage is at best an inconclusive test for identifying an upper GI source.

(2) Does NG lavage identify ongoing upper GI bleeding in patients who had coffee ground emesis or other prior hematemesis?

We’ve already seen that the sensitivity of NG lavage is not great – but maybe it’s of some use for patients who were having hematemesis and have since stopped. Should you do an NG lavage to see if they’re still bleeding? And, what does a ‘positive’ lavage in these patients mean? Aljebreen, et al. found that a bloody nasogastric aspirate  (not coffee grounds – frank blood!) had a relatively high specificity for identifying high-risk lesions (76%), which should prompt earlier endoscopy. On the other hand, their sensitivity was low (only 68% with a CI 57-78%), which means that a negative lavage shouldn’t be reassuring. Furthermore, the international consensus recommendations on the management of patients with upper GI bleeding emphasize early risk stratification on the basis of validated clinical criteria, such as the Glasgow-Blatchford score (GBS) and the Rockall risk assessment. These scales include clinical and laboratory criteria but make no mention of NG lavage findings. So what does this mean? If clinically your patient is at high risk for having a bad outcome and an NG lavage helps convince your gastroenterologist to come in faster, then it may be useful to facilitate their work up (i.e. – the middle of the night). There is also some utility in washing out blood and gastric contents to help with endoscopic visualization. However, the key to this question is that the decision should be made primarily on clinical grounds that have been well studied and validated. So, be familiar with the aforementioned scoring systems and be a good advocate for your patient.

(3) But I want the GI fellow to come now!! Will earlier endoscopy lead to better outcomes in my patient?

Huang, et al performed a thorough, retrospective analysis studying the correlation between NG tube placement and management. The study found that a bloody aspirate on NG lavage was positively associated with high-risk lesions as well as earlier endoscopy (CI, 1.09-2.04) but did not affect mortality. Furthermore, it did not affect other important measures such as length of hospital stay, transfusions, or surgery. Lim LG, et al. compared the mortality outcomes for upper GI bleeders who did and did not receive early endoscopy. They found that endoscopy within 13 hours of presentation was associated with lower mortality in high-risk patients but that it did not make a difference for low-risk groups. Furthermore, they used the GBS score to risk-stratify their patients, not an NG lavage. This brings us back to our original question – will early endoscopy improve your patient’s outcome?  If a patient is unstable or otherwise clinically high risk, they should have an urgent endoscopy, and if they’re not then a delayed approach doesn’t appear to affect mortality or other quality measures. Either way, an NG lavage doesn’t directly factor into the equation and its use just doesn’t pan out.

So, where does that leave us? What should you do? After reviewing the evidence, here are my recommendations:

  1. Use your clinical evaluation, judgment, and risk-stratification based on accepted and validated criteria such as the Glasgow-Blatchford and Rockall scores to decide whether your patient needs emergent endoscopy. This decision should be made together with your gastroenterology colleagues, but remember that it’s your patient – advocate for them.
  2. Treat your patient medically and provide supportive care – this includes blood as needed, PPI’s, octreotide and erythromycin if there is variceal bleeding, and platelets and FFP if there is an issue of coagulation abnormalities or if your patient is taking anti-platelet or anti-thrombin agents.
  3. Hold the NG tube!! NG lavage is a painful procedure and doing it won’t lead to an improved outcome for your patient. If you need to do it in order to convince your GI colleague to come in faster, try using clinical criteria first to make your case.

Ultimately, you must act in your patient’s best interest, and remember that it takes a long time for paradigms to change. If all of the above fail and an NG lavage seems to be the only way to get faster care for your patient, then do it… and then email your consultant a link to this blog.

References:

Palamidessi N, Sinert R, Falzon L, Zehtabchi S. Nasogastric aspiration and lavage in emergency department patients with hematochezia or melena without hematemesis. Acad Emerg Med. 2010 Feb; 17(2):126-132.

Aljebreen AM, Fallone CA, Barkun AN. Nasogastric aspirate predicts high-risk endoscopic lesions in patients with acute upper-GI bleeding. Gasrointest Endosc. 2004 Feb; 59(2):172-8.

Barkun AN, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med. 2010 Jan 19;152(2):101-13.

Lim LG, Ho KY, et al. Urgent endoscopy is associated with lower mortality in high-risk but not low-risk nonvariceal upper gastrointestinal bleeding. Endoscopy. 2011 Apr; 43(4):300-6.

Brennan MR, Spiegel MD, et al. Endoscopy for acute nonvariceal upper gastrointestinal tract hemorrhage: Is earlier better? Arch Intern Med. 2001 Jun;161(11):1393-1404.

Pallin DJ, Saltzman JR. Is nasogastric tube lavage in patients with acute upper GI bleeding indicated or antiquated? Gastrointest Endosc. 2001 Nov; 74(5):981-4.

 

 

The views expressed on this blog are the author's own and do not reflect the views of their employer. Please read our full disclaimer here. Any references to clinical cases refer to patients treated at a virtual hospital, Janus General Hospital.
The following two tabs change content below.

zinas

Latest posts by zinas (see all)

2 comments for “EBM: 4/2/2012

  1. jkhadpe
    April 4, 2012 at 1:49 pm

    Thanks Zina! I think this a great review of the literature concerning the use of NGT’s in GI bleeds. The evidence seems pretty clear that NG lavage is a poor diagnostic tool. I pretty much only place an NGT for patients that are actively vomiting blood to suck out that proemetic stimulus and hopefully make them feel better. Are others routinely dropping NGTs in their GI bleeders or if surgery requests NGT placement, are you putting them in? Just curious how others are approaching this topic.

  2. Ian deSouza
    April 12, 2012 at 5:57 pm

    I rarely place NG tubes for GI bleeding. The last time was probably in a patient who was intubated for airway protection first and required gastric decompression afterwards. I don’t believe it is useful in these cases. Performing an oftentimes brutal procedure (which you may not believe in to begin with) on a patient just to “placate/possible force the hand of” your consult seems unethical.

Comments are closed.