Few patients who come through our doors have the potential to crump in quite the quickly catastrophic way that those with upper gastrointestinal bleeding (UGIB) can and sometimes do. Take a pinch of active hemorrhage, swirl it together with poor underlying protoplasm (who needs a functional liver anyway?), and add a soupçon of difficult source control, and you have a recipe for disaster.
Cases like these make it incumbent on us as emergency physicians not to miss a trick as we resuscitate and treat these patients: Get large bore IV access, have blood on standby, secure the airway if indicated, and get GI involved early. Well-intentioned providers can also debate the utility of proton pump inhibitors and octreotide in the ED, though I personally think the evidence argues against their use. Giving antibiotics to cirrhotics with upper gastrointestinal bleeding (UGIB), in contrast, has an NNT of 22 for mortality and 4 for infection prevention, making it a demonstrably beneficial treatment. (Cochrane Database Syst Rev 2010 Sep 8;:CD002907.)
One intervention that you might have noticed was missing from the preceding paragraph: nasogastric lavage. NG lavage was considered by many back when I was training to be a routine part of the workup if a patient walked in the door with hematemesis or melena. Asking about the results of the NG lavage may be the first question from your local gastroenterologist, and their reaction if you haven't performed one may be akin to that of a cardiologist if you called a code STEMI without actually ordering an ECG.
But why the outrage? Does NG lavage have any diagnostic value? Is it necessary for successful endoscopy? Does it result in any improved patient important outcomes? Even if it doesn't, it makes our consultants happy, so is there any reason not to just get it done? The answers are no, no, no, no, and hell yes, respectively. Let's tackle each question in turn.
Can NG lavage rule in or rule out acute UGIB? A review article in JAMA says if you get back blood or coffee grounds from your lavage, you have a +LR of 9.6, which isn't too shabby. (JAMA 2012;307:1072.) Unfortunately, this seems to represent an absolute best-case scenario because these data come from a single center, retrospective cohort study. A review published in Academic Emergency Medicine in 2010 was more extensive and reveals much less impressive results.
Nasogastric Aspiration and Lavage in Emergency Department Patients with Hematochezia or Melena Without Hematemesis
Palamidessi N, Sinert R, et al.
Acad Emerg Med
This review found three studies where all the patients who had NG lavage also received endoscopy as the gold standard for diagnosis. Their results: sensitivity 42-84%, -LR 0.2-0.61, and specificity 54-91%, +LR 1.4-4.7. That low sensitivity isn't an outlier; it's consistent with what they found in the JAMA review article.
What does this mean for us? A positive NG lavage may help confirm the diagnosis, but even that's not something you can take to the bank. (Weird to think about false-positive lavages, but the evidence shows that they do indeed happen.) And a negative lavage absolutely, positively does not rule out UGIB. Don't believe me? Even the American Society for Gastrointestinal Endoscopy guidelines say that the “absence of blood ... does not exclude the presence of active UGIB.” (Gastrointest Endosc 2012;75:1132.) To be clear, a patient could be hemorrhaging into his stomach and be on the brink of crashing, and the NG lavage might come back negative. This is simply not a test we can rely on in this high-risk population.
So NG lavage has little to no diagnostic value, but does it result in better endoscopy (visualization or ability to treat) for our colleagues when they get in there to take a look?
Erythromycin Infusion or Gastric Lavage for Upper Gastrointestinal Bleeding: A Multicenter Randomized Controlled Trial
Pateron D, Vicaut E, et al.
Ann Emerg Med
This multicenter study randomized 253 patients with UGIB to receive IV erythromycin, NG lavage, or both prior to endoscopy. What did they find? No difference for satisfactory visualization among the groups and no difference in duration of the endoscopic procedure, rebleeding, the need for a second endoscopy, blood transfusion, or mortality. A safe IV medication can obviate the need for a painful intervention? Done and done.
If NG lavage doesn't help with diagnosis or improve visualization, does it have any effect on any other outcomes that might matter to our patients? If only there were a study designed to measure the impact of NG lavage on outcomes in acute UGIB. Oh, wait, actually there is.
Impact of Nasogastric Lavage on Outcomes in Acute GI Bleeding
Huang ES, Karsan S, et al.
Unlike some of the articles cited above, this is a GI study, done by GI docs, and published in a GI journal. This retrospective review of 632 patients at the Los Angeles VA found that getting an NG lavage in the ED did result in patients getting endoscopy a little bit sooner, but the results were pretty underwhelming. About 10 percent more patients had EGD performed at the 10-hour mark in the lavage group. That's nothing to write home about. What about the stuff that really matters? Mortality, need for blood transfusion, need for emergency surgery, and length of hospital stay were the same in both groups. Going 0 for 4 on important patient outcomes gives us another swing and a miss for our poor misbegotten NG lavage.
“OK, OK!” you say, “But what's the big deal? If doing the lavage makes my consultant happy and saves me from arguing over the phone, who is it really hurting?” It turns out we are hurting the patient, and maybe more than just a little bit.
Comparison of Patient and Practitioner Assessments of Pain from Commonly Performed Emergency Department Procedures
Singer AJ, Richman PB, et al.
Ann Emerg Med
This article looked at procedures performed on more than 1,100 patients, and had the patient and the practitioner rate how painful the procedure was (or looked) using a visual analogue scale score. Among the many procedures evaluated were ABGs, LPs, I&Ds, Foley placements, and fracture reductions. Care to take a guess which procedure was rated the most painful by both groups? That's right: NG tube placement was the winner and clear champion for things you would least like have done to you or have to do to someone else. Many great physicians and nurses have tricks and tips for making an NG tube go down Mary Poppins-style, but the decision is clear in this case if you're balancing the discomfort of annoying your consultant vs. torturing your patient.
Summing Things Up
- NG lavage has poor specificity and even worse sensitivity. Its diagnostic value is about the same as that Sacagawea dollar coin burning a hole in your pocket.
- It doesn't improve any important patient outcomes.
- It hurts. A lot.
I'll leave you with a quote from an editorial in the journal Gastrointestinal Endoscopy, “The jury is no longer out. NG lavage does not help patients in the emergency department with acute upper GI bleed.” (Gastrointest Endosc 2011;74:981.)
I couldn't have said it better myself.
All about Myths
This column will highlight dogmatic teachings in emergency medicine, beliefs that have been passed down from generation to generation of physicians. The authors will look at the evidence and research behind them and determine whether that information holds true.
The writing team for the column is comprised of Anand Swaminathan MD (@EMSwami), Natalie May, MD (@_NMay), Daniel Runde, MD (@Runde_MC), and Rory Spiegel, MD (@EMNerd_).
Have a practice, technique, or regimen you want them to analyze? Send it to firstname.lastname@example.org.
Share this article on Twitter and Facebook.
Access the links in EMN by reading this on our website or in our free iPad app, both available at www.EM-News.com.
Comments? Write to us at email@example.com.Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.