How helpful is NG lavage for diagnosing an upper GI bleed?

It has traditionally been taught that one of the first steps in managing an upper GI bleed is to place an NG tube to see if there are bloody aspirates. While it’s true that if you get a bloody aspirate, there is more likely to be an actively bleeding lesion, the data on whether the NG tube changes management is not too supportive.

In a Journal Watch post from 2011, David Bjorkman reviews a VA study of over 600 patients with suspected GIB. The study found that while placing an NG tube got patients faster care, it didn’t change clinical outcomes or rate of complications. He writes: “We already know that NGL cannot be used to exclude ongoing upper GI bleeding as one sixth of patients with active bleeding will have a negative NGL. Now, this study demonstrates that NGL results in no difference in a number of important clinical outcomes.”

However, if you are an ED provider, you may think differently about placing an NG tube. It may help you triage patients better in terms of who needs to be seen by GI first and whether there is evidence of a SEVERE bleed. Here is some evidence from the EM side of things.

What regimen can I give a patient going through opiate withdrawal? 

This is the typical cocktail our psychiatry consultants recommend:

  • Robaxin 500-700 mg q6h prn muscle cramps
  • Bentyl 50 mg prn abdominal pain
  • Clonidine 0.1 mg prn anxiety, hold for systolic blood pressure <90
  • Hydroxyzine 50 mg q6h prn anxiety or insomnia
  • Kaopectate prn diarrhea

Always make sure to check a tox screen to make sure they’re not going through concurrent alcohol or benzo withdrawal!

What can I give an opiate user or IV drug user for acute pain?

Imagine this scenario: you get a patient transferred from the surgery service. You glance at her home meds and with a flash of dismay, see 60 mg oxycontin BID with 10 mg oxycodone q4h PRN pain for chronic back pain. She just had hip surgery. What do you prescribe for breakthrough pain?

Or this: you are caring for a patient admitted for aspiration pneumonia. Although his breathing seems fine, he complains about the uncomfortable bed and requests something for pain. His social history notes that he currently buys Percs on the street and has used cocaine in the past. Initial doses of 5 mg oxycodone don’t seem to work, and he demands more and more. The nurses grumble about his neediness. What do you do?

I am, despite my bitterness and tough talk, an idealist and softie. I mean, I want to spend my professional life figuring out chronic abdominal pain and treating patients with malabsorption, dysmotility, and postsurgical anatomy. But in my daily practice, I have several fears when prescribing narcotics for inpatients:

  1. I’m feeding into an addiction. Giving them pain meds will make them worse off in the long run.
  2. They are taking advantage of me. They are lying to me or being nice to my face but laughing behind my back.
  3. When they realize they can get whatever they want, they’ll invent more excuses to stay.


Acknowledge these fears. Then let go.

What I’ve come to realize is that despite the patients who burn you, pain is real. Pain may stem from many different sources (deconditioning from being morbidly obese, psychiatric disease, dyspnea, social chaos, nerves that are wired differently and a low pain threshold). We may make us think, c’mon, there is NO way you are actually in that much pain…but they do need SOMETHING treated. Often people just don’t have enough insight to tell you exactly what that is. So your first formulation should be: is this actually pain? Is it anxiety? Is it fear at having to return to an abusive partner? Is it poor coping skills/functional status/dementia? Pain meds are not the only answer.

A final point: in those patients who want to be hooked on pain meds, make up stories, and manipulate the system, isn’t it sad that they would waste their lives trying to get addicted? A patient like this probably needs psychiatric help.

But sometimes you do need pain meds for people with complex pain histories. For instance, like patient #1 above, post-surgical patients have a really real reason to have pain: they just had freaking surgery! Many of your patients may have DEPENDENCE but not be ADDICTED. If you have a patient who is already on oxycodone, like #1 above, it’s a no-brainer they’ll require more than 2.5 mg oxycodone q6h.

NB: You should always make sure the whole team, including nurses, psychiatrists, pain specialists, etc are on the same page about an individual’s regimen. Here are some strategies I’ve seen used successfully:

  • Build trust with your patient and taper down gently. When you acknowledge the validity of your patient’s story and show that you care about them as a person, they are much more willing to listen to your suggestions–like, “how about we not do the extra IV dilaudid because I think you’re ready to try the pills? I really want to see you make progress.” In my experience this really separates the people in real pain from the people who have secondary gain.
  • Make sure you have a good non-narcotic foundation. Put people on standing Tylenol. Use flexeril if they specifically have spasms. Use lidocaine cream over painful joints. You wouldn’t believe how much relief a heat pack can give.
  • Use the appropriate meds for pain-like symptoms: benzodiazepines for anxiety, low-dose antipsychotics for agitation or insomnia, etc.
  • Start the patient on standing pain medications q3-4h instead of PRNs. Relying solely on PRNs is like putting a diabetic only on sliding scale insulin. If they ask for more, increase by 25% at a time until they are satisfied.
  • If one opiate isn’t working, try another. For some reason, different people respond differently to oxycodone or dilaudid or morphine…cross-tolerance is variable, but you can try using an equi-analgesic conversion.
  • I’ve seen methadone work well. It has a 4-8 hr half-life for analgesia (and 24-48 for withdrawal suppression). However it can take several days to titrate and you need to guarantee outpatient followup. If someone comes in already on methadone, you can give them extra narcotics if necessary but you must communicate any changes with their methadone clinic, otherwise, no dice.
  • I would not be so bold as to give known opioid abusers a PCA, but some people have found it works well.
  • Consider a social work consult (especially if you have a patient like #2 above) and offer substance abuse counseling or detox. If you know they’re going to snort or shoot up when they leave, consider writing a suboxone prescription.

References: a concise article from Anaesthesia
The school of hard knocks


Landmark Papers in Critical Care: the Rivers trial and management of sepsis

Rivers E, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368-77.

Question: In sepsis, it is critical to recognize the “golden hours” at which point systemic oxygen delivery cannot meet tissue demand –> global tissue hypoxia. Instead of monitoring vital signs and hemodynamic stability, can monitoring factors associated with tissue hypoxia–and titrating treatment to meet certain goals–improve outcomes?

The Rivers trial, at least in my mind, coined “early goal-directed therapy” in the management of sepsis. It put tissue hypoxia front and center.

263 patients who had 2+ SIRS criteria AND either a systolic blood pressure <90 that was unresponsive to one fluid bolus OR lactate >4.0 were enrolled in this trial. They were randomized into two groups. The control group got whatever the providers deemed to be “best treatment.” The intervention group was treated according to a protocol aimed at optimizing tissue perfusion:

Screen Shot 2015-07-27 at 9.48.06 PMThe resuscitation goals were:

  •  Central venous oxygen saturation (ScvO2) <70%, or add dobutamine to optimize cardiac index
  • Hct <30 were transfused RBCs until Hct was =30 or greater.

In-hospital mortality for the group receiving EGDT was 30.5% as compared with 46.5% in the group assigned to standard therapy (p=0.009). Furthermore, the EGDT group had higher ScvO2, lower lactate, and lower rates of organ failure. It should be noted that in addition to 99.2% of the EGDT patients meeting their hemodynamic goals, so did 86.1% of the control group, which speaks to the effectiveness of existing empiric therapy.

The authors made the interesting observation that the incidence of death due to sudden cardiovascular collapse was halved in the EDGT group, suggesting that an abrupt transition to severe disease is an important cause of early death. Early identification of patients proceeding down this route was important, as interventions that are started outside the “golden hours” may be too late.

There are a few controversies associated with the Rivers trial. One is that about one-third of the patients had CHF, so may have been in cardiogenic shock or mixed cardiogenic/septic shock, not pure septic shock. Thus, the use of dobutamine to increase the ScvO2 to > 70% in septic shock remains unclear. Additionally, the use of packed RBCs to a goal of Hct >30 to increase oxygen delivery remains controversial, since there are pros and cons (including cost, limited resources, transfusion reactions, and coagulation dysfunction) associated with using blood.

Bottom line:

When do you get an x-ray to evaluate an ankle injury?

At some point in your outpatient clinic, you will have a patient with an ankle injury. Maybe they tripped on the curb. Maybe they stepped wrong on a stair. Anyway, they’re here in your office, and the #1 thought on their mind is probably: is it broken?

Their ankle looks kind of swollen and bruised on exam, and they wince when you so much as put your fingertip on their skin. So do you get an x-ray? Or not?

Luckily, there are rules to guide us. The Ottawa Ankle Rules have a sensitivity of virtually 100% and okay specificity. This is what it boils down to: if they can’t bear weight on that leg AND they have pain around their ankle or the 5th metatarsal, get an x-ray.

What you do next depends on what the x-ray shows. If there was no fracture–only a sprain–I would have told her to move around to the best of her ability and RICE it. (RICE=rest, ice, compression, elevation) If there was a fracture, I would have made an urgent orthopedics appointment so that she could get a cast.