08 February 2010

You can't win 'em all

Seems like I've been on a real run of chest pain patients lately.  Which is fine -- it's part of the gig.  I did have a very interesting pair the other night.  They were seen in sequence, right next to one another, in room 7 and room 8.  They were both totally healthy woman in their mid-fifties.  And they were both over-the-edge, crazy, crawling-out-of-the-gurney anxious.

Anxiety is an awful red herring in the work-up of chest pain.  People who are having an anxiety attack often if not always manifest some chest pain (pressure, tightness, whatever) as a prominent symptom of their anxiety.  On the other hand, someone having a heart attack who is experiencing chest pain will also be anxious -- and for good reason!  I probably see ten patients with simple anxiety for every one patient I see with "real" chest pain, so just by probabilities and pattern recognition an ER doc might easily dismiss the anxiety cases, just blow them off.  Which would be a real mistake, since when you blow off these cases you absolutely will miss things.

So you have to pick and choose what you will work up and how far you will chase the zebras.  Most of this you are going to do by Bayesian analysis: looking at the patient's age and general state of health, incorporating some basic data like ECG, blood tests, and how well the patient's history fits with classic patterns of disease.  So the 25-year old who is hyperventilating and hysterical because her boyfriend was hitting on another girl probably will get minimal work-up, whereas the septuagenarian with a history of diabetes who has an elephant standing on his chest is a slam-dunk admit.   Those are the easy cases -- the extremes of probability.  We joke that it's the cases in the middle that are why we get paid the big bucks.

These two cases were real doozies as far as figuring out what to do with them.  The first patient was, as I mentioned, about 55 and completely healthy.  She had this sharp pain that had been migrating all over her chest for several days.  It was not exertional, nor was it associated with shortness of breath, fever, cough, etc.  She sometimes felt it between her shoulder blades, and sometimes it was worse with a deep breath or movement.  it was a very fleeting pain, and clearly seemed related to some situational stresses she was having at work.  ECG, labs and chest x-ray were completely normal, as was her exam.  The pain in the back did catch my attention: that can be a historical element associated with Aortic Dissections, a vascular catastrophe in which the aorta basically tears itself apart.  So I did a CT scan on her to evaluate the aorta, which was normal.  After a standard cardiac observation was completed, she went home with some xanax.

The second patient was even stranger in her presentation.  Her complaint was listed on the triage note as "chest pain" but she started telling me about this tooth that had been bothering her, and she had bitten down on something and it had gotten much worse.  The tooth pain was radiating over the top of her head and also down her neck into her shoulder and back.  She was so anxious that she could barely get out a coherent sentence and she would hyperventilate herself into vomiting.  She complained that after vomiting she felt a burning pain in her mid-chest, which was why she was billed as "chest pain."  It sounded pretty clearly esophageal.  As I took in this history, I wondered whether I should even work this up at all as chest pain, but the ECG and troponin had already been done (from triage) so I decided to roll with it.  She also had a normal exam, and felt much better after some ativan.

She slept through most of the night shift after that.  I would rouse her from time to time, and other than a deep conviction that she had something terribly wrong, she actually seemed to feel much better.  Never complained of any more pain.  I considered scanning her, but mindful of the one "wasted" un-indicated CT I had just done on room 7, I was feeling a little gun-shy.  I hate to shotgun tests, and this one seemed even less useful than the previous.  Eventually, I figured that "you just cannot scan everybody," and I put her in for the cardiac observation and discharge protocol, like the other woman.

By now, you probably can see where this is going. The next morning, when the patient's observation period was done and she was set for discharge, she still didn't feel right.  Fortunately, an alert nurse spoke up and asked the new doctor on shift to re-evaluate the patient.  He did, and really didn't see anything much different, but out of a general sense of "I should probably do something," he ordered a CT scan.

And it showed:

Yes, she had a Type A aortic dissection.  Quite lethal when not treated.  My partner reported that when he got the call from the radiologist, he got all sweaty and weak in the knees, it was so unexpected.  Fortunately, the diagnosis was made and she went to the operating room for repair.

I've reviewed this case with our medical director, and his words were, "If I had taken care of this patient, she would be dead, because I doubt I would have even kept her for observation."  I don't feel that it was a "miss" so much as a "Damn!  Who'd've thunk it!"  But still, the irony is maddening that I had two weird, anxious ladies and I picked the wrong one to scan.  Crap.

There are a number of good learning points to take from this "near miss," however.

1.  Listen to the nurses.  In my opinion, the hero of this story is Kathy, whose gut told her something was wrong and she brought it back to the doc.  Woe unto the physician who blows off a nurse in this context.  Whether or not you miss this case, if you disregard their advice, it's all the less likely that he or she will be willing to stick their neck out and ask for a re-evaluation in the future.  Nurses are so much closer to the patients that their input will save your ass if you're smart enough to listen.

2.  Beware sign-outs.  Most ER docs are reluctant to get involved with a patient dispositioned by a previous doc.  I get it -- who wants to re-open Pandora's box? But like it or not, they are your responsibility, and sometimes a fresh pair of eyes/ears are all that is needed to unlock the puzzle. 

3.  Keep an open mind.  In this case, I admit that I was a little annoyed at this patient for her strange behavior, and just for the injustice of the universe at subjecting me to this sort of thing, and that bias would have made it difficult to really re-evaluate her with an open mind.  I hope I could have, but I'll never know in this case.  My partner was able to do so, and that made the difference between getting the diagnosis and missing it.

4.  Dissections are strange beasts.  I've seen several now, and none of them had the classic presentation.  I've learned to respect the isolated neck/back pain, and, on reflection, the incredible anxiety tone of almost every dissection I have seen is a remarkably consistent feature.  Having a high index of suspicion is essential.

5.  D-dimers are useful to screen for dissections.  I did not originally order a d-dimer on this patient, but it was positive in retrospect.  It appears that the majority of cases of dissections do have a positive d-dimer, which makes sense when you think about the physiology of the study.  While the correlation does not seem to be strong enough to use d-dimer solely as a test to exclude dissection, it does appear to be useful as part of a rational strategy to determine which patients you might choose for further investigation.


  1. Dude - If she were my pt, I would be thanking whatever gods I could that the morning doc scanned her. I'd rather be good than lucky, but holy crap! I have scanned so many chests that turned out to be negative, I don't even know what my decision process is anymore!

  2. One more thing - had an old guy the other night that I would have bet my house had an aortic catastrophe: chest and back pain, diaphoretic, hypotensive, bradycardic - looked like crap. Sent him for a stat CT of chest/abd/pelvis - nada. Big red was OK....but his lipase was through the roof! Whaddya gonna do?

  3. Great post and I heartily agree with all the take home messages.

    The worst dissection I've ever seen (that survived) was a surgeon who presented with calf/foot pain...go figure...from the dissection compromising his peripheral circulation to the leg/foot.

    He had been skiing all day and if he hadn't been a surgeon we probably wouldn't have done the workup we did.

    Glad your case turned out as it did!

  4. It's nice to hear that MD's listen to the little people some times. I brought in a severe shortness of breath case to the ED once and the MD asked me if the patient was going to need to be intubated. I told him yes. He immediately placed and order for an intubation tray and RT and then went and did the assessment. The patient ended up coding an hour later. I really appreciated a doctor that respected my clinical skills even if I'm only a medic.

    Great post.

  5. Another take away is that we should always attempt to treat patients as statistically independent of each other, unless we have clear indications that they're not. These two patients appear to have presented statistically independent of each other, but your treatment made them statistically dependent. I wonder what would have happened if you had seen the two patients in reverse order.

  6. we should always attempt to treat patients as statistically independent of each other

    Ha! Great point, and totally correct!

  7. Hey my Twitter friend, my first time to your blog- most excellent! Bravo to you!

  8. These are the cases that affect the decision-making process of all of us who hear about them. Your first commenter says it all... What bugs me about the current health-care debate is that 99% of the media doesn't understand how imprecise medicine really is. As long as we have an environment where we cannot miss a diagnosis (no matter how improbable), our health care will be expensive. Naive writers like Ezra Klein (who I wish you would stop positively referencing) don't get this.

  9. As an aortic dissection patient and survivor, it is so interesting to read this blog and comments!
    Yes, you do have to listen to the patients. I have heard that we say different things than a heart attack patient. I know I felt a "burst in my chest" while driving. Nice, eh?. When the ambulance got me to the hospital, I was being treated for a heart attack. The pain continued to be pretty unbearable, even with pain killers. The heparin was on it's way and my Emergency room Doc was looking at my x-ray, but not looking at it. His "gut" told him to get back in the room and check for these other things, or they could lose me. He told me what he thought it might be, and moved aside as I went up to a CT scan. With the changes being made at MHI in MPLS, that CT happened right away. I had dissected to my legs, down my left arm and up my right carotid artery. Wow...in about an hour...
    Needless to say, when returning to my room, my husband said it looked like a swat team was there. They were preparing me for surgery.
    I write this because prior to being a key note to cardiovascular physicians at a symposium last year, I was worried about what I wanted to say. Kind of like preaching to the choir. I was encouraged to do it because apparently, physicians don't talk to many patients who survive.
    My ER doc said this to me " If I had stayed with my first diagnosis and let my ego get in the way, you may not be here today. Listening to patients is crucial and following through on my gut to change direction was what I needed to do at the time." Needless to say, if treated for a heart attack, I may not be here to comment on this blog.
    Keep listening to your patients. It is what saves us. Saves us enough so that we can continue our lives as moms, wives, gramas, friends, singers, etc. Also....lead us down paths to help spread awareness on this thing called "Aortic Dissection"..
    It is a new mission for me, for sure. That and heart health awareness. I have a mechanical valve and a partial mesh aorta, but as you know, I am still dissected. I don't live like I have that at all. I take really good care of my self but my attitude is one of " how much life can I live today?".

    This is a thank you to all of you who save people's lives. Thank you for listening to us. Thank you for using your gut as much as your wisdom. Than you for admitting things regarding certain diagnosis that could have been better. We are all here to learn....
    I am just so glad I get to be here because of the brilliant work being done in the medical community.
    Thank you...and keep it up.
    Patty Peterson

  10. until failure to diagnose dissection can no longer land us in court, we can and will scan everybody. or until we as a nation go bankrupt I guess.

  11. Well, there you have it.

  12. I remember seeing a patient late at night whose only complaint was left earache. The only reason I did was because I was near the consulting rooms, and I just happened to do an ECG after finding nothing else wrong, and stumbled on a HUGE inferior myocardial infarct.

    Patients clearly don't read the textbooks!

  13. Aortic dissection = Very Scary Thing. Points about medicine being imprecise and about seeing patients as statistically independent are both good points... boy, not what you want to think about if you're the patient!

  14. Dr. S, I think I sent you the link to that Japanese study on certain kinds of tissue damage within the body (i.e. rupturing aortas) causing elevated levels of the neurotransmitter dopamine in the amygdala region, which as anyone who's suffered an anxiety disorder will tell you (if he or she reads), makes one feel ill at ease and will even cause a sense of impending doom. The subject will fidget, feel extremely anxious without being able to pinpoint exactly why, and have an overall sense of being threatened. It was observed as such a marked behavior in pre-dissection patients, the researchers recommended adding anxiety itself to the list of diagnostic symptoms for aortic dissection. Very interesting, I think, how our brains are so exquisitely tuned chemical-wise. I wonder how this would in turn serve the survival instinct--what would an animal do if indeed there was a routine signal that its aorta was about to blow?


Note: Only a member of this blog may post a comment.