Friday, December 20, 2013

31 Year Old Male with RUQ Pain and a History of Pericarditis. Submitted by a Med Student, with Great Commentary on Bias!



This was submitted by a fantastic medical student who wished to remain anonymous:

A 31 year old male with a history of viral pericarditis one year ago presented with right upper quadrant pain. Here is his initial ED ECG:
The R-wave in V4 extends to 33 mm, the computerized QTc is 372 ms

















The only available previous ECG is from one year ago, during the admission when he was diagnosed with pericarditis:
1 year ago ECG, with clinician and computer interpretatioin of pericarditis



What do you think? What do these EKGs show? What is your plan for this patient?



Here was the story from my perspective, prospectively:
I was shown this ECG and told that it belonged to a 31 year old male with a history of pericarditis who has RUQ and back pain. Without seeing the patient, my interpretation of the first ECG was: likely normal variant ST-elevation (early repolarization), with a small possibility of pericarditis, and almost no possibility of acute coronary occlusion (STEMI). I was immediately suspicious that this patient had never actually had pericarditis in the past – it seemed much more likely to me this could simply be his baseline ECG which gets repeatedly misinterpreted as pericarditis, with more and more diagnostic momentum each time that he presented with his “history of pericarditis” and any nonspecific symptom near his chest.


So let’s take another look at the ECGs:
Sinus rhythm at around 70bpm. High voltage in the precordial leads. Obvious ST elevation is present in V2-V6, I, aVL, and II. There is obligate slight ST depression in aVR. The T waves are appropriate in size relative to the QRS. There is J-point notching in V4, V5, maybe V6, and II. There is T-wave inversion in III. If you care about S1Q3T3, there is a narrow Q in lead III with an inverted T-wave. It is fair to wonder whether the reciprocal T-wave inversion in III, with slight ST depression, is due to MI, but Dr. Smith says it just doesn't look right for MI.  There is PR depression in leads II, V4, V5, V6, and perhaps in other leads. The PR depression by my eyes is maximal in lead II, at about 0.5mm. There is obligate PR elevation in aVR, maybe also 0.5mm.  Such PR deviation is within the normal limits for atrial repolarization. Several leads do show the baseline sloping downward consistently after each QRS. In V6 the ST/T-wave ratio is something like 0.5mm/3.5mm = 0.14, depending on where exactly you measure. If we are considering anterior STEMI, the relevant values for the Smith formula (equation) are: STE in V3 60ms after J-point = 4mm, QTc = 372ms, and R in V4 = 33mm, for a resultant equation value of 15.974, obviously less than 23.4 and therefore highly unlikely to be STEMI.


Mostly the same findings and reasoning as the current ECG. The PR depression is slightly less, and the ST-elevation is slightly more in relation to the R wave amplitudes.



Here is the rest of the clinical information:

His vitals were within normal limits. His pain was located in the RUQ and curved around his chest wall to his right scapula. The pain began gradually during the previous evening, after the patient returned home from a 4-hour bus ride. He stated that his pain comes and goes, is made worse when sitting up and leaning forward, and is worse with inspiration. He said that his pain does not feel like his previous episode of pericarditis, and is not related to meals. He denied chest pain, shortness of breath, nausea, fever, chills, rashes, cough, and leg pain. However, the triage nurse wrote a note saying that he did admit to chest pain yesterday, which he vaguely endorsed when questioned again. He admitted to lifting some heavy objects a few days ago but denied trauma.

On exam the lungs were clear, there was no leg pain or swelling, and the pain with inspiration was accentuated when palpating under the right costal margin.  Heart and lung sounds were normal. There was no pericardial friction rub.


Bedside US for cholelithiasis was negative, and there was no pericardial effusion or wall motion abnormalities noted. CXR was negative. Amylase, lipase, and LFTs were also normal.  4 and 8-hour troponins were negative. We collected three more serial ECGs throughout his stay, all of which were identical to the presentation ECG.

Most of the team thought that this ECG was clearly diagnostic of pericarditis due to the PR depression with widespread ST elevation. At this point the team debated about whether to send him home with cardiology follow up, or to get a D-dimer (despite the fact that he was clearly PERC negative and Wells 0-1).

The D-dimer was elevated at 942, and the subsequent CT angiogram of the chest showed bilateral lower lobe subsegmental pulmonary emboli with a small right pleural effusion.

He was admitted for anticoagulation and workup for potential coagulation abnormalities. All of his coagulation studies were within normal limits, and his subsequent workup revealed no etiology of hypercoagulable state. He was discharged home on anticoagulation.

Review of old records showed no more definitive evidence of pericarditis (pericardial effusion, friction rub) during the initial “pericarditis” event. He had another ECG recorded at discharge, one day after admission:


The PR depression is even less. The ST/T ratio in V6, however, is slightly greater.


The final diagnosis on his ED note: pulmonary embolism AND pericarditis. Now, it is certainly possible that he had both recurrent pericarditis and simultaneous bilateral PEs, but this seems much less likely than having PE with a baseline ECG with early repolarization and a little PR depression. His physical exam findings, history, symptoms, ECG, and clinical course all support this as well. I worry that he will continue to be misdiagnosed with pericarditis in the future, and that one day when he has a potentially serious cause of chest pain his diagnosis and treatment may be missed or delayed because of this propagated error.


I feel that there are many excellent teaching points to this case:

1) It came to my attention during this case and afterward that there may be widespread misunderstanding in the ECG findings classically associated with pericarditis. Widespread PR depression (with reciprocal PR elevation in aVR) and/or widespread ST elevation are classically touted to be indicative of pericarditis, but what are the true diagnostic properties of such findings? How much PR depression should be considered indicative of pericarditis vs. a normal variant? We have seen many examples on this blog and others that ST elevation and/or PR depression can be normal variants:

Previous Posts on Atrial Repolarization:
Previous Posts on Pericarditis:

Causes of PR depression:
-Normal variant (atrial repolarization)
-Pericarditis

How much PR depression is there in pericarditis?
There does not seem to be good evidence to answer this question, unfortunately. It has previously been stated in this blog that 0.8mm of PR depression or greater may be somewhat specific for pericarditis (based on a 1973 Archives of Internal Medicine article by Charles et al, titled “Atrial Injury Current in Pericarditis”, which reviews even older literature and suggests that up to 0.5mm of PR depression is very common, while 0.8mm or more is suggestive of pericarditis).

How can we distinguish early repolarization from pericarditis using the ECG?
One potential way to differentiate the two is the ST/T-wave ratio in the lateral leads, usually V6. According to the following studies, a ST/T-wave ratio of 0.25 or greater favors pericarditis over early repolarization.



Keep in mind that ST segment and T-wave amplitudes must vary with lead placement, and we also know that the ST elevation of early repolarization can be dynamic, varying with many factors:



2) Cognitive errors pointed out by this case:

Anchoring bias/Diagnosis momentum:
His “history of pericarditis” undoubtedly influenced the perception and interpretation of his ECG. For most people it biases in favor of diagnosing pericarditis again. For myself (and I suspect other readers of this blog), however, it affected my perception in exactly the opposite way – it immediately made me suspicious of misdiagnosis and wary of the dangers of diagnosing pericarditis (and therefore missing other diagnoses).

Overconfidence inspired by scientific intrigue:
There is one thing that is potentially more worrisome than the widespread misunderstanding of the “classic” signs of pericarditis – the confidence with which they are misunderstood. We have all seen such blind confidence and contributed ourselves at one point or another. It is a flaw built into the nature of being human, and ECG interpretation can be a great example where this bias may occur. I will be the first to admit that it feels good to make a relatively rare diagnosis based on interpreting squiggly lines on a piece of paper. How much fun is it to take one look at a weird bradycardia with 3rd-degree block and unusual ST-elevation in V1-V2, in a room full of confused and concerned ER docs and cardiologists debating between taking the patient to the cath lab vs. inserting a transvenous pacer, to instantly know that the entire presentation is hyperkalemia and will likely respond to calcium??? The answer: it is almost blindingly intoxicating unless we can control our mind’s biases through metacognition. We must supervise our scientific excitement in order to keep it from blinding us to what is truly important.

In some situations when people are considering pericarditis on an ECG, it feels as though there is palpable excitement in the air: “Oh look at that PR depression with elevation in aVR! – classic for pericarditis.” Others are called in to look at the ECG, and everyone agrees: “cool case!” In these situations, we don’t just believe that it’s pericarditis – we want it to be pericarditis on some level. As doctors, it inherently feels good to think that we’ve figured out what is wrong with the patient; it feels good to think that our knowledge of the ECG was what made the interesting diagnosis. I believe that how we manage these feelings is what makes the difference in these cases. Do we unknowingly let that feeling control our thinking? Do we let it blind us from all future data that would refute our current thinking? Do we let it warp our perception of the physical exam, history, or lab findings? Or affect our course of management? Are we even able to perceive how the numerous known cognitive biases affect us in real time when there is a patient in front of us? Our minds seem to be designed such that, if we do not consciously attempt to understand the influences of our own thinking as we think, their influence will go unnoticed and their effect (whether helpful or harmful) will change our patient’s course.

When another resident on the team spoke up about the concern that this was not pericarditis, I witnessed many of these cognitive biases in action. The team told him without references that “any PR depression is very specific for pericarditis.” He pointed out that the patient’s pain is not located correctly for pericarditis and was described as being distinctly different from his previous episode of “pericarditis,” and that his pain was accentuated by sitting up and leaning forward, instead of alleviated – the team ironically replied that these were potentially unreliable clinical findings that were classic but not realistic in many cases.  It was pointed out (using examples from this blog) that his ST elevation was morphologically more consistent with early repolarization, with ST/T ratio < 0.25 in V6 and J-point notching. But it seemed they had already made up their minds to discredit or obviate any information that contradicted the diagnosis of pericarditis (the confirmation bias in action).

When you find yourself having to erase the stripes on your zebra to make it fit the clinical picture, maybe it’s time to consider the horses! And the opposite is true as well – if you find yourself having to paint stripes on your horses to keep yourself convinced of a supposed horse diagnosis, then you may have caught your own mind red-handed in a zebra-smuggling operation. More importantly, what do we care about most in emergency medicine? Amal Mattu would answer by saying, “lions, tigers, and bears” – the conditions which are life-threatening, regardless of their frequency. Which brings me to my next point:


3) The importance of doubting the electrocardiographic diagnosis of pericarditis in emergency medicine.
I believe that there are significant potential harms to diagnosing pericarditis, based on the ECG, in emergency medicine.

-STEMI can be misdiagnosed as pericarditis (in the words of Dr. Smith, “you diagnose pericarditis at your peril!http://hqmeded-ecg.blogspot.com/2013/05/24-yo-woman-with-chest-pain-is-this.html)

-Early repolarization can be misdiagnosed as pericarditis, causing the physician to prematurely attribute the symptoms to pericarditis, thus providing a decoy for any other cause of chest/back pain (dissection, PE, you name it) to go unnoticed and untreated.


As an aside, this case brings up many other important questions:
-How do we even know that anticoagulation is beneficial for PE? How do we know that the benefits outweigh the risks of anticoagulation?
-Does subsegmental pulmonary embolism matter?

For the first question I would refer you SMART EM’s incredible podcast on PE (http://www.smartem.org/podcasts/pulmonary-embolism-diagnosis-and-treatment) and for the second a recent article that may be of interest (http://www.ncbi.nlm.nih.gov/pubmed/23736701)



Take Home Points:
-Beware of the consequences of misdiagnosing any of the various important causes of chest pain as pericarditis

-The ECG can induce various cognitive errors, like any other piece of medical information, and the responsibility is on us to understand and monitor our own thinking to avoid these errors for the benefit of our patients



6 comments:

  1. Easiest way to sort this out is to obtain an EKG when he is asymptomatic. If EKG changes persist, then this is not a pericarditis EKG. Many, many thanks for your wonderful blog.

    ReplyDelete
    Replies
    1. Thanks, Jason. Unfortunately, they did not have this luxury at the time of evaluation. Also, early repol does change from EKG to EKG, so it is not necessarily so that a "baseline" ECG will make the diagnosis. Probably, but not definitely.

      Steve Smith

      Delete
  2. Cases like this keeps us docters humble and paranoid (better now a touch of paranoia, than later depression). Very interesting and educational case!

    ReplyDelete
    Replies
    1. "only the paranoid survive!" (an old quote of the former CEO of Intel). For us: "only the paranoid do the best for our patients!"

      Delete