Case No. 13: The Devil is in the Details

A 63 year-old white female with multiple medical problems presents to EMS via direct call from her nursing facility with complaints of worsening respiratory distress since awakening this AM. The patient’s history includes IDDM, hyperlipidemia, morbid obesity, HTN, and supplemental oxygen dependent COPD. The patient has had several previous MIs, suffers from CHF, and has refused consultation for bypass grafting after a catheterization six months ago revealed advanced CAD.

The patient is found sitting on the edge of her bed in tripod position, diaphoretic, globally cyanotic, audibly wheezing, tachypneic at 32 breaths per minute with “one-word” dyspnea, and markedly agitated. Lung sounds reveal minimal tidal exchange, diffuse high-pitched wheezes, and faint rales at the mid-thorax where her breath sounds disappear. The heart rate is 130, the SpO2 86%, and the BP is 190/90. She denies chest discomfort, nausea, or recent illness.

Prehospital rhythm strip and 12-lead EKGs

I remember explicitly thinking to myself that the diagnosis not to miss here was exacerbation of CHF due to new MI. The patient was hypoxic, agitated, and becoming combative. I glanced at the first 12-lead and I thought, “this is non-diagnostic, not a cath-lab activation,” and STEMI disappeared from my mind. The patient was treated with nitrates, broncho-dilators, aspirin, and furosemide; on arrival in the ED, she could speak in complete sentences and her respiratory function was significantly improved.

EKG obtained on arrival in the ED. Subtle ST elevation and deepened Q-waves are present in III and aVF. These findings are new by comparison to the 2007 tracing below, as is the additional ST depression present in aVL and I; new T-wave inversions are also seen in V2 and V3, as well as deepening of the inversions in I and aVL. The patient’s 2007 EKG showing an old inferior infarct, consistent with the previous cath report.

When the ED physician confronted me about missing the STEMI I was incredulous. Even the higher-quality hospital 12-lead seemed to me ambiguous for acute MI. If “you can’t make the diagnosis if you don’t think it,” how could I have missed this STEMI? The quality of the tracings was poor: the baseline wanders; there is substantial movement artifact. Yet the question of STEMI is there: consider the ST depression and T-wave inversion in aVL, I, V5, andV6—there are even hints of ST depression in V1 and V2. Minimal but significant ST elevation can be discriminated in III and aVF. Interestingly, the rhythm strip reveals the elevations in II and III most dramatically.

So what happened?

Medical error is responsible for substantial morbidity and mortality even in today’s climate of patient safety awareness. I want to use this case, in which I missed a critical diagnosis, to discuss cognitive error in clininical decision making with respect to electrocardiographic diagnosis.

Numerous taxonomies of cognitive error have been described; I will highlight several categories which I believe are of particular relevance in electrocardiography.

Premature Closure. This occurs when a clinician makes a rapid, confident diagnosis, often based on prior personal experience, and subsequently ceases to collect additional data or re-evaluate the initial diagnosis in light of new findings. Once I had satisfied myself that the EKG was non-diagnostic, I closed my mind to the possibility of revising this assessment; I made no attempts to improve the quality of the tracings or reconsider the presence of subtle indicators of AMI. The value of serial EKGs even when the initial 12-lead is normal is uncontested; simply because a patient is showing no signs of STE at one juncture does not mean that the EKG will again be negative later on—this is particularly relevant if there is a change in clinical presentation or symptomology. Tom Bouthillet has several superior case presentations highlighting this phenomenon.

Diagnostic Anchoring. This takes place when a clinician clings prejudicially to an initial diagnosis even when new, conflicting data surfaces. I made a decision that my patient was not having a STEMI; even in light of the more obviously pathological and less artifactual hospital 12-lead, it was difficult to unmoor myself from my misguided diagnosis.

Conformation Bias. This occurs when once a diagnosis has been formed the clinician proceeds to only attend to data which support or reaffirm the initial impression; conflicting evidence is often trivialized or explained away. Once I had anchored myself to the diagnosis of a “non-diagnostic EKG,” all abnormalities of subsequent 12-leads were taken as further non-specific evidence rather than viewed as potential revisions of the initial impression.

Framing. This occurs when demographic or prejudicial stereotypes cause the clinician to dismiss or trivialize certain diagnoses based on a-priori judgment rather than clinical assessment. In electrocardiography this is sometimes encountered in the context of relatively positive stereotypes: the patient is too young, too healthy, or too asymptomatic for their ST-segment abnormalities to be due to coronary occlusion. In these cases, the “healthy person” frame results in the exclusion of an important differential. Alternatively, a “sick person” frame can result in inappropriately minimizing acute results: “Given this patient’s complex medical history, this grossly pathological EKG is probably normal for them.” Most frequently, however, framing errors likely result in failure to perform a 12-lead in the first place rather than misinterpretation of an EKG result. Protocols can help with this: some systems have policies which mandate a 12-lead EKG for all patients complaining of certain symptoms or presenting with certain histories. While the sentiment is not unwise in certain respects, it is sometimes said that one should “treat the patient, not the monitor.” Framing a patient by their symptoms alone can be a grave oversimplification. Rather, a responsible assessment will address all components of a clinical scenario in appropriate proportions.

Finally, I believe the issue of distraction has some application here. “Distracting injuries” are dramatic, attention grabbing foci which can divert either the clinician or the patient from observing an often more serious but more subtle finding. In this case, I was distracted by the patient’s clinical acuity. Yet even in non-bedside electrocardiography a profoundly obvious finding such as dramatic ST-elevation can distract from more subtle diagnostic features such as chamber enlargement or conduction abnormalities. In the previous case as well as the case series of October 2010, there are significant derangements of rhythm together with ST-elevation. It can be difficult to adhere to an unwaveringly systematic approach to EKG diagnosis when a finding such as STEMI is jumping out at you. In radiology, there is a cognitive forcing strategy used to address this phenomenon: “If you see a fracture, look for another one!” There may be a role for a similar forcing strategy in electrocardiography.

This is a subtle case, yet it illustrates how one diagnostic oversight can lead to a cascade of cognitive errors. Although the ED physician activated the cath-lab after viewing these EKGs, the patient ultimately made an informed refusal of PCI and again restated her wishes to not be evaluated for bypass grafting. Due to the absence of chest pain, she ruled out for lytic therapies and was subsequently transferred to the ICU for continued observation and conservative management. A positive troponin was returned 4 hours after ED arrival. Laboratory assays were notable for elevated BNP, Glucose, BUN and creatinine. A review of the prior catheterization in 2007 indicated severe ostial RCA disease with a total distal occlusion combined with a totally occluded LCx and moderate, diffuse LAD disease becoming more severe in the distal portions.

This case involved a critical error on my part which resulted in failure to activate time-sensitive resources. Under other circumstances this oversight could have cost the patient her life. I carry the memory of my errors forward.


6 responses

  1. The first 12-Lead certainly doesn’t look good, but it doesn’t look new-STEMI. Old Q-wave MI for sure, but given the patient’s history it is to be expected. aVL doesn’t look right, but even still I wouldn’t jump on STEMI. As for the 3-Lead showing ST-E, being acquired in monitor mode you shouldn’t rely on it (unless it was diagnostic, but it doesn’t look like it was).

    I think you did the right thing treating the CHF exacerbation. Trending 12-Leads compared with an old tracing showed enough of a change to give the ED doc the ammunition needed to call the STEMI.

    I guess I’m not sure how you could have done it any better in the field!

    August 3, 2011 at 1:09 pm

    • Christopher, thank you for your comments and insight! I am inclined to agree with much of what you have said– certainly this is a highly ambigious case and although the 3-lead should have sent up a red flag, it is definately non-diagnostic. In retrospect what concerns me most about this case is the fact that STEMI was on the table and an assiduous effort was not made to aquire better tracings or establish trending in the pre-hospital setting. Given the clinical history and the pre-hospital 12-leads as they are, I do not think it would have been wise to activate the cath-lab. However, the cath-lab question was in my consciousness and I did not optimize my response to it; I turned away and lost focus. I made an imperfect judgment based on imperfect data and never rethought the issue.

      Thank you again for your thoughts!

      August 3, 2011 at 5:06 pm

  2. VinceD

    I’m very much on the fence about whether this is actually a STEMI. Even in retrospect I cannot convince myself to call it.

    True, the ED ECG, when compared with the old ECG, does have a significantly larger Q-wave in III and therefore some axis shift (and correspondingly what I may argue are secondary ST-changes from the changing QRS vector), along with some anterior T-wave inversions, but from what I can piece together your prehospital tracings very much resemble the old ECG from 2007. Compared with the ED ECG, in your tracings the inferior Q-waves are significantly smaller, there are no flipped T’s in V1-V3, and there is also no R-wave in V2 even though they were only taken a short time before the emergency department ran their tracings. I’m left wondering to what extent the changes on the ED ECG were due to lead placement because I highly doubt the patient “Q’d out” in such a short time.

    Also, it’s very hard to call the baseline and J-point when the patient is both tachycardic and moving, and even the ED ECG is not quite perfect. I am very wary of using the PR-interval as the baseline because PR-depression from atrial repolarization is such a common finding and can give the illusion of ST-elevation, however in this case it seems to be our only option. Now it is not a topic I have done much study of and can qualify with exact references, just something I’ve gathered from a couple texts and viewing hundreds and hundreds of ECGs, but without a usable TP-segment I am left wanting.

    I’m also very curious just how the ED physician determined this was a true STEMI. With no cath, we’re really left in the dark. You cite her abnormal lab results, but all of those, including the troponin, could be elevated chronically due to her multiple medical problems. Additionally, the acute pulmonary edema, exacerbation of CHF, and demand ischemia (which are all much easier to diagnose than STEMI in this patient) will elevate the troponin further, so unless the level was quite markedly elevated, we really don’t know if there was an acute occlusion.

    In the end I agree with your treatment and how you handled the case, and really don’t think there is much to be criticized. I applaud you for using it as an opportunity to examine how we make mistakes and improve your practice, but I don’t think it is something you can fault yourself for. Thanks for sharing!

    August 18, 2011 at 7:46 am

    • Vincent, thank for your detailed and very considerate assessment of this case. I appreciate your experience here and your willingness to look into this matter so closely. With regard to the EKGs, I have been back and forth on this for several years and even now I remain, as you seem to be, skeptical that activating the cath lab would have been most wise in this setting. The electrocardiograms are extremely subtle and invite justifiably conflicting interpretations. As I recall the troponin was as you correctly surmise a “low positive” and likewise may have invited attribution error of another variety. The diagnosis of record for both the EM consultant and the attending cardiologist was STEMI. Nonetheless I am reminded that, “often the most difficult diagnosis to make is one other than the diagnosis the patient already has.” Perhaps the entire story of premature closure, diagnostic anchoring, and attribution error could be told from the EM physician’s actions forward. Even the details of this case– clinical picture, EKGs, and lab results– do not add up to a definitively convincing picture.

      I want to thank both you and Christopher again for your kind attention and thoughtfulness in commenting on this case.

      August 18, 2011 at 9:26 pm

  3. Just stumbled onto your blog while looking for some hyperkalemia images. It looks good, and I can tell that you put a lot of research into each post.

    About this case and ECGs – stop flogging yourself! Prehospital activation is great for the clear-cut cases, and it speeds things along nicely. For the murkier cases, however, you need the emergency doc and all the resources of the hospital to figure things out. Without old ECGs, cath reports, bedside echo, you guys in the field are at a distinct disadvantage with regard to diagnosis. Sometimes all you can do is set the ED up for success, with multiple field ECGs, responses to treatment, and advocating for the proper triage disposition.

    So stop wearing the hairshirt. You can’t figure it all out in the field! The ED physician who took the patient was likely using a “retrospectoscope” to look at the 12-lead, something no EMS crew carries yet!

    Oh yeah, I’m an ED doc, former medic.

    November 10, 2011 at 5:26 pm

    • Dr. Walsh,

      Thank you for your kind words and support! This was and is, indeed, a very murky case– both in terms of the EKGs themselves and also the surrounding events. I appreciate your understanding, and I too believe that retrospective confirmation bias has played a significant role not only in my view of the case but most likely how the consulting physicians viewed this patient’s presentation once the STEMI label was on the table. Thank you again for your perspective.

      November 18, 2011 at 9:42 am

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s