Where Angels Fear to Tread
Case 1
A 66-year-old male patient presented to hospital with recent chest pain at rest. His cardiac history dated back nearly 10 years when he underwent coronary angiography elsewhere for accelerating angina, was found to have three distinct lesions, and underwent uncomplicated coronary bypass with a left internal mammary artery (LIMA) to the mid left anterior descending coronary artery (LAD), and saphenous grafts to obtuse marginal (OM)2 and the distal right coronary artery (RCA). He did well for about 2 years with only occasional symptoms, until he presented again at a second institution where it was found all three grafts had failed, but only one stent was placed. Over the next 5 years, he would eventually receive stents to the other two vessels at a third and fourth institution. Since the bypass, he had been maintained on aspirin and a statin, while his angina, though improved, had never completely resolved following bypass or the subsequent stents, though following the third stent, he had remained largely pain free on a beta blocker and long-acting nitrate, until he presented to our institution with accelerating episodes of rest pain. Admitted by the hospital service for unstable angina, his electrocardiogram (EKG) and troponin I remained normal, and he was referred to the catheterization lab on history without further testing.
Case 2
A 39-year-old man presented as a new patient for follow-up after his 6th angiogram, including two distant hospital transports, one by helicopter, for “STEMI” (ST-elevation myocardial infarction). He had originally presented with chest pain and undergone a successful percutaneous coronary intervention (PCI) of the mid circumflex (CFX) and proximal LAD. However, on 4 subsequent examinations, including the air transport case for apparent STEMI, angiography had failed to demonstrate a new, treatable, lesion. Recently, while undergoing inpatient therapy for substance abuse, he was transferred to a nearby institution for a 6th evaluation, due to persistent waxing and waning symptoms at rest, with a mildly elevated troponin and normal EKG.
Case 1. Angiographic Findings
Grafts:
LIMA: Atretic
Saphenous grafts: Occluded at origins and distal anastomoses
Native Vessels:
Left main: Smooth, no significant disease
Proximal LAD, CFX and mid RCA stents: All display smooth, 20% tubular in-stent restenoses
Distal CFX and RCA: No significant disease
Mid LAD: 31 mm 90% stenosis, TIMI-2/3 flow distally, with no symptoms or EKG changes
Case 2. Angiographic Findings
Left main: No disease
The LAD and CFX both were also free of significant disease, though mild ectasia was present. There was little to no restenosis in either the mid CFX or LAD stent. However, according to the patient, upon engagement of the RCA, the operator exclaimed excitedly, “I know what you’ve got!”
Board Question: What does the ‘P’ in HIPAA stand for?
In the lab, we are often faced by daunting challenges of technique, and practice problem solving in complex disease and complicated anatomy under emergency conditions, especially in the setting of acute coronary syndromes (ACS). In the 35 years since I graduated medical school, what has been accomplished in coronary angiography and interventions is nothing short of amazing. I know something about this, as for the first 21 years of my career, I was an interventional cardiologist in a large metropolitan heart center. I have spent the last 7 years caring for patients in an underserved area, primarily office consultation with minor interventions, such as the occasional cardioversion. We also provide high-quality echo services, including stress and dobutamine-stress echocardiography, which is helpful given the great variability of echo quality and interpretation. Yes, we’ve come a long way, but the one thing that has not changed during my first career or my second is the critical importance of the medical history.
These two presentations were purposely provided piecemeal, without any imaging, not even an EKG, so that the reader might experience the limited information garnered in real-life care situations, in the hope of engendering a moment of introspection to question, consider the past experience with similar cases, and form broader, alternative hypotheses as to the potential underlying pathology.
In Case 2, the patient reports having experienced one bona fide ACS (STEMI) with placement of 2 stents, yet he kept coming back and was retreated as a STEMI by immediate cath, even to the point of an air transport to a PCI-capable center. Yet out of 6 acute invasive evaluations, there are at least four negative troponins, only two stents, and a colossal waste of resources. In Case 1, the situation is also unusual, with full revascularization therapy by coronary artery bypass graft surgery (CABG) that failed early, followed by piecemeal PCI. What was really going on?
These cases might not have been decided on history alone, but I suspect had someone paused, or at least after the fact, obtained more records, these cases might have been settled much sooner. The investigator must consider which aspects are inconsistent with a typical ACS patient and how that changes the relative probabilities of the individual event as infarct/non-infarct ACS, due to underlying coronary disease vs spasm, or a Type II infarct. What could help source the pain from an epicardial lesion with plaque rupture and secondary, downstream spasm, vs a small-vessel lesion, or even non-epicardial, microvascular angina vs the myriad non-ischemic causes of chest discomfort from acute pericarditis, myocarditis, or pulmonary embolism, to Tietze syndrome and Zoster? What of those patients with the underappreciated, relatively flat pattern of marginally elevated, potentially false-positive troponin, due to heterophile antibodies and other interference molecules? How might that finding affect the interpretation of an atypical patient’s symptoms? How much more likely is the patient to have coronary artery disease (CAD), possibly even significant CAD, such they have a high-risk profile such as diabetes with renal insufficiency, if their pain is clearly non-cardiac in origin? What about the intermediate risk patient with equivocal noninvasive findings, but a very good story for angina, vs the same story in a very low probability patient?
In Case 1, the staff was already measuring lesion length for the stent they thought might follow when I pointed out how smooth and calcium-free the entire vascular bed appeared, then requested intracoronary nitroglycerin, after which the long, asymptomatic LAD stenosis relaxed, demonstrating a normal, nitrate-plump LAD.
In Case 2, troponins at the referral center were negative, as was the angiogram, and the patient was told he had coronary spasm. Coronary spasm, as Case 1 demonstrates, may occur in multiple vessels and without angina at the time of imaging. Fortunately, in Case 2, our patient stopped experiencing pain following the addition of a dyhydropyridine calcium channel blocker. Although we are still attempting to obtain his original angiogram in which the 2 stents were placed, that diagnosis seems likely as well.
In both of these cases, we must ask the question, what was really going on and what was overlooked? These patients were not managed by inexperienced caregivers; rather, the decision process was almost entirely determined by experienced, board-certified emergency physicians and cardiologists. The decision to proceed with angiography was accepted by board-certified interventional cardiologists and cardiothoracic surgeons, who were presented with similar information. Having cared for both of these patients, I believe they are illustrative of how in our haste to clear cases through the cath lab as soon as possible, our assessment may be compromised by tunnel vision. Do we accept a limited, even biased history provided in handoff, without much question as to the presence of any discordant findings? Do we too rapidly shift our focus exclusively to those incremental findings, which support the treatment path we have been committed to, while in the service of expedience, we may discredit, ignore, or fail to consider further, findings which do not fit the budding assessment? Hence, the patient arrives at the lab promptly, where we are stuck determining what happened and why the expected angiographic images typical of ACS are absent. Or worse, do we see what we expect to see without questioning, as in Case 1? In Case 2, following angiogram after angiogram, did no one ask why there was no new lesion, and simply move on as the tally of negative angiograms rose? Perhaps we may explain this as the cost of optimal efficiency. After all, you can’t be right every time or you won’t get your patients to the lab quickly enough. Besides, we’ve got a cardiogenic shock 5 minutes out, and need to turn over the room. Do we accept all of this costly and potentially risky rework, simply because a stent had been placed or bypass performed in the past? Or were these cases properly adjudicated earlier, but that information was lost, or never passed forward even by the patient?
Tunnel vision is part of the physiologic response to stress, something we as interventional team members face on a daily basis. So how do we use it to see what needs to be seen, without failing to understand the big picture? What else might cause individuals who do very well on board testing make yeoman’s errors in judgment or care?
Case 3.
I recently cared for a 76-year-old male patient, hypertensive and without clinical CAD, who had presented elsewhere with a reversible ischemic neurologic defect due to a confirmed stroke (though it was also called a transient ischemic attack [TIA] in the record). The acute computed tomography scan (CT) was negative, magenetic resonance angiography (MRA) demonstrated no great vessel disease, and the severe right-sided weakness on presentation had virtually resolved by the second day. However, an MRI report specified 3 small, new, ischemic lesions, in 2 different vascular distributions. This was a very well evaluated event. Despite this, the consulting neurologist, upon learning that the transthoracic echocardiogram including bubble contrast, was “negative for source”, recommended dual antiplatelet therapy, rather than an antithrombin, and the patient was sent to stroke rehabilitation (for only one day as he recovered so quickly), only to return within 10 days with a subacute, now permanent, right-sided deficit. The subsequent MRI demonstrated a total of 10 small lesions in multiple vascular distribution. Despite the evidence for significant additional injury and a now permanent neurologic defect, a transesophageal echocardiogram (TEE) performed the following day was still “negative for source”. Antithrombin therapy was finally provided, while an implanted loop recorder demonstrated the occult atrial fibrillation within less than 72 hours. How did this happen?
A cursory review of the echo report revealed plenty of “source”, with left ventricular hypertrophy (LVH) and left atrial enlargement. There had simply been no confirmation of atrial fibrillation or thrombus at the time of exam. It is so obvious, in retrospect, what should have happened, that it begs the question: what were they thinking? Where have you heard that before? I am convinced that board questions profiling this case would have been correctly answered by all involved. It was dispiriting to see the common response to a real-life situation and result: an echo request for “R/O source of Emboli”, being negative, could result in one forgetting or setting aside the fact that multiple emboli were present. Indeed, as soon as the 3 lesions in 2 vascular distributions were identified, the cause was less important than that fact and antithrombin therapy was unequivocally indicated, regardless of other findings. An echocardiogram can only prove the presence of thrombus at the time of examination and the negative echo report for source should routinely address that fact. The clinical finding of “No Source”, in a setting that bespeaks occult atrial fibrillation with cardioembolic stroke, proved far less than helpful. It provided discordant information, which misled the neurologist, such that despite the critical fact that brain emboli were present, dual antiplatelet therapy (DAPT), which is known to be inadequate in cardioembolic stroke, was recommended over an antithrombin. So that is what happened. But why?
I believe it is not due to a lack of training, knowledge, or experience. I know that has not been my own experience, and we all have the boards to prove otherwise.
Board examination tests integral knowledge, as a surrogate for experience and understanding. The examination questions are carefully vetted and scrubbed to eliminate ambiguity, such that they individually discriminate between those tested with the requisite knowledge and those who are missing a key learning point. Individually, any given answer means relatively little; however, taken together in numbers, and the context of each correct answer and question, a pattern of knowledge is established in the examinee that fits the board’s assessment of what a successful diplomate should possess. Unfortunately, however carefully constructed, this array of questions has a limited value as a confirmation of the understanding, correlation, and development of action that comes with experiential learning, that critical part of learning in which the number of neural connections is broadened to connect things learned in life and other areas of training (yes, even way back then), to broaden our understanding in ways we could never gain from training and boards alone. The boards certainly fail to assess expertise in teasing a detailed history out of patients who may lack insight, speak a foreign language, or suffer from cognitive dysfunction. They cannot provide insight into our own areas of daily work and interaction with our patients and peers, which can always be improved. Sadly, there is no test question that adequately provokes the candidate to really dig deeper into the complex and complicated patient, without perhaps giving the right answer away. In real life, we, too, underutilize the option to seek out outside records urgently/emergently, and due to transmission issues, outside angiograms must be copied and arrive by mail. Perhaps this occurs, in part, because it is just so easy to justify the next angiogram? That done, we just roll up our tools and look for our next case, because it is not mandated as part of the job to understand the big picture.
The successful diplomate has demonstrated ability that once provided with data and a question, they most frequently select the right answer by point and click from among a very small set of alternate possibilities. The possible answers are not random. Rather, they are carefully vetted alternatives, chosen with the explicit intent of finding those examinees who will most frequently select the correct answer under the testing conditions. If the question seems ambiguous between choice “a” and “c”, you can be certain there is a key piece of information in the text pointing to a key piece of knowledge the board has determined a successful examinee should possess. So you try to sort it out and in the end guess, if you must.
However, in real life, the table is turned and we must first generate our own set of questions, then develop the correct data, efficiently arrive at the best answer, and finally generate an action plan. Board competence implies that without supporting evidence that the successful examinee will always ask sufficient questions in clinical practice, whether the interview takes place in a crowded emergency department or darkened intensive care unit after a very long day. The conflict of a failed hypothesis (working diagnosis) or action plan with negative outcome should lead to more fruitful secondary questions or investigations, but when we are not presented with the vetted answers to our questions, aberrant processing may follow. Unfortunately, especially in the elderly complicated patient, should the first conjecture fail, left with ambiguity, the patient is discharged with multiple diagnoses unproven. Then, everyone indicating it is not “their” area of expertise, the patient is discharged unimproved, as further discussion of Case 3 may illustrate.
For some, DAPT may feel to be the safer option, but as the HAS-BLED score demonstrates, it is actually at least as, if not more risky than, antithrombin therapy done well in well-chosen patients. Had anyone calculated the CHA2DS2-VASc Stroke Risk Score in our stroke patient: 5 [AGE (2), HTN (1), CVA (2), they would realize even a stable patient with his history would carry a 10% annual risk of cerebrovascular accident (CVA). Knowing that, would they really have discharged him on aspirin
(ASA)/clopidogrel? A completed CVA and bleeding risk assessment might have also led to the conjecture that with the given acute, rather than stable, presentation, the CVA risk would actually have been even higher than the estimate, which applies to stable patients with atrial fibrillation, not a recent embolic stroke victim in whom atrial fibrillation is the most likely, by far, of the remaining possibilities.
In interventional cardiology, we are constantly forced to address real-life questions as to when, and not infrequently, when not to intervene. We must always remember that cath-stent is an unnatural reflex, which should also indicate “caution”. We get the answers correct on our board and recertification examinations, so that we may continue to practice as diplomates, but board questions are not real life. The room where real-life critical decisions are made is not quiet. We are not tested at a computer screen that provides both answers and questions. Rather, in real life, our test arises out of human suffering. There is, in fact, in a real life, a great pressure of time, perhaps self-imposed, to meet one standard of care, or to simply get the work done, but also because a life may hang in the balance, or more cynically: bragging rights at the next cine case review meeting. You don’t have time to meet in the real world? Do let me tell you why we must: not only for ourselves and our colleagues, but for our lab staff, who should also attend. They already know what went exceptionally right or wrong, and don’t be fooled: despite HIPAA, your colleagues hear about it as well. We should embrace the good and the bad of it together. Learn things together and from each other. Thus empowered, any one of our team members may ask a key question when we’re just about to stumble.
Misdirection-fueled situational bias may result in the misapplication of our most powerful diagnostic and therapeutic tool: the catheterization laboratory. It may also lead to failure to reach the correct diagnosis, or equally bad, with an obvious serious diagnosis staring us in the face, we may still blatantly fail in formulating the most appropriate action plan. Hence, it is vital that we explore how discordant test results create bias and how tunnel vision may lead us, force us even, to push down the wrong treatment path. We must learn how to pause and reassess the big picture, and control our tunnel vision, so that we don’t take a dehydrated patient on NSAIDs and an ACE inhibitor with diabetes and renal insufficiency to the lab unnecessarily, especially the morning following a pulmonary CTA.
Big data allows every aspect of the care each of us provides to be scrutinized down to the number of catheters used per case. I believe the health insurance system at large can no longer bear the waste associated with misdirection and tunnel vision errors, and it is only a matter of time before we are held accountable for our share as professionals. In fact, it is our churning of the complicated patient, repeating tests “as necessary” in excuse of the urgency of a night transfer from a nursing home, that has contributed to the poor economic state of the current healthcare system.
We must think each problem through and draw a line in the sand, not for ourselves, rather always making sure that line puts us on the patient’s side. As interventional cardiologists, we should push our referring colleagues to make a real effort to achieve optimal 3-4 drug medical therapy, as a prerequisite for elective coronary angiography in patients with stable angina, so that we will really know when to stent is best. It is critical that we ensure ourselves that medical failures are in fact medical failures, and not fall into the wrong treatment for the wrong condition, before we find ourselves 4 hours into a 40-minute case, trying open the 3rd ramus of a second branch supplying infarcted territory with minimal peri-infarct ischemia, only to learn later (if we ever hear about it at all), that despite our best efforts, the angina persisted, but resolved upon the addition of ranolazine. As we try to do the most good, did we really need to go interventional, and commit this 85-year-old with atrial fibrillation and shortness of breath to triple therapy/DAPT + antithrombin, rather than pushing harder for a medical solution in the first place?
The only way to avoid tunnel vision and misdirection is to better familiarize ourselves with the full patient history, understand the setting, causes, and appearance of false-positive noninvasive imaging, as well as the positive and negative predictive values of those results. We should examine the imaging ourselves whenever confronted by discordant results. We cannot leave all that to support staff alone; they are already running around from one thing to another as fast as they can, and in too many cases, we only receive snippets of the data set, not the whole download.
Our support staff, both within the lab and without, is critical to our very best outcomes. They can only learn from our good example to grow into the truly effective colleagues we desperately need in order to produce our very best outcomes, error free: everyone working together to get things right 100% of the time, with a zero tolerance for mistakes at all stages of care. We can accept no failure without the scrutiny, learning, and process improvement it must engender. We must strive for that level of careful thought and deliberation in the care of our patients that each of us would desire for ourselves or our own families. That is accountability.
Setting a practice mandate of applying the right intervention, at the right time, in the right patient, with the right precautions, only makes sense. Recognizing ASAP doesn’t only mean “as fast as possible”, rather that we should “Always Say A Prayer”, as we face the attendant risks of tunnel vision that may render us and our patients victims of misdirection and other sources of error, which might have been avoided had the necessary history been teased out, or the data better scrutinized. This means the ancillary studies themselves, and not just their reports reviewed or worse even, because of the snippets of history or study results that we may use as shorthand, i.e., “No Source”. There is a reason the medical examiner’s office does not permit the use of “heart failure” as a cause of death. It is often not quite accurate and we mislead ourselves when we fail to recognize that “heart failure” of the elderly may include obstructive sleep apnea or other cause of right more than left heart failure, even though the echo demonstrated diastolic abnormalities. The real diagnoses in the complicated patient are far more complex than we may admit.
As many of you likely recall, for the correct answer to the board question regarding HIPAA, the Health Insurance Portability and Accountability Act, the P stands for Portability, NOT privacy. We desperately need access to more portable records. Not just the reports of studies. We require prompt access to imaging. It is imperative that the federal government not just saddle us with the responsibility of ensuring every patient’s privacy. Patients, caregivers, and payers must insist, upon the necessary enhancements of secure data transfer in order to provide the critical data portability, so that every American’s caregiver of the moment has immediate and complete access to the all of the reports and imaging necessary to make fully educated decisions, based upon the most up-to-date information. The snippet-laden underdocumented or erroneous history creates a field of mines that smart, skillful, and well-tested operators stumble through all too frequently, as they are caught unaware. Yet it is our patients who will always suffer the most for our neglect in this area. We must Act to ensure their health and safety.
American healthcare simply can no longer afford to get this wrong.
Dr. Jackson Thatcher can be contacted at jackson.thatcher@va.gov.
