COMMENTARY

What Do We Do With a Positive Troponin Test?

Amal Mattu, MD

Disclosures

October 28, 2011

Introduction

A physician recently emailed me with a question (paraphrasing): "Our lab recently switched over to using the new high-sensitivity troponin test and we're confused. What do we do with a positive troponin (TN) test?"

A handful of years ago, I suppose my gut response to someone asking what to do with a positive TN test would have been to tell that person to do more continuing medical education before seeing any more patients. My belief now, however, has become one of complete empathy with this physician and his colleagues. It seems that we all are having a tough time figuring this out, including our own cardiologists. The most recent high-profile article pertaining to the high-sensitivity (HS)-TN test has prompted another flurry of emails, questions, and uncertainty. Here's a quick summary and then some further thoughts about the article, TN testing, and evaluating patients with possible acute coronary syndromes.

Rapid Exclusion of Acute Myocardial Infarction in Patients With Undetectable Troponin Using a High-Sensitivity Assay

Body R, Carley S, McDowell G, et al
J Am Coll Cardiol. 2011;58:1332-1339

Study Summary

The study authors evaluated 703 patients presenting with chest pain to a large hospital's emergency department (ED) in the United Kingdom. A total of 130 (18.5%) patients were diagnosed with acute myocardial infarction (AMI) using standard, accepted criteria for diagnosis with TN testing, which incorporates a serial rise (or fall) of TN. Serial testing in this case meant that the researchers obtained a TN level (a HS-TN-T assay was used in this study) at the time of presentation and a second level at 12 hours after the onset of symptoms. Nearly half (46%) of the patients presented within 3 hours of symptom onset, so the study did incorporate many early presenters.

The study authors found that of 130 patients diagnosed with AMI, 100% had an elevated HS-TN level at the time of presentation (95% confidence interval was 95.1%-100%). In other words, no patients who presented initially with a negative HS-TN level had a positive HS-TN on serial testing. The study authors suggested that this HS-TN level would obviate the need for serial testing and therefore allow early discharge of these patients from the ED. Sounds great! No more serial TN testing is needed, and we can now make dispositions after only 1 blood test and, at least slightly, decrease ED crowding. Right?

Viewpoint

Before we even begin to address the limitations associated with the study and its conclusions, let's first try to agree on a few things:

  • Nothing in medicine is 100%: Nothing is 100% positive, 100% negative, or 100% accurate. Well, death is 100% certain, but not much else. Just like every other study in history that claimed 100% accuracy, subsequent attempts to validate this study will demonstrate numbers less than 100%.

  • What we do in emergency medicine is not to rule in or rule out diseases, but we risk stratify everything. For example, if you see a patient with chest pain who has an ECG with ST-segment elevation in the anterior leads and ST-depression in the inferior leads, it risk stratifies that patient to a higher likelihood of AMI, but it doesn't definitely rule in AMI. In fact, the accuracy of such an ECG for AMI is only ~90%. Similarly, if a patient with chest pain has a normal ECG, it risk stratifies that patient to a lower likelihood of AMI but doesn't rule it out.

  • This same concept of risk stratification applies to TN testing. A negative TN risk stratifies a patient to a lower likelihood of AMI but not to 0%. Also, a positive TN risk stratifies a patient to a higher likelihood of AMI, but not to 100%. This last point is critically important to remember: Positive TNs are markers of myocardial injury, but AMI is only one cause of myocardial injury. There are numerous other causes of positive TN: cardiac trauma, congestive heart failure, aortic valve disease, dysrhythmias (tachy- or brady-), rhabdomyolysis, pulmonary embolism, acute neurologic disease (including stroke and subarachnoid hemorrhage), inflammatory diseases of the heart (pericarditis, myocarditis), infiltrative diseases of the heart, sepsis, toxins, and so on.[1] In fact, many cardiologists already think that the current TN tests are too sensitive!

  • The fact that there are so many causes of positive TNs is the exact reason why the international universal definition of AMI[1] incorporates serial TN testing rather than relying on an isolated positive TN. True atherosclerosis-related AMI produces a rise and fall of TNs, whereas the majority of other causes of positive TN do not produce this typical sequence of levels.

  • It's important to remember that our mandate in the ED -- at least in the United States -- is not to simply evaluate patients for AMI. We also need to evaluate patients for acute coronary syndrome. Elevated levels of TN are helpful for diagnosing myocardial injury, but they are nearly worthless for diagnosing ischemia (ie, unstable angina). Therefore, what we currently have in TN is a test that, when positive, doesn't rule in acute coronary syndrome and when negative doesn't rule out acute coronary syndrome.

Putting history and ECG aside, let's focus on just the TN. How exactly are we supposed to use TN testing? First, it's important that we all change our mind-set such that we don't think of TNs as a test to rule in or rule out disease. They are only useful at risk stratifying the disease and making decisions that are based on that risk. If the TN is positive, you and your consulting cardiologists might consider cardiac catheterization or instead consider performing serial testing to distinguish between the typical rise of true AMI vs the "smoldering" elevation that is often seen in the other conditions. On the other hand, if the TN is negative, it simply risk stratifies patients to a lower likelihood of acute coronary syndrome (but not zero). Essentially, negative TNs indicate that a patient is now stable for a provocative test; they do not exclude acute coronary syndrome. The current national American College of Cardiology/American Heart Association (ACC/AHA)[2,3] and American College of Emergency Physicians (ACEP)[4] guidelines allow us to discontinue the ACS workup only after a negative provocative test; they do not allow us to discontinue the workup after negative TNs.

Returning to the article by Body and colleagues, there are major limitations to incorporating these results into practice. First, this study was conducted in the United Kingdom, where it appears, on the basis of the study authors' discussion, that patients are routinely discharged from the ED after serial TNs have excluded myocardial injury. As noted above, in the United States our mandate in the emergency department is not to simply evaluate AMI but to also evaluate acute coronary syndrome. This did not seem to be a concern for the investigators. They stated, in fact, that use of HS-TN would allow for reduction of "unnecessary hospital admissions...and earlier ED discharges." This completely ignores the spectrum of acute coronary syndrome called "unstable angina"!

The second major limitation is the lack of specificity of the HS-TN level. Recall that a positive TN does not automatically equate to AMI. There are many other causes of TN elevation. In this study among 296 patients who had an HS-TN level of 3-14 ng/L, 277 (94%) were eventually judged to not have AMI using an international standard for diagnosis. Even among 212 patients who had HS-TN levels > 14 ng/L, 101 (48%) were eventually judged to not have AMI. The specificity of the test was abysmal, which begs the question with which we started this entire discussion: "What do we do with a positive TN?"

When we first started using TNs in the evaluation of patients with chest pain in the 1990s, it was believed that TN was so specific that a positive result was essentially a first-class ticket to the catheterization lab or at least to the coronary care unit. However, as the sensitivity of TN assays has increased in recent years with consequent reduction in specificity, we are finding with increasing frequency that cardiologists are demanding that we keep patients with positive TNs, who have equivocal histories and ECGs, in the ED for serial TN testing to distinguish between the true AMIs (who demonstrate typical rising levels) vs the non-AMI conditions. With the shortage of beds in most coronary care units and the increasing scrutiny and criticism of cardiologists for performance of catheterizations that turn out negative, it's difficult to blame cardiologists for their reluctance to let these patients automatically "board the plane." Ironically, what used to be a first-class ticket has become in many cases only a standby ticket, pending the serial levels.

If these HS-TN assays are widely adopted, I anticipate that they will only lead to increased ED length of stay and crowding in the United States. They cannot be considered an answer to any of our problems when they only produce more questions. In the words of a fellow emergency physician Stuart Swadron, "This just looks like a new, more expensive myoglobin test."

Abstract

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....