Author + information
- W. Frank Peacock, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. W. Frank Peacock, Emergency Medicine, Cleveland Clinic, 9500 Euclid Avenue, E-19, Cleveland, Ohio 44195
As an emergency department (ED) doctor, I liked this study. Why? Because emergency doctors send patients home. To send a patient home means that he or she lives or dies based on the negative predictive value of the test I just ordered. It is why the study by Body et al. (1), in this issue of the Journal, which presents data on the value of a low troponin, is important. Most studies report specificity. Highly specific troponin assays accurately predict the diagnosis of acute myocardial infarction and the need for cardiac catheterization and are clinically helpful. However, the consequence of high specificity is poor sensitivity. Tests with poor sensitivity are of little use to make a discharge decision, and honestly are dangerous if used as the sole basis for a discharge decision. The clinical consequence of a poorly sensitive test means that if I discharge Mr. Jones, who is sitting in front of me, on the basis of a negative test, he is sent home and perhaps dies of a myocardial infarction—not quite the outcome either one of us was hoping for.
Thus, the focus of this study, in that it attempts to define the value of a negative high-sensitivity troponin, represents an important consequence with significant clinical value. In fact, if you consider the annual 6 million Americans who seek treatment at the ED for chest pain and eventually are ruled out for myocardial infarction (2), an initially negative troponin (occurring in approximately 85%) will impact more patients than the positive result (3).
Highly specific but insensitive tests, such as the currently available United States troponin assays, scare me to death. This is because with the current generation of troponins, a negative result means absolutely nothing. Specificity is great in that it indicates who needs an intervention. But this is an interventional cardiologist's endpoint and does not work for those who must send patients home from the ED. Ponder this: if you are a cardiologist, most folks on your service will have at least some probability of having a cardiac reason for being there. Not so in the ED. Well described is the fact that few patients with chest pain seeking treatment in the ED will have positive troponin results (3). This is because the overwhelming majority do not even have coronary artery disease. Most have something more mundane, such as indigestion or a muscle cramp. So what good is a test that is negative in more than 85% and has no negative predictive value for excluding myocardial necrosis? The answer is “not much.” This is the reason why negative troponin results have not been helpful for ED clinical decision making. It is also why every single professional society guideline says “repeat if initially negative” (although there is no agreement as to the timing of when it should be repeated).
Enter the era of high-sensitivity troponin, with more positive test results than we have ever seen before. Although some bemoan that fact that in the setting of most patients having a detectable troponin, we can no longer have the knee-jerk response that it equates to a trip to the catheterization laboratory, the new reality is that every single troponin elevation above the 99th percentile is associated with increased mortality. It is only a matter of timing. There is no such thing as a “good” troponin elevation. So although it may not be a myocardial infarction in terms of the universal definition (4), from the ED point of view, in a patient with elevated troponin, an immediate discharge is neither the best nor safest strategy.
In this study by Body et al. (1), the consequences of low high-sensitivity troponin results are presented. The real value is the finding that if the initial troponin is <3 pg/ml, further troponin testing adds little to outcome prediction. Some will criticize the fact that only 17.5% of the chest pain cohort would be ruled out on the basis of this single test, saying this number is too small for clinical relevance. However, consider that the numbers of patients currently undergoing a rule-out strategy in the United States are so large that if applied here, there could be significant economic benefit to both patients and providers. It is a new era where a rule out means exactly that: acute myocardial infarction has been ruled out. Hence, we are witnesses to the beginning of the death of the serial biomarker strategy. But do not become too excited; important issues remain, including: 1) unstable angina still lurks; and 2) patients do not always get their stories straight.
Of import is that although troponin provides evidence of dead myocardium, it has yet to be proven that its absence excludes unstable angina. Whether ischemic, but not dead, myocardium can be detected by the newest high-sensitivity troponin assays remains a point of controversy. Currently, the safest strategy is probably that recommended by the Society of Chest Pain Centers and Providers: after necrosis has been excluded, an evaluation of myocardial perfusion provides the safest disposition information for the patient. Whether this means stress testing, echocardiography, or computed tomographic coronary angiography is unclear, and whether that should be carried out immediately or simply in the next several days is undefined, but it seems reasonable that the exclusion of necrosis alone does not provide adequate short-term safety.
The second problem is: can you trust your patient? With the redefinition of myocardial infarction, we now use chemistry to define it. Obviously, this is an improved concept over the old definition. But have we opened Pandora's box? This study demonstrates that we can exclude the presence of newly necrotic myocardium in patients with chest pain by the performance of an improved assay. But is that the end game? Are we done if part of the heart did not die? Thus, this is the challenge of the clinical evaluation in 2011. In the past, in a patient seeking treatment in the ED, chest pain begat an evaluation, an evaluation begat a diagnosis, and those who did not have chest pain were not included. But now that we can detect picogram levels of troponin, is this good enough? Historically, chest pain defined the pretest odds for the need to test for troponin level, and in the absence of chest pain, troponin testing was optional. What has occurred now, since we moved the definition, is that we also have moved the entry criteria. In a compelling study published by Canto et al. (5), nearly one third of more than 400,000 myocardial infarction patients did not have chest pain. Instead, they had nonspecific symptoms such as syncope, shortness of breath, or weakness. What would be the clinical impact of applying the strategy of Body et al. to an all-comers population? How their high-sensitivity test fits in the expanded catchment of a nonchest pain population is unclear, a point that will need to be answered by future studies.
Other critical limitations exist before the results of this study can be applied to U.S. clinical practice. The most important is that the assay that Body et al. (1) used is not available in the United States, and the assays that have been cleared by the Food and Drug Administration are not proven to meet this standard. Thus, we must wait for the regulators to catch up to that which we think we already know.
Although it is reassuring that in the Body et al. (1) analysis, only 1 patient had an adverse event in the 30 days after ED discharge, there are study design considerations in a European strategy that must be considered if it is to be applied to U.S. practice style. European studies consistently have markedly higher rule-in rates, and their litigation climate is not nearly as toxic compared with the United States. The implication of this is that the higher the rule-in rate, the easier is it for an assay to seem accurate in excluding disease (because there are fewer opportunities for misses), but as the rule-in rate declines, with greater numbers of patients having negative test results, the opportunity for false negatives increases. Thus, this study needs to be validated in the U.S. environment, where the rule-in rate commonly is lower than that reported by our European colleagues. In fact, in some U.S. studies of patients undergoing a chest pain unit evaluation, the rule-in rate is reported to be <2% (6–8). This represents a challenging population for any assay. If this new troponin assay is found to be successful for the identification of discharge candidates in this cohort, it could have even more impact on current hospital processes than projected for the European population.
Another consideration is the specific timing of the measurement of troponin. The authors indicate that nearly 50% of patients sought treatment within 3 h. This is important because it is reasonable to conclude that in the minutes immediately after a myocardial infarction, troponin levels are low. Future research will have to define better the relationship between symptom onset and troponin elevation to reassure the clinician that a low troponin, as short as 30 min after symptoms, has the same value as a test run 3 h later.
No discussion of ED chest pain is complete without considering the U.S. medicolegal climate, where tort reform seems to be the odd man out in health care planning. Thus, the notion that the initial troponin result can be used potentially to discharge patients may be a hazardous practice, given the litigation consequences of a missed acute myocardial infarction. Although not as significant as in Europe, this uniquely American challenge makes it difficult to implement such a strategy in the United States without further validation.
Finally, troponin cutoff points must be considered. Identifying a very low troponin cutoff point in the chest pain cohort for whom discharge is reasonable in approximately 15% creates a laboratory assay situation similar to that of the natriuretic peptides. Very low levels suggest an alternative diagnosis, very high levels in the appropriate presentation are consistent with myocardial infarction, and a middle gray zone is where neither exclusion nor diagnosis of acute myocardial infarction is appropriate, and further evaluation is required.
Ultimately, this study by Body et al. (1) represents a valuable examination of the potential for the newest troponin assays to identify patients at low risk of short-term adverse outcomes. Significant hurdles still must be overcome before it is available and appropriate for use in the United States population, but the value of a low high-sensitivity troponin result has great clinical potential.
Dr. Peacock has received research grants from Abbott, Alere, Brahms, Corthera, EKR, Nanosphere, and The Medicines Company; is a consultant for Abbott, Alere, Beckman Coulter, Electrocore, and The Medicines Company; is on the Speakers' Bureau for Abbott and Alere; and has ownership interest in Comprehensive Research Associates LLC, Vital Sensors, and Emergencies in Medicine LLC.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- Body R.,
- Carley S.,
- McDowell G.,
- et al.
- McCaig L.,
- Burt C.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction
- Mitchell A.M.,
- Garvey J.L.,
- Chandra A.,
- et al.