Authors: Neumann JT et al. JAMA Cardiol 2016 Jun 1. Carlton E et al. JAMA Cardiol 2016 Jun 1.
Two large studies show that troponin levels that are normal at presentation and not increased 1 hour later can rule out myocardial infarction.
Prior research has suggested that negative and nonincreasing high-sensitivity cardiac troponin (hsTn) levels measured at emergency department (ED) presentation and 1 hour later can rule out acute myocardial infarction (MI) in low-risk patients with nonischemic electrocardiograms (NEJM JW Emerg Med Feb 2016 and Ann Emerg Med 2016 Jan 12; [e-pub]). However, the negative predictive value in the cited study was 99.1%, and the sensitivity was 96.7%, meaning that 0.9% of negative tests were wrong and, importantly, 3.3% of MIs were missed. Two new studies assessed the sensitivity of hsTnI when a lower threshold is used.
In the first study, involving 1040 patients with acute chest pain presenting to a single ED in Germany, the outcome was an adjudicated diagnosis of acute MI or 12-month mortality. Non–ST-elevation MI (NSTEMI) type 1 (plaque rupture) was considered ruled out if hsTnI was ≤6 ng/L at presentation and 1 hour later. The negative predictive value was 99.8% and the sensitivity 99.1%, meaning that only 0.9% of MIs were missed. NSTEMI was considered ruled in if the 1-hour value was >6 ng/L and at least 12 ng/L higher or lower than the initial value. The positive predictive value was 82.8% and the specificity 98.0%. The investigators observed similar results when applying the algorithm to two large cohorts from prior studies. Accuracy was lower when the definition of MI also included demand ischemia (type 2).
In the second study, researchers pooled data from five prospective observational cohort studies in England, Australia, and New Zealand in which 3155 patients with chest pain suggestive of cardiac ischemia had hsTnI measured at ED presentation. The main outcome was MI within 30 days. MI was considered ruled out by a single hsTnI that was undetectable (i.e., below the assay’s 1.2 ng/L lower limit of detection). The negative predictive value was 99.5% and the sensitivity was 99.0%, meaning that 1% of MIs were missed.
COMMENT
As suggested by the different rule-in and rule-out criteria for MI used in these studies, more work is needed to determine the optimum cutoff and protocol. Nonetheless, these results suggest that hsTn will have a role in ruling out MI in low-risk chest pain patients. Unfortunately, hsTn assays are not yet available in the U.S., where conventional troponin assays at 0 and 3–6 hours must, for now, remain the diagnostic standard. Finally, beware high-risk unstable angina, which does not lead to an acute troponin rise but, rarely, can be the harbinger of a near-future MI.
On a technical note, these studies used a very low threshold or the limit of detection as the cutoff for an abnormal hsTn, which contrasts with the traditional use of the 99th percentile of the upper reference limit. This change results in higher sensitivity but lower specificity, with more possibility of false positives. Nevertheless, this approach might be an improvement over diagnosis based on conventional troponin assays; further work will be needed to determine how to risk stratify patients ruled in by these ultrasensitive protocols.