Reply: Against Another Nonspecific Marker of Perfusion and Troponin in Sepsis.

From the Authors:

We thank Siuba and Farkas for their interest in our article and for their thoughtful comments (1). Although respiratory infections can act as triggers for acute myocardial infarction (MI) (2), we agree that elevated troponin levels should not be misinterpreted as a sign of coronary artery disease in critically ill patients with pneumonia who present without clinical signs and symptoms suggesting cardiac ischemia, and we advise caution against performing invasive diagnostic procedures or starting treatment for myocardial injury without signs of MI in the intensive care unit (ICU) setting. Furthermore, we find the use of troponin clearance as a perfusion target during sepsis resuscitation, as suggested by Bonk and Meyer in the editorial that accompanied our publication (3), an interesting concept; yet, we concur that there is currently insufficient evidence to support such an approach. Moreover, mechanisms and kinetics of troponin release and decline during sepsis are complex and still ill defined, which renders troponin as a perfusion marker a challenging target.

The difficulty in providing a satisfying clinical interpretation of troponin release in critically ill patients also seems to underpin most of the critiques expressed by Aberegg and Kaufman (4). First, they suggest that higher cutoff levels for abnormal troponin values should be used in critically ill patients, basing this suggestion on the observation that abnormal values are known to be prevalent in this population and that the pretest probability of having type 1 MI is low. We agree that higher thresholds may increase test specificity for type 1 MI, but this argument seems to be beside the point. We used troponin to assess myocardial injury, not infarction, with the former defined as a troponin level above the 99th percentile upper reference limit in accordance with the universal definition of MI (5). This distinction is very important. It also renders mute the base rate fallacy argument put forward by Aberegg and Kaufman. Troponin release does not necessarily equal myocyte necrosis. In fact, troponin release during sepsis may result from a transiently increased membrane permeability releasing smaller troponin fragments from cytosolic pools into the systemic circulation without signifying cell death (6, 7). Furthermore, troponin release could be related to myocardial turnover and/or cell apoptosis, as may occur during acute increase in preload or ischemia (8). This uncertainty, in fact, underpins the very premise of our study. Troponin elevations in the ICU setting require more careful consideration than a knee-jerk response of MI versus no MI. We believe that increasing the threshold for what should be considered an abnormal troponin level in ICU patients (and thus for what is considered myocardial injury) would be particularly dangerous, because there is considerable evidence that even minor elevations of troponin are independently associated with increased morbidity and mortality (9, 10). Trivializing these findings by blindly raising the limit of what is considered normal seems unwise (11).

The second point raised by Aberegg and Kaufman claims that our data lend only little support to oxygen supply–demand mismatch as a potential cause of myocardial injury during sepsis. However, the authors seem to have overlooked the fact that this claim was not based simply on a logistic regression analysis yielding associations with preexisting risk factors for atherosclerosis but also on mixed model analyses in which time-dependent factors such as tachycardia and hypotension were independently associated with troponin release. These factors have been labeled as potential causes of type 2 myocardial ischemia in the fourth universal definition of MI (5).

We agree with Aberegg and Kaufman that the causes of troponin release in the absence of an acute coronary syndrome are most likely multifactorial and that the clinical significance of troponin release still requires further study. However, we strongly oppose the sentiment that elevated troponin concentrations during sepsis are nonspecific, merely representing yet another biomarker of general disease severity. This notion echoes a common frustration among clinicians that reflects their uncertainty about what to do with a positive troponin test result in a very sick patient without signs and symptoms of MI. This frustration should not lead to a disregard of the test. Given its clear association with mortality and how common it is, we should be motivated to find out why myocardial injury occurs during severe community-acquired pneumonia and sepsis; just disregarding it would be a poor approach to this clinical problem. Our study was one of the first to systematically investigate troponin release using a longitudinal approach, and this enabled us to identify several—potentially etiologic—factors. Disregarding these episodes and simply labeling them as “troponinemia,” “troponinitis,” or “troponin leak” would truly be a misadventure.