Author + information
- Lori B. Daniels, MD, MAS⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Lori B. Daniels, University of California at San Diego, Mail Code 7411, 9444 Medical Center Drive, La Jolla, California 92037-7411
“Be not simply good; be good for something.” —Henry David Thoreau, 1848 (1)
Troponin assays are getting better, so much so, in fact, that we as clinicians need to similarly evolve in our understanding of how we can use them most effectively. I have observed medical house staff cast off the significance of elevated troponin levels in a patient by diagnosing “troponinemia” and moving on, as if that were an explanation in and of itself. House staff are not alone in their misunderstanding of these test results. The ever-improving sensitivity of troponin assays has engendered skepticism and caused confusion for many, but the reality is that the modern troponin assays truly are better than earlier versions. They are providing accurate and precise information, and are providing it in abundance.
Early troponin assays, intended to aid in the diagnosis of acute myocardial infarction, were generally easy to use, providing simple “yes/no” answers because of their relative lack of sensitivity. They worked well, and troponins became an integral part of the universal definition of myocardial infarction (2). The universal definition also recommends minimum standards of precision for any troponin assay used to diagnose myocardial infarction, noting that the assay should have a coefficient of variation of ≤10% at the threshold concentration representing the 99th percentile upper limit of a normal reference population. As troponin assays have become more sensitive and precise in accordance with this mandate, they have improved the early diagnosis of myocardial infarction (3,4) but have left a lot of confusion in their wake. Most recently, “highly sensitive” troponin assays have emerged that can, by definition, detect troponin in more than 50% of the general population. The most sensitive of these assays can detect troponin in almost everyone (5). Thus, troponin assays are no longer simple binary “yes/no” variables but instead need to be interpreted as continuous variables, and within a greater context, with an understanding of the interplay of numerous factors such as age, sex, and renal function, among others.
In addition, now that troponin levels can be measured in nearly all individuals, they have become attractive candidates to be used for risk stratification in the general population. Initial studies of highly sensitive troponin assays in the general population have shown that even minimally elevated levels carry prognostic value. This was described in CHS (Cardiovascular Health Study), which involved elderly community dwellers aged older than 65 years (6), as well as in 30- to 65-year-olds from the Dallas Heart Study (7) and in middle-aged 54- to 74-year-old participants from the general population in the ARIC (Atherosclerosis Risk in Communities) study (8). Prevalence of detectable troponin in these studies ranged from 25% in the younger Dallas population to 66% in the other 2 studies. Together, these studies taught us that even minimally elevated cardiac troponin, at levels below the 99th percentile, are associated with an increased risk of adverse outcomes, including death, heart failure, and coronary heart disease.
In this issue of the Journal, Eggers et al. (9) extend these findings with their report on a community-based cohort of 70-year-olds from the PIVUS (Prospective Investigation of the Vasculature in Uppsala Seniors) study, in which cardiac troponin was detectable in nearly all subjects. For this study, highly sensitive cardiac troponin I (cTnI) was measured in 1,004 individuals from Uppsala, Sweden, and a second cTnI measurement was performed 5 years later in 81% of the cohort. Participants were followed up for a total of 8 years. The authors found detectable cTnI levels in 96% of participants at baseline and 99% at the 5-year follow-up, and these levels were independently associated with both all-cause and cardiovascular mortality. In addition, median cTnI levels increased over 5 years, and change in cTnI was also associated with all-cause mortality.
This study (9) is an important addition to the literature and will help us interpret troponin levels going forward. For the first time in a community-based cohort, troponin was detectable in almost all participants. With the prospect of ubiquitously detectable troponin levels, we need to learn more about what factors contribute to relative elevations. In this and previous studies, elevated troponin and changes in troponin levels were related to several cardiovascular risk factors (6–8). As consistently shown in community-based studies, troponin levels positively correlate with age and are higher in men than in women. This is also consistent with a recent study which showed that a lower, sex-specific cutoff is needed to optimally use high-sensitivity troponin assays for risk assessment among women in the general population (10). In PIVUS, body mass index (BMI) was associated with baseline troponin levels but only in analyses that did not adjust for echocardiographic variables. However, if troponin is to be used for risk stratification in the community, the majority of individuals would be unlikely to have echocardiograms a priori; it is probably reasonable, therefore, to consider BMI as an important covariate when interpreting troponin levels in this setting. The association of troponin with BMI in PIVUS is also consistent with results from both ARIC and the Dallas Heart Study (although not CHS) (6–8).
There are also several important differences in the current study (9) compared with previous studies. In ARIC, the Dallas Heart Study, and CHS (6–8), as well as in a cohort of 512 diabetic women from the Women's Health Study (11), diabetes was associated with higher troponin levels. In contrast, troponin levels were not associated with diabetes status in the current study, although as the authors point out, there were relatively few diabetic participants in PIVUS, and the study did not use a very sensitive means of assessing for diabetes. Also in contrast to previous studies, baseline troponin levels were not associated with renal function in PIVUS, even though a decline in renal function over 5 years was associated with a rise in troponin. The reason for the lack of association here is unclear but is contrary to established patterns (12). The study also did not find a significant association between renal function and mortality outcomes, which is in contrast to our current understanding of risk factors.
The study by Eggers et al. (9) has some weaknesses and raises several questions. Although median troponin levels increased over 5 years, it is unclear from the data presented the number of individuals who had an increase in troponin versus the number who had a decrease, and what the distribution was, although we do know that only 16 participants had a decrease of at least 50%. In addition, although the addition of troponin improved the integrated discrimination index compared with a model based on traditional cardiovascular risk factors, it did not significantly improve the C-statistic. Furthermore, when natriuretic peptides were included in the baseline model along with traditional cardiovascular risk factors, troponin did not significantly improve any measure of discrimination or reclassification. Although troponin levels are promising for assessing cardiovascular risk, other biomarkers may be more powerful and effective at present. Meanwhile, change in troponin did not significantly improve any of these metrics either; therefore, it is probably premature to conclude that changes in troponin levels, as currently modeled, can be used to meaningfully affect cardiovascular risk stratification. Previous studies have shown that elevated troponin levels in otherwise healthy individuals are associated more with incident heart failure than with incident coronary heart disease (8), so it is notable that heart failure was not specifically assessed in the current study. It is plausible that addition of this outcome measure would positively affect the incremental benefit of troponin. Finally, this community-based study had only approximately 50% participation. If a “healthy volunteer” bias is in effect, actual associations between troponin levels and outcomes may be even stronger than described here.
Community-based studies such as the current one (9) add to our collective understanding of how to interpret and contextualize these sensitive troponin assays that are destined to show up in our clinics, hospital wards, and emergency departments. Chronic troponin elevations likely reflect pathophysiology that is distinct from acute elevations because the former are associated with a higher risk of future heart failure, whereas troponin in acutely symptomatic patients carries an associated higher risk of coronary heart disease events (13). With the new highly sensitive assays, most patients will have detectable troponin levels, even “low-risk” patients with chest pain, but most will not be having, or even be at increased short-term risk for, an acute coronary syndrome. Understanding the characteristics that lead to elevated levels is therefore of paramount importance, and the study by Eggers et al. (9) helps with this. At the same time, there are limitations to its immediate translation into practice. At this point, it is not clear that measuring a single troponin level in otherwise healthy individuals is indicated, in the absence of symptoms, let alone measuring serial troponin levels. However, a cogent argument could be made for acquiring a “baseline” troponin level in individuals from certain intermediate or higher risk subgroups, especially in the setting of underlying renal or cardiac disease. Not only would this information aid with risk assessment, but it could also serve as a critical comparator if the individual ever presented with acute symptoms. In addition, because serial troponin levels carry some prognostic implications, it raises the question of whether high-sensitivity troponin assays may be useful for monitoring cardioprotective (or cardiotoxic) therapies. However, the optimal time interval between troponin measurements, and the optimal way to assess changing levels, remain questions that the current study does not answer.
In summary, the study by Eggers et al. (9) informs us about troponin levels in a community-based setting but has direct implications for our care of acutely symptomatic patients. With highly sensitive troponin assays, detectable troponin and acute myocardial infarction are no longer interchangeable concepts. More and more often we find that we need to evaluate troponin levels obtained in acute care settings in context, integrating the specifics of the troponin assay, the patient's demographic characteristics and comorbidities, and especially the acute clinical scenario, including symptoms, electrocardiogram findings, vital signs, and the temporal pattern of troponin levels (Table 1,Fig. 1). The current study adds to our understanding of the clinical factors associated with elevated troponin levels and their long-term significance; so that, as troponin assays get better, we can get better at interpreting them.
This work was supported, in part, by a grant to Dr. Daniels from the American Heart Association, National Affiliate. Dr. Daniels has received research support from Alere and Roche Diagnostics; has served as a consultant for Alere; has received speaking fees from Critical Diagnostics; and has served on an advisory board for Singulex.
↵⁎ 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
- Sanborn F.B.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- et al.
- Apple F.S.
- Saunders J.T.,
- Nambi V.,
- de Lemos J.A.,
- et al.
- Eggers K.M.,
- Venge P.,
- Lindahl B.,
- Lind L.
- Cramer G.E.,
- Brouwer M.A.,
- Vader H.L.,
- et al.
- Everett B.M.,
- Cook N.R.,
- Magnone M.C.,
- et al.
- Freda B.J.,
- Tang W.H.,
- Van Lente F.,
- Peacock W.F.,
- Francis G.S.