Author + information
- Received August 15, 2017
- Accepted August 30, 2017
- Published online October 23, 2017.
- Andreas Roos, MDa,b,
- Nadia Bandstein, MD, PhDa,b,
- Magnus Lundbäck, MD, PhDc,d,
- Ola Hammarsten, MD, PhDe,
- Rickard Ljung, MD, PhDf and
- Martin J. Holzmann, MD, PhDa,b,∗ ()
- aDepartment of Medicine, Karolinska Institutet, Solna, Stockholm, Sweden
- bFunctional Area of Emergency Medicine, Karolinska University Hospital, Huddinge, Stockholm, Sweden
- cDepartment of Clinical Sciences, Karolinska Institutet, Solna, Stockholm, Sweden
- dDepartment of Cardiology, Danderyd University Hospital, Danderyd, Stockholm, Sweden
- eDepartment of Clinical Chemistry and Transfusion Medicine, Sahlgrenska University Hospital at Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
- fUnit of Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Solna, Stockholm, Sweden
- ↵∗Address for correspondence:
Dr. Martin J. Holzmann, Functional Area of Emergency Medicine, C1:63, Karolinska University Hospital, Huddinge 14184 Stockholm, Sweden.
Background There is a paucity of data on the association between high-sensitivity cardiac troponin (hs-cTn) levels and outcomes in patients with chest pain but no myocardial infarction (MI), or any other condition that may lead to acutely elevated troponin levels.
Objectives The authors hypothesized that any detectable high-sensitivity cardiac troponin T (hs-cTnT) level is associated with adverse outcomes.
Methods All patients (N = 22,589) >25 years of age with chest pain and hs-cTnT analyzed concurrently in the emergency department of Karolinska University Hospital, Stockholm, Sweden from 2011 to 2014 were eligible for inclusion. After excluding all patients with acute conditions that may have affected hs-cTnT, or MI associated with the visit, or insufficient information to determine whether troponin levels were stable, Cox regression was used to estimate risks for all-cause, cardiovascular, and noncardiovascular mortality, MI, and heart failure at different levels of troponins.
Results A total of 19,460 patients with a mean age of 54 ± 17 years were included. During a mean follow-up of 3.3 ± 1.2 years, 1,349 (6.9%) patients died. Adjusted hazard ratios (with 95% confidence intervals) for all-cause mortality were 2.00 (1.66 to 2.42), 2.92 (2.38 to 3.59), 4.07 (3.28 to 5.05), 6.77 (5.22 to 8.78), and 9.68 (7.18 to 13.00) in patients with hs-cTnT levels of 5 to 9, 10 to 14, 15 to 29, 30 to 49, and ≥50 ng/l, respectively, compared with patients with hs-cTnT levels <5 ng/l. There was a strong and graded association between all detectable levels of hs-cTnT and risk for MI, heart failure, and cardiovascular and noncardiovascular mortality.
Conclusions Among patients with chest pain and stable troponin levels, any detectable level of hs-cTnT is associated with an increased risk of death and cardiovascular outcomes and should merit further attention.
High-sensitivity cardiac troponin (hs-cTn) assays have been established as key cardiac biomarkers for the diagnosis of myocardial infarction (MI) (1). The improved early diagnostic features have led to earlier rule-out and rule-in of MI in the emergency department (ED) (2–6).
Although older-generation troponin assays offered useful prognostic information in patients with cardiac disease (7–9), evidence is consistent that the newer high-sensitivity assays retain prognostic value at levels that were undetectable with the older assays (10,11).
The increased sensitivity has been associated with decreased specificity, causing a substantial proportion of patients without MI to have hs-cTn levels above the 99th percentile value, which is the cutoff level for the diagnosis of MI (12,13). As for patients with cardiac disease, troponin levels have been found to be independent determinants and predictors of adverse outcomes in patients with noncardiac acute medical conditions (14). Investigation and treatment of these patients are focused on the underlying conditions associated with elevated hs-cTn levels. Persistently elevated hs-cTn levels without a typical rise and/or fall pattern indicate chronic myocardial injury. Knowledge on how to investigate or treat these patients to prevent future adverse outcomes is limited (15). Moreover, in patients with chest pain but no MI, or any other condition that may affect troponin levels acutely, few data are available on the association between the high-sensitivity cardiac troponin T (hs-cTnT) level and long-term outcomes. Therefore, we conducted a large observational cohort study including all consecutive patients who sought medical attention for chest pain during a 4-year period at our hospital to investigate the association between stable hs-cTnT levels and long-term outcomes.
From January 1, 2011 through October 20, 2014, we included all patients >25 years of age with a principal report of chest pain and at least 1 hs-cTnT level analyzed in the ED at the Karolinska University Hospital, Stockholm, Sweden. Patients with an estimated glomerular filtration rate (eGFR) of <15 ml/min/1.73 m2, MI associated with the visit, and any other acute medical condition that could be related to an acute increase in the hs-cTnT level were excluded (Figure 1).
The Regional Ethical Review Board in Stockholm approved the study protocol. The study adhered to the principles in the Declaration of Helsinki.
We used our administrative database, which includes all visits to the ED, to identify patients with chest pain. Laboratory data were retrieved from the hospital’s Department of Information and Technology. The data were then sent to the Swedish National Board of Health and Welfare, where information about comorbidities and outcomes from the National Patient Register (16), medication use from the Prescribed Drug Register, and dates and causes of death from the Cause-of-Death register were retrieved. The creation of the database has been described previously (2). The Elecsys 2010 system (Roche Diagnostics, Mannheim, Germany) was used to analyze hs-cTnT levels. The assay has a limit of detection of 5 ng/l and a limit of blank of 3 ng/l. The 99th percentile cutoff point is 14 ng/l, and the coefficient of variation is <10% at 13 ng/l (4).
After excluding all patients with end-stage kidney disease (n = 131), and all patients with MI associated with the visit (n = 1,269), and all patients with a first hs-cTnT level of ≤12 ng/l and a change in the hs-cTnT level of >2 ng/l during the same visit as suggested by the European Society of Cardiology guidelines to identify patients at high risk for MI (1), all patients with at least 1 hs-cTnT level of >14 ng/l were identified (n=4,052). To mimic clinical practice all medical records of these patients were then scrutinized by 4 of the investigators with varied clinical experience (A.R., M.J.H., N.B., and M.L.) to identify and exclude all patients with a medical condition that could potentially explain an elevated hs-cTnT level (n = 1,327), as well as those with insufficient information to determine whether the hs-cTnT level was stable (n = 402). This process is described in detail in the Online Appendix. The numbers of patients and conditions associated with an acute elevation that led to exclusion are shown in Figure 1 and Online Table 1.
Briefly, all available information in a patient’s medical record regarding the index visit and other visits, including laboratory data, radiology, echocardiography, noninvasive stress testing, vital signs, and electrocardiography, was used to determine whether the hs-cTnT level was affected by an acute condition. For patients with at least 1 troponin level >14 ng/l, no specific delta-troponin value was used that would lead to exclusion of the patient. All patients with only 1 hs-cTnT level analyzed were excluded. The proportions of agreement between the assessment of A.R. and that of the other investigators were 0.90, 0.90, and 0.86, respectively, meaning that 90 of 100, 90 of 100, and 86 of 100 patients were assessed in the same way by different investigators. Finally, 100 patients were randomly chosen from the final cohort, and their medical records were scrutinized by 1 senior cardiologist and 1 attending internist who were blinded to the study protocol. These clinicians were asked to assess whether the elevated hs-cTnT level found in each patient was caused by an acute condition. Four patients were determined to have an acute condition leading to a higher than usual hs-cTnT level, meaning that 4% of patients were included in the stable troponin cohort instead of being excluded because of an acutely elevated troponin level.
Exposure was categorized according to the following hs-cTnT levels: 5 to 9, 10 to 14, 15 to 29, 30 to 49, and ≥50 ng/l. Patients with an hs-cTnT level <5 ng/l were used as references.
The day the patient sought medical attention for chest pain was defined as the index date. Delta-troponin was defined as the largest difference between the first troponin level measured at the index date and any other troponin level measured during 24 h from the first level measured. Diabetes was defined as ongoing medication with any hypoglycemic agent. In subgroup analyses, coronary artery disease (CAD) was defined as prior MI or cardiac revascularization, and established heart disease was defined as prior CAD, heart failure, or atrial fibrillation. Ongoing medication was defined as ≥2 dispensed medications during the year preceding the index date. eGFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration equation.
Outcomes and follow-up
Outcomes were retrieved from the Patient Register by using only diagnoses in the primary position. Cardiovascular death was defined as death caused by atherosclerotic disease (17). Only the underlying cause of death was used, and this information was retrieved from the Cause of Death Register in which all deaths in Sweden are registered. The end of follow-up for all-cause mortality was March 28, 2016, and then end of follow-up for all other outcomes was December 31, 2014.
Baseline characteristics are described as frequencies and percentages for categorical variables and means and SDs for continuous variables. Cox proportional hazards models were used to calculate the association, expressed as hazard ratios (HRs) with 95% confidence intervals (CIs), between the hs-cTnT level (<5 ng/l reference group and 5 to 9, 10 to 14, 15 to 29, 30 to 49, and ≥50 ng/l) and the following 5 outcomes: all-cause mortality, cardiovascular death, noncardiovascular death, MI, and hospitalization for heart failure. Two models for each outcome were conducted (unadjusted and adjusted) for the following covariates: age; sex; eGFR; prior MI; heart failure; stroke; chronic obstructive pulmonary disease; atrial fibrillation; hypertension; diabetes; and treatment with platelet inhibitors, beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, and statins. Moreover, 2 sensitivity analysis were conducted for the primary outcome of all-cause mortality unadjusted and adjusted for the same confounders as the main analysis: 1 where only patients with a delta-troponin level of 0 to 2 ng/l were included; and a second analysis where only patients with a change in delta-troponin of <20% were included. We used the Kaplan-Meier method to calculate the cumulative incidence of all-cause mortality. Moreover, time was introduced as an interaction term to account for the increased sensitivity of the hs-cTnT assay after April 24, 2012 (18) by dividing the follow-up period into 2 periods: January 1, 2011 to April 24, 2012 and April 25, 2012 to October 20, 2014.
For subgroup analyses, we evaluated potential effect modification by limiting our analyses to the following subgroups: age, sex, eGFR (>60 and 15 to 60 ml/min/1.73 m2), heart failure, established heart disease, CAD, atrial fibrillation, and time period.
The proportional hazard assumption was evaluated by calculating the correlation between Schoenfeld residuals for the covariates and the ranking of the failure times. The correlations were close to zero, and the p values were >0.05, which implies that the assumption was met. Data management was performed using the World Programming System, version 3.1 (World Programming Ltd., Romsey, Hampshire, United Kingdom). The software R, version 3.2.2 (R Foundation for Statistical Computing, Vienna, Austria) was used for the statistical analyses.
In total, 19,460 patients were included; of these patients, 62%, 21%, and 8.6% had hs-cTnT levels of <5, 5 to 9, and 10 to 14 ng/l, respectively (Table 1). A total of 7.9% of patients had hs-cTnT levels >14 ng/l. With increasing hs-cTnT levels, the patients were older, were more often men, had a lower eGFR, and had more comorbidities. Overall 89% of the study population had a delta-troponin level in the range of 0 to 2 ng/l (Online Table 2). The mean delta-troponin level was 1.4 ± 11.7 ng/l in the whole study population, with levels ranging from 0.23 ± 2.7 ng/l and 1.3 ± 8.1 ng/l in patients with baseline hs-cTnT levels of <5 ng/l and 5 to 9 ng/l, respectively, and increasing gradually to 9.8 ± 19.4 ng/l in patients with baseline hs-cTnT levels of ≥50 ng/l. The mean delta-troponin level in 1,327 patients who were excluded from the main analyses was 71 ± 330 ng/l (Online Table 2).
All-cause, cardiovascular, and noncardiovascular mortality
During a mean follow-up of 3.3 ± 1.2 years (65,196 person-years), 1,349 (6.9%) patients died (Table 2). The crude cumulative mortality rate is shown in the Central Illustration. The yearly rate of death was 0.5% among patients with hs-cTnT levels <5 ng/l, and this rate rose in a graded manner with increasing levels to 33% among patients with hs-cTnT levels ≥50 ng/l (Table 2). The adjusted risk of death during follow-up was doubled in patients with hs-cTnT levels of 5 to 9 ng/l, 3-fold higher in those with levels of 10 to 14 ng/l, and almost 10-fold higher in those with levels ≥50 ng/l compared with patients with hs-cTnT levels <5 ng/l.
In 2 sensitivity analyses we restricted the inclusion of patients to those with (1) a delta-troponin level of 0 to 2 ng/l; and (2) a change in delta-troponin of <20%. The associations were similar to those in the main analysis (Online Table 3).
In total, 288 (1.5%) cardiovascular and 630 (3.2%) noncardiovascular deaths occurred during a mean 2.1 ± 1.1 years (41,760 person-years) of follow-up. Similar to all-cause mortality, the yearly rates of cardiovascular and noncardiovascular mortality increased in a graded manner with increasing hs-cTnT levels. The adjusted risk for cardiovascular mortality was more than 3-fold in patients with hs-cTnT levels of 5 to 9 ng/l, and it increased to 27 times in patients with hs-cTnT levels ≥50 ng/l compared with patients with hs-cTnT levels <5 ng/l. Similarly, the association between the hs-cTnT level and an increased risk for noncardiovascular mortality was significant, but weaker, for all detectable hs-cTnT levels (Table 2).
The 3 most common groups of causes of death were as follows: cancer, with 325 deaths (35%); cardiovascular disease, with 313 deaths (34%); and lung disease, with 53 (5.8%) deaths. This was followed by 39 (4.2%) deaths from gastrointestinal disease and 39 (4.2%) deaths from neurological disease. Eight of the 10 most common specific causes of death were cancer related in patients with hs-cTnT levels <5 ng/l, and none were cardiovascular (Online Table 4). On the contrary, in patients with hs-cTnT ≥50 ng/l only 2 among the 10 most common causes of death were cancer related, and 4 were cardiovascular. The 10 most common causes of death in each stratum of hs-cTnT are shown in Online Table 4.
Myocardial infarction and heart failure
A total of 304 (1.6%) MIs occurred during a mean 2.1 ± 1.1 years of follow-up (Table 2). The yearly rate of MI increased from 0.3% among patients with hs-cTnT levels <5 ng/l to 4.5% in patients with hs-cTnT levels ≥50 ng/l. The adjusted risk for MI during follow-up was not increased among patients with hs-cTnT levels of 5 to 9 ng/l but was doubled among patients with hs-cTnT levels of 10 to 14 ng/l. A 2- to 3-fold increased risk for MI was found in patients with hs-cTnT levels of 15 to 29, 30 to 49, and ≥50 ng/l.
In total, 488 (2.5%) cases of heart failure occurred during a mean 2.1 ± 1.1 years of follow-up. There was a strong and graded increase in the yearly rate of hospitalization for heart failure from 0.1% to 1.0% and to 2.8% among patients with hs-cTnT levels of <5, 5 to 9, and 10 to 14 ng/l, respectively, and to 20% among patients with hs-cTnT levels ≥50 ng/l (Table 2). The adjusted risk of hospitalization for heart failure during follow-up was almost 4-fold higher among patients with hs-cTnT levels of 5 to 9 ng/l and 6-fold higher among those with levels of 10 to 14 ng/l. Patients with hs-cTnT levels >14 ng/l had an 11- to 13-fold increased adjusted risk of heart failure.
When time was added as an interaction term into our statistical models to account for the fact that the hs-cTnT assay became more sensitive after April 24, 2012 (18), the associations among different hs-cTnT levels and death, MI, and heart failure were similar (Online Table 5).
The association between different hs-cTnT levels and all-cause mortality was strong and increased in a graded manner in different age groups; men and women; patients with and without chronic kidney disease; patients with heart failure, CAD, and atrial fibrillation; and patients with no heart disease (Figure 2). When analyzed separately, the association between the hs-cTnT level and death was similar and independent of the change in sensitivity of the hs-cTnT assay after April 24, 2012.
Outcomes in patients with acute elevations of troponin levels
The 1,327 patients who were excluded from the main analyses because of acute illness that led to higher than usual troponin levels were on an average 14 years older, and they were 3 to 4 times more likely to have comorbidities such as chronic kidney disease, MI, heart failure, or atrial fibrillation (Online Table 6). The yearly rates of all-cause, cardiovascular, and noncardiovascular mortality were 11.2%, 4.5%, and 8.7%, respectively (Online Table 7). Corresponding yearly rates for MI and heart failure were 2.8% and 7.0%, respectively. The 30-day mortality, MI, and heart failure rates were 3.9%, 1.7%, and 2.5%, respectively.
This observational cohort study of 19,460 patients who sought medical attention for chest pain in the ED revealed a strong and graded association between all detectable levels of hs-cTnT and an increased risk of death, MI, and hospitalization for heart failure. An hs-cTnT level of 5 to 9 ng/l, which is well below the 99th percentile value of 14 ng/l, was associated with a doubled risk of death after adjustment for confounders.
Earlier studies involving community-based populations revealed consistent and graded associations between increasing hs-cTnT levels and the risk of future cardiovascular events, even at levels below the upper normal limit (19–21). Our results extend the findings from previous studies, which included mainly healthy individuals, to a population including patients with and without established cardiovascular disease, patients with chronic kidney disease, men and women, and patients with a wide range of ages. The association between increasing hs-cTnT levels and outcomes was similar in all subgroups of patients. In a recent study that evaluated an algorithm to diagnose MI in patients with chest pain in the ED, it was noted that patients with elevated hs-cTnT levels in the absence of MI had a high risk for all-cause mortality (22). In another study that included patients with chest pain to investigate negative predictive values for MI by using a high-sensitivity cardiac troponin I (hs-cTnI) assay, it was found that patients with hs-cTnI levels <5 ng/l had a 59% lower risk of MI or cardiac death compared with patients with values between 5 ng/l and the 99th percentile value (23). In these studies, patients without MI were not evaluated for any acute condition that may have led to an elevated troponin level. Therefore, we believe that our study extends the findings from previous studies in the respect that we evaluated all potential acute elevations of troponin levels and excluded patients not only with MI but also with any condition that potentially could explain an elevated troponin level.
Observation of patients with chest pain is currently focused on the exclusion of MI. In a recently published paper, we found that patients with MI were more than twice as likely to undergo echocardiography than were patients with elevated hs-cTnT levels and no MI, even though one-half of the patients with elevated hs-cTnT levels and no MI had no history of heart disease (13). In the current study, the yearly mortality rate was 12% in patients with a persistent hs-cTnT level of 15 to 29 ng/l. This finding corresponds to the first-year mortality rate among 97,254 patients who survived an acute MI in Sweden from 2006 to 2011 (24).
The current clinical guidelines on prevention of cardiovascular disease only briefly mention persistently elevated troponins (1,17). Additionally, no consensus has been reached regarding how “chronically” elevated troponin levels should be defined. These guidelines lack advice on how to investigate, follow-up, or treat patients with persistently elevated troponin levels. Thus, clinicians currently have no guidance regarding how to handle these patients even though our study and as other studies indicate a similar or even higher risk of premature death and cardiovascular outcomes than in patients with MI (19–21).
We found a markedly increased risk of hospitalization for heart failure with increasing hs-cTnT levels, even at concentrations within the normal range. This finding is consistent with data from animal models of induced left ventricular dysfunction, as well as in humans with heart failure, thus supporting continuous and small release of troponin from the myocardium (25). Another study showed that the risk of incident heart failure was 6 times higher in patients with hs-cTnT levels >14 ng/l than <14 ng/l (26).
The association between the hs-cTnT level and risk of hospitalization for heart failure was stronger than the association between the hs-cTnT level and MI. This finding may reflect important differences in the pathophysiological mechanisms of troponin release. A reduced eGFR has been associated with moderate troponin elevations (27). Although renal clearance of small troponin degradation products is the main source of measured troponin (28), increased release of troponin from the heart in the absence of necrosis may also contribute to persistently elevated troponin levels in many cases (29,30). However, the mechanism behind release of troponin from viable cardiomyocytes is unknown (31). Early electron microscopy studies suggested that transient holes in the plasma membrane called “cell wounds” form as a result of ischemia-induced swelling (32,33). Several other mechanisms have been suggested such as myocyte necrosis, apoptosis, normal myocyte turnover, cellular release of proteolytic troponin degradation products, and formation and release of membranous blebs (34). Several different mechanisms of troponin release may have been present among our patients, and some mechanisms may have been associated with worse prognosis than others.
In a population-based cohort comprising mainly healthy individuals, higher hs-cTnT levels were associated with increased left ventricular mass and wall thickness, an increased likelihood of left ventricular hypertrophy, and impaired left ventricular systolic function (20). In contrast, studies do not support an association between the hs-cTnT level and either myocardial hyperenhancement on magnetic resonance imaging as a marker of previous MI (35) or the coronary artery calcium score (20). In patients with stable CAD, the association between the hs-cTnT level and the risks of all-cause mortality and heart failure is stronger than that between the hs-cTnT level and MI (36). Accordingly, chronic troponin release may be mediated by functional and structural heart diseases rather than by ischemic heart disease. However, further investigation is needed regarding what adverse processes could be mediated by troponin leakage and whether the risk associated with detectable troponin levels in stable patients is modifiable.
In the absence of clinical guidelines, we strongly believe that persistently elevated hs-cTnT levels may be reason in itself to investigate patients for exclusion of previously undiagnosed heart disease. In addition, future research should focus on how to mitigate risk for patients with no detectable heart disease that may explain persistently elevated hs-cTnT levels.
The main strength of this study is that it was conducted in the setting where troponin measurement is most commonly performed, namely, in patients with chest pain in the ED. Thus, we believe that the external validity of our findings is high. The large sample size allowed us to analyze data in a large number of hs-cTnT categories and conduct appropriate subgroup analyses. Furthermore, we used validated national health care registers (16) with complete coverage of Sweden. Thus, we had no loss to follow-up, and all medications and comorbidities resulting in hospitalization were known to us.
Unlike some other studies that relied only on 1 hs-cTnT level to categorize patients (19–21), we used all levels available to us and never included patients with an elevated level on only a single measurement. We were able to exclude patients with acute medical conditions that may have led to high hs-cTnT levels. We observed high interobserver agreement in the assessment of whether the hs-cTnT level was affected by acute illness. In addition, when it was not possible to determine the presence of an acute or persistent elevation of the hs-cTnT level, those patients were excluded from the study. Moreover, only 4% of a random sample from the final cohort, which was assessed by 2 physicians not involved in the study, was determined to have acutely elevated hs-cTnT levels. Therefore, we believe that the risk of misclassification in terms of exposure was small.
The main limitation of our study is that we were not able to enroll patients prospectively on the basis of repeated measurements of the hs-cTnT level. Thus, although we believe that a selection bias that leads to 4% of patients being misclassified as exposed is rather small, a preferred strategy would have been to analyze hs-cTnT levels 1 to 2 weeks after the index visit. An alternative strategy for scrutinizing medical records to decide whether troponin levels were stable would have been to predefine a delta-troponin level that would have led to exclusion. However, because we wanted to mimic real-world conditions we chose to use all the information available to us in the medical records. Moreover, if patients with medical conditions related to an acute hs-cTnT elevation arrive late to the ED, their troponin levels may have plateaued. Consequently, even if delta-troponin levels were analyzed 12 to 24 h apart, patients with only minimal changes may have been considered to have falsely stable levels. In addition, if we had restricted our study population to a very narrow delta-troponin level, it would have led to a smaller study population and significantly reduced precision in our estimates. Although 89% of patients had delta-troponin levels between 0 and 2 ng/l, there were patients with rather large variability in hs-cTnT levels who were not excluded because we had no rule for excluding patients on the basis of absolute changes in hs-cTnT levels. With higher baseline hs-cTnT levels the variability became more pronounced, as expected, and as described previously in completely stable patients with elevated baseline troponin levels (37). However, in 2 sensitivity analyses where we restricted the included patients to those who had a delta-troponin level of 0 to 2 ng/l and <20% change in troponin levels, we found similar point estimates for the risk of all-cause mortality as in the main analysis. Therefore, we believe that our findings may be used in a clinical context because the basis of the process of determining whether patients had stable hs-cTnT levels was information available in clinical practice. There may have been differences in the proportion of MI type 1 or 2 cases during follow-up in different categories of hs-cTnT levels. However, because there was no code for subtypes of MI in the Swedish version of the International Classification of Diseases-Tenth Revision, we were not able to investigate whether there were any differences among the groups. We did not have information on left ventricular hypertrophy or reduced left ventricular ejection fraction, both of which were associated with elevated troponin levels previously (20). Finally, as in every observational cohort study, we cannot exclude that there was residual confounding.
In a large cohort of patients with chest pain without MI or other conditions that may have affected the hs-cTnT level, we found that an hs-cTnT level of 5 to 9 ng/l was associated with a 2-fold higher risk of long-term mortality. The risk of death and all cardiovascular outcomes increased in a graded manner with increasing hs-cTnT levels, starting at levels well below the normal upper limit. Our data indicate that any detectable hs-cTnT level is associated with an increased risk of death and cardiovascular events and may merit further attention.
COMPETENCY IN MEDICAL KNOWLEDGE: In patients presenting to the emergency department with chest pain, even slightly elevated blood levels of hs-cTnT are associated with adverse clinical outcomes.
TRANSLATIONAL OUTLOOK: More information is needed to define optimum evaluation and management strategies for patients with elevated hs-cTnT levels and the interventions that reduce the risk of adverse events.
The authors thank Fredrik Mattsson at FM Statistikkonsult for his assistance with data management and statistical analyses.
Dr. Roos holds a research position funded by the regional agreement on medical training and clinical research between Stockholm County Council and Karolinska Institutet (grant 20160644). Drs. Holzmann and Bandstein hold research positions funded by the Swedish Heart-Lung Foundation (grants 20150603 and 20150594). Dr. Holzmann has received consulting honoraria from Actelion and Pfizer. Dr. Ljung has received consulting fees from Pfizer. The sponsors had no role in the design or conduct of this study. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- emergency department
- estimated glomerular filtration rate
- hazard ratio
- high-sensitivity cardiac troponin
- high-sensitivity cardiac troponin T
- myocardial infarction
- Received August 15, 2017.
- Accepted August 30, 2017.
- 2017 The Authors
- Roffi M.,
- Patrono C.,
- Collet J.P.,
- et al.
- Bandstein N.,
- Ljung R.,
- Johansson M.,
- Holzmann M.J.
- Giannitsis E.,
- Kurz K.,
- Hallermayer K.,
- Jarausch J.,
- Jaffe A.S.,
- Katus H.A.
- Giannitsis E.,
- Becker M.,
- Kurz K.,
- Hess G.,
- Zdunek D.,
- Katus H.A.
- Eggers K.M.,
- Lagerqvist B.,
- Venge P.,
- Wallentin L.,
- Lindahl B.
- Latini R.,
- Masson S.,
- Anand I.S.,
- et al.
- Hammarsten O.,
- Fu M.L.,
- Sigurjonsdottir R.,
- et al.
- Roos A.,
- Hellgren A.,
- Rafatnia F.,
- et al.
- Ahmed A.N.,
- Blonde K.,
- Hackam D.,
- Iansavichene A.,
- Mrkobrada M.
- Chapman A.R.,
- Adamson P.D.,
- Mills N.L.
- European Society of Cardiology
- Saunders J.T.,
- Nambi V.,
- de Lemos J.A.,
- et al.
- Shah A.S.,
- Anand A.,
- Sandoval Y.,
- et al.
- Narula J.,
- Pandey P.,
- Arbustini E.,
- et al.
- Ndumele C.E.,
- Coresh J.,
- Lazo M.,
- et al.
- Carlsson A.C.,
- Bandstein N.,
- Roos A.,
- Hammarsten O.,
- Holzmann M.J.
- Fridén V.,
- Starnberg K.,
- Muslimovic A.,
- et al.
- Kashioulis P.,
- Hammarsten O.,
- Marcussen N.,
- Shubbar E.,
- Saeed A.,
- Guron G.
- Heyndrickx G.R.,
- Amano J.,
- Kenna T.,
- et al.
- Cooper S.T.,
- McNeil P.L.
- White H.D.
- Klinkenberg L.J.,
- van Dijk J.W.,
- Tan F.E.,
- van Loon L.J.,
- van Dieijen-Visser M.P.,
- Meex S.J.