- Email: [email protected]
Prediction of Outcome by Highly Sensitive Troponin T in Outpatients With Chronic Systolic Left Ventricular Heart Failure
Prediction of Outcome by Highly Sensitive Troponin T in Outpatients With Chronic Systolic Left Ventricular Heart Failure Michael Egstrup, MDa,*, Morten Schou, MD, PhDb, Christian D. Tuxen, MD, PhDd, Caroline N. Kistorp, MD, PhDe, Per R. Hildebrandt, MD, DMScf, Finn Gustafsson, MD, PhD, DMScb, Jens Faber, MD, DMSce, Jens-Peter Goetze, MD, DMScc, and Ida Gustafsson, MD, PhDg Our aim was to assess the prognostic impact of a high-sensitivity cardiac troponin T (hs-cTnT) assay in an outpatient population with chronic systolic left ventricular heart failure (HF). Four hundred sixteen patients with chronic HF and left ventricular ejection fraction <45% were enrolled in a prospective cohort study. In addition to hs-cTnT, plasma amino-terminal pro–B-type natriuretic peptide was measured at baseline. Mean age was 71 years, 29% were women, 62% had coronary artery disease (CAD), mean left ventricular ejection fraction was 31%, and 57% had abnormal level of hs-cTnT. During 4.4 years of follow-up, 211 (51%) patients died. In multivariate Cox regression models, hs-cTnT was categorized as quartiles or dichotomized by the 99th percentile of a healthy population. Adjusted hazard ratios for all-cause mortality for quartiles 2 to 4, with quartile 1 as reference, were 1.4 (95% confidence interval 0.9 to 2.4, p ⴝ 0.16) for quartile 2, 1.7 (0.9 to 2.5, p ⴝ 0.12) for quartile 3, and 2.6 (1.6 to 4.4, p <0.001) for quartile 4 and 1.7 (1.2 to 2.5, p ⴝ 0.003) for abnormal versus normal level of hs-cTnT. In patients without CAD, quartile 4 of hs-cTnT was associated with an adjusted hazard ratio of 6.8. In conclusion, hs-cTnT is increased in most outpatients with chronic systolic HF and carries prognostic information beyond clinical parameters and amino-terminal pro–B-type natriuretic peptide. Increased hs-cTnT indicated a particularly deleterious prognosis in patients without CAD. © 2012 Elsevier Inc. All rights reserved. (Am J Cardiol 2012;110:552–557) Since the development of high-sensitivity cardiac troponin T (hs-cTnT) assays, minute increases of plasma cTnT indicating low-grade myocardial damage can be detected, and increases in hs-cTnT have been linked to cardiovascular mortality in healthy populations.1,2 In view of possible incremental prognostic information provided by plasma hscTnT levels beyond clinical information and pro–B-type natriuretic peptide (pro-BNP) or amino-terminal pro-BNP (NT–pro-BNP), the aim of the present study was to investigate the association between hs-cTnT and mortality and morbidity in outpatients with chronic systolic heart failure (HF) managed in a specialized HF clinic. Furthermore, we intended to determine if hs-cTnT might be especially pre-
a Department of Cardiology and Endocrinology, Frederiksberg University Hospital, Frederiksberg, Denmark; bDepartments of Cardiology and c Clinical Biochemistry, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark; dDepartment of Cardiology, Bispebjerg University Hospital, Copenhagen, Denmark; eDepartment of Endocrinology, Herlev University Hospital, Herlev, Denmark; fDepartment of Cardiology, Glostrup University Hospital, Glostrup, Denmark; gDepartment of Cardiology, Gentofte University Hospital, Gentofte, Denmark. Manuscript received January 24, 2012; revised manuscript received and accepted April 17, 2012. Dr. Egstrup was supported by Grant 09-04-R72-A2426-22546 from the Danish Heart Foundation, Copenhagen, Denmark. Dr Gustafsson was supported by Grant 09-066331 from the Danish Agency for Science Technology and Innovation, Copenhagen, Denmark. *Corresponding author: Tel: 45 38-16-43-96; fax: 45 38-16-40-29. E-mail address: [email protected] (M. Egstrup).
0002-9149/12/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2012.04.033
dictive of outcome in a subgroup of patients with coronary artery disease (CAD). Methods We conducted a prospective cohort study in patients attending an outpatient HF clinic at Frederiksberg University Hospital (Copenhagen, Denmark) from 1999 through 2009. The design of the HF clinic has been described in detail previously.3 In brief, patients with suspected or diagnosed HF were referred from primary health care physicians or from the department of cardiology or internal medicine at Frederiksberg University Hospital. Inclusion criteria were chronic HF according to contemporary guidelines4 and left ventricular ejection fraction (LVEF) ⱕ45% documented by echocardiography. During the study period 416 patients provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Written informed consent was obtained before inclusion. At the baseline visit, information on New York Heart Association class, previous diagnosis of diabetes, pharmacologic treatment, height, weight, systolic and diastolic blood pressures, electrocardiogram, and description of echocardiogram including LVEF were recorded in a designated database. HF was considered caused by CAD if any of the following characteristics were present: previous diagnosis of myocardial infarction, percutaneous coronary intervention, coronary artery bypass grafting, coronary angiowww.ajconline.org
Heart Failure/Troponin T and Prognosis in Heart Failure
gram with significant stenosis, or signs of ischemia on myocardial perfusion image. A fasting blood sample was obtained at the baseline visit and directly analyzed for hemoglobin, creatinine, sodium, potassium, glycated hemoglobin A1c, and NT–pro-BNP. Concentrations of NT–pro-BNP were measured using the Elecsys 2010 platform (Roche Diagnostics, Mannheim, Germany). Plasma for hs-cTnT measurement was also obtained at baseline and collected in plastic vials containing ethylenediaminetetraacetic acid. Samples were placed on ice before chilled centrifugation at 3,000g and frozen at ⫺70°C, at which temperature all samples were stored until later analysis. After inclusion of the last patient, hs-cTnT concentrations were measured on the Elecsys 2010 platform using a new assay from Roche. This assay measures human troponin using monoclonal antibodies against amino acids 125 to 131 and 136 to 147, respectively. The assay working range is reported as 3 to 10.000 ng/L, with interassay coefficients of variation of 3.1% at 24 ng/L and 1.3% at 300 ng/L. The lower limit of quantification is 13 ng/L, the lowest limit of detection is 5 ng/L, and the limit of the blank is 3 ng/L as listed by the manufacturer. The gender-specific 99th percentile concentration limit as established in a healthy population (men 18 ng/L, women 8 ng/L) was adopted to differentiate normal from abnormal values of hs-cTnT.5 Estimated glomerular filtration rate was calculated from the 4-component Modification of Diet in Renal Disease equation incorporating age, race, gender, and serum creatinine level.6 The primary outcome was all-cause mortality. Information on time of death was obtained from the Danish Civil Registration System in April 2010. No patients were lost to follow-up. A secondary outcome of combined all-cause mortality or cardiovascular hospitalization was also applied. Dates of first hospitalizations after study inclusion registered with a primary cardiovascular disease diagnosis were obtained from the Danish National Patient Registry in April 2010. Patients were categorized in quartiles according to baseline values of hs-cTnT and in 2 groups with normal and abnormal hs-cTnT values. Numerical data with a normal distribution are presented as mean ⫾ SD and skewed numerical data as median (interquartile range). Categorical data are presented as proportions. Differences between patient groups based on quartiles of hs-cTnT were tested using 1-way analysis of variance or chi-square test. Mortality rates were compared using Kaplan–Meier curves and log-rank method to test for equality of survival distributions across hs-cTnT quartiles. Cox proportional hazards models were used to test the independent predictive power of hs-cTnT quartiles and concentrations ⬎99th percentile of a healthy population on mortality and on the combined end point. Three Cox proportional hazards models were constructed: (1) univariate model, (2) multivariate model adjusting for clinical risk factors and period of inclusion, and (3) multivariate model adjusting for logarithmic NT–pro-BNP) in addition to the variables in model 2. The most important confounders considered of potential prognostic impact by the authors were corrected for age, gender, LVEF, New York Heart Association class, diabetes, CAD, hemoglobin, and estimated glomerular filtration rate. To account for the
553
long inclusion period with potential differences in patient evaluation and treatment according to inclusion time, we divided the inclusion period into 4 intervals (1999 to 2001, 2002 to 2004, 2005 to 2007, 2008 to 2009) and adjusted for interval of inclusion in the multivariate Cox regression models. To assess whether the impact on mortality of hscTnT quartiles varied with the basic confounders gender and age, we performed interaction analyses within the multivariate Cox proportional hazards models. Because we hypothesized that hs-cTnT might be increased and have more pronounced prognostic impact in patients with CAD, interaction analyses with CAD and hs-cTnT quartiles were also performed. Statistical analyses were conducted by M.E. and I.G. All analyses were carried out using SPSS/PASW 19 (SPSS, Inc., Chicago, Illinois). A 2-sided p value ⱕ0.05 was considered statistically significant. Results During the study period 416 patients with chronic HF referred to the outpatient HF clinic were included and followed up for a median of 4.4 years (interquartile range 1.9 to 7.5). Seventy-one percent of the study population were men, mean age was 71 years, mean LVEF was 31%, and 62% had CAD as the cause of HF (Table 1). Distribution of hs-cTnT levels is depicted in Figure 1. Measurable hs-cTnT (ⱖ3 ng/L) was present in 95% of patients, and median hs-cTnT was 18 ng/L (10 to 31). Prevalence of patients with hs-cTnT levels ⬎99th gender-specific percentile of the hscTnT range in a healthy population was 237 (57%), prevalence in men was 148 (50%), and prevalence in women was 89 (73%). Several differences at baseline among groups stratified according to hs-cTnT quartiles were found (Table 1). With incremental quartiles of hs-cTnT, age, prevalence of diabetes and CAD, NT–pro-BNP concentrations, and daily loop diuretic doses increased, whereas estimated glomerular filtration rate decreased. During follow-up 211 patients (51%) died and 330 (79%) met the composite end point of all-cause mortality or cardiovascular hospitalization. Mortalities during follow-up in quartiles of hs-cTnT were 26% in quartile 1, 44% in quartile 2, 59% in quartile 3, and 77% in quartile 4. Differences in survival rates among groups were compared in a Kaplan–Meier analysis presented in Figure 2. A continuous divergence with decreased survival rates with higher hscTnT quartiles during long-term follow-up was found (p ⬍0.001). Univariate Cox regression analyses confirmed a marked increase in mortality risk with ascending hs-cTnT quartiles (quartile 2, hazard ratio [HR] 1.8, 95% confidence interval [CI] 1.1 to 2.9, p ⫽ 0.014; quartile 3, HR 2.7, 95% CI 1.7 to 4.2, p ⬍0.001; quartile 4, HR 4.6, 95% CI 3.0 to 7.0, p ⬍0.001; quartile 1 reference). The addition of clinical risk factors to the model decreased risk estimates, but high levels of hs-cTnT remained strongly associated with mortality risk as evidenced by approximately twofold and threefold increases in mortality risk for quartiles 3 and 4, respectively (quartile 2, HR 1.4, 95% CI 0.9 to 2.4, p ⫽ 0.15; quartile 3, HR 1.9, 95% CI 1.3 to 3.1, p ⫽ 0.008; quartile 4, HR 3.1, 95% CI 1.9 to 5.1, p ⬍0.001) The addition of NT–pro-BNP to the final models did not alter these estimates substantially (Table 2).
554
The American Journal of Cardiology (www.ajconline.org)
Table 1 Baseline characteristics according to quartiles of cardiac troponin T concentration Variable
Age (years) Men Coronary artery disease New York Heart Association Class 3 or 4 Left ventricular ejection fraction (%) Atrial fibrillation Diabetes Body mass index (kg/m2) Estimated glomerular filtration rate (ml/min/1.73 m2) Amino-terminal pro–B-type natriuretic peptide (pmol/L) Hemoglobin (mmol/L) Hemoglobin A1c (%) Angiotensin-converting enzyme inhibitor or angiotensin receptor blocker  Blocker Loop diuretic dose (mg/24 hours)
Total (n ⫽ 416)
Quartile 1 (⬍10 ng/L) (n ⫽ 112)
Quartile 2 (10–18 ng/L) (n ⫽ 100)
Quartile 3 (19–33 ng/L) (n ⫽ 104)
Quartile 4 (⬎33 ng/L) (n ⫽ 100)
p Value
71 ⫾ 11 295 (71%) 256 (62%) 105 (26%) 31 ⫾ 9 105 (26%) 79 (19%) 27.0 ⫾ 5.2 69 ⫾ 23
65 ⫾ 11 70 (63%) 52 (46%) 23 (21%) 32 ⫾ 8 25 (23%) 12 (11%) 27.1 ⫾ 5.3 82 ⫾ 21
71 ⫾ 10 77 (77%) 60 (61%) 19 (19%) 30 ⫾ 9 26 (26%) 17 (17%) 27.3 ⫾ 4.8 72 ⫾ 22
72 ⫾ 11 72 (69%) 72 (69%) 32 (31%) 32 ⫾ 9 25 (25%) 24 (23%) 27.1 ⫾ 6.0 66 ⫾ 22
75 ⫾ 9 76 (76%) 72 (72%) 33 (33%) 29 ⫾ 9 30 (30%) 26 (26%) 26.5 ⫾ 4.5 57 ⫾ 22
⬍0.001 0.07 ⬍0.001 0.06 0.06 0.71 0.02 0.76 ⬍0.001
137 (53–316)
63 (28–174)
115 (53–115)
200 (93–412)
302 (128–553)
⬍0.001
8.6 ⫾ 0.9 6.3 ⫾ 1.2 274 (66%)
8.7 ⫾ 1.0 5.9 ⫾ 0.8 77 (69%)
8.8 ⫾ 0.8 6.2 ⫾ 1.3 74 (74%)
8.4 ⫾ 1.0 6.4 ⫾ 1.2 61 (59%)
8.5 ⫾ 0.9 6.6 ⫾ 1.5 62 (62%)
0.007 ⬍0.001 0.09
184 (44%) 57 ⫾ 85
50 (45%) 30 ⫾ 51
48 (48%) 45 ⫾ 68
45 (43%) 66 ⫾ 92
41 (41%) 88 ⫾ 109
0.79 ⬍0.001
Data are presented as mean ⫾ SD or number (percentage) for continuous variables or median (interquartile range).
Figure 1. Frequency histogram of plasma concentrations of high-sensitivity cardiac troponin T in intervals of 10 ng/L from 0 to 200 ng/L.
The same models were constructed to evaluate the impact of quartiles of hs-cTnT on the composite end point, and similar, but less pronounced, risk estimates were found (quartile 2, univariate HR 1.2, 95% CI 0.9 to 1.6, p ⫽ 0.26; quartile 3, univariate HR 1.4, 95% CI 1.1 to 2.0, p ⫽ 0.02; quartile 4, univariate HR 1.7, 95% CI 1.3 to 2.4, p ⫽ 0.001). When adjusted for the 2 clinical risk factors and NT–proBNP, hs-cTnT did not significantly predict the combined end point (p ⬎0.5 for all comparisons). The risk attributable to having hs-cTnT ⬎99th genderspecific percentile for a healthy population was also examined in the models and, similarly, we found a strong association with mortality (Table 2) and a weaker association with the composite end point (univariate HR 1.3, 95% CI
Figure 2. Kaplan–Meier curves illustrating survival in groups stratified according to quartiles of high-sensitivity cardiac troponin T concentrations (p ⬍0.001, log-rank test).
1.1 to 1.6, p ⫽ 0.019; multivariate HR 1.1, 95% CI 0.9 to 1.5, p ⫽ 0.32). Apart from hs-cTnT, age, gender, NT–pro-BNP, and hemoglobin were found to be independent risk markers in this population (Table 2). Mortality rate was higher in men than in women (multivariate adjusted HR 1.7, p ⫽
Heart Failure/Troponin T and Prognosis in Heart Failure
555
Table 2 Impact of quartiles of cardiac troponin T or cardiac troponin T above 99th gender-specific percentile on all-cause mortality in multivariable analysis Variables in Model
Troponin T quartile 2 3 4 Troponin T ⬎99th percentile Age (per year) Male gender Coronary artery disease New York Heart Association class III or IV Diabetes Hemoglobin (per mmol/L) Left ventricular ejection fraction (per percentage) Estimated glomerular filtrations rate (per ml/min/1.73 m2) Log amino-terminal pro–B-type natriuretic peptide (per unit) Period of inclusion
HR (95% CI) for All-Cause Mortality Quartiles of hs-cTnT
p Value
1.4 (0.9–2.4) 1.7 (0.9–2.5) 2.6 (1.6–4.4)
0.16 0.12 ⬍0.001
1.05 (1.03–1.07) 1.7 (1.2–2.5) 1.0 (0.7–1.3) 1.5 (1.1–2.1) 1.6 (1.1–2.3) 0.9 (0.7–1.0) 1.0 (1.0⫺1.0) 1.01 (1.01–1.02) 2.1 (1.5–3.0) 1.1 (0.9–1.4)
⬍0.001 0.005 0.81 0.01 0.01 0.13 0.43 0.02 ⬍0.001 0.24
hs-cTnT ⬎99th Percentile
p Value
1.7 (1.2–2.5) 1.05 (1.03–1.07) 2.2 (1.5–3.2) 1.0 (0.7–1.4) 1.5 (1.1–2.1) 1.6 (1.1–2.3) 0.9 (0.8–1.1) 1.0 (1.0⫺1.0) 1.01 (1.0–1.01) 2.1 (1.5–2.9) 1.2 (1.0–1.4)
0.003 ⬍0.001 ⬍0.001 0.97 0.01 0.01 0.21 0.54 0.07 ⬍0.001 0.14
Multivariate Cox proportional hazard models; reference values are high-sensitivity cardiac troponin T concentration in quartile 1 or a concentration ⬍99th percentile.
0.005). There was a trend toward a slightly higher median concentration of hs-cTnT in men (19 vs 18 ng/L in women, p ⫽ 0.091). However, no gender modification of hs-cTnT-related risk was found in this study (p ⫽ 0.49 for interaction). To determine any modifying effect of age on the association between hs-cTnT and mortality, we divided the population into 5 age groups at 10-year intervals, with the youngest being 40 to 50 years old and the oldest being 80 to 90 years old. Median levels of hs-cTnT increased progressively with the older age groups (group 1 11 ng/L, group 2 11 ng/L, group 3 14 ng/L, group 4 22 ng/L, group 5 26 ng/L, p ⬍0.001). Adjusted mortality risk also increased progressively in the 3 older groups compared to the younger groups (HRs 2.5, 3.4, 5.4, respectively, p ⬍0.001 for all comparisons), but there was no interaction with hs-cTnT-related risk (p ⫽ 0.45 for interaction). Patients with and without CAD had similar mortality rates (multivariate adjusted HR 1.0, p ⫽ 0.81). Median concentrations of hs-cTnT were higher in patients with CAD (21 vs 15 ng/L, p ⬍0.001). CAD status modified the impact of hs-cTnT on mortality (p ⫽ 0.04 for interaction). The impact of hscTnT quartile 4 versus 1 on mortality was lower in the CAD group (univariate HR 3.2, 95% CI 1.9 to 5.4, p ⬍0.001; adjusted HR 1.6, 95% CI 0.8 to 2.9, p ⫽ 0.15) than in the group without CAD (univariate HR 8.0, 95% CI 3.8 to 16.9, p ⬍0.001; adjusted HR 6.8, 95% CI 2.6 to 17.5, p ⬍0.001). There was no interaction of CAD with the composite end point (p ⫽ 0.34). Discussion The main findings of this study are that hs-cTnT is increased in most outpatients with chronic systolic HF and that hs-cTnT is a strong and independent predictor of long-term all-cause mortality in these patients. The study population was comparable to other European HF populations attending similar HF clinics with respect to
age, gender distribution, prevalence of CAD, and mortality rate.7–10 The predictive ability of hs-cTnT remained significant for quartile 4, conferring a mortality HR of 2.6 independently of important clinical variables and of NT–proBNP, in itself a very powerful predictor of mortality in HF. When dichotomizing the population using the 99th gender-specific percentile, those with abnormal levels had a similar increase in adjusted risk of mortality (HR 1.7). The association of hs-cTnT with the composite end point of mortality and cardiovascular hospitalization was less pronounced. Although a trend toward increased risk estimates with increasing quartiles of hs-cTnT or ⬎99th percentile was noted, TnT did not add independent predictive power. Apparently hs-cTnT is a better predictor of all-cause mortality than cardiovascular hospitalization, which was also noted in the Valsartan Heart Failure Trial (Val-HeFT).11 Cardiovascular hospitalization can be expected to relate to worsening HF in most cases, whereas death outside the hospital often is sudden. Hs-cTnT might be especially suitable for estimating risk of sudden death in chronic HF. Future studies should address this issue. TnT concentrations were higher in patients with advanced age, which has been observed previously.12,13 A novel finding is the lack of a modifying effect of age and gender on the prognostic impact of hs-cTnT on mortality. Presence of CAD would be expected to increase hs-cTnT by ischemic mechanisms, and CAD was indeed associated with higher levels of hs-cTnT in the present study. However, the prognostic impact of hs-cTnT was much stronger in patients without CAD, conferring a sevenfold increase in adjusted mortality ratio in patients without CAD in quartile 4 versus 1. An increased hs-cTnT has been found in patients with HF and without CAD, which may be due to limited endocardial blood flow caused by increased wall stress or microvascular endothelial dys-
556
The American Journal of Cardiology (www.ajconline.org)
function.14 However, the reason for the interaction between hs-cTnT and CAD in relation to prognosis is unclear. One possible mechanism is undetected and untreated CAD in patients diagnosed with nonischemic cardiomyopathy (e.g., patients in whom only noninvasive testing was done). Furthermore, myocardial scarring may be present in patients with HF and normal coronary arteries. Extent of scarring has been associated with a cTnT increase in patients with chest pain and normal coronary arteries15 and may affect prognosis, but this has not been documented yet in patients with HF. Prevalence of hs-cTnT in ⬎99th percentile was 57%. Because this is also the cut-off value in the current universal definition of myocardial infarction,16 most outpatients with HF can be expected to meet this criterion. Emphasis must therefore be placed on the clinical context (e.g., relevant symptoms and electrocardiographic changes) and serial hs-cTnT measurements when distinguishing an HF-related increase in hs-cTnT from the characteristic increasing-and-decreasing pattern related to acute coronary syndromes. Our results confirm and extend the results of previous studies of cardiac troponin in HF populations in different clinical settings. An increased troponin level has been shown to be of prognostic value in hospitalized patients with acute HF17 and with decompensated chronic HF.18 Data from the Val-HeFT randomized clinical trial11 in 4,053 patients with stable HF revealed that hs-cTnT, which was measurable in 90% of participants, predicted all-cause mortality after adjustments for BNP. This population was very different from ours in several important parameters: The Val-HeFT population was approximately 10 years younger, more frequently men, follow-up time was considerably shorter (24 months), and only approximately 35% received a  blocker. A study in elderly primary care patients with suspected HF has shown that a substantial fraction (80%) presented with hs-cTnT above the detection limit.12 A direct comparison to this population is difficult because a large proportion had preserved LVEF (88%) and was in New York Heart Association class I (45%). Our study does, however, show comparable findings because 95% had detectable hs-cTnT. These findings suggest that HF in the absence of acute ischemia promotes leakage of myocyte troponins, and although mechanisms are still being explored, serial measurements of hs-cTnT could provide insight into disease progression in patients with and without CAD. Ishii et al19 demonstrated that an increase in cTnT from admission for decompensated HF to 2 months after initiation of medical therapy was associated with an increase in mortality and readmission rate independent of BNP. Also, data from the Val-HeFT study showed that serial measurements of hs-cTnT may be clinically relevant in outpatients with HF.11 There are some strengths and limitations to this study that should be recognized. Follow-up time was long, allowing for conclusions on the relation between hs-cTnT and risk beyond the time frame provided by Val-HeFT and, to our knowledge, by any other studies on hs-cTnT in outpatients with HF. The long inclusion period introduces a risk of bias because standards of treatment have
changed within this decade. Only about 30% of referrals to the HF clinic were included in the study period. Specific reasons for not including new referrals were not recorded. We tried to compensate for the long inclusion period by including a variable reflecting time of inclusion in multivariate models and this did not affect results significantly. We did not have data on implantable devices in the population, which could also contribute to differences in risk, and ideally should have been accounted for in the regression models. The study may have been underpowered in detecting risk related to modestly increased hs-cTnT. 1. Daniels LB, Laughlin GA, Clopton P, Maisel AS, Barrett-Connor E. Minimally elevated cardiac troponin T and elevated N-terminal pro– B-type natriuretic peptide predict mortality in older adults: results from the Rancho Bernardo Study. J Am Coll Cardiol 2008;52:450 – 459. 2. deFilippi CR, de Lemos JA, Christenson RH, Gottdiener JS, Kop WJ, Zhan M, Seliger SL. Association of serial measures of cardiac troponin T using a sensitive assay with incident heart failure and cardiovascular mortality in older adults. JAMA 2010;304:2494 –2502. 3. Kistorp C, Galatius S, Gustafsson F, Faber J, Corell P, Hildebrandt P. Prevalence and characteristics of diabetic patients in a chronic heart failure population. Int J Cardiol 2005;100:281–287. 4. Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJ, Ponikowski P, Poole-Wilson PA, Strömberg A, van Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, Nieminen M, Priori SG, Swedberg K; ESC Committee for Practice Guidelines (CPG). ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the task force for the diagnosis and treatment of acute and chronic heart failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur J Heart Fail 2008;10:933–989. 5. Mingels A, Jacobs L, Michielsen E, Swaanenburg J, Wodzig W, van Dieijen-Visser M. Reference population and marathon runner sera assessed by highly sensitive cardiac troponin T and commercial cardiac troponin T and I assays. Clin Chem 2009;55:101–108. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease study group. Ann Intern Med 1999;130:461– 470. 7. Gustafsson F, Schou M, Videbaek L, Dridi N, Ryde H, Handberg J, Hildebrandt PR. Danish Heart Failure Clinics Network. Incidence and predictors of hospitalization or death in patients managed in multidisciplinary heart failure clinics. Eur J Heart Fail 2009;11:413– 419. 8. de la Porte PW, Lok DJ, van Veldhuisen DJ, van Wijngaarden J, Cornel JH, Zuithoff NP, Badings E, Hoes AW. Added value of a physician-and-nurse-directed heart failure clinic: results from the Deventer-Alkmaar heart failure study. Heart 2007;93:819 – 825. 9. Senni M, De Maria R, Gregori D, Gonzini L, Gorini M, Cacciatore G, Gavazzi A, Pulignano G, Porcu M, Maggioni AP. Temporal trends in survival and hospitalizations in outpatients with chronic systolic heart failure in 1995 and 1999. J Card Fail 2005;11:270 –278. 10. Pons F, Lupón J, Urrutia A, González B, Crespo E, Díez C, Cano L, Cabanes R, Altimir S, Coll R, Pascual T, Valle V. Mortality and cause of death in patients with heart failure: findings at a specialist multidisciplinary heart failure unit. Rev Esp Cardiol 2010;63:303–314. 11. Latini R, Masson S, Anand IS, Missov E, Carlson M, Vago T, Angelici L, Barlera S, Parrinello G, Maggioni AP, Tognoni G, Cohn JN. Val-HeFT Investigators. Prognostic value of very low plasma concentrations of troponin T in patients with stable chronic heart failure. Circulation 2007;116:1242–1249. 12. Alehagen U, Dahlström U, Rehfeld JF, Goetze JP. Prognostic assessment of elderly patients with symptoms of heart failure by combining high-sensitivity troponin T and N-terminal pro–B-type natriuretic peptide measurements. Clin Chem 2010;56:1718 –1724. 13. Koerbin G, Tate JR, Hickman PE. Analytical characteristics of the Roche highly sensitive troponin T assay and its application to a cardio-healthy population. Ann Clin Biochem 2010;47:524 –528.
Heart Failure/Troponin T and Prognosis in Heart Failure 14. Thygesen K, Mair J, Katus H, Plebani M, Venge P, Collinson P, Lindahl B, Giannitsis E, Hasin Y, Galvani M, Tubaro M, Alpert JS, Biasucci LM, Koenig W, Mueller C, Huber K, Hamm C, Jaffe AS. Study Group on Biomarkers in Cardiology of the ESC Working Group on Acute Cardiac Care. Recommendations for the use of cardiac troponin measurement in acute cardiac care. Eur Heart J 2010;31:2197–2204. 15. Christiansen JP, Edwards C, Sinclair T, Armstrong G, Scott A, Patel H, Hart H. Detection of myocardial scar by contrast-enhanced cardiac magnetic resonance imaging in patients with troponin-positive chest pain and minimal angiographic coronary artery disease. Am J Cardiol 2006;97:768 –771. 16. Thygesen K, Alpert JS, White HD. Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction. Universal
557
definition of myocardial infarction. J Am Coll Cardiol 2007; 50:2173–2195. 17. Peacock WFt, De Marco T, Fonarow GC, Diercks D, Wynne J, Apple FS, Wu AH. Cardiac troponin and outcome in acute heart failure. N Engl J Med 2008;358:2117–2126. 18. Koide K, Yoshikawa T, Nagatomo Y, Kohsaka S, Anzai T, Meguro T, Ogawa S. Elevated troponin T on discharge predicts poor outcome of decompensated heart failure. Heart Vessels 2010;25:217–222. 19. Ishii J, Cui W, Kitagawa F, Kuno T, Nakamura Y, Naruse H, Mori Y, Ishikawa T, Nagamura Y, Kondo T, Oshima H, Nomura M, Ezaki K, Hishida H. Prognostic value of combination of cardiac troponin T and B-type natriuretic peptide after initiation of treatment in patients with chronic heart failure. Clin Chem 2003;49:2020 –2026.
a Department of Cardiology and Endocrinology, Frederiksberg University Hospital, Frederiksberg, Denmark; bDepartments of Cardiology and c Clinical Biochemistry, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark; dDepartment of Cardiology, Bispebjerg University Hospital, Copenhagen, Denmark; eDepartment of Endocrinology, Herlev University Hospital, Herlev, Denmark; fDepartment of Cardiology, Glostrup University Hospital, Glostrup, Denmark; gDepartment of Cardiology, Gentofte University Hospital, Gentofte, Denmark. Manuscript received January 24, 2012; revised manuscript received and accepted April 17, 2012. Dr. Egstrup was supported by Grant 09-04-R72-A2426-22546 from the Danish Heart Foundation, Copenhagen, Denmark. Dr Gustafsson was supported by Grant 09-066331 from the Danish Agency for Science Technology and Innovation, Copenhagen, Denmark. *Corresponding author: Tel: 45 38-16-43-96; fax: 45 38-16-40-29. E-mail address: [email protected] (M. Egstrup).
0002-9149/12/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2012.04.033
dictive of outcome in a subgroup of patients with coronary artery disease (CAD). Methods We conducted a prospective cohort study in patients attending an outpatient HF clinic at Frederiksberg University Hospital (Copenhagen, Denmark) from 1999 through 2009. The design of the HF clinic has been described in detail previously.3 In brief, patients with suspected or diagnosed HF were referred from primary health care physicians or from the department of cardiology or internal medicine at Frederiksberg University Hospital. Inclusion criteria were chronic HF according to contemporary guidelines4 and left ventricular ejection fraction (LVEF) ⱕ45% documented by echocardiography. During the study period 416 patients provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Written informed consent was obtained before inclusion. At the baseline visit, information on New York Heart Association class, previous diagnosis of diabetes, pharmacologic treatment, height, weight, systolic and diastolic blood pressures, electrocardiogram, and description of echocardiogram including LVEF were recorded in a designated database. HF was considered caused by CAD if any of the following characteristics were present: previous diagnosis of myocardial infarction, percutaneous coronary intervention, coronary artery bypass grafting, coronary angiowww.ajconline.org
Heart Failure/Troponin T and Prognosis in Heart Failure
gram with significant stenosis, or signs of ischemia on myocardial perfusion image. A fasting blood sample was obtained at the baseline visit and directly analyzed for hemoglobin, creatinine, sodium, potassium, glycated hemoglobin A1c, and NT–pro-BNP. Concentrations of NT–pro-BNP were measured using the Elecsys 2010 platform (Roche Diagnostics, Mannheim, Germany). Plasma for hs-cTnT measurement was also obtained at baseline and collected in plastic vials containing ethylenediaminetetraacetic acid. Samples were placed on ice before chilled centrifugation at 3,000g and frozen at ⫺70°C, at which temperature all samples were stored until later analysis. After inclusion of the last patient, hs-cTnT concentrations were measured on the Elecsys 2010 platform using a new assay from Roche. This assay measures human troponin using monoclonal antibodies against amino acids 125 to 131 and 136 to 147, respectively. The assay working range is reported as 3 to 10.000 ng/L, with interassay coefficients of variation of 3.1% at 24 ng/L and 1.3% at 300 ng/L. The lower limit of quantification is 13 ng/L, the lowest limit of detection is 5 ng/L, and the limit of the blank is 3 ng/L as listed by the manufacturer. The gender-specific 99th percentile concentration limit as established in a healthy population (men 18 ng/L, women 8 ng/L) was adopted to differentiate normal from abnormal values of hs-cTnT.5 Estimated glomerular filtration rate was calculated from the 4-component Modification of Diet in Renal Disease equation incorporating age, race, gender, and serum creatinine level.6 The primary outcome was all-cause mortality. Information on time of death was obtained from the Danish Civil Registration System in April 2010. No patients were lost to follow-up. A secondary outcome of combined all-cause mortality or cardiovascular hospitalization was also applied. Dates of first hospitalizations after study inclusion registered with a primary cardiovascular disease diagnosis were obtained from the Danish National Patient Registry in April 2010. Patients were categorized in quartiles according to baseline values of hs-cTnT and in 2 groups with normal and abnormal hs-cTnT values. Numerical data with a normal distribution are presented as mean ⫾ SD and skewed numerical data as median (interquartile range). Categorical data are presented as proportions. Differences between patient groups based on quartiles of hs-cTnT were tested using 1-way analysis of variance or chi-square test. Mortality rates were compared using Kaplan–Meier curves and log-rank method to test for equality of survival distributions across hs-cTnT quartiles. Cox proportional hazards models were used to test the independent predictive power of hs-cTnT quartiles and concentrations ⬎99th percentile of a healthy population on mortality and on the combined end point. Three Cox proportional hazards models were constructed: (1) univariate model, (2) multivariate model adjusting for clinical risk factors and period of inclusion, and (3) multivariate model adjusting for logarithmic NT–pro-BNP) in addition to the variables in model 2. The most important confounders considered of potential prognostic impact by the authors were corrected for age, gender, LVEF, New York Heart Association class, diabetes, CAD, hemoglobin, and estimated glomerular filtration rate. To account for the
553
long inclusion period with potential differences in patient evaluation and treatment according to inclusion time, we divided the inclusion period into 4 intervals (1999 to 2001, 2002 to 2004, 2005 to 2007, 2008 to 2009) and adjusted for interval of inclusion in the multivariate Cox regression models. To assess whether the impact on mortality of hscTnT quartiles varied with the basic confounders gender and age, we performed interaction analyses within the multivariate Cox proportional hazards models. Because we hypothesized that hs-cTnT might be increased and have more pronounced prognostic impact in patients with CAD, interaction analyses with CAD and hs-cTnT quartiles were also performed. Statistical analyses were conducted by M.E. and I.G. All analyses were carried out using SPSS/PASW 19 (SPSS, Inc., Chicago, Illinois). A 2-sided p value ⱕ0.05 was considered statistically significant. Results During the study period 416 patients with chronic HF referred to the outpatient HF clinic were included and followed up for a median of 4.4 years (interquartile range 1.9 to 7.5). Seventy-one percent of the study population were men, mean age was 71 years, mean LVEF was 31%, and 62% had CAD as the cause of HF (Table 1). Distribution of hs-cTnT levels is depicted in Figure 1. Measurable hs-cTnT (ⱖ3 ng/L) was present in 95% of patients, and median hs-cTnT was 18 ng/L (10 to 31). Prevalence of patients with hs-cTnT levels ⬎99th gender-specific percentile of the hscTnT range in a healthy population was 237 (57%), prevalence in men was 148 (50%), and prevalence in women was 89 (73%). Several differences at baseline among groups stratified according to hs-cTnT quartiles were found (Table 1). With incremental quartiles of hs-cTnT, age, prevalence of diabetes and CAD, NT–pro-BNP concentrations, and daily loop diuretic doses increased, whereas estimated glomerular filtration rate decreased. During follow-up 211 patients (51%) died and 330 (79%) met the composite end point of all-cause mortality or cardiovascular hospitalization. Mortalities during follow-up in quartiles of hs-cTnT were 26% in quartile 1, 44% in quartile 2, 59% in quartile 3, and 77% in quartile 4. Differences in survival rates among groups were compared in a Kaplan–Meier analysis presented in Figure 2. A continuous divergence with decreased survival rates with higher hscTnT quartiles during long-term follow-up was found (p ⬍0.001). Univariate Cox regression analyses confirmed a marked increase in mortality risk with ascending hs-cTnT quartiles (quartile 2, hazard ratio [HR] 1.8, 95% confidence interval [CI] 1.1 to 2.9, p ⫽ 0.014; quartile 3, HR 2.7, 95% CI 1.7 to 4.2, p ⬍0.001; quartile 4, HR 4.6, 95% CI 3.0 to 7.0, p ⬍0.001; quartile 1 reference). The addition of clinical risk factors to the model decreased risk estimates, but high levels of hs-cTnT remained strongly associated with mortality risk as evidenced by approximately twofold and threefold increases in mortality risk for quartiles 3 and 4, respectively (quartile 2, HR 1.4, 95% CI 0.9 to 2.4, p ⫽ 0.15; quartile 3, HR 1.9, 95% CI 1.3 to 3.1, p ⫽ 0.008; quartile 4, HR 3.1, 95% CI 1.9 to 5.1, p ⬍0.001) The addition of NT–pro-BNP to the final models did not alter these estimates substantially (Table 2).
554
The American Journal of Cardiology (www.ajconline.org)
Table 1 Baseline characteristics according to quartiles of cardiac troponin T concentration Variable
Age (years) Men Coronary artery disease New York Heart Association Class 3 or 4 Left ventricular ejection fraction (%) Atrial fibrillation Diabetes Body mass index (kg/m2) Estimated glomerular filtration rate (ml/min/1.73 m2) Amino-terminal pro–B-type natriuretic peptide (pmol/L) Hemoglobin (mmol/L) Hemoglobin A1c (%) Angiotensin-converting enzyme inhibitor or angiotensin receptor blocker  Blocker Loop diuretic dose (mg/24 hours)
Total (n ⫽ 416)
Quartile 1 (⬍10 ng/L) (n ⫽ 112)
Quartile 2 (10–18 ng/L) (n ⫽ 100)
Quartile 3 (19–33 ng/L) (n ⫽ 104)
Quartile 4 (⬎33 ng/L) (n ⫽ 100)
p Value
71 ⫾ 11 295 (71%) 256 (62%) 105 (26%) 31 ⫾ 9 105 (26%) 79 (19%) 27.0 ⫾ 5.2 69 ⫾ 23
65 ⫾ 11 70 (63%) 52 (46%) 23 (21%) 32 ⫾ 8 25 (23%) 12 (11%) 27.1 ⫾ 5.3 82 ⫾ 21
71 ⫾ 10 77 (77%) 60 (61%) 19 (19%) 30 ⫾ 9 26 (26%) 17 (17%) 27.3 ⫾ 4.8 72 ⫾ 22
72 ⫾ 11 72 (69%) 72 (69%) 32 (31%) 32 ⫾ 9 25 (25%) 24 (23%) 27.1 ⫾ 6.0 66 ⫾ 22
75 ⫾ 9 76 (76%) 72 (72%) 33 (33%) 29 ⫾ 9 30 (30%) 26 (26%) 26.5 ⫾ 4.5 57 ⫾ 22
⬍0.001 0.07 ⬍0.001 0.06 0.06 0.71 0.02 0.76 ⬍0.001
137 (53–316)
63 (28–174)
115 (53–115)
200 (93–412)
302 (128–553)
⬍0.001
8.6 ⫾ 0.9 6.3 ⫾ 1.2 274 (66%)
8.7 ⫾ 1.0 5.9 ⫾ 0.8 77 (69%)
8.8 ⫾ 0.8 6.2 ⫾ 1.3 74 (74%)
8.4 ⫾ 1.0 6.4 ⫾ 1.2 61 (59%)
8.5 ⫾ 0.9 6.6 ⫾ 1.5 62 (62%)
0.007 ⬍0.001 0.09
184 (44%) 57 ⫾ 85
50 (45%) 30 ⫾ 51
48 (48%) 45 ⫾ 68
45 (43%) 66 ⫾ 92
41 (41%) 88 ⫾ 109
0.79 ⬍0.001
Data are presented as mean ⫾ SD or number (percentage) for continuous variables or median (interquartile range).
Figure 1. Frequency histogram of plasma concentrations of high-sensitivity cardiac troponin T in intervals of 10 ng/L from 0 to 200 ng/L.
The same models were constructed to evaluate the impact of quartiles of hs-cTnT on the composite end point, and similar, but less pronounced, risk estimates were found (quartile 2, univariate HR 1.2, 95% CI 0.9 to 1.6, p ⫽ 0.26; quartile 3, univariate HR 1.4, 95% CI 1.1 to 2.0, p ⫽ 0.02; quartile 4, univariate HR 1.7, 95% CI 1.3 to 2.4, p ⫽ 0.001). When adjusted for the 2 clinical risk factors and NT–proBNP, hs-cTnT did not significantly predict the combined end point (p ⬎0.5 for all comparisons). The risk attributable to having hs-cTnT ⬎99th genderspecific percentile for a healthy population was also examined in the models and, similarly, we found a strong association with mortality (Table 2) and a weaker association with the composite end point (univariate HR 1.3, 95% CI
Figure 2. Kaplan–Meier curves illustrating survival in groups stratified according to quartiles of high-sensitivity cardiac troponin T concentrations (p ⬍0.001, log-rank test).
1.1 to 1.6, p ⫽ 0.019; multivariate HR 1.1, 95% CI 0.9 to 1.5, p ⫽ 0.32). Apart from hs-cTnT, age, gender, NT–pro-BNP, and hemoglobin were found to be independent risk markers in this population (Table 2). Mortality rate was higher in men than in women (multivariate adjusted HR 1.7, p ⫽
Heart Failure/Troponin T and Prognosis in Heart Failure
555
Table 2 Impact of quartiles of cardiac troponin T or cardiac troponin T above 99th gender-specific percentile on all-cause mortality in multivariable analysis Variables in Model
Troponin T quartile 2 3 4 Troponin T ⬎99th percentile Age (per year) Male gender Coronary artery disease New York Heart Association class III or IV Diabetes Hemoglobin (per mmol/L) Left ventricular ejection fraction (per percentage) Estimated glomerular filtrations rate (per ml/min/1.73 m2) Log amino-terminal pro–B-type natriuretic peptide (per unit) Period of inclusion
HR (95% CI) for All-Cause Mortality Quartiles of hs-cTnT
p Value
1.4 (0.9–2.4) 1.7 (0.9–2.5) 2.6 (1.6–4.4)
0.16 0.12 ⬍0.001
1.05 (1.03–1.07) 1.7 (1.2–2.5) 1.0 (0.7–1.3) 1.5 (1.1–2.1) 1.6 (1.1–2.3) 0.9 (0.7–1.0) 1.0 (1.0⫺1.0) 1.01 (1.01–1.02) 2.1 (1.5–3.0) 1.1 (0.9–1.4)
⬍0.001 0.005 0.81 0.01 0.01 0.13 0.43 0.02 ⬍0.001 0.24
hs-cTnT ⬎99th Percentile
p Value
1.7 (1.2–2.5) 1.05 (1.03–1.07) 2.2 (1.5–3.2) 1.0 (0.7–1.4) 1.5 (1.1–2.1) 1.6 (1.1–2.3) 0.9 (0.8–1.1) 1.0 (1.0⫺1.0) 1.01 (1.0–1.01) 2.1 (1.5–2.9) 1.2 (1.0–1.4)
0.003 ⬍0.001 ⬍0.001 0.97 0.01 0.01 0.21 0.54 0.07 ⬍0.001 0.14
Multivariate Cox proportional hazard models; reference values are high-sensitivity cardiac troponin T concentration in quartile 1 or a concentration ⬍99th percentile.
0.005). There was a trend toward a slightly higher median concentration of hs-cTnT in men (19 vs 18 ng/L in women, p ⫽ 0.091). However, no gender modification of hs-cTnT-related risk was found in this study (p ⫽ 0.49 for interaction). To determine any modifying effect of age on the association between hs-cTnT and mortality, we divided the population into 5 age groups at 10-year intervals, with the youngest being 40 to 50 years old and the oldest being 80 to 90 years old. Median levels of hs-cTnT increased progressively with the older age groups (group 1 11 ng/L, group 2 11 ng/L, group 3 14 ng/L, group 4 22 ng/L, group 5 26 ng/L, p ⬍0.001). Adjusted mortality risk also increased progressively in the 3 older groups compared to the younger groups (HRs 2.5, 3.4, 5.4, respectively, p ⬍0.001 for all comparisons), but there was no interaction with hs-cTnT-related risk (p ⫽ 0.45 for interaction). Patients with and without CAD had similar mortality rates (multivariate adjusted HR 1.0, p ⫽ 0.81). Median concentrations of hs-cTnT were higher in patients with CAD (21 vs 15 ng/L, p ⬍0.001). CAD status modified the impact of hs-cTnT on mortality (p ⫽ 0.04 for interaction). The impact of hscTnT quartile 4 versus 1 on mortality was lower in the CAD group (univariate HR 3.2, 95% CI 1.9 to 5.4, p ⬍0.001; adjusted HR 1.6, 95% CI 0.8 to 2.9, p ⫽ 0.15) than in the group without CAD (univariate HR 8.0, 95% CI 3.8 to 16.9, p ⬍0.001; adjusted HR 6.8, 95% CI 2.6 to 17.5, p ⬍0.001). There was no interaction of CAD with the composite end point (p ⫽ 0.34). Discussion The main findings of this study are that hs-cTnT is increased in most outpatients with chronic systolic HF and that hs-cTnT is a strong and independent predictor of long-term all-cause mortality in these patients. The study population was comparable to other European HF populations attending similar HF clinics with respect to
age, gender distribution, prevalence of CAD, and mortality rate.7–10 The predictive ability of hs-cTnT remained significant for quartile 4, conferring a mortality HR of 2.6 independently of important clinical variables and of NT–proBNP, in itself a very powerful predictor of mortality in HF. When dichotomizing the population using the 99th gender-specific percentile, those with abnormal levels had a similar increase in adjusted risk of mortality (HR 1.7). The association of hs-cTnT with the composite end point of mortality and cardiovascular hospitalization was less pronounced. Although a trend toward increased risk estimates with increasing quartiles of hs-cTnT or ⬎99th percentile was noted, TnT did not add independent predictive power. Apparently hs-cTnT is a better predictor of all-cause mortality than cardiovascular hospitalization, which was also noted in the Valsartan Heart Failure Trial (Val-HeFT).11 Cardiovascular hospitalization can be expected to relate to worsening HF in most cases, whereas death outside the hospital often is sudden. Hs-cTnT might be especially suitable for estimating risk of sudden death in chronic HF. Future studies should address this issue. TnT concentrations were higher in patients with advanced age, which has been observed previously.12,13 A novel finding is the lack of a modifying effect of age and gender on the prognostic impact of hs-cTnT on mortality. Presence of CAD would be expected to increase hs-cTnT by ischemic mechanisms, and CAD was indeed associated with higher levels of hs-cTnT in the present study. However, the prognostic impact of hs-cTnT was much stronger in patients without CAD, conferring a sevenfold increase in adjusted mortality ratio in patients without CAD in quartile 4 versus 1. An increased hs-cTnT has been found in patients with HF and without CAD, which may be due to limited endocardial blood flow caused by increased wall stress or microvascular endothelial dys-
556
The American Journal of Cardiology (www.ajconline.org)
function.14 However, the reason for the interaction between hs-cTnT and CAD in relation to prognosis is unclear. One possible mechanism is undetected and untreated CAD in patients diagnosed with nonischemic cardiomyopathy (e.g., patients in whom only noninvasive testing was done). Furthermore, myocardial scarring may be present in patients with HF and normal coronary arteries. Extent of scarring has been associated with a cTnT increase in patients with chest pain and normal coronary arteries15 and may affect prognosis, but this has not been documented yet in patients with HF. Prevalence of hs-cTnT in ⬎99th percentile was 57%. Because this is also the cut-off value in the current universal definition of myocardial infarction,16 most outpatients with HF can be expected to meet this criterion. Emphasis must therefore be placed on the clinical context (e.g., relevant symptoms and electrocardiographic changes) and serial hs-cTnT measurements when distinguishing an HF-related increase in hs-cTnT from the characteristic increasing-and-decreasing pattern related to acute coronary syndromes. Our results confirm and extend the results of previous studies of cardiac troponin in HF populations in different clinical settings. An increased troponin level has been shown to be of prognostic value in hospitalized patients with acute HF17 and with decompensated chronic HF.18 Data from the Val-HeFT randomized clinical trial11 in 4,053 patients with stable HF revealed that hs-cTnT, which was measurable in 90% of participants, predicted all-cause mortality after adjustments for BNP. This population was very different from ours in several important parameters: The Val-HeFT population was approximately 10 years younger, more frequently men, follow-up time was considerably shorter (24 months), and only approximately 35% received a  blocker. A study in elderly primary care patients with suspected HF has shown that a substantial fraction (80%) presented with hs-cTnT above the detection limit.12 A direct comparison to this population is difficult because a large proportion had preserved LVEF (88%) and was in New York Heart Association class I (45%). Our study does, however, show comparable findings because 95% had detectable hs-cTnT. These findings suggest that HF in the absence of acute ischemia promotes leakage of myocyte troponins, and although mechanisms are still being explored, serial measurements of hs-cTnT could provide insight into disease progression in patients with and without CAD. Ishii et al19 demonstrated that an increase in cTnT from admission for decompensated HF to 2 months after initiation of medical therapy was associated with an increase in mortality and readmission rate independent of BNP. Also, data from the Val-HeFT study showed that serial measurements of hs-cTnT may be clinically relevant in outpatients with HF.11 There are some strengths and limitations to this study that should be recognized. Follow-up time was long, allowing for conclusions on the relation between hs-cTnT and risk beyond the time frame provided by Val-HeFT and, to our knowledge, by any other studies on hs-cTnT in outpatients with HF. The long inclusion period introduces a risk of bias because standards of treatment have
changed within this decade. Only about 30% of referrals to the HF clinic were included in the study period. Specific reasons for not including new referrals were not recorded. We tried to compensate for the long inclusion period by including a variable reflecting time of inclusion in multivariate models and this did not affect results significantly. We did not have data on implantable devices in the population, which could also contribute to differences in risk, and ideally should have been accounted for in the regression models. The study may have been underpowered in detecting risk related to modestly increased hs-cTnT. 1. Daniels LB, Laughlin GA, Clopton P, Maisel AS, Barrett-Connor E. Minimally elevated cardiac troponin T and elevated N-terminal pro– B-type natriuretic peptide predict mortality in older adults: results from the Rancho Bernardo Study. J Am Coll Cardiol 2008;52:450 – 459. 2. deFilippi CR, de Lemos JA, Christenson RH, Gottdiener JS, Kop WJ, Zhan M, Seliger SL. Association of serial measures of cardiac troponin T using a sensitive assay with incident heart failure and cardiovascular mortality in older adults. JAMA 2010;304:2494 –2502. 3. Kistorp C, Galatius S, Gustafsson F, Faber J, Corell P, Hildebrandt P. Prevalence and characteristics of diabetic patients in a chronic heart failure population. Int J Cardiol 2005;100:281–287. 4. Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJ, Ponikowski P, Poole-Wilson PA, Strömberg A, van Veldhuisen DJ, Atar D, Hoes AW, Keren A, Mebazaa A, Nieminen M, Priori SG, Swedberg K; ESC Committee for Practice Guidelines (CPG). ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the task force for the diagnosis and treatment of acute and chronic heart failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur J Heart Fail 2008;10:933–989. 5. Mingels A, Jacobs L, Michielsen E, Swaanenburg J, Wodzig W, van Dieijen-Visser M. Reference population and marathon runner sera assessed by highly sensitive cardiac troponin T and commercial cardiac troponin T and I assays. Clin Chem 2009;55:101–108. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease study group. Ann Intern Med 1999;130:461– 470. 7. Gustafsson F, Schou M, Videbaek L, Dridi N, Ryde H, Handberg J, Hildebrandt PR. Danish Heart Failure Clinics Network. Incidence and predictors of hospitalization or death in patients managed in multidisciplinary heart failure clinics. Eur J Heart Fail 2009;11:413– 419. 8. de la Porte PW, Lok DJ, van Veldhuisen DJ, van Wijngaarden J, Cornel JH, Zuithoff NP, Badings E, Hoes AW. Added value of a physician-and-nurse-directed heart failure clinic: results from the Deventer-Alkmaar heart failure study. Heart 2007;93:819 – 825. 9. Senni M, De Maria R, Gregori D, Gonzini L, Gorini M, Cacciatore G, Gavazzi A, Pulignano G, Porcu M, Maggioni AP. Temporal trends in survival and hospitalizations in outpatients with chronic systolic heart failure in 1995 and 1999. J Card Fail 2005;11:270 –278. 10. Pons F, Lupón J, Urrutia A, González B, Crespo E, Díez C, Cano L, Cabanes R, Altimir S, Coll R, Pascual T, Valle V. Mortality and cause of death in patients with heart failure: findings at a specialist multidisciplinary heart failure unit. Rev Esp Cardiol 2010;63:303–314. 11. Latini R, Masson S, Anand IS, Missov E, Carlson M, Vago T, Angelici L, Barlera S, Parrinello G, Maggioni AP, Tognoni G, Cohn JN. Val-HeFT Investigators. Prognostic value of very low plasma concentrations of troponin T in patients with stable chronic heart failure. Circulation 2007;116:1242–1249. 12. Alehagen U, Dahlström U, Rehfeld JF, Goetze JP. Prognostic assessment of elderly patients with symptoms of heart failure by combining high-sensitivity troponin T and N-terminal pro–B-type natriuretic peptide measurements. Clin Chem 2010;56:1718 –1724. 13. Koerbin G, Tate JR, Hickman PE. Analytical characteristics of the Roche highly sensitive troponin T assay and its application to a cardio-healthy population. Ann Clin Biochem 2010;47:524 –528.
Heart Failure/Troponin T and Prognosis in Heart Failure 14. Thygesen K, Mair J, Katus H, Plebani M, Venge P, Collinson P, Lindahl B, Giannitsis E, Hasin Y, Galvani M, Tubaro M, Alpert JS, Biasucci LM, Koenig W, Mueller C, Huber K, Hamm C, Jaffe AS. Study Group on Biomarkers in Cardiology of the ESC Working Group on Acute Cardiac Care. Recommendations for the use of cardiac troponin measurement in acute cardiac care. Eur Heart J 2010;31:2197–2204. 15. Christiansen JP, Edwards C, Sinclair T, Armstrong G, Scott A, Patel H, Hart H. Detection of myocardial scar by contrast-enhanced cardiac magnetic resonance imaging in patients with troponin-positive chest pain and minimal angiographic coronary artery disease. Am J Cardiol 2006;97:768 –771. 16. Thygesen K, Alpert JS, White HD. Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction. Universal
557
definition of myocardial infarction. J Am Coll Cardiol 2007; 50:2173–2195. 17. Peacock WFt, De Marco T, Fonarow GC, Diercks D, Wynne J, Apple FS, Wu AH. Cardiac troponin and outcome in acute heart failure. N Engl J Med 2008;358:2117–2126. 18. Koide K, Yoshikawa T, Nagatomo Y, Kohsaka S, Anzai T, Meguro T, Ogawa S. Elevated troponin T on discharge predicts poor outcome of decompensated heart failure. Heart Vessels 2010;25:217–222. 19. Ishii J, Cui W, Kitagawa F, Kuno T, Nakamura Y, Naruse H, Mori Y, Ishikawa T, Nagamura Y, Kondo T, Oshima H, Nomura M, Ezaki K, Hishida H. Prognostic value of combination of cardiac troponin T and B-type natriuretic peptide after initiation of treatment in patients with chronic heart failure. Clin Chem 2003;49:2020 –2026.