Cardiac troponin T (cTnT) and B-type natriuretic peptide (BNP) have been used to estimate prognosis in heart failure. However, most studies have evaluated decompensated patients using single measurements. To determine the value of serial measurements, we evaluated 190 stable chronic heart failure patients every 3 months during 2 years.
Clinical events do not always accompany the advancement of chronic heart failure.1 To identify chronic heart failure patients at highest risk, the assessment of cardiac troponin T (cTnT) and B-type natriuretic peptides (BNP and NT-proBNP) levels has been used.2,3 Previous studies have evaluated BNP and cTnT measurements in acutely decompensated hospitalized patients at a single point in time. To obtain additional information about the prognosis of clinically stable patients and to establish the most favorable monitoring pattern, we evaluated cardiac cTnT and BNP levels every 3 months for 2 years in nonhospitalized stable subjects with chronic heart failure.
Subjects and methods
Two hundred subjects with New York Heart Association class III and IV heart failure were prospectively enrolled in our study. After a baseline evaluation, subjects were followed up every 3 months (±3 weeks) for 24 months. Subjects who had experienced acute heart failure decompensation 30 days before enrollment were excluded from the study. We also excluded subjects who were scheduled to undergo cardiac transplantation or revascularization within 6 months. The Mayo Foundation Institutional Review Board approved the study.
Following established protocols, cTnT and BNP levels were assessed at the start of the study and every 3 months for 2 years. All investigators and clinicians were blinded to the cTnT and BNP results. A physical examination was performed at each follow-up appointment, along with assessment of any changes since the last visit, including hospitalizations. Subjects were followed up for 18.9 ± 7.8 (range, 0.13-26.4) months.
Cardiac troponin T levels were assessed using the Roche Elecsys third-generation troponin T immunoassay (Roche Diagnostics, Indianapolis, IN). Levels equal to or above the 99th percentile of the normal reference population (≥0.01 ng/mL) were considered to be elevated, and the analyses were also performed using the conventional clinical practice cut-off value of >0.03 ng/mL. B-type natriuretic peptide levels were evaluated using the Shionoria assay method. Levels above the 95th percentile of normal based on sex and age were considered elevated.4 We used the Modification of Diet in Renal Disease equation for estimated glomerular filtration rate to assess renal function at baseline and at all follow-up appointments.
SAS software (version 9; SAS Institute Inc., Cary, NC) was used to perform statistical analyses. Categorical variables are reported as a percent of the total values and continuous variables as the mean ± standard deviation (SD) and the median with 25% and 75% percentiles. Time to death or cardiac transplantation was the primary end point, with a secondary end point of time to first hospitalization for decompensated heart failure. B-type natriuretic peptide levels were divided into 1 of 2 groups: either elevated or not elevated. Cardiac troponin T levels were divided into 1 of 3 groups: <0.01, 0.01-0.03, and >0.03 ng/mL. We calculated univariate Cox proportional hazard models at baseline for the end points, which are represented as hazard ratios (HRs) and 95% confidence intervals. We used Shoenfeld residuals to test proportional hazard assumptions. We also used forward stepwise methods to create multivariable models and time-dependent analysis methods to evaluate serial values over time. Survival curves reflect serial measurements and changes demonstrated over time. We employed time-dependent Cox models to assess changes between the 3-month periods.
Data were reported for 190 subjects because 10 subjects did not have sufficient information. Baseline BNP levels were elevated in 64.2% of subjects (n = 122). Eighty-seven subjects (45.8%) had baseline cTnT levels <0.01 ng/mL, 50 subjects (26.3%) had levels from 0.01 to 0.03 ng/mL, and 53 subjects (27.9%) had levels >0.03 ng/mL. The baseline mean duration of heart failure was 41.9 ± 44.2 (median, 31) months. The characteristics of the subjects who participated in the study are shown in Table 1.
Table 1. Baseline patient characteristics.
During the follow-up period, 55 subjects (29%) died and 6 (3%) underwent cardiac transplantation. The highest risk of death or cardiac transplantation was shown to be associated with a cTnT level ≥0.01 ng/mL and an elevated BNP level (HR = 6.37) in an interaction model using baseline cTnT and BNP values. One hundred three subjects (54%) initially had elevated cTnT levels, and 31 (36%) developed cTnT elevation, with 70.5% having increased cTnT levels overall (103 + 31/190). Baseline BNP levels were elevated in 122 subjects (64%), and 31 subjects developed increased BNP levels. As shown in Figure 1, survival for subjects with a cTnT level >0.03 ng/mL at any point during the study was significantly worse (P <.001). Survival was also worse for subjects with elevated BNP levels at the last follow-up visit preceding events (P <.001).
Figure 1. (A) Kaplan-Meier curves for time-dependent events of death or cardiac
transplantation for troponin T levels grouped by category-based serial follow-up
values. The number of patients at risk initially for each group is shown in
parentheses. (B) Kaplan-Meier curves for time-dependent events of death or
cardiac transplantation for B-type natriuretic peptide (BNP) levels grouped by
category-based serial follow-up values. (Reprinted with permission from Miller WL,
Hartman K, Burrit M, et al. Serial biomarker measurements in ambulatory patients
with chronic heart failure: the importance of change over time. Circulation. 2007;
There was a 3.38-fold greater risk of death or cardiac transplantation among subjects with an increase in cTnT level >0.01 ng/mL from one follow-up visit to the next or if subjects had a persistently elevated cTnT level (Table 2). Greater increases in cTnT level accentuated the risk (<20% increase, HR = 3.40; ≥20% increase, HR = 2.02). Elevated BNP levels at any time predicted an increased risk of death or cardiac transplantation, and further changes in BNP level, whether increases or decreases, did not reduce the risk (HR = 4.94; Table 3).
Table 2. Effect of change in serial troponin T values on the outcome of death/
cardiac transplantation during the study period.
Table 3. Effect of change in serial B-type natriuretic peptide (BNP) levels on
outcome of death/cardiac transplantation during the study period.
The risk of the end point increased significantly when cTnT alone was stratified at levels of ≥0.01 ng/mL and >0.03 ng/mL, especially at values of >0.03 ng/mL (Figure 2). Although less powerful, BNP elevations alone also predicted greater risk; when both increased cTnT and BNP levels were used, the risk increased markedly. The HR was 5.01 when a cTnT cut-off level of >0.01 ng/mL was used. When a cut-off level of >0.03 ng/mL was used, the HR was 8.26. The risk did not increase when an elevated BNP level was added (HR = 8.58). During the study, 71 subjects were hospitalized for decompensated heart failure, and there were 150 heart-related hospitalizations; the risk pattern was the same for these subjects as for nonhospitalized subjects.
Figure 2. Risk of death/cardiac transplantation. Time-dependent multivariate model using serial follow-up cardiac troponin T (cTnT) and B-type natriuretic peptide (BNP) values. + Indicates elevated BNP or cTnT level ≥0.01 ng/mL or >0.03 ng/mL; -, not elevated (hazard ratio [HR] = 1.0 for elevation of neither BNP nor cTnT). (Reprinted with permission from Miller WL, Hartman K, Burrit M, et al. Serial biomarker measurements in ambulatory patients with chronic heart failure: the importance of change over time. Circulation. 2007;116:249-257.)
Our findings add to current data on the significance of biomarkers in the assessment and prognosis of patients with heart failure. These results concur with the findings from prior studies3-6 showing the prognostic effects of single baseline elevations in cTnT and BNP levels. In our study, 103 subjects (54%) had a baseline cTnT level >0.01 ng/mL, and 53 (28%) had levels >0.03 ng/mL. Subsequent elevations in cTnT (particularly >0.30 ng/mL) and BNP levels during the follow-up period also independently correlated with greater short-term risk. Thirty-one subjects (16%) had new elevations of cTnT, which transferred them into higher risk groups. When the cTnT level returned to normal, the risk was decreased. These findings show that the use of baseline and follow-up cTnT levels are valuable, and methods to prevent new elevations or to intercede should they occur are needed.
Over time, increased BNP levels did not have the same effect as increased cTnT levels. Although subjects with new elevations above normal had greater risk, once the BNP level was elevated, further increases or decreases did not make any difference. Long-term observation of BNP levels, therefore, might not be effective, as previously proposed.7,8 This could be a result of the fact that the subjects in our study had chronic stable heart failure instead of acutely decompensated heart failure. The types of BNP fragments elaborated may also be of critical importance.9
The greater risk seen with our subjects is comparable to that seen during acute decompensation resulting in hospitalization. In the setting of hospitalization, the acute event significantly affects the prognosis.10 The majority of the biomarker increases seen in our study did not result in hospitalization, but they may represent comparable pathophysiology without overt decompensation; therefore, these biomarker elevations may have prognostic value similar to hospitalization. In addition, these results indicate that left ventricular volume overload and detectable and subclinical myocardial injury may not occur gradually over time but instead may take place in a stepwise manner. These events may indicate myocardial apoptosis,11 and are in accordance with the notion that heart failure—related infarctions are a factor in the progressive decompensation of myocardial function. This may be why there is an association between elevated cTnT levels and adverse ventricular remodeling.12
Increased BNP levels resulted in a 4-fold greater risk of death or cardiac transplantation, and a cTnT level >0.30 ng/mL resulted in a 4- to 8-fold greater risk. The combination of increased cTnT and BNP levels resulted in an 8-fold greater risk. When cTnT levels decreased to normal, the risk was decreased; however, this was not the case for BNP levels. This finding may indicate that a vulnerable phase, in which patients are at increased risk, may exist, and strategies to prevent such events should be adopted. Assessment of cTnT levels would benefit these patients. Although similar results might be expected with monitoring of BNP levels, this was not borne out in the results of our study. Once the BNP level was elevated, increases or decreases in the level did not affect the prognosis. This may be because of the high variability in BNP13 or because high BNP levels may have prognostic value in some patients, whereas low levels may have prognostic value in others.14
The results of our study showed that elevations of cTnT (>0.03 ng/mL) or BNP levels among clinically stable, nonhospitalized, chronic heart failure subjects over time were associated with an increased risk of death, cardiac transplantation, and/or hospitalization. Combined elevations of cTnT and BNP levels defined the highest risk group. The ability to monitor changes in cTnT and BNP levels by serial measurements adds to the assessment of risk.
This study was supported in part by a grant from Dade-Behring, Deerfield, IL.