Author + information
- Received April 1, 2003
- Revision received June 8, 2003
- Accepted July 29, 2003
- Published online December 3, 2003.
- Tomas Jernberg, MD, PhD*,§,* (, )
- Bertil Lindahl, MD, PhD*,§,
- Agneta Siegbahn, MD, PhD†,
- Bertil Andren, MD, PhD‡,
- Gunnar Frostfeldt, MD, PhD*,
- Bo Lagerqvist, MD, PhD*,
- Mats Stridsberg, MD, PhD†,
- Per Venge, MD, PhD† and
- Lars Wallentin, MD, PhD*,§
- ↵*Reprint requests and correspondence:
Dr. Tomas Jernberg, Department of Cardiology, Cardiothoracic Center, University Hospital, 751 85 Uppsala, Sweden.
Objectives We sought to examine whether measurements of N-terminal pro-brain natriuretic peptide (NT-proBNP), in addition to cardiac troponin T (cTnT) and interleukin-6 (IL-6), improve the ability to identify high-risk patients who benefit from an early invasive strategy.
Background Biochemical indicators of cardiac performance (e.g., NT-proBNP), inflammation (e.g., IL-6), and myocardial damage (e.g., cTnT) predict mortality in unstable coronary artery disease (UCAD) (i.e., unstable angina or non–ST-segment elevation myocardial infarction [MI]). In these patients, an early invasive treatment strategy improves the outcome.
Methods Levels of NT-proBNP, cTnT, and IL-6 were measured in 2,019 patients with UCAD randomized to an invasive or non-invasive strategy in the FRagmin and fast revascularization during InStability in Coronary artery disease (FRISC-II) trial. Patients were followed up for two years to determine death and MI.
Results Patients in the third NT-proBNP tertile had a 4.1-fold (95% confidence interval [CI] 2.4 to 7.2) and 3.5-fold (95% CI 1.8 to 6.8) increased mortality in the non-invasive and invasive groups, respectively. An increased NT-proBNP level was independently associated with mortality. In patients with increased levels of both NT-proBNP and IL-6, an early invasive strategy reduced mortality by 7.3% (risk ratio 0.46, 95% CI 0.21 to 1.00). In patients with lower NT-proBNP or IL-6 levels, the mortality was not reduced. Only elevated cTnT was independently associated with future MI and a reduction of MI by means of an invasive strategy.
Conclusions N-terminal proBNP is independently associated with mortality. The combination of NT-proBNP and IL-6 seems to be a useful tool in the identification of patients with a definite survival benefit from an early invasive strategy. Only cTnT is independently associated with future MI and a reduction of MI by an invasive strategy.
The FRagmin and fast revascularization during InStability in Coronary artery disease (FRISC-II) trial is the only large-scale study that has demonstrated an improved survival and a reduction in myocardial (re)infarction (MI) by means of an early invasive strategy in patients with unstable coronary artery disease (UCAD) (i.e., unstable angina or non–ST-segment elevation MI) (1–3). These results were recently supported by two other large trials with lower rates of composite events in the invasive cohorts after four to six months (4,5). However, patients with UCAD vary regarding both prognosis and response to treatment.
In many trials, indicators of myocardial ischemia (e.g. ST-segment depression) (6)and myocardial necrosis (e.g., troponin) (7)have been associated with raised mortality and risk of MI. In the FRISC-II trial and in another trial, patients with ST-segment depression, elevated troponin, or a high-risk score had a more severe prognosis and a greater benefit from an invasive strategy (4,8). Increased levels of inflammatory markers, such as C-reactive protein and interleukin-6 (IL-6), have also been associated with an adverse outcome in this population (7,9). In the FRISC-II trial, IL-6 was an independent predictor of mortality, even when measurements with a high-sensitivity C-reactive protein method were included in the analysis (9). Moreover, the mortality was markedly reduced by an invasive strategy in patients with increased IL-6 levels, whereas there was no effect on mortality in patients with lower levels of IL-6 (9).
Brain natriuretic peptide (BNP) and the N-terminal part of its prohormone, NT-proBNP, are released mainly from the cardiac ventricles in response to increased stretch and wall tension (10). Recently, several studies have shown that BNP and NT-proBNP are also useful in the prediction of outcome in patients with UCAD (11–15). However, none of these studies has related the prognostic value of a natriuretic peptide to other indicators of left ventricular dysfunction, myocardial necrosis, and inflammation or the effect of an early invasive strategy.
The aim of this study was twofold: 1) to investigate the prognostic value of NT-proBNP in relation to other important predictors of outcome; and 2) to examine whether NT-proBNP, in addition to other prognostic markers, improves the identification of high-risk patients who benefit from an early invasive strategy.
Patients and study design
This study was a substudy of the invasive/non-invasive part of the FRISC-II trial. The protocol and results after 6, 12, and 24 months have been published elsewhere (1–3). In brief, the objective was to compare an early invasive to an early non-invasive strategy in 2,457 patients with UCAD. Myocardial ischemia had to be verified by an electrocardiogram (ST-segment depression ≥0.10 mV or T-wave inversion ≥0.10 mV) or by raised biochemical markers. Excluded were patients with an indication for or treated within the past 24 h with thrombolysis, angioplasty within the last six months, previous open-heart surgery, advanced age, or other conditions that made randomization to early revascularization inappropriate. All patients received aspirin and open-label dalteparin for at least five days and, in the invasive group, always until the evening before revascularization. Informed, written consent was obtained from all patients, and the protocol was approved by all local ethics committees.
In the invasive strategy, the aim was to perform coronary angiography and, if appropriate, revascularization within seven days of admission, whereas patients randomized to a non-invasive strategy underwent coronary angiography only if refractory or recurrent symptoms occurred or if they showed signs of severe ischemia on a pre-discharge exercise test. In each group, patients were also randomized to continued treatment with subcutaneous dalteparin or placebo for three months. In this substudy, patients were followed up for two years to determine death and MI.
Samples of EDTA-plasma were obtained at randomization and stored frozen at −70°C in aliquots. Analyses were performed at the Department of Clinical Chemistry, Uppsala, Sweden. Plasma NT-proBNP was determined using the Elecsys proBNP sandwich immunoassay on an Elecsys 2010 (Roche Diagnostics, Uppsala). The analytical range extends from 20 to 35,000 ng/l. At our laboratory, the total coefficient of variation was 3.3% (n = 21) at a level of 209 ng/l and 3.0% (n = 21) at a level of 7,431 ng/l. Men and women were, according to the NT-proBNP level, separately classified into tertiles (Table 1). Plasma cardiac troponin T (cTnT) was determined by the third-generation cTnT assay on an Elecsys 2010 (Roche Diagnostics). A cut-off level of 0.03 μg/l was used. Levels of IL-6 were analyzed with Immulite (Diagnostic Products Corp.). The lower detection limit of this assay is 5 ng/l, which was used as a cut-off level.
The study protocol recommended a standard echocardiographic examination performed in the hospital and before coronary angiography. All examinations were stored on videotape and analyzed at each participating center. From the parasternal and apical views, the left ventricular ejection fraction (LVEF) was visually estimated and graded as normal (LVEF >0.55), mildly depressed (0.45 to 0.55), moderately depressed (0.30 to 0.44), and severely depressed (<0.30). In patients randomized to an early invasive treatment, coronary angiography was performed within seven days of inclusion. The assessment was performed by the operator and was based on a detailed evaluation form. Significant stenosis was considered present if there was a reduction of ≥50% in the vessel diameter.
Differences in proportions were judged by chi-square analysis. If not stated otherwise, continuous data are given as the median value (25th to 75th percentile). The Kruskal-Wallis test was used to test the equality of distributions in groups according to the level of NT-proBNP. Further pairwise multiple comparison tests were not performed. A significant difference was considered to exist at p < 0.05. The Kaplan-Meier method was used to illustrate the timing of events. The risk ratio (RR; i.e., ratio of proportions having the event in each group) with 95% confidence interval (CI) was used to express a difference in outcome. To examine whether NT-proBNP level was independently associated with mortality, logistic regression analysis was used. Deaths not caused by MI resulted in a shorter follow-up time regarding the risk of MI. Therefore, Cox regression analysis was performed regarding the risk of future MI. A backward selection was used for both regression analyses. The following variables were entered: age ≥70 years (yes/no), male gender (yes/no), diabetes mellitus (yes/no), previous MI (yes/no), ST-segment depression (yes/no), cTnT ≥0.03 μg/l (yes/no), IL-6 ≥5 ng/l (yes/no), and third-tertile NT-proBNP (yes/no).
IL-6 and cTnT in relation to NT-proBNP
Of 2,457 patients, 2,019 (82%), 2,288 (93%), and 2,329 (95%) had NT-proBNP, IL-6, and cTnT measured at randomization, respectively, after a median time of 39 h (range 27 to 54) from the last episode of symptoms. The median value of NT-proBNP was 535 ng/l (range 212 to 1,192) for men and 672 ng/l (283 to 1,820) for women. A total of 635 patients (28%) had IL-6 ≥5 ng/l and 1,587 (65%) had cTnT ≥0.03 μg/l. Clinical background factors in relation to tertiles of NT-proBNP levels in men and women, respectively, are listed in Table 1, whereas their relations to IL-6 and cTnT have previously been published (9,16). Patients with IL-6 <5 ng/l had lower NT-proBNP levels than those with IL-6 ≥5 ng/l (446 [range 181 to 1,055] vs. 1,060 [441 to 2,217] ng/l) In the latter group, there was a weak correlation between NT-proBNP and IL-6 (rho 0.29, p < 0.001). Patients with cTnT <0.03 μg/l had lower NT-proBNP levels than those with cTnT ≥0.03 μg/l (219 [range 107 to 480] vs. 860 [396 to 1,720] ng/l). In the latter group, there was a moderate correlation between NT-proBNP and cTnT (rho 0.53, p < 0.001).
Echocardiography and coronary angiography in relation to NT-proBNP
Echocardiographic evaluation of LVEF was available in 1,944 patients (79%), 1,633 of whom had NT-proBNP determined. The median time from admission to examination was three days (range 2 to 4). The prevalence of left ventricular dysfunction increased with increasing NT-proBNP levels (Table 1). In the invasive group, 1,011 had NT-proBNP measured at randomization. Of these, 994 (98%) had coronary angiography performed. The proportion of patients with more severe coronary artery disease was higher at higher NT-proBNP levels (Table 1).
Prognostic value in the non-invasive group
In the non-invasive group, patients in the first and second tertiles had about the same mortality, whereas patients in the third tertile had a 4.1-fold (95% CI 2.4 to 7.2) higher mortality (Fig. 1A). Concerning MI, there were no significant differences between each tertile or groups of tertiles (Fig. 2). The outcome in relation to cTnT and IL-6 is summarized in Table 2(part A). Patients with cTnT ≥0.03 μg/l had a non-significant increase in mortality (RR 1.59, 95% CI 0.88 to 2.92) and a significantly increased risk of future MI (RR 2.35, 95% CI 1.55 to 3.57). Patients with IL-6 ≥5 ng/l had a higher mortality (RR 2.72, 95% CI 1.68 to 4.42), but no raised risk of MI (RR 0.98, 95% CI 0.70 to 1.37).
Prognostic value in the invasive group
Also in the invasive group, patients in the first and second tertiles had similar mortality, whereas patients in the third tertile had a 3.5-fold (95% CI 1.8 to 6.8) higher mortality (Fig. 1A). Concerning MI, there were no significant differences between tertiles (Fig. 2). The outcome in relation to cTnT and IL-6 is summarized in Table 2(part B). Patients with cTnT ≥0.03 μg/l had no significant increase in mortality (RR 1.30, 95% CI 0.66 to 2.57), but still had an increased risk of MI (RR 1.59, 95% CI 1.04 to 2.46) (Fig. 2). Patients with IL-6 ≥5 ng/l had neither raised mortality (RR 1.24, 95% CI 0.65 to 2.37) nor a higher risk of MI (RR 1.11, 95% CI 0.74 to 1.65).
In a multivariate analysis including only non-invasively treated patients, increasing age, diabetes mellitus, previous MI, and elevated IL-6 were independently associated with mortality (Table 3, model 1). When NT-proBNP was added to the model, NT-proBNP was independently associated with mortality, whereas the association between IL-6 and mortality was weakened (p = 0.09) (Table 3, model 2). In the invasive group, increasing age, male gender, diabetes mellitus, previous MI, and ST-segment depression were independent predictors of death (Table 3, model 1). When NT-proBNP was added, NT-proBNP, but not increasing age, was an independent predictor of death (Table 3, model 2). A categorization of cTnT into quartiles did not affect the final models.
The LVEF did not affect the association between NT-proBNP and mortality. Of the patients with LVEF <0.45, only those with raised levels of NT-proBNP had a higher mortality, whereas all patients with raised levels of NT-proBNP had a higher mortality, regardless of LVEF and treatment strategy (Fig. 3).
Concerning the risk of subsequent MI, cTnT was an independent predictor in the non-invasive (RR 2.02, 95% CI 1.30 to 3.14) and invasive (RR 1.61, 95% CI 1.03 to 2.52) groups. N-terminal proBNP and IL-6 were not independently associated with the risk of MI in any group.
Effect of an early invasive strategy
When patients were stratified according to tertiles of NT-proBNP level, the mortality in the invasive group was numerically reduced by 3.6% (RR 0.67, 95% CI 0.41 to 1.10) in the third tertile and by 0.6% (RR 0.78, 95% CI 0.39 to 1.57) in the first and second tertiles (Fig. 1A). When the patient data were divided according to IL-6 level, the invasive strategy reduced mortality by 5.7% (RR 0.42, 95% CI 0.22 to 0.78) at IL-6 ≥5 ng/l, but by only 0.3% at IL-6 <5 ng/l (RR 0.91, 95% CI 0.55 to 1.52).
The effect of a combination of these to risk markers on mortality is shown in Figure 1B and 1C. In patients with NT-proBNP in the first or second tertile, the mortality was low regardless of the IL-6 level and not influenced by an invasive strategy. In patients with NT-proBNP in the third tertile, but with IL-6 <5 ng/l, the mortality was intermediate but still not influenced by an invasive strategy. In patients in the third tertile of NT-proBNP with IL-6 ≥5 ng/l, the mortality was highest, but with a reduction of 7.3% (RR 0.46, 95% CI 0.21 to 1.00) by the invasive strategy.
Concerning the risk of subsequent MI, the effect of an invasive strategy was greater in the second and third tertiles of NT-proBNP (Fig. 2), whereas the level of IL-6 was unrelated to the risk of MI. However, there remained no difference in treatment effect in relation to NT-proBNP when patients were stratified according to the cTnT level (Fig. 2). Thus, the greater effect of an invasive strategy on the risk of MI in the second and third tertiles of NT-proBNP was explained by the underlying relation of both factors to the cTnT level.
Recently, the acute-phase levels of BNP and NT-proBNP have been associated with mortality and an increased risk of future MI in patients with UCAD (11–15). However, also other biochemical markers (e.g., troponin, C-reactive protein, IL-6) and LVEF are related to the outcome in UCAD. Furthermore, to be useful in clinical practice, a predictor should also improve the selection of the most effective treatment strategy. When evaluating NT-proBNP, we wanted to challenge it against the best predictors of outcome and treatment efficacy. As a marker of inflammation, we therefore chose IL-6, which in the FRISC-II material had been identified as the inflammatory marker with the best predictive value (9). As a marker of myocardial damage, we chose cTnT, as this marker had been measured in more patients than other markers and with a similar sensitivity and specificity as that of cardiac troponin I (17).
This study demonstrates that the level of NT-proBNP predicts mortality in UCAD, independent to other known predictors of outcome and regardless of treatment strategy. Moreover, the combination of NT-proBNP and IL-6 seems to be a useful tool in the identification of patients with a definite survival benefit from an early invasive strategy. Finally, the study demonstrates that neither of these indicators, but only the cTnT level, identifies patients with a raised risk of MI and an effect of an invasive strategy on subsequent MI.
In healthy subjects, women have higher levels of natriuretic peptides than men (18). Therefore, it seems appropriate to use different cut-off values for men and women when identifying different risk groups. There is also an increase in levels of natriuretic peptide with age (18). However, because this probably reflects a higher prevalence of subclinical cardiac conditions in the elderly (19), it seems less appropriate to adjust the cut-off values for age.
In this study, higher levels of NT-proBNP in the acute phase of UCAD were associated with several indicators of poor cardiac performance and increased risk, which is in accordance with previous studies (11,14). There was a significant correlation between cTnT and NT-proBNP, whereas the correlation between IL-6 and NT-proBNP was weaker. This indicates that the NT-proBNP level is influenced by the size of myocardial damage, whereas its relation to the inflammatory response is more variable.
The present study shows that NT-proBNP is independently associated with mortality in both non-invasively and invasively treated patients. In multivariate analyses, NT-proBNP, but not LVEF, was independently associated with mortality. When the methods were compared, patients with raised levels of NT-proBNP had a high mortality, regardless of LVEF, whereas LVEF <0.45 resulted in a high mortality only if NT-proBNP was elevated.
There are several possible reasons why NT-proBNP is a more important predictor of outcome than LVEF determined by echocardiography in the present population. Previous studies have shown that the neurohormonal response, including raised levels of NT-proBNP, is independently associated with mortality in patients with left ventricular dysfunction (20). Patients with high levels of NT-proBNP may have had temporary left ventricular dysfunction secondary to transient ischemia, jeopardizing a large part of the myocardium. Elevations of BNP have also been demonstrated shortly after percutaneous coronary intervention (21). A raised level of NT-proBNP in patients with normal or slightly depressed LVEF may also indicate diastolic dysfunction (22), which can also contribute to a worse outcome. Finally, the performance and interpretation of echocardiography are operator dependent, with considerable interobserver variability, which might have contributed to the superiority of NT-proBNP in the present setting.
The mortality benefit of an early invasive strategy was greater in patients with raised NT-proBNP levels. An obvious reason is the low mortality at low NT-proBNP levels, making further improvements difficult. Another reason is probably that a high NT-proBNP level was associated both with more severe coronary artery disease and reduced LVEF. Also, in previous randomized trials of an invasive versus non-invasive strategy, subgroups with multivessel disease and reduced LVEF obtained improved survival by the invasive approach (23,24).
Interestingly, the improved survival of an invasive strategy was found only in patients with simultaneous high levels of NT-proBNP and IL-6. This suggests that patients with increased NT-proBNP and an inflammatory response have a reversible condition that can improve by revascularization. Those with high NT-proBNP levels, but without an inflammatory response, on the other hand, may have irreversible myocardial dysfunction that will not be affected by an intervention. Several possible mechanisms might explain these findings. Increased inflammatory activity has been associated with both stunned and hibernating myocardium (25–28). Therefore, it can be hypothesized that patients with increased NT-proBNP and IL-6 levels more often have areas of stunned or hibernating myocardium that may be reversed by revascularization. The inflammatory response to the ischemic insult may also have long-lasting negative effects on myocardial performance (29–31)that might be reduced by revascularization.
In this study, neither the NT-proBNP nor IL-6 level was an independent predictor of future MI. Concerning NT-proBNP, this is in accordance with Omland et al. (12), who found no differences in natriuretic peptide levels between patients with and without subsequent MI. Other studies have unfortunately not presented any multivariate analyses regarding the relationship between BNP or NT-proBNP and the risk of MI (11,13–15). In accordance with previous studies, elevation of cTnT was the biochemical marker with the strongest relation to future MI. Even if there was a more pronounced reduction of MI after an early invasive strategy in patients with higher NT-proBNP levels, this was entirely explained by the connection to the cTnT level.
There are some limitations to this study. First, because of previous analyses, plasma samples for analyzing NT-proBNP were not available in 18% of the study group. The baseline characteristics and outcomes were similar to those in the main study, and any selection bias should be small. However, some subgroups, especially when adding echocardiographic data, are small, which may influence the stability of the estimates. Second, details on the causes of death are not known. Such knowledge would have shed more light on the mechanisms explaining our findings.
There are several indicators of increased risk and a more beneficial effect of an early invasive strategy. To meet the need of more comprehensive risk stratification, scoring systems related to both risk and the effect of treatment have been constructed (4,5). Based on previous data (11–15)and the present findings, it is evident that measurement of natriuretic peptides improves the early assessment of patients with UCAD and should therefore be included in future risk scores and guidelines.
We thank Lars Berglund for statistical advice and Birgitta Fahlström for excellent technical assistance. We also gratefully acknowledge the dedicated work of the research nurses, laboratory assistants, monitors, investigators, coordinators, and Data Safety and Monitoring Board.
☆ This study was supported by grants from the Swedish Heart and Lung Foundation, Uppsala County Association Against Heart and Lung Diseases, and Swedish Research Council, Pharmacia AG, and Roche Diagnostics, all in Sweden. Dr. Gottlieb Friesinger acted as the Guest Editor for this manuscript.
- brain natriuretic peptide
- cardiac troponin T
- FRagmin and fast revascularization during InStability in Coronary artery disease
- left ventricular ejection fraction
- myocardial infarction
- N-terminal pro-brain natriuretic peptide
- odds ratio
- risk ratio
- unstable coronary artery disease
- Received April 1, 2003.
- Revision received June 8, 2003.
- Accepted July 29, 2003.
- American College of Cardiology Foundation
- the FRISC II Investigators,
- Wallentin L.,
- Lagerqvist B.,
- Husted S.,
- Kontny F.,
- Stahle E.,
- Swahn E.
- the FRagmin during Instability in Coronary artery disease (FRISC) Study Group,
- Lindahl B.,
- Toss H.,
- Siegbahn A.,
- Venge P.,
- Wallentin L.
- Sabatine M.S.,
- Morrow D.A.,
- de Lemos J.A.,
- et al.
- Jernberg T.,
- Stridsberg M.,
- Venge P.,
- Lindahl B.
- Omland T.,
- Persson A.,
- Ng L.,
- et al.
- Lindahl B.,
- Diderholm E.,
- Lagerqvist B.,
- Venge P.,
- Wallentin L.
- the Australia–New Zealand Heart Failure Group,
- Richards A.M.,
- Doughty R.,
- Nicholls M.G.,
- et al.
- Lubien E.,
- DeMaria A.,
- Krishnaswamy P.,
- et al.
- Kalra D.K.,
- Zhu X.,
- Ramchandani M.K.,
- et al.
- Lagrand W.K.,
- Niessen H.W.,
- Wolbink G.J.,
- et al.
- Bozkurt B.,
- Kribbs S.B.,
- Clubb F.J. Jr.,
- et al.