Author + information
- Received August 13, 2001
- Revision received February 20, 2002
- Accepted February 25, 2002
- Published online May 15, 2002.
- Christian Zugck, MD*,
- Armin Haunstetter, MD*,
- Carsten Krüger, MD*,
- Robert Kell, MD*,
- Dieter Schellberg, MS*,
- Wolfgang Kübler, MD, FRCP* and
- Markus Haass, MD*,* ()
- ↵*Reprint requests and correspondence:
Dr. Markus Haass, Department of Cardiology, University of Heidelberg, Bergheimerstrasse 58, D-69115 Heidelberg, Germany.
Objectives This prospective study tested the impact of beta-blocker treatment on currently used risk predictors in congestive heart failure (CHF).
Background Given the survival benefit obtained by beta-blockade, risk stratification by factors established in the “pre–beta-blocker era” may be questioned.
Methods The study included 408 patients who had CHF with left ventricular ejection fraction (LVEF) <45%, all treated with an angiotensin-converting enzyme inhibitor or angiotensin type 1 receptor antagonist, who were classified into those receiving a beta-blocker (n = 165) and those who were not (n = 243). In all patients, LVEF, peak oxygen consumption (peakVo2), plasma norepinephrine (NE) and N-terminal pro-brain natriuretic peptide (NT-proBNP) levels were determined.
Results Although the New York Heart Association functional class (2.2 ± 0.7 vs. 2.3 ± 0.7), peakVo2(14.4 ± 5.2 ml/min per kg vs. 14.4 ± 5.5 ml/min per kg) and NT-proBNP (337 ± 360 pmol/l vs. 434 ± 538 pmol/l) were similar in the groups with and without beta-blocker treatment, the group with beta-blocker treatment had a lower heart rate (68 ± 30 beats/min vs. 76 ± 30 beats/min), lower NE (1.7 ± 1.2 nmol/l vs. 2.5 ± 2.2 nmol/l) and higher LVEF (24 ± 10% vs. 21 ± 9%; all p < 0.05). Within one year, 34% of patients without beta-blocker treatment, but only 16% of those with beta-blocker treatment (p < 0.001), reached the combined end point, defined as hospital admission due to worsening CHF and/or cardiac death. A beneficial effect of beta-blocker treatment was most obvious in the advanced stages of CHF, because the end-point rates were markedly lower (all p < 0.05) in the group with beta-blocker treatment versus the group without it, as characterized by peakVo2<10 ml/min per kg (26% vs. 64%), LVEF ≤20% (25% vs. 45%), NE >2.24 nmol/l (18% vs. 40%) and NT-proBNP >364 pmol/l (27% vs. 45%), although patients with beta-blocker treatment received only 37 ± 21% of the maximal recommended beta-blocker dosages.
Conclusions The prognostic value of variables used for risk stratification of patients with CHF is markedly influenced by beta-blocker treatment. Therefore, in the beta-blocker era, a re-evaluation of the selection criteria for heart transplantation is warranted.
Beta-blockade has been shown in several controlled trials to improve the prognosis and to reduce the cardiovascular hospitalization rate in patients with symptomatic congestive heart failure (CHF) (1–4). On the basis of these trials, beta-blocker treatment has been included in the current guidelines for CHF therapy and is now strongly recommended as standard therapy for all patients with CHF and systolic dysfunction (5,6). Long-term beta-blockade has consistently been shown to increase left ventricular ejection fraction (LVEF) in patients with CHF, whereas functional capacity, as determined by the New York Heart Association (NYHA) functional class, the 6-min walk test or peak oxygen consumption (peakVo2), derived from cardiopulmonary exercise testing, is not or only marginally improved (7,8). The plasma concentrations of norepinephrine (NE) decrease slightly with long-term beta-blocker treatment (9), whereas reports on plasma concentrations of brain natriuretic peptides are inconsistent (10). Left ventricular ejection fraction, peakVo2and NE (11), as well as brain natriuretic peptide and the N-terminal fragment of its precursor molecule (NT-proBNP) (12), have been proposed as independent predictors of survival in patients with CHF. However, most studies evaluating the prognostic variables in CHF have been carried out in study groups in which only a minority of patients were treated with beta-blockers. Because of the nonuniform effect of beta-blocker treatment on survival, on the one hand, and on potential risk predictors, on the other hand, the prognostic value of established risk factors era may be questioned in the beta-blocker.
The purpose of the present study was to evaluate the impact of beta-blocker treatment on the utility of established variables for risk stratification in ambulatory patients with CHF. Specifically, the study sought to examine whether the prognostic impact of peakVo2, LVEF, NE and NT-proBNP is altered by concomitant beta-blocker treatment.
The study included 408 consecutively enrolled patients with chronic CHF (NYHA functional classes I to IV, <70 years old) and LVEF <45%. All patients were referred to the Department of Cardiology at the Medical Clinic of the University of Heidelberg for assessment of their heart failure status and/or evaluation of their potential candidacy for heart transplantation between November 1994 and January 2000. The etiology of CHF (dilated or ischemic cardiomyopathic) was confirmed by cardiac catheterization before inclusion into the study. Treatment (number of patients, mean percentage ± SD the recommended daily dose) with an angiotensin-converting enzyme (ACE) inhibitor (n = 397, 53 ± 37%) or an angiotensin type 1 (AT1) receptor antagonist (n = 11, 48 ± 21%) was defined as a mandatory inclusion criterion. The subgroup of patients with additional beta-blocker treatment (n = 165) received the following beta-blockers (number of patients, mean percentage ± SD the recommended daily dose): metoprolol (n = 75, 200 mg/day, 35 ± 17%), carvedilol (n = 65, 25 mg twice daily, 33 ± 22%), sotalol (n = 18, 160 mg twice daily, 57 ± 20%), bisoprolol (n = 5, 5 mg twice daily, 40 ± 14%) and celiprolol (n = 2, 200 mg/day, 75 ± 35%). All patients had to be in a stable condition for at least four weeks, with the medication individually optimized by the referring physician, before assessment of the variables outlined subsequently. The study conformed with the principles outlined in the Declaration of Helsinki (13)and was approved by the institutional Ethics Committee. All patients gave written, informed consent.
Equilibrium rest radionuclide ventriculography was performed with a multicrystal gamma camera (Orbiter, Siemens, Erlangen, Germany) in the left anterior oblique view. Twenty minutes after pretreatment of red blood cells with 2 ml of stannous pyrophosphate, 30 mCi of technetium-99m (DuPont, Bad Homburg, Germany) was rapidly injected, followed by data acquisition and analysis (RNV, version 2.1, Elscint Medical Technology, Vienna, Austria). The LVEF was derived from time-activity curves as: (end-diastolic − end-systolic counts) / end-diastolic counts (14–16).
A symptom-limited cardiopulmonary exercise test was performed to determine peakVo2. The cardiopulmonary exercise test equipment included a metabolic cart (Oxycon-alpha, Jaeger, Würzburg, Germany) with a bicycle ergometer (Ergoline, Jaeger). Details of the test protocol have been published previously (14–16).
After insertion of an intravenous cannula and 30 min in the supine position, 2 ml of venous blood was drawn (EDTA-containing tubes, Sarstedt, Nümbrecht, Germany) and immediately centrifuged, and the plasma was stored at −30°C. The plasma concentrations of NT-proBNP and NE were analyzed by enzyme-linked immunosorbent assay (12)and radioenzymatic assay (14), respectively.
The combined study end point was defined as progressive heart failure requiring hospital admission, with intravenous inotropic, diuretic or mechanical support and/or cardiac death within one year. Patients who received a heart transplant during the one-year follow-up period were considered as survivors until the date of transplantation (15). Information on hospital periods and the circumstances of death was obtained from the hospital’s medical records or the referring physician.
The data are presented as the mean value ± SD, except where otherwise specified. Repeated measures analysis of variance on ranks was performed. The design included the intergroup factor (i.e., with or without beta-blocker treatment) and repeated measurements of the intragroup factors (i.e., NYHA functional class, rest heart rate, mean arterial pressure, peakVo2, LVEF, NE and NT-proBNP). Analysis of available cases was done. The highly significant interaction term of “group–measure” (F = 7.17, df = 7.1995; p < 0.0001; Greenhouse-Geisser–corrected p value) strongly indicated that differences between the groups were restricted to some of the measures studied. To work out the locus of the interaction (i.e., at which measure the groups differed), a series of nonparametric two-sample Wilcoxon tests was performed. To test for significant differences between mean values, the two-sample Wilcoxon test was also used. To compare frequencies, chi-square analysis was performed. A p value <0.05 was considered statistically significant. Univariate and multivariate Cox regression analyses were performed. Differences in event-free survival were detected by the Kaplan-Meier product limit method and compared by the Petro-Prentice generalized log-rank test. Calculations were performed with SAS version 6.12.
The clinical characteristics of the enrolled patients are summarized in Table 1. Fifteen percent of the patients were in NYHA functional class I, 41% in class II, 42% in class III and 2% in class IV. The study group was divided into two cohorts: one with (n = 165, 40%) and one without (n = 243, 60%) beta-blocker treatment, in addition to an ACE inhibitor or an AT1receptor antagonist. Age, gender, NYHA functional class, mean blood pressure, peakVo2and NT-proBNP were similar in both groups. However, patients with beta-blocker therapy were characterized by a reduced rest heart rate, increased LVEF and lower plasma concentrations of NE (Table 1).
Independent of additional beta-blocker treatment, patients reaching the study end point were characterized by a lower peakVo2and lower LVEF, as well as a higher NYHA functional class and higher NT-proBNP plasma levels, whereas NE plasma concentrations tended to be higher (Table 2). However, beta-blocker–treated patients had a significantly lower cardiac event rate, as compared with patients without beta-blocker treatment (16% vs. 34%, p < 0.0001) (Fig. 1), although patients reaching the combined study end point in both groups showed no significant differences with respect to NYHA functional class, peakVo2, LVEF and NT-proBNP plasma levels (Table 2). Only 10 (6%) of 165 beta-blocker–treated patients, as compared with 40 (17%) of 243 patients without additional beta-blocker treatment, died of cardiac causes within one year (p < 0.001). Similarly, a significantly lower proportion of beta-blocker–treated patients were hospitalized due to worsening heart failure (12% vs. 24%, p < 0.005). Within the beta-blocker group, the mean daily equivalent dosages of the diverse beta-blockers (35 ± 17% vs. 38 ± 21%) and ACE inhibitors (52 ± 32% vs. 59 ± 34%) was similar between those patients with and those without a cardiac event. Patients treated with a beta-blocker for more than three months and those treated less than three months were not different in terms of their baseline values (data not shown) or end point rate (15% vs. 16%). The beneficial effect of beta-blocker therapy was seen throughout the total patient group and was most obvious in patients with severe CHF (Fig. 2). Compared with patients without beta-blocker treatment, the cardiac event rates in beta-blocker–treated patients were reduced not only in NYHA functional classes I and II (10% vs. 20%, p < 0.05), but also in classes III and IV (18% vs. 30%, p < 0.05). The beneficial beta-blocker effect was independent of the CHF etiology (data not shown).
Univariate and multivariate predictors of event-free survival
In the total group, univariate Cox regression analysis revealed peakVo2, LVEF, NE and NT-proBNP as strong risk predictors (Table 3). Because beta-blocker treatment also predicted outcome, on univariate analysis,univariate Cox regression analyses were separately performed for patients with and those without additional beta-blocker treatment. Hence, only NE failed to persist as a risk predictor in both groups. PeakVo2, LVEF and NT-proBNP continued to be independent predictors of a cardiac event after multivariable Cox regression analysis in the total study group. However, only peakVo2and NT-proBNP remained predictive in patients without beta-blocker treatment. In contrast, in the beta-blocker group, only peakVo2and LVEF remained as independent risk predictors. Furthermore, Kaplan-Meier analyses revealed that for all strata, beta-blocker–treated patients showed a significantly better outcome than did patients without beta-blocker treatment. Most importantly, the beta-blocker–treated patients with an unfavorable risk profile (e.g., peakVo2≤14 ml/min per kg) showed a cardiac event rate comparable to that of patients without beta-blocker treatment, as well as a relatively low risk profile (e.g., peakVo2>14 ml/min per kg) (Fig. 3).
This study indicates that the predictive value of variables used for risk stratification in CHF is markedly influenced by the use of beta-blocker treatment. In particular, those patients with advanced CHF and an unfavorable risk profile benefited the most from beta-blocker treatment.
Total study group
Our study group, in which most patients were in NYHA functional class II or III and had a mean LVEF 22 ± 10%, was comparable to those study groups in several controlled trials, such as the Studies Of Left Ventricular Dysfunction, Vasodilator-Heart Failure Trial II and MEtoprolol CR/XL Randomized Intervention Trial in congestive Heart Failure (3,17,18). This is also reflected by a similar one-year mortality rate of 12% and a hospitalization rate of 19% for worsening CHF in the total study group. Furthermore, the classic predictors of CHF prognosis—LVEF (11)and peakVo2(19–22)—allowed us to independently predict the cardiac event rate in the total patient group. In agreement with previous studies on the predictive value of natriuretic peptides (23), multivariate analysis revealed NT-proBNP as another independent risk predictor.
Comparison of two subgroups with respect to additional beta-blocker treatment
Although the present study does not fulfill the criteria of a randomized, controlled trial, an equivalent risk may be assumed for both subgroups, as they did not differ in mean NYHA functional class, age, gender or peakVo2. The modestly lower heart rate and the slightly higher LVEF in the beta-blocker group are in close agreement with the responses previously observed under treatment with beta-blockers at dosages (i.e., between 25% and 50% of the maximal recommended study doses) equivalent to the mean dosages used in the present study (24). As patients with CHF and ischemic heart disease have a worse prognosis than those with dilated cardiomyopathy, the higher proportion of patients with ischemic cardiomyopathy in the beta-blocker–treated group may have slightly increased the risk in this subgroup. Consistent with previous reports on patients with CHF, the mean plasma concentrations of NE (25)were about 25% lower in beta-blocker–treated patients than in those without beta-blocker treatment. Mean NT-proBNP was 22% lower in the beta-blocker group, but failed to reach statistical significance, because of marked variations within both groups.
Impact of beta-blocker treatment on cardiac event rate
The present study was based on the assumption that the beneficial effects of beta-blocker treatment on mortality and morbidity are still demonstrable under routine clinical conditions. Although patients treated with beta-blockers received only ∼40% of the maximal recommended study dosages, beta-blocker–treated patients had a 65% lower one-year mortality rate than did patients without beta-blocker therapy (6% vs. 17%). This low mortality rate closely corresponds to that achieved in patients included into three major beta-blocker trials (7). Likewise, the annual rate of hospital admission due to worsening CHF was 50% lower in patients with than in those without beta-blocker treatment (12% vs. 24%). Because of the low event rate with beta-blocker therapy, a combined end point—cardiac death or hospital admission for worsening CHF—was used for further analysis. The combined end point was observed in 16% of the patients receiving a beta-blocker and in 34% of the patients without beta-blocker treatment. The beneficial effect of beta-blocker therapy on the hospitalization rate and cardiac mortality was seen throughout the total patient group (Fig. 3).
Prognostic variables and beta-blocker treatment
PeakVo2is currently used as a “reference standard” for selection of heart transplant candidates (6,16), with peakVo2<14 ml/min per kg usually regarded as a critical value. However, despite a similar mean peakVo2in both groups, the end-point rate was significantly lower in beta-blocker–treated patients. For any given peakVo2value, beta-blockade was associated with a better prognosis. Even in patients with the worst prognosis (i.e., peakVo2<10 ml/min per kg), one-year mortality was only 17% in beta-blocker–treated patients, which is much lower than the 38% annual mortality rate in corresponding patients without beta-blocker treatment.
Given the low mortality rates in high-risk patients treated with a beta-blocker, a new definition of the optimal cut-off value of peakVo2may be warranted in the beta-blocker era. For the other risk predictors also (i.e., LVEF, NE and NT-proBNP), the impact of beta-blocker treatment on prognosis was higher than expected from its effect on each variable, as for any given value, the end-point rate was markedly lower in beta-blocker–treated patients.
The present study is based on the prospective collection of risk predictors in consecutive patients in a single-center CHF registry. Thus, there might be a selection bias, as the center was a transplant referral site. The mean age of the included patients was relatively low, and the proportion of patients with ischemic heart disease was lower than the proportion of patients who had dilated cardiomyopathy. Because beta-blocker therapy has become mandatory in symptomatic patients with CHF (5,6), it is no longer ethically feasible to conduct a randomized trial. Furthermore, with the emphasis on the assessment of the impact of beta-blockade on the prognostic value of risk predictors, a strict randomization is not a prerequisite, as the major conclusions are based on the end-point rate observed for a certain range of a risk predictor (e.g., peakVo2<10 ml/kg per min). In addition, it was an important aim of the current study to draw conclusions from “real-life” practice, which is also reflected by beta-blocker doses below the maximal study doses. However, the present study does not allow us to rule out that a higher beta-blocker dose might have been associated with an even better prognosis. Finally, as the follow-up was limited to one year, no reliable statements on long-term prognosis are possible.
In patients admitted to the hospital with refractory heart failure and dependent on intravenous inotropic agents or short-term mechanical circulatory support, there is little doubt that heart transplantation improves their survival. However, heart transplantation is still associated with a 15% mortality rate in the first year (26). Thus, it does not necessarily improve the outcome in every patient with severe CHF (27). The present study indicates that the prognostic impact of variables used for risk stratification in CHF is markedly influenced by the use of beta-blocker treatment. Therefore, a re-evaluation of the current criteria for risk stratification in patients with CHF, including treatment of CHF, is warranted.
The assistance of Karin Hornig, RN, is gratefully acknowledged.
☆ This study was supported by a grant from the Faculty for Clinical Medicine of the University of Heidelberg, project 32/95. Measurement of N-terminal pro-brain natriuretic peptide was determined in collaboration with and with the support of Roche Diagnostics, Penzberg, Germany.
- angiotensin-converting enzyme
- angiotensin type 1 receptor
- congestive heart failure
- left ventricular ejection fraction
- N-terminal pro-brain natriuretic peptide
- New York Heart Association
- peak oxygen consumption
- Received August 13, 2001.
- Revision received February 20, 2002.
- Accepted February 25, 2002.
- American College of Cardiology Foundation
- The Task Force of the Working Group on Heart Failure of the European Society of Cardiology
- American College of Cardiology/American Heart Association Task Force
- Gilbert E.M,
- Anderson J.L,
- Deitchman D,
- et al.
- Luchner A,
- Burnett J.C,
- Jougasaki M,
- Hense H.W,
- Riegger G.A,
- Schunkert H
- Cohn J.N,
- Johnson G.R,
- Shabetai R,
- et al.
- Karl J,
- Borgya A,
- Gallusser A,
- et al.
- ↵World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 2000;284:3043–5
- Zugck C,
- Krüger C,
- Dürr S,
- et al.
- Meyer F.J,
- Borst M.M,
- Zugck C,
- et al.
- Hansen A,
- Haass M,
- Zugck C,
- et al.
- Mancini D.M,
- Eisen H,
- Kussmaul W,
- Mull R.Edmunds L.H Jr..,
- Wilson J.R
- Richards A.M,
- Nicholls M.G,
- Yandle T.G,
- et al.
- Bristow M.R,
- Gilbert E.M,
- Abraham W.T,
- et al.
- Gilbert E,
- Abraham W,
- Olsen S,
- et al.
- Deng M.C,
- De Meester J.M.J,
- Smits J.M.A,
- Heinecke J,
- Scheld H.H