Neurohumoral Prediction of Benefit From Carvedilol in Ischemic Left Ventricular Dysfunction
Background—Plasma neurohormones were analyzed for prediction of adverse outcomes and response to treatment in 415 patients with ischemic left ventricular dysfunction randomly assigned to receive carvedilol or placebo.
Methods and Results—Atrial natriuretic peptide, brain natriuretic peptide (BNP), or norepinephrine (NE) levels above the group median were associated with increased mortality rates and heart failure. On multivariate analysis, both BNP and NE interacted with treatment to predict death or heart failure independent of age, New York Heart Association class, and left ventricular ejection fraction. For placebo, supramedian levels of BNP were associated with 3-fold the mortality rate of inframedian levels (20/104; 19% vs 6/99; 6%; P<0.01). For carvedilol, mortality rate was comparable in these 2 subgroups (12/109; 11% vs 8/94; 9%; NS). Corresponding rates for heart failure were 29/104 (28%) versus 3/99 (3%; P<0.001) for placebo and 16/109 (15%) versus 7/94 (7%; NS) for carvedilol. High NE levels did not predict additional benefit from carvedilol, which significantly reduced heart failure admissions only in those with NE levels below the median (13.1% to 4.0%; P<0.01). In the 23% of the study population with supramedian BNP but inframedian levels of NE, carvedilol reduced hospital admission with heart failure by >90% (P<0.001).
Conclusions—Carvedilol reduced mortality rates and heart failure in those with higher pretreatment BNP levels but lesser activation of plasma NE. Neurohumoral profiling may guide introduction of β-blockade in heart failure.
Neurohumoral responses to cardiac injury participate in the pathogenesis of heart failure.1 2 3 4 Plasma levels of neurohumoral factors offer an indication of cardiovascular prognosis1 4 5 6 and may aid risk stratification and choice of therapy. We hypothesized that plasma levels of ≥1 of norepinephrine (NE), the cardiac natriuretic peptides atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), and arginine vasopressin (AVP) would identify patients likely to benefit from treatment with carvedilol. We report the results from 415 patients with ischemic left ventricular impairment randomly assigned to treatment with carvedilol or placebo for congestive heart failure.
The study organization and the effects of carvedilol on left ventricular function, exercise tolerance, symptoms, morbidity, and mortality rates have been reported.7 8 Twenty centers in Australia and New Zealand participated (see “Appendix”). The study was coordinated by the Clinical Trials Research Unit, University of Auckland. Neurohormonal assays and related analyses were conducted by the Christchurch Cardioendocrine Group, Christchurch School of Medicine.
Patients (n=415) were recruited with chronic stable heart failure caused by ischemic heart disease, left ventricular ejection fraction (LVEF) by radionuclide ventriculography <45%, and current New York Heart Association class II or III or previous NYHA class II-IV. Exclusion criteria included current NYHA class IV; heart rate <50 bpm; sick sinus syndrome; second- or third-degree heart block; blood pressure <90 mm Hg systolic or >160/100 mm Hg; treadmill exercise duration <2 or >18 minutes (modified Naughton protocol); coronary event or procedure within 4 weeks; primary myocardial or valve disease; insulin-dependent diabetes; chronic airways disease; hepatic disease (serum transaminase >3 times normal); renal impairment (creatinine >250 μmol/L); and life-threatening noncardiac disease or current treatment with β-blocker, β-agonist, or verapamil.
Patients tolerating 6.25 mg BID carvedilol were randomly assigned, double-blind, to receive continued carvedilol (titrated to a maximum dose of 25 mg BID) or matching placebo for a trial period of 18 months. Measurements included LVEF (radionuclide scan), exercise tolerance (modified Naughton treadmill test and 6-minute walk test), and NYHA class. End points included change in LVEF, exercise tolerance, symptoms, morbidity (hospitalization with heart failure, coronary ischemic syndromes, or noncardiovascular illnesses), and mortality rates.
Neurohumoral sampling was conducted before randomization. Venous blood (EDTA) was collected (9 am to noon) after intravenous cannulation and the patient having been seated for 30 minutes. The sample was separated within 20 minutes, and the plasma was stored at −80°C before transport on dry ice to the assay laboratory for assay within 6 weeks of sampling. Assays for ANP, BNP, NE, and AVP were conducted according to established methods.9 10 11 12 Interassay and intra-assay coefficients of variation were 5% to 12%.
Plasma hormones were compared (2-sample t tests) to age-matched normal subjects (n=168). Event rates were compared with the use of χ2 tests with risk ratios (and 95% CI) and Kaplan-Meier curves calculated for groups with admission levels above and below the median of individual neurohumoral factors, LVEF, exercise time, and distance. Analyses were conducted separately for groups receiving carvedilol and placebo as well as for the total patient population. Cox proportional hazards analysis was used to examine potential indicators (NYHA class, LVEF, history of heart failure or previous myocardial infarction, and BNP/NE by treatment interactions) for independent prediction of death and admission with heart failure.
P<0.05 (2-tailed) was taken to indicate statistical significance.
Seventy percent of patients were in NYHA class II-III at randomization (43% had previously been class IV at some time). Eighty-eight percent had had a previous myocardial infarction and 43% a previous hospital admission for heart failure. Eighty-five percent were receiving an ACE inhibitor, 75% diuretics, and 38% digoxin. The average LVEF at entry was 29%, treadmill duration 10.5 minutes, and mean 6-minute walk distance 392 meters.
The effects of treatment on morbidity and mortality rates and surrogate end points have been reported.7 8 Carvedilol improved LVEF without deterioration in symptoms or exercise capacity. The combined end point of death or hospital admission for any reason was reduced by 26% (104 carvedilol vs 131 placebo; 95% CI −43% to −5%; 2P=0.02). Falls in death (20 vs 26; 95% CI −58% to +36%; 2P>0.1) and all admissions (99 vs 120; 95% CI −41% to 0%; 2P=0.054) were not significant. Fewer admissions for heart failure occurred in the carvedilol group (23 vs 33; NS). Numbers of events in patients receiving carvedilol compared with placebo did not differ significantly for acute coronary syndromes (25 vs 34), other cardiovascular events (40 vs 38), or noncardiovascular admissions (69 vs 51).
Markers of Morbidity/Mortality
Plasma BNP, ANP, and NE but not AVP exceeded normal values (30±23 vs 5.5±2.5 pmol/L [mean±SD], P<0.001; 43±36 vs 12.5±7 pmol/L, P<0.001; 4400±2100 vs 1550±650 pmol/L, P<0.001; 3.3±4.1 vs 2.3±2.5 pmol/L, NS, respectively). Tables 1⇓ and 2⇓ give event rates (death and heart failure) for potential prognostic markers divided according to the median levels of candidate markers (Table 1⇓) and then further divided according to administration of placebo or carvedilol (Table 2⇓).
Kaplan-Meier curves for ANP and BNP levels above and below the group median (irrespective of whether carvedilol or placebo was given) indicated that patients with higher levels of ANP and BNP were at increased risk of death and heart failure (Figures 1⇓ and 2⇓). Norepinephrine showed less clear-cut separation of survival curves, significant (P<0.01) for heart failure but not death (P=0.0508). Survival curves for LVEF did not separate significantly for death or heart failure. Patients with treadmill times below the group median were at significantly increased risk of heart failure (P<0.001) but not death. Six-minute walk distance results paralleled those for treadmill time. Event curves for AVP levels above and below the group median were not significantly separated for any clinical end point. No marker predicted acute coronary ischemic syndromes by either risk ratio or Kaplan-Meier analyses.
Death and heart failure were more frequent in patients with cardiac peptide levels above the group median who received placebo. In such patients randomly assigned to carvedilol, event rates did not differ significantly from patients with inframedian peptide levels (Table 2⇑ and Figures 1⇑ and 2⇑). The ratio of mortality in those with ANP and/or BNP levels above the median to those with inframedian levels was ≈2:1 overall, 3:1 for placebo, and 1:1 for carvedilol (Table 2⇑ and Figure 1⇑). For admissions with heart failure, the corresponding ratios were ≈3:1, 8:1, and 2:1, respectively (Table 2⇑ and Figure 2⇑).
In contrast to the cardiac peptides, higher plasma levels of NE did not predict benefit from carvedilol. Instead, carvedilol reduced heart failure in those with inframedian levels of plasma NE (Table 2⇑), and Kaplan-Meier analysis indicated an advantage for such patients receiving carvedilol compared with placebo (P<0.05). Similar trends for the interaction of NE, treatment, and mortality rates were not significant (Table 2⇑).
Twenty-three percent (n=97) of the total group had supramedian plasma BNP with concurrent inframedian plasma NE levels. Cumulative heart failure rates were identical in this subgroup (13/97, 13%) as in the remainder of the group (39/303, 13%; Figure 3⇓). However, the response to carvedilol clearly differed in this defined neurohumoral subgroup. For placebo, heart failure rates were higher in this subgroup (12/48, 25%) compared with others (19/153, 12%, P<0.05), whereas carvedilol reversed this pattern (1/49, 2% vs 20/153, 13%, P<0.03). Hence, within this neurohumoral subgroup, carvedilol reduced heart failure admissions from 12/48 (25%) to 1/49 (2%; P<0.001; Figure 3⇓). Increased benefit from carvedilol was also seen for mortality rate (placebo 7/48, 15% vs carvedilol 2/46, 4%, P=0.098), though mortality rate was too low for these differences to achieve statistical significance. Therefore, nearly all the reduction in heart failure admissions (23 vs 33, ie, 10 fewer events) and deaths (20 vs 26) potentially attributable to carvedilol occurred within this neurohormonal subgroup.
Relations between indicator/treatment interactions and clinical events were statistically weak or nonsignificant by both risk ratio (Table 2⇑) or Kaplan-Meier (not shown) analyses for LVEF, treadmill time, 6-minute walk distance, and pretreatment blood pressures and heart rate.
Cox proportional hazards analyses indicated that interactions of treatment (ie, carvedilol or placebo) with BNP and NE predicted death and heart failure independent of age, NYHA class, and LVEF (Table 3⇓). Both were predictive of death independent of previous myocardial infarction and for heart failure independent of previous heart failure. ANP/treatment in place of BNP/treatment interaction yielded similar though slightly weaker results. Rotation of hypertension, diabetes, sex, blood pressure, and heart rate through the model produced similar results, with BNP/treatment and NE/treatment terms remaining independent predictors for heart failure (P<0.03-P<0.001) but showing only marginal statistical significance in some models for predicting death (P≤0.06 to 0.03). In an all-inclusive analysis with 13 variables, both BNP/treatment and NE/treatment interactions remained independently predictive of death (P=0.04 for both) and heart failure (P=0.003 for both).
Beyond confirming the known relations between circulating neurohormones and cardiovascular death in heart failure,1 4 5 6 the current study is the first to demonstrate neurohormonal prediction of response to carvedilol.
Recent trials of carvedilol in heart failure indicate that it is a significant advance in therapy for a common and lethal condition.7 8 13 14 15 However, appropriate patient selection and prediction of benefit have remained areas of uncertainty. The current report throws new light on these issues. Prerandomization plasma concentrations of cardiac natriuretic peptides above compared with below the median level for the group predicted a far greater risk of death (>3-fold) or heart failure (>7-fold) in those randomly assigned to placebo, but such patients receiving carvedilol exhibited marked reductions in those ratios to <2-fold (NS).
Plasma NE exhibited a different pattern in terms of predicting benefit from carvedilol. As expected, increased levels of plasma NE were associated with an increased risk of death and heart failure in the patient group overall. However, higher plasma NE levels did not predict additional treatment benefit. Carvedilol significantly reduced heart failure admissions only in patients with lower NE levels. This surprising finding concurs with that reported by Vantrimpont et al16 in their recent discussion of the additive beneficial effects of β-blockers above and beyond converting enzyme inhibition in the Survival and Ventricular Enlargement (SAVE) study. In this retrospective analysis in >500 patients, β-blockers “were not found to have a greater effect in patients with neurohumoral activation at the time of hospital discharge.” In fact, increased NE levels were associated with significantly reduced efficacy of β-blockers in prevention of cardiovascular death, whereas (in common with the current report), elevated ANP levels predicted benefit. These data are subject to several limitations that do not apply to the current study. The SAVE trial assessed the impact of ACE inhibition in patients selected for early postinfarction asymptomatic left ventricular dysfunction rather than the efficacy of β-blockers. Within the neurohormonal substudy group, those receiving β-blockers had a higher incidence of hypertension and higher average LVEF. The neurohormonal subgroup was younger, had less previous myocardial infarction, and had greater use of thrombolysis and revascularization than the overall SAVE population. That is, potential confounders may have distorted the apparent impact of β-blockade, and it is questionable whether findings within the subgroup could be extended to SAVE trial participants in general. The Australia–New Zealand Heart Failure Study was specifically designed to assess the effect of carvedilol in stable ischemic left ventricular dysfunction, the randomization achieved matching of groups for relevant characteristics, and all participants underwent neurohormonal sampling. The current data confirm the findings of the SAVE neurohormonal substudy and extend them from standard β-blockers to carvedilol.
Mechanisms underlying restriction of benefit to those with less activation of NE are unknown. Markedly increased plasma NE may indicate dependence on increased sympathetic traffic to sustain cardiac output. Alternatively, Vantrimpont et al16 suggested that the β-blockade was insufficient, only effectively blocking harmful cardiac sympathetic drive in those with milder sympathetic activation. However, plasma NE is a poor surrogate measurement for cardiac sympathetic activity or for sympathetic nerve traffic as measured by peripheral microneurography.17 Further, lower levels of NE may simply have been a marker of less advanced (or better compensated) clinical heart failure, indicating those patients more likely to respond to many possible additional treatments.
What underlies the predictive power of ANP and BNP for benefit from carvedilol is also unknown. Carvedilol, in common with other β-blockers, enhances plasma concentrations of ANP.18 This may reflect drug-induced downregulation of natriuretic peptide clearance receptors.19 It is possible that some of the benefit of carvedilol is derived from enhancing plasma ANP (and/or BNP) with promotion of natriuresis, vasodilation, and relative suppression of sympathetic and renin-angiotensin-aldosterone systems.
The current data are the first to interrelate pretreatment neurohumoral status and response to carvedilol. Existing reports relate neurohormones and response to converting enzyme inhibition. Swedberg et al20 reported that mortality benefit from enalapril in NYHA class IV heart failure was greater in patients with catecholamines, angiotensin II, aldosterone, and ANP above the group median. In the SAVE trial, elevated plasma renin activity predicted a moderate increase in efficacy of captopril in reducing 1-year but not total cardiovascular mortality rates.21 Enalapril offered a greater survival benefit in those with more markedly increased NE and renin levels in the Veterans Administration Heart Failure Trial II (V-HeFT II).4 The relation of neurohormonal status to treatment effect on rates of worsening heart failure was not addressed in the latter report. CONSENSUS (Cooperative North Scandinavian Enalapril Survival Study), SAVE, and V-HeFT II did not provide data on BNP.
Therefore, available evidence suggests elevated pretreatment renin, NE, and cardiac peptide levels are positively associated with benefit from ACE inhibitor therapy, whereas ANP and BNP are positive predictors and plasma NE (surprisingly) a negative predictor of benefit when a β-blocker is added. The specific experimental setting of the current study does not indicate whether plasma cardiac peptide levels would be useful in predicting benefit from carvedilol without prior introduction of ACE inhibitor treatment or from other antifailure therapy (such as endopeptidase inhibitors, angiotensin I receptor blockers, central sympathoinhibitory agents, or endothelin antagonists), and these questions should be addressed in future randomized trials of therapy in heart failure.
In summary, the current findings are consistent with previous reports indicating important associations of plasma concentrations of the cardiac peptides and plasma NE with left ventricular function and cardiac prognosis.1 2 4 Our data support a growing literature indicating that BNP in particular is a powerful prognostic indicator.6 22 23 This may reflect the direct relation between ventricular wall stress and secretion of BNP. BNP is the only truly “ventricular hormone” examined in the current neurohumoral panel. It is possible that greater stability of BNP than ANP on storage24 partially accounts for its better performance, although the degree of instability of ANP on freezing is disputed.25 BNP is also less responsive than ANP to short-term stimuli. The current report does not include data on N-terminal ANP,5 a stable product that may be superior to ANP as a prognostic marker. However, our experience23 suggests that it is still probably weaker than BNP.
The current results cannot be extrapolated to other treatments or patient groups. Examination of subgroups entails loss of statistical power and generates hypotheses rather than conclusive findings. Therefore, the current report requires confirmation. However, measurement of plasma levels of BNP and NE may usefully complement symptomatic status and/or left ventricular imaging as triggers for initiation of β-blocker therapy in heart failure.
Australia–New Zealand Heart Failure Research Collaborative Group: Participating Investigators
Australia: J. Bradbury, R. Burton, D. Colquhoun, M. Croot, D. Cross, T. Davidson, J. Garrett, C. Hall, A. Hamer, B. Hicks, J. Horowitz, B. Jackson, I. Jeffrey, V. Kinber, H. Krum, G. Lane, S. Leslie, D. Owensby, G. Rudge, J. Ryan, J. Shepherd, B. Singh, J. Stephensen, P. Taverner, P. Thompson, A. Thomson, A. Tonkin, A. Trotter, J. Turner, W. Walsh, S. Woodhouse, and Y. Zhang.
New Zealand: C. Bond, G. Brown, J. Bruning, A. Clayton, J. Crawford, R.N. Doughty, C. Hall, H. Ikram, G. Lewis, C. Low, H. McAlister, J. Murphy, L. Nairn, M. Richards, D. Scott, and N. Sharpe.
This study was funded by a grant from SmithKline Beecham. Dr MacMahon is the recipient of Senior Research Fellowship from the Health Research Council of New Zealand. Dr Richards holds the New Zealand National Heart Foundation Chair of Cardiovascular Studies. Supplementary support for the neurohormonal assays was provided by the Health Research Council. Assistance with graphics was provided by Leanne Liggett. Barbara Griffin provided secretarial assistance.
- Received June 24, 1998.
- Revision received October 14, 1998.
- Accepted October 26, 1998.
- Copyright © 1999 by American Heart Association
Francis GS, Benedict C, Johnstone DE, Kirlin PC, Nicklas J, Liang CS, Kubo SH, Rudin-Toretsky E, Yusuf S. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure: a substudy of the studies of left ventricular dysfunction (SOLVD). Circulation. 1990;82:1724–1729.
Rouleau JL, de Champlain J, Klein M, Bichet D, Moyé L, Packer M, Dagenais GR, Sussex B, Arnold JM, Sestier F, Parker JO, McEwan P, Bernstein V, Cuddy E, Lamas G, Gottlieb SS, McCans J, Nadeau C, Delage F, Hamm P, Pfeffer MA. Activation of neurohumoral systems in postinfarction left ventricular dysfunction. J Am Coll Cardiol. 1993;22:390–398.
Francis GS, Cohn JN, Johnson G, Rector TS, Goldman S, Simon A, for the V-HeFT VA Cooperative Studies Group. Plasma norepinephrine, plasma renin activity, and congestive heart failure: relations to survival and the effects of therapy in V-HeFT II. Circulation 1993;87(suppl VI):VI-40–VI-48.
Omland T, Aakvaag A, Bonarjee VVS, Caidahl K, Lie RT, Nilsen DW, Sundsfjord JA, Dickstein K. Plasma brain natriuretic peptide as an indicator of left ventricular systolic function and long-term survival after acute myocardial infarction. Circulation. 1996;93:1963–1969.
Australia–New Zealand Heart Failure Research Collaborative Group. Effects of carvedilol, a vasodilator–β-blocker, in patients with congestive heart failure due to ischemic heart disease. Circulation. 1995;92:212–218.
Yandle T, Espiner E, Nicholls M, Duff H. Radioimmunoassay and characterisation of atrial natriuretic peptide in human plasma. J Clin Endocrinol Metab. 1986;61:71–78.
Packer M, Colucci WS, Sackner-Bernstein JD, Liang C-S, Goldscher DA, Freeman I, Kukin ML, Kinhal V, Udelson JE, Klapholz M, Gottlieb SS, Pearle D, Cody RJ, Gregory JJ, Kantrowitz NE, LeJemtel TH, Young ST, Lukmas MA, Shusterman NH, for the PRECISE Study Group. Double-blind, placebo-controlled study of the effects of carvedilol in patients with moderate to severe heart failure: the PRECISE Trial. Circulation. 1996;94:2793–2799.
Bristow MR, Gilbert EM, Abraham WT, Adams KF, Fowler MB, Hershberger RE, Kubo SH, Narahara KA, Ingersoll H, Krueger S, Young S, Shusterman N, for the MOCHA Investigators. Carvedilol produces dose-related improvements in left ventricular function and survival in subjects with chronic heart failure. Circulation. 1996;94:2807–2816.
Vantrimpont P, Rouleau JL, Wun C-C, Ciampi A, Klein M, Sussex B, Arnold MO, Moye L, Pfeffer M, for the SAVE Investigators. Additive beneficial effects of beta-blockers to angiotensin-converting enzyme inhibitors in the Survival and Ventricular Enlargement (SAVE) Study. J Am Coll Cardiol. 1997;29:229–236.
Rundqvist B, Elam M, Bergmann-Sverrisdottir Y, Eisenhofer G, Friberg P. Increased cardiac adrenergic drive precedes generalized sympathetic activation in human heart failure. Circulation. 1997;95:169–175.
Nakaoka H, Kitahara Y, Amano M, Imataka K, Fujii J, Ishibashi M, Yamaji T. Effect of β-adrenergic receptor blockade on atrial natriuretic peptide in essential hypertension. Hypertension. 1987;10:221–225.
Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L. Hormones regulating cardiovascular function in patients with severe congestive heart failure and their relation to mortality. Circulation. 990;82:1730–1736.
Rouleau J-L, Packer M, Moyé L, de Champlain J, Bichet D, Klein M, Rouleau JR, Sussex B, Arnold JM, Sestier F, Parker JO, McEwan P, Bernstein V, Cuddy TE, Lamas G, Gottlieb SS, McCans J, Nadeau C, Delage F, Wun C-C, Pfeffer MA. Prognostic value of neurohumoral activation in patients with an acute myocardial infarction: effect of captopril. J Am Coll Cardiol. 1994;24:583–591.
Tsutamoto T, Wada A, Maeda K, Hisanaga T, Maeda Y, Fukai D, Ohnishi M, Sugimoto Y, Kinoshita M. Attenuation of compensation of endogenous cardiac natriuretic peptide system in chronic heart failure: prognostic role of plasma brain natriuretic peptide concentration in patients with chronic symptomatic left ventricular dysfunction. Circulation. 1997;96:509–516.
Richards AM, Nicholls MG, Yandle TG, Frampton C, Espiner EA, Turner JG, Buttimore RC, Lainchbury JG, Elliott JM, Ikram H, Crozier IG, Smyth DW. Plasma N-terminal pro–brain natriuretic peptide and adrenomedullin: new neurohormonal predictors of left ventricular function and prognosis after myocardial infarction. Circulation. 1998;97:1921–1929.
Nelesen RA, Dimsdale JE, Ziegler MG. Plasma atrial natriuretic peptide is unstable under most storage conditions. Circulation. 1992;86:463–466.
Buckley MG, Sagnella GA, MacGregor GA. Stability of atrial natriuretic peptide during storage at −20°C. J Hum Hypertens. 1996;10:431–432.In 415 patients with ischemic left ventricular dysfunction, carvedilol reduced mortality rates (19.2% to 11.0%; P<0.05) and heart failure (27.9% to 14.7%; P<0.01) in those with supramedian brain natriuretic peptide (BNP) levels. In patients with inframedian norepinephrine levels, carvedilol reduced heart failure from 13.1% to 4.0% (P<0.01). In the subgroup (23%) with supramedian BNP but inframedian norepinephrine, carvedilol reduced heart failure admissions by >90% (P<0.001). Pretreatment neurohumoral profiling with plasma BNP and norepinephrine may identify patients who will benefit from carvedilol.