American Heart Journal
Volume 141, Issue 6 , Pages 884-889, June 2001

Which β-blocker for heart failure?☆☆☆

Departments of Medicine and Radiology, University of North Carolina at Chapel Hill, Chapel Hill, NC

Article Outline

 

See related article on page 899 .

Establishing the efficacy of drugs for human disease remains a difficult enterprise characterized by major turns in conceptual thinking about specific pharmaceutical agents often after long periods of apparently appropriate use. β-Blockade for heart failure well illustrates this phenomenon because these agents once considered being contraindicated are now highly recommended therapy for heart failure.1, 2 Administration of these drugs results in substantial reductions in mortality by decreasing the risk of both progressive heart failure and sudden death, and important morbidity benefits have been unequivocally established for 3 agents in this class.3, 4, 5 Once physicians have accepted the efficacy of at least one agent in a class, interest typically shifts to whether a “class effect” is present that has obvious importance to clinicians in the field. Prescribing practices develop in the context of practical issues other than efficacy, such as cost, familiarity, and convenience, which often dominate the choice of drug therapy. In the specific case of β-blockade for heart failure, a vigorous debate concerning the relative efficacy of “second-generation” β1-selective agents (metoprolol controlled release or long acting or metoprolol) versus the “third-generation” nonselective β-blocker carvedilol.6 The work of Packer et al elsewhere in the Journal examines this issue by considering the relative effect of these agents on left ventricular function as assessed by change in left ventricular ejection fraction after initiation of therapy. The authors are to be commended for their effort to improve our understanding of potential differences between these agents that are being increasingly prescribed for heart failure. They conclude, on the basis of their meta-analysis, that carvedilol administration results in a significantly greater improvement in ejection fraction compared with β1-selective agents. Several considerations help assess the likelihood that these findings are correct and evaluate their potential clinical implications: (1) how strong is the conclusion that carvedilol produces greater improvement in left ventricular ejection fraction, (2) if present, what mechanism(s) could account for a greater effect of carvedilol, and (3) should these results influence the choice of β-blocker for patients with heart failure?

Back to Article Outline

Study findings 

Despite the best methodologic arguments of statisticians, clinicians will continue to seek answers to important therapeutic questions that influence their daily practice through reconsideration of available clinical trial data.7 Such analyses, often post-hoc in nature and involving subgroups of trial patients, are conducted regardless of the original intent of a specific clinical trial.8 This approach has been extended to combine trials to create a so-called meta-analysis of pooled clinical data.9 Although there are methodologic standards to guide meta-analyses, clinical observers differ on the value of this approach.10 The current work uses classical methods of meta-analysis and conforms to accepted statistical standards for this approach. The study’s primary strength is the classical one for meta-analysis where pooling trials provides a very substantial amount of data concerning the study measure, left ventricular ejection fraction, which can help discern differences despite noise. The analysis includes and is consistent with the results of the well-designed prospective, randomized study directly comparing carvedilol and metoprolol,11 which had sufficient sample size to test for differences in the response of left ventricular ejection fraction during therapy.

Despite the strength of the study analysis, 2 issues concerning its validity can be entertained. The first involves concerns that could arise during the review of any meta-analysis. The pooling of data from individual clinical trials necessary for such an analysis often comes into question, particularly as in this case, when the results are accumulated across a relatively long period of time, in a syndrome in which patient populations may differ in etiology and background therapy, and when the primary study measure is obtained by a variety of different methods. In this regard, data from the RESOLVD trial stand out among the placebo trials of metoprolol.12 In this study the angiotensin receptor blocker candesartan was combined with metoprolol controlled release or long acting in a substantial number of study patients. Data from the Evaluation of Losartan in the Elderly (ELITE) II study and recent as-yet-unpublished data from the Valsartan-Heart Failure Trial (Val-HeFT) study suggest the potential of an unfavorable interaction between angiotensin receptor blockers and β-blockers in patients with heart failure resulting from systolic dysfunction.13, 14 Although it is unknown whether this interaction includes an influence on ejection fraction benefit with β-blockade, no studies with carvedilol and angiotensin receptor blockers are included in the meta-analysis. Inclusion of the study of DiLendra et al15 could also be questioned because the study population for this randomized comparison was obtained by selecting nonresponders to immediate release metoprolol. The second issue concerns the adjustment that is made for the effect of ischemic etiology on ejection fraction response. The data presented do support a relationship between ischemic etiology and ejection fraction response to β-blockade that would bias the analysis against carvedilol. However, etiology may be only one of a number of factors that could influence the ejection fraction response, which might differ between the carvedilol- and metoprolol-treated patients. Development of a more general model of ejection fraction response within the study population could have been attempted. This modeling might have shown that etiology was the only readily apparent predictor or identified other characteristics in need of adjustment. The presence of differing etiologic patterns in the trials of carvedilol versus metoprolol is a warning concerning the possibility of other unknown biases that could have influenced the outcome of the meta-analysis.

Back to Article Outline

Origin of differential ejection fraction improvement 

Despite elegant studies in transgenic animals defining the role of specific receptors, the fundamental molecular mechanisms responsible for the beneficial effect of β-blockade in heart failure remain elusive. The presumption remains that their salutary effects are predominantly due to restraint of adrenergic stimulation. In the case of improvement in ejection fraction, this likely results from a complex interplay between augmentation of factors that enhance function and lessening processes detrimental to function. Several specific mechanisms and their relative inhibition by carvedilol and metoprolol can be considered.

Relative bradycardia may be a critical mechanism for β-blocker benefit.16, 17 Slowing the heart rate seems particularly important in the dilated, failing human heart where energetic balance is disrupted and an abnormal inverse relationship is present between heart rate and contractile force.18 Heart rate is determined predominantly by β1-receptors. Although carvedilol may produce greater reductions in heart rate during maximal exercise, the decline in resting heart rate is similar between selective and nonselective β-blockers.19 Apoptosis has been proposed as an important part of the pathophysiologic mechanisms of progressive heart failure.20, 21 Recent studies by Communnal et al22 have shown that norepinephrine promotes programmed cell death. Their in vitro studies indicate that this effect is mediated through the β1-receptor and that activation of the β2-receptor may be protective not harmful in this regard.23 Reducing the rate of apoptosis would be expected to retard the progressive loss of functioning myocardium and could contribute to a net improvement in ejection fraction.

Perhaps the most interesting potential mechanism for β-blocker augmentation of ejection fraction concerns the ability of these agents to restore activity of the adrenergic receptor–adenyl cyclase system and downstream signaling pathways. The β1-selective agents are well known to up-regulate adrenergic receptor function, but presumably this would contribute little because these receptors would be inhibited by drug action.19 In contrast, carvedilol does not up-regulate β-receptor function but may enhance signaling downstream from surface receptors.24 Carvedilol increases production of the β-adrenergic receptor kinase (BARK) inhibitor protein, which results in increased signaling in pathways downstream from the β-receptors.25 Likewise, studies with milrinone suggest that metoprolol may also increase the activity of signaling distal to the β-receptors.26 Although counterintuitive based on current models, the most parsimonious method to improve contractile function is to restore the activity of existing pathways responsible for contractility.27

Finally, the greater effect of carvedilol on ejection fraction could result, not from improvement in intrinsic cardiac function, but from a salutary effect on loading conditions. Loading conditions are well known to be an important determinant of left ventricular emptying, and reduction in afterload during pharmacologic therapy has often been shown to enhance ejection fraction. Given the ability of carvedilol to produce α-blockade, the possibility exists that any additional improvement in ejection fraction observed with this agent may be related to ventricular unloading rather than an intrinsic improvement in contractile state.

Back to Article Outline

Implications of differences in ejection fraction response 

Can change in ejection fraction be accepted as a clinical surrogate sufficient to dictate therapeutic choices for β-blockade in heart failure? Initial consideration suggests ejection fraction improvement as a plausible surrogate in congestive heart failure. However, although reversal of cardiac dysfunction, a primary defect in heart failure, should translate into better outcome, studies in human heart failure suggest a more complex relationship. The problem is that pharmacologic therapy does not necessarily reverse cardiac dysfunction by restoring normal physiologic mechanisms. Positive inotropic agents that are known to cause improvement in ventricular function but do so by augmenting cytosolic calcium are associated with either neutral or adverse effects on survival and morbidity in congestive heart failure.28, 29 Pure unloading drugs enhance function but activate hormones and provide less benefit than do neurohormonal antagonists.30 However, a direct correlation between change in ejection fraction and favorable outcome could be proposed if (1) the improved cardiac function arises from a favorable biologic effect on the myocardium specifically and (2) the favorable effect on ejection fraction is not countered by other effects such as activation of neurohormones or calcium loading that would negate the benefit derived from improvement in ejection fraction.

Thus the origin of the superior effect of carvedilol on ejection fraction is likely to be central to the clinical implications of the current study. If carvedilol produced a favorable biologic influence on the myocardium in excess of metoprolol—less remodeling and improved function without cost—important clinical outcomes should differ as well. Definitive data are not available on this point. The trial of Metra et al11 does provide rare comparative data on the change in volumes during therapy with metoprolol and carvedilol. Both agents caused a decline in left ventricular end-diastolic volume to a similar degree. Thus, in this head-to-head randomized comparison, no difference in myocardial remodeling, at least as indicated by changes in end-diastolic volume, was evident between the agents.

Finally, there is justifiable concern that focusing on ejection fraction response as a guide to β-blocker choice could also lead clinicians to determine whether to use or to continue any form of β-blocker therapy by these criteria. It seems likely that the clinical benefits of β-blockers arise from many mechanisms (ie, reduction in sudden death) that may be independent of any measurable effect on ventricular function. Current evidence suggests that these agents should be continued if they are tolerated clinically regardless of the ejection fraction response during therapy.

Back to Article Outline

Pan receptor blockade 

If there is an important role for all 3 adrenergic receptors in mediating the effectiveness of β-blockade, then carvedilol could be superior to selective blockers in the treatment of heart failure.31 However, the story of receptor effects being developed in the laboratory suggests a complex picture that calls into question a simple model based on complete blockade of all adrenergic receptors. An important pathophysiologic role for the β1-adrenergic receptor has been established with use of transgenic animal models, which have demonstrated that even mild overexpression of β1-receptors in an otherwise healthy animal results in myocardial hypertrophy and subsequent failure.32, 33 In contrast, many-fold increases in β2-receptor expression in wild-type mice are required for adverse myocardial effects.34 Even more intriguing experiments suggest that overexpression of the β2-receptor may ameliorate cardiomyopathy in some, if not all, animal models of heart failure.35, 36, 37 The experimental effects of α1 overexpression also appear much less impressive.38 Although significant myocardial hypertrophy has been noted, no decrement in life span or heart failure has been described. Studies of naturally occurring polymorphisms are beginning to relate these findings to clinical congestive heart failure.39 Liggett et al39 found that patients referred to an advanced heart failure and transplant program who carried a polymorphism associated with reduced β2-receptor function were at substantially higher risk of death during follow-up. Caution is needed in interpreting these findings from a single center in a relatively small number of patients. However, they support results from transgenic animal models concerning the relative toxicity of β1- versus β2-receptors.

Even assuming that all adrenergic receptors are equally toxic, the net in vivo effect and thus the expected clinical response to different β-blockers will be significantly influenced by receptor distribution, number, and affinity to agonist. β2-Receptors are present on presynaptic sympathetic nerves where their activation could act to increased neuronal outflow.40 This could be hypothesized as a theoretic benefit of carvedilol independent of its effect on β2-receptor toxicity. Preliminary studies suggest, however, that in vivo carvedilol does not reduce sympathetic flow, possibly because of its blockade of α2-receptors.41

As pointed out by Bristow,6 in the heart, the key target organ, β1-receptors predominate over β2- and α1-receptors even in the human failing myocardium. In addition, norepinephrine, the agonist responsible for mediating the effects of excessive sympathetic activity at the cellular level, is β1-selective as well. These considerations would argue that the β1-receptor would be the primary pharmacologic target and that only marginal benefit would be gained from antagonizing other receptors. The dose used could also influence the potential for differences between carvedilol and metoprolol, given that carvedilol has some degree of selectivity for the β1-receptor at low doses.42, 43

Back to Article Outline

Thought experiment 

Thought experiments have undoubtedly occurred for ages and, although popularized by Einstein, may potentially be used by lesser lights to useful ends, much in the way Newton’s calculus may be.44 A simple thought experiment may shed light on the potential likelihood of important differences in effectiveness of β-blockers for heart failure. At some risk of oversimplification, mortality and morbidity trials can be viewed as an example of elementary probability theory involving dichotomous outcomes.45 Only two outcomes are possible in such trials (much like a coin toss), and patients can be considered survivors or event free (for example, heads) or deceased or hospitalized (therefore tails). In this analogy, drugs may be conceptualized as altering the likelihood of a heads or tails during the coin toss. As with experiments involving the tossing of a coin, trials consisting of a small number of tests may give misleading results, whereas trials involving a large number of observations will give an estimate of the true probability, by a theory first successfully formulated by the 16th century mathematician Jakob Bernouli. Taken as a whole, the combined positive trials, with a very large total number of independent tests, suggest convergence toward a true risk reduction for 35% for effective β-blockers in heart failure. Comparison of large-scale trials with metoprolol controlled release and long acting and carvedilol (Metoprolol Controlled-Release Randomised Intervention Trial [MERIT] and Carvedilol Prospective Randomized Cumulative Survival [COPERNICUS] trial) suggests a similar risk reduction of approximately 35% for these agents. Compared with results seen with other cardiovascular drugs, this represents a very substantial benefit that is rarely observed and is unlikely to be exceeded by any particular β-blocker.

Back to Article Outline

Decision making in heart failure management 

A major focus on mechanistic differences between drugs is appropriate during drug development. However, this limited perspective may be detrimental when new agents are applied in general practice. Lack of attention and understanding of the many factors that influence use of drugs in practice may limit the application of life-saving therapies. Clearly cost, familiarity, and convenience play a significant role in the practical determination of prescribing patterns. These considerations become critical for the ultimate public health impact of any treatment advance. A particular drug may even have less intrinsic benefit but a more significant public health impact when multiple factors other than efficacy result in more widespread application in patients at risk.

Undoubtedly, spirited debate will continue on the relative merits of particular β-blockers for heart failure. Packer et al are appropriately cautious in interpreting the implications of their findings as they apply to drug selection. Perhaps the Carvedilol or Metoprolol Evaluation Trial (COMET) will provide more definitive data concerning the relative benefits of β1- versus nonselective β-blockade with α-blockade.46 At this juncture, the majority of effort should be focused on increasing the use of β-blockers for heart failure among cardiologists, internists, and family practitioners. Recent data from our multicenter heart failure database of predominantly academically based heart failure specialty clinics still suggest a substantial gap in the use of β-blockade even in this setting.47 Although the picture is changing rapidly in specialty clinics and elsewhere, patients will be well served by additional education and physician-to-physician interchange on this issue.

Back to Article Outline

References 

  1. Bristow MR. Beta adrenergic receptor blockade in chronic heart failure. Circulation. 2000;101:558–569
  2. HFSA guidelines for management of patients with heart failure caused by left ventricular systolic dysfunction—pharmacological approaches. J Cardiac Failure. 1999;5:357–382
  3. MERIT-HF Study Group . Effect of metoprolol CR/XL in chronic heart failure: metoprolol CR/XL randomized intervention trial in congestive heart failure (MERIT-HF). Lancet. 1999;353:2001–2007
  4. Packer M, Bristow MR, Cohn JN, et al.  The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med. 1996;334:1349–1355
  5. CIBIS Investigators and Committees . The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomized trial of beta-blockade in heart failure. Lancet. 1999;353:9–13
  6. Bristow MR. What type of β-blocker should be used to treat chronic heart failure?. Circulation. 2000;102:484–486
  7. Assmann SF, Pocock SJ, Enos LE, et al.  Subgroup analysis and other (mis)uses of baseline data in clinical trials. Lancet. 2000;355:1064–1069
  8. Adams KF. Post hoc subgroup analysis and the truth of a clinical trial. [editorial] Am Heart J. 1998;136:751–758
  9. Davey Smith G, Egger M. Meta-analysis: unresolved issues and future developments. BMJ. 1998;316:221–225
  10. LeLorier J, Gregoire G, Benhaddad A, et al.  Discrepancies between meta-analyses and subsequent large randomized, controlled trials. N Engl J Med. 1997;337:536–542
  11. Metra M, Giubbini R, Nodari S, et al.  Differential effects of beta-blocker in patients with heart failure: a prospective, randomized, double-blind comparison of the long-term effects of metoprolol versus carvedilol. Circulation. 2000;102:546–551
  12. RESOLVD Investigators . Effects of metroprolol CR in patients with ischemic and dilated cardiomyopathy: the Randomized Evaluation of Strategies for Left Ventricular Dysfunction pilot study. Circulation. 2000;101:378–384
  13. Pitt B, Poole-Wilson PA, Segal R, et al.  Effects of losartan compared with captopril on mortality in patients with symptomatic heart failure: randomized trial. The losartan heart failure survival study ELITE II. Lancet. 2000;355:1568–1579
  14. Cohn JN. Proceedings of the the 73rd annual meeting of the American Heart Association. In: New Orleans, La. New Orleans: The Association; 2000 Nov 15;p. 2000
  15. DiLendra A, Sabbadini G, Salvatore L, et al.  Long-term effects of carvedilol in idiopathic dilated cardiomyopathy with persistent left ventricular dysfunction despite chronic metoprolol. J Am Coll Cardiol. 1999;33:1926–1934
  16. Lechat P, Escolano S, Golmard JL, et al.  Prognostic value of bisoprolol-induced hemodynamic effects in heart failure during the Cardiac Insufficiency Bisoprolol study (CIBIS). Circulation. 1997;96:2197–2205
  17. Kjekshus JK. Importance of heart rate in determining beta-blocker efficacy in acute and long-term acute myocardial infarction intervention trials. Am J Cardiol. 1986;57:43F–49F
  18. Alpert NR, Mulieri LA. Human heart failure: determinants of ventricular dysfunction. Adv Exp Med Biol. 1997;430:97–108
  19. Gilbert EM, Abraham WT, Olsen S, et al.  Comparative hemodynamic, left ventricular functional, and antiadrenergic effects of chronic treatment with metoprolol versus carvedilol in the failing heart. Circulation. 1996;94:2817–2825
  20. Narula J, Haider N, Virmani R, et al.  Apoptosis in myocytes in end-stage heart failure. N Engl J Med. 1996;335:1182–1189
  21. Olivetti G, Abbi R, Quaini F, et al.  Apoptosis in the failing human heart. N Engl J Med. 1997;336:1131–1141
  22. Communnal C, Singh K, Pimentel DR, et al.  Norepinephrine stimulates apoptosis in adult rat ventricular myocytes by activation of the β-adrenergic pathway. Circulation. 1998;98:1329–1334
  23. Communal C, Singh K, Sawyer DB, et al.  Opposing effects of beta-1 and beta-2 adrenergic receptors on cardiac myocyte apoptosis: role of a pertussis toxin-sensitive G protein. Circulation. 1999;100:2210–2212
  24. Yoshikawa T, Port JD, Asanoo K, et al.  Cardiac adrenergic receptor effects of carvedilol. Eur Heart J. 1996;17:8–16
  25. Iaccarino G, Tomhave ED, Lefkowitz RJ, et al.  Reciprocal in vivo regulation of myocardial G protein–coupled receptor kinase expression by β-adrenergic receptor stimulation and blockade. Circulation. 1998;98:1783–1789
  26. Bohm M, Deutsch HJ, Hartmann D, et al.  Improvement of postreceptor events by metoprolol treatment in patients with chronic heart failure. J Am Coll Cardiol. 1997;30:992–996
  27. Lefkowitz RJ, Rockman HA, Koch WJ. Catecholamines, cardiac beta-adrenergic receptors, and heart failure. Circulation. 2000;101:1634–1637
  28. Digitalis Investigation Group . The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med. 1997;336:525–533
  29. Packer M, Carver JR, Rodeheffer RJ, et al.  Effects of oral milrinone on mortality in severe chronic heart failure: the PROMISE Study Research Group. N Engl J Med. 1991;325:1468–1475
  30. Cintron G, Johnson G, Francis G, et al.  Prognostic significance of serial changes in left ventricular ejection fraction in patients with congestive heart failure. Circulation. 1993;87(Suppl):VI17–VI23
  31. Packer M. Beta-adrenergic blockade in chronic heart failure: principles, progress, and practice. Prog Cardiovasc Dis. 1998;41:39–52
  32. Engelhardt S, Hein L, Wiesmann F, et al.  Progressive hypertrophy and heart failure in beta-1 adrenergic receptor transgenic mice. Proc Natl Acad Sci U S A. 1999;96:7059–7064
  33. Bisognano JD, Weinberger HD, Bohlmeyer TJ, et al.  Myocardial-directed overexpression of the human beta-1 adrenergic receptor in transgenic mice. J Mol Cell Cardiol. 2000;32:817–830
  34. Liggett SB, Tepe NM, Lorenz JN, et al.  Early and delayed consequences of beta-2 adrenergic receptor overexpression in mouse hearts: critical role for expression level. Circulation. 2000;101:1707–1714
  35. Dorn GW, Tepe NM, Lorenz JN, et al.  Low- and high-level transgenic expression of beta-2 adrenergic receptors differentially affect cardiac hypertrophy and function in G αq-overexpressing mice. Proc Natl Acad Sci U S A. 1999;96:6400–6405
  36. Rockman HA, Chien KR, Choi D, et al.  Expressions of a beta-adrenergic receptor kinase-1 inhibitor prevents the development of myocardial failure in gene-targeted mice. Proc Natl Acad Sci U S A. 1998;95:7000–7005
  37. Arber S, Hunter JJ, Ross , et al.  MLP-deficient mice exhibit a disruption of cardiac cytoarchitectural organization, dilated cardiomyopathy, and heart failure. Cell. 1997;88:393–403
  38. Milano CA, Dolber PC, Rockman HA, et al.  Myocardial expression of a constitutively active alpha1B-adrenergic receptor in transgenic mice induces cardiac hypertrophy. Proc Natl Acad Sci U S A. 1994;91:10109–10113
  39. Liggett SB, Wagoner LE, Craft LL, et al.  The Ile 164 beta-2 adrenergic receptor polymorphism adversely affects the outcome of congestive heart failure. J Clin Invest. 1998;102:1534–1539
  40. Vincent HH. Ma in’t Veld AJ, Boomsma F, et al. Compound ICI 118,551, a beta-2 adrenoceptor antagonist, lowers blood pressure. J Hypertens. 1985;S3:S247–S249
  41. Kurz T, Richardt D, Adler S, et al.  Presynaptic effects of carvedilol in the human heart. [abstract] Circulation. 2000;102:II629
  42. Lindenfeld J, Lowes BD, Bristow MR. Hypotension with dobutamine: β-adrenergic antagonist selectivity at low doses of carvedilol. Ann Pharmacother. 1999;33:1266–1269
  43. Bristow MR, Abraham WT, Gilbert EM, et al.  Relationships between carvedilol stereoisomer plasma concentration and beta receptor occupancies in subjects with chronic heart failure in the Multicenter Oral Carvedilol Heart Failure Assessment (MOCHA). [abstract] Eur Heart J. 1996;17:135A
  44. Guillen M. Curiosity killed the lights. Five equations that changed the world. New York: Hyperion; 1995;
  45. Dunham W. Bernoulli trials. The mathematical universe: an alphabetical journey through the great proofs, problems, and personalities. New York: John Wiley; 1994;
  46. Krum H. Beta-blockers in heart failure: the ‘new wave’ of clinical trials. Drugs. 1999;58:203–210
  47. Oren RM, Ventura HO, Pina I, et al.  Prevalence of beta blocker use in heart failure clinics: report from the UNITE-HF multicenter heart failure database. [abstract] J Cardiac Failure. 2000;6:2A–79A

☆ Reprint requests: Kirkwood F. Adams, Jr, MD, Division of Cardiology, University of North Carolina at Chapel Hill, CB #7075, Burnett-Womack Building, Chapel Hill, NC 27599-7075. E-mail: kfa@med.unc.edu

☆☆ Am Heart J 2001;141:884-8.

PII: S0002-8703(01)68115-8

doi:10.1067/mhj.2001.115583

Refers to article:

  • Comparative effects of carvedilol and metoprolol on left ventricular ejection fraction in heart failure: Results of a meta-analysis

    Milton Packer, George V. Antonopoulos, Jesse A. Berlin, Jesse Chittams, Marvin A. Konstam, James E. Udelson
    American Heart Journal June 2001 (Vol. 141, Issue 6, Pages 899-907)

American Heart Journal
Volume 141, Issue 6 , Pages 884-889, June 2001

Article Tools