.gif){kind=spacer}
Treatment of chronic heart failure (CHF), an important cause of global morbidity and mortality, has evolved in the last three decades.[1-3] Activation of neurohormonal systems plays a key role in the pathophysiology and progression of the disease. Therapeutic strategies directed towards their inhibition have reduced morbidity and mortality.[3] The major mechanism seems to be related to the inhibition or reversal of remodelling.[4] Angiotensin-converting enzyme (ACE) inhibitors are known to reduce symptoms and improve prognosis.[5] The benefit of beta blockers in patients with heart failure has been amply demonstrated by comparing outcomes in patients prescribed a beta blocker or a placebo in patients on optimal treatment with diuretics and ACE inhibitors.[6]
Guidelines advise the initial use of ACE inhibitors and subsequently beta blockers.[7] Adherence to guidelines brings benefit.[8] One reason for this strategy is that the use of diuretics and ACE inhibitors may optimise haemodynamics and improve ventricular function[9] prior to the introduction of a beta blocker, which may for a short period have an adverse effect on haemodynamics.[10] The converse argument would be that adrenergic activity is an earlier phenomenon in heart failure than stimulation of the renin-angiotensin system.[3] Consequently, starting beta blockers before ACE inhibitors might be better.
Study Aim
The results of the Cardiac Insufficiency Bisoprolol Study (CIBIS) were recently published[11] and commented on in this journal.[12] This trial was a bold attempt to answer an important clinical question relating to the treatment of heart failure, namely whether patients with heart failure could be treated equally well with either a beta blocker or an ACE inhibitor as initial treatment after the use of diuretics. The first objective of the study was to show non-inferiority of starting beta blockers before ACE inhibitors.
Some might question whether this was an important question. The answer is potentially yes, because the benefits of beta blockers are dose-dependent and titration to maximal doses may frequently be limited by low pre-titration blood pressures caused by ACE-inhibitor treatment or for other reasons. Despite evidence of efficacy of the use of beta blockers in CHF, many physicians are still resistant to the use of this class of drug.[8]
Outcome
In CIBIS III the outcomes were essentially similar regardless of whether a beta blocker or an ACE inhibitor was the initial treatment. A superficial conclusion might be that both drugs were equally effective. The authors reported a trend toward a higher frequency of the end point of 'worsening of CHF requiring hospitalisation or occurring in hospital' in patients starting on beta blockers. The authors had no explanation for this and thought it unlikely to be due to rapid up-titration of bisoprolol as they used a slower up-titration protocol than in CIBIS II.[13] One explanation may be that beta blockade in patients with persistent neurohormonal activation may have further worsened left ventricular systolic function and sodium and water retention. If this explanation is correct, then therein is the vindication for the conventional practice of starting CHF therapy with ACE inhibitors and diuretics.
Limitations
The CIBIS III trial is not without considerable limitations. Looked at in the harshest manner, it was a negative study in that it failed to fulfil its primary end point of demonstrating the non-inferiority of bisoprolol compared to enalapril when analysed per protocol, which is the correct form of analysis for this type of study. But the study came close to demonstrating non-inferiority (HR 0.97; 95% CI 0.78 to 1.21). The upper limit of 1.17 was the criterion for non-inferiority and a more reasonable judgement might be that no difference was shown, especially since on an intention-to-treat analysis the study did demonstrate non-inferiority.
One key inclusion criterion was euvolemia but it is frequently impossible to be sure of a patient's volume status by clinical examination alone prior to starting drug treatment. The authors begin their article with the challenging remark that initial use of an ACE inhibitor followed by a beta blocker is the mainstay of the management of heart failure. The European Society of Cardiology guidelines[7] also convey and appear to support this view. That seems curious. Most physicians believe the initial objective in patients with heart failure is to control fluid volumes with diuretics and that takes precedence over the early use of other drugs. Several trials have shown that congestive heart failure cannot be treated satisfactorily with an ACE inhibitor alone.[14,15]
Study Design
A more important limitation was the design of the study. Patients were randomised to either bisoprolol or enalapril for a period of six months and then placed on both drugs. The results are presented over a period of 1.22 years, the mean follow-up time. Thus, part of the result might be the consequence of the allocated drug in the initial six-month treatment period. The situation is similar to that pertaining, for example, to a surgical procedure where the benefit is seen long after the intervention. But to answer the real question, which is whether a beta blocker and an ACE inhibitor are similar, the results should only be analysed during the six-month period of randomised treatment. That is far too short a time to study a treatment strategy in heart failure. Many early studies have shown benefits of drugs at such a time period in terms of symptoms and haemodynamics and later demonstrated an increase of mortality.
Thus, the trial could not have answered the important clinical question. A key issue is why the study was designed in that way. One possible reason is that ethical committees would not allow the trial to proceed with one drug against another for more than six months. If the reality were that beta blockers are equivalent to or even better than ACE inhibitors when given alone, or that an ACE inhibitor has little if any benefit if a patient has been on a beta blocker, knowledge of that fact will be concealed forever if proper trials are not permitted. The decision of the ethical committee would then be preventing advances in medicine which might affect large numbers of patients over many years and exposing many patients to the adverse events from use of an unnecessary drug. It is possible that the decisions of ethics committees may be influenced by the ease of adopting the path of least resistance or the path least likely to give rise to criticism, and to accept readily the current position with regard to the optimal treatment of heart failure. But in doing so, ethical committees block the possibility of gaining knowledge about treatment which could be of advantage to many patients. This conundrum is common in heart failure but not in other disease entities. For example, in testing a new antibiotic it would not be added to the existing antibiotics but be put against current antibiotics.
The authors argue that in early heart failure, treatment should be aimed at the prevention of sudden cardiac death. That is a perfectly reasonable concept since sudden death is more common in mild heart failure than in severe heart failure, where deaths are usually associated with deteriorating ventricular function. On the other hand, determining whether a sudden death is due to an opportunistic arrhythmia or whether it is associated with a myocardial event or other rarer causes is immensely difficult. For some, the key question is whether deaths are reduced regardless of the mechanism. Thus, the logic that a beta blocker is preferable because such drugs may preferentially reduce sudden death is not entirely convincing.
One great merit of this study is that patients were older than in many earlier studies, with a mean age of 72 years. Of the trial population, 30% were female, which is a greater percentage than in previous studies.
The authors defend the prospective randomised open label blinded end point evaluation (PROBE) design on the grounds that it would have been impossible to undertake the dose escalation in a double-blinded fashion. They are almost certainly correct and even if not, a clinician can easily measure the heart rate and form an opinion as to which drug the patient is taking. Undoubtedly the open design limits the certainty with which one can accept the result although it is impossible to be certain which way any bias might have affected the result.
Conclusion
CIBIS III, for the reasons stated, has not moved the clinical care of patients with heart failure forward. We still do not know, and probably never will, whether ACE inhibitors bring about a further advantage to patients with heart failure who are already being treated with diuretics and beta blockers. Physicians in the area of heart failure have become prisoners of history. The traditional approach based on the temporal order in which clinical trials have been undertaken is that treatment should be initiated with a diuretic and an ACE inhibitor, or possibly an angiotensin receptor blocker; a beta blocker is introduced later. The results of CIBIS III do provide some comfort to physicians who for one reason or another wish to introduce the drugs in the reverse order over a short period of time.
References
|
Correspondence to: Professor PA Poole-Wilson (email: p.poole-wilson@imperial.ac.uk )
.gif){kind=spacer}
Fernando A Botoni, Academic Visitor from Minas Gerais, Brazil
Cardiac Medicine, National Heart and Lung Institute, Imperial College London, Dovehouse Street, London, SW3 6LY