Beta blockers first in heart failure?

September 04, 2005

Stockholm, Sweden - Initiating chronic heart-failure (CHF) treatment with the beta blocker bisoprolol is as safe and effective as starting treatment with guideline-recommended ACE-inhibitor therapy, according to the results of the Cardiac Insufficiency Bisoprolol Study (CIBIS III). Among patients 65 years and older with mild to moderately symptomatic CHF and impaired left ventricular ejection fraction (LVEF), initiation of treatment with bisoprolol followed by enalapril, as compared with the opposite sequence of drugs, showed similar rates of combined death and all-cause hospitalization, the study's primary end point.

Presenting the findings here during the first hotline session of the European Society of Cardiology (ESC) Congress 2005, lead investigator Dr Ronnie Willenheimer (Lund University, Malmö, Sweden) said the results of the study support the free choice of initial treatment of chronic heart failure with the ACE inhibitor enalapril or the beta blocker bisoprolol.

"The use of an ACE inhibitor before a beta blocker is based on the fact that the ACE-inhibitor trials were performed first, with the beta-blocker trials done with the drugs used on top of the ACE inhibitor. It's stayed that way, and no one has ever really looked at the importance of the order," Willenheimer told journalists during a press conference announcing the results. "Two small surrogate-end-point studies have found that the use of the beta blocker first seems better, but this is the first study to look at mortality and hospitalization as an end point. I always like to turn things upside down and say if the prevailing tradition had been to start with a beta blocker, adding ACE inhibitors, and then we did this study to challenge that, we wouldn't have been able to find anything in these results that would have challenged that order."

Dr Kenneth Dickstein (University of Bergen, Stavanger, Norway), the discussant for CIBIS III during the hotline session, told the assembled audience that the study addressed a clinically important question but had limitations, adding that he "did not quite share the same enthusiasm" about the findings. Dickstein pointed out that the study failed to prove noninferiority when the original study protocol was used to assess the primary end point of noninferiority for all-cause mortality or hospitalization. Despite concerns, Dickstein said he believed the two approaches show comparable safety and "probably comparable efficacy" and that the findings would likely have an impact on clinical practice.

"I think it will change practice because it will give us data now that will support the initial use of beta blockers," said Dickstein. "It suggests that we do now have a choice." Taken together with the positive findings supporting beta-blocker therapy first in the Carvedilol ACE Inhibitors Remodelling Mild Heart Failure Evaluation (CARMEN) trial, the studies provide a larger database supporting the two approaches as comparable, thereby allowing clinicians to select the best approach for their patients, said Dickstein.

To coincide with the ESC presentation, the results of the study are published online September 4, 2005 in Circulation[1].

What treatment approach works best?

Current guidelines recommend the combination of ACE-inhibitor and beta-blocker therapy as standard treatment for CHF, advising that beta-blocker therapy be given on top of standard ACE-inhibitor background therapy.

There are theoretical considerations suggesting that it might be beneficial to initiate treatment with beta blockers rather than with an ACE inhibitor, said Willenheimer. For example, the sympathetic nervous system is activated earlier than the renin-angiotensin-aldosterone system (RAAS), making beta blockers, an inhibitor of both the sympathetic nervous system and RAAS, a better choice to initiate therapy, he said. Also, sudden death is the most prevalent cause of death in the early course of CHF, he noted.

CIBIS III was a multicenter, prospective, randomized, open-label, blinded study in 1010 CHF patients designed to test the concept that initial monotherapy with bisoprolol followed by enalapril was comparable to the usual order of ACE inhibitor first, followed by a beta blocker. Randomized treatment with 1.25 mg bisoprolol once daily or 2.5 mg enalapril twice daily was initiated, with the drugs titrated upward every two weeks until the target maintenance dose of 10 mg was reached. After a six-month monotherapy phase, during which time the addition of the complementary ACE inhibitor or beta blocker was not allowed, the second drug was added and titrated upward in a manner similar to the first drug. The duration of the combination phase was between six months for the last recruited patients and 24 months (for the first recruited patient).

In the study protocol, bisoprolol-first treatment was considered noninferior to enalapril-first treatment with regard to the primary end point if the upper limit of the 95% confidence interval for the between-group difference was <5%, corresponding to a hazard ratio of 1.17. Based on this protocol, bisoprolol was not proven noninferior to enalapril-first treatment. In the intention-to-treat analysis, however, the data showed that starting with the beta blocker was not inferior to starting with ACE-inhibitor therapy.

Bisoprolol-first vs enalapril-first treatment

Analyses Hazard ratio 95% CI
Per-protocol noninferiority analysis 0.97 0.78-1.21
Intention-to-treat noninferiority analysis 0.94 0.77-1.16

"The result was quite similar in the per-protocol analysis and the intention-to-treat analysis, so there is one possible explanation as I see it," said Willenheimer. "We lost patients more rapidly in the per-protocol population. In heart-failure trials, over time, you're going to lose patients from the protocol analysis, as it is so strictly defined. If you look at the numbers, you can see that at all time points there are fewer patients in the protocol analysis than in the intention-to-treat analysis. So you lose statistical significance over time, the confidence interval gets larger and reaches above 1.17, which is what we defined as the upper limit."

In a post hoc analysis of the first year of treatment in which all patients were followed up, 155 bisoprolol-first patients compared with 165 enalapril-first patients had a primary end point, and there was a trend toward less mortality among patients treated with beta-blocker therapy first. However, the bisoprolol-first strategy was associated with a worsening of heart failure requiring hospitalization, although the finding was not statistically significant.

According to Willenheimer, the trend toward worsening heart failure is likely indicative of the lack of experience clinicians have in starting the drug prior to ACE inhibitor therapy. He said the worsening heart failure could have been caused by up-titrating too quickly.

"We have so little experience in titrating the beta blocker without the ACE inhibitor in hospitals, which is one of the reasons, I think, we saw a trend toward increased hospitalizations for heart failure with the beta-blocker-first therapy," said Willenheimer. "No one is experienced in titrating the drug in a heart-failure population without the ACE inhibitor."

Although higher doses of bisoprolol are often used in clinical practice, Willenheimer told heart wire that a starting dose larger than 1.25 mg might have caused a negative inotropic effect, possibly resulting in some patients refusing treatment. Regarding safety, both treatment combinations were safe and well tolerated. There was no statistically significant difference in serious adverse events reported during the monotherapy phase or combination-therapy phase of the trial.

Open-label design might have introduced bias

Discussing the trial, Dickstein explained some of his concerns about the study, specifically taking issue with the design of the study. He questioned whether the CIBIS III study more adequately addressed a comparison between enalapril and bisoprolol as monotherapy, because the heart-failure patients in this study were on only one drug for six months before the second agent was added.

"Twenty-six weeks on monotherapy is not too impressive and certainly not in the heart-failure guidelines," said Dickstein.

Another issue with the trial was the open-label design, Dickstein said. All-cause hospitalization is a highly subjective end point when clinicians are aware that the patient is undergoing a prolonged period of monotherapy and could have introduced bias into the results. He added that the noninferiority end point based on the intention-to-treat analysis might also introduce bias, as fewer patients taking the study medication makes it more likely that noninferiority will be shown.

Willenheimer defended the open-label approach, writing in Circulation that "it would have been virtually impossible to separately adjust the two doses of the two study drugs during the combined study phase, eg, in response to side effects, had [we] masked both treatments." This design is relevant for beta-blocker studies in CHF, when the titration schedule depends on response to treatment, write the investigators.

Despite his concerns, Dickstein still agreed with the conclusions reached by the CIBIS III investigators. According to Willenheimer and Dickstein, the findings leave clinicians with a choice of starting either with a beta blocker or an ACE inhibitor with their CHF patients.

"For a large majority of patients, many of them could be started with beta-blocker therapy first," said Willenheimer. "I think if you asked a thousand of my colleagues, a lot of them will say that if they have a patient with an abnormal heart rhythm like tachycardia, they will likely go with a beta blocker. If they have a patient with primary idiopathic cardiomyopathy, they'll likely start, especially if it is a younger patient, with an ACE-inhibitor strategy."

Merck KgaA (Darmstadt, Germany) sponsored CIBIS III. Willenheimer has received honoraria from Merck for lectures.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.