Beta-Blockers for Advanced Heart Failure — How Far can you go?

Two recent large-scale clinical trials have demonstrated conclusive evidence of the beneficial effects of beta-blocker therapy on survival in chronic heart failure 1,2. CIBIS-II enrolled 2647 patients with NYHA class III or IV symptoms and left ventricular (LV) ejection fraction ≤35%. All-cause mortality was reduced by 32% with bisoprolol, from 17.3% in the placebo group to 11.8% in the bisoprolol group over a mean follow-up of 1.3 years (risk ratio 0.68, 95% confidence interval: 0.56–0.82) 1. The MERIT-HF trial enrolled 3991 patients with NYHA class II-IV symptoms and LV ejection fraction <40%. All-cause mortality was reduced by 34% with metoprolol CR/XL, from 11% in the placebo group to 7.2% in the metoprolol group over one year 2. This benefit is in addition to that gained with ACE inhibitor therapy. As a result of these data beta-blocker therapy has become part of standard therapy for patients with chronic heart failure (in addition to ACE inhibitors and diuretics). The treatment is, in general, well tolerated. However, there are several issues that should be considered when using beta-blocker therapy in clinical practice. One important aspect is at what stage of the disease process should patients with heart failure be started on beta-blockers, in particular what is the role of this treatment in patients with advanced heart failure. This article reviews the clinical trials of beta-blockers in heart failure with particular regard to the patient selection process. Consideration of this will allow the safe use of these agents in clinical practice. The severity of heart failure can be assessed in a number of ways, including the patients clinical condition (often with the NYHA functional class, peak VO2 or with the 6-min walk distance), left ventricular function (usually LV ejection fraction), and the likelihood of morbidity and mortality. To guide clinical use of beta-blocker therapy it is appropriate to look at the inclusion criteria for selection of patients for the clinical trials. The main inclusion criteria for the larger trials are shown in Table 1. The patients enrolled were, in general, receiving standard therapy for heart failure, which in most cases included ACE inhibitors and diuretics. The patients had to be clinically stable at entry to the trials. In CIBIS II 1, to be eligible for the trial patients had to be ambulant, clinically stable for 6 weeks and have had no changes to their cardiovascular medications for 2 weeks prior randomisation. In the MERIT-HF trial 2, patients were excluded from entry to the trial if they had unstable, decompensated heart failure and their clinical condition had to remain stable during a 2-week placebo run-in phase 2. Thus beta-blocker therapy should not be initiated in patients who are clinically unstable with overt pulmonary congestion or who are requiring frequent alterations in their medical treatment. In general, there is little clinical experience with beta-blocker therapy in patients with low blood pressure. Both CIBIS II 1 and MERIT-HF 2 excluded patients with systolic blood pressure <100 mmHg. Mean systolic blood pressure at entry to these two trials was approximately 130 mmHg. The US carvedilol trials 3 allowed entry of patients with systolic blood pressure as low as 85 mmHg. While mean systolic blood pressure at randomisation (115 mmHg) was lower than in CIBIS II 1 or MERIT-HF 2. Thus most clinical trial experience, at present, is for patients without low blood pressure. If beta-blockers are to be used in patients with lower blood pressure then extreme care is required with frequent monitoring. Assessment of NYHA functional class is a clinical assessment of the patients reported symptoms and has considerable intra- and inter-observer variability. Although the inclusion criteria for the trials have allowed NYHA class IV patients (Table 1), these patients represent a relatively small proportion of the total patients at the time of randomisation (Table 2). In reality, patients with NYHA functional class IV symptoms are often hospitalised, not ambulant, and require frequent alterations in medical therapy. These factors, as discussed above, were exclusions for the beta-blocker trials and would in reality have excluded many patients with class IV symptoms. In addition, patients with class IV symptoms have high mortality rates, despite ACE inhibitor therapy. The annual mortality rates in the placebo treated groups of CIBIS II 1 and MERIT-HF 2 were 13.2% and 11% , respectively, consistent with that observed in the ACE inhibitor treated group of the SOLVD treatment trial 4 (12.4%). By way of comparison, the annual mortality rate in the ACE inhibitor treated group of the CONSENSUS I trial 5 (NYHA class IV patients) was 36%. Thus, it would appear that the majority of the population of patients in the major beta-blocker trials represented those with mild to moderate heart failure, a similar group to that involved in the SOLVD treatment trial. Many of the patients in the beta-blocker trials had previously had Class IV symptoms. For example, in the ANZ Carvedilol Trial 6, which is often described as involving patients with ‘milder’ heart failure, 43% of the patients had previously had Class IV symptoms although none did at entry to the study. Thus, NYHA class IV symptoms do not preclude therapy with beta-blockers for ever. Patients should be re-evaluated when further medical therapy has improved their clinical status. A further marker of severity of disease is the LV ejection fraction. The beta-blocker trials only included patients with impaired LV systolic function, and thus there is no experience with these agents in patients with heart failure and preserved LV systolic function (or ‘diastolic dysfunction’). The highest LV ejection fraction allowed at entry to the clinical trials was 45% 6, and none of the trials stipulated a lower cut-off of LV ejection fraction. The mean LV ejection fraction at entry to both CIBIS II 1 and MERIT-HF 2 was approximately 28%, while that in the other 24 trials was approximately 24% 7 (Table 2). Retrospective analyses from CIBIS-I 8 have suggested that prognostic improvement was associated with improvement in LV ejection fraction. However, this relationship could not be demonstrated when the analysis was restricted to patients who actually received the beta-blocker therapy. The clinical trial data at present supports the use of beta-blockade in patients with heart failure due to systolic impairment. A very low LV ejection fraction should not exclude the consideration of beta-blocker therapy as these patients also have the potential to gain from such treatment. In summary, selection of patients for beta-blocker therapy should be based on the clinical characteristics of the patients involved in the clinical trials. Most of the patients appear to have had mild to moderate heart failure, for the reasons outlined above, although some patients were labeled as having class IV symptoms at entry to these trials. Patients should be on appropriate doses of ACE inhibitors and diuretics and importantly should be clinically stable for several weeks prior to initiation of the beta-blocker. Clinical stability appears to be the key to the safe use of beta-blockers in patients with heart failure. The lack of clinical trial data in patients with NYHA class IV symptoms signifies that these patients should not be offered beta-blocker therapy. However, patients should be re-evaluated for suitability for beta-blockade if appropriate therapy has improved their clinical status. Some patients with a systolic blood pressure <100 mmHg may tolerate beta-blocker therapy, although great care should be taken in these patients. There appears to be no lower limit of LV ejection fraction where patients do not have the potential for clinical improvement with beta-blocker therapy. Monitoring of response in LV ejection fraction to beta-blocker therapy is not required since lack of improvement in LVEF does not imply lack of potential long-term clinical benefit (according to current data). In CIBIS II 1, permanent withdrawals from study treatment occurred in 15% of both the bisoprolol and placebo treatment groups. Similar withdrawal rates were seen in the MERIT-HF trial 2, 13.9% and 15.3% in the metoprolol and placebo treated patients, respectively. Interestingly, in the US carvedilol trials 3 the commonest reason for discontinuation of double-blind treatment was worsening heart failure, which occurred more frequently in the placebo group. In these trials, only 5.7% of the carvedilol treated patients discontinued study treatment because of adverse reactions, compared with 7.8% of the placebo treated patients. These data compare favourably with the withdrawal rates from ACE inhibitor therapy in the CONSENSUS I trial 5, and are better than that in SOLVD 4, although this study involved considerably longer follow up. These data suggest that if beta-blockers are used in patients with similar characteristics to those enrolled in the clinical trials and patients are carefully monitored during therapy as occurred in the clinical trials, then the treatment can be well tolerated with acceptable withdrawal rates due to adverse effects. The benefits to be gained from beta-blocker therapy are considerable. There is always the potential for increase in risk associated with the translation of therapies from the clinical trial setting to clinical practice. This partly relates to patient selection for clinical trials and the less frequent monitoring of patients in clinical practice than in the controlled trials. Careful selection of patients will allow the safe use of beta-blockers in patients with heart failure outside of the clinical trials. Further clinical trials will report over the next few years and help to define the relative risks and benefits of beta-blockers in specific subgroups of patients, particularly those with severe symptoms (NYHA class IV).