Initiating treatment for chronic heart failure

Current guidelines for therapy for chronic heart failure (CHF) recommend beginning with an angiotensin-converting enzyme (ACE) inhibitor, which should be titrated up to the target dose before adding a beta-blocker.^1,2 This recommendation is not evidence based, but is a consequence of the fact that ACE inhibitors were first shown to improve survival and reduce morbidity in CHF.³ In light of this, it has become considered unethical to withhold ACE inhibitors from patients with CHF, and the beneficial effects of beta blockers have been assessed as an addition to an ACE inhibitor regimen.^4-6 Subsequently, therapeutic tradition has evolved based on this sequence of initiating CHF therapy. However, before the Cardiac Insufficiency Bisoprolol Study (CIBIS) III, the best sequence for initiating medications in terms of mortality and hospitalization had not been examined.^7,8

Initial monotherapy with bisoprolol fumarate (Zebeta; starting dose, 1.25 mg once daily; target dose, 10 mg once daily) was compared with initial therapy with enalapril maleate (Vasotec; starting dose, 2.5 mg twice daily; target dose, 10 mg twice daily) for 6 months, followed by a combination of the 2 drugs for 6 to 24 months in the CIBIS III study, which was an open-label trial. Patients were 65 years of age or older and had stable, mildly symptomatic, or moderately symptomatic CHF corresponding to New York Heart Association (NYHA) functional class II or III and a left ventricular ejection fraction (LVEF) of ≤35%. All patients were randomly assigned to 1 of the 2 treatment groups and were essentially naive to treatment with ACE inhibitors, beta blockers, and angiotensin receptor blockers. The 2 strategies were blindly compared with regard to the combined primary end point of all-cause mortality or hospitalization, and each of these components individually. Treatment with bisoprolol first was regarded as noninferior to treatment with enalapril first if the upper limit of the 95% confidence interval (CI) for the absolute between-group difference was below 5%, corresponding to a hazard ratio (HR) of 1.17.

Patients and methods

Analyses were done by intention to treat. Assessment of noninferiority for the primary end point of mortality or hospitalization was also performed per protocol. All randomized patients were analyzed by intention to treat. Per-protocol analysis eliminated pa­tients who had never received the study drugs, did not meet inclusion criteria, met any of the exclusion criteria, did not take the second medication during the combined treatment stage, illegitimately discontinued randomized treatment too soon, illegitimately took the second drug before 6 months after randomization, or were not compliant. Data were collected in the per-protocol analysis up until the protocol was breached, as determined by the blinded end point committee.

Results

The mean follow-up period for the 1010 included patients was 1.22 years. The 2 groups—bisoprolol-first and enalapril-first—were no different regarding baseline features. Female patients made up 31.8% of the sample, and the mean age was 72.4 years. Ischemic heart disease was the most common cause of CHF, and hypertension was considered a cause in 36.5% of patients. The mean LVEF was 28.8%. Half of the patients were in NYHA functional class II, and half were in class III. The proportion of patients with previous diseases was as follows: peripheral vascular disease, nearly 10%; cerebrovascular disease, 10%; diabetes, nearly 20%; and acute myocardial infarction, about 50%.

Baseline cardiovascular medication was comparable between the 2 groups. About 10% of patients were taking potassium-sparing diuretics, almost 15% were receiving an aldosterone antagonist, about 15% were taking antidiabetic medication, about 33% were receiving cardiac glycosides, about 66% were receiving anti­platelet medication, and 84.3% were taking diuretics, predominantly loop diuretics. Throughout the course of the study, medications remained nearly the same.

P

P

P

P

P

The 2 strategies were similarly effective in terms of the combined primary end point of mortality or all-cause hospitalization (Figure 1). The primary end point occurred in 35.2% of patients (n = 178) in the bisoprolol-first group compared with 36.8% of patients (n = 186) in the enalapril-first group, according to intention-to-treat analysis (absolute difference = -1.6%; 95% CI, -7.6% to 4.4%; HR = 0.94; 95% CI, 0.77 to 1.16; noninferiority, = .02). These results show that in the intention-to-treat analysis, bisoprolol-first treatment was noninferior to enalapril-first treatment. In the per-protocol analysis, the primary end point occurred in 32.4% of patients (n = 163) in the bisoprolol-first group compared with 33.1% of patients (n = 165) in the enalapril-first group (absolute difference = -0.7%; 95% CI, -6.6% to 5.1%; HR = 0.97; 95% CI, 0.78 to 1.21; noninferiority, = .46). These results show that in the per-protocol analysis, bisoprolol-first treatment was not quite proven to be noninferior to enalapril-first treatment. In the enalapril-first group, 73 patients died compared with 65 patients in the bisoprolol-first group (HR = 0.88; 95% CI, 0.63 to 1.22; between-group difference, = .44). In the enalapril-first group, 157 patients were hospitalized compared with 151 in the bisoprolol-first group (HR = 0.95; 95% CI, 0.76 to 1.19; between-group difference, = .66); 51 patients in the enalapril first group had a worsening of CHF occurring in the hospital or necessitating hospital admission compared with 63 patients in the bisoprolol-first group (HR = 1.25; 95% CI, 0.87 to 1.81; between-group difference, =. 23).

P

P

The primary end point occurred in 108 patients in the enalapril-first group compared with 109 patients in the bisoprolol-first group following the monotherapy stage of the study (HR = 1.02; 95% CI, 0.78 to 1.33; between-group difference, = .90). In the enalapril-first group, 32 patients died compared with 23 patients in the bisoprolol-first group (HR = 0.72; 95% CI, 0.42 to 1.24; between-group difference, = .24).

P

P

Among the enalapril-first group, 92 patients experienced a hospital admission compared with 99 patients in the bisoprolol-first group during the monotherapy stage (HR = 1.08; 95% CI, 0.81 to 1.43; between-group difference, = .59). The bisoprolol-first approach showed a 31% lower mortality rate at the end of the first year (Figure 2), the longest time during which all patients were followed-up (42 deaths in the bisoprolol-first group vs 65 deaths in the enalapril-first group; HR = 0.69; 95% CI, 0.46 to 1.02; between-group difference, = .06).

Fifty-nine patients in the enalapril-first group and 60 patients in the bisoprolol-first group were formally eliminated from the study. Sixteen patients in the enalapril-first group and 19 patients in the bisoprolol-first group could not be contacted concerning their vital status at study end. Of these patients, 2 in the enalapril-first group and 3 in the bisoprolol-first group were lost to follow-up.

P

P

During the monotherapy stage of the study, the dosage was increased to at least 5 mg twice daily in the en­alapril-first group in 94% of patients (n = 476) and to the target dose of 10 mg twice daily in 84% of patients (n = 425). In the bisoprolol-first group, the dosage was increased to at least 5 mg once daily in 82% of patients (n = 415) and to the target dose of 10 mg once daily in 65% of patients (n = 326). At the end of the study, the first drug prescribed was prescribed at a dose of at least 50% of the target dose to significantly more patients in both treatment groups (Figure 3). The second drug was introduced prior to the end of the 6-month monotherapy stage because of poor control of CHF among 7.3% of patients (n = 37) in the enalapril-first group and 7.7% of patients (n = 39) in the bisoprolol-first group ( = .81). Over the course of the study, 17.6% of patients (n = 89) in the enalapril-first group discontinued randomized treatment compared with 20% of patients (n = 101) in the bisoprolol-first group (= .33). The rates of adverse events were similar between the 2 groups. In the enalapril-first group, 37.3% of patients (n = 187 patients) had at least 1 serious adverse event compared with 36.5% of patients (n = 184) in the bisoprolol-first group. There was no difference between groups regarding the effects on blood pressure in both the monotherapy and combined stages of treatment. Although heart rate was lower in the bisoprolol-first group during the monotherapy phase, there were no differences between the 2 groups in heart rate at the conclusion of the combined stage.

P

P

P

With regard to the primary end point, there was uniformity across subgoups in the predetermined subgroup analysis, except for an association with LVEF (= .001). In patients with an LVEF < 28% (HR = 0.61; 95% CI, 0.44 to 0.85; = .001), administering bisoprolol first had markedly better results than administering enalapril first; however, an opposite trend was seen among patients with an LVEF of 28% to 35% (HR = 1.23; 95% CI, 0.94 to 1.61; = .13). This association was almost entirely explained by a difference in hospitalization for noncardiovascular reasons during the monotherapy stage.

Discussion

Although conventional therapy for patients with CHF and left ventricular systolic dysfunction involves beta blockers and ACE inhibitors, treatment must be started first with 1 or the other of the 2 drugs. There is no evidence supporting a strategy of up-titrating an ACE inhibitor and a beta blocker simultaneously. By tradition, guidelines recommend starting with an ACE inhibitor.^1,2 However, the optimum sequence has previously only been tested in 2 surrogate end point trials,^9,10 indicating that starting with a beta blocker is at least as good as starting with an ACE inhibitor. CIBIS III was the first study to assess which strategy is the best or whether both strategies are similarly effective in terms of mortality, hospitalization, and safety. In patients at least 65 years of age with stable CHF, NYHA functional class II or III, and an LVEF of 35%, the 2 strategies showed similar results regarding combined mortality and all-cause hospitalization. Treatment with bis&shy;oprolol first was proven noninferior to treatment with enalapril first by intention-to-treat analysis. Noninferiority was not proven by per-protocol analysis. However, the results were similar, with Kaplan-Meier curves that were able to be superimposed on those of the intention-to-treat analysis, indicating that the lack of significance of noninferiority was caused by less statistical power in the per-protocol analysis.

With bisoprolol-first therapy, there was a trend toward more worsening of CHF occurring in the hospital or causing hospitalization, especially early during the trial. Currently, we are analyzing whether this worsening was mostly mild and temporary, as would be expected during up-titration of a beta blocker. There was a clear trend toward a survival benefit for bisoprolol-first therapy, with a 28% hazard reduction at the end of the monotherapy stage and a 31% hazard reduction at 1 year. This survival benefit placed more patients at risk of worsening of CHF in the bisoprolol-first group, which might account for the entire difference in worsening of CHF events. In terms of safety, the 2 strategies were similar. With regard to age, CIBIS III was representative of a true CHF population in clinical practice. The mean age of the patients was approximately 74 years at the end of the study.

Conclusion

The results of CIBIS III indicate that CHF therapy may be started with bisoprolol or enalapril in patients with CHF, NYHA functional class II or III, and depressed LVEF. The choice must be based on individual assessment of each patient. The interesting survival trend in favor of bisoprolol-first therapy might suggest that more patients could survive to later stages of CHF if therapy were started with bisoprolol rather than with enalapril. This should be examined in future trials.