Chronic heart failure (CHF) is associated with high. Heart Failure

Transcription

1 Heart Failure Effect on Survival and Hospitalization of Initiating Treatment for Chronic Heart Failure With Bisoprolol Followed by Enalapril, as Compared With the Opposite Sequence Results of the Randomized Cardiac Insufficiency Bisoprolol Study (CIBIS) III Ronnie Willenheimer, MD, PhD; Dirk J. van Veldhuisen, MD, PhD; Bernard Silke, MD, PhD; Erland Erdmann, MD, PhD; Ferenc Follath, MD, PhD; Henry Krum, MD, PhD; Piotr Ponikowski, MD, PhD; Allan Skene, PhD; Louis van de Ven, MD, PhD; Patricia Verkenne, MD; Philippe Lechat, MD, PhD; on behalf of the CIBIS III Investigators Background In patients with chronic heart failure (CHF), a -blocker is generally added to a regimen containing an angiotensin-converting-enzyme (ACE) inhibitor. It is unknown whether -blockade as initial therapy may be as useful. Methods and Results We randomized 1010 patients with mild to moderate CHF and left ventricular ejection fraction 35%, who were not receiving ACE inhibitor, -blocker, or angiotensin receptor blocker therapy, to open-label monotherapy with either bisoprolol (target dose 10 mg QD; n 505) or enalapril (target dose 10 mg BID; n 505) for 6 months, followed by their combination for 6 to 24 months. The 2 strategies were blindly compared with regard to the combined primary end point of all-cause mortality or hospitalization and with regard to each of these end point components individually. Bisoprolol-first treatment was noninferior to enalapril-first treatment if the upper limit of the 95% confidence interval (CI) for the absolute between-group difference was 5%, corresponding to a hazard ratio (HR) of In the intention-to-treat sample, the primary end point occurred in 178 patients allocated to bisoprolol-first treatment versus 186 allocated to enalapril-first treatment (absolute difference 1.6%, 95% CI 7.6 to 4.4%, HR 0.94; 95% CI 0.77 to 1.16). In the per-protocol sample, 163 patients allocated to bisoprolol-first treatment had a primary end point, versus 165 allocated to enalapril-first treatment (absolute difference 0.7%, 95% CI 6.6 to 5.1%, HR 0.97; 95% CI 0.78 to 1.21). With bisoprolol-first treatment, 65 patients died, versus 73 with enalapril-first treatment (HR 0.88; 95% CI 0.63 to 1.22), and 151 versus 157 patients were hospitalized (HR 0.95; 95% CI 0.76 to 1.19). Conclusion Although noninferiority of bisoprolol-first versus enalapril-first treatment was not proven in the per-protocol analysis, our results indicate that it may be as safe and efficacious to initiate treatment for CHF with bisoprolol as with enalapril. (Circulation. 2005;112: ) Key Words: heart failure drugs adrenergic beta-antagonists angiotensin-converting enzyme inhibitors Chronic heart failure (CHF) is associated with high morbidity and mortality rates as well as considerable healthcare costs. 1,2,3 Increased activity of neurohormonal systems plays an important role in the pathophysiology of CHF and has become an important target for drug treatment. 4 Editorial p 2380 Blocking the renin-angiotensin-aldosterone system (RAAS) by angiotensin-converting enzyme (ACE) inhibitors effectively delays the progression of CHF and improves clinical Received August 10, 2005; revision received August 29, 2005; accepted August 29, From Lund University, Department of Cardiology, University Hospital, Malmö, Sweden (R.W.); Thoraxcenter, Department of Cardiology, University Hospital Groningen, the Netherlands (D.J.v.V.); Department of Pharmacology and Therapeutics, Trinity Centre, St James Hospital, Dublin, Ireland (B.S.); Medizinische Klinik III, University of Cologne, Germany (E.E.); Medicine A, University Hospital Zürich, Switzerland (F.F.); Departments of Epidemiology and Preventive Medicine and Medicine, Monash University, Alfred Hospital, Melbourne, Australia (H.K.); Cardiology Department, Clinical Military Hospital, Wroclaw, Poland (P.P.); The Nottingham Clinical Research Limited, Nottingham, UK (A.S.); Merck KGaA, Darmstadt, Germany (L.v.d.V., P.V.); and Service de Pharmacologie, Hopital Pitié-Salpetriere, Paris, France (P.L.). The Appendix, which lists the CIBIS III Investigators, Steering Committee, Data Safety Monitoring Board, and End Point Committee, is available in the online-only Data Supplement at Correspondence to Dr Ronnie Willenheimer, Department of Cardiology, University Hospital, S Malmö, Sweden. ronnie.willenheimer@med.lu.se 2005 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 Willenheimer et al Initiation of Heart Failure Therapy 2427 Figure 1. Titration of study drugs. outcomes. 5 The use of a -blocker in addition to an ACE inhibitor further reduces morbidity and decreases mortality by about 35%. 6 8 Current guidelines for the treatment of CHF recommend that treatment be started with an ACE inhibitor and that a -blocker be given in addition 9,10 ; however, this sequence seems to result primarily from the fact that the beneficial effects of ACE inhibitors were documented first. 5 8 After that, it was widely considered unethical to withhold ACE inhibitors in CHF, and any new treatment for CHF, including -blockers, has been evaluated in addition to ACE inhibitors. Currently, the combination of both classes of drugs is recommended standard treatment for CHF. The sequence of initiating these agents may be important, however. Most patients with CHF do not receive both agents in adequate doses, and in many instances patients will take only one drug for extended periods of time before they receive the second agent. 11,12 The first initiated treatment, ie, an ACE inhibitor, is more likely to be uptitrated to its target dose, whereas the second agent, ie, a -blocker, is often given in suboptimal doses, if given at all. 11,12 Moreover, there are theoretical considerations suggesting that it may be more beneficial to initiate treatment for CHF with a -blocker than with an ACE inhibitor. The sympathetic system is systemically activated at an earlier stage than is the RAAS in CHF. 4 -Blockers effectively inhibit the activation of the sympathetic system and also of the RAAS, but ACE inhibitors have a less pronounced sympathoinhibitory effect. 13,14 Sudden death is the most prevalent cause of death in the early course of CHF and in mildly symptomatic CHF, and although there is limited evidence that ACE inhibitors prevent sudden death in CHF patients, -blockers are well documented in this regard. 5 8,18 21 So far, only 2 small studies have examined the importance of the order of drug initiation in patients with CHF. In the CARMEN (Carvedilol and ACE-inhibitor Remodelling Mild heart failure EvaluatioN) study, 22 treatment with the -blocker carvedilol was at least as effective as enalapril in terms of left ventricular volumes on echocardiography, whereas the combination therapy appeared superior; however, interpretation of that study was limited by the large proportion of patients previously receiving ACE inhibitors. In a more recent study by Sliwa et al, 23 starting therapy for newly diagnosed heart failure with the -blocker carvedilol followed by the ACE inhibitor perindopril, as compared with the opposite order, had a superior effect on New York Heart Association (NYHA) class, left ventricular ejection fraction (LVEF), plasma N-terminal pro-brain natriuretic peptide concentration, and LV volumes. In summary, sufficient data do not currently exist to establish the optimum order of initiating CHF therapy. This was the objective of the CIBIS III trial, in which we compared the effect on mortality and hospitalization of initial monotherapy with either bisoprolol or enalapril for 6 months, followed by their combination for 6 to 24 months. Methods Study Design CIBIS III was an investigator-initiated, multicenter, prospective, randomized, open-label, blinded end point evaluation (PROBE) trial, 24 with 2 parallel groups. It was performed at 128 centers in 18 European countries (see the Appendix in the online-only Data Supplement at LATIONAHA /DC1) and in Australia and Tunisia between October 2002 and May The design and protocol of CIBIS III have been described previously. 25 After initial physician screening, patients were randomized, via random numbers produced by a dynamic balancing algorithm, 26 at the independent statistical center and were stratified according to NYHA class (II or III). Treatment was allocated by telephone from the statistical center. Randomized treatment with bisoprolol (1.25 mg QD) or enalapril (2.5 mg BID) was initiated, and study drugs were progressively uptitrated at 2-week intervals (Figure 1). Titration was mandatory, unless prohibited because of intolerance, but could be slower according to individual tolerance. The target maintenance dose for bisoprolol was 10 mg QD, and for enalapril it was 10 mg BID. The titration phase was followed by a maximum maintenance period of 16 weeks for the bisoprolol-first group and 22 weeks for the enalapril-first group. During the 6-month monotherapy phase, initiation of adjuvant therapy with an angiotensin-receptor blocker or an aldosterone-receptor blocker was not permitted, whereas continued preexisting aldosterone-receptor blocker therapy was. Open treatment with a -blocker or an ACE inhibitor was prohibited through-

3 2428 Circulation October 18, 2005 out the study period. The combination-therapy period immediately followed completion of the monotherapy phase. Early introduction of the second drug was allowed if medically legitimate, as judged by the end point committee. At the beginning of the combined-therapy phase, the complementary drug was uptitrated in a manner similar to the first uptitration of the initiated drug. Thereafter, patients were followed up on maintenance doses of the combination therapy. During the combined phase, initiation of adjuvant therapy with an angiotensin-receptor blocker or an aldosterone-receptor blocker was permitted if considered appropriate. Study medications and diuretics could be adjusted at the investigators discretion at any time point after randomization. The duration of the combination phase was between 6 months (for the last recruited patient) and 24 months (for the first recruited patient). Because patients were recruited over the course of 18 months, the total study duration was between 1 and 2.5 years. The primary end point was time-to-the-first-event of combined all-cause mortality or all-cause hospitalization throughout the study period. Secondary end points included the combined end point at the end of the monotherapy phase and the individual components of the primary end point, not only at study end, but also at the end of the monotherapy phase. Further secondary end points were cardiovascular death (CHF, myocardial infarction, ineffective internal defibrillator shock, other cardiovascular death, or unknown cause of death) and cardiovascular hospitalization (myocardial infarction, worsening of CHF, or other cardiovascular hospitalization). Safety end points included permanent treatment cessation and the need for early introduction of the second drug as indicators of drug tolerability. An independent steering committee was responsible for study design, management, data analysis, and interpretation. An independent data safety monitoring board continuously reviewed study data. A masked independent end point committee adjudicated all end points and validated the period of follow-up in each patient for the per-protocol analysis. An independent clinical trial data center was responsible for the random allocation, maintained the study database, performed statistical analysis, and was responsible for providing regular safety reports to the data safety monitoring board. The role of the sponsor was restricted to study medication supply, on-site monitoring, and collection of data. Two representatives of the sponsor were nonvoting members of the steering committee. The study was approved by the regional ethics committees of the participating centers and was conducted in accordance with the Declaration of Helsinki. Patients Eligible patients were male or female patients aged at least 65 years, with mild to moderate CHF (NYHA class II or III), and LVEF 35% according to echocardiography assessment within the 3 months before randomization. CHF was required to have been clinically stable, without clinically relevant fluid retention or diuretic adjustment within the last 7 days before randomization. All patients gave written informed consent before enrollment into the study. The main exclusion criteria were treatment with an ACE inhibitor, an angiotensin-receptor blocker, or a -blocker for 7 days during the 3 months before randomization; acute coronary syndrome within 3 months before randomization; percutaneous coronary intervention or coronary bypass surgery planned or performed within 3 months before randomization; stroke within 1 month or with permanent neurological sequelae within 6 months before randomization; heart rate at rest 60 beats per minute without a functioning pacemaker; supine systolic blood pressure 100 mm Hg at rest; important electrolyte disturbances; serum-creatinine 220 mol/l; atrioventricular block greater than first degree without a functioning pacemaker; and obstructive lung disease contraindicating bisoprolol treatment. Statistical Analysis Sample size estimation and interim analysis procedures have been described elsewhere. 25 Two interim analyses of safety were performed. Given that the study could not be stopped prematurely Figure 2. Illustration of the definition of noninferiority. The upper line illustrates that superiority for bisoprolol-first vs enalapril-first treatment is shown if the entire 95% CI for the between-group difference is below HR 1.0. As demonstrated by the second line, inferiority is shown if the entire 95% CI is above HR 1.0. Noninferiority is illustrated by line 3, showing the actual result of the intention-to-treat analysis in CIBIS III. Line 4 depicts a result that is not superior, inferior, or noninferior ie, the actual result of the per-protocol analysis in CIBIS III. because of between-group differences with regard to the primary end point or other efficacy end points, no adjustment of probability values for the primary end point or other efficacy end points was necessary. The first objective of this study was to examine noninferiority for bisoprolol-first versus enalapril-first treatment with regard to the primary end point. We postulated a 40% event rate over the entire study period in the enalapril-first group and a 34% event rate in the bisoprolol-first group, leading to a probability of at least 92% at the 2.5% significance level to show noninferiority of bisoprololfirst versus enalapril-first treatment with 450 patients per study group available for the per-protocol analysis. As prespecified in the 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 (CI) for the betweengroup difference in absolute risk was 5% (40% with enalapril-first versus 45% with bisoprolol-first treatment), corresponding to a hazard ratio (HR) of 1.17, in the per-protocol sample (Figure 2). The secondary objective was to examine superiority with regard to primary and secondary end points. All analyses were performed in both the intention-to-treat sample and the per-protocol sample with the SAS statistical software. The intention-to-treat sample included all randomized patients. The per-protocol sample included all patients treated according to protocol, excluding those who never took any study medication; were not clinically stable for at least 7 days before randomization; were treated with an ACE inhibitor, an angiotensin-receptor blocker, or a -blocker for 7 days during the 3 months before randomization; met any other exclusion criteria; showed inadequate compliance on 3 consecutive visits; or contravened protocol (illegitimately received the second agent before 6 months after randomization, illegitimately withdrew from randomized therapy prematurely, or never started the second drug during the combined-therapy phase). Patient data were included in the per-protocol analysis until the time of protocol violation, as judged by the masked end-point committee. Figure 3B depicts the number of patients included in the per-protocol sample at various time points. The primary end point, mortality and hospitalization separately, and relevant serious adverse events were analyzed by the log-rank test. 27 Survival curves were calculated with Kaplan- Meier estimates. Differences between groups in proportions of patients with events or numbers of events were analyzed by 2 test. Vital signs were analyzed with 1-way ANCOVA with a factor for treatment group and screening value as covariate. P 0.05 generally indicated statistical significance, but P denoted statistical significance for noninferiority because this analysis was 1 sided. The primary end point results are presented for the intention-to-treat

4 Willenheimer et al Initiation of Heart Failure Therapy 2429 diabetes. Baseline cardiovascular medication was similar in the 2 groups (Table 1). A total of 84.3% of the patients were receiving diuretic treatment, mostly a loop diuretic; 10% were taking potassium-sparing diuretics; and 15% received an aldosterone-receptor blocker. Around one third received a cardiac glycoside, and two thirds were treated with antiplatelet agents. Around 15% were on hypoglycemic treatment. During the study, the use of aldosterone-receptor blockers and digoxin changed very little. In the bisoprolol-first group, aldosterone-receptor blockers were taken by 14.3% of the patients at baseline and by 14.1% at the last registration, whereas 32.9% and 31.1%, respectively, used digoxin. In the enalapril-first group, 12.3% used an aldosterone-receptor blocker at baseline and 11.1% at the last registration, whereas 30.7% and 32.7% used digoxin, respectively. Around 2% in both groups used an angiotensin-receptor blocker at the last registration. Figure 3. Kaplan-Meier plot of the combined primary end point (death or hospitalization) in the intention-to-treat sample (A) and in the per-protocol sample (B). Bisoprolol-first group is indicated by the solid line; enalapril-first group, the dotted line. In the intention-to-treat sample, the primary end point occurred in 178 patients allocated to bisoprolol-first vs 186 allocated to enalapril-first treatment (absolute difference 1.6%, 95% CI 7.6 to 4.4%, HR 0.94; 95% CI 0.77 to 1.16; noninferiority for bisoprolol-first vs enalapril-first treatment P 0.019). In the perprotocol sample, 163 patients allocated to bisoprolol-first treatment had a primary end point, vs 165 allocated to enalapril-first treatment (absolute difference 0.7%, 95% CI 6.6 to 5.1%, HR 0.97; 95% CI 0.78 to 1.21; noninferiority for bisoprolol-first vs enalapril-first treatment P 0.046). sample and for the per-protocol sample. All other efficacy results are presented only for the intention-to-treat sample. Results A total of 1010 patients were enrolled into the study and followed up for a mean of years (maximum of 2.10 years). The patient flow through the study is shown in Figure 4. Four subjects, 1 randomized to bisoprolol-first and 3 to enalapril-first treatment, did not take any study medication. Baseline characteristics were similar in the 2 groups (Table 1). Mean age of the patients was 72.4 years, and 68.2% were male. The most common cause of CHF was ischemic heart disease (62.4%), and in 36.5% the cause was judged to be hypertension. Mean LVEF was 28.8%, and patients were evenly distributed between NYHA classes II and III. Approximately half of the patients had a history of acute myocardial infarction, 10% had a history of peripheral vascular disease, 10% had a history of cerebrovascular disease, and 20% had Morbidity and Mortality In the intention-to-treat sample, there were 178 patients (35.2%) with a primary end point in the bisoprolol-first group, as compared with 186 (36.8%) in the enalapril-first group (absolute difference 1.6%, 95% CI 7.6 to 4.4%; HR 0.94; 95% CI 0.77 to 1.16; noninferiority for bisoprololfirst versus enalapril-first treatment, P 0.019) (Figures 2 and 3A). Thus, bisoprolol-first was noninferior to enalapril-first treatment in the intention-to-treat sample. In the per-protocol sample, there were 163 patients (32.4%) allocated to bisoprolol-first treatment with a primary end point, versus 165 (33.1%) allocated to enalapril-first treatment (absolute difference 0.7%; 95% CI 6.6 to 5.1%; HR 0.97; 95% CI 0.78 to 1.21; noninferiority for bisoprolol-first versus enalapril-first treatment, P 0.046) (Figures 2 and 3B). Thus, bisoprolol-first was not proven to be noninferior to enalaprilfirst treatment in the per-protocol sample. There were 65 deaths in the bisoprolol-first group, as compared with 73 in the enalapril-first group (HR 0.88; 95% CI 0.63 to 1.22; between-group difference P 0.44) (Figure 5). In the bisoprolol-first group, 151 patients were hospitalized, as compared with 157 in the enalapril-first group (HR 0.95; 95% CI 0.76 to 1.19; between-group difference P 0.66). The treatment groups were similar with regard to cardiovascular death: 55 with bisoprolol-first versus 56 with enalapril-first treatment (HR 0.97; 95% CI 0.67 to 1.40; between-group difference P 0.86). In the bisoprolol-first group, 63 patients had a worsening of CHF requiring hospitalization or occurring in hospital, as compared with 51 in the enalapril-first group (HR 1.25; 95% CI 0.87 to 1.81; between-group difference P 0.23) (Figure 6). At the end of the monotherapy phase, 109 bisoprolol-first patients had a primary end point, versus 108 enalapril-first patients (HR 1.02, 95% CI 0.78 to 1.33, between-group difference P 0.90); 23 versus 32 patients died (HR 0.72, 95% CI 0.42 to 1.24, between-group difference P 0.24); and 99 versus 92 patients had a hospitalization (HR 1.08, 95% CI 0.81 to 1.43, between-group difference P 0.59). In post hoc analysis of the first year, during which all patients were followed up, 155 bisoprololfirst patients versus 165 enalapril-first patients had a primary end point (HR 0.94, 95% CI 0.76 to 1.17, between-group

5 2430 Circulation October 18, 2005 Figure 4. Illustration of the patient flow through the study. PP indicates perprotocol. Of the 4 patients not receiving any study medication, 3 were allocated to enalapril-first and 1 to bisoprolol-first. Of the 9 patients excluded from the perprotocol sample already at study start, 7 were allocated to enalapril-first and 2 to bisoprolol-first. difference P 0.59), and there were 42 versus 60 deaths (HR 0.69, 95% CI 0.46 to 1.02, between-group difference P 0.065) (Figure 5). Safety During the monotherapy phase, in the bisoprolol-first group 326 patients (65%) were uptitrated to the target dose (10 mg QD) and 415 (82%) to at least 5 mg QD, whereas in the enalapril-first group 425 patients (84%) were uptitrated to the target dose (10 mg BID) and 476 (94%) to at least 5 mg BID. In the bisoprolol-first group 39 patients (7.7%) were subject to early introduction of the second drug, versus 37 (7.3%) in the enalapril-first group (P 0.81). During the monotherapy phase, 35 (6.9%) bisoprolol-first patients permanently discontinued randomized treatment, as compared with 49 (9.7%) patients in the enalapril-first group. Reasons for permanent treatment cessation are shown in Table 2. During the combined-therapy phase, in the bisoprolol-first group, 19 patients (4.2%) permanently discontinued bisoprolol and 47 (10.4%) discontinued enalapril. In the enalapril-first group, 24 patients (5.5%) permanently discontinued bisoprolol and 16 (3.7%) discontinued enalapril. During the combined phase, according to protocol, 448 (88.7%) patients allocated bisoprolol-first treatment were started on additional enalapril, and 431 (85.3%) of the enalapril-first patients were started on bisoprolol (P 0.13). Table 3 shows first and last prescribed study drug dose during the combined-therapy phase by treatment group. During the entire trial, 119 patients were formally withdrawn from the study, 60 with bisoprolol-first and 59 with enalapril-first treatment, and 19 bisoprolol-first patients and 16 enalapril-first patients could not subsequently be contacted about vital status. Five of these were lost to follow-up, 3 with bisoprolol-first and 2 with enalapril-first treatment. Reasons for protocol violation are shown in Table 4. The 2 strategies were comparable in terms of adverse events (Table 5). Both regimens equally affected blood pressure, during both the monotherapy and combined-therapy phases, whereas bisoprolol-first treatment had a more pronounced heart rate lowering effect during the monotherapy phase, with similar effect of both regimens at the end of the combined phase (Table 6). Only 3 patients received a biventricular pacemaker during the study, and no patient received an internal defibrillator. Subgroup Analysis In the prespecified subgroup analysis of LVEF, there was a significant interaction with regard to the primary end point Figure 5. Kaplan-Meier plot of all-cause death in the intentionto-treat sample. Bisoprolol-first group is shown by the solid line; enalapril-first group, the dotted line. There were 65 deaths in the bisoprolol-first group, as compared with 73 in the enalapril-first group (HR 0.88; 95% CI 0.63 to 1.22, between-group difference P 0.44).

6 Willenheimer et al Initiation of Heart Failure Therapy 2431 TABLE 1. Baseline Data (P 0.001). Among those with LVEF 28%, bisoprolol-first was significantly superior to enalapril-first treatment (HR 0.61, 95% CI 0.44 to 0.85, P 0.003), whereas an opposite trend was seen among those with LVEF 28% (HR 1.23, 95% CI 0.94 to 1.61, P 0.13) (Figure 7). This interaction was mainly due to a difference in noncardiovascular hospitalization during the monotherapy phase: Of the bisoprololfirst patients, 35 had a noncardiovascular hospitalization, versus 39 enalapril-first patients (9 versus 22 in the LVEF 28% group and 26 versus 17 in the LVEF 28% group). The same pattern was seen if the median LVEF value of 30% was used as the cutoff (interaction P 0.036). With this exception, analyses showed homogeneity across subgroups. Bisoprolol-First (n 505) Enalapril-First (n 505) Age, y Males 333 (65.9) 356 (70.5) NYHA class II/III 245/260 (48.5/51.5) 250/255 (49.5/50.5) Median duration of CHF, mo Left ventricular ejection fraction, % Serum creatinine, mol/l Heart rate, bpm Systolic blood pressure, mm Hg Cause of CHF* Coronary artery disease 309 (61.2) 321 (63.6) Hypertension 197 (39.0) 172 (34.1) Valvular heart disease 11 (2.2) 15 (3.0) Primary cardiomyopathy 49 (9.7) 51 (10.1) Other 68 (13.5) 50 (9.9) History of Hypertension 354 (70.1) 314 (62.2) Myocardial infarction 254 (50.3) 243 (48.1) Angina pectoris 259 (51.3) 255 (50.5) Peripheral vascular disease 37 (7.3) 56 (11.1) Cerebrovascular disease 52 (10.3) 49 (9.7) Diabetes 95 (18.8) 113 (22.4) Renal disease 93 (18.4) 89 (17.6) Anemia 10 (2.0) 8 (1.6) Prior percutaneous coronary intervention 22 (4.4) 18 (3.6) Prior coronary artery bypass grafting 45 (8.9) 40 (7.9) Pacemaker 38 (7.5) 33 (6.5) Baseline diuretic treatment 430 (85.1) 421 (83.4) Thiazide diuretics 97 (19.2) 115 (22.8) Loop diuretics 361 (71.5) 338 (66.9) Potassium-sparing diuretics 52 (10.3) 53 (10.5) Aldosterone-receptor blockers 72 (14.3) 62 (12.3) Baseline antiplatelet medication 345 (68.3) 334 (66.1) Baseline cardiac glycoside treatment 166 (32.9) 155 (30.7) Baseline hypoglycemic medication 72 (14.3) 86 (17.0) Values are mean SD, n (%), or median. *More than 1 cause may be given for each patient. Discussion ACE inhibitors and -blockers constitute standard treatment for CHF. 9,10 Although it is essential to combine an ACE inhibitor and a -blocker, even CHF patients treated by physicians experienced in CHF often do not receive this combination, especially not in adequate doses, and many patients remain on the initial drug for extended periods of time. 11,12,28 This is one reason why the order of initiating -blockers and ACE inhibitors may be important. However, with the exception of 2 smaller surrogate end point studies, 22,23 the optimum sequence of initiating these agents has not been examined until now. The CIBIS III trial showed that, in patients with NYHA class II or III CHF and impaired

7 2432 Circulation October 18, 2005 TABLE 3. First and Last Prescribed Study Drug Dose During the Combined-Therapy Phase According to Treatment Bisoprolol-First Enalapril-First Figure 6. Kaplan-Meier plot of patients with worsening heart failure requiring hospitalization or occurring while in hospital in the intention-to-treat sample. Bisoprolol-first group shown by solid line; enalapril-first group, dotted line. In the bisoprolol-first group, 63 patients had a worsening of CHF requiring hospitalization or occurring in hospital, as compared with 51 in the enalapril-first group (HR 1.25; 95% CI 0.87 to 1.81; betweengroup difference P 0.23). LVEF, there was no difference in terms of efficacy and safety between initiation of treatment with bisoprolol followed by enalapril and initiation of treatment with enalapril followed by bisoprolol. In terms of combined mortality and hospitalization, the 2 strategies led to a similar result, with superimposable Kaplan-Meier curves. The prespecified statistical criterion for noninferiority of bisoprolol-first versus enalapril-first treatment was met in the intention-to-treat sample but not in the per-protocol sample. However, the observed upper value of the 95% CI in the per-protocol sample was very close to the predefined noninferiority limit, TABLE 2. Reasons for Permanent Cessation of Study Medication During the Entire Study Period Reason* Bisoprolol-First (n 101) Enalapril-First (n 89) Total (n 190) Partly or fully withdrew consent Met an exclusion criterion Needed nonpermitted medication Lost to follow-up Nonspecified Adverse events Cardiac disorders Respiratory and thoracic disorders Vascular disorders Nervous system disorders Gastrointestinal disorders Various events *Reasons are not mutually exclusive. First Dose* Last Dose* First Dose* Last Dose* Enalapril dose n mg, % mg, % mg, % Bisoprolol dose n mg, % mg, % mg, % mg, % mg, % mg, % In the bisoprolol-first group, last prescribed doses of bisoprolol were significantly higher as compared with the enalapril-first group (P by Wilcoxon rank-sum test). In the enalapril-first group, last prescribed doses of enalapril were significantly higher as compared with the bisoprolol-first group (P 0.001). *During the combined-therapy phase. whereas it was clearly below the noninferiority limit in the intention-to-treat sample. The per-protocol analysis is traditionally the preferred and most conservative approach with regard to noninferiority. Nevertheless, in trials with relatively long follow-up periods, such as CIBIS III, the per-protocol sample becomes difficult to define. Despite efforts to avoid the introduction of any bias in this regard eg, by blinded definition of the per-protocol sample by the end point committee this is probably impossible to preclude. Therefore, the intention-to-treat analysis may be equally relevant in terms of noninferiority, especially because only 4 patients in the intention-to-treat sample did not take any study medication. The substantial number of protocol violations is likely to have affected our analyses of efficacy. Because the number of patients in the per-protocol sample rapidly diminished with time (as seen in Figure 3B) as a result of various protocol violations and withdrawals, this analysis had less statistical power as compared with the intention-to-treat analysis. This is also indicated by the wider 95% CI in the per-protocol sample, which suggests that inadequate statistical power might have been important to the lack of statistical significance for noninferiority in the per-protocol sample. No difference was seen between the 2 strategies in terms of all-cause hospitalization, and at the end of the monotherapy phase, the 2 strategies were similar with regard to the primary end point. The bisoprolol-first strategy tended to be more favorable in terms of survival; with bisoprolol-first treatment, fewer patients died, deaths occurred later, and the HR for mortality was 28% lower at the end of the monotherapy phase and 31% lower at 1 year, as compared with the enalapril-first strategy. In both study groups, the first-initiated therapy was given in higher doses during the combined-therapy phase, which may be relevant to survival. This observation also

8 Willenheimer et al Initiation of Heart Failure Therapy 2433 TABLE 4. Protocol Violations Adjudicated Reason Bisoprolol-First (n 505, 100%) Enalapril-First (n 505, 100%) Total (n 1010, 100%) Entirely excluded from per-protocol analysis 2 (0.4) 7 (1.4) 9 (0.9) Some data excluded from per-protocol analysis* 83 (16.4) 78 (15.4) 161 (15.9) Unstable CHF within 7 days before randomization 0 (0.0) 1 (0.2) 1 (0.1) Nonpermitted treatment before randomization 1 (0.2) 2 (0.4) 3 (0.3) Failure to meet inclusion criteria 0 (0.0) 1 (0.2) 1 (0.1) Never took any study medication 1 (0.2) 3 (0.6) 4 (0.4) Never started second drug 4 (0.8) 7 (1.4) 11 (1.1) Inadequate study medication compliance 3 (0.6) 2 (0.4) 5 (0.5) Medically illegitimate early introduction of second drug 33 (6.5) 21 (4.2) 54 (5.3) Medically illegitimate permanent treatment cessation 48 (9.5) 48 (9.5) 96 (9.5) Values are n (%). More than 1 cause may be given for each patient. There were no statistically significant between-group differences. *Data were included in the per-protocol analysis until the time of the protocol violation. underscores the findings of prior surveys that the firstinitiated therapy stands a better chance of being uptitrated to the target dose. 11,12,28 In contrast to the effects on survival, the bisoprolol-first strategy was associated with a trend toward a higher frequency of the end point worsening of CHF requiring hospitalization or occurring in hospital, especially during the early course of the study. We do not have a definitive explanation for this finding. It is unlikely to have been caused by too-fast uptitration of bisoprolol. We used a slower initial uptitration than used previously. 6 Although -blockers have been shown to decrease CHF hospitalization, 6 8 it is well recognized that initiation and uptitration of a -blocker may cause minor and temporary deterioration of CHF. 9,10 This is likely due to an early and temporary negative inotropic effect, which in the clinical setting might appear differently in patients not receiving background ACE inhibition. In clinical practice, this usually is handled by temporarily increasing the diuretic dose, which in some hospitalized patients in CIBIS TABLE 5. Serious Adverse Events and Adverse Events in the Safety Sample* Bisoprolol-First (n 504) Patients Reporting, n (%) No. of Reports Enalapril-First (n 502) Patients Reporting, n (%) No. of Reports Monotherapy phase Serious adverse 113 (22.4) (22.1) 163 events Adverse events 316 (62.7) (63.5) 861 Entire study period Serious adverse 184 (36.5) (37.3) 354 events Adverse events 396 (78.6) (78.7) 1769 There were no statistically significant between-group differences. *The safety sample included all patients who had received at least one dose of randomized treatment. III might have been sufficient reason for reporting this as an end point. The generally limited prior experience of uptitrating a -blocker in CHF patients not on an ACE inhibitor may also be of importance in this regard. With more experience, it is possible that worsening of CHF during initial uptitration of the -blocker can be avoided more often. Prespecified subgroup analysis showed an interaction with regard to LVEF. However, this seemed to be mainly due to an imbalance between groups in noncardiovascular hospitalizations during the monotherapy phase and might, therefore, not be of clinical significance. We chose an open-label trial design because it would have been virtually impossible to separately adjust the doses of the 2 study drugs during the combined study phase, eg, in response to side effects, had we masked both treatments. The design is especially relevant for -blocker studies in CHF, in which the titration schedule depends on individual response to treatment. TABLE 6. Effects on Blood Pressure and Heart Rate Bisoprolol-First Enalapril-First n Mean SD n Mean SD Systolic blood pressure, mm Hg Baseline End of monotherapy At 1 year Diastolic blood pressure, mm Hg Baseline End of monotherapy At 1 year Heart rate, bpm Baseline End of monotherapy * * At 1 year *Between-group difference P (besides this, there were no significant between-group differences).

9 2434 Circulation October 18, 2005 Figure 7. Prespecified subgroup analyses with regard to the primary end point in the intention-to-treat sample. BP indicates blood pressure. The interaction with regard to LVEF was significant (P 0.003), but all other analyses showed homogeneity across subgroups. Individual dose adaptations of -blockers do not seem to lead to an inferior effect on survival and morbidity, as long as the -blocker is uptitrated according to the patient s true tolerance. 29,30 The central telephone randomization procedure and the blinded end-point evaluation by the independent end-point committee limited any possible investigator bias. The generalizability of our findings to routine clinical practice might depend on several factors, with the patient sample and the titration schedule probably being among the most important. Patients had to be at least 65 years of age to be included in the CIBIS III study, and the mean age of the included patients was 72 years, which is close to the mean age of 75 years in the general population of CHF patients in clinical practice. 1 We applied a moderately aggressive uptitration to ensure that both study drugs were uptitrated to target doses in a timely fashion. A different titration schedule, whether less or more aggressive for either or both drugs, might have produced another result. In conclusion, in patients at least 65 years of age, with mildly to moderately symptomatic CHF and impaired LVEF, initiation of treatment for CHF with bisoprolol followed by combined therapy with enalapril, as compared with the opposite order, showed similar rates of combined death and all-cause hospitalization and a trend toward improved survival. There is no previously proven superior order of initiating these agents in CHF and, although noninferiority of bisoprolol-first versus enalapril-first treatment was not proven in the per-protocol analysis, our results indicate that it may be as safe and efficacious to initiate treatment for CHF with bisoprolol as with enalapril. Acknowledgments We thank all patients who participated in CIBIS III, as well as the study investigators (see the Appendix in the online-only Data Supplement), members of the Data Safety Monitoring Board and the End Point Committee (see Appendix), study nurses, and coordinators, whose work made the trial possible. This study was supported by Merck KGaA, Darmstadt, Germany. Disclosure All authors except A. Skene and Drs van de Ven and Verkenne have received honoraria from Merck KGaA for lectures and have also received honoraria from other pharmaceutical companies and/or served as consultants or on advisory boards. Drs van de Ven and Verkenne are full-time employees of Merck KGaA. A. Skene is an owner of The Nottingham Clinical Research Limited, Nottingham, UK. References 1. Erhardt LR, Cline CM. Organisation of the care of patients with heart failure. Lancet. 1998;352(suppl I): Young JB, Dunlap ME, Pfeffer MA, Probstfield JL, Cohen-Solal A, Dietz R, Granger CB, Hradec J, Kuch J, McKelvie RS, McMurray JJ, Michelson EL, Olofsson B, Ostergren J, Held P, Solomon SD, Yusuf S, Swedberg K; Candesartan in Heart failure Assessment of Reduction in Mortality and morbidity (CHARM) Investigators and Committees. Mortality and morbidity reduction with candesartan in patients with chronic heart failure and left ventricular systolic dysfunction: results of the CHARM low-left ventricular ejection fraction trials. Circulation. 2004; 110: Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L, Klein W, Tavazzi L; Cardiac Resynchronization-Heart Failure (CARE-HF) Study Investigators. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med. 2005;352: Packer M. Pathophysiology of heart failure. Lancet. 1992;340: Garg R, Yusuf S. Overview of randomized trials of angiotensinconverting enzyme inhibitors on mortality and morbidity in patients with

10 Willenheimer et al Initiation of Heart Failure Therapy 2435 heart failure. Collaborative on ACE Inhibitor Trials. JAMA. 1995; 273: CIBIS II Investigators and Committees. The Cardiac Insufficiency Bisoprolol Study II (CIBIS II): a randomized trial. Lancet. 1999;353: 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: Packer M, Coats AJ, Fowler MB, Katus HA, Krum H, Mohacsi P, Rouleau JL, Tendera M, Castaigne A, Roecker EB, Schultz MK, DeMets DL; Carvedilol Prospective Randomized Cumulative Survival Study. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med. 2001;344: Swedberg K, Cleland J, Dargie H, Drexler H, Follath F, Komajda M, Tavazzi L, Smiseth OA, Gavazzi A, Haverich A, Hoes A, Jaarsma T, Korewicki J, Levy S, Linde C, Lopez-Sendon JL, Nieminen MS, Pierard L, Remme WJ; Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005). Eur Heart J. 2005;26: Hunt SA, Baker DW, Chin MH, Cinquegrani MP, Feldman AM, Francis GS, Ganiats TG, Goldstein S, Gregoratos G, Jessup ML, Noble RJ, Packer M, Silver MA, Stevenson LW, Gibbons RJ, Antman EM, Alpert JS, Faxon DP, Fuster V, Gregoratos G, Jacobs AK, Hiratzka LF, Russell RO, Smith SC Jr; American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1995 Guidelines for the Evaluation and Management of Heart Failure); International Society for Heart and Lung Transplantation; Heart Failure Society of America. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1995 Guidelines for the Evaluation and Management of Heart Failure): developed in collaboration with the International Society for Heart and Lung Transplantation; endorsed by the Heart Failure Society of America. Circulation. 2001;104: Cleland JG, Cohen-Solal A, Aguilar JC, Dietz R, Eastaugh J, Follath F, Freemantle N, Gavazzi A, van Gilst WH, Hobbs FD, Korewicki J, Madeira HC, Preda I, Swedberg K, Widimsky J; IMPROVEMENT of Heart Failure Programme Committees and Investigators. Improvement programme in evaluation and management; Study on Diagnosis of the Working on Heart Failure of The European Society of Cardiology Management of heart failure in primary care (the IMPROVEMENT of Heart Failure Programme): an international survey. Lancet. 2002;360: Davies M, Hobbs F, Davis R, Kenkre J, Roalfe AK, Hare R, Wosornu D, Lancashire RJ. Prevalence of left-ventricular systolic dysfunction and heart failure in the Echocardiographic Heart of England Screening study: a population based study. Lancet. 2001;358: Campbell DJ, Aggarwal A, Esler M, Kaye D. Beta-blockers, angiotensin II, and ACE inhibitors in patients with heart failure. Lancet. 2001;358: Teisman AC, van Veldhuisen DJ, Boomsma F, de Kam PJ, Tjeerdsma G, Pinto YM, de Zeeuw D, van Gilst WH. Chronic beta-blocker treatment in patients with advanced heart failure: effects on neurohormones. Int J Cardiol. 2000;73: Uretsky BF, Sheahan RG. Primary prevention of sudden cardiac death in heart failure: will the solution be shocking? J Am Coll Cardiol. 1997;30: Huirkuri HV, Castellanos A, Myerburg RJ. Sudden death due to cardiac arrhythmias. N Engl J Med. 2001;345: Lane RE, Cowie MR, Chow AW. Prediction and prevention of sudden cardiac death in heart failure. Heart. 2005;91: Krum H. -Adrenoceptor blockers in chronic heart failure a review. Br J Clin Pharmacol. 1997;44: Heidenreich PA, Lee TT, Massie BM. Effect of beta-blockade on mortality in patients with heart failure: a meta analysis of randomized clinical trials. J Am Coll Cardiol. 1997;30: Lechat P, Packer M, Chalon S, Cucherat M, Arab T, Boissel JP. Clinical effects of beta-adrenergic blockade in chronic heart failure: a meta-analysis of double blind, placebo controlled, randomized trials. Circulation. 1998;98: Cleland JG, Massie BM, Packer M. Sudden death in heart failure: vascular or electrical? Eur J Heart Fail. 1999;1: Remme WJ, Riegger G, Hildebrandt P, Komajda M, Jaarsma W, Bobbio M, Soler-Soler J, Scherhag A, Lutiger B, Ryden L. The benefits of early combination treatment of carvedilol and an ACE-inhibitor in mild heart failure and left ventricular systolic dysfunction. The carvedilol and ACEinhibitor remodelling mild heart failure evaluation trial (CARMEN). Cardiovasc Drugs Ther. 2004;18: Sliwa K, Norton GR, Kone N, Candy G, Kachope J, Woodiwiss AJ, Libhaber C, Sareli P, Essop R. Impact of initiating carvedilol before angiotensin-converting enzyme inhibitor therapy on cardiac function in newly diagnosed heart failure. J Am Coll Cardiol. 2004;44: Hansson L, Hedner T, Dahlöf B. Prospective Randomized Open Blinded End-point (PROBE) study: a novel design for intervention trials. Blood Press. 1992;1: Willenheimer R, Erdmann E, Follath F, Krum H, Ponikowski P, Silke B, van Veldhuisen DJ, Van De Ven L, Verkenne P, Lechat P, on behalf of the CIBIS-III investigators. Comparison of treatment initiation with bisoprolol versus enalapril in chronic heart failure patients: rationale and design of CIBIS-III. Eur J Heart Fail. 2004;6: Signorini DF, Leung O, Simes RJ, Beller E, Gebski VJ, Callaghan T. Dynamic balance randomisation for clinical trials. Stat Med. 1993;12: Cox DR. Regression models and life-tables. J R Stat Soc. 1972;B34: Komajda M, Lapuerta P, Hermans N, Gonzalez-Juanatey JR, van Veldhuisen DJ, Tavazzi L, Poole-Wilson P, Le Pen C. Adherence to guidelines is a predictor of outcome in chronic heart failure: the MAHLER survey. Eur Heart J. 2005;26: Wikstrand J, Hjalmarson A, Waagstein F, Fagerberg B, Goldstein S, Kjekshus J, Wedel H, MERIT-HF Study. Dose of metoprolol CR/XL and clinical outcomes in patients with heart failure: analysis of the experience in metoprolol CR/XL randomized intervention trial in chronic heart failure (MERIT-HF). J Am Coll Cardiol. 2002;40: Simon T, Mary-Krause M, Funck-Brentano C, Lechat P, Jaillon P. Bisoprolol dose-response relationship in patients with congestive heart failure: a subgroup analysis in the cardiac insufficiency bisoprolol study (CIBIS II). Eur Heart J. 2003;24: