Reasons for Review

Since the publication of the 2003 guidelines on chronic heart failure, evidence on the use of ACEI in the management of patients with HFPEF, especially the elderly, has been published.

Populations

• LVEF ≥ 40% ^53,54
• Mean age range: 75–78 years ^53,54
• >50% female ^53,54

Background medication

• Beta Blockers >60% ⁵⁴
• Beta Blockers <20% ⁵³

Intervention

• Quinapril (up to 40mg) ⁵³
• Perindopril (4mg) ⁵⁴

Comparison

• Placebo ^53,54

Relative value placed on the outcomes considered

In the two appraised trials^54;53 compared to placebo, ACEI had no effect on all cause mortality at 6–12 months or on the rate of adverse events at 6–18 months. In the small study by Zi et al, there was no impact on quality of life at 6 months or on the rate of improvement of patients with NYHA Class III to II at 6 months. In PEP-CHF trial⁵⁴, treatment with ACEI resulted in significant (35%) reduction in the rate of heart failure hospitalisation at 1 year, while it had no impact on cardiovascular mortality at 1 year.

There was no difference between those given placebo and those given ACEI in terms of the side effects, quality of life or the New York Heart Association functional class.

However, at completion of the PEP-CHF study by Cleland et al⁵⁴, there was an insignificant trend towards reduced hospitalisation at 5 years. The significant reduction in heart failure hospitalisation at 1 year in PEP-CHF was derived from a post-hoc analysis. The GDG felt both trials were underpowered with wide confidence intervals around the results. Therefore, the GDG believed that there was insufficient evidence of effectiveness of ACEI in HFPEF to recommend their general use in patients with HFPEF.

Quality of evidence

The evidence reported on all the parameters alluded to above from the two trials was of moderate quality.

Trade-off between clinical benefits and harms

While the GDG did not consider that a post hoc finding of a reduction in heart failure hospitalisation at one year was sufficient to recommend the widespread use of ACEI in HFPEF in the absence of any other significant benefit, it was noted that there was no evidence of significant harm either, with adverse event rates similar in active treatment and placebo arms of the two trials.

Trade-off between net health benefits and resource use

No relevant economic analysis was identified from our review assessing the cost-effectiveness of ACEI in patients with heart failure and preserved LVEF. From clinical trials, net resource use would be likely to be low given that hospital admissions might be reduced, and ACEI therapy is of relatively low cost. However, the GDG noted that the pre-specified hospitalisation endpoint was non-significant and the GDG therefore did not attach weight to the reduction of hospitalisation at one year.

Use of ACEI in left ventricular systolic dysfunction

The evidence base for use of ACEI in left ventricular systolic dysfunction was not formally reviewed. The GDG noted the 2003 recommendations. The GDG endorsed that ACEI doses should be up-titrated slowly up to the target doses used in randomised controlled trials (RCTs). The safety of treatment with ACEI is best achieved by adhering to the protocols used in the clinical trials and proposed in the 2003 guidelines as practical recommendations, as well as the recommendations of the NICE chronic kidney disease guideline. It is particularly important to measure the serum urea, electrolytes, creatinine and eGFR before the initiation of ACEI, following each dose increment, and then at regular intervals.

Reasons for Review

Since the 2003 guidelines, randomised clinical trials have been published looking at comparing selective and non-selective beta-blockers in the treatment of heart failure, at the order of therapeutic strategies (ACEI/BB), and at the use of other beta-blockers in elderly patients with heart failure. There may also be some indirect evidence of the use of these agents in patients with heart failure with preserved left ventricular ejection fraction (HFPEF).

a. Beta blockers versus placebo in older adults with chronic heart failure

Five papers were identified comparing beta-blockers with placebo in older adults with chronic heart failure ^{55; 56; 57; 58; 59}. Two of these papers were in a sub-population derived from RCTs carried out on all patients with chronic heart failure ^{55; 56}. Table 5.1 below summarises the patient population and intervention for each study. Patients with COPD were excluded in all studies except one study ⁵⁸.

Table 5.1

Patient population and intervention: beta blockers in older adults with heart failure.

b. Evidence profile: Beta blockers versus placebo for patients with LVEF > 35%

One paper pre-specified subanalysis analysis from SENIORS exploring the efficacy of beta-blockers in patients with LVEF > 35%.

Table 5.2

Population and intervention: efficacy of beta blockers in patients with LVEF >35%.

NOTE: The study reported the following statistically significant differences between patients with reduced LVEF and those with preserved LVEF at baseline:

• Proportion of women: LVEF ≤ 35% 29.8%; LVEF > 35% 49.9%
• NYHA functional class II: LVEF ≤ 35% 52.8%; LVEF > 35% 62.5%
• NYHA functional class III: LVEF ≤ 35% 42.5%; LVEF > 35% 32.2%
• Sitting systolic blood pressure (mm Hg): LVEF ≤ 35% 135.5; LVEF > 35% 145.4
• Sitting diastolic blood pressure (mm Hg): LVEF ≤ 35% 79.2; LVEF > 35% 82.9
• Proportion on diuretic: LVEF ≤ 35% 87.9%; LVEF > 35% 83.1%
• Proportion on Angiotensin converting enzyme inhibitor: LVEF ≤ 35% 80.5%; LVEF > 35% 85.9%
• Proportion on Angiotensin II antagonist: LVEF ≤ 35% 9.9%; LVEF > 35% 5.6%
• Proportion of Aldosterone antagonist: LVEF ≤ 35% 32.1%; LVEF > 35% 5.6%

c. Selective vs non-selective beta blockers in chronic heart failure in reduced LVEF?

Three papers were identified comparing selective with non-selective β blockers for chronic heart failure^60;61;62. One of the papers⁶¹ reported on additional data from the main study⁶⁰. Both studies excluded patients with COPD. Table 5.3 below summarises the patient population and interventions by study.

Table 5.3

Patient population and interventions: selective vs non-selective beta blockers.

d. Beta blockers then ACEI compared with ACEI then beta blockers in reduced LVEF

One study was identified comparing beta-blockers then ACEI with ACEI then beta-blockers ⁶³. Patients with COPD were excluded. Table 5.4 below summarises the patient population and intervention for each study.

Table 5.4

Patient population and intervention: BB then ACEI vs ACEI then BB.

a. Beta blockers versus placebo in older adults with chronic heart failure

Compared with placebo, beta-blockers had a significant reduction on

• Mortality – all cause up to 27 months [low quality]
• Sudden death – up to 24 months [low quality]

Compared with placebo, beta-blockers were associated with no significant differences for:

• All cause hospitalisation – up to 27 months [moderate quality]
• Quality of life – Minnesota Living with Heart Failure at 40 weeks [low quality]
• Adverse events – no. of patients at 40 weeks [low quality]
• Adverse events – no. of patients (leading to withdrawal of study medication) at 12 months [low quality]

The evidence profile below summarises the quality of evidence and outcome data from five papers^{55,56,57,58,59} comparing beta-blockers with placebo in older adults with chronic heart failure.

Evidence profile

Comparison of beta-blockers with placebo in older adults. Bibliography: Deedwania PC, Gottlieb S, Ghali JK et al. Efficacy, safety and tolerability of beta-adrenergic blockade with metoprolol CR/XL in elderly patients with heart failure. European Heart (more...)

b. Beta blockers versus placebo for patients with preserved LVEF (LVEF > 35%)

For patients with LVEF > 35%, there was no significant difference between beta-blockers and placebo for:

• All cause hospitalisation or CV hospitalisation (no of patients) at 21 mths [moderate quality]
• All cause mortality (no of patients) at 21 mths [moderate quality]
• All cause mortality –at 21 mths [moderate quality]
• All cause hospitalisation - (no of patients) at 21 mths [moderate quality]

The evidence profile below summarises the quality of evidence and outcome data from the one paper comparing beta-blockers with placebo for chronic heart failure and preserved LVEF

Evidence profile

Comparison of beta-blockers with placebo for chronic heart failure and preserved LVEF (LVEF > 35%). Question: Should Beta blockers be used for Chronic heart failure - older adults? Bibliography: van Veldhuisen DJ, Cohen SA, Bohm M et al. Beta-blockade (more...)

c. Selective vs non-selective beta blockers in chronic heart failure?

Compared to non-selective beta blockers, selective beta-blockers were associated with a significant increase in:

• Mortality – all cause mean follow-up 58 months [moderate quality]
• Sudden death - mean follow-up 58 months [moderate quality]

Compared to non-selective beta blockers, selective beta-blockers were associated with no significant differences for:

• Mortality and hospitalisation – all cause mean follow-up 58 months [high quality]
• Quality of life – Minnesota Living with Heart Failure follow-up 12 weeks [moderate quality]
• Adverse events – no. of patients experiencing mean follow-up 58 months [high quality]

The evidence profile below summarises the quality of evidence and outcome data from three papers comparing selective with non-selective beta blockers for chronic heart failure ^{60; 61; 62}.

Evidence profile

Comparison of selective vs non-selective beta blockers. Question: Should Selective BB vs non-selective BB be used for chronic heart failure? Bibliography: Bibliography: Poole-Wilson PA SK. Comparison of carvedilol and metoprolol on clinical outcomes in (more...)

d. Evidence profile: Beta blockers then ACEI compared with ACEI then beta blockers

Compared to ACEI then beta blockers, beta blockers then ACEI were associated with no significant differences for:

• Mortality and hospitalisation – all cause mean follow-up 1.22 years [high quality]
• Mortality – all cause mean follow-up 1.22 years [moderate quality]
• Hospitalisation – all cause mean follow-up 1.22 years [high quality]
• Sudden death - mean follow-up 1.22 years [moderate quality]
• Adverse events – no. of patients experiencing mean follow-up 58 months [high quality]

The evidence profile below summarises the quality of evidence and outcome data from one study comparing beta blockers then ACEI with ACEI then beta blockers ⁶³.

Evidence profile

Comparison of BB then ACEI vs ACEI then BB. Bibliography: Willenheimer R, van Veldhuisen DJ, Silke B et al. Effect on survival and hospitalisation of initiating treatment for chronic heart failure with bisoprolol followed by enalapril, as compared with (more...)

Relative value placed on the outcomes considered

The GDG noted that the new evidence concerned the use of beta-blockers in older people with heart failure, the relative effectiveness of the non-selective beta-blocker Carvedilol compared with the selective beta-blocker Metoprolol tartrate, and the sequencing of therapy: ACEI followed by beta-blockers compared with beta-blockers followed by ACEI.

Quality of evidence

The GDG noted the evidence for the use of Nebivolol in older people with heart failure in the SENIORS study ⁵⁸. The GDG reviewed the post-hoc analyses of two randomised controlled studies of the older adult population using Bisoprolol or Metoprolol CR/XL ^55,56. The consistency of the results of the post-hoc analyses in the elderly sub-groups (reduction in all cause mortality and sudden death) with the randomised controlled trial that had specifically looked at this population was noted.

Trade-off between clinical benefits and harms

The GDG noted that there were no studies that specifically looked at the use of beta-blockers in the treatment of HFPEF. One third of the population recruited into the SENIORS study of Nebivolol in heart failure in older adults ⁵⁸ were patients with a left ventricular ejection fraction >40%. While the size of effect in the sub-group with an ejection fraction >35% was of similar magnitude to that seen in patients with LVSD, the effect in the subgroup with higher ejection fraction was non-significant. ⁵⁹ The GDG considered that there was insufficient evidence to recommend using beta-beta-blockers in the treatment of HFPEF and that further research was required.

The GDG reviewed the COMET trial ⁶⁰ comparing the impact of the non-selective beta-blocker Carvedilol to the selective beta-blocker Metoprolol tartarate in the treatment of heart failure. Although the study suggested that Carvedilol was superior at reducing all cause mortality and sudden death, the GDG were not convinced that this difference between Carvedilol and the short acting Metoprolol tartrate was necessarily applicable to other beta-blockers. The GDG noted that the MERIT-HF trial used the long-acting Metoprolol Succinate, and CIBIS II trial used Bisoprolol. Both Metoprolol Succinate and Bisoprolol are selective beta-blockers with outcomes in heart failure not dis-similar to those achieved in the trials that used Carvedilol. The GDG concluded that the implication is that the best results can be achieved by using the beta-blocking agents of proven efficacy in heart failure, namely: Carvedilol, Metoprolol Succinate, Bisoprolol, and Nebivolol.

The GDG considered the CIBIS III trial⁶³, and noted that heart failure patients derived similar outcome of therapy with ACEI followed by beta-blockers, to those treated with beta-blockers followed by ACEI. The GDG accepted that both agents should be given in the absence of contra-indications irrespective of the sequence they are given. The GDG agreed that either agent (or both) could be commenced first (see Section 5.2.1 on ACEI). The clinical decision to use one of these two agents before the other, or to commence both of them simultaneously depends on the clinical status of the patient. Several factors could affect the choice, including the patient s blood pressure, heart rate, the presence of symptomatic ischaemia, arrhythmias and other co-morbidity.

The GDG expressed concern that certain subgroups of patients with heart failure continue to be under-treated with beta-blockers. These include patients with chronic obstructive pulmonary disease (COPD), peripheral vascular disease, diabetes mellitus, erectile dysfunction and older adults. Patients with asthma and reversible airway obstruction were excluded from the trials of beta blockers in heart failure. The remaining patients with COPD should be able to tolerate beta blockers, and are likely to benefit significantly from their use. These patients are undertreated when they develop heart failure, and their outcomes are worse than the average heart failure patient. There is no evidence that selective beta-blockers will worsen these patients pulmonary function. (Salpeter 2005)⁶⁶. Beta-blockers are often avoided in patients with peripheral vascular disease for fear of exacerbating intermittent claudication, but this concern is unfounded (Radack 1991)⁶⁷. Although patients with recently unstable diabetes mellitus were excluded from some trials of beta-blockers in heart failure, significant numbers of diabetic patients have been included in beta-blocker trials such as COMET with no evidence that diabetes adversely influenced the effectiveness of the beta-blocker ^{68; 69}(COMET, MERIT-HF). Erectile dysfunction can be caused by some beta-blockers, but there are many causes including other medications and vascular disease. Discussion of these factors with the patient and explanation of the symptomatic and prognostic impact of beta-blockers in heart failure will better inform the decisions made by the patient and the health professional regarding these agents.

There is now sufficient evidence to justify the use of beta-blockers licensed for heart failure in patients in these groups, with the exception of patients who have COPD with reversible obstructive pulmonary disease. This group was excluded from the trials using selective beta-blockers such as bisoprolol (CIBIS II) ⁵⁶ and Metoprolol CR/XL (MERIT-HF)⁵⁵. The GDG noted that beta-blockers can be used in irreversible COPD. Moreover, in a meta-analysis of the trials on cardio-selective beta-blockers used in mild to moderate reversible COPD, no clinically significant adverse respiratory effects were demonstrated. (Salpeter 2005)⁶⁶.

The GDG suggested that if practitioners have particular concerns about side effects in patients with heart failure who also have irreversible COPD or peripheral vascular disease, then a selective beta-blocker licensed for heart failure could be considered.

The GDG considered the issue of managing patients who develop heart failure while on a beta-blocker not licensed for heart failure for another indication such as angina, hypertension, or arrhythmia. Contrary to the 2003 guidance, the GDG felt that it would be appropriate to switch to an agent licensed for use in heart failure, given the demonstrated significant impact these agents have on morbidity and mortality.

The GDG endorsed the 2003 practical recommendations. It is important, during the uptitration of beta-blockers, to monitor the patient s pulse rate, blood pressure and the clinical status, to avoid side effects such as symptomatic bradycardia and symptomatic hypotension. The uptitration should be undertaken gradually and slowly to achieve the target doses used in the clinical trials, if tolerated. The patient needs to be informed that transient pulmonary congestion could occur at times during uptitration of beta-blockers.

Trade-off between net health benefits and resource use

From the 2003 Guideline²², economic evidence on beta-blockers consistently showed beta-blockers to be cost effective. Our review added a study⁶⁴ that addressed the use of beta blockers in older adults with heart failure. This study⁶⁴ demonstrated that these agents are also cost-effective for this specific population.

Reason for review

Since the publication of the 2003 guideline, new evidence for the use of Aldosterone Antagonists in heart failure has been published. NICE guidance on the management of patients with myocardial infarction includes advice on the use of aldosterone antagonists in patients with heart failure following acute myocardial infarction ⁷⁰.

In patients on ACEI and beta-blockers who remain symptomatic, aldosterone antagonists as well as other options may be indicated.

Aldosterone antagonist + optimal medical management vs. placebo + optimal medical management

Three papers from the EPHESUS trial programme were identified comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with heart failure post-MI ^71,72;73

PITT 2003 compared eplerenone with placebo in patients 3–14 days after acute myocardial infarction (MI) with left ventricular dysfunction. PITT 2005 was a post-hoc analysis reporting further outcomes at 30 days and PITT 2006 reported results for the subgroup of patients included in the EPHESUS trial with severe left ventricular impairment (LVEF ≤30%).

Two studies were identified comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with heart failure due to severe left ventricular systolic dysfunction (LVEF <35%) ^74,75.

PITT 1999 was a part of the RALES study comparing spironolactone with placebo in patients with heart failure and severe LVSD (LVEF <35%). Patients were included with a history of NYHA class II through IV, a left ventricular ejection fraction ≤ 35%, and a history of NYHA class III or IV within the prior six months of enrolment. ANON 1996 ⁷⁵ was performed by the RALES investigators, this trial was intended as a dose finding trial for spironolactone in patients with HF due to severe LVSD (LVEF <35%).

The results from the EPHESUS severe heart failure subgroup were not meta-analysed with these results due to severe heterogeneity for the outcome heart failure hospitalisation, which may have been caused by the different populations (heart failure vs. heart failure post-MI), the different type of aldosterone antagonist used (spironolactone vs. eplerenone) or the difference in outcome (nonfatal HF hospitalisation vs. HF hospitalisation).

Three studies were identified comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with chronic heart failure ^76–78.

Barr (1995) ⁷⁸ compared spironolactone with placebo in a population with chronic heart failure (CHF) secondary to coronary heart disease. Macdonald (2004) ⁷⁷ compared spironolactone with placebo in a population with mild heart failure, defined as patients who at diagnosis their CHF had been at least NYHA class II, but optimising their treatment had improved the patients condition substantially into a stable and less symptomatic one. Agostoni (2005) ⁷⁶ compared spironolactone with placebo in a population with CHF and reduced lung diffusion.

a. Aldosterone antagonists plus optimal medical management vs. placebo plus optimal medical management in patients with heart failure post-MI

Compared with placebo, aldosterone antagonists resulted in a significant reduction of:

• Mortality all cause at 30 days [moderate quality]
• Mortality all cause at 16 months [high quality]
• Mortality all cause at 16 months – subgroup: severe LVSD / LVEF <35% [moderate quality]
• Sudden death at 16 months [moderate quality]
• Sudden death at 16 months – subgroup: severe LVSD / LVEF <35% [moderate quality]

Compared with placebo, aldosterone antagonists significantly increased:

• Hyperkalaemia at 16 months [high quality]

Compared with placebo, aldosterone antagonists had a non-significant effect on:

• Sudden death at 30 days [high quality]
• HF hospitalisation at 30 days [moderate quality]
• Nonfatal HF hospitalisation at 16 months – subgroup: severe LVSD / LVEF <35% [moderate quality]
• All hospitalisation at 16 months [high quality]

The evidence profile below summarises the quality of the evidence and outcome data from 3 studies ^71–73 comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with heart failure post-MI.

Evidence profile

Aldosterone antagonists plus optimal medical management vs. placebo plus optimal medical management in patients with heart failure post-MI. Question: Should aldosterone antagonist vs placebo be used for chronic heart failure post-MI? Bibliography: Pitt (more...)

a. Aldosterone antagonists plus optimal medical management vs. placebo plus optimal medical management in patients with heart failure due to severe LVSD (LVEF <35%)⁹

Compared with placebo, aldosterone antagonists had a significant reduction on:

• Mortality all cause at 24 months [moderate quality]
• HF hospitalisation at 24 months [moderate quality]

Compared with placebo, aldosterone antagonists had a significant increase on:

• Gynecomastia in men at 24 months [high quality]

Compared with placebo, aldosterone antagonists a non-significant increase on:

• Hyperkalaemia at 3 to 24 months [low quality]

The evidence profile below summarises the quality of the evidence and outcome data from 2 studies ^74,75 comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with heart failure due to severe LVSD (LVEF <35%).

Evidence profile

Aldosterone antagonists plus optimal medical management vs. placebo plus optimal medical management in patients with heart failure due to severe LVSD (LVEF <35%). Question: Should aldosterone antagonist vs placebo be used for heart failure due (more...)

b. Aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with chronic heart failure

Compared with placebo, aldosterone antagonists non-significantly increased:

• Hyperkalaemia >5.5 mmol/l at two months [low quality]
• Raised creatinine >300 umol/l at 8 weeks [low quality]

Compared with placebo, aldosterone antagonists non-significantly worsened:

• Quality of life- Minnesota Living with Heart Failure Questionnaire (MLWHFQ) score at 6 months [low quality]

Compared with placebo, aldosterone antagonists had a non-significant reduction on:

• Creatinine mean change at 6 months [low quality]

The evidence profile below summarises the quality of the evidence and outcome data from 3 studies ^76–78 comparing aldosterone antagonists plus optimal medical management with placebo plus optimal medical management in patients with chronic heart failure.

Evidence profile

Aldosterone antagonists plus optimal medical management vs. placebo plus optimal medical management in patients with chronic heart failure. Question: Should aldosterone antagonist vs placebo be used for all chronic heart failure? Bibliography: Agostoni (more...)

UK study assessing eplerenone

Duerden et al. (2008)⁷⁹ presented a cost-effectiveness analysis conducted from a UK NHS perspective with a 3-year time horizon (reporting cost per life-year gained). This analysis was based on the EPHESUS trial and assessed the addition of eplerenone to optimal medical treatment in patients with heart failure and left ventricular systolic dysfunction post acute myocardial infarction. For the placebo cohort, resource use estimates were calculated using data from the Office of National Statistics, data from the England and Scotland NHS, and probabilities published by the NICE clinical guideline on secondary prevention of myocardial infarction⁷⁰. In addition for the placebo cohort, survival estimates were derived from an 18-month epidemiological study assessing patients with all-cause heart failure and carried out in West London (Cowie 2000)⁹). Survival estimates from this study were extrapolated to 3 years (predicting a 48% survival). For the eplerenone cohort, additional resource use and additional survival were taken from EPHESUS (16-month follow-up) and extrapolated to 3 years. Costs considered in this assessment were the hospitalisation cost and the cost of eplerenone (additional drug cost for the treatment cohort). A 100% adherence and compliance to eplerenone was assumed. Future costs and benefits were discounted at 3.5% per annum. The sensitivity analysis varied mortality rates (increasing by 10%, 15%, and 20%). Table 5.7 presents the quality and applicability assessment of this economic analysis.

Table 5.7

Economic study assessment.

Irish study assessing spironolactone

Tilson et al. (2003)⁸⁰ conducted a cost-effectiveness analysis reporting cost per life-year gained and was based on the RALES trial. The analysis was developed from an Irish perspective and for a 10-year time horizon. The assessed population were patients with severe chronic heart failure (NYHA class III & IV) and left ventricular systolic dysfunction with a mean age of 65 years. Adding spironolactone to optimal medical management was compared to optimal medical treatment only (might include diuretics, ACEI, digoxin, BB, or a combination of these). Probabilities of death and hospitalisation for the placebo cohort were taken from a cohort of patients followed over 12 months in an Irish teaching hospital. The differences in probabilities of death and hospitalisation for the treatment cohort were taken from RALES. It was assumed that no difference in death and hospitalisation rates occurred between the cohorts after the 2-year mean duration of follow-up for RALES. Costs incorporated in the analysis were spironolactone treatment cost, hospitalisation cost for severe heart failure, and outpatient visit cost. A two-way sensitivity analysis varied probabilities of death and hospitalisation, and one-way sensitivity analyses varied the hospitalisation cost and added outpatient visits to the spironolactone cohort. Future costs and outcomes were discounted at 5% and 1.5% respectively. Table 5.7 presents the quality and applicability assessment of this economic analysis.

UK study assessing eplerenone

Results of the Duerden et al. (2008) cost-effectiveness analysis⁷⁹ are presented in Table 5.8. These results showed that adding eplerenone to optimal medical treatment in patients with heart failure and left ventricular systolic dysfunction post acute myocardial infarction is cost-effective in the UK. Limitations of this study were that the analysis used a short time horizon (3 years) to assess a long-term treatment for a chronic disease, the analysis did not estimate QALYs, and the sensitivity analysis did not vary resource use estimates.

Table 5.8

Results - Duerden 2008 economic analysis.

Irish study assessing spironolactone

Results of the cost-effectiveness analysis by Tilson et al. (2003)⁸⁰ are presented in Table 5.9. Considering a cost-effectiveness threshold of £13,000 per life-year gained, we concluded that adding spironolactone to optimal medical treatment is highly cost-effective in Ireland. Limitations of the study were that it did not incorporate quality of life, and the mean age of the population of patients in the RALES study was lower than in the Irish population of patient with chronic heart failure (65 vs 76 years).

Table 5.9

Results - Tilson 2003 economic analysis.

Relative value placed on the outcomes considered

The GDG reviewed the evidence of using aldosterone antagonists in the treatment of chronic heart failure. Two agents were assessed: Spironolactone and Eplerenone. The GDG noted that there was no direct comparison made between the two agents in the treatment of heart failure.

In the RALES study aldosterone antagonist spironolactone was added to loop diuretics and an ACEI in patients with moderate to severe chronic heart failure (NYHA Class III–IV) who remained symptomatic. The GDG noted the significant 30% reduction of both all cause mortality and heart failure hospitalisation at 24 months of therapy with spironolactone. This treatment also resulted in a significant rise in the incidence of gynaecomastia in males, with no significant rise in the risk of hyperkalaemia. However, subsequent observational evidence suggests that the rise in use of spironolactone following the publication of the RALES study was associated with a significant rise in the number of hospitalisations and mortality related to hyperkalaemia and renal failure in patients with chronic heart failure over the age of 66 years treated with ACEI and spironolactone. (Juurlink 2004)⁸². The GDG took this to highlight the importance of strict monitoring in such patients and of strict adherence to the inclusion and exclusion criteria used in the clinical trial.

In the EPHESUS study, use of the aldosterone antagonist eplerenone was tested in the treatment of symptomatic heart failure (LVEF<40%) after myocardial infarction, or in asymptomatic heart failure caused by left ventricular systolic dysfunction (LVEF<40%) after myocardial infarction in diabetic patients. Eplerenone was used in addition to conventional medical therapy (loop diuretics, ACEI/ARB, beta-blockers). The GDG debated whether the cohort of patients with heart failure after myocardial infarction could be considered as relevant to recommendations on the treatment of chronic heart failure. While the heart failure in this cohort had resulted from acute myocardial infarction, patients continued to display evidence of left ventricular systolic dysfunction (LVEF<40%, with symptoms unless diabetic) some 3–14 days after myocardial infarction and management continued beyond the acute phase of the infarction. Therefore, the GDG decided that the evidence from this group of trials was relevant. Eplerenone therapy resulted in 14%, and 20% reductions of all cause mortality and sudden death, respectively, at 16 months. The GDG noted the evidence from subgroup analysis of the same trial suggesting better outcomes when the agent is started in the first 7 days following the acute event (Adamopoulos 2009)⁸³. Not surprisingly, the impact of therapy was larger in the subgroup of patients with the more severe left ventricular systolic dysfunction (LVEF<30%). There was also a significant reduction of non-fatal heart failure hospitalisations at 16 months, for this group in a post-hoc analysis.

Quality of evidence

The GDG noted the greater weight to be given to results of pre-specified analyses of randomised controlled trials as opposed to post-hoc analyses of randomised controlled trials. The GDG felt it was not appropriate to combine the post-hoc analysis of the outcomes in the sub-group of patients with LVEF<30% treated with eplerenone ⁷², with the study of patients with LVEF<35% treated with spironolactone⁷⁴ in a meta-analysis since the two cohorts received different medical therapies and had different backgrounds.

The GDG looked at the small trials that assessed the impact of adding these agents in heart failure patients on quality of life, hyperkalaemia and renal failure. These results are superseded by the larger studies.

Trade-off between clinical benefits and harms

The general side effects of this class of drug are hyperkalaemia and renal impairment.

Since the initiation of aldosterone antagonists is a decision to be made by a specialist, and the decision whether to stop or reduce dose of aldosterone antagonists in the light of rises in serum creatinine and potassium or decline of eGFR is also to be made by a specialist, it is not appropriate to give detailed recommendations on how frequently to monitor renal function or when to stop these agents. The GDG recognised the value of the practical recommendations in the previous guideline, and were happy to support these, recognising that they would have a useful role in implementation of the guidance. In addition, the GDG accepts the NICE guidance on the diagnosis and management of Chronic Kidney Disease, recommending the addition of estimating GFR to the routine assessment and monitoring of renal function. Thus urea, electrolytes, creatinine and eGFR should be checked at 1 week, and at 1, 2, 3, and 6 months and 6 monthly thereafter. They also recommended that the aldosterone antagonist dose should be halved if the potassium rises to 5-5-5.9 mmol/l and stopped if the potassium rises above 6 mmol/l or the creatinine above 220 μmol/l. The latter is based on the evidence from the clinical trials of aldosterone antagonists in heart failure.

There are other side-effects that are pertinent to the non-selective aldosterone antagonist spironolactone, namely gynaecomstia and mastodynia

Trade-off between net health benefits and resource use

The GDG considered the health economic analysis⁷⁹ assessing eplerenone based on the EPHESUS trial⁷³. On a three-year time horizon, the incremental cost-effectiveness ratio (ICER) was less than £7000 per life-year gained, making the use of eplerenone in heart failure after myocardial infarction already treated with beta-blockers and ACEI, a cost-effective therapy.

In the cost-effectiveness study by Tilson et al, conducted from an Irish perspective and based on the RALES study⁸⁰, the use of spironolactone was also cost effective (ICER of £291 per life-year gained).

There is no comparative study between the two aldosterone antagonists. The GDG felt that the two agents are probably comparable. From a health economic point of view, the substantially lower cost of spironolactone compared to eplerenone was noted. The current evidence reviewed suggests that spironolactone should be used in severe chronic heart failure (NYHA Class III–IV), and eplerenone should be used in the patients with heart failure following myocardial infarction. The latter is in keeping with the guidance of NICE on the management of myocardial infarction complicated by heart failure.

The GDG are aware of two other trials: the EMPHASIS trial assessing the use of Eplerenone in mild heart failure (NYHA Class II), and the TOPCAT trial looking at the use of smaller doses of spironolactone in patients with heart failure and preserved left ventricular ejection fraction. The EMPHASIS trial was expected to complete recruitment in October 2011, however early termination in May 2010 is said by the sponsors to be due to superiority of eplerenone compared to placebo. The GDG did not have access to the data to analyse.

Another potential use of eplerenone might be where side effects specific to spironolactone (painful gynaecomastia) preclude the continuation of therapy.

The GDG agreed with the 2003 recommendation that a specialist should initiate spironolactone. The same applies to eplerenone.

The GDG suggested as a research recommendation a study investigating the best third agent in the treatment of heart failure, comparing AA vs. ARB in the treatment of heart failure patients who remain symptomatic after optimal therapy with ACEI and BB.

Reason for review

Since the publication of the guideline in 2003 new evidence in relation to ethnicity has emerged.

a. Isosorbide dinitrate/ hydralazine vs. placebo in addition to optimal medical management in the black population

Four studies (2 RCTs) were identified comparing isosorbide and hydralazine combination versus placebo in addition to optimal medical management in the black population with heart failure^84–87. In one RCT the patients self-identified as black (defined as of African descent)⁸⁵ and in one RCT the patients were defined as ‘black’ but no further details of ethnic origin were provided. Two of the studies reported on different outcomes from the main RCT study^86;84. The studies by Carson ⁸⁷ and Taylor ⁸⁴ are analysed separately to reflect the differences in the background medications the patients were receiving. Table 5.10 below presents a summary of the patient population, background medications and interventions for each study.

Table 5.10

Population and interventions for studies.

b. Isosorbide dinitrate plus hydralazine vs. ACE I in the black population

One RCT was identified comparing isosorbide dinitrate + hydralazine vs ACEI in the black population ⁸⁷.

c. Isosorbide dinitrate plus hydralazine vs. placebo in different age groups

One post hoc sub-group analysis of an RCT was identified comparing isosorbide dinitrate + hydralazine versus placebo in addition to optimal medical management in different age groups⁸⁸. Table 5.11 below summarises the patient population and intervention for this study.

Table 5.11

Population and interventions for RCT (Cohn et al.).

d. Isosorbide dinitrate plus hydralazine vs. ACE I in different age groups

One post hoc sub-group analysis of an RCT was identified comparing isosorbide dinitrate + hydralazine versus ACE I in different age groups⁸⁹. Table 5.12 below summarises the patient population and intervention for this study.

Table 5.12

Population and intervention RCT (Johnson et al.).

a. Isosorbide + hydralazine vs. placebo + optimal medical management in the black population

TAYLOR ⁸⁴

Compared with placebo (+optimal medical therapy), the combined isosorbide dinitrate plus hydralazine (+optimal medical therapy) has a significant reduction in:

• All cause mortality 0 to 18 months [moderate quality]
• Hospitalisation for heart failure mean 12.8 months [moderate quality]
• Cardiovascular death mean 10 months [moderate quality]

Compared with placebo (+optimal medical therapy), the combined isosorbide dinitrate plus hydralazine (+optimal medical therapy) has a significant improvement in:

• Composite score follow-up range 0 to 18 months [high quality]
• Quality of life [moderate quality]

Compared with placebo (+optimal medical therapy), the combined isosorbide dinitrate plus hydralazine (+optimal medical therapy) was associated with a:

• significant increase in headache [high quality] and dizziness [high quality]

Compared with placebo (+optimal medical therapy), the combined isosorbide dinitrate plus hydralazine (+optimal medical therapy) had no significant effect on:

• The number of unplanned emergency room admissions or unscheduled office visits [moderate quality]

The evidence profile below summarises the quality of the evidence and outcome data from 4 studies (2 RCTs) ^84–87 comparing isosorbide dinitrate + hydralazine versus placebo in addition to optimal medical management in the black population (patients self-identified as black: defined as of African descent). Two of the studies reported on different outcomes

Evidence profile

Isosorbide dinitrate + hydralazine (+ optimal medical management) versus placebo (+ optimal medical management) in the black population. Question: Should isosorbide dintrate and hydralazine (vs. placebo) be used in addition to optimal medical therapy (more...)

CARSON ⁸⁷

Compared with placebo, the combined isosorbide dinitrate plus hydralazine had no significant effect on:

• All cause mortality up to 66 months (5.5 yrs) [moderate quality]
• Hospitalisation for heart failure 66 months [moderate quality]

The evidence profile below summarises the quality of the evidence and outcome data from the one RCT⁸⁷ comparing isosorbide dinitrate + hydralazine versus placebo in the black population (patients self-identified as black: defined as of African descent).

Evidence profile

Question: Should Isosorbide + hydralazine be used vs placebo? Bibliography: Carson P, Ziesche S, Johnson G et al. Racial differences in response to therapy for heart failure: Analysis of the Vasodilator-Heart Failure Trials. Journal of Cardiac Failure. (more...)

b. Isosorbide dinitrate plus hydralazine vs. ACE I in the black population

Compared with ACEI, isosorbide dinitrate plus hydralazine had no significant effect on:

• All cause mortality follow-up 0 to 66 months [moderate quality]
• Hospitalisations for chronic heart failure follow-up 0 to 66 months [moderate quality]

The evidence profile below summarises the quality of the evidence and outcome data from 1 RCT ⁸⁷ comparing isosorbide + hydralazine versus ACE I in the black population.

Evidence profile

isosorbide dinitrate + hydralazine versus ACE I in the black population. Question: Should isosorbide dinitrate + hydralazine vs ACE I be used for chronic heart failure in black population? Bibliography: Carson P, Ziesche S, Johnson G et al. Racial differences (more...)

c. Isosorbide dinitrate plus hydralazine vs. placebo in different age groups

Compared with placebo, the post-hoc sub group analysis did not detect a significant difference for isosorbide plus hydralazine compared with placebo in the > 60 yrs or < 60 yrs for:

• all cause mortality [low quality]

The evidence profile below summarises the quality of the evidence and outcome data from 1 RCT (post-hoc sub group analysis) ⁸⁸ comparing isosorbide dinitrate + hydralazine versus placebo in different age groups. The table below summarises the patient population and intervention for this study.

Evidence profile

Isosorbide dinitrate + hydralazine versus placebo in different age groups. Question: Should isosorbide dinitrate + hydralazine vs placebo be used for chronic heart failure in different age groups? Bibliography: Cohn JN, Archibald DG, Francis GS. Veterans (more...)

d. Isosorbide plus hydralazine vs. ACE I in different age groups

The evidence profile below summarises the quality of the evidence and outcome data from 1 RCT (post-hoc subgroup analysis) ⁸⁹ comparing isosorbide dinitrate + hydralazine versus ACE I in different age groups. The table below summarises the patient population and intervention for this study.

Evidence profile

Comparing isosorbide dinitrate + hydralazine versus ACE I in different age groups. Author(s): Date: 2009-03-11

Compared with ACEI, the post-hoc sub group analysis did not detect a significant difference for isosorbide dinitrate plus hydralazine compared with ACEI in the over 60 yrs or < 60 yrs for:

• all cause mortality at 2 yrs [low quality]

Relative value placed on the outcomes considered

The GDG reviewed the statement from the 2003 guidelines concerning the use of the combination of Hydralazine and Isosorbide Dinitrate in heart failure, and felt that the 2003 conclusions were valid, even though they were based on a trial when the baseline therapy in 1986 was diuretics and digoxin only. The GDG noted that the main studies related to this subject, since the publication of the 2003 guidelines, were on the use of the combination in the black population, who were found to be less responsive than non-blacks to treatment with Angiotensin Converting Enzyme Inhibitors (ACEI).^85,84,86,87 Subsequent evidence of benefit for the combination of hydralazine and nitrates was found from subgroup analyses of prospective studies using the combination either against placebo or in comparison to ACEI. Evidence for the impact of the combination in black patients with moderate to severe heart failure (mainly NYHA III) came from the AHEFT study where the combination of hydralazine and nitrates was given in addition to optimal therapy that included ACEI/ARB, BB and Aldosterone antagonists. The GDG noted that adding the combination to optimal therapy (ACEI, BB and AA) in such patients reduced morbidity and mortality.

The response was related to the treatment with the combination rather than with one of the two drugs. It was felt that patients should be simultaneously commenced on both drugs, and that the doses should be increased gradually according to tolerance, aiming to achieve the target doses used in the clinical trials.

The GDG considered the use of the term ‘black’ as used in these studies. Black patients of African and Carribean descent have been found to derive less benefit than non-blacks from ACEI in both heart failure and hypertension trials, and it is this group in the UK to which this evidence is applicable.

Quality of evidence

The RCT evidence on the treatment of black people with heart failure with hydralazine and nitrate vs. placebo and vs ACEI was of moderate to high quality. The evidence for the age groups analysis was of low quality due to the inclusion of post-hoc subgroup analysis from RCT data ^88,89.

Trade-off between clinical benefits and harms

In a post-hoc analysis in blacks, the treatment of black people with heart failure with hydralazine and nitrate vs. placebo resulted in reduced morbidity and mortality, better quality of life but with more headache and dizziness. The comparison with ACEI was associated with wide confidence intervals. The GDG noted that the patients included in the AHEFT trial ⁸⁵ were already treated with ACEI, beta-blockers (BB), and aldosterone antagonists suggesting that earlier concerns about the safety of the combination in the presence of treatment with ACEI and BB could be allayed.

The GDG noted that the effect of the combination is not limited to an age group. The GDG also noted that side-effects could limit some patients’ tolerance of the treatment with the combination.

The GDG discussed the potential use of the combination in heart failure patients with renal dysfunction, in whom ACEI and ARB could not be used. The GDG noted the publication of the Chronic Kidney Disease Guideline No. 73 (2008) that gives recommendations on the management of patients with impaired renal function who may be on ACEI, ARB and/or aldosterone antagonists⁹¹.

There is no evidence on the use of this combination in non-black patients who remain symptomatic after treatment with ACEI and beta-blockers. In the absence of such evidence, one could consider adding these agents on the pathophysiological basis of the helpful vasodilatation offered by these agents in such patients. In addition to the lack of evidence in this regard, and to the potential for intolerance related to side-effects, the introduction of these agents requires the patient’s blood pressure to be adequate or raised. It may be that non-black hypertensive patients with heart failure who remain symptomatic after treatment with ACEI and beta-blockers and who could not have ARB or aldosterone antagonists could benefit from the introduction of this combination. The GDG noted that international guidelines (ESC/ACC/AHA) made such a recommendation but felt that, in the absence of firm evidence to support this, a research recommendation was more appropriate.

The addition of this combination should be initiated by a specialist.

Trade-off between net health benefits and resource use

The GDG noted that the health economic review suggested that the addition of this combination in black patients, who remain symptomatic of heart failure while on ACEI and beta-blockers, is cost saving over 18 months. It is likely to be cost-effective over the lifetime as long as the effects observed in trials continue for some months beyond the 18 month trial follow-up. The GDG noted that the cost-effectiveness analysis⁸⁶ was developed from a US perspective, so may be of limited applicability to the UK NHS. The GDG felt that the short time horizon was not a significant limitation given that life expectancy is short in patients who remain at NYHA class III (94% of the cohort) despite treatment with ACEI and beta-blockers.

Reasons for Review

New randomised clinical trials have reported on the use of ARBs in the treatment of heart failure due to left ventricular systolic dysfunction as an add-on to ACEI, in the treatment of heart failure due to left ventricular systolic dysfunction where ACEI are not tolerated, in heart failure with preserved left ventricular ejection fraction and in heart failure due to left ventricular systolic dysfunction caused by myocardial infarction. Some of the trials looking at similar populations produced different results. Thus, there is a need for a review and appraisal of the evidence.

Traditionally, when ACEI are not tolerated due to side effects (such as cough), an ARB is used. However, the question arises as to whether ARBs exert the same effect as ACEI. In addition, another question is whether all ARBs exert the same effect. Clarification is needed on the potential risks from combining ACEI, ARB and beta-blockers. Another issue is whether patients who remain symptomatic despite therapy with ACEI and beta blockers should be additionally treated with ARBs, aldosterone antagonists or the combination of hydralazine and nitrates.

Angiotensin-II receptor antagonists (ARBS) vs. placebo

a. ARBs vs. placebo in heart failure with left ventricular systolic dysfunction (LVSD)

Compared with placebo, angiotensin-II receptor antagonists had a significant reduction on:

• HF hospitalisation [moderate quality]
• Composite score (CV mortality and HF hospitalisation) [moderate quality]

Compared with placebo, angiotensin-II receptor antagonists significantly increased:

• Hyperkalaemia [moderate quality]
• Raised creatinine [moderate quality]
• Hypotension [moderate quality]

Compared with placebo, angiotensin-II receptor antagonists significantly improved:

• Quality of life scores (MLWHFQ) [moderate quality]

Compared with placebo, angiotensin-II receptor had a non-significant affect on:

• All cause mortality [high quality]
• All cause mortality post-hoc subgroup [low quality]
• Hypotension post-hoc subgroup [low quality]
• Mean increase in creatinine post-hoc subgroup [moderate quality]

Change in NYHA class was reported in one study ⁹⁵:

• Improved: placebo: 28/201 (14%); 4mg: 39/203 (19%); 8mg 41/202 (20%); 16mg: 34/201 (17%); Total on Candesartan: 114/606 (24%)
• No change: placebo: 170/201(85%); 4mg: 162/203 (80%); 8mg: 161/202 (80%); 16mg: 165/201 (82%); Total on Candesartan: 488/606 (81%)
• Deterioration: placebo: 3/210 (1%); 4mg: 2/203 (1%); 8mg: 0/202 (0%); 16mg: 2/201 (1%); Total on Candesartan: 4/606 (0.7%)

The evidence profile below summarises the quality of the evidence and outcome data from 5 studies ^92–96 comparing ARBs vs. placebo in heart failure with left ventricular systolic dysfunction (LVSD)

Evidence profile

ARBs vs. placebo in heart failure with left ventricular systolic dysfunction (LVSD). Question: Should angiotensin II receptor blockers (ARBs) vs. placebo be used for chronic heart failure? Bibliography: Matsumori A. Efficacy and safety of oral candesartan (more...)

b. ARBs vs. placebo in heart failure with preserved ejection fraction (HFPEF)

Compared with placebo, angiotensin-II receptor antagonists significantly increased:

• Hyperkalaemia [moderate quality]
• Raised creatinine [moderate quality]

Compared with placebo, angiotensin-II receptor antagonists had a non-significant effect on:

• All cause mortality [high quality]
• CV mortality [high quality]
• Hypotension [low quality] - however there was serious heterogeneity (I² 82%) seen when meta-analysing the results from I-PRESERVE and CHARM-preserved for this outcome. A possible cause for the inconsistency of results could be due to the use of the stronger drug candersartan in CHARM-preserved compared to irbesartan in I-PRESERVE.
• HF hospitalisation [high quality]
• Composite score (CV mortality and HF hospitalisation) [high quality]

Compared with placebo, angiotensin-II receptor antagonists made no difference to:

• Mean increase in creatinine [high quality]

The evidence profile below summarises the quality of the evidence and outcome data from 2 studies ^97,98 comparing ARBs vs. placebo in heart failure with preserved ejection fraction (HFPEF).

Evidence profile

ARBs vs. placebo in heart failure with preserved ejection fraction. Question: Should angiotensin II receptor blockers (ARBs) vs. Placebo be used for HFPEF? Bibliography: Massie BM, Carson PE, McMurray JJ et al. Irbesartan in patients with heart failure (more...)

Relative value placed on the outcomes considered

Compared to placebo, ARB did not reduce all cause mortality. However, treatment with ARB led to significant reduction in the rate of heart failure hospitalisation (CHARM-Alternative and the ARCH-J trials)^92,96. There was also a significant reduction of the composite end-point of cardiovascular mortality and heart failure hospitalisation (CHARM-alternative trial) ⁹².

Only one trial (Val-HeFT post-hoc analysis) ⁷⁷ showed an improved quality of life score, and another trial (STRETCH) ⁹⁵ showed an increased number of patients with improved NYHA functional class when treated with ARB.

Treatment with ARB resulted in significant increase in hyperkalaemia, hypotension and raised creatinine level.

The GDG were aware of two trials (ELITE II and OPTIMAAL) that provided direct comparison of ARB with ACEI in heart failure due to left ventricular systolic dysfunction^102,103. ELITE-II compared Losartan to Captopril in patients > 60 years with LVEF < 40% and found similar morbidity and mortality associated with treatment with either agent. OPTIMAAL compared Losartan to Captopril in patients with significant left ventricular systolic dysfunction following Q wave myocardial infarction and found a trend for reduced mortality in the captopril arm, and no difference in morbidity. All the placebo-controlled ACEI trials except CONSENSUS-II (which was in a unique early AMI phase using intravenous ACEI), have consistently shown reduction of morbidity and mortality in heart failure due to left ventricular systolic dysfunction, in contrast with the results of the placebo-controlled ARB trials. However, such indirect comparison can be misleading. The ARB trials were performed in a different era in heart failure patients with better prognosis as a result of treatment with other effective agents such as beta-blockers. Therefore, it will have been more difficult to demonstrate survival benefit in these studies. The more recent HEAAL study¹⁰⁴, which was published after the cut off date for the literature searches for this guideline, did find reduced mortality and heart failure hospitalisation in people on high dose (150 mg/day) losartan as compared to low dose (50 mg/day) losartan in the treatment of heart failure due to LVSD in patients intolerant of ACEI (85% due to cough). However, the higher dose was associated with increased renal complications and hyperkalaemia.

The GDG explored the current practice of readily switching patients with heart failure with LVSD from ACEI to ARB whenever side-effects are encountered. Intractable dry cough is the only side-effect that remains unique to ACEI and is readily relieved by switching treatment to ARB. The GDG felt that in light of the the stronger evidence base (and lower cost) of ACEI, treatment should only be switched when ACEI are not tolerated.

Angio-oedema reflects true intolerance to ACEI. It can, however, occur with ARB, albeit much less frequently. The occurrence of renal impairment, hypotension or hyperkalaemia while on ACEI should initially call for reduction (when significant) of the dose of ACEI rather than an immediate switch to ARB (see Appendix J on practical recommendations). The GDG advises that every attempt is made not to stop ACEI in the presence of side-effects, and that education is provided for patient and carers. The GDG noted that some of the patients recruited into the CHARM-Alternative trial had hypotension, hyperkalaemia or renal impairment as the reason for stopping ACEI, and that many were able to tolerate the ARB candesartan. However, candesartan itself led to significantly more patients than placebo discontinuing the study medication due to hypotension, hyperkalaemia and renal impairment.

The GDG considered the impact of treatment of heart failure associated with preserved left ventricular ejection fraction, with ARB.

Two large randomised controlled trials were reviewed.^97,98 CHARM-Preserved and I-PRESERVE. ARB had no impact in this group of patients on all cause mortality, cardiovascular mortality, heart failure hospitalisation and the composite score of cardiovascular mortality and heart failure hospitalisation. These agents did not significantly cause hypotension resulting in symptoms or in withdrawal from the trial. However, they significantly increased the incidence of hyperkalaemia and the number of patients with raised serum creatinine (though not the mean creatinine level between the placebo and the ARB treated groups). Taken alone, CHARM-Preserved trial showed a reduction of hospitalisation, but not when combined with I-PRESERVE in meta-analysis.

Quality of evidence

In trials looking at the impact of ARB therapy on patients with heart failure and reduced left ventricular ejection fraction, the evidence of lack of effect on all cause mortality was of high quality in all the trials, except the Val-HeFT-post-hoc analysis⁹³, where the evidence on the effect of all cause mortality was of low quality.

Moderate quality evidence was observed for the significant reduction in heart failure hospitalisation and in the composite score of cardiovascular mortality and heart failure hospitalisation.

Moderate quality of evidence from single trials showed that ARB therapy in these patients leads to improved quality of life scores, and increased number of patients with improved NYHA functional class. Similarly, moderate quality of evidence was observed for ARB therapy increasing the rates of hyperkalaemia, hypotension and raised creatinine level.

The appraisal of trials looking at the impact of ARB on patients with heart failure and preserved left ventricular ejection fraction produced high-quality evidence that these agents have no impact on: all cause mortality, cardiovascular mortality, composite score of cardiovascular mortality and heart failure hospitalisation or on the mean increase in serum creatinine. However, moderate quality evidence was observed for the ARB therapy resulting in a significant rise in the number of patients with raised creatinine, and in the significant increase in the incidence of hyperkalaemia.

Trade-off between clinical benefits and harms

The use of ARB is not justifiable in patients with heart failure and preserved left ventricular ejection fraction as there is no evidence of benefit, with evidence of potential harmful side effects (hyperkalaemia and raised creatinine level).

An ARB could be prescribed to patients with heart failure and preserved left ventricular ejection fraction if there is another indication to prescribe them, such as systemic hypertension or diabetes mellitus.

In patients with heart failure and left ventricular systolic dysfunction, the use of ARB is helpful in reducing hospitalisation, improving quality of life and improving heart failure functional class. There is also evidence from some trials of a reduction in the combined endpoint of mortality and hospitalisation. However, treatment with these agents requires frequent monitoring of serum urea, electrolytes, creatinine and eGFR to guard against the potential side effects of the drugs.

Trade-off between net health benefits and resource use

The use of ARB in patients with heart failure and left ventricular systolic dysfunction was found to be cost effective, however the GDG noted the broad confidence interval of the results of the cost-effectiveness analysis of the CHARM-Alternative trial ¹⁰⁰. The breadth of the confidence interval reflects the uncertainty around the mortality reduction.

For the cost-effectiveness of ARBs in patients with heart failure and preserved ejection fraction, the GDG agreed that the evidence is not clear or conclusive in this population.

a. Angiotensin-II receptor antagonists (ARBs) plus Angiotensin Converting Enzyme Inhibitors (ACEI) in comparison to ACEI plus placebo

Two studies were identified comparing ARB plus ACEI with Placebo plus ACEI (Houghton et al; Krum et al)

b. ARBs + ACEI + betablockers (BB) vs placebo + ACEI + BB in the medical management of adults with chronic heart failure?

c. ARBs + ACEI + betablockers (BB) vs placebo + ACEI + BB in the medical management of adults with heart failure post myocardial infarction

The VALIANT trial ¹⁰⁵ was designed differently to the trials used in patients with chronic heart failure (see above). Patients were not on a background of ACEI but were randomised to ARB + ACEI vs ACEI vs ARB, and most patients were on a background of beta blockers.

a. Angiotensin-II receptor antagonists (ARBS) plus Angiotensin Converting Enzyme Inhibitors (ACEI) in comparison to ACEI plus placebo in chronic heart failure

Compared with ACEI + placebo, ARBs + ACEI significantly reduced:

• First hospitalisation [low quality]

Compared with ACEI + placebo, ARBs + ACEI significantly improved:

• QoL (MLHQ) [moderate]

Compared with ACEI + placebo, ARBs + ACEI significantly increased:

• Hyperkalaemia [high quality]

Compared with ACEI + placebo, ARBs + ACEI had no difference on:

• Mortality [moderate quality]
• Increased serum creatinine (μmol/L) [low quality]

The evidence profile below summarises the quality of the evidence and outcome data from 2 studies^106,107. Krum 2004 was a subgroup analysis of the Val-HeFT RCT. Both studies compared ARBs + ACEI vs. ACEI + placebo in heart failure with left ventricular systolic dysfunction (LVSD). Patients in both arms in both studies were not on a background of BB.

Evidence Profile

ARBs + ACEI vs. ACEI + placebo in heart failure with left ventricular systolic dysfunction (LVSD). Question: Should ARB + ACEI (no BB) vs Placebo + ACEI (no BB) be used for CHF? Bibliography: H. Krum, P. Carson, C. Farsang, A. P. Maggioni, R. D. Glazer, (more...)

b. ARB + ACEI + BB vs placebo + ACEI + BB in patients with chronic heart failure

Compared with placebo + ACEI + BB, ARB + ACEI + BB had a significant reduction on:

• HF hospitalisation [moderate quality]
• Composite score (CV mortality and HF hospitalisation) [high quality]

Compared with placebo + ACEI + BB, ARB + ACEI + BB had a significantly fewer number of cases with:

• Worsened NYHA class [low quality]

Compared with placebo + ACEI + BB, ARB + ACEI + BB had no significant effect on:

• All cause mortality [moderate quality]
• Improved NYHA class [low quality]
• Unchanged NYHA class [low quality]

Compared with placebo + ACEI + BB, ARB + ACEI + BB were significantly worse for:

• Hypotension [moderate quality]
• Hyperkalaemia [high quality]
• Increased serum creatinine (number of patients) [high quality]

NYHA class

The results of one study that could not be incorporated into the meta-analysis showed¹⁰⁸:

• Patient NYHA class II
• Candesartan 13.8% got worse vs 23.8% improved
• Placebo 20.8 got worse vs 18.7% improved
• NYHA III–IV
• Candsartan 4.2% got worse vs 45.7% improved
• Placebo 5.5% got worse vs 45.8% improved

The evidence profile below summarises the quality of the evidence and outcome data from three studies ^101,109,110 comparing ARBs + ACEI + BB vs. placebo + ACEI + BB in heart failure with reduced left ventricular ejection fraction (LVEF).

Evidence profile

Question: Should ARB + ACEI + BB vs Placebo + ACEI + BB be used for CHF? Bibliography: McMurray et al Effects of candesartan in patients with chronic heart failure and reduced left ventricular systolic function taking angiotensin-converting-enzyme inhibitors: (more...)

Evidence Profile: ARBs + ACEI + BB vs. placebo + ACEI + BB in heart failure with left ventricular systolic dysfunction (LVSD)

c. ARB + ACEI + BB vs placebo + ACEI + BB in chronic heart failure post myocardial infacrtion

The evidence profile below summarises the quality of the evidence and outcome data from one study ¹⁰⁵ comparing ARBs + ACEI + BB vs. placebo + ACEI + BB in post-MI patients with heart failure with reduced left ventricular ejection fraction (LVEF).

Evidence profile

Question: Should ARB + ACEI + BB vs Placebo + ACEI + BB be used for post-MI and CHF? Bibliography:. M. A. Pfeffer, J. J. McMurray, E. J. Velazquez, J. L. Rouleau, L. Kober, A. P. Maggioni, S. D. Solomon, K. Swedberg, Werf F. Van de, H. White, J. D. Leimberger, (more...)

Compared with placebo + ACEI + BB, ARB + ACEI + BB had a significant reduction on:

• HF hospitalisation [high quality]

Compared with placebo + ACEI + BB, ARB + ACEI + BB had no difference on:

• All cause mortality [high quality]
• Hyperkalaemia [moderate quality]

Compared with placebo + ACEI + BB, ARB + ACEI + BB were significantly worse for:

• Hypotension [high quality]

Relative value placed on the outcomes considered

The question was considered in two stages: adding ARB to the combination of ACEI and beta- blockers and combining ARB with ACEI.

For the first part, there were four appraised studies. Three of the studies were of similar design adding candesartan¹⁰¹ (CHARM-Added study), or Valsartan^109,110 (Val-HeFT, Cocco et al) to treatment with an ACEI that was given to 92–100% of participants. In addition, beta-blockers were given to 100% of the patients in the Cocco et al study, 55% in CHARM-Added and 35% in the Val-HeFT study. The fourth study¹⁰⁵ (VALIANT) was in patients with heart failure due to LVSD following myocardial infarction. The design was more complex in that there were three arms in the study: ACEI, Valsartan or ACEI + Valsartan. In the VALIANT study 77% of the patients were on beta-blockers.

The addition of ARB to the combined ACEI and BB in patients with heart failure and LVSD did not affect all cause mortality but did significantly reduce heart failure hospitalisation, and the combined score of heart failure hospitalisation and mortality.

This intervention led to significantly less chance of worsening NYHA functional class. Adding ARB to this combination significantly increased the incidence of hyperkalaemia, hypotension and raised serum creatinine.

There was some concern raised after the publication of the Val-HeFT study ¹¹⁰ about the safety of combining ARB with beta- blockers in patients with heart failure. This led to a safety warning in the 2003 NICE guidelines on heart failure. However, given the results of the other studies that used both Candesartan ¹⁰¹ (CHARM-Added) and Valsartan ¹⁰⁵ (VALIANT), the GDG concluded this combination could be used safely.

The second part of the question addressed combining ARB with ACEI. Two studies were appraised: Krum et al (sub-study of Val-HeFT trial)¹⁰⁶, and Houghton et al¹⁰⁷. These used Valsartan and Losartan, respectively.

Compared to placebo, the addition of Valsartan to ACEI in the Krum et al trial¹⁰⁶ did not impact on all cause mortality, but it significantly reduced the rate of first hospitalisation. This addition also resulted in significant improvement in the quality of life. There was no significant impact of adding Losartan in the Houghton et al study¹⁰⁷ on the incidence of hyperkalaemia or increased serum creatinine.

Quality of evidence

There is high-quality evidence that adding ARB to the combination of ACEI and beta-blockers results in significantly reduced combined score of cardiovascular mortality and heart failure hospitalisation; and for increased risk of hyperkalaemia.

With regards to the impact of this addition on all cause mortality, heart failure hospitalisation, hypotension, and the number of patients with raised serum creatinine, the evidence is of moderate quality.

The evidence supporting the remainder of the statements was of low quality. The evidence behind the statements derived from the Houghton et al study¹⁰⁷ of the addition of ARB to ACEI was of moderate quality. The main statements derived from the results of the Krum study¹⁰⁶ were based on low quality evidence. The latter is particularly related to the fact that this study was a post-hoc analysis.

Trade-off between clinical benefits and harms

The addition of ARB to other drugs for heart failure with LVSD did not reduce all cause mortality, but the trials were not powered to detect such an effect. However, another analysis (Young et al) from the CHARM programme combined the results of CHARM-Added and CHARM-Alternative. This was powered to look at the impact of ARB on mortality in heart failure patients with reduced left ventricular ejection fraction. It showed a statistically significant reduction of all cause mortality and cardiovascular mortality. This was in addition to the significant reduction of heart failure hospitalisation.

ARBs reduce the combined score of cardiovascular mortality and heart failure hospitalisation, as well as reducing the rate of hospitalisation and improving quality of life score. Against these benefits are the potential risks of hyperkalaemia, hypotension and raised serum creatinine. The latter three potential harms call for frequent checks to be made on the renal profile and the electrolyte balance when patients are given these agents. These harms have also to be considered when prescribing these agents to heart failure patients with significant renal dysfunction or borderline low systolic blood pressure. Further details regarding the issue of monitoring and adjusting the doses of ARB are in Appendix J. The GDG considered whether some patients with heart failure and LVSD might be prescribed ACEI, beta-blockers, ARB and an aldosterone antagonist. Although some patients in the CHARM-Added trial were on quadruple therapy, these were the minority. The GDG does not believe there is sufficient evidence to support the widespread use of quadruple therapy. Similarly, even in the absence of beta-blockers the GDG does not recommend using triple therapy of ACEI with ARB and aldosterone antagonists for safety concerns (risks of hyperkalaemia and renal impairment). A similar view was adopted by the Chronic Kidney Disease NICE guidance where such combination of ACEI/ACEI and AA was discouraged.

Trade-off between net health benefits and resource use

The use of ARB in patients with heart failure and left ventricular systolic dysfunction added to ACEI and beta-blockers was found to be cost-saving in the reviewed cost-effectiveness analysis based on CHARM-Added¹⁰⁰.

A confidence interval was not reported with this result. The all-cause mortality reduction in the CHARM-Added trial, although not statistically significant, was larger than that recorded in our meta-analysis when the study was combined with other trials (RR=0.91 vs RR=0.98) (Section 5.2.5.4). Had the meta-analysis been used ARBs might not appear cost-effective.

First-line treatment

See also recommendations R30 – R34 on the use of ACE inhibitors and beta-blockers for first-line treatment. See recommendations R39 – R40 for alternative first-line treatments for patients who are intolerant of ACE inhibitors. See recommendation R38 for alternative first-line treatments for patients who are intolerant of ACE inhibitors and ARBs.

R27.

Offer both angiotensin-converting enzyme (ACE) inhibitors and beta-blockers licensed for heart failure to all patients with heart failure due to left ventricular systolic dysfunction. Use clinical judgement when deciding which drug to start first. [new 2010] KPI

Second-line treatment

See also recommendations R35 – R37 and R40 on second-line treatments.

R28.

Seek specialist advice before offering second-line treatment to patients with heart failure due to left ventricular systolic dysfunction. [new 2010]

R29.

Seek specialist advice and consider adding one of the following if a patient remains symptomatic despite optimal therapy with an ACE inhibitor and a beta-blocker:

• an aldosterone antagonist licensed for heart failure (especially if the patient has moderate to severe heart failure [NYHA¹⁴ class III–IV], or has had an MI within the past month) or
• an angiotensin II receptor antagonist (ARB) licensed for heart failure¹⁵ (especially if the patient has mild to moderate heart failure [NYHA class II–III]) or
• hydralazine in combination with nitrate (especially if the patient is of African or Caribbean origin¹⁶ and has moderate to severe heart failure [NYHA class III–IV]). [new 2010] KPI

Diuretics

R42.

Diuretics should be routinely used for the relief of congestive symptoms and fluid retention in patients with heart failure, and titrated (up and down) according to need following the initiation of subsequent heart failure therapies. [2003]

R43.

The diagnosis and treatment of heart failure with preserved ejection fraction should be made by a specialist, and other conditions that present in a similar way may need to be considered. Patients in whom this diagnosis has been made should usually be treated with a low to medium dose of loop diuretics (for example, less than 80 mg furosemide per day). Patients who do not respond to this treatment will require further specialist advice. [2003]

Calcium channel blockers

R44.

Amlodipine should be considered for the treatment of comorbid hypertension and/or angina in patients with heart failure, but verapamil, diltiazem or short-acting dihydropyridine agents should be avoided. [2003]

Amiodarone

R45.

The decision to prescribe amiodarone should be made in consultation with a specialist. [2003]

R46.

The need to continue the amiodarone prescription should be reviewed regularly. [2003]

R47.

Patients taking amiodarone should have a routine 6-monthly clinical review, including liver and thyroid function test, and including a review of side effects. [2003]

Anticoagulants²³

R48.

In patients with heart failure in sinus rhythm, anticoagulation should be considered for those with a history of thromboembolism, left ventricular aneurysm, or intracardiac thrombus. [2003]

Aspirin

R49.

Aspirin (75–150 mg once daily) should be prescribed for patients with the combination of heart failure and atherosclerotic arterial disease (including coronary heart disease). [2003]

Inotropic agents

R50.

Intravenous inotropic agents (such as dobutamine, milrinone or enoximone) should only be considered for the short-term treatment of acute decompensation of chronic heart failure. This will require specialist advice. [2003]

Heart failure due to valve disease

R51.

Patients with heart failure due to valve disease should be referred for specialist assessment and advice regarding follow-up. [2003]

R52.

ACE inhibitor therapy should not be initiated in a patient with a clinical suspicion of haemodynamically significant valve disease, until the valve disease has been assessed by a specialist. [2003]

Age

R53.

The management of heart failure should be determined by clinical criteria, irrespective of the age of the patient. [2003]

R54.

Tolerance of drugs may be lower and side effects require closer and more frequent monitoring in older patients. [2003]

Gender

R55.

The principles of pharmacological management of heart failure should be the same for men and women. [2003]

R56.

In women of reproductive age who have heart failure, contraception and pregnancy should be discussed. If pregnancy is being considered or occurs, specialist advice should be sought. Subsequently, specialist care should be shared between the cardiologist and obstetrician. [2003]

R57.

The potential teratogenic effects of drugs should be considered. [2003]

Comorbidities

R58.

Manage co morbidities according to:

• ‘Hypertension’, NICE clinical guideline 34
• ‘MI: secondary prevention’, NICE clinical guideline 48
• ‘Type 2 diabetes’, NICE clinical guideline 87

and other relevant NICE guidance. This is particularly important in heart failure with preserved ejection fraction. [new 2010]