NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Chronic Conditions (UK). Chronic Heart Failure: National Clinical Guideline for Diagnosis and Management in Primary and Secondary Care. London: Royal College of Physicians (UK); 2003. (NICE Clinical Guidelines, No. 5.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Chronic Heart Failure

Chronic Heart Failure: National Clinical Guideline for Diagnosis and Management in Primary and Secondary Care.

Show details

Appendix DHealth economics of pharmacological therapies

Comparing the results of cost-effectiveness studies

There are numerous difficulties in comparing the results of studies that employ heterogeneous data and methods. All of the studies reported here are reasonably explicit about the data and methods used, and this eases comparisons. Studies with insufficient detail were omitted from consideration.

In relation to the clinical data the trials vary greatly in the characteristics of the patients studied in terms of: age, sex, symptoms of heart failure, duration and degree of heart failure, association with myocardial infarction, and in the end points of hospitalisation, death and progression to heart failure.

Economic analyses vary for a number of reasons including:

Source of cost and benefit data While efficacy data is usually taken from the named trials, the sources of resource use (and cost) data vary; some are prospective trial data and some are collected retrospectively. Either method can be collected for the whole trial sample or a sub-sample. In some analyses resource use data from an observational source is combined with trial data on effectiveness. In all cases the validity of comparisons from different countries is also an issue.

Which costs and benefits are included? The studies reported all take the health service of the country in question as the relevant perspective. Therefore they focus on direct costs (largely medication and hospitalisation) and direct benefits (life years gained).

Period of analysis This can be restricted to the trial period or extended beyond the trial.

If analysis is extended beyond the trial period, costs and benefits can be projected in a number of ways. This will usually involve a decision analytic model that follows a hypothetical cohort of patients (usually with characteristics similar to the trial sample) through time. Various assumptions are necessary about disease progression, resource use and the flow of benefits over time. These are subject to varying degrees of uncertainty and a good economic analysis will include sensible sensitivity analyses to reflect this uncertainty.

Location of study

While much of the effectiveness data on pharmacological therapies is obtained from multinational trials, the economic analyses usually adopt the perspective of one particular healthcare system. The costs of providing an intervention are specific to the healthcare system in question and non-UK results may be of limited use in informing this guideline. In particular different healthcare systems will have different policies and practice on length of stay, the criteria for hospitalisation, outpatient arrangements, the relative distribution of care between primary and secondary services etc. Nevertheless, all good quality cost effectiveness studies are reported here for completeness and to inform discussion. It is important to note that where cost figures have been converted into sterling, this is for guidance only. The figures are not directly comparable due to the different years in which the source data was collected.

Discounting of costs and benefits

When costs and benefits accrue over a number of years these are usually discounted to convert future values into current values. The base case figures in the tables usually report figures for costs discounted at 5% or 6%. Some ‘best’ and ‘worst’ figures include varying assumptions relating to the discounting of costs and benefits. In all of the studies reported, variations in the discount rate make no substantive difference to the results presented.

General economic considerations

Potential cost savings

These are likely to occur if a treatment improves QoL and slows disease progression

The magnitude of direct costs (and hence potential savings) will depend on:

  • the morbidity burden relating to the condition,
  • the specific costs related to this morbidity (for example, GP visits are relative cheap and hospitalisation is relatively expensive),
  • the efficacy of treatment in relation to morbidity (for example, how big the impact is on hospitalisation).

The slowing of the progression of CHF from mild (NYHA I and II) to more severe states (NYHA III and IV) will reduce all forms of healthcare utilisation including hospitalisation.

Potential additional costs

Direct costs of the treatment itself (drugs and related, eg monitoring, treating adverse effects).

If the treatment extends life, survivors will continue to utilise healthcare resources.

The studies reviewed here concentrate largely on direct costs (those arising from provision of treatment). Indirect costs (loss of earnings and pension payments for those who survive) are not likely to be as great as in some other chronic conditions since CHF is concentrated in older people. Intangible costs, that reflect the drawbacks of CHF such as pain, depression and reduced QoL, are very difficult to quantify. If indirect and intangible costs were included in the economic analyses of pharmacological therapies such as ACE inhibitors and beta-blockers they would be likely to produce even more favourable cost effectiveness ratios.

Increasing the use of effective pharmacological therapies in eligible patients will increase treatment costs but some, or all, of these will be offset, largely by reduced hospitalisation. If the savings are not large enough to offset the increase in cost, the net increase in costs has to be considered against benefits such as life years gained (LYG), or quality adjusted life years (QALYs).

The focus here is on LYG, as there is little data on QoL in CHF that can be used in economic evaluation (for example SF-36 or EQ-5D scores) to apply a quality weighting to LYG. The years of life saved are still associated with significant disability. This is important when comparing the cost-effectiveness of beta-blockers and ACE inhibitors with treatments, like statins, that prevent the transition from full health to disease.

1. Ace inhibitors


Evidence shows that ACE inhibitors in CHF reduce hospitalisation by slowing disease progression, improving quality of life (QoL) and extending length of life.

ACE inhibitors appear to be cost effective. They can be cost saving and have very favourable cost effectiveness ratios even when conservative assumptions are employed.

ACE inhibitors are under prescribed and doses in clinical practice are lower than in trials and therefore may be too low to achieve optimum benefits in terms of cost effectiveness.273–275

A recent survey273 suggests that while more than 90% of doctors believe there is strong evidence for mortality benefits of ACE inhibitors in CHF, only 47–62% of doctors actually prescribe them. The recent BHF Heart Failure Supplement1 reports that around 53% of patients with heart failure receive ACE inhibitors.

Under prescribing may be due to (i) an exaggerated perception of risk, especially renal failure and hypotension;276 (ii) the incorrect perception that lower doses retain mortality benefits while reducing risk of side effects277 (the ATLAS study shows this to be false; with higher doses of lisinoprol being more effective and cost effective than lower doses);278 (iii) diagnostic uncertainty.

Guidelines may have important role in encouraging the use of ACE inhibitors in primary care.


The clinical evidence in this paper has been taken primarily from two systematic reviews70,279 and the summaries of trials provided in the economics literature.

The economic analysis is largely based on three trials: SOLVD, SAVE and AIRE.

Due to time constraints the use of ACE inhibitors across different patient groups has not been specifically considered. There appears to be little evidence on this area.

Existing guidance that has considered economic implications

  1. Trent Working Group on Acute Purchasing guidance note for purchasers on ACE inhibitors in heart failure.280 This report took the cost effectiveness of ACE inhibitors as given and focused on the effects of increased ACEI use on hospitalisation. It concluded that the effects on reduced hospitalisation mean that the health gain associated with the use of ACE inhibitors can be achieved with potential resource savings.
  2. Inherited NICE guideline on Prophylaxis for patients who have experienced a MI.281 This considers the use of ACE inhibitors in patients with prior MI and CHF. The economic analysis used a ‘profiling’ approach to list the potential costs and consequences of ACEI use. The relevant recommendation for patients with prior MI and CHF is that ‘all patients should be offered long term treatment with an ACEI and then a beta-blocker … All of these treatments are cost effective.’ (A)
  3. North of England Evidence Based Development Project (1997) guideline on ACE inhibitors in primary care management of adults with symptomatic heart failure.282 The economic analysis used a ‘profiling’ approach to list the potential costs and consequences of ACEI use. The relevant recommendations are:
    • All patients with symptomatic heart failure and evidence of impaired left ventricular function should be treated with an ACE inhibitor. (A)
    • Patients with recent myocardial infarction and evidence of left ventricular dysfunction should be treated with an ACE inhibitor. (A)
    • Treatment of heart failure with ACE inhibitors is cost-effective. (A)
    • Statement: ACE inhibitors appear to be a cost effective use of resources when compared with other common health service interventions. (III)

Drug costs

Eight ACE inhibitors are licensed for the treatment of CHF in the UK. There are large differences in prices across drugs and between branded and generic versions.

There does not seem to be any evidence to suggest that one ACEI is better than another;279 although most clinical evidence comes from trials of enalapril. In addition a recent study by Schneeweiss et al 283 found no adverse effects resulting from the introduction of reference pricing for ACE inhibitors in British Columbia, which encourages switching from more expensive to cheaper drugs.

Existing evidence of the cost effectiveness of ace inhibitors

Fifty-two papers considering the economics of ACE inhibitors in the treatment of heart failure were identified by the economic review. Of these, only 12 were cost-effectiveness analyses of sufficient quality to report in the evidence tables; three were UK studies (Table D1) and the remainder were from other countries (Table D2).

Table D1. Existing evidence on cost-effectiveness of ACE inhibitors in the treatment of CHF in the UK.

Table D1

Existing evidence on cost-effectiveness of ACE inhibitors in the treatment of CHF in the UK.

Table D2. Existing international evidence on cost-effectiveness of ACE inhibitors in the treatment of CHF.

Table D2

Existing international evidence on cost-effectiveness of ACE inhibitors in the treatment of CHF.

The weight of evidence of the cost effectiveness of ACEIs

The cost effectiveness results cover a broad range. One UK study shows ACE inhibitors in the treatment of CHF to be cost saving in the base case scenario (as do four of the international studies); in other words ACE inhibitors represent a dominant technology in cost-effectiveness terms producing greater benefits that the comparator (usual care) at lower cost.

The majority of cost effectiveness ratios calculated on UK data are well within the boundaries of what would normally be considered a cost effective treatment.

The one very high (non-UK) C/QALY figure (£149,000)271 reflects very conservative assumptions about benefits and drug costs and is the figure estimated for patients in the youngest age group, who benefit the least from treatment (as shown in the SAVE trial).

The evidence therefore suggests that ACE inhibitors are cost effective in the treatment of CHF. However it should be noted that evidence from the UK context is limited.

Hospitalisation in more detail

Hospitalisation accounts for 60–70% of all expenditure on CHF in the UK.8,284,285

Of the major ACE inhibitor trials seven have included data on hospitalisation for CHF. These trials consistently show a reduction in admissions to hospital for progressive heart disease in patients taking ACE inhibitors. On average a GP would expect four patients with CHF to be admitted to hospital each year. The SOLVD results suggest that one of these admissions might be prevented (or delayed) through the use of ACE inhibitors.

Since hospitalisation is most frequent in the latter stages of the disease, treatment with ACE inhibitors may simply postpone hospitalisation, beyond trial follow-up (maximum four years). McMurray and Davie284 argue that this is probably not a justifiable concern since ACE inhibitors extend life for a relatively short period (eg 0.16 LYs in SOLVD Treatment Trial). Also the economic analysis of the SOVLD data shows no evidence of ‘catch-up’ costs in the later stages of the trial period.286

While HF hospitalisations decrease, it may be that other causes of morbidity mean that overall hospitalisation rates are not reduced. This does not seem to be reflected in the trial data, with ACE inhibitors appearing to reduce (but not significantly) other-cause hospitalisation in symptomatic patients.

Will the reductions in hospitalisation found in trials convert to clinical practice? The admission rate in the SOLVD control arm matches that found in general practice in England. Also McMurray et al 285 compare hospitalisation rates from official data sources, primary care surveys hospital survey and clinical trials and find them broadly in agreement.

Reductions in hospitalisation will not automatically translate to financial savings. However, with 90% occupancy Cornell et al 280 estimate that reductions in hospitalisation amount to around 13.3 beds (about half an average ward) in a typical district.

Other concerns about the generalisability of trial data

The trials are all ‘intention to treat’ analyses, therefore the calculations do take into account those who withdraw due to adverse effects or intolerance. However they do not take into account those excluded at the screening stage.

Patients tend to be younger in trials (early 60s). CHF is more common in older patients who have shorter life expectancy and are more susceptible to side effects.

Doses in clinical practice tend to be lower than in trials.287–290 There are estimates that only 14–22% of patients in clinical practice met the target doses suggested by the large trials.

Side effects and risks

Evidence suggests that ACE inhibitors are low in side effects with few people unable to tolerate them, including the elderly though doses may need to be reduced. Less than 2% of patients in SOLVD (both treatment and prevention trials) had side effects severe enough to stop treatment.

Increased use of ACE inhibitors may lead to increased incidence of renal failure in the elderly resulting from combination therapy of ACE inhibitors with NSAIDS. This could have considerable resource consequences.

Different groups of patients

This has not been explicitly considered. Evidence from the studies reviewed here suggests:

  • Reduction in mortality is greater in higher risk groups. In severe CHF, treating six patients for one year would prevent one death, whereas in less severe disease it would be necessary to treat 62 patients.291 Even in patients with less severe disease it is only necessary to treat nine patients for one year to prevent one admission to hospital.
  • Analysis of the SAVE data showed that ACE inhibitors can be cost effective as a preventative measure in post acute MI, ie selective use of a prevention strategy in a high-risk asymptomatic group.269,271
  • Analysis of SAVE data also showed that in this group, cost effectiveness was greatest in the older age groups, and the CER may be high (> £30k) for patients of 50yrs at MI.271

Trent WGAP report: Guidance note for purchasers on ace inhibitors in heart failure

This report is concerned specifically with the economics of the use ACE inhibitors for the treatment CHF in a ‘typical’ English health district with a population of 500,000. It starts from the premise that ACE inhibitors are cost effective, and focuses on modelling potential cost savings in the form of reduced hospitalisation.

Key assumptions of the model:

a simple (non-age specific) prevalence rate for the whole population

a proportion of these patients are assumed to be prescribed ACE inhibitors already

of the remaining population, a proportion will be assumed to be unsuitable for treatment with ACE inhibitors, due to contraindications

the identified marginal population are the CHF patients who are suitable for treatment with ACE inhibitors, but who currently do not take the drug

the cost of treating these patients with ACE inhibitors is estimated by estimating the annual drug treatment cost and making assumptions about the additional number of GP consultations required to undertake U&E tests at initiation and during monitoring

each patient to be considered for treatment with an ACE inhibitor is also assumed to require an echocardiogram to confirm the diagnosis of heart failure

the model requires an assumption about the proportion of heart failure patients who require hospitalisation for ACE initiation

the model assumes that cost savings result only from reduced demand for secondary sector in-patient hospitalisation. Any reduced demand for GP consultations, outpatient visits, non-ACE drug costs etc, is excluded

in this respect, the model will underestimate the potential cost savings of ACE inhibitor treatment for the given population

the model examines only the ‘year one’ costs and savings, which are expected to be higher than in subsequent years.


The model parameters, assumptions and results for the base case scenario are described in Table D3. This suggests that in the base case, extending the use of ACE inhibitors could result in net cost savings of £418,762. Using the evidence on effectiveness from SOLVD, for a population of 500,000, an additional 42 premature deaths could be prevented per annum.

Table D3. Central scenario from WGAP report and updated information.

Table D3

Central scenario from WGAP report and updated information.

The right hand column reports more recent evidence enabling an updating of some of the parameters. The figures in bold denote changes to the original estimates, and the results in the right hand column are produced when the model is run with these updated parameters. In most cases the revised parameter estimates are within the ranges considered in the original sensitivity analysis (see below). Two issues warrant further consideration.

Length of stay – The revised base case scenario suggests a cost saving but this is substantially reduced from the original. The main reason is the shorter mean length of stay for CHF suggested by the 2001 NHS Reference Costs database

Revising all other parameters and leaving LoS at 12 days, results in a net saving of £389,157, very similar to the original. The reference costs database distinguishes between mean length of stay for those aged under 70 (seven days), and those of 70 and older (11 days). For the younger age group the net saving is small at around £5,000.

Drug costs – The model is run assuming the cheapest ACE inhibitor (currently captopril at £52/year) is used. However, drugs costing up to £339/year are available. At this cost the model estimates a net cost increase of £616,000.

Sensitivity analysis indicates that the assumptions of the central case scenario would have to change substantially in order to eliminate the net saving result of the central scenario.

Discussion of the WGAP model and results

Modelling has indicated that managing heart failure patients according to good clinical practice, that is, using echocardiographic investigation and treatment using ACE inhibitors, can be achieved with potential marginal cost savings.

The updated central case scenario implies that, for a population of 500,000, an additional 42 premature deaths could be prevented per annum with a potential net cost saving of around £160,000 in ‘year one’.

The exclusion of potential savings from reduced demand for out-patient and GP attendances through the use of ACE inhibitors implies that the modelled savings are under-estimated in the central case.

The demand and costs for U&E tests and echocardiography have been ‘front loaded’ into ‘year one’, which means that the actual net costs could be even lower than those indicated by the results of the modelled central scenario.

The model does not allow for the possibility of induced extra demand for echocardiography for patients subsequently confirmed not to have heart failure. This induced demand will mean increased costs of echocardiography per patient with heart failure detected.

The increased use of echocardiography could result in potential savings from the reduction or avoidance of inappropriate or over-prescribing of diuretics and ACE inhibitors for misdiagnosed patients.

Recent work on screening out the need for echocardiography with pre-tests using electrocardiography and BNP testing may have implications for the assumptions relating to echocardiography employed here (see Appendix G).

The analysis has been confined to the ‘year one’ costs and benefits of treating heart failure patients with ACE inhibitors. The model does not fully address issues associated with treatment over the lifetime of the patient, eg changes in the reduction in risk of hospitalisation over time, costs of treating patients who live longer because of their improved health state.

The studies reported in tables D1 and D2, some of which do model the lifetime effects of ACE inhibitor treatment, indicate that the cost-effectiveness of ACE inhibitors, in terms of discounted LYG, is very favourable. Life years may even be gained with cost savings.

The SOLVD study121 excluded frail elderly patients and patients with severe heart failure. If the hospitalisation gains from treatment with ACE inhibitors are lower for these groups of patients than those found in the SOLVD sample, then the hospitalisation savings indicated by the model have been overestimated.

Modelled bed day savings, whilst a true opportunity for cost saving, will not automatically accrue to the purchasers as cash savings. The benefits are more likely to manifest themselves in the form of reduced waiting times and increased bed capacity in wards no longer occupied by heart failure patients.

2. Beta-blockers


Trial evidence suggests that beta-blockers can improve symptoms, prolong survival and reduce hospitalisation.

Therefore, while increasing the use of beta-blockers in eligible patients will increase treatment costs some, or all, of these will be offset, largely by reduced hospitalisation.

The majority of patients in beta-blocker trials are also being treated with ACE inhibitors, so the benefits of beta-blocker therapy can be considered as additional to those achieved with first line use of ACE inhibitors.

The existing economic evidence suggests that the use of beta-blockers is cost effective, or may be cost saving.

Trial follow-up is relatively short so we know little about the longer-term cost and benefit profiles of beta-blockers.

Recently published BHF statistics1 suggest that in 1998 only 11% of heart failure patients were receiving beta-blockers in England and Wales. However, this had increased from around 8% in 1994. A number of papers suggest that low prescribing rates are caused by concerns on tolerability, but no evidence is presented to support this claim.

Drug costs

Only two beta-blockers are licensed for the treatment of CHF in the UK.*

Both convenience of administration and cost are important factors in determining the appropriate drug to use.

There may be differences in therapeutic efficiency, symptomatic benefit and adverse effects between these two drugs, but no trial has carried out a direct comparison.

Existing guidance that has considered economic implications

The inherited NICE guideline on Prophylaxis for patients who have experienced a MI (North of England Guidelines Group, 2001)265 states:

  • beta-blockers are associated with a substantial reduction in all cause mortality in patients with symptoms of heart failure being treated with an ACE inhibitor, who may or may not have experienced an MI.

The relevant recommendation for patients with prior MI and CHF is:

  • ‘All patients should be offered long term treatment with an ACE inhibitor and then a beta-blocker… All of these treatments are cost effective.’ (A)

Existing economic evidence on the use of beta-blockers

Only six studies of adequate quality were identified by the economic review; these are summarised in Table D4. Five are based on data from the CIBIS trials (bisoprolol) and one employs data from the Carvedilol Heart Failure Trials Program.

Four of the studies are cost-effectiveness analyses; two of the studies look only at costs. Only three of the studies consider the UK context.

The studies by Varney292 and Delea et al 293 attempt to extrapolate the cost and benefits profile beyond the end of the trial by using decision analytic models. These are subject to a large degree of uncertainty. For the UK study (Varney) all CERs are well within acceptable limits of cost-effectiveness. For the US study (Delea et al) the worst case CERs are relatively high (close to £30,000) but this may reflect the high cost of Carvedilol in comparison with bisoprolol and a very conservative benefits scenario where benefits persist for six months only.

Like ACE inhibitors, there is some concern that treatment with beta-blockers may simply delay resource utilisation, or even increase it due to increased survival. There is no evidence to inform this view and trial follow-up is short.

Table D4. Existing evidence on cost-effectiveness of beta-blockers in the treatment of CHF.

Table D4

Existing evidence on cost-effectiveness of beta-blockers in the treatment of CHF.

Initiation and up-titration

In clinical practice achieving and maintaining dose is more difficult than in a trial.

Achievement of maintenance dose may be protracted, so considerable resource burden occurs early during initiation and up-titration.

Both beta-blockers are licensed for initiation in a community setting.

The proportion of patients dealt with in the community rather than the hospital may be an important determinant of costs.

The Varney study suggests that initiation by a specialist nurse in the community will be more expensive than initiation in hospital as the nurse is expected to spend more time with the patient.

3. Digoxin

Only one US-based study99 was identified, which presents an economic analysis of data from the PROVED and RADIANCE trials.

This study estimates costs and outcomes of continuation v. withdrawal of digoxin therapy in patients with normal sinus rhythm, NYHA II–III and LVEF <= 35%.

Trial data is supplemented with epidemiological information from published reports and expert opinion.

The trials suggest that digoxin may reduce the rate of heart failure related hospitalisation in patients who are (RADIANCE) and are not (PROVED) receiving ACE inhibitors.

The trials find no effect on life expectancy.

Continuation of digoxin therapy involves risks from toxicity, hence monitoring costs are an important component of resource use along with the costs of the drug itself, and the costs of treating side effects.

The base case scenario suggests that digoxin therapy is cost saving, that is the extra costs of the drug, drug administration, monitoring and toxicity are more than outweighed by savings from reduced hospitalisation.

Extensive sensitivity analysis was performed and the vast majority of cases showed digoxin to be cost saving.

This is a relatively old study and is based in the US, hence applicability to the UK context is unclear.

4. Diuretics

Five studies were identified but none consider the UK context (four US, one Germany), these are summarised in Table D5.

All of the studies compare torasemide (the newer, more expensive drug) with furosemide.

None of the studies constitute a formal economic evaluation although one does attempt a cost effectiveness calculation using improvement in NYHA class as the outcome measure.

None of the studies are of very high quality, in particular because they report very little sensitivity analysis of the results and they only consider costs over a relatively short time period.

Most of the studies show that the higher drug acquisition costs of torasemide are more than outweighed by lower healthcare resource use as a result of reduced hospitalisation.

However, in one of the prospective US community-based studies294 the total costs of treatment with torasemide and furosemide are shown to be very similar. The benefits from torasemide treatment are greater but no formal QoL outcomes or comparison between costs and benefits is undertaken.

The potential cost savings associated with torasemide are solely derived from reduced hospital admissions and readmissions, therefore these savings could only be realised if hospitals can capitalise on this reduction in demand for beds.

Table D5. Existing economic evidence on diuretics in the treatment of CHF.

Table D5

Existing economic evidence on diuretics in the treatment of CHF.

Quality of life and side effects

Evidence suggests that these drugs are most effective at relieving symptoms, hence we would expect the QoL benefits to be substantial.

Relatively small numbers of adverse events have been reported, and the majority are mild or transitory and do not require discontinuation of treatment.

Information for the UK

Diuretics are generally much cheaper than ACE inhibitors and beta-blockers.

Evidence from the General Practice Research Database suggests that most patients with heart failure (>90%) are already being treated with diuretics, but there appears to be no readily available information on the distribution of molecules used.

Diuretics are rarely used in heart failure as a monotherapy but there is no evidence on their incremental benefit or cost effectiveness when used alongside ACE inhibitors and beta-blockers.



A substantial number of people with heart failure are treated with the β-blocker atenolol for CHD, but this is not licensed for the treatment of heart failure.

Copyright © 2003, Royal College of Physicians of London.
Bookshelf ID: NBK65602


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...