Family Practice Vol. 20, No. 5, 570-574
© Oxford University Press 2003
Health Services Research |
The use of brain natriuretic peptide as a screening test for left ventricular systolic dysfunction— cost-effectiveness in relation to open access echocardiography
Department of Medicine, Llandough Hospital, Vale of Glamorgan,
a Department of Clinical Biochemistry and
c Department of Cardiology, Royal Gwent Hospital, Newport and
b Centre for Health Economics and Policy Studies, School of Health Science, University of Wales Swansea, Singleton Park, Swansea, UK.
Correspondence to Dr Victor Sim, Directorate of Medicine, On Call Landing, Llandough Hospital, Penlan Road, Penarth, Vale of Glamorgan CF64 2XX, UK; E-mail: victor.sim{at}uhw-tr.wales.nhs.uk
Sim V, Hampton D, Phillips C, Lo S-N, Vasishta S, Davies J and Penney M. The use of brain natriuretic peptide as a screening test for left ventricular systolic dysfunction—cost-effectiveness in relation to open access echocardiography. Family Practice 2003; 20: 570–574.
Received 23 October 2002; Accepted 19 May 2003.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. Heart failure due to left ventricular systolic dysfunction (LVSD) has a high prevalence in the adult population but is difficult to diagnose accurately on clinical grounds in the community. Early diagnosis is important as effective treatments are available to reduce morbidity and mortality. Echocardiography is widely used to assess heart failure; however, this technology is relatively expensive and of limited availability. A potential diagnostic aid in primary care is the measurement of plasma brain natriuretic peptide (BNP).
Objective. This study was performed to assess the value of BNP measurement as a selective pre-screen for breathless patients referred for open access echocardiography.
Methods. BNP was measured by radioimmunoassay with prior extraction in 83 breathless subjects (age range 37–87 years, mean 72). Standard echocardiography was performed and left ventricular systolic function was assessed.
Results. The prevalence of LVSD was 31% in this group. At cut-off values chosen to give negative predictive values for LVSD of >98% (BNP = 19 pg/ml), the sensitivity, specificity and positive predictive value for BNP were 100, 49.1 and 46.9%. Using this BNP threshold as a pre-screen for echocardiography would make a net saving of £964.20 without compromising the diagnostic accuracy.
Conclusion. BNP measurement appears to have a significant cost-effective benefit for the selection of patients for echocardiography.
Keywords. Brain natriuretic peptide, cost-effectiveness, echocardiography, heart failure, open access.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Chronic heart failure (CHF) due to left ventricular systolic dysfunction (LVSD) has a high morbidity and mortality. A recent UK population-based echocardiographic study reported LVSD (left ventricular ejection fraction
30%) with an overall prevalence of 3%.1 This prevalence rises to >10% of those aged 70 years and over.1,2 The costs of this syndrome, in both economic and personal terms, are considerable. Health care expenditure on CHF in developed countries consumes 1–2% of the total health care budget.3 Effective treatments for patients with CHF due to left ventricular dysfunction are now available;4,5 however, the condition is often difficult to diagnose in the primary care setting, and many patients remain undiagnosed and untreated.6 Echocardiography plays a major role in diagnosing heart failure, assessing not only ventricular function but also cardiac valvular function. Performing and interpreting echocardiography, however, requires considerable skill and training: echocardiography is currently of limited availability in the UK.7
Natriuretic peptides are a family of peptides with similar secondary structure. Three peptides have been described so far in human physiology—atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP) and C-type natriuretic peptide (CNP).8 Both ANP and BNP are secreted from the heart myocytes, predominantly ANP from the low-pressure atria and BNP from the high-pressure ventricles. The stimulus for secretion is muscle stretch, and both peptides share common receptors. The immediate physiological effect of ANP and BNP secretion is an increase of renal sodium and chloride secretion with concomitant inhibition of renin and aldosterone.9 The natriuretic peptides are stored within the myocytes as pro-hormones and, when released, are cleaved to the natriuretic C-terminal fragment and the non-natriuretic N-terminal fragment.
The measurement in plasma of cardiac natriuretic peptides and the N-terminal fragments has been proposed as methods of assessing left ventricular dysfunction,10 and hence as a potential screening tests for CHF. This approach is particularly attractive in the primary care setting as blood samples can be taken from patients in the community and transported to the central hospital laboratory for analysis.11
An open access echocardiography service has been available for GPs in the Newport, South Wales catchment area, serving 110 GP practices with a population of 556 622, since 1996, with >400 referrals per annum in recent years. The aim of this study was to determine if the measurement of BNP, the most stable natriuretic peptide in whole blood, could provide evidence of a feasible cost-effective prior stratification of patients referred to the open access echocardiography service.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
An open access echocardiography service has been established at the Royal Gwent Hospital since 1995. The most common reasons for requesting echocardiography in this service are to assess patients with breathlessness and heart murmur.12 Between 1997 and 1998, 83 patients with symptoms of breathlessness underwent echocardiography via the open access service and a 20 ml sample of venous blood was drawn from each patient for natriuretic peptide measurements. Patients with symptoms of both breathlessness and heart murmur were excluded. Ethical approval for this study was granted by the local health authority, and all patients provided consent.
Echocardiography assessment
Two-dimensional images, M-mode recordings and Doppler ultrasound interrogation were obtained using a Hewlett Packer 2000 scanner equipped with a 2.5 MHz phased array transducer. The subjects were examined in the left lateral position, and standard parasternal, short-axis and apical views were obtained. All the echocardiograms were analysed independently by two echocardiographers. Disagreements between the echocardiographers were adjudicated by the third observer.
Ejection fraction by Simpson’s rule was measured whenever possible, and left ventricular systolic function was considered impaired when an ejection fraction was 35%. When the endocardium was not clearly defined, the global left ventricular systolic function was assessed qualitatively by ‘eyeball’ estimate and categorized as normal, mildly, moderately or severely impaired.
Measurement of plasma BNP
Blood was obtained by venepuncture after the patient had been seated quietly for 30 min. Samples were transferred into specially prepared EDTA tubes containing 500 KIU/ml aprotinin (Trasylol, Bayer plc, Newbury, UK) and then immediately centrifuged. The plasma was stored frozen at –70°C until assayed.
BNP was measured by radioimmunoassay. Briefly BNP was extracted from plasma using Sep-pak C18 cartridges (Waters Ltd, Elstree, UK). Following elution and evaporation to dryness, the reconstituted extracts were assayed using anti-BNP antibody purchased from Bachem Ltd, St Helens, UK. Purified [125 I]BNP (also supplied by Bachem) was used as tracer, and BNP standards were prepared from material supplied by the Sigma Chemical Company. The recovery of BNP from the extraction procedure was 85% (±5%, n = 10). The intra-assay coefficient of variation for control of samples was 12% at 44 pg/ml (n = 10) with a detection limit of 2 pg/ml.
Assessment of costs
The cost of echocardiography was based on a published article in the British Society of Echocardiography Newsletter.7 The cost per patient, based on one trained full-time technician performing 1400 investigations per year, amounted to £42.07. In this study, the costs are based on a staff grade doctor performing the same number of investigations per year.
The cost of a plasma BNP assay was estimated using a standard laboratory costing package (Data Tree International Limited, Dublin, Ireland). The costing was based on a single plasma extraction with assay in duplicate.
Cost analysis
The approach adopted was to determine the costs of the service ‘currently provided’—all patients undergo echocardiography to determine the presence or absence of LVSD—and compare this with the costs of a service where all patients receive BNP, which then determines the need for echocardiography. Obviously, the effectiveness of BNP in identifying those patients who require echocardiography and consequently the number of patients with LVSD is a crucial factor in this process and thus it was necessary to assess the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) at different BNP thresholds. This information, together with the costs of performing an echocardiogram and the costs of the BNP assay, would enable the incremental cost-effectiveness ratios, based on different threshold levels of BNP, to be calculated.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Cost of echocardiography
The purchasing price of an echocardiography machine is
£125 000 (1999).7 While these machines do not wear out easily physically, they rapidly become technically obsolete. An echocardiography machine has a useful technical life of <5 years. The major manufacturers provide upgrades to keep machines up to date as technology develops, and extend the overall life of the machine to 12 years. As a rough guide, plan on at least 5% of the purchase price per year, this will cost £30 000 to upgrade the machine every 3–5 years. Since bank interest could be earned on the capital sum, the ‘cost of money’ needs to be included. In the UK, the current (1999) Treasury rate of the Government is 6%. The annual charge for servicing the machine is 6–8% of the purchasing price.7 The annual cost over machine life is £32 700 as illustrated in Table 1
.
|
Therefore, the cost per echocardiography study performed by a staff grade doctor as in this study is £54.07, as illustrated in Table 2
.
|
Cost of BNP measurement
The total cost of each BNP result was calculated to be £6.62 (fixed cost £2.83, variable cost £3.78) (Table 3
). There were no additional costs involved in terms of patient management, as the test was done on blood routinely collected as part of the diagnosis of the patient’s condition.
|
The prevalence of impaired left ventricular systolic function
Eighty-three patients with breathlessness, 43 female and 40 male, aged between 37 and 87 years (mean age 72) referred to the open access echocardiography service were identified. Twenty-six patients had LVSD.
The sensitivity, specificity, PPV and NPV of BNP
The sensitivity, specificity, PPV and NPV at different BNP thresholds in identifying those subjects who subsequently were confirmed as having LVSD by echocardiography are shown in Figures 1 and 2.
|
|
Cost-effectiveness analysis
The number of patients assessed as requiring echocardiography by plasma BNP at a threshold of 19 pg/ml was 55, and of these 26 were assessed as having LVSD. At a plasma BNP threshold of 20 pg/ml, 49 patients required echocardiographic assessment, and of these 25 were assessed as having LVSD, one less than the current service.
Details of the cost analysis are shown in Table 4. With a BNP threshold of 19 pg/ml and echocardiography performed by a doctor, there would be a net saving of £964.20 without compromising the diagnostic accuracy. At the threshold of 20 pg/ml, however, there would be an additional false negative if BNP were used as a prior stratification, but it would cost £1288.20 less than having all patients assessed using echocardiography.
|
Sensitivity analysis
If the original cost per echocardiogram (echocardiography performed by technician only) of £42.07 were used, the difference in costs at the BNP threshold of 19 pg/ml would have been £630.16 and, when at the BNP threshold of 20 pg/ml, the cost per additional patient diagnosed as having LSVD using the current service would be £882.58.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
An open access echocardiography service is popular among GPs, and many cardiology departments are providing this service.12 With the increasing prevalence of patients with heart failure,1,13 open access schemes may overstretch existing service. It has been suggested that BNP levels could be used to pre-select patients for echocardiography.10
McDonagh et al. demonstrated that a cut-off value of BNP 
17.9 pg/ml was able to detect LVSD in subjects aged over 55 years with ischaemic heart disease at a sensitivity of 92% and specificity of 72%.14 The corresponding NPV and PPV were 98.5% and 32%, respectively, with a prevalence of LVSD in the study population of 12.1%. In another study of 122 patients referred to a rapid-access heart failure clinic, BNP 22.2 pmol/l (76.8 pg/ml) had a sensitivity of 97%, a specificity of 84%, and a corresponding NPV of 98% and PPV of 70% in identifying patients with clinical heart failure.10 The prevalence of heart failure syndrome was 29% in this study.
Our study demonstrates an NPV of 100% and a PPV of 46.9% using a BNP value of 19 pg/ml in detecting LVSD. The PPV is higher than that of the McDonagh et al. study (46.9% versus 32%), and this is probably due to the higher prevalence of LVSD in the present study (31% versus 12.1%). The relatively low PPV is probably due to the fact that natriuretic peptide levels reflect not only systolic dysfunction but also diastolic dysfunction and possibly other cardiac abnormalities.15 Diastolic heart failure is thought to account for >50% of heart failure in patients aged over 75 years.16 This perhaps explains why the study of Cowie et al. has a relatively high PPV of 70% because they used BNP to detect heart failure syndrome, which included diastolic heart failure and other causes of heart failure in addition to systolic heart failure.10
Various medications have been shown to affect the natriuretic peptide levels, and these include diuretics, ß-blockers, digoxin and vasodilators.17–20 Since most patients referred to the open access echocardiography service were taking one or more of these medications, this could further complicate the use of natriuretic peptides in diagnosing LVSD.
Using the BNP threshold of 19 pg/ml as prior stratification of the 83 patients referred to the open access echocardiography service, there is a net saving of £964.20 with no false-negative result, i.e. the sensitivity of BNP at this threshold in identifying LVSD is 100%. At the BNP threshold of 20 pg/ml, there is a net saving of £1288.20 but at the cost of missing one patient with LVSD. These data suggest that it is cost-effective to use BNP at the threshold of 19–20 pg/ml as a pre-screen for echocardiography. Increasing the thresholds of BNP values to >20 pg/ml will compromise the sensitivity of BNP as a pre-screen test and increase the clinical risk of missed diagnosis. Screening efficiency may be considerably enhanced if age-related cut-offs are used. Natriuretic peptide levels in plasma rise throughout adulthood20 so that higher cut-off values will need to be applied to older aged groups.
The strength of this study is that it was performed based on a ‘real world’ established clinical practice, and one can therefore be more confident to extrapolate the results and its implications into our clinical practice.
Access to echocardiography is still limited in many regions in the UK7 because of the lack of qualified echocardiographers, but measurement of plasma BNP is relatively easy, inexpensive and could, with adequate resources, be geared up to supply a service with little lead-in time. However, many UK laboratories would no longer be equipped to provide a radioimmunoassay service and there is an urgent need for reliable, robust and inexpensive non-isotopic assays to be provided which would be suitable for analysis on large automated analytical platforms. In addition, there is an urgent need for international standardization of BNP and the development of accredited external quality assurance programmes to support such laboratory assays. BNP is also a very powerful prognostic marker for patients with chronic heart failure,21 and thus once the assay was introduced into the laboratory repertoire demand would increase further to support this monitoring role.
In clinical practice, diagnosing heart failure syndrome is difficult, and physicians rarely rely on one single diagnostic modality. BNP measurement is an important tool in diagnosing this syndrome and appears to be a cost-effective prior stratification of patients referred to the open access echocardiography service. Our data suggest that BNP thresholds of 19–20 pg/ml may be cost-effective as a pre-screen for echocardiography. A study with a larger cohort is needed to confirm the use of these thresholds as a pre-screen.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 McDonagh TA, Morrison CE, Lawrence A et al. Symptomatic and asymptomatic left-ventricular systolic dysfunction in an urban population. Lancet 1997; 350: 829–833.[CrossRef][ISI][Medline]
2 Kannel WB, Belanger AJ. Epidemiology of heart failure. Am Heart J 1991; 121: 951–957.[CrossRef][ISI][Medline]
3 Berry C, Murdoch DR, McMurray JJ. Economics of chronic heart failure. Eur J Heart Failure 2001; 3: 283–291.[CrossRef][ISI][Medline]
4 Garg R, Yusuf S. Overview of randomised trials of angiotensin-converting enzyme inhibitors on mortality and morbidity in patients with heart failure. J Am Med Assoc 1995; 273: 1450–1456.
5 Bristow MR. Beta-adrenergic receptor blockade in chronic heart failure. Circulation 2000; 101: 558–569.
6 Morgan S, Smith H, Simpson I et al. Prevalence and clinical characteristics of left ventricular dysfunction among elderly patients in general practice setting: cross sectional survey. Br Med J 1999; 318: 368–372.
7 Leech G. Optimising our resources to meet the clinical demand. Br Soc Echocardiogr Newslett September 1998: 7–9.
8 Koller KJ, Goeddel DV. Molecular biology of the natriuretic peptides and their receptors. Circulation 1992; 86: 1081–1088.
9 Levin ER, Gardner DG, Samson WK. Natriuresis peptides. N Engl J Med 1998; 339: 321–328.
10 Cowie MR, Struthers AD, Wood DA, Coat AJS, Thompson SG, Poole-Wilson PA. Value of natriuretic peptides in assessment of patients with possible new heart failure in primary care. Lancet 1997; 350: 1347–1351.
11 Murdoch DR, Byrne J, Morton JJ et al. Brain natriuretic peptide is stable in whole blood and can be measured using a simple rapid assay: implications for clinical practice. Heart 1997; 78: 594–597.
12 Sim MFV, Davies J. GP open access echocardiography—Newport experience. Br J Cardiol 1998; 5: 286–290.
13 Cowie MR, Wood DA, Coats AJS et al. Incidence and aetiology of heart failure: a population based study Eur Heart Jl 1999; 20: 421–428.
14 McDonagh TA, Robb SD, Murdoch DR et al. Biochemical detection of left ventricular systolic dysfunction. Lancet 1998; 351: 9–13.[CrossRef][ISI][Medline]
15 Yamamoto K, Burnett J Jr, Jougasaki M et al. Superiority of brain natriuretic peptide as a hormonal marker of ventricular systolic dysfunction and diastolic dysfunction and ventricular hypertrophy. Hypertension 1996 ; 28: 988–994.
16 Vasan RS, Benjamin EJ, Levy D. Prevalence, clinical features and prognosis of diastolic heart failure: an epidemiologic perspective. J Am Coll Cardiol 1995; 26: 1565–1574.[Abstract]
17 Struthers AD. Effect of ACE inhibitors on the atrial natriuretic peptides in heart failure. Br Heart J 1994; 72: 504.
18 Northridge DB, Newby DE, Rooney E, Norrie J, Dargie HJ. Comparison of the short-term effects of Candoxatril, an orally active neutral endopeptidase inhibitor, and frusemide in the treatment of patients with chronic heart failure. Am Heart J 1999; 138: 1149–1157.[CrossRef][ISI][Medline]
19 Sanderson JE, Chan WW, Hung YT et al. Effects of low dose betablockers on atrial and ventricular (B-type) natriuretic factor in heart failure: a double blind, randomized comparison of metoprolol and a third generation vasodilating betablocker. Br Heart J 1995; 74: 502–507.
20 Tsutamato T, Wada A, Maeda K et al. Digitalis increases brain natriuretic peptide in patients with severe congestive cardiac failure. Am Heart J 1997; 134: 910–916.[CrossRef][ISI][Medline]
21 Tsutamato T, Maeda Y, Wada A et al. Plasma brain natriuretic concentration as a prognostic predictor in patients with chronic congestive cardiac failure. Circulation 1988: [Suppl 4.2]: 321–322.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||