Heart failure is common, causes major disability and often shortens life. Two UK guidelines advocate the measurement of plasma concentrations of B-type natriuretic peptide (BNP) in diagnosis of chronic heart failure, either in combination with, or as an alternative to, an ECG.1,2 Here we review the evidence for BNP testing in the diagnosis of chronic heart failure, and discuss the implications in terms of availability of the test.
Statistics from Altmetric.com
About heart failure
Heart failure is a syndrome resulting from a structural or functional cardiac disorder that reduces the functional ability of the heart to maintain normal circulation. Chronic heart failure may be caused by myocardial dysfunction, valve abnormalities or pericardial disease or may be induced by rhythm disturbances.3 Left ventricular systolic dysfunction (LVSD) is present in around half of cases.4 For example, in 2007/2008, the prevalence of heart failure and left ventricular dysfunction of patients registered to GPs in the UK was 0.84 and 0.47 per 100, respectively.5 Heart failure also has a high mortality rate: overall, 40% of patients admitted to hospital with heart failure die or are readmitted within 1 year and 50% of patients with the condition have died by 4 years.3
Conventional diagnosis of heart failure
Heart failure is characterised by symptoms and signs, such as breathlessness or fatigue, either at rest or during exercise, or ankle swelling, and objective signs of cardiac dysfunction at rest;3 these are the common presenting features in primary care. A recent UK Health Technology Assessment (HTA; discussed further below) considered five studies (involving a total of 2,187 patients) on diagnosis of heart failure (published up to July 2006); it concluded that breathlessness was the only symptom or sign with high sensitivity for identifying heart failure (89%); however, its specificity was poor (51%)4,*.These data highlight the fact that symptoms similar to those of heart failure can be presenting features of many other conditions, including obesity; chest disease; venous insufficiency in the lower limbs; drug-induced ankle swelling (e.g. dihydropyridine calcium channel blockers); drug-induced fluid retention (e.g. NSAIDs); hypoalbuminaemia; intrinsic renal or hepatic disease; pulmonary embolic disease; depression and/or anxiety disorders; severe anaemia; thyroid disease; and bilateral renal artery stenosis.1
Echocardiography can confirm a diagnosis of heart failure if it shows that the heart's ejection fraction is reduced (which can be an indication of LVSD). It can also identify cardiac abnormalities or myopathies that may be either causing or aggravating heart failure symptoms. Speed of access to echocardiography has increased in the UK in recent years; however, patients may still have to wait several weeks for the investigation. For example, data from the Department of Health in England indicates that in July 2009, of the 28,880 patients awaiting echocardiography, only 193 had been waiting more than 6 weeks but 7,093 had been waiting longer than 3 weeks.6 The varied or delayed access to echocardiography has meant that primary care physicians have traditionally relied on alternative investigations, such as an ECG, to help in deciding which patients should be referred for echocardiography. However, changes in ECG may be subtle, and the lack of skills in interpreting this investigation may still require referral for specialist opinion.4 These issues underlie the potential place of BNP in the diagnosis of chronic heart failure.
Characteristic findings in heart failure include increases in left ventricular volume and/or pressure. These changes stimulate ventricular muscle to increase production of a precursor protein, called pro-B-type natriuretic peptide (proBNP).7 This is then broken down (either within, or during secretion from, the muscle) to form two peptides that enter the bloodstream: BNP (which causes natriuresis, diuresis, vasodilation and muscle relaxation); and N-terminal pro BNP (NT-proBNP, which is biologically inactive).7,8 This explains the basis for measuring concentrations of these peptides as potential markers of heart failure, and there are several assays now available in the UK for BNP and NT-proBNP.9 The current widely used assays all have cut-offs of 100pg/mL for BNP and 125pg/mL for NT-proBNP.9
What guidelines say about BNP
The National Institute of Health and Clinical Excellence (NICE) guideline on chronic heart failure advises that if the condition is suspected, a 12-lead ECG, and/or measurement of BNP or NT-proBNP where available, should be used to indicate whether the patient needs to be referred for echocardiography (i.e. when either the ECG and/or the peptide concentration is abnormal).1 Guidelines from the Scottish Intercollegiate Guideline Network (SIGN) on chronic heart failure also advise ECG and/or BNP or NT-proBNP measurement in the diagnosis of suspected chronic heart failure, specifically advising that samples should ideally be taken before starting therapy, as BNP concentrations tend to fall afterwards.2 Similarly, European Society of Cardiology guidelines state that plasma concentrations of BNP and NT-proBNP are “useful” in the diagnosis of heart failure (with a low-normal concentration in an untreated patient making heart failure unlikely to be the cause of the symptoms).3
Potential limitations of the tests
Age and gender
A systematic review investigating the role of BNP in the diagnosis of heart failure noted that the performance of both BNP and NT-proBNP tests decreased with the age of patients.10 The diagnostic odds ratio (DOR–see Box 1)11 for BNP compared to standard clinical reference criteria for heart failure declined by a factor of 2.0 (95% CI 1.0 to 4.2) for each decade of increasing age. However, the evidence was based upon the average age of patients within the included studies, rather than on individual patient data, so the reliability of the result is questionable. Evidence also suggests that in the general population, BNP concentrations increase with age, and are higher in women than men.12
Diagnostics odds ratio (dor)
The DOR is the odds of positive test results in participants with disease divided by the odds of positive test result in those without disease.11
For example, if the DOR is 3, the odds of having a positive test are three times higher in someone with the disease than in someone without the disease.
Potential causes of high BNP concentrations in the absence of heart failure (false-positive results), include renal failure, myocardial infarction, acute coronary syndrome and acute pulmonary embolism.13 Potential causes of low concentrations in patients with heart failure (false-negative results) include rapid-onset pulmonary oedema, acute mitral regurgitation, mitral stenosis, atrial myoma and fully treated low-grade heart failure.
Diet, drugs and exercise
Evidence indicates that increasing dietary sodium intake causes plasma BNP concentrations to rise.14 Previous studies have suggested that some medicines reduce the concentration of BNP and NT-proBNP (including potassium-sparing diuretics, ACE-inhibitors and angiotensin-II receptor blockers),15 16 17 although as these drugs are used routinely in the treatment of patients with heart failure, it is unclear whether this reduction was an effect of the drug per se, or is due to improvement of the condition. Other studies have suggested that certain drugs raise BNP concentrations (e.g. digoxin).18 Although exercise has been shown to cause an increase in BNP concentrations in patients with heart failure,19,20 these increases are small.20
Handling of test samples
One study has suggested that BNP is stable in whole blood for 3 days,21 while other reviewers suggest that accuracy of the assays may depend on the duration between collection of the blood and performance of the assay.4
Current evidence on BNP testing
A recent HTA (mentioned above) has assessed evidence on, and strategies for, the use of plasma concentrations of BNP and NT-proBNP in the diagnosis of heart failure.4
Accuracy as a diagnostic marker
Twenty studies in the HTA, including a total of 5,030 patients, examined the accuracy of BNP for a diagnosis of heart failure defined according to standard clinical criteria. The results showed BNP had a consistently high sensitivity (0.93, 95% CI 0.91 to 0.95) but varying specificity (0.74, 95% CI 0.63 to 0.83), with a DOR of 39.5 (95% CI 21.4 to 72.6); a positive likelihood ratio of 3.6 (95% CI 2.4 to 5.2); and a negative likelihood ratio of 0.09 (95% CI 0.06 to 0.13)*. In subgroup analysis of four studies involving a total of 883 patients in (or referred from) general practice, the sensitivity of BNP was slightly lower (0.84, 95% CI 0.72 to 0.92), but the specificity was similar (0.73, 95% CI 0.65 to 0.80).
The HTA reviewed 16 studies involving a total of 4,513 patients, which examined the accuracy of NT-proBNP for the diagnosis of heart failure defined by standard clinical criteria. The results for NT-proBNP showed generally high sensitivity (0.93, 95% CI 0.88 to 0.96) but specificity was lower than for BNP (0.65, 95% CI 0.56 to 0.74), giving a DOR of 24.6 (95% CI 14.4. to 42.2); a positive likelihood ratio of 2.7 (95% CI 2.1 to 3.4); and a negative likelihood ratio of 0.11 (95% CI 0.07 to 0.18). Results for a subgroup of eight studies involving a total of 2,733 patients in (or referred from) general practice were similar to the overall results for NT-proBNP, with slightly lower specificity.
In the past, there has been conflicting evidence as to whether BNP and NT-proBNP are comparable in terms of diagnostic accuracy in heart failure.9,22 The HTA reviewed six studies including a total of 1,623 patients, which directly compared the diagnostic accuracy of BNP with that of NT-proBNP for the clinical diagnosis of heart failure and found no statistical difference between them. There was also no clear evidence of the superiority of one specific assay over any another for either BNP or NT-proBNP tests.
Comparisons with ECG
In four studies including a total of 1,889 patients comparing BNP to ECG for the diagnosis of heart failure, BNP was shown to have greater diagnostic accuracy (relative DOR of ECG/BNP 0.32, 95% CI 0.12 to 0.87). However, there was no difference in the diagnostic accuracy of NT-pro BNP compared to ECG in these same four studies.
Predicting the likelihood of heart failure
The authors of the HTA report used analysis of individual patient data from studies in their systematic review to develop and test a model for predicting the likelihood of a patient having heart failure.4 This model was based on using a simple clinical scoring system, used alongside a BNP concentration, to assess the need for echocardiography. Testing of the model allowed the authors to develop a simple clinical decision rule, stating that any patient presenting with symptoms such as breathlessness and suspected of having heart failure should be referred directly for echocardiography, if the person had a history of myocardial infarction; had basal crepitations; or was a male with ankle oedema. Such factors automatically gave them a sufficiently high risk of heart failure. Otherwise, the BNP concentration should be measured and the decision on whether to refer should be based on the test result.
Addition of an ECG to the model (i.e. in combination with clinical assessment and BNP testing) did not improve the model's diagnostic accuracy. This finding, and the superior diagnostic accuracy of BNP to ECG, led the HTA's authors to conclude that there “is no need to perform an ECG as part of the assessment of whether or not heart failure is present”.
Crucially, unlike earlier studies and systematic reviews, the HTA analysis found “no consistent evidence” of an association between the performance of BNP or NT-proBNP and patients' age or gender. There was also no consistent evidence of an association between the assay concentrations and the presence or not of chronic obstructive pulmonary disease, obesity, ischaemic heart disease, atrial fibrillation, diabetes mellitus or existing pharmacological therapy at the time of the diagnostic test.
Analysis by the authors of the HTA report indicated that use of their clinical decision rule in terms of heart failure diagnosis would be cost-effective to the NHS in terms of cost per additional case detected. Furthermore, when the analysis was extended to consider potential benefits to patients in terms of increased life expectancy (rather than just costs to the NHS), the results led the authors to conclude that “the optimum strategy would be to refer all patients with symptoms suggestive of heart failure directly for echocardiography”. However, in recognising the resource implications in adopting such a policy, they suggested that “in the presence of a limited supply” of echocardiography, their clinical rule decision rule should be used as set out below (see Box 2).
Clinical decision rule for bnp testing in chronic heart failure4
The clinical decision rule in the recently published HTA offers a practical approach in assessing patients with suspected heart failure. However, adoption of this practice raises key issues. For example, measurement of BNP concentration for diagnosis should probably be done before instigation of any treatment that could improve cardiac function (and so alter BNP concentration). This may be problematic if the doctor or patient is unwilling to delay treatment for suspected heart failure until a blood sample can be taken. Also, recent surveys have suggested that only a third of primary care trusts have access to BNP testing for aiding diagnosis of heart failure (although the number is increasing).23
Heart failure is a common condition in the UK, causing significant morbidity and mortality. It has been traditionally diagnosed both clinically and by ECG and subsequent referral for echocardiography. However, measurement of plasma concentrations of B-type natriuretic peptide (BNP) or N-terminal pro-BNP (NT-proBNP) assays have also been recommended by guidelines as a first-line diagnostic tool as an alternative to, or in combination with, ECG. Current evidence suggests that BNP assays as a diagnostic tool of exclusion in the diagnosis of heart failure are superior to ECG, irrespective of the individual assay used. Also, evidence suggests that it is a cost-effective use of NHS resources to apply a simple clinical decision rule to identify those who should be referred directly for echocardiography, as well as those who should first undergo BNP testing to see whether their concentration justifies such referral. Evidence also suggests that if there were unlimited access to echocardiography, referral of all patients with suspected heart failure would be cost-effective, if the benefits in terms of patient survival (and not just NHS resources) were taken into account.
Currently, access to BNP tests is not uniform throughout the UK. We believe that the tests should be available to all to aid the diagnosis of chronic heart failure, particularly where there is no reliably prompt access to echocardiography.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.