Family Practice Advance Access originally published online on December 3, 2008
Family Practice 2009 26(1):69-74; doi:10.1093/fampra/cmn096
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clinically relevant diagnostic research in primary care: the example of B-type natriuretic peptides in the detection of heart failure
a Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, PO Box 85500, Stratenum 6.131, 3508 GA Utrecht
b Department of Cardiology, St. Antonius Hospital, PO Box 2500, 3430 EM Nieuwegein, The Netherlands
Correspondence to Frans H Rutten, Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, PO Box 85500, Stratenum 6.131, 3508 GA Utrecht, The Netherlands; Email: f.h.rutten{at}umcutrecht.nl
Received 6 June 2008; Revised 23 October 2008; Accepted 11 November 2008.
 |
Abstract |
|---|
 Top Abstract IntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
With the emergence of novel diagnostic tests, e.g. point-of-care tests, clinically relevant empirical evidence is needed to assess whether such a test should be used in daily practice. With the example of the value of B-type natriuretic peptides (BNP) in the diagnostic assessment of suspected heart failure, we will discuss the major methodological issues crucial in diagnostic research; most notably the choice of the study population and the data analysis with a multivariable approach. BNP have been studied extensively in the emergency care setting, and also several studies in the primary care are available. The usefulness of this test when applied in combination with other readily available tests is still not adequately addressed in the relevant patient domain, i.e. those who are clinically suspected of heart failure by their GP. Future diagnostic research in primary care should be targeted much more at answering the clinically relevant question ‘Is it useful to add this (new) test to the other tests I usually perform, including history taking and physical examination, in patients I suspect of having a certain disease’.
Keywords. Diagnosis, primary care, heart failure, B-type naturiuretic peptide.
|
Introduction |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
Diagnosing is a major and challenging task for GPs, and setting a diagnosis is the starting point for prognostication and adequate treatment. In daily practice, the diagnostic process starts with a patient presenting himself with certain complaints and after history taking and physical examination, including a patient’s medical history, the GP suspects a particular disorder or disease. This process is often implicit, as are the estimates of the probability of the suspected disease. The latter is partly due to the lack of evidence from adequate quantitative diagnostic research.1 Estimated disease probabilities would guide the GP in choosing whether an additional test is needed or whether the GP gained sufficient assurance to rule in or exclude the suspected disease. When subsequent tests provide information beyond already available test results (i.e. have added value), they ‘update’ the probability of the presence or absence of the disease. This process continues until the GP is confident enough to take clinical (treatment) decisions, for example because the GP’s treatment threshold for a specific disease, say 75% or 90% probability is reached.
With the example of B-type natriuretic peptide (BNP) in the diagnostic assessment of patients suspected of heart failure in primary care, we will discuss which evidence from diagnostic research should be available (and how such studies should be designed) before such a new diagnostic test can be qualified as an useful additional tool in every day primary care practice.
We will focus on the following design issues of such clinically relevant diagnostic research: (i) inclusion of the relevant patient population (domain); (ii) the application of a multivariable approach; (iii) the natural clinical hierarchy of diagnostic tests with emphasis on the ‘added’ value of the new test and (iv) does the study enable the prediction of probabilities of the disease? (Boxes 1 and 2).
| BOX 1. An example A company representative offers you some specimen of a new diagnostic test to detect heart failure in your practice; a bedside test to measure BNP. In short, BNP is a peptide that is mainly produced in the myocardium of the left ventricle in response to stretch, generally caused by increased volume and pressure overload. The latter is the mainstay of heart failure.2 The representative shows you tables with very high values of sensitivity and specificity of both more than 90%! Just a few drops of capillary blood on a test strip were put in the machine, and a few minutes later, the test result is available. ‘Something every GP should have: You save the patient and yourself the inconvenience of referral for echocardiography.’ The next day, you suspect one of the patients presenting himself at your office of having heart failure: a man, 78 years of age, 30 pack-years of cigarette smoking, who receives medication for diabetes and hypertension and who had a myocardial infarction 10 years ago. He now complains of slowly progressive breathlessness while walking with his dog. Since 3 weeks he sleeps with an extra pillow. On physical examination, the blood pressure is 164/92 mmHg, the pulse frequency is regular with 92 beats/minutes and the patient has minor peripheral oedema at both legs. On palpation, the apex beat is broadened and sustained in left lateral position and with auscultation a holosystolic cardiac murmur at the apex, suggestive of mitral valve insufficiency, is audible. No abnormal pulmonary sounds are audible. Unable to rule in or exclude heart failure, you cannot decide about the initiation of treatment and you want to perform a simple non-burdening additional investigation. You wonder whether the new BNP test would offer help at this very moment. During coffee break, you do a quick internet search with Google and PubMed. You find impressive numbers, 98% diagnostic accuracy, and you have more than 1000 hits with the search terms ‘diagnosis, BNP, heart failure’. Can you now conclude that this is a good test that perfectly fits as a first additional step in your diagnostic assessment of this type of patient?
|
| BOX 2. What is BNP? Apart from its mechanical role, the heart also has an endocrine function. Myocytes produce natriuretic peptides, a family of vasoactive hormones, in response to myocyte stretch and increased wall tension, primarily caused by pressure and/or volume overload.2 The atria mainly produce atrial natriuretic peptide, while BNP is produced mainly in the ventricles. The prehormone proBNP is cleaved upon release in the circulation in equal portions of the biological active hormone BNP and inactive split product amino-terminal proBNP (NTpro-BNP). Because of different ways of clearance, the half-life of BNP is 20 minutes and that of NTpro-BNP is 120 minutes. The bio-active hormone BNP augments urinary volume and urinary sodium excretion, relaxes vascular smooth muscles and thus counterbalances the sympathetic nervous and renin–angiotensin system, trying to maintain circulatory homeostasis.2 Both BNP can be measured in plasma, serum or full blood by fully automated and commercially available assays from different companies, including reliable point-of-care tests. Nowadays, immunoassays are based on two monoclonal antibodies, and differences between the most often used assays of different companies are small and generally not of clinical relevance.3 Blood concentrations of BNP and NTpro-BNP are expressed in pg/ml or pmol/l. The conversion factor for BNP is 1 pg/ml = 0.289 pmol/l and for NTpro-BNP it is 1 pg/ml = 0.118 pmol/l. Nowadays, pg/ml is most often used. BNP can be used for diagnostic and prognostic purposes in different clinical settings, in particular in patients with (suspected) heart failure.4–6 More recently, it has also been suggested that BNP can be helpful in titrating drug treatment in heart failure patients.7,8
|
|
Research in the relevant clinical domain |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
First of all, the population under study should be the population where there is a diagnostic dilemma. There is a diagnostic problem neither in patients with overt (or established) disease nor in healthy people. Studies that ‘screen’ ‘all’ patients in the community or those in the community with certain risk factors for a disease also lack the ‘diagnostic dilemma’. The aforementioned examples do not represent the domain of interest for clinically relevant diagnostic research, namely patients suspected of a certain disease by the doctor, based on symptoms and signs, in whom the doctor considers additional diagnostic testing. Thus, diagnostic accuracy data obtained in other populations, such as from the population at large or of those with established disease, are not applicable to individual patients from the domain of interest, i.e. those who consult the practising clinician with symptoms suggestive of the disease of interest. Consequently, diagnostic accuracy results with cut-off levels for BNP from population studies or studies in high-risk patients for heart failure (without suspicion of having heart failure) were not very helpful.9,10
In addition, results from diagnostic research from secondary care cannot be applied uncritically to the primary care setting, even when performed in patients in whom the same diagnostic problem exists. On the one hand, this is caused by higher prevalence rates (that is, higher pretest likelihood or prior probability) of the relevant diseases in suspected patients in secondary care compared to those in primary care, due to the selection process of referral. As a result, positive predictive values of tests in secondary care are higher than in primary care, while on the other hand, the negative predictive values of tests are lower than in primary care. Moreover, more advanced disease stages with higher levels of diagnostic markers will be presented in secondary care compared to primary care. This will in general provide higher sensitivity and lower specificity in secondary care compared to primary care, because sensitivity and specificity are not ‘fixed’ test characteristics, and vary according to differences in severity of disease.11 In laboratory tests with continuous values such as with BNP, this will lead to (much) higher values. When BNP is digitomized, the ‘optimal’ (highest sum of sensitivity and specificity) cutpoint will be much higher in the secondary care setting compared to primary care (see also Box 3 about cutpoints of BNP).
| BOX 3. Cutpoints of BNP Blood levels of BNP and NT-proBNP are continuous. To facilitate the use of BNP and NT-proBNP, single cutpoints have been proposed. The disadvantage, however, is the dramatic reduction of available information since levels in daily practice can range from nearly zero to ten thousands of pg/ml. A partial solution to this problem is to use two cutpoints, one exclusionary (rule out) and one confirmatory (rule in) threshold. Thus, more patients can be correctly classified as no heart failure on the one hand and (probable) heart failure on the other. In between the cutpoints, however, a grey zone is then created in which (combinations of) other diseases and phenomena such as renal disease, pulmonary embolism, pulmonary artery hypertension, cardiac ischaemia, atrial fibrillation, left ventricular hypertrophy and advanced age (>75 years) have to be considered as causes of increased blood BNP or NT-proBNP values.14 Many cutpoints for BNP and NT-proBNP have been published, mainly depending on the patient domain. Studies from secondary care assessing patients with acute dyspnoea in the emergency department published ‘ruling out’ cutpoints of NT-proBNP of 300 pg/ml (or 400 pg/ml) and BNP of 100 pg/ml.12,13,15 Studies from primary care assessing patients with slowly progressive dyspnoea and suspected of slow onset new heart failure published (much) lower values, with ‘optimal single’ cutpoints of about 125 pg/ml for NT-proBNP and 35–77 pg/ml for BNP, ruling out heart failure >95%.16–20 For the ‘ruling in’ cutpoint 1800 pg/ml (or 2000 pg/ml) for NT-proBNP and 400 pg/ml for BNP has been suggested for the secondary care population,12,15 while such a cutpoint is lacking for primary care patients.21 The biological basis for different cutpoints between secondary and primary care is disease severity, with high intracardiac pressures and volume overload and thus high plasma levels of BNP and NT-proBNP in acute dyspnoeic patients (suspected of acute new onset heart failure) assessed in the secondary care setting versus low intracardiac pressures and volumes in patients with chronic or slowly progressive dyspnoea patients from primary care and suspected of slow onset new heart failure.
|
In the example of heart failure, the population under study should be patients from primary care suspected of heart failure because of typical symptoms and signs such as (slowly progressive) shortness of breath, fatigue and/or peripheral oedema. In those patients, a diagnostic challenge exists and additional tests such as BNP and electrocardiography can be considered.
Because heart failure is a chronic progressive disease and effective morbidity and mortality reducing treatment is available, the GP has the difficult task to detect early stages of the disease, at a point of time when symptoms and signs are less overt and specific and plasma levels of BNP or the N-terminal fraction of BNP (NT-proBNP) often may not differ much from normal values3; a situation very different from secondary care, where BNP and NT-proBNP are typically measured in acute dyspnoeic patients visiting the emergency department, with typical signs and symptoms of fluid overload and often very high values of BNP and NT-proBNP in those with heart failure.12,13
Thus, the usefulness of a diagnostic test such as BNP and NT-proBNP in primary care can only be assessed when the diagnostic research is performed in close adherence to daily primary care practice. For this, consecutive patients suspected of heart failure by the GP should be investigated with as few as possible exclusion criteria to ensure wide applicability of the findings.
|
Multivariable approach, natural clinical hierarchy and the prediction of absolute probabilities |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
In daily practice, a diagnosis is never based on a single diagnostic test. Rather, multiple tests are applied, including age, gender, history taking, physical examination and most often simple laboratory tests. As a consequence, also diagnostic research should take multiple tests into account. This ‘multivariable’ diagnostic process has a natural hierarchy. First, age, sex and earlier diagnoses (such as myocardial infarction) will be considered since these parameters are known at the time of or even before the presentation of the complaints. Subsequently, history taking and physical examination will be performed, followed by easily available additional test (e.g. electrocardiogram, lung function tests or some laboratory tests). Finally, more complicated, expensive and patient-burdening tests may be considered (e.g. chest X-ray, echocardiography) that often require patient referral.22 The data analysis in diagnostic research should therefore incorporate multivariable techniques in which multiple diagnostic tests in their proper sequence (natural clinical hierarchy) can be assessed.22 Thus, before a new diagnostic test can be applied in every day practice, its value in ‘improving’ the prediction of the ‘absolute’ probability of the suspected disease in a clinically relevant manner should be established ‘beyond’ what can be derived from readily available information obtained from the clinical assessment (notably signs and symptoms). In other words, a new diagnostic test should prove to have added value.
Apart from showing that a (new) tests independently contributes to the detection or exclusion of a disease, the extent of this contribution can and should be quantified. This can be done by using multivariable techniques (usually multiple logistic regression analysis) with the disease as the dependent variable and the diagnostic tests (including sign and symptoms and the new tests as independent variables). Once the optimal combination of tests is established, each test gets a ‘weight’ according to its contribution to detect or exclude a disease, and thus a diagnostic score can be calculated in which an individual score translates to an absolute probability of the disease (ranging from 0% to 100%). In a ‘diagnostic or clinical prediction rule’, diagnostic score thresholds can be related to subsequent patient management, e.g. ruling out of the disease/assuring the patient in case of low scores and considering the disease to be present/initiating targeted drug therapy in case of high scores. Applying such a (simplified) rule in daily practice may assist the physician in his decision making. Although prediction rules are important tools, the physician’s own judgement (based on clinical assessment of the patient and the doctor’s clinical experience) should not be discarded. Examples of widely used decision rules are the Well’s rule to diagnose or exclude deep venous thrombosis, and a similar, but adapted rule for the primary care setting by Oudega et al. and the Ottawa Ankle rule.23–25
In the example of heart failure, the practising GP should not be interested to know how good a single BNP test diagnoses or rules out a disease (nor in the test's accuracy characteristics such as sensitivity, specificity, positive and negative predictive value), but needs to know whether BNP has added value beyond the clinical assessment in patients he suspects of having heart failure and whether electrocardiography and chest X-ray are viable alternatives. The additional tests either decrease or increase the estimated probability of heart failure in case these additional tests are negative or positive, respectively. When such probabilities then reach a threshold below which the physician is confident that the disease can be ruled out or beyond which he/she considers the disease present (and initiates therapy), the clinical relevance of the additional tests is evident.26 In addition, he may wish to have an easy to use diagnostic rule to calculate the absolute risk of heart failure in order to estimate the level of certainty, end the diagnostic process and enter the prognostic and therapeutic realm.
|
The ‘evidence’: diagnostic studies with BNP to detect heart failure in the primary care setting |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
We did not apply a standard systematic review methodology, but used available information of three recently published systematic reviews by Doust et al.,27 Hill et al.28 and Davenport et al.6 of the diagnostic accuracy of BNP (or NTpro-BNP) in detecting heart failure or reduced left ventricular ejection fraction (LVEF). We selected studies in the relevant domain, that is, patients suspected of heart failure in the primary care setting (leaving out studies that screened high risk or subpopulations or studies performed in the hospital setting). Furthermore, we selected studies in which the outcome was heart failure (with or without reduced LVEF), that is, signs and/or symptoms of heart failure in combination with objective evidence of (systolic and/or diastolic) ventricular dysfunction, preferably using a panel of clinicians to assess the final diagnosis.22,29,30 Thus, studies with as outcome heart failure with reduced LVEF (and thus missing about half of all patients with heart failure, namely those with a preserved LVEF) were excluded. Only four studies met the two selection criteria.16,18,20,31 Table 1 presents relevant characteristics of the four studies. The study patients were predominantly older than 70 years of age, as is typical for patients suspected of heart failure in primary care.32 The pretest likelihood of the study patients ranged from 23% to 34%, representing the prior probability for a patient the GP suspects of having heart failure.
|
None of the studies performed a complete multivariable analysis, providing data of the added value of either BNP or NT-proBNP on top of signs and symptoms. As a consequence, none of the studies provided absolute probabilities of heart failure or a diagnostic prediction rule applicable in the primary care setting.
|
Conclusions |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
Diagnostic studies in primary care should provide answers to diagnostic dilemmas of practising GPs. For this, they should be performed in patients suspected of a certain disease by their GP usually as a result of prior history, signs and symptoms. A multivariable data approach should be undertaken, quantifying the added value of a test on top of readily available information such as demographic data and the clinical assessment, that is, taking the natural hierarchy of the diagnostic process into account. Although this seems reasonable and straightforward, with the example of heart failure and BNP (or NTpro-BNP) in the primary care setting, we showed that studies executed in the aforementioned preferable way are lacking in literature.
BNP have been studied extensively since their first description in 1988 in many different patient populations, including primary care patients. However, the majority of studies were ‘test research’, providing accuracy data. BNP (or NT-proBNP) showed to have high negative predictive values as a single test, both in the primary and secondary care setting. Importantly, however, at completely different ‘exclusionary’ cutpoints. As such, the test may be useful as a ruling out test, however, at different cutpoints in primary care compared to secondary care.
Importantly, a straightforward answer to the question how should GPs use this test in combination with his clinical assessment and other readily available tests such as electrocardiography is still lacking.
|
Declaration |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
Funding: None.
Ethical approval: None.
Conflicts of interest: FHR has participated as lecturer and speaker in scientific meetings under the sponsorship of Chiesi and Roche Inc.
|
Notes |
|---|
Kelder JC, Rutten FH and Hoes AW. Clinically relevant diagnostic research in primary care: the example of B-type natriuretic peptides in the detection of heart failure. Family Practice 2009; 26: 69–74.
|
References |
|---|
|
TopAbstractIntroductionResearch in the relevant...Multivariable approach, natural...The 'evidence': diagnostic...ConclusionsDeclarationReferences |
|---|
1 Richardson WS. Five uneasy pieces about pre-test probability. J Gen Intern Med (2002) 17:882–883.[CrossRef][Medline]
2 Levin ER, Gardner DG, Samson WK. Natriuretic peptides. N Engl J Med (1998) 339:321–328.
3 Rutten FH, Cramer MJ, Zuithoff NP, et al. Comparison of B-type natriuretic peptide assays for identifying heart failure in stable elderly patients with a clinical diagnosis of chronic obstructive pulmonary disease. Eur J Heart Fail (2007) 9:651–659.
4 Battaglia M, Pewsner D, Juni P, Egger M, Bucher HC, Bachmann LM. Accuracy of B-type natriuretic peptide tests to exclude congestive heart failure: systematic review of test accuracy studies. Arch Intern Med (2006) 166:1073–1080.
5 Struthers A, Lang C. The potential to improve primary prevention in the future by using BNP/N-BNP as an indicator of silent ‘pancardiac’ target organ damage: BNP/N-BNP could become for the heart what microalbuminuria is for the kidney. Eur Heart J (2007) 28:1678–1682.
6 Davenport C, Cheng EY, Kwok YT, et al. Assessing the diagnostic test accuracy of natriuretic peptides and ECG in the diagnosis of left ventricular systolic dysfunction: a systematic review and meta-analysis. Br J Gen Pract (2006) 56:48–56.[Web of Science][Medline]
7 Troughton RW, Frampton CM, Yandle TG, Espiner EA, Nicholls MG, Richards AM. Treatment of heart failure guided by plasma aminoterminal brain natriuretic peptide (N-BNP) concentrations. Lancet (2000) 355:1126–1130.[CrossRef][Web of Science][Medline]
8 Jourdain P, Jondeau G, Funck F, et al. Plasma brain natriuretic peptide-guided therapy to improve outcome in heart failure: the STARS-BNP Multicenter Study. J Am Coll Cardiol (2007) 49:1733–1739.
9 Hobbs FD, Davis RC, Roalfe AK, Hare R, Davies MK, Kenkre JE. Reliability of N-terminal pro-brain natriuretic peptide assay in diagnosis of heart failure: cohort study in representative and high risk community populations. Br Med J (2002) 324:1498.
10 McDonagh TA, Robb SD, Murdoch DR, et al. Biochemical detection of left-ventricular systolic dysfunction. Lancet (1998) 351:9–13.[CrossRef][Web of Science][Medline]
11 Moons KG, van Es GA, Deckers JW, Habbema JD, Grobbee DE. Limitations of sensitivity, specificity, likelihood ratio, and bayes’ theorem in assessing diagnostic probabilities: a clinical example. Epidemiology (1997) 8:12–17.[Web of Science][Medline]
12 Januzzi JL, van Kimmenade R, Lainchbury J, et al. NT-proBNP testing for diagnosis and short-term prognosis in acute destabilized heart failure: an international pooled analysis of 1256 patients: the International Collaborative of NT-proBNP Study. Eur Heart J (2006) 27:330–337.
13 Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med (2002) 347:161–167.
14 De Lemos JA, McGuire DK, Drazner MH. B-type natriuretic peptide in cardiovascular disease. Lancet (2003) 362:316–322.[CrossRef][Web of Science][Medline]
15 Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008. Eur Heart J (2008) 29:2388–2442.
16 Zaphiriou A, Robb S, Murray-Thomas T, et al. The diagnostic accuracy of plasma BNP and NTproBNP in patients referred from primary care with suspected heart failure: results of the UK natriuretic peptide study. Eur J Heart Fail (2005) 7:537–541.
17 Fuat A, Murphy JJ, Hungin AP, et al. The diagnostic accuracy and utility of a B-type natriuretic peptide test in a community population of patients with suspected heart failure. Br J Gen Pract (2006) 56:327–333.[Web of Science][Medline]
18 Nielsen LS, Svanegaard J, Klitgaard A, Egeblad H. N-terminal pro-brain natriuretic peptide for discriminating between cardiac and non-cardiac dyspnoea. Eur J Heart Fail (2004) 6:63–70.
19 Gustafsson F, Steensgaard-Hansen F, Badskjaer J, Poulsen AH, Corell P, Hildebrandt P. Diagnostic and prognostic performance of N-terminal ProBNP in primary care patients with suspected heart failure. J Card Fail (2005) 11:S15–S20.[CrossRef][Web of Science][Medline]
20 Cowie MR, Struthers AD, Wood DA, et al. Value of natriuretic peptides in assessment of patients with possible new heart failure in primary care. Lancet (1997) 350:1349–1353.[CrossRef][Web of Science][Medline]
21 Rutten FH, Hoes AW. B-type natriuretic peptide assays for detecting heart failure in the elderly: same value as those in the younger? Int J Cardiol (2008) 125:161–165.[CrossRef][Web of Science][Medline]
22 Moons KG, Grobbee DE. Diagnostic studies as multivariable, prediction research. J Epidemiol Community Health (2002) 56:337–338.
23 Wells PS, Anderson DR, Bormanis J, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet (1997) 350:1795–1798.[CrossRef][Web of Science][Medline]
24 Stiell IG, Greenberg GH, McKnight RD, Nair RC, McDowell I, Worthington JR. A study to develop clinical decision rules for the use of radiography in acute ankle injuries. Ann Emerg Med (1992) 21:384–390.[CrossRef][Web of Science][Medline]
25 Oudega R, Moons KG, Hoes AW. Ruling out deep venous thrombosis in primary care. A simple diagnostic algorithm including D-dimer testing. Thromb Haemost (2005) 94:200–205.[Web of Science][Medline]
26 Pauker SG, Kassirer JP. The threshold approach to clinical decision making. N Engl J Med (1980) 302:1109–1117.[Abstract]
27 Doust JA, Glasziou PP, Pietrzak E, Dobson AJ. A systematic review of the diagnostic accuracy of natriuretic peptides for heart failure. Arch Intern Med (2004) 164:1978–1984.
28 Hill SA, Balion CM, Santaguida P, et al. Evidence for the use of B-type natriuretic peptides for screening asymptomatic populations and for diagnosis in primary care. Clin Biochem (2008) 41:240–249.[Medline]
29 Swedberg K, Cleland J, Dargie H, et al. Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): The Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. Eur Heart J (2005) 26:1115–1140.
30 Rutten FH, Moons KG, Cramer MJ, et al. Recognising heart failure in elderly patients with stable chronic obstructive pulmonary disease in primary care: cross sectional diagnostic study. Br Med J (2005) 331:1379.
31 Wright SP, Doughty RN, Pearl A, et al. Plasma amino-terminal pro-brain natriuretic peptide and accuracy of heart-failure diagnosis in primary care: a randomized, controlled trial. J Am Coll Cardiol (2003) 42:1793–1800.
32 Rutten FH, Grobbee DE, Hoes AW. Differences between general practitioners and cardiologists in diagnosis and management of heart failure: a survey in every-day practice. Eur J Heart Fail (2003) 5:337–344.
Read all E-letters
CiteULike 
Connotea 
Del.icio.us What's this?
E-letters:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||