Could BNP screening of acute chest pain cases lead to safe earlier discharge of patients with non-cardiac causes? A pilot study
From the Division of Medicine and Therapeutics, Ninewells Hospital and Medical School, Dundee DD1 9SY, UK
Address correspondence to Professor A.D. Struthers, Division of Medicine and Therapeutics, Ninewells Hospital and Medical School, Dundee DD1 9SY, UK. email: a.d.struthers{at}dundee.ac.uk
Received 29 August 2007 and in revised form 3 October 2007
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Background: The assessment of chest pain relies on clinical assessment and markers of cell necrosis such as Troponin T (TnT). B-type natriuretic peptide (BNP) is also raised in myocardial ischaemia and therefore may be useful in deciding if acute chest pain is of cardiac origin or not.
Aim: To investigate the role of BNP in the assessment of unselected patients presenting with acute chest pain.
Methods: A prospective observational study of 100 patients presenting with chest pain to the Acute Medical Admissions Unit was carried out. All patients had BNP and TnT levels measured. The primary outcome was categorization of chest pain as cardiac or non-cardiac. This was determined by the discharge diagnosis. BNP cutoffs were derived from a receiver operated characteristic (ROC) curve. The sensitivity, specificity, positive predictive accuracy (PPA) and negative predictive accuracy (NPA) were all calculated for BNP, TnT and for the composite of BNP and TnT.
Results: Mean BNP in patients with cardiac chest pain was significantly greater than mean BNP for patients with non-cardiac chest pain (P 
0.0001). BNP was significantly more sensitive than TnT (P = 0.003). However TnT was more specific than BNP (98% vs. 75%, P 0.0001). Combining BNP and TnT increased sensitivity from 55.6% to 95.6%.
Conclusion: Our findings suggest that BNP is more sensitive but less specific than TnT in the diagnostic assessment of acute chest pain. However, combining BNP and TnT was a very satisfactory rule out test (negative predictive accuracy 96%) for excluding chest pain that had a cardiac origin.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Annually around 500 000 patients attend Accident and Emergency departments in the United Kingdom with chest pain and 20–30% of acute medical admissions are for acute chest pain.1 Currently clinicians rely on the history, clinical examination, an electrocardiogram (ECG) and blood tests detecting irreversible cardiac myocyte damage such as TnT to determine whether the pain is cardiac in origin or not.
Less than half of patients presenting with chest pain have a final diagnosis of myocardial infarction or unstable coronary disease.2 Overall, this means that more than half of the patients presenting with chest pain do not have significant coronary heart disease and these patients represent over 10% of all acute medical admissions in the United Kingdom.3
However, it has also been reported that 
2–4% of acute myocardial infarctions are missed by emergency departments in the United States.4,5 The exact figure in the United Kingdom is unknown. However, a recent study estimated that around 6% of patients discharged from A&E have had a prognostically significant myocardial infarction.6 The discharge of such patients is unsafe and associated with an increased mortality.5 Efforts are therefore continually being made to improve the sensitivity of the risk stratification techniques and the underlying diagnostic methods for unselected patients presenting with acute chest pain.
B-type natriuretic peptide (BNP) is one of a family of structurally similar peptide hormones that also includes atrial natriuretic peptide (ANP), C-type natriuretic peptide (CNP), Dendroaspis natriuretic peptide (DNP), adrenomedullin and urodilatin.7 BNP is produced by myocytes as a result of the cleavage of a precursor molecule (pro-BNP), which also produces NT-proBNP in an equimolar ratio in response to ventricular stretch.8 It is this potential of BNP to detect haemodynamic stress in the left ventricle that has led to most of its initial use as an aid to the diagnosis and management of congestive cardiac failure. As a result BNP now has an established role as a ‘rule out’ test for patients presenting with suspected heart failure.7 This use of BNP is endorsed by its inclusion in the current UK NICE and SIGN guidelines on heart failure.9,10
Recent evidence suggests that lower levels of BNP may also be able to detect myocardial ischaemia, irrespective of haemodynamics.11 Indeed there is a large body of evidence which shows that BNP levels increase with chronic myocardial ischaemia and may be proportional to the extent and severity of ischaemia.12–15 It has also been shown that BNP levels on admission are a powerful prognostic marker in patients admitted with an acute coronary syndrome16 and it is now believed that myocardial ischaemia itself leads to increased cardiac expression of BNP.11 Bassan et al.17 reported that BNP is certainly able to identify infarction in patients presenting with chest pain. However, whether BNP identifies myocardial ischaemia in unselected patients admitted acutely with chest pain has not been hitherto assessed.
The traditional biomarkers such as the cardiac troponins used for the assessment of acute chest pain are only released when there is irreversible damage to the cardiac myocytes. However BNP is released by intact ischaemic myocytes.11 Therefore BNP has the potential to detect areas of ischaemic myocardium that have not infarcted and would otherwise be missed by the traditional indices of myocyte necrosis.
The purpose of this study was therefore to prospectively investigate the diagnostic usefulness of measuring serum BNP in patients admitted with acute chest pain alone or in combination with other clinical variables. The possibility exists that if BNP was a good rule-out test for acute cardiac chest pain, then it could lead to safe and early discharge of patients with acute chest pain cases who would otherwise be admitted overnight for observation and further assessment.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Study design
A prospective observational study to assess the role of BNP in the detection of ischaemic chest pain was performed. The study was carried out in the acute medical admissions unit in Ninewells Hospital, Dundee.
Study population
All patients admitted with a principal complaint of chest pain were potentially eligible for inclusion into the study. Patients were identified at random by reviewing the medical notes of all patients admitted to the unit on a daily basis. Exclusion criteria included a current or past history of renal failure, chronic obstructive pulmonary disease with cor pulmonale, left ventricular hypertrophy on ECG, arrhythmias, cardiomyopathy, rheumatic fever or valvular heart disease. According to the clinical presentation, any patients with past history of cardiac failure or radiographic evidence on admission of cardiac failure or a diagnosis of cardiac failure by the admitting doctor were also excluded, as these patients are known to have raised BNP levels.
Data collection
Having received Caldicott Guardian approval, data were collected from the case notes using a standard data collection sheet and included demographic variables, the characteristics of the chest pain, associated symptoms, cardiac risk factors, significant past medical history (e.g. previous ischaemic heart disease or any exclusion criteria), treatment received and the results of initial investigations including blood results, an ECG and a chest X-ray. Subsequent interventions, the results of definitive investigations and the discharge diagnosis were collected by reviewing the medical notes post-discharge.
Biochemical analysis
Blood samples were taken for TnT measurement in accordance with the Tayside protocol for the assessment of chest pain. After analysis the serum from these blood samples was stored in a refrigerator at 4°C. Within 3 days, 0.5 ml of the remaining serum was aliquoted for the measurement of BNP. BNP was measured on a Triage Meter PlusTM (Biosite, San Diego, California, USA). The laboratory personnel measuring the BNP were blinded to all the clinical characteristics and the TnT result.
Outcome measures
The main outcome was whether or not the patients had a final diagnosis of cardiac chest pain. Cardiac diagnoses included stable angina, acute coronary syndrome (with myocyte necrosis or with unstable angina) or acute myocardial infarction. Patients with alternative diagnoses were classified as having non-cardiac chest pain.
Statistical analysis
Receiver Operating Characteristic (ROC) curves were constructed for BNP, TnT and for a composite of BNP and TnT. In accordance with the British Cardiac Society guidelines, a TnT cutoff value of 0.01 ng/ml was used to define an abnormal test.18 The BNP cutoff value was determined from the BNP ROC curve. For the composite of BNP and TnT, the test was deemed positive if either one of the markers was elevated above its cutoff value. The sensitivity, specificity, positive predictive accuracy (PPA) and negative predictive accuracy (NPA) were all calculated for BNP, TnT and for the composite of BNP and TnT. Differences between the sensitivity and specificity of BNP, TnT and the composite of BNP and TnT were calculated with 95% confidence intervals. McNemar's test was performed to compare these sensitivities and specificities of BNP, TnT and the composite of BNP and TnT.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
One hundred patients with chest pain were included in the study. The background information including demographics, cardiac risk factors and medications on admission is shown in Table 1. The patients’ outcome and breakdown of diagnosis are shown in Figure 1 and Table 2, respectively.
|
|
|
For BNP levels above 10 pg/ml, the levels of BNP measured in patients with cardiac chest pain were consistently higher than in patients with non-cardiac chest pain (Figure 2).
|
The mean BNP concentration for patients with cardiac chest pain was 36.50 pg/ml. This was significantly greater than the mean BNP concentration for patients with non-cardiac chest pain which was only 7.88 pg/ml (P
0.0001). The mean BNP concentration for patients diagnosed as having myocardial infarction was 55.14 pg/ml in comparison to 22.14 pg/ml for patients diagnosed as having stable angina (Table 3).
|
The AUC constructed for BNP, TnT and the composite of BNP and TnT were 0.858, 0.772 and 0.919, respectively (Figure 3). The AUC for the composite of BNP and TnT was significantly greater than the AUC for TnT (P = 0.012).
|
Due to the potential consequences of missing someone with an acute coronary syndrome, a value that maximized BNP's sensitivity for detecting ischaemic heart disease was considered essential. A BNP value of 5 pg/ml was determined from the BNP ROC curve. The resulting sensitivities, specificities, positive predictive accuracies and negative predictive accuracies along with 95% CIs are shown in Table 4.
|
As a single marker, BNP was significantly more sensitive than TnT for detecting cardiac chest pain. The sensitivity was 88.6% with BNP compared to 56.8% with TnT [difference 31.8% (95% CI 13.5–48.7%), P = 0.003]. The addition of BNP to TnT increased the sensitivity from 56.8% with TnT alone to 90.0% with BNP and TnT in combination [difference 36.8% (95% CI 21.5–54.0%), P
0.0001].
The increase in sensitivity observed with BNP was associated with a decrease in specificity. TnT was significantly more specific than BNP. The specificity with TnT was 98.2% compared to 76.8% with BNP [difference 21.4% (95% CI 11.6–35.7%), P 0.0001]. TnT was also significantly more specific than the composite of BNP and TnT. The specificity with TnT was 98.2% compared to 76.8% with the composite of BNP and TnT [difference 21.4% (95% CI 11.6–35.7%), p 0.0001].
BNP increased the NPA but decreased the PPA for detecting cardiac chest pain. The NPA with BNP alone was 89.6% (95% CI 75.3%–94.7%) compared to 74.3% (95% CI 62.0%–82.2%) with TnT. This NPA with TnT alone increased to 95.6% (95% CI 82.6–98.2%) with the addition of BNP.
The PPA with BNP alone was 75.0% (95% CI 60.6–84.1%) compared to 96.2% (95% CI 82.5–99.8%) with TnT alone. This PPA with TnT decreased to 76.4% (95% CI 62.5–85.0%) with the addition of BNP.
If BNP and TnT were both measured in patients with chest pain, then a normal result in both tests would lead to 43/45 (95.6%) of such cases being safely discharged home with huge cost savings although at the price of 2/45 (4.4%) of those discharged turning out to have a cardiac cause for their chest pain (Table 5). Of these two patients, one has been lost to follow-up and the other had no further admissions into hospital for chest pain in the following year.
|
As BNP naturally rises with age,19 we performed a regression analysis that showed that BNP still predicts the final diagnosis independent of a potentially confounding variable such as age (P < 001).
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
This study has shown that BNP is significantly more sensitive for detecting cardiac chest pain than TnT and markedly increases the negative predictive accuracy for diagnosis of chest pain. This supports the theory that BNP levels are elevated by myocardial ischaemia and that as a result, it has the potential to assist in deciding if acute chest pain is of cardiac origin or not.11.
This increase in sensitivity and negative predictive accuracy associated with the use of BNP for detecting ischaemic heart disease is also demonstrated by a number of other studies. Bassan et al.17 was the first to investigate the diagnostic potential of admission BNP by assessing whether the addition of BNP to troponin I and creatine kinase MB increased the sensitivity for detecting myocardial infarction. Using a BNP cutoff of 100 pg/ml, they found the sensitivity of BNP (70.8%) to be significantly higher than troponin I (50.7%).
Funck et al.20 examined the diagnostic role of BNP in patients with unstable angina. Although the authors did not specify heart failure as an exclusion criterion, the authors reported a negative predictive value of 90% for a BNP of <10 pg/ml. The association of a BNP <10 pg/ml and a normal Troponin I increased the negative predictive value to 94%. Our study extends these observations from using BNP to detect infarcted or threatened myocardium in acute cases to its ability to detect all cardiac causes of acute chest pain.
Brown et al.21 recently investigated whether the addition of BNP to troponin I, creatine kinase-MB and myoglobin increased the sensitivity for detecting not only acute myocardial infarctions but also acute coronary syndromes and 30-day adverse events in patients with potential acute coronary syndrome. For all three of these outcomes, the addition of BNP increased the sensitivity and negative predictive accuracy. However, they did not exclude patients with heart failure who are known to have a high BNP, Furthermore, in the non-heart failure patients, only 18% had a definite cardiac cause of their chest pain. This may not be representative of the UK acute chest pain cases. Our study extends this to the key clinical question of whether BNP could be used to safely discharge otherwise unselected patients presenting with chest pain who were felt to require at least an overnight stay in hospital.
This recent evidence combined with the results of our investigation supports the possible use of BNP as a ‘rule out’ test for patients presenting with potential acute coronary syndromes. A low BNP level on admission could well be used to establish that a patient does not have ischaemic chest pain and hence could be used to safely discharge them from hospital earlier, and even from the Emergency Medicine department itself.
The significant increase in sensitivity and marked increase in negative predictive accuracy associated with the use of BNP on its own for detecting ischaemic heart disease observed in this study came at the cost of a reduced positive predictive accuracy and a significantly reduced specificity. It should also be noted at this point that the low specificity observed for BNP was obtained despite exclusion criterion that excluded many other conditions, which are known to raise BNP levels. In clinical practice, BNP levels may be raised secondary to mild congestive heart failure or tachyarrhythmias, which may present in a similar way to acute coronary syndrome thus further reducing its specificity.7 This substantial reduction in specificity would not necessarily be a major limitation for using BNP as a diagnostic test for ischaemic heart disease in practice because clinical examination should readily detect these confounding diseases and using TnT in conjunction with this would improve the diagnostic accuracy of BNP.
Numerous studies have demonstrated that BNP levels on admission provide important prognostic information in patients admitted with suspected acute coronary syndrome.16,22 Hence, admission BNP could be used not only diagnostically but also prognostically. This is particularly useful because it infers that the very few cardiac patients in our study with a normal BNP and TnT are likely to have a good prognosis even if their ultimate diagnosis is confirmed as cardiac in origin.
Study limitations
This study was carried out to investigate the assessment of patients admitted with chest pain. However, instead of a formal scoring system, we use an overall clinical impression supplemented by appropriate tests where necessary. Therefore some patients were discharged before definitive investigations, as these were deemed clinically unnecessary. Therefore, in theory, some patients with ischaemic heart disease may have been missed and this could have increased the number of false negatives observed in this study. However, our study was ‘piggy-backed’ onto normal clinical practice, which is potentially a strength rather than a weakness.
There were a large number of patients with non-cardiac chest pain and BNP levels of <5 pg/ml which resulted in a high NPA value. This value would be lower in a population with higher median BNP values.
As BNP is not stored in the myocardium but rather expressed secondary to DNA transcription over several hours,23 a single BNP measurement may not detect the ischaemic origin of symptoms in patients presenting early after an ischaemic event. We were unable to collect information on what happened to all of the patients after discharge and therefore were unable to confirm BNP as a prognostic marker. Such data has been previously published by various groups.24,25
|
Conclusion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
In conclusion, in patients with acute chest pain, we have shown that BNP performs well to rule-out a cardiac cause. This is particularly so when BNP is used in conjunction with TnT. Such an approach could potentially lead to the safe early discharge of a large number of patients who would otherwise be admitted overnight.
|
Acknowledgements |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
The authors would like to thank Jack McIntosh and Lesley McFarlane for carrying laboratory assistance and Dr Simon Ogston for statistical advice.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
1. Clancy M. Chest pain units. Br Med J (2002) 325:116–7.
2. Blatchford O, Capewell S, Murray S, Blatchford M. Emergency medical admissions in Glasgow: general practices vary despite adjustment for age, sex, and deprivation. Br J Gen Pract (1999) 49:551–4.[Web of Science][Medline]
3. Capewell S, McMurray J. "Chest pain-please admit": is there an alternative? A rapid cardiological assessment service may prevent unnecessary admissions. Br Med J (2000) 320:951–2.
4. McCarthy BD, Beshansky JR, D’Agostino RB, Selker HP. Missed diagnoses of acute myocardial infarction in the emergency department: results from a multicenter study. Ann Emerg Med (1993) 22:579–82.[CrossRef][Web of Science][Medline]
5. Pope JH, Aufderheide TP, Ruthazer R, Woolard RH, Feldman JA, Beshansky JR, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med (2000) 342:1163–70.
6. Collinson PO, Premachandram S, Hashemi K. Prospective audit of incidence of prognostically important myocardial damage in patients discharged from emergency department. Br Med J (2000) 320:1702–5.
7. Cowie MR, Jourdain P, Maisel A, Dahlstrom U, Follath F, Isnard R, et al. Clinical applications of B-type natriuretic peptide (BNP) testing. Eur Heart J (2003) 24:1710–8.
8. Cameron SJ, Green GB, White CN, Laterza OF, Clarke W, Kim H, et al. Assessment of BNP and NT-proBNP in emergency department patients presenting with suspected acute coronary syndromes. Clin Biochem (2006) 39:11–8.[Web of Science][Medline]
9. Excellence NIoC. Management of chronic heart failure in adults in primary and secondary care. In: NICE Guideline No 5 (2003) July.
10. Network SIG. SIGN 95: Management of Chronic Heart Failure. (2007).
11. Struthers AD, Davies J. B-type natriuretic peptide: a simple new test to identify coronary artery disease? QJM (2005) 98:765–9.
12. Bibbins-Domingo K, Ansari M, Schiller NB, Massie B, Whooley MA. B-type natriuretic peptide and ischemia in patients with stable coronary disease: data from the Heart and Soul study. Circulation. (2003) 108:2987–92.
13. Davidson NC, Pringle SD, Pringle TH, McNeill GP, Struthers AD. Right coronary artery stenosis is associated with impaired cardiac endocrine function during exercise. Eur Heart J (1997) 18:1749–54.
14. Goetze JP, Christoffersen C, Perko M, Arendrup H, Rehfeld JF, Kastrup J, et al. Increased cardiac BNP expression associated with myocardial ischemia. FASEB J (2003) 17:1105–7.
15. Sadanandan S, Cannon CP, Chekuri K, Murphy SA, Dibattiste PM, Morrow DA, et al. Association of elevated B-type natriuretic peptide levels with angiographic findings among patients with unstable angina and non-ST-segment elevation myocardial infarction. J Am Coll Cardiol (2004) 44:564–8.
16. de Lemos JA, Morrow DA, Bentley JH, Omland T, Sabatine MS, McCabe CH, et al. The prognostic value of B-type natriuretic peptide in patients with acute coronary syndromes. N Engl J Med (2001) 345:1014–21.
17. Bassan R, Potsch A, Maisel A, Tura B, Villacorta H, Nogueira MV, et al. B-type natriuretic peptide: a novel early blood marker of acute myocardial infarction in patients with chest pain and no ST-segment elevation. Eur Heart J (2005) 26:234–40.
18. Fox KA, Birkhead J, Wilcox R, Knight C, Barth J. British Cardiac Society Working Group on the definition of myocardial infarction. Heart (2004) 90:603–9.
19. Munagala VK, Burnett JC Jr, Redfield MM. The natriuretic peptides in cardiovascular medicine. Curr Probl Cardiol (2004) 29:707–69.[CrossRef][Web of Science][Medline]
20. Funck F, Jourdain P, Guillard N, Fulla Y, Sadeg N, Bellorini M, et al. Utility of BNP measurement in the emergency room in patients with suspected unstable angina with a normal ECG. Arch Mal Coeur Vaiss (2003) 96:181–5.[Web of Science][Medline]
21. Brown AM, Sease KL, Robey JL, Shofer FS, Hollander JE. The impact of B-type natriuretic peptide in addition to troponin I, creatine kinase-MB, and myoglobin on the risk stratification of emergency department chest pain patients with potential acute coronary syndrome. Ann Emerg Med (2007) 49:153–63.[CrossRef][Web of Science][Medline]
22. de Lemos JA, Morrow DA, Gibson CM, Murphy SA, Sabatine MS, Rifai N, et al. The prognostic value of serum myoglobin in patients with non-ST-segment elevation acute coronary syndromes. Results from the TIMI 11B and TACTICS-TIMI 18 studies. J Am Coll Cardiol (2002) 40:238–44.
23. Hama N, Itoh H, Shirakami G, Nakagawa O, Suga S, Ogawa Y, et al. Rapid ventricular induction of brain natriuretic peptide gene expression in experimental acute myocardial infarction. (1995) 1558–64.
24. Kistorp C, Raymond I, Pedersen F, Gustafsson F, Faber J, Hildebrandt P. N-terminal pro-brain natriuretic peptide, C-reactive protein, and urinary albumin levels as predictors of mortality and cardiovascular events in older adults. JAMA (2005) 293:1609–16.
25. Kragelund C, Kistorp C, Pedersen F, Raymond I, Hildebrandt P. Biochemical cardiac risk markers in the general population, hypertension and coronary artery disease. Scand J Clin Lab Invest Suppl (2005) 240:138–42.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||