Value of natriuretic peptides in assessment of patients with possible new heart failure in primary care

THE LANCET 46 Florig KH. China’s air pollution risks. Environ Sci Tech 1997; 31: 274A-79A. 47 Murray CJL, Lopez AD. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Cambridge: Harvard University Press, 1996. 48 National Center for Health Statistics. Deaths and death rates for the 10 leading causes of death in specified age groups: report of final mortality statistics. Hyattsville, MD: NCHS, 1995. 49 New York Times. Southeast Asia chokes on Indonesia’s forest fires. September 25, 1997: 1. 50 Brimblecombe P. The Big Smoke. London: Routledge Chapman and Hall, 1987.

51 World Resources Institute. World Resources Report 1996-97: the urban environment. New York: Oxford University Press, 1996. 52 Saldiva PHN, Pope III CA, Schwartz J, et al. Air pollution and mortality in elderly people: a time series study in Sao Paulo, Brazil. Arch Environ Health 1995; 50: 159-64. 53 Katsouyanni K, Pershagen G. Ambient air pollution exposure and cancer. Cancer Causes Control 1997; 8: 284-91. 54 MacKenzie JJ. Climate protection and the national interest: the links among climate change, air pollution and energy security. Washington, DC: World Resources Institute, 1997. 55 McMichael AJ. Healthy world, healthy people. People Planet, 1997; 6: 6-9.

Value of natriuretic peptides in assessment of patients with possible new heart failure in primary care

Martin R Cowie, Allan D Struthers, David A Wood, Andrew J S Coats, Simon G Thompson, Philip A Poole-Wilson, George C Sutton

Summary Background The reliability of a clinical diagnosis of heart failure in primary care is poor. Concentrations of natriuretic peptides are high in heart failure. This population-based study examined the predictive value of natriuretic peptides in patients with a new primary-care diagnosis of heart failure. Methods Concentrations of plasma atrial (ANP and Nterminal ANP) and B-type (BNP) natriuretic peptides were measured by radioimmunoassay in 122 consecutive patients referred to a rapid-access heart-failure clinic with a new primary-care diagnosis of heart failure. On the basis of clinical assessment, chest radiography, and transthoracic echocardiography, a panel of three cardiologists decided that 35 (29%) patients met the case definition for new heart failure. ANP and NT-ANP results were available for 117 patients (34 with heart failure) and BNP results for 106 (29 with heart failure). Findings Geometric mean concentrations of natriuretic peptides were much higher in patients with heart failure than in those with other diagnoses (29·2 vs 12·4 pmol/L for ANP; 63·9 vs 13·9 pmol/L for BNP; 1187 vs 410·6 pmol/L for NT-ANP; all p<0·001). At cut-off values chosen to give negative predictive values for heart failure of 98% (ANP 肁18·1 pmol/L, NT-ANP 肁537·6 pmol/L, BNP 肁22·2 pmol/L), the sensitivity, specificity, and positive predictive value for ANP were 97%, 72%, and 55%; for NT-ANP 97%, 66%, and 54%; and for BNP 97%, 84%, and 70%. Addition of ANP or NT-ANP concentration or both did not improve

Cardiac Medicine, Imperial College School of Medicine, National Heart and Lung institute, London SW3 6LY, UK (M R Cowie MRCP, D A Wood FRCP, A J S Coats FRCP, P A Poole-Wilson FRCP); Department of Clinical Pharmacology and Therapeutics, Ninewells Hospital and Medical School, Dundee (A D Struthers FRCP); Department of Medical Statistics and Evaluation, Royal Postgraduate Medical School, London (S G Thompson DSc); and Hillingdon Hospital, Uxbridge, Middlesex (G C Sutton FRCP) Correspondence to: Dr Martin R Cowie

Vol 350 • November 8, 1997

the predictive power of a logistic regression model containing BNP concentration alone. Interpretation In patients with symptoms suspected by a general practitioner to be due to heart failure, plasma BNP concentration seems to be a useful indicator of which patients are likely to have heart failure and require further clinical assessment.

Lancet 1997; 350: 1347–51

Introduction Heart failure is commonly misdiagnosed, and the validity of the diagnosis in primary care is poor.1,2 The symptoms are non-specific and the clinical signs, although reasonably specific, are not at all sensitive. Consequently, even experienced physicians disagree on the diagnosis in individual cases, especially when the heart failure is mild.3 Natriuretic peptides are released in response to increased intracardiac volume or pressure.4 They have a natriuretic and vasodilatatory effect and suppress the renin-angiotensin-aldosterone system.5 A prohormone, stored in the atria, on release is cleaved into the active Cterminal atrial natriuretic peptide (ANP) and the inactive and less rapidly cleared N-terminal atrial natriuretic peptide (NT-ANP). B-type natriuretic peptide (BNP) is secreted mainly by the ventricle. The plasma concentration of these peptides is higher than normal in patients with heart failure and also, but to a lesser extent, in patients with symptomless cardiac impairment.6–11 Measurements of these peptides could therefore provide valuable information about underlying cardiac function.5 Previous work has focused on their role in detecting symptomless left-ventricular dysfunction after myocardial infarction,12,13 but little attention has been paid to the role of the natriuretic peptides in assessment of whether a patient’s symptoms are due to heart failure at the time of first presentation to primary care.14,15 Measurement of natriuretic peptides might indicate whether referral for further cardiological assessment is necessary. NT-ANP and BNP are more stable than ANP4,16 and therefore are more suitable for use in primary care. Both of these peptides seem to be more sensitive and specific 1349

THE LANCET

Methods Study population and identification of cases This study was part of the Hillingdon Heart Failure Study, which identified incident (new) cases of clinical heart failure developing in a population of 151 000 served by 81 general practitioners in 31 practices in Hillingdon District, west London.21 The general practitioners agreed to refer all suspected cases of new heart failure to a rapid-access study clinic, preferably before patients had started treatment. 122 patients were referred within the 15-month study period (April 6, 1995, to July 5, 1996). Ethical approval for this study was granted by the local health authority ethics committee. One cardiologist (MRC) took a standardised medical history and clinically examined all the patients. At the same visit, electrocardiography, chest radiography, and transthoracic echocardiography were done, and blood samples were drawn. The echocardiogram was done to a standard protocol and according to accepted guidelines22–24 by the same cardiologist or by one of two experienced cardiac technicians.

Case definition There is no gold standard for the diagnosis of heart failure, and there has been much variation in the diagnostic criteria used in previous studies. We adopted the criteria recommended by the Working Group on Heart Failure of the European Society of Cardiology.25 To meet the case definition of heart failure patients had to have appropriate symptoms (shortness of breath, fatigue, fluid retention, or any combination of these symptoms) with clinical signs of fluid retention (pulmonary or peripheral) in the presence of an underlying abnormality of cardiac structure and function. If an element of doubt remained, a beneficial response to therapy for heart failure (eg, a brisk diuresis accompanied by substantial improvement in breathlessness) was taken to confirm the diagnosis. All data collected (excluding the concentrations of natriuretic peptides) were presented to a panel of three cardiologists (AC, GCS, DAW) who determined on the basis of a majority decision whether the case definition had been met, and the aetiology. An underlying abnormality of cardiac structure or function was necessary to confirm a case as heart failure, but echocardiographic abnormalities were not sufficient in themselves to diagnose heart failure; patients had to satisfy the full case definition. Patients’ hospital and general-practitioner notes were checked to ensure that only those without a previous history of heart failure were included in the study. If a patient had been assessed at another hospital in the past, we contacted the physician in charge of his or her care at that hospital for details of the consultation and investigations undertaken.

Collection, storage, and assay of samples A 10 mL sample of venous blood was drawn from each patient after he or she had rested supine for 10 min. Blood was drawn into a tube containing edetic acid as an anticoagulant and aprotonin to prevent breakdown of the natriuretic hormones. The sample was centrifuged at 1700⫻ g for 15 min; the plasma was divided into several parts and stored at ⫺40°C before being transferred for longer-term storage to a ⫺80°C freezer. At the end of the study period, the samples were transported at ⫺40°C to the Department of Clinical Pharmacology and Therapeutics at Ninewells Hospital and Medical School, Dundee, for assays. ANP-like immunoreactivity, BNP-like-immunoreactivity, and NT-ANP-like immunoreactivity were measured in the samples in two assay runs. C8 solid-phase extraction columns were pretreated with 4 mL methanol, 4 mL distilled water, and 4 mL

1350

100 80 Sensitivity (%)

than ANP for detection of ventricular dysfunction.13,17–20 We report the utility of measurements of ANP, NT-ANP, and BNP in predicting the presence of heart failure in a consecutive series of 122 patients with a new primarycare diagnosis of heart failure.

60 40 BNP ANP NT-ANP CTR

20 0 0

10

20

30

40

50

60

70

1–specificity (%) Receiver-operating-characteristic curves for natriuretic peptides and cardiothoracic ratio (CTR) on posteroanterior chest radiography 1% trifluoroacetic acid. Plasma was applied and the columns were washed with 9 mL 1% trifluoroacetic acid, and samples eluted with 4 mL 95% methanol and 1% trifluoroacetic acid. After drying of the samples, radioimmunoassays for ANP, BNP, and NT-ANP were done with standard commercial kits (Peninsula Laboratories Europe Ltd, St Helens, Merseyside, UK). Recovery was 80% for ANP, 86% for BNP, and 78% for NT-ANP. The between-assay and within-assay coefficients of variability were 11·8% and 12·6% for ANP, 14·8% and 9·9% for BNP, and 11·7% and 7·7% for NT-ANP. The laboratory reference range is 2·4–10·5 pmol/L for ANP, 2·3–4·4 pmol/L for BNP, and 37·9–152·3 pmol/L for NT-ANP.

Data analysis The natriuretic peptide values showed a right skewed distribution, which was normalised by log transformation before the concentrations in patients with and without heart failure were compared by an unpaired t test. The sensitivity, specificity, and positive predictive value of each peptide were described for cutoff concentrations at which the negative predictive value was 98%. The receiver-operating-characteristic curves for the three peptides were drawn, and the area under the curves calculated. A logistic regression model with the log-transformed data was used to assess the independent contributions of the three natriuretic peptides in predicting the presence of heart failure.

Results 122 patients (59 male, 63 female; age range 24–87 years) were referred to the study clinic with a new diagnosis of heart failure suspected by the general practitioner. The median delay between the general practitioner’s suspicion of heart failure and the patient’s attendance at the clinic was 4 days (IQR 2–15). 38 of these patients were receiving long-term diuretic treatment for hypertension or long-standing ankle oedema. Of the remaining 84 patients, 26 (31%) were started on diuretic treatment by the general practitioners before referral to the clinic. Of the 122 patients referred, 35 (29%) were judged to satisfy the case definition for heart failure. Of the 87 with other diagnoses, 19 (22%) had chronic obstructive pulmonary disease, 17 (20%) obesity, 12 (14%) angina, seven (8%) venous insufficiency, seven (8%) anxiety, five (6%) palpitation/arrhythmias, two (2%) pulmonary fibrosis, two (2%) malignant disorders, and nine (10%) other diagnoses; the diagnosis was not known in the remaining seven. A higher proportion of men than of women referred to the clinic had heart failure (25/49 [41%] vs 11/63 [17%]; p<0·01). Among the 35 patients with heart failure, the Vol 350 • November 8, 1997

THE LANCET

Natriuretic peptide and cut-off value

Sensitivity (%)

ANP 肁18·1 pmol/L NT-ANP 肁537·6 pmol/L BNP 肁22·2 pmol/L

97 97 97

Specificity (%) 72 66 84

Positive predictive value (%) 55 54 70

Table 1: Diagnostic value of natriuretic peptides in detecting heart failure when negative predictive value is 98%

aetiology was coronary artery disease in seven, hypertension in six, atrial fibrillation in five, valve disease in three, complete heart block in two, chronic alcohol misuse in one, and unknown in 11. Ten of these patients were on long-standing diuretic therapy for hypertension or long-standing ankle swelling, and 12 of the remaining 25 patients had been given a diuretic by the general practitioner before they were seen in the clinic. 30 (86%) of the 35 cases had symptoms on mild to moderate exertion, and only five had symptoms at rest. No patient had a blood urea concentration greater than 20 mmol/L or serum creatinine greater than 240 µmol/L. Results for ANP and NT-ANP were available for 117 (95%; 34 meeting the case definition for heart failure) patients, but results for BNP were available for only 106 patients (86%; 29 patients meeting the case definition for heart failure) because of insufficient remaining sample and difficulties in elution. The geometric mean concentrations of natriuretic peptides were much higher in patients with heart failure than in those without heart failure (29·2 vs 12·4 pmol/L for ANP; 63·9 vs 13·9 pmol/L for BNP; and 1187 vs 410·6 for NT-ANP; all p<0·001). There was no significant association between the concentrations of the natriuretic peptides and age, sex, previous diuretic therapy (short-term or long-term), or aetiology of heart failure. The receiver-operating-characteristic curves for the natriuretic peptides are shown in the figure, with that for cardiothoracic ratio on posteroanterior chest radiography (available for 111 patients) for comparison. The area under the curve was greatest for BNP (0·96), compared with 0·93 for ANP and 0·89 for NT-ANP. This finding suggests that BNP combined higher sensitivity and higher specificity over a range of different cut-off values than the other peptides. The area under the curve for cardiothoracic ratio was 0·79. If natriuretic peptides are to be useful in aiding the clinical decision of a general practitioner to refer patients for further assessment and investigation, the cut-off value chosen should combine a very high negative predictive value (ie, a very high proportion of patients with concentrations below the cut-off value do not have heart failure) with a high positive predictive value (ie, a high proportion of patients with concentrations above the cutoff value do have heart failure). Table 1 compares the three peptides for the negative predictive value of 98%. In this study, BNP appeared the most useful with a higher positive predictive value than the other peptides; 70% (28/40) of patients with a plasma BNP concentration of 22·2 pmol/L or higher had heart failure compared with 55% (29/53) of patients above the cut-off for ANP and 54% (33/61) for NT-ANP. To calculate the predictive values of the peptides independently of each other, we added ANP and NTANP concentrations in turn to a logistic regression model including only BNP concentration, with the presence of heart failure as the dependent variable. Neither ANP nor NT-ANP added any independent predictive power to

Vol 350 • November 8, 1997

Analysis 1 BNP Analysis 2 BNP ANP

Odds ratio for heart failure (95% CI)*

p

11·2 (4·41–28·6)

<0·001

8·39 (3·09–22·7) 2·27 (0·69–7·44)

<0·001 0·18

Analysis 3 BNP NT-ANP

9·21 (3·04–27·9) 1·39 (0·47–4·12)

<0·001 0·55

Analysis 4 BNP NT-ANP ANP

7·97 (2·58–24·6) 1·11 (0·36–3·42) 2·19 (0·64–7·55)

<0·001 0·86 0·21

29 of these 106 patients had heart failure; 77 did not. *Odds of heart failure per doubling in concentration of peptide.

Table 2: Prediction of presence of heart failure in logistic regression analyses of 106 referred patients for whom concentrations of all three natriuretic peptides were available

that provided by BNP alone (table 2). The size of the heart on chest radiograph is often used clinically in the process of deciding whether heart failure is likely to be present in a patient with appropriate symptoms, especially in general practice. In a logistic regression model including both cardiothoracic ratio and BNP concentration, only the BNP concentration was independently predictive of the presence of heart failure (table 3).

Discussion Heart failure is difficult to diagnose in primary care. Our finding that fewer than 30% of referred patients had heart failure accords with previous studies in Finland1 and Scotland,2 which suggested that only 25–50% of patients with a primary-care diagnosis of heart failure have evidence of this disorder on further cardiological assessment. The false-positive rate was higher in women than in men in our study and in the Finnish study.1 Without further assessment (such as echocardiography), more than 50% of patients given a primary-care diagnosis of heart failure will receive inappropriate (and often expensive) treatment to no benefit. In routine clinical practice not all patients are referred for such further assessment,26–28 and if they were, many centres would not be able to cope with the large number of extra referrals (estimated at 160 new patients per year for the average district hospital serving a population of 250 000). BNP and NT-ANP are stable in whole blood at room temperature for many hours. Blood samples for measurement of these peptides could therefore be taken in general practice and transported to the local hospital laboratory without any special handling.4,16 The radioimmunoassays are cheap and simple,29 and this approach may be an effective method of rationalising the referral of patients with symptoms possibly due to heart Odds ratio for heart failure (95% CI)* CTR alone BNP alone Combined analysis CTR BNP

6·06 (2·44–15·0) 10·3 (4·04–26·2) 2·65 (0·67–10·8) 8·34 (3·21–21·7)

p <0·001 <0·001 <0·17 <0·001

25 of these 96 patients had heart failure; 71 did not. *Odds of heart failure per doubling in concentration of BNP or increase in cardiothoracic ratio by 0·10 (eg, from 0·50 to 0·60).

Table 3: Prediction of presence of heart failure in logistic regression analyses of 96 referred patients for whom cardiothoracic ratio (CTR) and BNP concentration were available

1351

THE LANCET

failure. A study in a group of elderly people found that plasma ANP concentrations were high in those who subsequently developed heart failure.30 NT-ANP had a high sensitivity in detecting ventricular dysfunction in a series of patients referred for radionuclide angiography, with higher values in patients with heart failure than in those with symptomless ventricular dysfunction.17 A study in patients admitted urgently to hospital with acute breathlessness suggested that measurement of BNP can distinguish heart failure and lung disease as causes of breathlessness.31 In a population-based study of incident heart failure similar to our study (but among people aged 45–74 years only) in eastern Finland, ANP concentrations were significantly higher in 19 patients with heart failure than in the 26 who were unlikely to have heart failure.14 Our study strengthens the evidence, with BNP showing the best characteristics for a screening test. Our findings are directly relevant to patients who present with symptoms possibly due to heart failure in primary care: the study patients were drawn from general practice and were not already undergoing investigation in hospital; although some patients were on diuretic treatment for hypertension, most had not started treatment for heart failure; and the final diagnosis of heart failure was not based merely on a measure of underlying cardiac function such as ejection fraction. The simple blood test of BNP could be used to screen out patients in whom heart failure is extremely unlikely from those in whom the probability of heart failure is high. For patients with low BNP concentrations, the heart is unlikely to be the cause of the symptoms, but those with higher concentrations should be referred for further assessment. Hospitals will benefit from the lower rate of inappropriate referrals. Confirmation of the value of BNP (and the other peptides) in this role in other populations is necessary. The diagnostic criteria for heart failure in this series were strict, and may have excluded a few patients with very mild symptoms and signs. Since there is a historical interest in heart failure within the study district,27 the local general practitioners may be better informed about heart failure than those elsewhere. This issue is important, since the predictive value of any test is related to the prevalence of the disorder among the population being screened. In a district where general practitioners are less able to diagnose heart failure correctly, the predictive value of BNP may be lower. However, the rate of false-positive diagnoses in our study is similar to that reported from other centres.32 We do not propose that BNP should be used as a diagnostic tool in its own right. Much individual misclassification would result, with too many falsepositive diagnoses of heart failure still being made. Furthermore, certain groups of patients presenting with breathlessness and fatigue should be referred for further cardiological assessment irrespective of the BNP concentration—eg, patients presenting with a murmur or breathlessness after a myocardial infarction. The blood test would not be a substitute for clinical judgment. However, our findings suggest that the inexpensive assay of natriuretic peptide concentrations may be a useful addition to the tests available to a general practitioner to help in decisions about referral and further cardiological assessment. 1352

Contributors All the investigators contributed to the design of this study, the funding application, and the writing of the paper. Martie Cowie led the project, drafted and amended the paper, and analysed the data in conjunction with Simon Thompson.

Acknowledgments This study, undertaken by the Clinical Epidemiology Group in Cardiac Medicine, was supported by the British Heart Foundation and the Wellcome Trust (M R Cowie). We thank Helen Penston (research nurse, Hillingdon Hospital), Wendy Coutie and Lesley McFarlane (Clinical Pharmacology, Ninewells Hospital and Medical School, Dundee) for technical support, and the patients and general practitioners who made this study possible.

References 1

2

3

4 5 6

7 8 9 10

11

12

13

14 15

16

17

18

19

20

21

22

Remes J, Miettinen H, Reunanen A, Pyorala K. Validity of clinical diagnosis of heart failure in primary health care. Eur Heart J 1991; 12: 315–21. Wheeldon NM, MacDonald TM, Flucker CJ, McKendrick AD, McDevitt DG, Struthers AD. Echocardiography in chronic heart failure in the community. Q J Med 1993; 86: 17–20. Hlatky MA, Fleg JL, Hinton PC, et al. Physician practice in the management of congestive heart failure. J Am Coll Cardiol 1986; 8: 966–70. Struthers AD. Prospects for using a blood sample in the diagnosis of heart failure. Q J Med 1995; 88: 303–06. Struthers AD. Ten years of natriuretic peptide research: a new dawn for their diagnostic and therapeutic use? BMJ 1994; 308: 1615–19. Francis GS, Benedict C, Johnstone DE, et al. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure: a substudy of the studies of left ventricular dysfunction (SOLVD). Circulation 1990; 82: 1724–29. Richards AM, Cleland JGF, Tonolo G, et al. Plasma ␣ natriuretic pepetide in cardiac impairment. BMJ 1986; 293: 409–12. Cody RJ, Atlas SA, Laragh JH, et al. Atrial natriuretic factor in human subjects and heart failure patients. J Clin Invest 1986; 78: 1362–74. Burnett JC, Kao PC, Hu DC, et al. Atrial natriuretic peptide elevation in congestive heart failure in the human. Science 1986; 231: 1145–47. Mukoyama M, Nakao K, Hosoda K, et al. Brain natriuretic peptides as a novel cardiac hormone in humans. J Clin Invest 1991; 87: 1402–12. Richards AM, Crozier IG, Yandle TG, Espiner EA, Ikram H, Nicholls MG. Brain natriuretic factor: regional plasma concentrations and correlations with haemodynamic state in cardiac disease. Br Heart J 1993; 69: 414–17. Motwani JG, McAlpine H, Kennedy N, Struthers AD. Plasma brain natriuretic peptide as an indicator for angiotensin-converting-enzyme inhibition after myocardial infarction. Lancet 1993; 341: 1109–13. Choy AJ, Darbar D, Lang CC, et al. Detection of left ventricular dysfunction after acute myocardial infarction: comparison of clinical, echocardiographic, and neurohormonal methods. Br Heart J 1994; 72: 16–22. Remes J, Tikkanen I, Fyhrquist F, Pyorala K. Neuroendocrine activity in untreated heart failure. Br Heart J 1991; 65: 249–55. Clarkson PBM, Macleod C, Coutie W, MacDonald TM. Use of natriuretic peptide assays in the assessment of chronic dyspnoea. International Meeting of the Working Group on Heart Failure, Amsterdam, April 1995. European Society of Cardiology. Hall C, Aaberge L, Stokke O. In vitro stability of N-terminal proatrial natriuretic factor in unfrozen samples: an important prerequisite for its use as a biochemical parameter of atrial pressure in clinical routine. Circulation 1995; 91: 911. Lerman A, Gibbons RJ, Rodeheffer RJ, et al. Circulating N-terminal ANP as a marker for symptomless left ventricular dysfunction. Lancet 1993; 341: 1105–08. Kettunen RVJ, Leppaluoto J, Jounela A, Vuolteenaho O. Plasma Nterminal atrial natriuretic peptide in acute myocardial infarction. Am Heart J 1994; 127: 1449–55. Omland T, Aakvaag A, Vik-Mo H. Plasma cardiac natriuretic peptide determination as a screening test for the detection of patients with mild left ventricular impairment. Heart 1996; 76: 232–37. Davidson NC, Naas AA, Hanson JK, Kennedy NSJ, Coutie WJ, Struthers AD. Comparison of atrial natriuretic peptide, and Nterminal proatrial natriuretic peptide as indicators of left ventricular systolic dysfunction. Am J Cardiol 1996; 77: 828–31. Cowie MR, Penston H, Wood DA, et al. A population survey of the incidence and aetiology of heart failure. Eur Heart J 1996; 17: 131 (abstr). Devereux RB, Liebson PR, Horan MJ. Recommendations concerning use of echocardiography in hypertension and general population research. Hypertension 1987; 9: 97–104.

Vol 350 • November 8, 1997

THE LANCET 23 Sahn DJ, DeMaria A, Kisslo J, Weyman A. The Committee on M-mode standardization of the American Society of chocardiography: recommendations on quantitation in M-mode echocardiographic measurements. Circulation 1978; 58: 1072–82. 24 Henry WL, DeMaria A, Gramiak R, et al. Report of the American Society of Echocardiography Committee on Nomenclature and Standards in two-dimensional echocardiography. Circulation 1980; 62: 212–17. 25 The Task Force on Heart Failure of the European Society of Cardiology. Guidelines for the diagnosis of heart failure. Eur Heart J 1995; 16: 741–51. 26 Clarke KW, Gray D, Hampton JR. Evidence of inadequate investigation and treatment of patients with heart failure. Br Heart J 1994; 71: 584–87. 27 Parameshwar J, Shackell MM, Richardson A, Poole-Wilson PA, Sutton GC. Prevalence of heart failure in three general practices in

north west London. Br J Gen Pract 1992; 42: 287–89. 28 Mair FS, Crowley TS, Bundred PE. Prevalence, aetiology and management of heart failure in general practice. Br J Gen Pract 1996; 46: 77–79. 29 Struthers AD. Plasma concentrations of brain natriuretic peptide: will this new test reduce the need for cardiac investigations? Br Heart J 1993; 70: 397–98. 30 Davis KM, Fish LC, Elahi D, Clark BA, Minaker KL. Atrial natriuretic peptide levels in the prediction of congestive heart failure in frail elderly. JAMA 1992; 267: 2625–29. 31 Davis M, Espiner E, Richards G, et al. Plasma brain natriuretic peptide in assessment of acute dyspnoea. Lancet 1994; 343: 440–44. 32 Colquhoun MC, Waine C, Monaghan MJ, Struthers AD, Mills PG. Investigation in general practice of patients with suspected heart failure: how should the essential echocardiographic service be delivered? Br J Gen Pract 1995; 45: 517–19.

Randomised trial of epidural bupivacaine and morphine in prevention of stump and phantom pain in lower-limb amputation

Lone Nikolajsen, Susanne Ilkjaer, Jørgen H Christensen, Karsten Krøner, Troels S Jensen

Summary Background Epidural analgesia before limb amputation is commonly used to reduce postamputation pain. But there have been no controlled studies with large numbers of patients to prove such a pre-emptive effect. We investigated whether postamputation stump and phantom pain in the first year is reduced by preoperative epidural blockade with bupivacaine and morphine. Methods In a randomised, double-blind trial, 60 patients scheduled for lower-limb amputation were randomly assigned epidural bupivacaine (0·25% 4–7 mL/h) and morphine (0·16–0·28 mg/h) for 18 h before and during the operation (29 patients; blockade group) or epidural saline (4–7 mL/h) and oral or intramuscular morphine (31 patients; control group). All patients had general anaesthesia for the amputation and were asked about stump and phantom pain after 1 week and then after 3, 6, and 12 months by two independent examiners. Study endpoints were rate of stump and phantom pain, intensity of stump and phantom pain, and consumption of opioids. Findings Two patients in each group were withdrawn before amputation. The groups were well matched in baseline characteristics. Median duration of preoperative saline treatment was 18·5 h (IQR 17–20). Median duration of preoperative epidural blockade in the blockade group was 18 h (15–20·3). The combined median duration of postoperative epidural pain treatment in both groups was 166 h (89·3–308·3). After 1 week, 14 (52%) patients in the blockade group and 15 (56%) in the control group had phantom pain (95% CI ⫺30·6 to 22·7, p=0·9). The figures Departments of Neurology (L Nikolajsen MD, Prof T S Jensen MD) and Anaesthesiology (S Ilkjaer MD), Aarhus University Hospital, and Danish Pain Research Center (L Nikolajsen, K Kroner MD, Prof T S Jensen), University of Aarhus, DK-8000 Aarhus C, Denmark Correspondence to: Dr Lone Nikolajsen, Danish Pain Research Center, Bygning IC, Aarius Kommunehospital, 8000 Aarhus C, Denmark

Vol 350 • November 8, 1997

for blockade versus control group were: 14 (82%) vs ten (50%; 4·0 to 60·8, p=0·09) at 3 months; 13 (81%) vs 11 (55%; ⫺2·7 to 55·3, p=0·2) at 6 months; and nine (75%) vs 11 (69%; ⫺27·0 to 39·6, p=1·0) at 12 months. Intensity of stump and phantom pain and consumption of opioids were similar in both groups at all four postoperative interviews. Interpretation Perioperative epidural blockade started a median of 18 h (15–20·3) before the amputation and continued into the postoperative period does not prevent phantom or stump pain.

Lancet 1997; 350: 1353–57 See Commentary page 1264

Introduction Epidural administration of local anaesthetics and opioids, which is effective to treat postoperative pain, has also been introduced before the operation to prevent postoperative pain.1–6 The observations that severe preamputation pain is associated with the development of phantom pain7,8 prompted Bach and colleagues9 to carry out a controlled study in which patients scheduled for amputation received epidural bupivacaine, morphine, or both for 3 days before the amputation or conventional pain treatment. After 6 months, the rate of phantom pain was significantly lower among 11 patients on epidural treatment for pain than among 14 patients on conventional pain treatment. This finding corroborated and, in turn, stimulated a series of experimental and clinical studies on the effect of analgesic treatment before a surgical incision for the subsequent development of central hyperexcitability and postoperative pain.1,2,10–20,46 Subsequent clinical trials have confirmed Bach and coworkers’ study.21–23 Trials of the effect of postoperative perineural or intraneural block on phantom pain have found different results.24-26 However, design flaws, such as small sample sizes, no or insufficient randomisation, and non-blinded assessment of treatment and pain, limit the validity of these studies. 1353