Cowie's editorial envisages a clinical role for measurement of brain natriuretic peptide (BNP) in the near future.[1] However, while the potential for risk stratification and monitoring of treatment in patients with heart failure is encouraging, we would urge caution regarding the diagnostic utility of plasma BNP concentration.

There is little doubt that the diagnosis of heart failure is difficult - only about a third of patients referred for echocardiography or to a cardiology clinic with suspected heart failure turn out to have left ventricular systolic dysfunction.[2, 3] Nor is there any doubt that a definitive diagnosis is important, as the institution of appropriate treatments may have a significant impact on survival as well as symptoms, while exclusion of left ventricular systolic dysfunction as a cause of symptoms allows other investigations and treatment to be planned. However, many echocardiography departments are becoming overwhelmed. Thus a simple test that can "rule in" or "rule out" left ventricular systolic dysfunction has considerable appeal.

In the assessment of any diagnostic test, it is useful to consider both the properties of the test (sensitivity and specificity) and the baseline probability of the diagnosis (ie, the probability before the test result is known). The use of decision analysis, based upon Bayesian statistics is invaluable in this regard and is recommended in standard textbooks of evidence-based medicine.[4]

Plasma natriuretic peptides are highly sensitive (over 90%) but only moderately specific (approximately 60%) tests for left ventricular systolic dysfunction.[5] As a consequence low plasma concentrations of natriuretic peptides will be helpful in "ruling out" the diagnosis of left ventricular systolic dysfunction but high concentrations will be of little value as they may be associated with a number of other diagnoses.[1] Hence, while plasma BNP concentrations will be very helpful in screening large numbers of patients where the overall prevalence of left ventricular systolic dysfunction is low,[5] their use in the investigation of patients with suspected heart failure will be much more limited.

In a study of patients referred to outpatients with suspected heart failure, we found that 15% of those with a normal plasma brain natriuretic peptide concentration had left ventricular systolic dysfunction.[3]

A further important question is whether BNP provides additional information aborve and beyond that given by current investigations. In our study, the use of plasma BNP concentration did not add significantly to more simple "tests" - a history of previous myocardial infarction and an electrocardiogram.[3] While history and an electrocardiogram are mainstays of the investigation of the breathless patient, a false negative rate for plasma BNP concentration of approximately 15% is too high to obviate the need for echocardiography in such individuals.

We believe that in the setting of patients with "suspected heart failure", an echocardiogram rather than plasma brain natriuretic peptide concentration remains the investigation of choice. Despite the exciting potential for estimating prognosis and adjusting treatment, there is currently no role for measurement of brain natriuretic peptide in routine clinical practice.

References

(1) Cowie MR. BNP: soon to become a routine measure in the care of patients with heart failure? Heart 2000:83:617-18.

(2) Francis CM, Caruana L, Kearney P, et al. Open access echocardiography in management of heart failure in the community. BMJ 1995;310:634-6.

(3) Landray MJ, Lehman R, Arnold IR. Measuring brain natriuretic peptide in suspected left ventricular systolic dysfunction in general practice: cross-sectional study. BMJ 2000;320:985-6.

(4) Sackett DL, Richardson WS, Rosenberg W, Haynes RB. Evidence-based medicine. How to practice and teach EBM. Churchill Livingstone 1998:118-28.

(5) McDonagh TA, Robb SD, Murdoch DR, et al. Biochemical detection of left- ventricular systolic dysfunction. Lancet 1998;351:9-13.

Martin J Landray
Lecturer in Medicine
University of Birmingham
Birmingham B15 2TH, UK

Richard Lehman
General Practice Principal
Hightown Surgery
Banbury OX16 9DB, UK

Ian Arnold
Consultant Cardiologist
Horton Hospital
Oxford Radcliffe Hospitals NHS Trust
Banbury OX16 9AL, UK

The recent article by Ghuran and Nolan is a valuable review about cardiovascular effects of recreational drugs.[1] We would like to provide additional information, not included in the review, about the induction of changes in the QT interval by some recreational substances.

The long QT syndrome has been associated with the occurrence of ventricular tachyarrhythmias (torsades de pointes). Considering the substances of abuse, to our knowledge, at least cocaine, 3,4-methylenedioxymethamphetamine (MDMA, ecstasy) and alcohol[2] can induce changes in the QT/QTc interval. There are published clinical case reports describing this kind of change for MDMA,[3, 4] and for cocaine it has also been described the presence of torsades de pointes in a patient with idiopathic long QT syndrome.[5]

In addition we would like to present results of the ECG-recording obtained during a clinical trial where MDMA was administered. Eight healthy male recreational users of MDMA participated in a randomized, double blind, cross-over and placebo controlled study. All subjects received, in different experimental sessions, a single oral dose of placebo, MDMA 125 mg; MDMA 75 mg and d,l-amphetamine 40 mg. A 12 lead-electrocardiogram (50 mm/sec) was performed prior treatment (baseline measure) and 1, 2, 4, 8 and 24 hours after drug administration. Cycle length (RR) and QT interval were measured manually by the same blinded investigator. Both measures were performed in all 12 leads of each ECG recording. Corrected QT was calculated with Bazett's formula. Interlead QTc dispersion was determined as the longest minus the shortest QTc on the 12-lead ECG. Changes in the T wave and appearance of U waves were also considered. Individual changes after each treatment conditions were evaluated according to criteria suggested by the Committee for Proprietary Medicinal Products (CPMP).[6] Other details of the trial have been previously published.[7] The table shows the number of subjects (maximal 8) who fulfil any criteria for prolonged cardiac repolarisation:

Mean QTc values (ms; n = 8) for each administered substance
 

*Following CPMP criteria

The maximal effects were observed two hours after MDMA administration and returned to normal values at 4-8 hours. For amphetamine the maximal effects were observed lately, 8 to 24 hours after drug administration. Although the magnitude of mean QTc increase was not clinically relevant, individual data of some subjects might be interpreted as important "warning" signs of possible malignant arrhythmia in accordance to the CPMP recommendation. The mean QTc increase observed for MDMA is similar to that described after the administration of therapeutic doses of prescription drugs as terfenadine or cisapride in clinical trials.

The administration of a single oral dose of MDMA, in the range of those consumed recreationally, produce a slight increase of QTc. Although these changes are mild in magnitude, MDMA could enhance the risk of ventricular arrhytmia in subjects with genetic predisposition to present an acquired long QT syndrome, and/or when used in combination with other well known arrhythmogenic drugs.

Supported by grants FIS 97/1198, CIRIT 1997SGR00077, ISC-III-97/4344 and Plan Nacional sobre Drogas (Madrid).

Magí Farré
Marta Mas
Lluis Molina
Jordi Camí
Units of Pharmacology and Cardiology
IMIM-Hospital del Mar, UAB - UPF
Barcelona, Spain

References

(1) Ghuran A, Nolan J. Recreational drug misuse: issues for the cardiologist. Heart 2000;83:627-33.

(2) Rossinen J, Sinisalo J, Partanen J, et al. Efects of acute alcohol infusion on duration and dispersion of QT interval in male patients with coronary artery disease and in healthy controls. Clin Cardiol 1999;22:591-4.

(3) Maxwell DL, Polkey MI, Henry JA. Hyponatremia and catatonic stupor after taking "ecstasy". BMJ 1993;307:1399.

(4) Drake WM, Broadhurst PA. QT-interval prolongation with ecstasy. S Afr Med J 1996;86:180-1.

(5) Schrem SS, Belsky P, Schwartzman D, et al. Cocaine-induced torsades de pointes in a patient with the idiopathic long QT syndrome. Am Heart J 1990;120:980-4.

(6) Committee for Proprietary Medicinal Products (CPMP). The assessment of the potential for QT interval prolongation by non-cardiovascular medicinal products. London: Commission of the European Communities, 1997 (Available from URL: http://www.eudra.org/en_home.htm).

(7) Mas M, Farré M, de la Torre R, et al. Cardiovascular and neuroendocrine effects, and pharmacokinetics of MDMA in humans. J Pharmacol Exp Ther 1999;290:136-45.

The article by Viskin et al[1] adds some confusion to the definition of "the typical" electrocardiographic pattern associated with the syndrome of sudden death and right bundle branch delay described by Nava and Martini in 1988-1989,[2, 3] by Aihara in 1990, and by Brugada in 1992. As clearly documented,[3] three (and not one as discussed by Viskin et al) of six patients with aborted sudden death showed different degrees of right bundle branch block, left axis deviation, prolonged PR interval and ST segment elevation. Two more patients had mild isolated ST elevation with T inversion. The same polymorphic patterns were lately documented either with standard electrocardiography or after class 1C drugs (in subjects with normal ECG or minor ST-T abnormalities), either in Brugada's or other series.[4] Thus a typical sign of the syndrome is a fascinating but not confirmed statement.

Viskin probably wants to emphasise the first and the most popular electrocardiogram of the syndrome, which is the one published in 1988[1] and 1989 (patient 3 of figure 1 in reference 2), and reprocessed by Brugada on his web site.

This electrocardiogram shows an R' wave in V1 (incorrectly called J wave), very similar to an incomplete right bundle branch block, followed by a coved ST elevation. It must be remembered that the typical J wave is those recorded in leads II, III, aVF, V5-V6 in vagotonia, and other pathological conditions.[5] Thus R' in V1 is an atypical J wave, or something else.

Different from a right bundle branch delay, the atypical J wave seen in V1 shows neither a late and slurred S wave in V6, or the typical vectorcardiographic terminal loop.[1] Late potentials often characterise this pattern, but are absent in pure incomplete right bundle branch block. Thus the trace proposed by Viskin et al in figure 2 as the typical sign of the syndrome is, in reality, a typical incomplete right bundle branch block, while the trace presented in figure 3A closely resembles a true atypical J wave with coved ST elevation, and not a questionable sign.

This does not means that Viskin et al's patients do not have the syndrome, because this is not characterised by a single electrocardiographic pattern, but by polymorphic right bundle branch delays with ST elevation.[2] The latter can be the only basal abnormality, with the atypical J wave appearing spontaneously, or being induced after class 1C drugs.

The significance of all the electrocardiographic patterns of the syndrome, particularly of the atypical J wave and coved ST elevation, is strongly debated. Brugada favours a functional repolarisation abnormality, while we think, according to electrical and anatomic demonstration, that it represents an organic depolarisation abnormality.[4] This can be easily demonstrated by recording late potentials. As mentioned before, in patients with the discussed "typical" electrocardiogram late potentials are almost always present, clearly demonstrating that the atypical J wave and coved ST elevation is due to a delayed QRS activity, probably of the outflow tract of the right ventricle.[4]

Please contact Dr Martini for a copy of the figure described below (unfortunately we cannot support illustrations in Rapid Responses)
The atypical J wave with coved ST elevation in V1, is identical in these two subjects with the syndrome. The trace on the left is the patient of Brugada described in Heart[6] after ajmaline; this pattern was ascribed to a functional repolarisation abnormality. The trace on the right is that of an apparently normal patient, belonging to a family affected by a minor form of right ventricular cardiomyopathy. This patient has a very positive late potentials study, indicating that the atypical J wave and ST elevation, both spontaneous and drug inducible, represents a late depolarisation abnormality of the cardiac activity, probably at the right ventricular outflow tract level. A repolarisation abnormality can never induce late potentials.

Bortolo Martini MD*
Andrea Nava MD+
Sergio Cannas MD*
*Department of Cardiology, Ospedale Civile di Thiene
+Department of Cardiology, University of Padua

Correspondence to:
Bortolo Martini MD
Via Gioberti 9
36016 Thiene (VI), Italy
Fax +39 445 368537

References

(1) Viskin S, Fish R, Eldar M, et al. Prevalence of the Brugada sign in idiopathic ventricular fibrillation and healthy controls. Heart 2000;84:31-6.

(2) Nava A, Canciani B, Schiavinato ML, Martini B, Buja GF, La repolarisation precoce dans le precordiales droites: trouble de la conduction intraventriculaire droite? Correlations de l'electrocardiographie- vectorcardiographie avec l'electro- physiologie. Mises a Jour Cardiologiques 1988;17:157-9.

(3) Martini B, Nava A, Thiene G, et al. Ventricular fibrillation without apparent heart disease: description of six cases. Am Heart J 1989;11:1203-9.

(4) Martini B, Corrado D, Nava A, Thiene G. Syndrome of right bundle branch block, ST segment elevation and sudden death: evidence of an organic substrate. In: Nava A, Thiene G, eds. Arrhythmogenic right ventricular cardiomyopathy. Elsevier 1997:438-53.

(5) Yan GX, Antzelevitch C. Cellular basis for the electrocardiographic J wave. Circulation 1996;93:372-9.

(6) Brugada P. Brugada syndrome: an electrocardiographic diagnosis not to be missed. Heart 2000;84:1-3.

We thank Lawrance et al for their letter that raises several important points about the implementation of the national audit of myocardial infarction.[1] Until the numerator and denominator can be confidently stated there is little point in producing case fatality data either nationally or locally. The data that they quote in their letter strongly support this view. There is an urgent need to address this problem.

The definition of myocardial infarction which can be used in hospital, and which we propose to use for the national audit relies primarily on the demonstration of an elevation of a cardiac specific enzyme such as creatine kinase to a level greater than twice the upper limit of normal for the local laboratory, or the elevation of serial non-specific enzymes to a level greater than twice the upper limit of normal with an appropriate pattern of release in relation to the relevant symptoms. This elevation must be accompanied by a history of symptoms consistent with a myocardial infarction and/or cardiographic changes consistent with infarction. Rarely it is necessary for a clinician to need to make a judgment about a patient dying in hospital before significant enzyme release has occurred.

As we point out in the Acute Myocardial Infarction - A Core data Set,[2] the recognition of all patients admitted with myocardial infarction may be difficult for some hospitals. Norris, in an unpublished study from four South Coast Hospitals demonstrated that between 68% and 98% of patients admitted with myocardial infarction to those hospitals were admitted to the Cardiac Care Unit (CCU). How patients having infarction not going to CCU should be identified is a problem which requires a local solution. However this is achieved, accurate coding remains of paramount importance.

Adjustment for case mix in myocardial infarction is a unresolved problem. The work outlined by Lawrance and his colleagues may be a solution. We did not include clinical variables in the initial data set because there were no compelling data to show they would add to the analysis. However, the dataset is not static, and the addition of further fields was always anticipated. It may be that the simple and objective measures proposed by the EMMACE study group will be the first to be added. We will be shortly start collecting data on large numbers of patients with myocardial infarction in the National Audit, together with planned mortality flagging. The introduction of a robust case mix model will be potentially of great value to the National Audit.

J S Birkhead
Clinical Lead National Audit of Myocardial Infarction
Clinical Effectiveness and Evaluation Unit
Royal College of Physicians
London NW1 4LE, UK

References

(1) Lawrance RA. National Service Framework – core data set [Rapid response] Heart 18 August 2000.

(2) Birkhead JS, Norris RM, Quinn T, Pearson M, on behalf of the Coronary Heart Disease National Service Framework for coronary disease: Acute Myocardial Infarction - a core data set for monitoring standards of care. Royal College of Physicians, London. December 1999.