Correspondence
B-Type Natriuretic Peptide in Heart Failure
N Engl J Med 2002; 347:1976-1978December 12, 2002
- Article
To the Editor:
Maisel et al. (July 18 issue)1 suggest the usefulness of measurement of B-type natriuretic peptide in the diagnosis of left ventricular dysfunction in patients with acute dyspnea, but their conclusions must be challenged. With a cutoff level of 50 pg per milliliter for B-type natriuretic peptide, the negative likelihood ratio was 0.04, allowing them to rule out the diagnosis. Conversely, the specificity (83 percent) and the positive likelihood ratio (5.1) obtained with a cutoff level of 150 pg per milliliter seem too low to confirm the diagnosis definitely, because the post-test probability of symptomatic left ventricular dysfunction increases from 49 percent to 83 percent, leaving 17 percent of cases still undiagnosed. A very similar positive likelihood ratio has been obtained in patients with a lower pretest probability of left ventricular dysfunction (32 percent),2 thus supporting less optimistic conclusions concerning the diagnostic performance of the B-type natriuretic peptide value.
Furthermore, to overcome the imperfect sensitivity of echocardiography (the current gold standard), the authors based their diagnoses of left ventricular dysfunction on the clinical judgment of two cardiologists. However, it is unclear why a review of medical records should be more reliable than the probability assigned by the physician involved in the patient's care. In addition, the rate of agreement between these two diagnostic assessments, as well as the rate of agreement between the two cardiologists, should have been clearly evaluated and reported. Finally, the lack of longitudinal data makes it impossible to estimate the relation between B-type natriuretic peptide levels and the severity of left ventricular dysfunction or to evaluate the prognostic effectiveness of the measurement of B-type natriuretic peptide or its superiority over the easier and cheaper New York Heart Association classification.
2 ReferencesAgostino Colli, M.D.
Ospedale A. Manzoni, 22053 Lecco, ItalyMirella Fraquelli, M.D., Ph.D.
Dario Conte, M.D.
Istituto di Ricovero e Cura a Carattere Scientifico, Ospedale Maggiore, 20122 Milan, Italy
dario.conte@unimi. it 1
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
Full Text | Web of Science | Medline2
Landray MJ ,Lehman R ,Arnold I . Measuring brain natriuretic peptide in suspected left ventricular systolic dysfunction in general practice: cross-sectional study. BMJ 2000;320:985-986
CrossRef | Web of Science | Medline
To the Editor:
In their recent report, Maisel et al. have calculated their predictive values with a pretest probability of 47 percent. This is an extraordinarily high prevalence of disease and is reflective of the sick nature of the population studied for this report. It should be pointed out that with a different prevalence of disease, perhaps in a different clinical setting, the predictive values would be completely different. It is therefore important to clarify that these numbers would only be reproducible in clinical situations with a high prevalence of disease.
It should also be pointed out that the negative predictive value for a B-type natriuretic peptide level of 50 pg is associated with a false positive rate of nearly 40 percent. This makes it effective only as an initial screening test, with a more specific test required to confirm the diagnosis.
It is also obvious that the large standard deviations around the various means for different disease groups indicate the huge variance in B-type natriuretic peptide levels within the same disease group. This variation makes B-type natriuretic peptide less useful as a diagnostic tool, because the range is not narrow enough to include a normal population within 2 SD of the mean. It does not, however, preclude the use of the test for prognostic purposes, with the use of the base-line measurement for an individual patient serving as a control value. Finally, it would have been much more informative to compare the odds ratio associated with B-type natriuretic peptide measurements with a cumulative odds ratio associated with history, physical findings, and radiographic abnormalities.
Yusuf Hassan, M.D.
Adam R. Shapira, M.D.
Southwestern Medical School, Dallas, TX 75390-9030Saira Hassan, M.D.
University of Texas Health Science Center, Houston, TX 77025To the Editor:
Maisel et al. present data supporting previous observations that B-type natriuretic peptide levels are elevated in patients with left ventricular dysfunction. The question of the usefulness of this test in diagnosing heart failure in symptomatic patients is more complex.
As the authors note, the diagnosis of heart failure remains largely clinical, based on a pattern of signs, symptoms, and imaging studies. The important practical question is whether the measurement of B-type natriuretic peptide adds to the accuracy of the clinical diagnosis when it is in doubt. The authors do not report the diagnosis given to their study patients by emergency physicians; without this information, it is not possible to determine whether the measurement of B-type natriuretic peptide added diagnostic accuracy. If the emergency physician was already correct, no additional diagnostic information was provided by the B-type natriuretic peptide level.
Further reason for caution in interpreting B-type natriuretic peptide levels is that recent investigations have found elevated levels in patients with right ventricular dysfunction.1,2 Thus, the attribution of elevated B-type natriuretic peptide levels in patients with dyspnea to left ventricular failure might lead to dangerous misdiagnosis in, for example, cases of pulmonary embolism.
Finally, the authors' analysis of odds ratios for clinical signs of heart failure appears to violate the maxim that studies of the accuracy of a test cannot examine tests that are diagnostic criteria for the disease in question.3 For example, rales and cephalization of vessels are both included in the criteria for heart failure used in the Framingham Heart Study and the National Health and Nutrition Examination Surveys. Despite the promise of B-type natriuretic peptide as a marker of heart failure, it remains unclear at present which patients will benefit from its measurement and what the clinical implications of a single value — particularly if it is in the moderately elevated range — ought to be.
3 ReferencesRobert S. Foote, M.D.
Justin D. Pearlman, M.D., Ph.D.
Dartmouth Hitchcock Medical Center, Lebanon, NH 037561
Tulevski II ,Groenink M ,van Der Wall EE , et al. Increased brain and atrial natriuretic peptides in patients with chronic right ventricular pressure overload: correlation between plasma neurohormones and right ventricular dysfunction. Heart 2001;86:27-30
CrossRef | Web of Science | Medline2
Tulevski II ,Hirsch A ,Sanson BJ , et al. Increased brain natriuretic peptide as a marker for right ventricular dysfunction in acute pulmonary embolism. Thromb Haemost 2001;86:1193-1196
Web of Science | Medline3
Newman TB, Browner WS, Cummings SR. Designing studies of medical tests. In: Hulley SB, Cummings SR, Browner WS, Grady D, Hearst N, Newman TB, eds. Designing clinical research: an epidemiologic approach. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2001:175-93.
Author/Editor Response
The authors reply:
To the Editor: Four important points have been raised regarding our study and the use of B-type natriuretic peptide in the emergency diagnosis of congestive heart failure. In response to Hassan et al., we would note that sampling was an important issue, in that our population was broadly inclusive of adults who presented with acute dyspnea as a primary symptom. The prevalence of 47 percent for congestive heart failure exceeded our prediction of 30 percent and, hence, makes the important point that congestive heart failure was quite common as an underlying cause of dyspnea in one of the largest studies conducted in this area. The prevalence of the condition does drive the positive and negative predictive values, and given the frequency of congestive heart failure in our population, we believe our predictive-value statistics are quite stable.
Second, as Colli et al. and Hassan et al. point out, there are multiple issues raised by the forcing of a continuous measure such as B-type natriuretic peptide into binary “positive” and “negative” categories. As the B-type natriuretic peptide level decreases through its measurable range, the likelihood that heart failure is present in a given population also decreases. Conversely, as the B-type natriuretic peptide level increases, the likelihood that heart failure is present increases. Certainly, we embrace an integrative approach to the use of the B-type natriuretic peptide level along with clinical judgment in the diagnosis of congestive heart failure and do not support turning B-type natriuretic peptide into a binary variable for clinical use.1
Third, in response to Foote and Pearlman, the reference or gold standard for the diagnosis of congestive heart failure was agreement on that diagnosis by two cardiologists who independently reviewed the case, with all the clinical data assembled, 30 days after the encounter. These data included the response to treatment, additional diagnostic tests, and follow-up data that were not available to the emergency physician. Agreement between the two cardiologists was 90 percent, leaving 10 percent of cases to be adjudicated by a panel seeking consensus on the final diagnosis. The reference standard did not depend solely on standardized congestive heart failure scores; hence, the multivariate analysis for the outcome of congestive heart failure as determined by the reviewing cardiologists is valid.
Fourth, we have demonstrated in a subsequent publication how the B-type natriuretic peptide measurement would add to clinical judgment in the diagnosis of congestive heart failure.1 The blood test will not replace the clinician but will enhance the speed and accuracy of the diagnosis in one of the most difficult settings — the emergency department.
1 ReferencesAlan S. Maisel, M.D.
Veterans Affairs Medical Center, San Diego, CA 92161
amaisel@ucsd.edu Paul Clopton, B.S.
San Diego VA Healthcare System, San Diego, CA 92161Peter A. McCullough, M.D., M.P.H.
William Beaumont Hospital, Royal Oak, MI 480731
McCullough PA ,Nowak RM ,McCord J , et al. B-type natriuretic peptide and clinical judgment in the emergency diagnosis of heart failure: analysis from Breathing Not Properly (BNP) Multinational Study. Circulation 2002;106:416-422
CrossRef | Web of Science | Medline
- Citing Articles (2)
Citing Articles
1
P. Alter, H. Rupp, M.B. Rominger, A. Vollrath, F. Czerny, K.J. Klose, B. Maisch. (2007) Relation of B-type natriuretic peptide to left ventricular wall stress as assessed by cardiac magnetic resonance imaging in patients with dilated cardiomyopathy. Canadian Journal of Physiology and Pharmacology 85:8, 790-799
CrossRef2
Mathavakkannan Suresh, Ken Farrington. (2005) Natriuretic Peptides and the Dialysis Patient. Seminars in Dialysis 18:5, 409-419
CrossRef
