Research article

Natriuretic peptide vs. clinical information for diagnosis of left ventricular systolic dysfunction in primary care

Janka Koschack^{* 1} , Martin Scherer^{* 1} , Claus Lüers² , Michael M Kochen¹ , Dirk Wetzel¹ , Sibylle Kleta² , Claudia Pouwels³ , Rolf Wachter² , Christoph Herrmann-Lingen⁴ , Burkert Pieske² and Lutz Binder³

¹Department of General Practice, Georg-August-University Göttingen, Germany

²Department of Cardiology and Pneumology, Georg-August-University Göttingen, Germany

³Department of Clinical Chemistry, Georg-August-University Göttingen, Germany

⁴Department of Psychosomatic Medicine and Psychotherapy, Georg-August-University Göttingen, Germany

author email corresponding author email* Contributed equally

BMC Family Practice 2008, 9:14doi:10.1186/1471-2296-9-14

Published:	25 February 2008

Abstract

Background

Screening of primary care patients at risk for left ventricular systolic dysfunction by a simple blood-test might reduce referral rates for echocardiography. Whether or not natriuretic peptide testing is a useful and cost-effective diagnostic instrument in primary care settings, however, is still a matter of debate.

Methods

N-terminal pro-brain natriuretic peptide (NT-proBNP) levels, clinical information, and echocardiographic data of left ventricular systolic function were collected in 542 family practice patients with at least one cardiovascular risk factor. We determined the diagnostic power of the NT-proBNP assessment in ruling out left ventricular systolic dysfunction and compared it to a risk score derived from a logistic regression model of easily acquired clinical information.

Results

23 of 542 patients showed left ventricular systolic dysfunction. Both NT-proBNP and the clinical risk score consisting of dyspnea at exertion and ankle swelling, coronary artery disease and diuretic treatment showed excellent diagnostic power for ruling out left ventricular systolic dysfunction. AUC of NT-proBNP was 0.83 (95% CI, 0.75 to 0.92) with a sensitivity of 0.91 (95% CI, 0.71 to 0.98) and a specificity of 0.46 (95% CI, 0.41 to 0.50). AUC of the clinical risk score was 0.85 (95% CI, 0.79 to 0.91) with a sensitivity of 0.91 (95% CI, 0.71 to 0.98) and a specificity of 0.64 (95% CI, 0.59 to 0.67). 148 misclassifications using NT-proBNP and 55 using the clinical risk score revealed a significant difference (McNemar test; p < 0.001) that was based on the higher specificity of the clinical risk score.

Conclusion

The evaluation of clinical information is at least as effective as NT-proBNP testing in ruling out left ventricular systolic dysfunction in family practice patients at risk. If these results are confirmed in larger cohorts and in different samples, family physicians should be encouraged to rely on the diagnostic power of the clinical information from their patients.