NT-proBNP is associated with coronary heart disease risk in healthy older women but fails to enhance prediction beyond established risk factors: Results from the British Women's Heart and Health Study

Abstract

ObjectiveLimited evidence suggests NT-proBNP improves prediction of coronary heart disease (CHD) events but further data are needed, especially in people without pre-existing CHD and in women. MethodsWe measured NT-proBNP in serum from 162 women with incident CHD events and 1226 controls (60–79 years) in a case–control study nested within the prospective British Women's Heart and Health Study. All cases and controls were free from CHD at baseline. We related NT-proBNP to CHD event risk, and determined to what extent NT-proBNP enhanced CHD risk prediction beyond established risk factors. ResultsThe odds ratio for CHD per 1 standard deviation increase in logeNT-proBNP was 1.37 (95% CI: 1.13–1.68) in analyses adjusted for established CHD risk factors, social class, CRP and insulin. However, addition of logeNT-proBNP did not improve the discrimination of a prediction model including age, social class, smoking, physical activity, lipids, fasting glucose, waist:hip ratio, hypertension, statin and aspirin use, nor a standard Framingham risk score model; area under the receiver operator curve for the former model increased from 0.676 to 0.687 on inclusion of NT-proBNP (p=0.3). Furthermore, adding NT-proBNP did not improve calibration of a prediction model containing established risk factors, nor did inclusion more appropriately re-classify participants in relation to their final outcome. Findings were similar (independent associations, but no prediction improvement) for fasting insulin and CRP. ConclusionThese results caution against use of NT-proBNP for CHD risk prediction in healthy women and suggest a need for larger studies in both genders to resolve outstanding uncertainties.