Methods

In a cross-sectional study, the authors evaluated 559 community subjects aged> or= 45 years. The association between plasma BNP concentration and the hypothesized determinants was assessed by linear regression using the natural logarithm of BNP plasma concentration as the dependent variable and separate models for each gender.

Results

Plasma BNP was significantly higher in women. BNP increased with age, more steeply in men than women (p< 0.05 for the interaction). In univariate analysis, BNP was associated with left ventricular systolic dysfunction (LVSD), moderate to severe valvular abnormalities, left atrial diameter/body surface area, left ventricular mass index, systolic blood pressure, lower creatinine clearance and jugular venous distension. In women it was also associated with left ventricular end-diastolic diameter/body surface area and atrial fibrillation, and in men with changes in the segmental wall motion index. The final multivariate models included age, LVSD and left atrial size as significant independent determinants of BNP level, explaining 16.8% and 32.3% of BNP variability in women and men respectively.

Conclusion

The association between BNP and left atrial size independently of systolic dysfunction suggests that BNP might reflect subclinical diastolic abnormaltes. False positive BNP levels for the diagnosis of left ventricular systolic dysfunction and heart failure are probably due to high BNP levels in people with cardiac abnormalities not detected by echocardiography.