Can pro–brain natriuretic peptide be used as a noninvasive predictor of elevated left ventricular diastolic pressures in patients with normal systolic function?

Abstract

This study was sought to investigate whether plasma N-terminal pro-brain natriuretic peptide (proBNP) can help identify patients with an elevated left ventricular end-diastolic pressure (LVEDP) or filling pressures in patients with a normal systolic function. The proBNP is a good predictor of an elevated LVEDP in patients with a systolic dysfunction. However, whether proBNP can predict an elevated LVEDP in patients with a normal systolic function remains to be determined. The LV pressures were measured by fluid-filled catheters in 216 patients (125 men, mean age 60 +/- 10 years) with a normal systolic function (ejection fraction 66% +/- 8%, range 50%-81%) who were undergoing diagnostic cardiac catheterization. The proBNP was sampled at the time of cardiac catheterization and was measured using a quantitative electrochemiluminescence immunoassay. The log-transformed proBNP levels correlated significantly with the LVEDP (r = 0.33, P = .001) and LV pre-A-wave pressure (pre-A pressure) (r = 0.31, P = .001). An elevated proBNP, defined as >315 pg/mL, predicted an LVEDP > or = 15 mm Hg with a sensitivity of 16% and a specificity of 95% as well as a pre-A pressure > or = 15 mm Hg with a sensitivity of 36% and a specificity of 95%. However, among the 93 patients with an LVEDP > or = 15 mm Hg, 77 (83%) patients had a normal proBNP concentration (< 315 pg/mL). The proBNP level showed weak correlations with the LVEDP and LV pre-A pressure in patients with a normal systolic function. Although high proBNP levels can predict an elevated LV diastolic pressure with high specificity, the sensitivity was quite low. Because the majority of patients with an elevated LVEDP had a normal proBNP, the proBNP level may not be suitable as a screening test for assessing LV filling pressures in the presence of normal systolic function.