Prognostic value of serial B-type natriuretic peptide measurement in asymptomatic organic mitral regurgitation.

ABSTRACT

AIMS: Optimal timing of surgery is crucial in mitral regurgitation (MR) to avoid excess mortality and morbidity. The role of brain-type natriuretic peptide (BNP) in this setting remains controversial. We evaluated the value of serial BNP measurements for early prediction of deterioration in asymptomatic MR. METHODS AND RESULTS: Eighty-seven consecutive asymptomatic patients with severe organic MR, normal left ventricular (LV) function (ejection fraction ≥ 60%, end-systolic diameter index < 26 mm/m²), systolic pulmonary artery pressure (sPAP) <50 mmHg, and no atrial fibrillation underwent clinical assessment, echocardiography, and measurement of BNP and N-terminal pro-BNP (NT-proBNP) at 6-month intervals. The primary endpoint was the development of symptoms and/or LV dysfunction. The secondary endpoint was the occurrence of atrial fibrillation or sPAP ≥ 50 mmHg. Over a mean follow-up of 786 ± 454 days, 20 patients reached the primary endpoint and 5, the secondary endpoint. By univariate analysis, age, BNP, NT-proBNP, and sPAP were significant predictors of reaching the primary endpoint during the 6 months following testing, whereas LV function and dimensions were not. By multivariate analysis, only BNP (P = 0.03) and sPAP (P = 0.04) remained independent predictors. When secondary endpoints were additionally considered, results remained unchanged. Receiver operator curve analysis yielded AUC-values of 0.90, 0.84, and 0.80 for BNP, NT-proBNP, and sPAP, but 0.60 and 0.57 for left ventricular ejection fraction and end-systolic diameter. The negative predictive value for normal neurohormone levels and sPAP was high (98-100%). A BNP of 145 pg/mL had a positive predictive value of 36%. CONCLUSION: Brain natriuretic peptide and NT-proBNP independently predict outcome in asymptomatic MR. Serial measurements may help to improve timing of surgery. Low plasma levels with their high negative predictive values appear to be particularly helpful by identifying low-risk individuals.