Biological, electrical and echocardiographic indices versus cardiac magnetic resonance imaging in diagnosing left ventricular hypertrophy

Abstract

The aim of this study was to compare the diagnostic performance of N-terminal pro-brain natriuretic peptide (NT-proBNP), electrocardiographic (ECG) criteria and transthoracic echocardiography (TTE) versus cardiac magnetic resonance imaging in detecting left ventricular hypertrophy (LVH). The study included 42 hypertensive subjects with mean±s.d. age 48.1±12.3 years, 57.1% men, 24-h ambulatory blood pressure 144/89 mm Hg, left ventricular ejection fraction >50%, without symptoms of heart failure, and not taking any drugs that interfere with hormonal regulation. The accuracies of the methods in detecting LVH were compared at two diagnostic LVH cutoffs: low, 83 g m(-2) in men and 67 g m(-2) in women; and high, 96 g m(-2) in men and 81 g m(-2) in women. With the low and high LVH cutoffs, the areas under the receiver-operating characteristic curves and the optimal values for NT-proBNP were 0.761, 0.849, 200 and 421 pg ml(-1), respectively. An NT-proBNP level under 30 pg ml(-1) ruled out LVH with 100% sensitivity. The optimal values and literature-based values of NT-proBNP allowed a correct classification of 73-81% of the subjects. In 80-90% of the cases, the diagnostic accuracy of NT-proBNP was close to that of ECG criteria but lower than that of TTE criteria. Interestingly, combining ECG criteria and NT-proBNP level improved the diagnostic performance to be at least comparable to that of TTE: the percentages of correctly classified subjects were 73-95% vs. 67-86%, respectively. Of note, the range considers both diagnostic LVH cutoffs. The simultaneous use of ECG criteria and NT-proBNP plasma levels seemed to be powerful enough to detect LVH in most hypertensive subjects.