Possible role of BNP for an early diagnosis of asymptomatic left ventricular dysfunction: a retrospective study.

AIM: The use of haematochemical markers of intra-ventricular myocardial strain, known as B-type natriuretic peptides, has been proposed as an initial diagnostic step for early screening programs targeting asymptomatic left ventricular dysfunction (ALVD) in populations of subjects at risk. METHODS: The diagnostic accuracy of BNP for the identification of ALVD was assessed by adopting Doppler echocardiography supplemented by Doppler Tissue Imaging (DTI) as the gold standard method. For this purpose, a retrospective analysis of the medical records of patients judged at risk of ALVD, mostly hypertensive and/or diabetic patients, was carried out. For admission into the study, at least one determination of plasma BNP and one Doppler echocardiographic assessment complemented by DTI were required for each patient. Systolic ALVD was diagnosed in the presence of left ventricular ejection fraction(LVEF) of <50 %. Diastolic ALVD was assessed according to the recommendations of the American Society of Echocardiography using two-dimensional Doppler echocardiography and TDI. Thus, in the presence of LVEF of ≥50%, concomitant findings of lateral mitral annular e' velocity <10 cm/s and septal mitral annular e' velocity <8 cm/s with left atrial volume index ≥34 ml/m2 were required to substantiate a diagnosis of diastolic ALVD . It was subsequently graded by measuring the ratio of early (E) to late (A) transmitral flow velocity (E/A ratio), mitral deceleration time and E/average e' ratio. Several Receiver Operating Characteristic (ROC) curves were built in the entire study population as well as in some subsets, to identify the BNP values exhibiting the best profile of sensitivity, specificity and positive and negative likelihood ratios with respect to ALVD. RESULTS: Among 88 patients at risk enrolled on the whole, 33 cases (37.5%) of mild diastolic ALVD and 10 cases (11.4%) of moderate-to-severe diastolic ALVD were found. No cases of systolic ALVD were detected. The ROC plots for the diagnostic accuracy of serum BNP in early detection of ALVD in the entire cohort of patients at risk of heart failure showed an area under the curve of 0.761 [standard error=0.0523; P<0.0001 ] . Moreover, among the various BNP values located across the ROC curve, an optimal cut-off value of 156 pg/ml was found (sensitivity=83.7%; specificity=46.7%; positive likelihood ratio=1.57; negative likelihood ratio=0.35). Conversely, a value of 290 pg/ml was shown to be more specific at the cost of a remarkable loss in sensitivity (sensitivity=53.5%;specificity=95.5%; positive likelihood ratio=12; negative likelihood ratio=0.5). CONCLUSION: The sensitivity of BNP in the identification of patients with ALVD by adopting a cut-off of 156 pg/ml was equal to 83.7%, which is in keeping with the literature values. In case of programs targeted to ALVD screening, the choice of adopting a relatively low threshold value for BNP seems to comply with the goal of enhancing sensitivity , so as to exclude from the echocardiography the individuals found negative at BNP assay and thereby reduce the number of unnecessary echocardiographic examinations in individuals with low probability of having ALVD.