N-terminal proB-type natriuretic peptide testing: practical diagnostic value in a emergency department

Abstract

Purpose: N-terminal proBNP (NT-proBNP) testing has been proposed as a part of the evaluation of patients with dyspnea in Emergency Department (ED). Observational studies demonstrated that NT-proBNP improved diagnosis and reduced costs. Nevertheless, recent randomized studies failed to confirm the utility of this laboratory test in ED. We sought to evaluate the diagnostic impact of NT-proBNP measurement in patients presenting with acute dyspnea in ED, taking into account clinical and chest x-ray results routinely obtained. Methods: We enrolled 498 consecutive patients presented at our ED for dyspnea (male/female 248/250; mean±SD age 77.3±12) with a single NT-proBNP determination (Elecsys Roche Diagnostics, cutoff: 450 ng/L) at admission. For each patient enrolled, a score was calculated by evaluating the presence or absence of six clinical and radiological findings, such as: 1) history of heart failure; 2) history of ischemic cardiomyopathy; 3) pulmonary rales; 4) leg edema; 5) chest x-ray cardiac enlargement; 6) chest x-ray pulmonary congestion. According to the clinical and radiological score, the patients were divided in three groups: low (A-group), intermediate (B-group), and high (C-group) probability of heart failure. On a second time two cardiologists, rather than knowledge of the results of NT-proBNP, reviewed all other clinical data and made the final diagnosis. Results: NT-proBNP mediane value was 2445 ng/L (lower quartile 631 and upper quartile 5847 ng/L), and the area under the receiver-operating characteristic curves (AUC) was 0.854 for NT-proBNP, 0.921 for clinical/radiological score, and 0.936 for the two in combination (logistic model). The likelihood ratio positive (LR+) was not determinable in the group A, 1.63 in the group B and 1.22 in the group C. The LR- was 1.26 in the group A, 0.16 in the group B and 0.05 in the group C In the B- group (intermediate) NT-proBNP test added correct diagnostic information in 126 subjects with HF and in 53 subjects without a final diagnosis of HF. In A and C-group NT-proBNP test added correct diagnostic information in 1 patient. Conclusions: NT-proBNP did not substantially enhance diagnostic accuracy in all patients with shortness of breath in ED. In patients presenting to an ED with acute dyspnea classifiable in the low or high probability groups this marker keeps its strong discriminating power, but gives poor additional information to clinical standard evaluation. However, in patients with not conclusive clinical and radiological results NT-proBNP determinations improved the percentage of correct diagnosis.