Comparison of the diagnostic characteristics of two B-type natriuretic peptide point-of-care devices

Abstract

Background: B-type natriuretic peptide (BNP) is used to diagnose heart failure (HF).

Objective: To compare the accuracy of two commercially available point-of-care (POC) devices for measuring B-type natriuretic peptide (BNP) in emergency department (ED) patients with suspected heart failure using the central laboratory testing results as the criterion standard.

Methods: Venous blood samples were collected from adults with suspected heart failure and split into three samples for BNP analysis: central laboratory (Siemens ADIVA Centaur; Siemens, Deerfield, IL), Triage BNP POC device (Biosite, San Diego, CA), and i-STAT BNP POC device (Abbott, East Windsor, NJ). The criterion standard for BNP levels was the central laboratory.

Results: Two hundred fifty patients were enrolled. Mean (SD) age was 70.7 (13.8) years; 200 (80%) were over age 55 years; 146 (58.4%) were male. A final hospital discharge diagnosis of heart failure was made in 108 (42%) patients. The i-STAT system yielded a result within a median of 9 min (interquartile range [IQR] 9-10 min). The Triage device yielded a result within a median of 19 min (IQR 15-22 min); p < 0.001. The device failure rate for the central laboratory (8 failures, 3.2%) was significantly higher than that of the i-STAT device (1 failure, 0.4%, p = 0.04), but not statistically different than the Triage device (3 failures, 1.2%). Neither the Triage nor the i-STAT were statistically different than the central laboratory result in terms of sensitivity; the i-STAT was less specific than the Triage result (p = 0.003). The area under the curve for the Triage device was 0.95 (95% confidence interval [CI] 0.91-0.98), whereas the area under the curve for the i-STAT device was 0.98 (95% CI 0.96-0.99; p < 0.01).

Conclusions: Both POC devices tested were accurate and rarely failed; however, the i-STAT was faster with single use.