Use of a solid-phase extraction with radioimmunoassay to identify the proportional bias of clinical B-type natriuretic peptide immunoassay: the impact of plasma matrix and antibody multispecificity

BACKGROUND: Circulating immunoreactive B-type natriuretic peptide-32 (ir-BNP-32) has diagnostic and prognostic values in heart failure. We compared, in parallel, a point-of-care (POC) test (Triage((R)) BNP Test) of whole plasma and radioimmunoassay (RIA) of solid-phase extracted (SPE) plasma (SPE/RIA) utilizing a novel copolymer column, in the measurement of patient ir-BNP-32 concentrations. METHODS: Approximately 0.25 mL thawed plasma was transferred to a BNP test device and inserted in a Triage Meter Plus, which gave ir-BNP-32 concentration in pg/mL. Concurrently, for the SPE/RIA measurement, 1.0 mL plasma was acidified and extracted with an OASIS column; eluate dried, reconstituted and quantified by RIA. RESULTS: Inter-day coefficient of variation for both methods were <15%. Plasma SPE recovery was 75.2%. POC correlated with recovery corrected SPE/RIA for ir-BNP-32, r=0.843 (p<0.0001) and the Passing-Bablok model was POC ir-BNP-32=1.43x (recovery corrected SPE/RIA ir-BNP-32)+9.75 ng/L (n=81). A proportional bias was also evident from the Bland-Altman plot, r=0.716 (p<0.0001). CONCLUSIONS: A proportional bias is responsible for plasma ir-BNP-32 concentration differences between whole plasma POC test and recovery corrected SPE/RIA measurements. Ir-BNP-32 assays are influenced by plasma matrix and antibody multispecificity. Consequently, consistent analytical accuracy between immunoassays is necessary to attain a single ir-BNP-32 concentration threshold for diagnosis. Clin Chem Lab Med 2007;45:1353-9.