Comparing the digoxin test with the Architect® i1000sr system with respect to the AxSYM® system.

OBJECTIVE To assess the technique employed by the autoanalyser Architect® i1000sr to determine digoxin in serum samples, compared with the assay developed for AsSYM® using microparticle enzyme immunoassay (DigoxinII). METHOD A prospective analysis of the samples from 100 requests to monitor patients being treated with digoxin. The samples were processed in AxSYM® and Architect®. The techniques were assessed using the linear regression coefficient, determination coefficient, mean absolute error, mean squared prediction error and the Bland-Altman method. RESULTS The serum levels showed a correlation coefficient of 0.93. There was nearly a 40% difference for the concentrations between 0.8 and 2 ng/ml and nearly 20% in the rest of the samples analysed. CONCLUSIONS The Architect® system is precise; however, from a clinical monitoring point of view, it is unacceptably inaccurate when compared with the AxSYM®.