Background The assessment of biological drug levels and immunogenicity might be essential in terms of a more effective and rational use of biological therapies and it is dependent upon the establishment of efficient standardized assays or a consensus that could allow a direct comparison of drug levels and anti-drug antibodies (ADA) data with clinical outcome.
Objectives To determine whether an enzyme-linked immunosorbent assay (ELISA) performed with two versions of a commercial kit to assess IFX and ADL levels and ADA yield similar results.
Methods The diagnostic capability of two ELISA versions [Promonitor® IFX R1 and R2 (V.1), Promonitor® IFX and Anti-IFX (V.2); Promonitor® ADL R1 and R2 (V.1), Promonitor® ADL and Anti-ADL (V.2) kits (Progenika Biopharma, Spain)] was evaluated in patients with RA treated either with infliximab (IFX; n=24) or adalimumab (ADL; n=24) by three different laboratories. The reliability was determined using the Cohen's Kappa coefficient (K), the Pearson's r, the intraclass correlation coefficient (ICC) and the Lin's concordance correlation coefficient (CCC). The Bland-Altman plots of differences between V.1 and V.2 were drawn to compare values of each assay.
Results The qualitative discordant results for IFX levels V.1 were 9/24 samples (K= good) and 2/23 V.2 (K= very good), for ADL levels V.1 were 7/24 (K= moderate) and for V.2 were 0/24 (K= very good). For IFX-ADA were 0/24 in V.1 and V.2 (K= excellent), while for ADL-ADA were 1/24 in V.1 and 4/24 in V.2 (K= from very good to good). The quantitative agreement is shown in table 1, we found a good linear association using the Pearson's r, the ICC was questionable to excellent for both versions in IFX and ADL levels. The CCC results were poor in all determinations. Bland-Altman plots for both, IFX and ADL levels demonstrate the differences between both versions (Fig. 1 and 2, respectively).
Conclusions We observed a better agreement in the qualitative than in the quantitative results in both, for IFX and ADL levels and for IFX-ADA and ADL-ADA. We have found a good linear association (Pearson's r) but a low agreement (K, ICC and CCC) comparing results for V.1 and V.2. Further interlaboratory investigations are necessary to improve results and determine their possible clinical application.
Disclosure of Interest L. Valor: None declared, D. Hernández Flόrez: None declared, I. de la Torre: None declared, F. Llinares: None declared, J. Rosas: None declared, J. Yaque: None declared, E. Naredo Grant/research support: UCB and MSD, Consultant for: Abbvie, Roche Farma, Bristol-Myers Squibb, Pfizer, UCB, General Electric Healthcare, and Esaote, C. Gonzalez: None declared, J. Lόpez-Longo: None declared, I. Monteagudo Consultant for: Abbvie, Roche Farma, Bristol-Myers Squibb, Pfizer, UCB, General Electric Healthcare, MSD and Esaote, M. Montoro: None declared, L. Carreño Perez: None declared