Background The options for treatment of rheumatic diseases is constantly growing. Tocilizumab (TCZ) is an anti-interleukin 6 receptor monoclonal antibody indicated for the treatment of severe active and progressive rheumatoid arthritis (RA) and active systemic juvenile idiopathic arthritis (sJIA). Monitoring of drug levels (DL) and anti-drug antibodies (ADA) in clinical practice is highly advisable in order to establish normal therapeutic ranges, complement clinical assessment and optimise patient treatment. As a consequence several tests and technologies have been developed. However, a common objection that limits implementation of biological drug monitoring in medical practice is the claim that the magnitude of measurements varies among assays and that assay standardisation is lacking.

Objectives The purpose of this study was to evaluate the correlation and agreement between different commercially available tests to measure TCZ and antibodies to TCZ (ATT) in patients with rheumatic diseases.

Methods Trough levels of TCZ and ATT were analyzed in a retrospective cohort of 26 rheumatic patients (42 serum samples, 88% RA and 12% sJIA) under maintenance therapy with TCZ for up to two years. Samples were assayed using Promonitor-TCZ and Promonitor-ANTI-TCZ ELISA kits (PR, Progenika, Spain), LISA-TRACKER Duo Tocilizumab ELISA kit (LT, Theradiag, France), and by Sanquin Blood Supply testing service (SQ, Amsterdam, The Netherlands) that uses ELISA and RIA technologies to measure TCZ and ATT levels, respectively. One-way analysis of variance (ANOVA) and Spearman's test was used to study correlation between the DL tests; the difference (bias) in values obtained with DL assays was also assessed with Bland-Altman analysis. Percentage of positive and negative agreement was used to study the association between the ADA assays.

Results TCZ trough level in patients are much higher than other biological treatments like anti-TNF infliximab and adalimumab. ANOVA analysis showed that the differences in the median values among the three DL assays are not statistically different (median 4.4 vs 4.7 vs 5.4 μg/mL, IQR 0 – 8.7 μg/mL vs 0 – 7.8 μg/mL vs 0 – 9.4 μg/mL for Promonitor-TCZ, LISA-TRACKER and Sanquin ELISA tests, respectively, p=0.693). Spearman's coefficient showed a very high correlation between the each pair of assays (r=0.951, 0.935 and 0.944 for PR vs LT, PR vs SQ and LT vs SQ, respectively, p<0.0001). Bland-Altman analysis shows an excellent association between each pair of DL assays with no significant difference in bias (-0.3, -0.5 and -0.2 μg/mL for PR vs LT, PR vs SQ and LT vs SQ, p>0.05). Antibodies to TCZ were not detected with any of the assays in the patient cohort confirming the low immunogenicity of TCZ.

Conclusions Despite different reagents, assay formats and technologies, all tests studied here provide exactly the same results and measure the same magnitude of TCZ in rheumatic patients. Assays can be safely exchanged in a clinical setting even in the absence of a gold standard test, and should lead to similar drug therapeutic ranges and actions.

Acknowledgement This work was supported by the Basque Country Government (Dep. Education, Universities and Research, Ref. IT687–13) and ZabaldUZ 2012, UPV/EHU.

Disclosure of Interest S. Martín Grant/research support from: PhD student research support from Progenika, A. Ruiz del Agua Employee of: Employee of Progenika Grifols, N. Torres Employee of: Employee of Progenika Grifols, D. Pascual-Salcedo: None declared, C. Plasencia: None declared, T. Jurado: None declared, A. Martínez: None declared, A. Balsa: None declared, B. Ruiz-Argüello Employee of: Employee of Progenika Grifols, A. Martínez Employee of: Employee of Progenika Grifols, R. Navarro: None declared, D. Nagore Employee of: Employee of Progenika Grifols