OP0025 Antidrug antibodies detection is strongly influenced by the type of assay used

Background Direct comparison of immunogenicity data is hampered due to different assays used with different sensitivity for drug interference. This is the first study to compare detection of antidrug antibodies (ADA) with different assays in ankylosing spondylitis (AS) patients. Studying immunogenicity in AS patients with a drug tolerant assay may contribute to a better understanding of development of ADA. Objectives To compare the detection of ADA in different assay techniques with differences in drug interference, acid-dissociation-radioimmunoassay (ARIA) antidrug binding test (ABT) and the more frequently used 2-sided Enzyme-linked Immuno Sorbent Assay (ELISA) in AS patients. Second, to study the relation of adalimumab drug levels with the detection of ADA. Methods In this study, the detection of the ADA in ARIA and ABT was compared with the detection of ADA in the 2-sided ELISA in 84 consecutive AS patient over a period of 24 weeks; at 4, 12 and 24 weeks of treatment. Adalimumab drug levels were measured using an ELISA. The assays were designed by Sanquin, Amsterdam. For the difference in drug levels we divided the patients in four different groups; all assays negative (group 0), only ARIA positive (group 1), ARIA and ABT positive, 2-sided ELISA negative (group 2) and all assays positive (group 3). We used last observation carried forward to imputate missing data. Results As shown in Figure 1, 26% of the patients tested positive for ADA in the ARIA compared to 14% in the ABT and no detection in the 2-sided ELISA at week 4. At weeks 12 and 24 respectively, cumulative 46% and 69% of patients, tested positive in the ARIA for ADA, compared to 19% and 35% in the ABT and 2% and 4% with the 2-sided ELISA. Median adalimumab levels at week 24 in group 0, 1, 2 and 3 were 9.5 (5.3–13.3), 8.4 (5.3–11.0), 2.8 (0.9–4.3) and 0.002 (0.0–1.3) respectively. No significant differences were found in median adalimumab levels between patients with no ADA detection and patients tested positive in only the ARIA, respectively, 8.4 (5.3–11.0) and 9.5 (5.3–13.3) p=0.385. However, when both ARIA and ABT tested positive, drug levels significantly differed from patients with no ADA detection, 2.8 (0.9–4.3) and 9.5 (5.3–13.3), p=0.000. When patients tested positive in the ARIA, ABT and the 2-sided ELISA almost all patients had undetectable drug levels. Conclusions The fast majority of the AS patients develop ADA. The ARIA detects more ADA compared to the assays susceptible for drug interference, the ABT and the more frequently used 2-sided ELISA. Immunogenicity data should be interpreted with the knowledge of the assay and in context of drug levels. Disclosure of Interest J. Ruwaard: None declared, A. Marsman: None declared, M. Nurmohamed: None declared, H. te Velthuis: None declared, K. Bloem: None declared, T. Rispens Speakers bureau: Pfizer, AbbVie, G. J. Wolbink Grant/research support from: Pfizer, Speakers bureau: Pfizer, AbbVie, UCB