LAUSR

It is not currently possible to accurately predict tumour necrosis factor inhibitor (TNF-i) response in psoriatic arthritis (PsA). In rheumatoid arthritis, suboptimal adherence to TNF-i and reduced drug levels have been associated with decreased response. Suboptimal treatment adherence (16%-81%) has been reported in PsA although little is known about the impact on TNF-i response. Self-reported adherence may be biased, but direct drug measurements offer an objective adherence evaluation. The objective of this study is to assess the relationship between self-reported adherence, drug levels, and drug response to TNF-i in PsA. Blood samples and adherence questionnaires were collected at 3, 6 and 12 months for PsA patients prescribed biologic or targeted synthetic anti-rheumatic drugs in a UK multi-centre observational study (OUTPASS). Self-reported adherence was measured using two previously validated questionnaires (MARS5 and Treatment). Adherence was classified as self-reporting never missing or delaying a dose, unless medically advised. Adalimumab (ADA) and etanercept (ETN) drug levels were measured at 3 and 6 months using commercially available ELISAs. Clinical response was assessed using PsARC and ΔDAS28 between baseline and 3, 6 and 12 months. Univariate regression and general estimating equations were used to assess associations across timepoints. In the OUTPASS cohort (46% ADA, 40% ETN), 42% of patients self-reported non-adherence within the first 12 months of treatment. Patients with self-reported non-adherence at 6 months using the MARS5 questionnaire were significantly less likely to achieve a PsARC response at 6 months [odds ratio (OR)=0.40, P = 0.047] in univariate regression. Non-adherence, reported using the Treatment questionnaire, was associated with decreased PsARC response over 12 months [OR = 0.52, P = 0.031]. There was no association (P > 0.05) between whether a patient returned a self-reported adherence questionnaire and response. Drug levels at 3 months (n = 70 (ADA), 73 (ETN)) were associated with drug levels at 6 months (n = 61 (ADA), 67 (ETN)) for both ADA [β = 0.47, P = 0.001] and ETN [β = 0.41, P = 0.018] paired samples. Higher 3-month ADA levels were associated with increased ΔDAS28 response at 3 months [β = 0.11, P = 0.036] and PsARC response at 12 months [OR = 1.39, P = 0.028]. Higher 3-month ETN levels were associated with increased PsARC response at 3 months [OR = 1.50, P = 0.026]. Increased ADA levels were associated with improved ΔDAS28 over 12 months [β = 0.07, P = 0.001] and higher ETN levels associated with enhanced PsARC response [OR = 1.17, P = 0.022] over 12 months. Levels of neither drug were associated (P > 0.05) with any self-reported adherence measures at any time point. There was no association between self-reported adherence and TNF-i drug levels in ADA or ETN-treated PsA patients. Objectively measured adherence via drug level measurements were more strongly associated with drug response than self-reported measures, particularly using ΔDAS28 as the measure of response. Disclosure P.D.K. Curry: None. A.P. Morris: None. M. Jani: None. H. Chinoy: Grants/research support; HC received grant support from Eli Lilly and UCB. A. Barton: Grants/research support; AB has received fees/grant awards from Galapagos, Scipher Medicine, Roche-Chugai, Pfizer and BMS. J. Bluett: Grants/research support; JB has received a research grant award from Pfizer. Other; JB has received travel/conference fees from UCB, Pfizer and Eli Lilly.