Background Response to treatment varies widely between RA-patients. Our means to predict disease course and treatment response is limited, leading to substantial over- as well as under- treatment. Whereas female gender and smoking have been identified as predictors of non-response, the impact of autoimmune co-morbidities remains largely unknown. Autoimmune thyroid disease (AITD) is one of the most frequent autoimmune diseases in the population. AITD is more prevalent in RA-patients and has also been identified as a risk factor for RA. AITD can be readily identified via thyroxin substitution. We aimed at assessing the impact of prevalent AITD in relation to 3- and 6-month EULAR response to methotrexate in early RA.

Objectives To investigate whether thyroxin substitution impacts response to methotrexate as the first-line therapy in RA.

Methods We identified patients with incident RA (symptom duration <1 year), included in the Swedish Rheumatology Quality Register, July 2006 through 2015 (n=7009). All patients starting treatment with methotrexate and who had a follow-up visit at 3 months (n=4364) and/or at 6 months (n=3148) were included. Prevalent AITD was defined as prescription of thyroxin substitution before RA-diagnosis (n=347), based on linkage to the Swedish Prescribed Drug register, and excluding participants with prescriptions for iodine-containing drugs or history of thyroid cancer (e.g. non-autoimmune cause of thyroxin use). We used a case-control design with thyroxin substitution as exposure, cases defined as EULAR DAS28 3- and 6-month non-responders, and controls defined as moderate/good responders at these time-points. Odds Ratios (OR) were calculated adjusted for sex, age, and HAQ.

Results At 3 months, the proportion of thyroxin users did not differ between responders and non-responders 12% vs. 11%, (OR non-response =1.1, 95% CI 0.9–1.4). At 6 months, the corresponding figures were 13% vs. 11%, respectively, (OR non-response =1.3, 95% CI 1.0–1.7). However, a significant difference was observed in RF and/or ACPA positive patients at 6 months, where 15% of non-responders and 10% of responders used thyroxin (OR non-response=1.6, 95% CI 1.1–2.1), while no such difference was observed for seronegative RA. When stratified for gender, thyroxin substitution was significantly associated with non-response in men but not in women.

Conclusions This large real-life study of response to methotrexate in early RA suggests that AITD, measured as thyroxin replacement therapy, may be linked to treatment response in seropositive patients (and among males). Exploratory by nature, these findings call for replication.

Disclosure of Interest K. Waldenlind: None declared, S. Saevarsdottir: None declared, C. Bengtsson: None declared, J. Askling Grant/research support from: Abbvie, Pfizer, Lilly, Samsung, MSD, UCB, Roche, Janssen