Background The last decade brought major advances in the available treatment choices for rheumatoid arthritis (RA) patients. Also, new treatment algorithms propose rapid therapeutic adaptations, and, if needed, a quick step-up to biological DMARDs in patients at risk for bad outcome.

Objectives We aimed to establish the temporal trends in disease modifying antirheumatic drug (DMARD) applications in the 21st century.

Methods Beginning in 2001, we identified an inception cohort of patients who had at least 3 complete follow-up visits, and received a DMARD therapy for at least one month. We divided DMARDs into 5 groups and used survival analysis, cox regression and comparative analyses.

Results We identified 309 patients who fulfilled inclusion criteria (75% female, 56% rheumatoid factor positive, median symptom duration 0.8 (0.4-2.1) years; median SDAI was 17.7 (9.8-30.0). 69% of the patients showed moderate to high clinical disease activity at start of their first treatment. As first treatment, 294 patients (95%) received a conventional synthetic (cs) DMARD strategy, in 77.3% of patients this was methotrexate (MTX), in 8.7% salazopyrine and in 6.5% leflunomide. As second treatment after MTX most patients received either another csDMARD (47%) or a biologic (b) DMARD (39%). Biological DMARDs were used increasingly after failure of increasing numbers of csDMARD therapies (Table).

Median retention of the first five treatment segments were 35.7 months (95% confidence interval: 20.1-51.3), 25.2 (16.5-33.9), 14.2 (4.9-23.5), 14.4 (11.4-17.4), 10 (6.2-13.9) months; the rate of censoring (still ongoing therapies) was between 44% and 48% in the first 3 segments and 26% and 32% for segments 4 and 5. In analyses of all treatment segments, csDMARDs (30.7 months; 22.8-38.6) and tumor necrosis factor alpha-inhibitor (TNFi)-combinations (29.4; 14.8-44.0) showed the highest retention rates, followed by combinational therapy of synthetic DMARDs (15.3; 2.4-28.2), non-TNFi-bDMARDs (13.5; 10.2-16.8) and biological-monotherapy (11.7; 1.9-21.5). When adjusting for number of previous treatment segments (p<0.001), the total observation period (p=0.61) and CDAI levels (p=0.03) at the respective treatment starts, TNFi bDMARDs showed the longest retention rates and csDMARDs and non-TNFi bDMARDs were very similar. After the first treatment course bDMARDs showed longer survival than csDMARDs (2nd segment: 22.6 vs. 39.1 p=0.087).

Disease activity markers at baseline were not significantly different between different treatment segments, except for acute phase reactant levels, which decreased with each treatment course (p<0.001 for CRP; p=0.037 for ESR). With higher number of previous treatment courses, more patients were rheumatoid factor and ACPA positive.

Conclusions MTX and leflunomide are clearly dominating in early treatment phases, but already as a second treatment choice biological DMARDs are of significance and show longer retention rates. After each previous treatment the retention of the treatment course decreases significantly.

Disclosure of Interest None declared

DOI 10.1136/annrheumdis-2014-eular.3812