D. Di Giuseppe

Biological treatment in ankylosing spondylitis in the Nordic countries during 2010–2016: a collaboration between five biological registries

Objectives: Large-scale observational cohorts may be used to study the effectiveness and rare side effects of biological disease-modifying anti-rheumatic drugs (bDMARDs) in ankylosing spondylitis (AS), but may be hampered by differences in baseline characteristics and disease activity across countries. We aimed to explore the research infrastructure in the five Nordic countries regarding bDMARD treatment in AS. Method: This observational cohort study was based on data from biological registries in Denmark (DANBIO), Sweden (SRQ/ARTIS), Finland (ROB-FIN), Norway (NOR-DMARD), and Iceland (ICEBIO). Data were collected for the years 2010–2016. Registry coverage, registry inventory (patient characteristics, disease activity measures), and national guidelines for bDMARD prescription in AS were described per country. Incident (first line) and prevalent bDMARD use per capita, country, and year were calculated. In AS patients who started first line bDMARDs during 2010–2016 (n = 4392), baseline characteristics and disease activity measures were retrieved. Results: Registry coverage of bDMARD-treated patients ranged from 60% to 95%. All registries included extensive prospectively collected data at patient level. Guidelines regarding choice of first line drug and prescription patterns varied across countries. During the period 2010–2016 prevalent bDMARD use increased (p < 0.001), whereas incident use tended to decrease (p for trend < 0.004), with large national variations (e.g. 2016 incidence: Iceland 10.7/100 000, Finland 1.7/100 000). Baseline characteristics were similar regarding C-reactive protein, but differed for other variables, including the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) (range 3.5–6.3) and Ankylosing Spondylitis Disease Activity Score (ASDAS) (2.7–3.8) (both p < 0.0001). Conclusion: Collaboration across the five Nordic biological registries regarding bDMARD use in AS is feasible but national differences in coverage, prescription patterns, and patient characteristics must be taken into account depending on the scientific question.

FRI0213 Comparative effectiveness of abatacept, rituximab, tocilizumab and anti-tnf biological dmards in ra: results from the nationwide swedish register

Background Many current guidelines rank abatacept (ABA), rituximab (RTX), tocilizumab (TOC), and the TNFi bDMARDs as equal in effectiveness for the treatment of RA, at least as second line therapies. This is mainly based on evidence from separate RCTs, with few direct comparisons and limited comparative effectiveness data from clinical practice. Objectives To describe outcomes in clinical practice among RA patients starting different bDMARDs as first bDMARD, and after switch from initial TNFi. Methods The Swedish Rheumatology Register was linked to nationwide registers with data on demographics and medical history. We included all patients with RA starting a first ever bDMARD, or switching to a new bDMARD after a TNFi as first bDMARD, in 2010 - 2014, with follow-up through 2015. Effectiveness was assessed at 1 year (±90 days) after starting therapy, and measured as 1) the proportion remaining on therapy, or the proportion remaining on therapy and with 2) Good EULAR response, 3) HAQ improvement >0.2, 4) no swollen or tender joints. Relative response was estimated with log-binomial regression adjusting for potential confounders. Results Patients starting non-TNFi were older than those starting a TNFi, had lower socioeconomic status, and more often a history of diseases including malignancy, serious infections, and diabetes. After switch from TNFi, those starting non-TNFi also had higher disease activity. Non-TNFi were associated with better drug survival and higher proportion reaching response outcomes compared to TNFi as first bDMARD. After switch from TNFi, RTX and TOC, but not ABA, were associated with significantly better drug survival and response. Differences remained after adjusting for identified potential confounders.Table 1. Status at 12 months among all patients with RA initiating a biologic DMARD 2010–2014 in Sweden TNFi RTX TOC ABA % % RR† % RR† % RR† First bDMARD N=5568 N=654 N=202 N=240 On drug 68.4 87.8 1.34 (1.27–1.41) 75.5 1.20 (1.09–1.31) 77.7 1.15 (1.05–1.27) On drug + EULAR Good resp. 26.1 31.1 1.42 (1.19–1.69) 53.1 2.03 (1.70–2.42) 34.3 1.37 (1.10–1.72) On drug + HAQ Improvement 26.7 39.2 1.64 (1.40–1.93) 45.0 1.54 (1.27–1.87) 36.8 1.37 (1.09–1.71) On drug + 28 Joint count = 0 20.3 22.4 1.13 (0.89–1.43) 30.9 1.60 (1.21–2.11) 22.8 1.26 (0.91–1.74) Switch from TNFi N=1840 N=408 N=320 N=256 On drug 57.7 80.2 1.48 (1.37–1.60) 73.0 1.36 (1.23–1.49) 65.1 1.11 (0.98–1.26) On drug + EULAR Good resp. 11.4 24.0 1.87 (1.41–2.49) 36.8 3.06 (2.37–3.94) 14.6 1.16 (0.76–1.76) On drug + HAQ Improvement 16.6 34.3 1.85 (1.49–2.30) 32.4 1.71 (1.33–2.19) 20.4 1.10 (0.78–1.53) On drug + 28 Joint count = 0 12.3 20.8 1.96 (1.43–2.70) 19.9 2.12 (1.48–3.02) 11.2 0.86 (0.48–1.52) †Adj. for region, sex, age, birth country, RF, dis. dur., HAQ, DAS28, co-medication, recent history of malignancy, infection, SSRI, and hospital days last 5 yrs. Conclusions Despite channeling of older and sicker individuals to non-TNFi-bDMARDs, treatment outcomes were in general better in these groups, particularly for TOC and RTX. In interpreting this, the risk of residual confounding should be remembered, and that we did not include safety or long term outcomes. Acknowledgements The ARTIS registry has been, or is, supported by agreements with Abbvie, BMS, MSD, Pfizer, Roche, Samsung, and UCB. Disclosure of Interest T. Frisell: None declared, M. Dehlin: None declared, D. Di Giuseppe: None declared, N. Feltelius: None declared, A. Kastbom Consultant for: Bristol-Myers Squibb, Pfizer, Roche, UCB, Paid instructor for: Bristol-Myers Squibb, Pfizer, Roche, UCB, C. Turesson Grant/research support from: Abbvie, Pfizer, Roche, Consultant for: MSD, Pfizer, Roche, Paid instructor for: Abbvie, Bristol-Myers Squibb, Janssen, MSD, Pfizer, Roche and UCB, J. Askling Grant/research support from: Abbvie, UCB, Pfizer, Merck, Samsung, Roche, Lilly

THU0608 Visualization and Analysis Platform in The Swedish Rheumatology Register for Real-Time Data Feedback

Background The Swedish Rheumatology Quality register (SRQ) started in 1995 and is one of the largest register collecting data on rheumatoid arthritis and other rheumatic diseases in the world. The aim of the register has always been to support health care along with collect data for research. The national coverage is 82.7% in rheumatoid arthritis. However, the extraction of useful real-time data in every day practice have been so far only limited. Objectives To design and implement a tool connected to the SRQ that allows users to monitoring their local data and results of care in comparison to national data. Methods We designed a Visualization and Analysis Platform (VAP) based on R language, using the web framework of Shiny (© RStudio, Inc.). The platform is based on predefined types of analyses such as flexible tabular presentations, cross-sectional and longitudinal comparisons. A user interface with reactive programming was implemented to control the appearance of interactive graphs and tables, and to explore the changes in outcome measures. Results A web-based platform for live visualization of data was developed and linked to the SRQ, a register containing ca. 75,000 patients and 500,000 registered visits, as long as 60,000 prescriptions of biological treatments. The users is able to control the specifics of the data analysis using a flexible interface, and it can visualize the results in a graph as long as in a table, that can easily be download. An explanation of the graph is included, to support the user in the understanding of what is represented. The VAP tool has been used to visualize data on the so called Open Comparison diagrams, 5 quality indicators that have been discussed at national level and that have been used to evaluate performances of care across counties in Sweden. We also applied this tool to the analysis of use of biological treatments in Stockholm (Fig.1). The user is therefore able to visualize how many times a specific biological treatment has been prescribed in his unit and compare it to the national value, as long as select different periods of time of interest and different diagnosis. Conclusions The VAP offers a flexible tool for visualization and analysis of real-time SRQs data and it is aimed to satisfy the growing request of data by clinicians, patients, researchers, care providers and pharma companies. Moreover, the VAP tool by its nature allows the possibility of fast changes in this structure, following the changes in everyday rheumatology care in real-time. Disclosure of Interest None declared

THU0607 Effectiveness and Persistence of Golimumab across Rheumatic Disease Indications in Clinical Practice: Results from The National Swedish Register: Table 1.

Background Real life evidence regarding the effectiveness and discontinuation of golimumab across indications is limited. Objectives To describe the effectiveness and persistence of golimumab in a real-life setting, in rheumatoid arthritis (RA), ankylosing spondylitis (AS), psoriatic arthritis (PsA), and other spondyloarthropathies (SpA), in relation to disease activity at treatment initiation. Methods The national Swedish Rheumatology Quality Register (SRQ) was used to identify patients starting treatment with golimumab between 2009 and 2014. Effectiveness was assessed using the difference between start of treatment and at the 6 months visits (defined as 110–240 days from start of treatment) of different measures of disease activity. Persistence was evaluated using Kaplan Meier curves and Cox proportional hazard model. Results A total of 2984 patients (1366 RA (69% were RF positive), 655 PsA, 499 AS and 464 SpA) started treatment with golimumab. The percentage of bionaïve patients was similar across indications (from 46% in AS to 48% in RA). There was positive response at 6 months across all indications, according to DAS28, HAQ, tender and swollen joint counts (TJC and SJC), patient global health, pain, EQ-5D, CRP and SR (Table 1). Proportion of persistence was highest in AS (69% at 1 year, 56% at 2, and 50% at 3), and was lowest in PsA (60%, 49%, and 42%). Across all indications, the hazard ratio of discontinuation was higher among women, lower among bionaïve compared to non-bionaïve patients, and 34% higher among RA patients >65 years old compared to RA patients 56–65 years old. There was no association between age and discontinuation among patients with other indications, and between DAS28 and HAQ and discontinuation in RA and SpA.Table 1. Baseline characteristics of the golimumab treated population, stratified by indication. Reported numbers are median and interquartile range in parenthesis All (n=2984) RA (n=1366) PsA (n=655) AS (n=499) SpA (n=464) Symptom duration (years) 6 (1–15) 3 (1–13) 6 (2–14) 12 (5–23) 7 (2–13) Concurrent therapy  NSAID 989 (38.98%) 441 (36.78%) 182 (45.16%) 209 (37.06%) 157 (42.32%)  MTX 1252 (49.35%) 793 (66.14%) 71 (17.62%) 271 (48.05%) 117 (31.54%)  Non-MTX DMARD 310 (12.22%) 184 (15.35%) 21 (5.21%) 64 (11.35%) 41 (11.05%)  Percentage bionaïve 47% 48% 47% 46% 48% Among patients switching from previous biological therapy:  Time from start of first biologic (months) 41 (15–80) 48 (16–93) 35 (14–69) 42 (19–72) 30 (13–67)  Number of biologics before golimumab 1 (1–2) 1 (1–2) 2 (1–2) 1 (1–2) 2 (1–2) Difference between start of treatment and the 6 months visits:  ΔDAS28 −1.41 (−2.33; −0.48) −1.38 (−2.41; −0.53) NA NA  ΔHAQ −0.13 (−0.50; 0) −0.13 (−0.50; −0) NA NA  ΔTJC −3 (−6; −1) −3 (−6; 0) NA NA  ΔSJC −3 (−6; 0) −1 (−4; 0) NA NA  ΔPatient global assessment (VAS) −16 (−39; 0) −20 (−40; 2) −22 (−43; −3) −23 (−47; −8)  ΔPain in general (VAS) −18 (−37; 0) −18 (−40; 3) −28 (−48; −2) −26 (−49; −7)  ΔEQ-5D 0.04 (0; 0.26) 0.07 (0; 0.23) 0.09 (0; 0.38) 0.17 (0; 0.5)  ΔCRP −2 (−10; 0) −1.40 (−8.60; 0) −6 (−14.1; 0) −1.40 (−10.40; 0)  ΔSR −4 (−12; 1) −3 (−13; 0) −7 (−19; −1) −4 (−15; 0) Conclusions In this real-life national cohort, patients on golimumab experienced clinical improvement. Persistence is dependent on contextual factors rather than disease activity at start, and is better in AS than in RA. Disclosure of Interest None declared

AB0118 Physical Activity and Risk of Rheumatoid Arthritis in Women

Background Only one study has evaluated the association between exercise and risk of developing RA, showing no association [1]. Objectives To examine the association between physical activity and risk of developing rheumatoid arthritis (RA) in middle age and elderly women from the Swedish Mammography Cohort, a population-based prospective study [2]. Methods Data on physical activity were collected in 1997 by self-administrated food-frequency questionnaire (FFQ). Risk of RA associated with physical activity was estimated using Cox proportional hazard regression models. Results Among 30,112 women born between 1914 and 1948 followed-up from January 1, 2003 to December 31, 2010, 201 RA cases were identified (226,477 person-years). There was a statistically significant 35% lower risk of RA (relative risk (RR), 0.65; 95% confidence interval (CI), 0.43-0.96) among women in the highest category of leisure-time activity (combining more than 20 minute per day of walking/bicycling (median 40-60 minute per day) and more than 1 hour per week of exercise (median 2-3 hours per week)) as compared to women in the lowest category (less than 20 minute per day of walking/bicycling and less than 1 hour per week of exercise). A non-statistically significant decreased risk was observed for household work (-32%) and work/occupation (-15%), while an increased risk was observed for leisure-time physical inactivity (+27%). Daily energy expenditure was not associated with risk of RA. Conclusions This prospective population-based cohort study of women supports the hypothesis that physical activity can be a protective factor in the etiology of rheumatoid arthritis. Our results add to accumulated evidence on benefits of modifiable leisure-time physical activity for prevention of many other chronic diseases. References Cerhan JR, Saag KG, Criswell LA, Merlino LA, Mikuls TR: Blood transfusion, alcohol use, and anthropometric risk factors for rheumatoid arthritis in older women. J Rheumatol 29(2), 246-254 (2002). Harris H, Håkansson N, Olofsson C, Julin B, Åkesson A, Wolk A. The Swedish mammography cohort and the cohort of Swedish men: Study design and characteristics of 2 population-based longitudinal cohorts. OA Epidemiology 2013 Oct 01;1(2):16. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.1737