OBJECTIVES To characterise the incidence rate of skin cancer associated with methotrexate and hydroxychloroquine in older adults with rheumatoid arthritis (RA). METHODS RA patients aged ≥65 years who initiated methotrexate or hydroxychloroquine as their first disease modifying antirheumatic drugs (DMARDs). The primary outcome was new occurrence of any skin cancer (i.e. malignant melanoma or non-melanoma skin cancer; NMSC) based on validated algorithms (positive predictive value >83%). Secondary outcomes were malignant melanoma, NMSC, basal cell carcinoma (BCC), and squamous cell carcinoma (SCC). We estimated the incidence rates (IRs) and hazard ratios (HRs) for each outcome in the 1:1 propensity score (PS)-matched methotrexate and hydroxychloroquine groups. RESULTS We included 24,577 PS-matched pairs of methotrexate and hydroxychloroquine initiators. Compared with hydroxychloroquine (IR 25.20/1,000 person-years), methotrexate initiators (IR 26.21/1,000 person-years) had a similar risk of any skin cancer [HR 1.03 -(95%CI 0.92, 1.14)] over a mean follow-up of 388 days. The HR (95%CI) associated with methotrexate was 1.39 (0.87, 2.21) for malignant melanoma, 1.01(0.90, 1.12) for NMSC, 1.37 (1.13, 1.66) for BCC, and 0.79 (0.63, 0.99) for SCC compared with hydroxychloroquine. CONCLUSIONS In this large cohort of older RA patients initiating methotrexate or hydroxychloroquine as their first DMARD, we found no difference in the risk of skin cancer including malignant melanoma and NMSC. However, for specific components of NMSC, methotrexate initiators had higher risk of BCC but lower risk of SCC compared with hydroxychloroquine initiators.

Background: The multi-biomarker disease activity (MBDA) blood test measures 12 protein biomarkers (IL-6, CRP, SAA, EGF, VEGF, VCAM, MMP-1, MMP-3, leptin, resistin, TNF-RI and YKL40). It uses a validated algorithm to provide a score on a scale of 1-100 for assessing disease activity in patients with rheumatoid arthritis (RA). The MBDA score reflects several molecular aspects of inflammation, including cytokines, acute phase reactants, growth factors, molecular adhesion, metalloproteinases and hormones. Insights gained by understanding how vaccination affects these biomarkers in healthy subjects - in whom the level of inflammation prior to vaccination should be low and stable - may aid the understanding of how vaccination affects patients with RA. Objectives: The goal of this study was to understand how immunization of healthy subjects with the influenza vaccine affects the assessment of inflammation with the MBDA score and its 12 biomarkers. Methods: A 4-strain influenza virus vaccine (Fluarix Quadrivalent, GlaxoSmithKline) was administered intramuscularly to 22 healthy volunteer subjects on October 24, 2018. Serum samples were obtained immediately prior to vaccination (baseline) and 1, 2 and 3 weeks after vaccination. No restrictions were placed on subject activity. Samples were stored at -80 o C until measurement of the 12 MBDA biomarkers for determination of the adjusted MBDA score, hereafter called the MBDA score. (Adjustment accounts for the effects of age, sex and adiposity 1 ). MBDA scores (natural scale) and biomarker concentrations (log scale) were modeled using generalized estimating equations (GEE) that account for correlations between measurements from the same subject at multiple timepoints. Significance of MBDA score change or biomarker concentration change over time was determined by a likelihood ratio test of timepoints. Results: Of the 22 healthy subjects receiving the influenza virus vaccine, 14 (63.6%) were female, with mean (SD) age of 40.0 years (8.9). MBDA scores were low (<30), moderate (30-44) or high (>44) for 15 (68%), 6 (27%) and 1 (5%) subjects at baseline, and this distribution was stable over time (Figure 1). Overall, MBDA scores did not change significantly over time (p=0.48, Figure 2). Mean changes in MBDA score (95% CI) from baseline to weeks 1, 2 and 3 were 0.32 (-3.07, 3.71), 0.82 (-3.03, 4.67) and 2.86 (-1.10, 6.82), respectively (Figure 2); the week 3 value becomes 0.95 (-1.78, 3.68) if the week 3 outlier is removed. Among the 66 post-baseline measurements of change in MBDA score (Figure 2), 3 (5%) exceeded the 95% CI for change in MBDA score in this study (i.e., 14). When assessing the entire cohort across all timepoints, EGF was the only biomarker that demonstrated statistically significant change over time (p=5.6 x 10 -7 ). At weeks 1, 2 and 3, the mean relative concentrations of EGF, compared with baseline, were 0.62 (0.52, 0.74), 0.86 (0.70, 1.06) and 0.62 (0.50, 0.76), respectively. Figure 1 Figure 2 Conclusion: Immunization of 22 healthy subjects with a quadrivalent influenza vaccine did not have a statistically significant effect on MBDA scores during a 3-week observation, and it had minimal effect on the component biomarkers. References: [1]Curtis et al. Rheumatology [Oxford] 2018;58:874 Disclosure of Interests: Daniel Furst Grant/research support from: AbbVie, Actelion, Amgen, BMS, Corbus Pharmaceuticals, the National Institutes of Health, Novartis, Pfizer, and Roche/Genentech, Consultant of: AbbVie, Actelion, Amgen, BMS, Cytori Therapeutics, Corbus Pharmaceuticals, the National Institutes of Health, Novartis, Pfizer, and Roche/Genentech, Speakers bureau: CMC Connect (McCann Health Company), Lauren Lenz Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Megan Horton Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Darl Flake Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Eric Sasso Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Michael E. Weinblatt Grant/research support from: BMS, Amgen, Lilly, Crescendo and Sonofi-Regeneron, Consultant of: Horizon Therapeutics, Bristol-Myers Squibb, Amgen, Abbvie, Crescendo, Lilly, Pfizer, Roche, Gilead

Background: The multi-biomarker disease activity (MBDA) score, adjusted for age, sex and adiposity (MBDA adj ), has been shown to be better than several conventional disease activity measures for predicting risk for radiographic progression (RP) in patients with rheumatoid arthritis (RA). 1 Serologic status and other non-disease activity measures are also predictive of RP risk. Combining them with the MBDA adj should result in a stronger prognostic test for RP than any one measure alone. Objectives: Develop a multivariate model for predicting risk for RP that includes the adjusted MBDA score and other known predictors of RP. Methods: Four RA cohorts were used, two for training (OPERA and BRASS, n=555) and two for validation (SWEFOT and Leiden, n=397). Each pair of cohorts was heterogeneous in disease duration and treatment history. BMI data were not available for one validation cohort, so a BMI surrogate was modeled using forward selection with the two training cohorts and 3 others (CERTAIN, InFoRM, RACER) (N=1411). An RP risk score was then trained using forward selection in a linear mixed-effects regression, considering disease-related and demographic variables as predictors of change in modified total Sharp score over one year (ΔmTSS), with a random effect on cohort. The RP risk score was validated as a predictor of RP with two cutoffs (ΔmTSS >3 and >5) using logistic mixed-effects regression. Odds ratios (OR) and 95% profile likelihood-based confidence intervals (CI) were calculated from the models and significance was assessed by likelihood ratio tests. Risk curves were generated to show probability of RP as a function of the RP risk score. Results: The BMI surrogate included leptin, sex, age and age 2 and correlated well with BMI (ρ = 0.76). In training, the most significant independent predictors of RP were MBDA adj (p = 0.00020), seropositivity (p = 9.3 x 10 -5 ), BMI surrogate score (p = 0.013) and use of targeted therapy (p = 0.0026). The final model was: RP risk score = 0.024 x MBDA adj + 0.093 if seropositive – 0.063 x BMI surrogate score – 0.61 if using a targeted therapy. In validation, the OR (95% CI) of the RP risk score for predicting ΔTSS >3 or >5 were 2.2 (1.6, 3.2) (p = 2.6 × 10 -6 ) and 3.1 (2.0, 5.0) (p = 5.7 × 10 -8 ), respectively (Figure 1). The odds of a patient having RP increases by 50% for each 21-unit or 15-unit increase in MBDA adj , for RP defined as ΔTSS >3 or >5, respectively. Figure 1. Conclusion: A multivariate model containing adjusted MBDA score, seropositivity, a BMI surrogate and use of targeted therapy has been trained and validated as a prognostic test for radiographic progression in RA. References: [1]Curtis, et al. Rheumatology [Oxford]. 2018;58:874 Disclosure of Interests: Thomas Huizinga Grant/research support from: Ablynx, Bristol-Myers Squibb, Roche, Sanofi, Consultant of: Ablynx, Bristol-Myers Squibb, Roche, Sanofi, Michael E. Weinblatt Grant/research support from: BMS, Amgen, Lilly, Crescendo and Sonofi-Regeneron, Consultant of: Horizon Therapeutics, Bristol-Myers Squibb, Amgen, Abbvie, Crescendo, Lilly, Pfizer, Roche, Gilead, Nancy Shadick Grant/research support from: Mallinckrodt, BMS, Lilly, Amgen, Crescendo Biosciences, and Sanofi-Regeneron, Consultant of: BMS, Cecilie Heegaard Brahe: None declared, Mikkel Ǿstergaard Grant/research support from: AbbVie, Bristol-Myers Squibb, Celgene, Merck, and Novartis, Consultant of: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Speakers bureau: AbbVie, Bristol-Myers Squibb, Boehringer Ingelheim, Celgene, Eli Lilly, Hospira, Janssen, Merck, Novartis, Novo Nordisk, Orion, Pfizer, Regeneron, Roche, Sandoz, Sanofi, and UCB, Merete L. Hetland Grant/research support from: BMS, MSD, AbbVie, Roche, Novartis, Biogen and Pfizer, Consultant of: Eli Lilly, Speakers bureau: Orion Pharma, Biogen, Pfizer, CellTrion, Merck and Samsung Bioepis, Saedis Saevarsdottir Employee of: Part-time at deCODE Genetics/Amgen Inc, working on genetic research unrelated to this project, Megan Horton Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Brent Mabey Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Darl Flake Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Rotem Ben-Shachar Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Eric Sasso Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Alexander Gutin Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Elena Hitraya Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Jerry Lanchbury Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Jeffrey Curtis Grant/research support from: AbbVie, Amgen, Bristol-Myers Squibb, Corrona, Janssen, Lilly, Myriad, Pfizer, Regeneron, Roche, UCB, Consultant of: AbbVie, Amgen, Bristol-Myers Squibb, Corrona, Janssen, Lilly, Myriad, Pfizer, Regeneron, Roche, UCB

Background: Rheumatoid arthritis (RA) has been shown a strong genetic association with particular HLA–DRB1 alleles containing shared epitope (SE). However, whether SE is clinically useful in treatment choices is insufficiently investigated 1 and previous studies have presented mixed findings in the role of SE in the response of TNFi therapies 2,3 . Objectives: To assess the role of SE in response to TNFi treatment in real-world RA patients (pts). Methods: Pts enrolled in a large RA registry, Brigham and Women’s Hospital RA Sequential Study, with known SE and received TNFi therapies were included for the analysis. TNFi pts were identified by the first-time use of the drugs between March 2003 to June 2018. For this analysis, all pts were followed up to 1 year. Summary statistics are reported for demographics, serostatus and disease activity (DA) at baseline and follow-up, stratified by SE status. Given the strong association of SE and anti-citrullinated protein antibody (ACPA), the analysis was further stratified by ACPA status. The effect of SE on change in DA was assessed using linear regression model with age, gender, RA disease duration, baseline DA, smoking status, SE, ACPA and ACPA-SE interaction as covariates. Results: Of the 484 TNFi pts included in the study, 68.8% were SE+. SE+ pts (vs SE-) were more likely to be rheumatoid factor positive, have erosive disease and a higher disease duration, irrespective of ACPA status. No difference in the change of DA was observed by SE. In SE- pts, ACPA+ pts had a greater reduction of DA than ACPA- pts (Table 1). After accounting for baseline differences, there was no significant effect of SE status on the mean change from baseline in any of the 3 DA measures.(Figure 1) The change in DA was not associated with ACPA but was significantly affected by disease duration and baseline DA. Table 1. Disease Activity in TNFi Patients, Stratified by SE and ACPA Status Parameter SE+ (1 & 2 alleles, n=333 ) SE- (n=151 ) ACPA+ ACPA– Overall ACPA+ ACPA - Overall (n=264 ) (n=69 ) (n=333 ) (n=90 ) (n=61 ) (n=151 ) Baseline, Mean (SD ) DAS28 CRP 3.94 (1.69) 3.57 (1.61) 3.86 (1.67) 3.85 (1.49) 3.45 (1.65) 3.69 (1.57) CDAI 23.06 (18.13) 18.95 (15.96) 22.25 (17.78) 21.91 (15.96) 17.72 (17.06) 20.26 (16.48) SDAI 24.08 (18.82) 19.96 (16.59) 23.27 (18.45) 22.58 (16.34) 18.55 (17.87) 20.99 (17.01) Follow-up, Mean (SD ) DAS28 CRP 3.42 (1.55) 2.69 (1.32) 3.27 (1.53) 3.19 (1.43) 3.11 (1.53) 3.16 (1.47) CDAI 17.61 (15.53) 12.11 (12.65) 16.51 (15.14) 15.15 (13.35) 14.94 (14.73) 15.07 (13.84) SDAI 18.35 (15.73) 12.45 (12.78) 17.15 (15.34) 15.31 (13.81) 15.71 (15.45) 15.46 (14.38) Change, Mean (SD ) DAS28 CRP -0.48 (1.31) -0.65 (1.53) -0.52 (1.36) -0.52 (1.50) -0.24 (0.93) -0.42 (1.34) CDAI -4.29 (13.16) -4.79 (13.13) -4.39 (13.12) -6.45 (13.56) -2.63 (9.58) -4.99 (12.28) SDAI -4.74 (14.13) -5.07 (13.90) -4.80 (14.05) -6.87 (14.21) -2.97 (10.32) -5.41 (12.98) Figure 1. Linear Regression Model for Change in Disease Activity *Estimates, p-values are shown as data labels on the graphs; The above model is adjusted for age, gender, RA duration, smoking status, SE status, baseline DA, ACPA and ACPA*SE status Conclusion: This real-world study validates the finding from previous studies conducted in clinical settings that SE does not differentiate treatment response for TNFi therapies. References: [1]Saruhan-Direskeneli G, et al. Rheumatology (Oxford ) 2007;46(12):1842-44 [2]Skapenko A, et al. Clin Exp Rheumatol 2019;37(5):783-790 [3]Rigby W, et al. Annals of the Rheumatic Diseases 2019;78(2):263-264 Disclosure of Interests: Joe Zhuo Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, Joshua Bryson Shareholder of: I own shares of Bristol-Myers Squibb Company, Employee of: I am a paid employee of Bristol-Myers Squibb Company, Qian Xia Shareholder of: I own shares of Bristol-Myers Squibb Company, Employee of: I am a paid employee of Bristol-Myers Squibb Company, Niyati Sharma Consultant of: I work as a consultant for Bristol-Myers Squibb Company, Chidananda Samal Consultant of: I work as a consultant for Bristol-Myers Squibb Company, Sonie Lama Shareholder of: I own shares of Bristol-Myers Squibb Company., Employee of: I am a paid employee of Bristol-Myers Squibb Company., Michael E. Weinblatt Grant/research support from: BMS, Amgen, Lilly, Crescendo and Sonofi-Regeneron, Consultant of: Horizon Therapeutics, Bristol-Myers Squibb, Amgen, Abbvie, Crescendo, Lilly, Pfizer, Roche, Gilead, Nancy Shadick Grant/research support from: Mallinckrodt, BMS, Lilly, Amgen, Crescendo Biosciences, and Sanofi-Regeneron, Consultant of: BMS

Objective SB4, SB2, and SB5 are biosimilars of etanercept (ETN), infliximab (INF), and adalimumab (ADA), respectively. This pooled analysis evaluated the immunogenicity of these treatments across three phase III randomized controlled trials of patients with rheumatoid arthritis (RA). Methods Patients had to have at least one anti-drug antibody (ADAb) assessment up to the time of the primary endpoint from each study (week 24 in SB4 and SB5 studies; week 30 in SB2 study). The effect of ADAbs on American College of Rheumatology 20% (ACR20) response and the incidences of injection-site reactions (ISRs)/infusion-related reactions (IRRs) were evaluated. Results The study included 1709 patients. The cumulative incidences of ADAbs were 30.3% in the all-treatments-combined group, 29.1% in the biosimilars combined group, and 31.5% in the reference products combined group. ACR20 response rates were significantly lower in ADAb-positive patients in the all-treatments-combined [odds ratio (95% confidence interval) 1.77 (1.37, 2.27), p < 0.0001], biosimilars combined [2.24 (1.53, 3.30), p < 0.0001], and reference products combined [1.49 (1.06, 2.09), p = 0.0225] groups. ADAb-positive patients also had a higher likelihood of developing ISRs/IRRs in the all-treatments-combined group [0.56 (0.31, 1.01), p = 0.0550], predominantly due to the results observed with SB2 + INF combined rather than with SB4 + ETN or SB5 + ADA combined. Conclusion In this pooled analysis, ADAbs were associated with reduced efficacy in patients with RA treated with biosimilars (SB4, SB2, and SB5) or their reference products (ETN, INF, and ADA). ADAbs were associated with an increased incidence of ISRs/IRRs in those treated with SB2 + INF. Clinical trial registration numbers: NCT01936181 (SB2 study), NCT01895309 (SB4 study), and NCT02167139 (SB5 study).

Background: The mechanistic association of HLA-DRB1 alleles that code a “shared epitope” (SE) with rheumatoid arthritis (RA) is not yet clear. Previous data has suggested the carriage of SE is associated with the production of cyclic citrullinated peptide antibodies (anti-CCP) 1 and severe RA 2-4 . The interrelationship among SE, anti-citrullinated protein antibody (ACPA) positivity and disease outcomes is not fully understood. Objectives: To assess the RA prognosis associated with the carriage of SE, in relation to ACPA positivity. Methods: Pts enrolled in a large RA registry, Brigham and Women’s Hospital RA Sequential Study between March 2003 to June 2018, with known SE and ACPA status were included in the analysis. HLA-DRB1 SE status was determined by allele-specific polymerase chain reaction and DNA sequencing for most of the subjects and by GWAS-based imputation for the rest. Disease activity (DA) was measured at baseline (BL) and 1-year follow-up by DAS28(CRP), CDAI and SDAI. Pts were stratified by SE+ (1 or 2 SE alleles) and SE- (0 alleles) and ACPA status. We analyzed the relationship of SE with ACPA positivity and change in DA by a linear regression model separately. A mediation analysis was used to examine the mediating effect of ACPA on association between SE and change in DA. Results: Out of 926 pts included in the analysis, 65.1% were SE+, of whom 75.6% were ACPA+. In comparison, 51.7% were ACPA+ in SE- pts. SE+ pts were similar with SE- pts in age, gender, BMI and smoking status, but had longer disease duration, were more likely to be rheumatoid factor positive, have erosive disease and higher comorbidity burden irrespective of ACPA status. The differences were more pronounced if the pts were also ACPA+. Adjusting for BL differences, pts with SE 1 and 2 alleles (vs 0) had an odd ratio of 1.97 (95% CI:1.36-2.84; p=0.0003) and 3.82 (95% CI: 2.44-5.98; p<.0001) to be ACPA +, respectively. The regression analysis suggests that SE+ (vs SE-) pts had an average increase in DAS28 (CRP) of 0.22 (p=0.033), CDAI of 2.07 (p=0.045) and SDAI of 2.43 (p=0.029) over a year (Fig 1). Using a mediation analysis, the direct effect of SE+ account for 78.8% to 81.0% of total effect in the increase in DAS28 (CRP), CDAI and SDAI, and the indirect effect mediated by ACPA account for 19.0% to 21.2% (Table 1). Table 1. Mediation Analysis for SE and ACPA Association with Change in DA Parameter Change in DAS28 CRP (N=666 ) Change in CDAI (N=653 ) Change in SDAI (N=629 ) Estimate P-value Estimate P-value Estimate P-value Total Effect of SE on DA chang e 0.22 0.034 2.05 0.047 2.40 0.030 Direct effect of SE on DA change excluding mediation of ACPA 0.17 0.101 1.57 0.140 1.89 0.098 Indirect effect of SE on DA change due to ACPA mediation and interaction 0.04 0.183 0.48 0.133 0.51 0.143 The model is adjusted with other covariates: Age, Gender, Charlson comorbidity score; baseline biologic use, Smoking status, baseline DA, Interaction term (ACPA*SE ) Figure 1. Linear Regression Model for SE Association with Change in Disease Activity *Estimates, p-values are shown as data labels on the graphs; Change in disease activity (DA) = (follow-up DA- baseline DA); The above model is adjusted for age, gender, CCI, baseline DA, baseline biologic use, SE status and smoking status Conclusion: SE is strongly related to ACPA and a greater burden of disease in RA pts. In pts receiving standard treatments including biologics, SE is predictive of a greater increase in DA, which is partially mediated by the presence of ACPA. References: [1] Dayan I, et al., Arch of Rheumatology , 2010;25:012-018. [2] Gregerson PK, et al, Arthritis Rheum . 1987;30:1205-1213. [3] Turesson C, et al. Arthritis Res Ther . 2005;7:R1386-1393. [4] Moreno I, et al. J Rheumatol . 1996;23:6-9. Disclosure of Interests: Joe Zhuo Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, Joshua Bryson Shareholder of: I own shares of Bristol-Myers Squibb Company, Employee of: I am a paid employee of Bristol-Myers Squibb Company, Qian Xia Shareholder of: I own shares of Bristol-Myers Squibb Company, Employee of: I am a paid employee of Bristol-Myers Squibb Company, Niyati Sharma Consultant of: I work as a consultant for Bristol-Myers Squibb Company, Sheng Gao Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, Sonie Lama Shareholder of: I own shares of Bristol-Myers Squibb Company., Employee of: I am a paid employee of Bristol-Myers Squibb Company., Michael E. Weinblatt Grant/research support from: BMS, Amgen, Lilly, Crescendo and Sonofi-Regeneron, Consultant of: Horizon Therapeutics, Bristol-Myers Squibb, Amgen, Abbvie, Crescendo, Lilly, Pfizer, Roche, Gilead, Nancy Shadick Grant/research support from: Mallinckrodt, BMS, Lilly, Amgen, Crescendo Biosciences, and Sanofi-Regeneron, Consultant of: BMS

Background: Novel antibody systems including anti-Carbamylated Protein Antibody (anti-CarP IgG) and anti-Peptidyl Arginine Deiminase Antibody (anti-PAD4 IgG) are emerging as independent diagnostic and prognostic biomarkers for Rheumatoid Arthritis (RA) (ref, 1-3). As such, these antibody systems may add value to rheumatoid factor (IgM) and anti-citrullinated peptide antibody (ACPA IgG), the hallmark antibodies in RA. Objectives: We evaluated the diagnostic performance of Rheumatoid Factor (RF) with antibody systems targeting citrullinated, carbamylated and PAD4 autoantigens in RA. Methods: The cohort consisted of 638 consenting subjects with RA (fulfilling the 1987 or 2010 ACR classification criteria, mean age: 59.8±0.5 years [SEM], 80% female) and a control group of 775 subjects (mean age: 44.7±0.5 years, 85% females, including Systemic Lupus Erythematosus [n=369], primary Sjogren’s Syndrome [n=64], Primary Fibromyalgia [n=85], other connective tissue diseases [n=63], and a group of normal healthy donors [n=194]). Autoantibodies titers from serum were measured using fluoroenzyme immunoassays (anti-RF [IgM] and anti-CCP [IgG]; Phadia Upsala, Sweden), ELISA (anti-CarP [IgG], research use only [RUO], Inova Diagnostics, San Diego) and bead-based AptivaTM technology (anti-PAD4 [IgG], RUO, Inova Diagnostics) in a clinical laboratory accredited by the College of American Pathologists. For each positive antibody (above each cutoff) a score of 1 was assigned and the cumulative presence of the 4 antibodies was determined [range 0-4]. The ability of the biomarkers to distinguish RA from controls was calculated using sensitivity, specificity and interval likelihood ratio (LR). Positive Predictive Value (PPV) was estimated at 10% pre-test probability. Statistics consisted of Mann-Whitney and Chi-square tests. Results: In this cohort anti-CarP IgG (>20 Units) yielded 33.5% sensitivity and 77.9% specificity. Anti-PAD4 (>1000 Units) yielded 35.0% sensitivity and 95.0% specificity. RF IgM (>5 Units/ml) and anti-CCP (>10 Units/ml) were 67.4% and 66.5% sensitive, respectively (87.5% and 97.0% specific, respectively). RA presented 5-fold higher 4-antibody system scores (2.02±0.05) than controls (0.42±0.02); (p<0.01). Scores greater than 2 yielded 42% sensitivity and 98.8% specificity. A total of 82 subjects presented with the full-house 4 antibodies (score = 4) while 81 of them had RA (99.9% specific). Interval LR and PV for each of the 4-antibody score are presented in the Table. There was no difference in the 4-antibody score between RA who fulfilled the 1987 ACR or 2010 ACR criteria (1.99±0.07 vs 2.08±0.09; p=0.40). In the subset of subjects newly diagnosed (less than one year), the average 4-antibody system score for RA (n=33) was 1.72±0.22 (36.3% with score greater than 2) and 0.58±0.12 for other diseases (0% with score greater than 2, 100% specific); (p<0.01). Conclusion: This cumulative combination of antibody systems targeting citrullinated, carbamylated, PAD4 and Fc autoantigens (RF IgM) is highly specific for RA. It may be useful in diagnosing and classifying RA even in symptomatic patients who present early in the course of disease. References: [1] Shi, et al. Proc Natl Acad Sci USA. 2011 108(42):17372-7 [2] Darrah E et al. Sci Transl Med. 2013 2;5186ra65 [3] Verheul, et al. Arthritis Rheumatol. 2018 Nov;70(11):1721-1731 Disclosure of Interests: Thierry Dervieux Shareholder of: Exagen (a diagnostics company not a pharmaceutical company), Employee of: Exagen (a diagnostics company not a pharmaceutical company), John Conklin Employee of: Exagen (a diagnostics company not a pharmaceutical company), Tyler O’Malley Employee of: Exagen (a diagnostics company not a pharmaceutical company), Kelley Brady Employee of: Exagen (a diagnostics company not a pharmaceutical company), Roberta Alexander Employee of: Exagen (a diagnostics company not a pharmaceutical company), Jing Shi Employee of: Exagen (a diagnostics company not a pharmaceutical company), Claudia Ibarra Shareholder of: Exagen (a diagnostics company not a pharmaceutical company), Employee of: Exagen (a diagnostics company not a pharmaceutical company), Michael Mahler Employee of: Inova Diagnostics (Not pharmaceutical, diagnostics company), Joel Kremer Shareholder of: Corrona, Consultant for: AbbVie, Amgen, Bristol-Myers Squibb, Genentech, GSK, Lilly, Pfizer, Regeneron and Sanofi, Employee of: Corrona, Michael E. Weinblatt Shareholder of: Stock option: CanFite, Lycera, Scipher, Inmedix, Grant/research support from: Crescendo Bioscience, Bristol Myers Squibb, Sanofi, Consultant for: AbbVie, Amgen, Bristol-Myers Squibb, CanFite, Corrona, Crescendo, GlaxoSmithKline, Gilead, Horizon, Lilly, Lycera, Merck, Novartis, Pfizer, Roche, Samsung, Scipher, Set Point, Arthur Weinstein Shareholder of: Exagen (a diagnostics company not a pharmaceutical company), Consultant for: Exagen (a diagnostics company not a pharmaceutical company)

Introduction In RA-BEAM (NCT01710358), baricitinib (BARI), an oral selective inhibitor of Janus kinase (JAK) 1 and JAK 2, showed significant improvements in patients (pts) with active RA who had an inadequate response to methotrexate compared to placebo (PBO) or adalimumab (ADA). Objectives To analyse pathways modulated by BARI compared with ADA (both relative to PBO) through 12 wks of treatment. Methods Pts (n=1307) were randomised 3:3:2 to PBO, BARI 4 mg QD, ADA 40 mg q 2 wks. Total RNA extracted from whole blood drawn at baseline (BL), wk4, and wk12 was analysed using the GeneChip Human Transcriptome Array 2.0 (Affymetrix). Data were analysed using a mixed effects model on a log2 transformed response. Results There was little overlap of the immune pathways modulated by both BARI and ADA at wk4 with no significant overlap by wk12. BARI downregulated JAK/Signal Transducer and Activator of Transcription (STAT) signalling pathways, like those induced by IFNs, IL-6, GM-CSF, IL-5, and IL-3. Expression of interferon responsive genes (IRGs) was downregulated by BARI and upregulated by ADA. BARI reduced IRGs by 75% at wk4 in pts that had high IFN gene expression at BL. ADA modulated complement pathways. Of interest, STAT transcripts were reduced at wk4 by BARI (STAT1, 2, 3, 5A, 5B, 6); by wk12 several STATs (STAT 1, 2, 5A) did not differ from PBO. Additional differences were noted in the number of genes modulated by each treatment. BARI modulated more genes than ADA at wks 4 and 12; BARI resulted in more gene modulation at wk12 than at wk4, whereas ADA gene modulation was similar at wks 4 and 12. Both the numbers and types of genes modulated by BARI diverged further from ADA at wk12 than at wk4. Conclusions Gene expression profiling showed significant differences between BARI and ADA treatments. BARI and ADA modulated JAK/STAT or complement pathways, respectively, and the drugs had opposite effects on interferons, indicating different and possibly complementary mechanisms of action of each targeted therapy. Acknowledgements Study support: Eli Lilly and Company and Incyte Corporation. Encore of ACR/ARHP-2017 Annual Scientific Meeting, Nov 4–8, 2016; San Diego, CA, USA. Disclosure of interest None declared

Background Prior studies have demonstrated challenges in developing and validating claims-based algorithms that accurately predict RA disease activity.1 2 The ability to adjust for and predict RA disease activity would be a powerful epidemiological tool for studies that lack direct disease activity measures such as the DAS28. Objectives We used machine-learning methods to incorporate claims and electronic medical record (EMR) data to develop models to predict DAS28 (CRP) as a continuous measure, and to distinguish moderate-to-high disease activity from low activity/remission. Methods We identified 300 adults (≥18 years of age) with RA enrolled in a single academic centre cohort with ≥1 year of linked Medicare insurance claims preceding a DAS28 (CRP) measurement between 2006 and 2010. Of these, 95 had Medicare Part D pharmacy data. From claims we included demographics, co-morbidities, joint replacement surgery, physical therapy visits, numbers of RA-related codes, laboratory values and imaging studies, and healthcare utilisation. For those with Part D pharmacy data we included medications (steroids, analgesics, DMARDs) and switches between drugs. From the EMRs we obtained smoking status, BMI, blood pressure, medication use, laboratory values for seropositivity (RF or anti-cyclic citrullinated peptide antibodies), haematocrit, ESR and CRP. We constructed models with claims only, claims with medications and claims with EMR data. We examined these models with DAS28 (CRP) as a continuous measure and as a binary outcome (moderate/high activity vs low activity/remission). We used adaptive least absolute shrinkage and selection operator (LASSO), which avoids model overfitting by penalising large coefficients and selects a subset of variables by shrinking some coefficients to zero. We used adjusted R2 to compare continuous model fit and C-statistics to compare binary models. Results In models that included DAS28 as a continuous measure, using claims alone explained 11% of the DAS28 variability. Adding medications and EMR data to claims improved the adjusted R2 by 6% (table 1). In models that included DAS28 as a binary outcome (moderate/high activity vs low activity/remission), our claims-only model yielded a C-statistic of 0.68, which increased to 0.79 after inclusion of medications and EMR data. Conclusions Incorporating medications, EMR data and laboratory values into a claims-based index did not significantly improve the ability to predict DAS28 scores as a continuous measure. However, models that include claims, medications and EMR data may be used to reasonably distinguish moderate-to-high disease activity from low disease activity/remission. References [1] Sauer BC, et al. Arthritis Res Ther2017;19:86. [2] Desai RJ, et al. Arthritis Res Ther2015;17:83. Disclosure of Interest C. Feldman Grant/research support from: Bristol-Myers Squibb, Pfizer, K. Yoshida Grant/research support from: Tuition support from Harvard T.H. Chan School of Public Health (partially supported by training grants from Pfizer, Takeda, Bayer and PhRMA)., B. Pan: None declared, M. Frits: None declared, N. Shadick Grant/research support from: BRASS registry, Amgen, Bristol-Myers Squibb, and Mallinckrodt, Consultant for: Bristol-Myers Squibb, M. Weinblatt Grant/research support from: Bristol-Myers Squibb, Amgen, Crescendo Bioscience, Sanofi, Consultant for: Bristol-Myers Squibb, Amgen, Crescendo Bioscience, AbbVie, Eli Lilly, Pfizer, Roche, Merck, Samsung, Novartis, S. Connolly Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, E. Alemao Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, D. Solomon Grant/research support from: Bristol-Myers Squibb, Pfizer, Amgen, Genentech

Background Compared to the general population, RA patients (pts) are less likely to be employed and have lower HRQoL.1,2 However, data on impact of work status on HRQoL in RA pts are limited. Objectives Evaluate association between work status and HRQoL in RA pts. Methods We analysed data from adult pts enrolled in a large sequential RA registry. Physicians assessed pt demographics, clinical characteristics, disease activity and laboratory parameters at baseline (BL) and then annually. Follow-up questionnaires to assess pt-reported outcomes were administered every 6 months and included HRQoL measures (12-Item Short-Form Health Survey physical and mental component summary [SF-12 PCS, MCS], EuroQoL-5 Dimension [EQ-5D]) and work status; higher score indicates better health for all 3 HRQoL measures. General linear mixed models with repeated measures were used for SF-12 analysis and finite mixture models for the EQ-5D analysis, controlling for BL covariates. Results A total of 974 RA pts with HRQoL information were included: 49.3% (n=480) ‘employed for pay’, 38.9% (n=379) ‘employed not for pay’ (retired, homemaker or student), 11.8% (n=115) ‘not employed or on disability’. Pts employed for pay were younger and had lower disease activity compared with other groups (table 1). Compared with pts ‘not employed or on disability’, pts employed had significantly higher PCS (mean [SE] 7.17 [0.82]; p<0.001), MCS (5.39 [0.80]; p<0.001) and EQ-5D (0.48 [0.16]; p=0.0031) scores. Similar results were observed comparing pts ‘employed not for pay’ to pts ‘not employed or on disability’ (Table 2). Conclusions Work status and household income in RA are independently associated with HRQoL. More studies are needed to evaluate impact of work on changes in HRQoL in RA pts. References [1] Uhlig T, et al. J Rheumatol2007;34:1241–7. [2] Chorus AM, et al. Ann Rheum Dis2003;62:1178–84. Disclosure of Interest E. Alemao Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, A. Boonen: None declared, Z. Guo Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, C. Iannaccone: None declared, M. Frits: None declared, M. Weinblatt Grant/research support from: Amgen, Bristol-Myers Squibb, Crescendo Bioscience, Sanofi, Consultant for: AbbVie, Amgen, Bristol-Myers Squibb, Crescendo Bioscience, Lilly, Merck, Novartis, Pfizer, Roche, Samsung, N. Shadick Grant/research support from: Amgen, BRASS registry, Bristol-Myers Squibb, Mallinckrodt, Consultant for: Bristol-Myers Squibb

Background Accurate patient stratification is critical if medical professionals are to adopt a precision medicine approach when planning clinical trials or prescribing medication. This approach results in a superior level of drug response in the target group, a reduction in adverse effects and reduced costs for payers. Best practice treatment recommendations and disease activity markers such as DAS28, as opposed to a treat to target approach, guide current treatment of arthritic disease. There is currently a lack of tools to enable patient stratification, in part due to traditional biomarkers reflecting systemic inflammation rather than the target tissue. Objectives In this paper, we explore the use of a combination of novel tissue specific biomarkers for patient clustering with the objective of identifying different disease profiles. Methods Four biomarker substudy cohorts were pooled for this study, including two RA studies; LITHE (n=574) and OSKIRA-1 (n=131) and two OA studies; SMC1 (n=447) and SMC2 (n=81) all of which have been described in detail before. Whilst the principle focus was to examine RA patient profiles, OA studies were included to enrich the cohort with a non-RA population. OSKIRA and LITHE both had measurements at 24 weeks with additional measurements at 52 week s in LITHE. Several serological biomarkers were measured in each cohort, selected due to the specific tissue metabolite they represent. These included: C2M (cartilage degradation); CTX-I and PINP (bone resorption and formation); C1M and C3M (interstitial matrix degradation); CRPM (CRP metabolite) and VICM (macrophage activity). Each biomarker was log transformed and min-max normalised in order to allow for direct comparison of each of the variables. Patient clustering was performed using Ward hierarchical clustering and the number of clusters determined using the GAP statistic. ANOVA test was used to identify differences in delta change in radiographic scores at 24 and 52 weeks in the RA placebo groups (n=271) only. Results Clustering analysis resulted in five different clusters (A-E). Cluster A and B were both comprised of >98% RA patients. Cluster D was comprised mainly of OA patients whilst clusters C and E were a mix of OA and RA patients. Clusters A and B were characterised by high levels of all biomarkers compared to other clusters except for VICM, which is significantly lower in cluster A than in cluster B (Tukey test p<0.001). Biomarker levels in Cluster C were all close to the median. Cluster D was characterised by low levels of all biomarkers compared to other clusters with significantly lower C2M levels, whilst cluster E also had low levels of markers, yet with significantly higher levels of CTX-1 compared to cluster D. When looking at the RA placebo groups there were no difference in change in SHP score at 24 weeks between the groups, (n=271, LITHE, OSKIRA), but a significant difference in SHP change 52 weeks (n=83, p<0.05, LITHE). Conclusions We have identified putative RA profiles based on novel serological biomarker status. Whether patients in particular clusters may benefit from specific targeted treatments, according to their tissue turnover profile, will be investigated further. Disclosure of Interest None declared

Background Strong genetic association has been reported between RA and human leukocyte antigen (HLA) regions, particularly HLA-DRB1 alleles with the shared epitope (SE). SE alleles are associated with seropositivity, erosions and higher disease activity (DA) in RA. Objectives To evaluate the association between SE alleles and the presence of multiple poor prognostic factors (PPFs) of seropositive (anti-citrullinated protein antibody [ACPA] and/or RF) and erosive RA; as well as changes in DA. Methods We analysed patients (pts) enrolled in a large sequential RA registry established in 2003; most had established RA and annual clinical evaluations. Pts with baseline (BL) data on SE status were included. A commercially available kit (Qiagen, USA) was used for HLA genotyping. HLA-DRB1 serotypes were assessed from DNA sequences using allele-specific polymerase chain reaction methods and categorised as pts with no, 1 or 2 SE alleles. Changes from BL were compared in pts with vs without SE alleles. The association of multiple PPFs and SE status was evaluated using multinomial logistic models. The association between change in DA and SE status was analysed using linear regression models with age, sex, disease duration (DD), co-morbidities and biologic DMARDs as covariates. Results Of 689 RA pts included, no, 1 and 2 SE alleles were reported in 241 (35.0%), 275 (40.0%) and 173 (25.1%) pts, respectively. At BL, pts with SE alleles (vs no SE) were more likely to have PPFs, and had longer DD and higher DA (table 1). The odds ratio (OR) for seropositive erosive RA in pts with 2 and 1 SE alleles (vs no SE) was 5.44 (95% CI 2.39, 12.39) and 2.87 (1.32, 6.23; Fig), respectively. The OR for double seropositivity in pts with 2 and 1 SE alleles (vs no SE) was 4.27 (95% CI 2.51, 7.28) and 2.56 (1.66, 3.94), respectively. A total of 551 pts had DA measures at BL and 1 year follow-up. After controlling for BL covariates, pts with SE (vs no SE) had an average increase in DAS28 (CRP) of 0.24 (p=0.031), CDAI of 2.71 (p=0.027) and SDAI of 3.25 (p=0.013; Table 2). Conclusions Pts with (vs without) SE alleles are more likely to have multiple PPFs; pts with 2 SE alleles are 5 times more likely to be seropositive with erosive RA and 4 times more likely to be double positive. Pts with (vs without) SE alleles also experienced an increase in DA over time with standard-of-care treatment. Disclosure of Interest E. Alemao Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, Z. Guo Employee of: Bristol-Myers Squibb, J. Bryson Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, C. Iannaccone: None declared, M. Frits: None declared, N. Shadick Grant/research support from: BRASS registry, Amgen, Bristol-Myers Squibb, and Mallinckrodt, Consultant for: Bristol-Myers Squibb, M. Weinblatt Grant/research support from: Bristol-Myers Squibb, Amgen, Crescendo Bioscience, Sanofi, Consultant for: Bristol-Myers Squibb, Amgen, Crescendo Bioscience, AbbVie, Eli Lilly, Pfizer, Roche, Merck, Samsung, Novartis

To the Editor: The syndrome of inappropriate secretion of antidiuretic hormone (SIADH) has been reported in diseases involving the central nervous system (CNS). However, SIADH is rarely reported in patients with systemic lupus erythematosus (SLE)1,2,3,4,5,6. We present a case of a 34-year-old woman with a history of SLE who was diagnosed with SLE in 2009 with polyarthritis, leukopenia, oral ulcers, alopecia, photosensitive rash, diffuse proliferative glomerulonephritis, positive antinuclear antibody, and elevated dsDNA. Two months prior to presentation, she was transitioned from mycophenolate mofetil (MMF) 500 mg bid to azathioprine (AZA) 50 mg daily because of the desire to attempt conception. One month later, she started having daily fevers, sore throat, photosensitive rash, tender posterior cervical lymphadenopathy, abdominal pain, and leukopenia. Because of the leukopenia, she was switched from AZA back to MMF. Her worsening leukopenia and elevated inflammatory markers were suggestive of an SLE flare and she was given a higher dose of prednisone. One month later, she developed abdominal pain, nausea, and vomiting. Her medications included … Address correspondence to Dr. N. Yang, 60 Fenwood Road, Suite 3032X, Boston, Massachusetts 02115, USA. E-mail: nyang2{at}partners.org

Background Vobarilizumab is a Nanobody® consisting of an anti-IL-6R domain and an anti-human serum albumin domain in development for treatment of RA. The efficacy and safety were assessed in a 24-week double-blind global phase 2b study in patients with active RA on a stable background of MTX. Main efficacy and safety results were previously reported [1]. Objectives To report the impact of treatment with vobarilizumab on secondary efficacy endpoints including SDAI and CDAI remission and the sustained response at 4 consecutive visits based on ACR50, ACR70 and DAS28CRP. Methods Patients were randomized to receive subcutaneously administered placebo or 1 of 4 dose regimens of vobarilizumab in addition to MTX. SDAI and CDAI remission at Week 24 was evaluated, as was maintenance of efficacy as defined by sustained DAS28CRP<2.6 responses at 4 consecutive visits (i.e., at Weeks 12, 16, 20 and 24). In addition, a post-hoc analysis was performed on sustained ACR50 and ACR70 responses from Week 12 through Week 24. Proportions of patients achieving response for these endpoints were summarized by treatment group. Subjects with missing values were analyzed as non-responders. Results A total of 345 patients were randomized. Demographics and baseline characteristics were similar across groups with mean baseline DAS28CRP between 5.8 and 6.2. At Week 24, up to 19% and 20% in the vobarilizumab groups reached CDAI and SDAI remission, respectively vs. 10% and 9% who received placebo (Table 1). At Week 24, up to 61% and 45% of the patients in the vobarilizumab groups achieved an ACR50 or ACR70 response, respectively (39% and 17% on placebo). Approximately one third of the randomized patients in the 3 highest treatment groups had a sustained ACR50 response from Week 12 through Week 24 (Table 2). Sustained remission defined by DAS28CRP<2.6 at 4 consecutive visits, i.e. at weeks 12, 16, 20 and 24, was observed in 20% to 25% of the patients in the 3 highest dosing arms compared with 3% of those receiving placebo. Conclusions In patients with active RA, treatment with vobarilizumab at the 3 highest dose regimens in addition to MTX had a positive and sustained impact on disease activity through Week 24 as defined by clinically relevant efficacy endpoints. References Weinblatt et al. (Annual Scientific Meeting, Canadian Rheumatology Association, 2017). Disclosure of Interest T. Dörner Consultant for: Ablynx, M. Weinblatt Consultant for: Ablynx, P. Durez Consultant for: Ablynx, R. Alten Consultant for: Ablynx, K. Van Beneden Employee of: Ablynx, E. Dombrecht Employee of: Ablynx, K. De Beuf Employee of: Ablynx, P. Schoen Employee of: Ablynx, R. Zeldin Employee of: Ablynx