Background The inflammatory reflex is a centrally integrated physiological mechanism whereby vagus nerve signaling regulates the production of proinflammatory cytokines that mediate systemic inflammation. The safety and efficacy of a neuroimmune modulation device were assessed in a first-in-human, double-blind pilot study in subjects with multi-drug refractory rheumatoid arthritis (RA). Previously reported data showed that vagus nerve stimulation (VNS) for 12 weeks using the neuroimmune modulation device, a miniature, leadless neurostimulator that is surgically implanted in the neck on the left vagus nerve, improved clinical outcomes, was safe and well-tolerated.[1] We now report 36-month long-term data from this pilot study. Objectives Determine the long-term safety and efficacy of neuroimmune modulation using a novel neuroimmune modulation device. Methods Subjects (N=14) with active RA and prior insufficient response to ≥2 biological or targeted synthetic (b/ts) DMARDs with ≥2 modes of action were implanted with the neuroimmune modulation device.[1] All subjects who completed Week 12 (N=14) were enrolled in an open-label, long-term extension study where subjects continued or if assigned to the sham group, began to receive treatment. Of these subjects, 11 completed 36 months of treatment as follows: VNS for 1 minute either once per day (QD, N=4) or four times per day (QID, N=3); subjects randomized to sham treatment during the first 12 weeks were crossed over to active VNS for 1 minute, randomized to either QD (N=3) or QID (N=1). The addition of a concomitant b/tsDMARD to study treatment was allowed at the discretion of the investigator and the subject. The safety and clinical effectiveness of VNS using the Clinical Disease Activity Index (CDAI) were evaluated. Results The 11 subjects completing the Month 36 visit had previously failed, on average, five different agents, with 73% (8/11) having failed a tsDMARD. The average CDAI change from baseline (i.e., initiation of stimulation) for the 11 subjects completing Month 36 was -17.8 (SD 16.3). At Month 36, 64% (7/11) of subjects achieved a CDAI MCID response from baseline (by the 12pt/6pt/1pt criteria[1]), of whom two (2) were treated with VNS monotherapy and five (5) were on continuous VNS with concomitant b/tsDMARD. The number of subjects in low disease activity (LDA) increased from 9% (1/11) at baseline to 45% (5/11) at Month 36, while the number of subjects in high disease activity (HDA) decreased from 64% (7/11) at baseline to 36% (4/11) at Month 36. Through Month 36, 81% (9/11) subjects elected to combine VNS with a b/tsDMARD, with 8/11 continuing concomitant b/tsDMARD at Month 36. Two adverse events related to VNS therapy occurred during the long-term follow-up period. These events were non-serious, anticipated, and reported by the same subject. One event was a mild sore throat, and the other was moderate tenderness near the implant site. Both events resolved without sequelae following a reduction in stimulation strength. There were no related infections, surgical revisions, or device explantations. Conclusion Despite small numbers, data demonstrate long term use of VNS was safe and effective treatment for RA to maintain or lower disease activity including when added to a b/tsDMARD. Reference [1]Genovese MC, et al. The Lancet Rheumatology. 2020 Sep 1;2(9):e527-38. Acknowledgements: NIL. Disclosure of Interests Norman Gaylis: None declared, David Sikes: None declared, Alan Kivitz: None declared, Diane M Horowitz: None declared, Melissa Evangelista Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, Yaakov Levine Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, David Chernoff Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, Mark C. Genovese: None declared.

Background An urgent need exists for differentiated RA therapies that are safer and cost-effective to expand treatment approaches for non-responders to disease-modifying anti-rheumatic drugs (DMARDs). Electrical stimulation of the vagus nerve activates the inflammatory reflex and has been shown to inhibit the production and release of inflammatory cytokines and decrease clinical signs and symptoms in chronic inflammatory diseases, including rheumatoid arthritis [1]. Objectives The RESET-RA Study (NCT04539964) was designed to determine the safety and efficacy of a novel neuroimmune modulation device for treating rheumatoid arthritis. Presented here are data on the safety of the surgical implantation and use of this device in the first 60 human subjects enrolled in the study. Methods The RESET-RA study is a randomized, double-blind, sham-controlled, multi-center, two-stage pivotal study to evaluate the safety and efficacy of a novel neuroimmune modulation device in patients with moderate-to-severe RA who are incomplete responders or are intolerant to one or more biologic or targeted synthetic DMARDs. The device system (SetPoint Medical, Valencia, CA) consists of 2 implanted components: a miniature, rechargeable, leadless pulse generator that is surgically implanted in the neck on the left vagus nerve and a silicon sleeve referred to as a positioning and orientation device (POD) that holds the generator in close approximation to the nerve; and two external components: a wireless charger and an iPad application for programming the pulse generation. All subjects were implanted with the study device. One to three weeks after the implant procedure, subjects were randomly assigned (1:1) to receive either active or sham stimulation (control). The safety of the surgical procedure, device, and device stimulation was blindly assessed after 12 weeks of stimulation therapy. Results All device implant procedures were completed with no intraoperative complications, infections, or surgical revisions. No unanticipated adverse events (AEs) were reported during the perioperative period and at the end of 12 weeks of follow-up. No study discontinuations were due to AEs, and no subjects died during the study. There were no serious AEs related to the device, stimulation, or explant procedures. There were two serious AEs related to the implant procedure: vocal cord paresis and prolonged hoarseness were reported in two subjects and are known risks of implanting a device on the vagus nerve. The vocal cord paresis resolved following vocal cord augmentation with injectable filler and speech therapy; the other SAE is ongoing and improving with speech therapy. Conclusion Initial results demonstrated that implantation and programming of the novel neuroimmune modulation device was safe, and the surgical procedure and device were well tolerated. Full results from this study, including the clinical efficacy, will be presented after the study is fully enrolled and data is analyzed to determine potential of neuroimmune modulation for treating rheumatoid arthritis. Reference [1]Genovese MC, et al. The Lancet Rheumatology. 2020 Sep 1;2(9):e527-38. Acknowledgements: NIL. Disclosure of Interests Daniel Peterson: None declared, Mark Van Poppel: None declared, Warren Boling: None declared, Perry Santos: None declared, Jason Schwalb: None declared, Howard Eisenberg: None declared, Ashesh Mehta: None declared, Heather Spader: None declared, James Botros: None declared, Frank Vrionis: None declared, Andrew Ko: None declared, David Adelson: None declared, Bradley Lega: None declared, Peter Konrad: None declared, Yaakov Levine Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, David Chernoff Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, Mark Richardson: None declared.

Background: Rheumatoid arthritis (RA) is a disease with significant remaining unmet medical needs for better treatments. Vagus nerve stimulation (VNS) to activate the inflammatory reflex (cholinergic anti-inflammatory pathway) represents a novel experimental therapy for RA. 1 Previously, we reported that inflammatory reflex activation by VNS reduced pro-inflammatory cytokine production and improved disease activity in a 17-patient rheumatoid arthritis (RA) proof-of-concept study using a reprogrammed epilepsy stimulator2; clinical improvement was sustained for 24 months without untoward safety signals. 3 Here we report the 48 months results from this long-term observational study. Objectives: Determine the long-term safety and efficacy of VNS for the treatment of RA Methods: In the primary study, a VNS device was implanted into 17 RA patients, mostly with insufficient response to multiple conventional and biologic DMARDs, on stable background of methotrexate (≤25 mg weekly) therapy 2 . The device electrically stimulated the vagus nerve, 1-4 min/day, over a 12-week open label period. On completion, subjects were offered to enroll into a follow-up study, where the study physicians were given flexibility to alter VNS dosing parameters and/or to add a biologic disease-modifying antirheumatic drug (DMARD) to the treatment regimen to induce disease remission. Clinical disease activity measures and safety were accessed over 4 years. Results: All patients electively continued VNS treatment in the long-term follow-up study, 4 subjects withdrew prior to month 48. Reasons for discontinuation were withdrawal of consent (N=3) and adverse event due to device discomfort (N=1). At the start of the follow-up study, the mean DAS28-CRP, CDAI and HAQ-DI were significantly reduced compared to the pre-implant baseline (mean difference± SD: DAS28-CRP=-1.60± 1.13, p<0.001; CDAI=-21.19± 13.5, p<0.001; HAQ-DI=-0.44± 0.49, p<0.01), and this effect was retained through 48 months. Patients using VNS monotherapy and those using a combination of VNS with biologic DMARDs exhibited stable improvements in DAS28-CRP, CDAI and HAQ-DI at month 48 (Table 1). Improvements were observed for patients who both previously had an insufficient response to targeted biological therapies as well those who had an insufficient response to standard DMARDs. No association was seen between DAS28-CRP and stimulation frequency (Range= 1X-8X/day). There was no difference in the adverse events profile between the two groups. Table 1. Efficacy of VNS treatment. Treatment Reinitiated N=9 VNS Monotherapy N=8 Total N=17 Mo. 24 Mo. 36 Mo.48 Mo. 24 Mo. 36 Mo. 48 Mo. 24 Mo. 36 Mo. 48 Mean change from baseline (SD) DAS28-CRP -2.58 (1.0)*** -2.40 (1.0)** -2.28 (1.3)** -2.61 (1.3)* -1.77 (1.8) -2.0 (1.7) -2.59 (1.1)*** -2.19 (1.2)** -2.17 (1.4)** CDAI -24.06 (8.3)*** -18.02 (13.3) * -16.2 (19.6) -33.5 (11.1)*** -27.8 (16.0)* -27.9 (12.7)* -28.20 (10.5)*** -21.93 (14.5)* -20.83 (17.5)* HAQ-DI -.60 (0.64)* -.63 (0.45)* -.31 (0.60) -.89 (0.69)* -.88 (0.92) -.88 (0.61) -0.73 (0.66)*** -0.73 (0.64)** -0.54 (0.64)* *P<0.05, **P<0.01, ***P<0.001 versus primary study baseline (month -3.5) Conclusion: VNS was safe, well-tolerated, and resulted in significant and clinically important improvements in disease activity measures that were maintained over 48 months. These results support development of VNS devices as a new therapeutic option for RA treatment. References: [1]van Maanen MA, et al. Nat Rev Rheum 2009 [2]Koopman FA, et al. PNAS 2016 [3]Koopman FA, et al. Arthritis Rheum 2018 Disclosure of Interests: Frieda Koopman: None declared, Anne Musters: None declared, Marieke Backer: None declared, Danielle Gerlag Shareholder of: GlaxoSmithKline, Employee of: GlaxoSmithKline, Sanda Miljko: None declared, Simeon Grazio Speakers bureau: Abbvie., Roche, MSD, Eli Lilly, Pfizer, Mylan, Amgen, Fresenius Kabi, Stada, Berlin-Chemie, Sekib Sokolovic: None declared, Yaakov Levine Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, Emmett Glass Employee of: SetPoint Medical, David Chernoff Shareholder of: SetPoint Medical, Adamas Pharmaceuticals, Olly Nutrition, NAIA Pharma, Aquinox Pharma, Consultant of: Adamas Pharmaceuticals, Olly Nutrition, NAIA Pharma, Aquinox Pharma, Crescendo Bioscience, Employee of: SetPoint Medical, Niek de Vries Grant/research support from: AbbVie, Janssen, Ergomed Clinical Research, GlaxoSmithKline, Pfizer, Boehringer Ingelheim, Roche, Consultant of: MSD, Pfizer, Paul P. Tak Shareholder of: GlaxoSmithKline, Employee of: GlaxoSmithKline

Background: Rheumatoid arthritis (RA) is a debilitating chronic disease with an unmet need for additional therapeutic approaches. Activating neuro-immune reflex pathways by stimulation of the vagus nerve (VNS) could represent a novel means of treating RA [1] and other immune-mediated inflammatory diseases. Last year we reported a 12-week proof-of-concept study using a VNS device, approved for drug-resistant epilepsy, showing reduction in the DAS28-CRP clinical disease activity score, with concomitant reductions in TNF and IL-6 levels [2]. Objectives: To understand the long term safety and efficacy of this novel treatment approach, we followed the patients in a 24 months long-term extension study and report on the safety and clinical efficacy data. Methods: VNS devices were implanted into 17 RA patients, mostly with insufficient response to multiple conventional and biologic disease-modifying antirheumatic drugs (DMARDs), on stable background of methotrexate (≤25 mg weekly) therapy. The devices electrically stimulated the vagus nerve, 1–4 min/day, over a 12 week open label period. On completion, subjects were offered to enroll into a follow-up study, where the study physicians were given flexibility to alter VNS dosing parameters and/or to add a biologic DMARD to the treatment regimen. DAS28-CRP and Health Assessment Questionnaire-Disability Index (HAQ-DI) were collected over 2 years. Results: All subjects electively continued on VNS treatment through 24 months of the long term follow-up study. Biologic DMARDs were started in 1 and restarted in 8 of 17 subjects; of these, 4 were non-responders to VNS, and 5 had stable improvement but had not yet achieved disease remission on VNS alone (table 1). At the start of the follow-up study, the mean DAS28–28 and HAQ-DI were significantly reduced compared to the pre-implant baseline (mean difference±SE in DAS28-CRP=-1.60±0.37, p<0.0001; mean difference±SE in HAQ-DI = -0.44±0.21, p<0.037), and the depth of effect was retained through 24 months. At 24 months, there was no significant difference in DAS28-CRP between the subjects using VNS monotherapy or those using a combination of VNS and biologic DMARDs (VNS monotherapy= 3.76±1.77 vs. VNS and biologic DMARD= 3.21±1.44, p<0.54). No difference in the adverse events profile between the two groups was seen.Table 1 Two Year Efficacy of VNS Treatment. Mean DAS28-CRP at primary study baseline (month -3-5) and at visits over 2 years of long term follow up (months 0-24). Conclusions: The data presented here demonstrate that VNS in subjects with RA is associated with a substantial reduction in disease activity that is sustained for 24 months without untoward safety signals. In addition, the data suggest that biological DMARDs can be initiated safely in combination with VNS treatment, though this requires further study in larger cohorts. These results support further development of VNS devices as an alternative therapeutic approach for RA treatment, which potentially can safely be combined with biologic DMARDs. References [1]Van Maanen M, et al. The cholinergic anti-inflammatory pathway: towards innovative treatment of rheumatoid arthritis. Nat Rev Rheumatol2009;5:229–32. [2]Koopman FA, et al. Vagus nerve stimulation inhibits cytokine production and attenuates disease severity in rheumatoid arthritis. Proc Natl Acad Sci USA2016;113(29):8284–9. Disclosure of Interest: F. Koopman: None declared, A. Musters: None declared, M. Backer: None declared, D. Gerlag Shareholder of: GlaxoSmithKline, Employee of: GlaxoSmithKline, which has an interest in SetPoint, S. Miljko: None declared, S. Grazio: None declared, S. Sokolovic: None declared, Y. Levine Shareholder of: SetPoint Medical, Employee of: SetPoint Medical, D. Chernoff Shareholder of: SetPoint Medical, Adamas Pharmaceuticals, OLLY Nutrition, NAIA Pharma, Aquinox Pharma, Consultant for: Adamas Pharmaceuticals, OLLY Nutrition, NAIA Pharma, Aquinox Pharma, Crescendo BioScience, Employee of: SetPoint Medical, N. de Vries Grant/research support from: Abbvie, Janssen Biologics BV, Ergomed Clinical Research, GlaxoSmithKline, Pfizer, Boehringer Ingelheim, Roche, Consultant for: MSD, Pfizer, P.-P. Tak Shareholder of: GlaxoSmithKline, Employee of: GlaxoSmithKline, which has an interest in SetPoint

Objectives: To investigate the multi-biomarker disease activity (MBDA) score by comparison with imaging findings in an investigator-initiated rheumatoid arthritis (RA) trial (HURRAH trial, NCT00696059). Method: Fifty-two patients with established RA initiated adalimumab treatment and had magnetic resonance imaging (MRI), ultrasonography (US), computed tomography (CT), and radiography performed at weeks 0, 26, and 52. Serum samples were analysed using MBDA score assays and associations between clinical measures, MBDA score, and imaging findings were investigated. Results: The MBDA score correlated significantly with MRI synovitis (rho = 0.65, p < 0.001), MRI bone marrow oedema (rho = 0.36, p = 0.044), and US power Doppler (PD) score at week 26 (rho = 0.35, p = 0.039) but not at week 0 or week 52. In the 15 patients who had achieved a Disease Activity Score based on C-reactive protein (DAS28-CRP) < 2.6 at week 26, MRI and/or US detected subclinical inflammation and 13 (87%) had a moderate/high MBDA score. For the cohort with available data, none of the four patients in MBDA remission (score ≤ 25) at week 26 had progression of imaging damage from baseline to week 52 whereas progression was observed in three out of nine (33%) and seven out of 21 (33%) patients with moderate (30–44) and high (> 44) MBDA scores, respectively. Conclusions: In this cohort, the MBDA score correlated poorly with MRI/US inflammation. However, the MBDA score and MRI/US were generally concordant in showing signs of inflammation in most patients in clinical remission during anti-tumour necrosis factor (anti-TNF) therapy. MBDA scores were elevated in all patients with structural damage progression.

Background The Multi-Biomarker Disease Activity (MBDA) score is a validated tool that quantifies 12 biomarkers to assess disease activity in rheumatoid arthritis (RA) patients. Many studies have demonstrated usefulness of the score for assessing RA disease activity. Objectives To determine minimum clinically important change in MBDA score (ΔMBDA) from baseline (BL) to Month 3 (M3) associated with clinical improvement (decrease in DAS-ESR >1.2) in early RA patients after initiating methotrexate (MTX). Methods We evaluated the MBDA test in patients from one of the sites participating in the Solna Epidemiological Investigation of RA (EIRA) cohort. EIRA patients were eligible if they were ≥18 years; RA diagnosis within 12 months of symptom duration; had serum and clinical assessments at BL and M3; and clinical follow-up data in the Swedish Rheumatology Quality Register. Patients naïve to disease modifying anti-rheumatic drugs who received MTX were included. Kruskal-Wallis was used to test the null hypothesis that medians of ΔMBDA scores of 3 EULAR response groups are equal. Receiver operating characteristic (ROC) analysis was performed. The optimal threshold of ΔMBDA associated with DAS28-ESR improvement (decrease in DAS-ESR >1.2 at M3) was determined by Youden criterion maximizing sum of sensitivity and specificity. Results 176 patients were included: 72% women, mean age 51 (SD: 11.7) years, mean DAS28-ESR score 5.6 (SD: 0.99); 51% had ESR <28 mm/hr, 66% were anti-CCP2+, and 22% received prednisone. Mean BL MBDA score was 56.8 (SD: 14.7) with 8 (5%) patients in low (<30), 29 (16%) patients in moderate (30–44) and 139 (79%) patients in high MBDA disease activity categories. Median MBDA scores for patients with no EULAR response worsened by 2 points and for patients with moderate and good response improved by 12 and 16 points, respectively (p<0.0001 across groups, Fig 1A). Median MBDA scores improved by 10 points for all patients and 15 points in patients with a DAS28-ESR decrease >1.2. The best combination of sensitivity and specificity to achieve a DAS28-ESR decrease >1.2 was provided by a ≥8 point MBDA score improvement (Fig 1B). A similar result was obtained using the bootstrap method. AUROC was 0.77 (95% CI: 0.71, 0.84). 125 patients (71%) had concordance between DAS28-ESR improvement and ΔMBDA improvement at the optimal threshold (Table 1). Conclusions The optimal threshold of ΔMBDA score associated with a clinically relevant decrease of DAS28 was 8 points. Using this threshold, the MBDA test is informative to detect clinical improvement. Thus, based on these results improvement in MBDA score ≥8 points at M3 after initiating MTX is indicative of meaningful clinical improvement. Disclosure of Interest K. Chatzidionysiou Consultant for: AbbVie, Pfizer, Eli Lilly, UCB, Roche, A. Hensvold: None declared, S. Saevarsdottir: None declared, R. Bolce Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., D. Chernoff Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., C. Hwang Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., X. Wang Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., A. Catrina Grant/research support from: Roche, Abbvie, Consultant for: BMS, GSK, Pfizer, Roche, Lilly, Abbvie

Background The Multi-Biomarker Disease Activity (MBDA) score has been validated as a disease activity metric in rheumatoid arthritis (RA) patients. Patients initiating new therapy or changing therapy frequently have moderate to high MBDA scores. Understanding short term biological variation of MBDA scores in these patients is important in order to determine a minimally important difference (MID). Objectives To evaluate biological variation in MBDA scores over a 24-hour period and from day to day in patients with clinically stable RA with moderate to high MBDA scores at baseline and to determine the MID in these patients. Methods We performed an analysis of 22 RA patients with moderate or high baseline MBDA scores. Adults with clinically stable seropositive RA (>8 weeks without DMARD and/or biologic medication changes and ≤10 mg prednisone per day) who had MBDA scores of moderate (MBDA 30–44) or high (MBDA >44) were eligible. Serum samples were obtained 5 times over the first 24-hour period (8 AM, 12 PM, 4 PM, 8 PM, and 8 AM); at 12 PM in the next 24-hour period; and at 8 AM the next 2 consecutive days, for a total of 8 timepoints. An additional midnight sample was excluded from the analysis because this timepoint is not relevant to normal clinical practice hours. Diurnal variation was calculated using 5 timepoints over the first 24 hours. Daily variation was determined using 4 timepoints taken at 8 AM on successive days. Combined diurnal and daily variation was calculated using 8 timepoints over 4 days. For each patient, absolute changes in MBDA scores were calculated for all possible pairs of timepoints for: a) diurnal variation (total 220 pairs), b) daily variation (total 132 pairs) and c) diurnal and daily variation (total 616 pairs). MID was calculated as the 90th percentile of absolute changes using all diurnal and daily variation data. Bootstrapping was used to validate the result. Results Of patients included in the analysis, 13 had moderate MBDA scores and 9 had high MBDA scores at baseline. Baseline demographics were: 73% women, mean age 61.8 (SD: 12.1) years, mean MBDA score 43.9 (SD: 8.3), and mean CDAI 20.2 (SD: 17.1). No patients were on glucocorticoids. Based on the analysis of the absolute change of MBDA score in the data with daily and diurnal variation combined, the mean was 3.4 (SD: 3.8), the median (Q1, Q3) was 2 (1, 5), and the MID was calculated as 7. Similar results were obtained using a bootstrap method. Minimal variability in mean MBDA scores was observed over 4 days for patients with moderate and high baseline MBDA scores (Figure 1). Conclusions Based upon the short term biologic variability of moderate and high MBDA scores, the MID was 7 units. An absolute change exceeding this threshold is unlikely due to diurnal and daily biological variation of the MBDA scores. Disclosure of Interest D. Chernoff Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., R. Bolce Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., C. Hwang Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., X. Wang Shareholder of: Myriad Genetics, Inc., Employee of: Crescendo Bioscience Inc., A. Kivitz Consultant for: Genentech, Pfizer, UCB, Janssen, J. Curtis Grant/research support from: Crescendo Bioscience Inc., Consultant for: Crescendo Bioscience Inc.

Background In our previous study (1) from the Swedish Farmacotherapy (SWEFOT) trial (2), we showed that a change of multi-biomarker disease activity (ΔMBDA) score (3) in methotrexate incomplete responders (MTX-IR) was predictive for subsequent response to non-biological triple therapy (TT; a combination of MTX with sulfasalazine and hydroxychloroquine) or to anti-tumour necrosis factor (anti-TNF) therapy. Objectives To evaluate further how ΔMBDA score could be used to predict optimal choice of second-line treatment, by investigating different cut-offs and by comparing this to using the C-reactive protein (CRP) for prediction. Methods 157 MTX-IR patients from the SWEFOT trial with complete data were grouped into those with and without big ΔMBDA (big decrease >22 and non-big decrease ≤22 respectively from baseline to Month 3) based on highest quartile. The change of CRP (ΔCRP) was studied in parallel using a cut-off based on the highest quartile (>28 & ≤28). Analysis was done by last observation carried forward. The proportion of clinical responders (DAS28 ≤3.2) across ΔMBDA groups between the two therapy arms was assessed by Breslow-Day test. Results Among patients with ΔMBDA>22, 76% responded to TT and 47% to anti-TNF, while among the others more responded to anti-TNF (55% vs 38%; cross-comparison for all 4 groups p=0.016; figure 1). When based on ΔCRP responses were 50% and 39% versus 45% and 56% (p=0.226). Conclusions In MTX-IR patients, an improvement in the MBDA scores by >22 predicts better response to subsequent Triple Therapy, and lack thereof predicts a better response to anti-TNF. The same analyses using CRP did not reveal similar results. Thus, monitoring patients during MTX treatment using the MBDA score may guide optimal therapy choice in MTX-IRs on the individual patient level. References Hambardzumyan K, Bolce R, Saevarsdottir S, et al. FRI0005 In Early RA Patients with Non-Response to Methotrexate Monotherapy the Change in Multi-Biomarker Disease Activity Score is Differentially Associated with Subsequent Response to Non-Biological versus Biological Therapy. Annals of the rheumatic diseases. 2014 June 1, 2014;73(Suppl 2):382-3. van Vollenhoven RF, Geborek P, Forslind K, et al. Conventional combination treatment versus biological treatment in methotrexate-refractory early rheumatoid arthritis: 2 year follow-up of the randomised, non-blinded, parallel-group Swefot trial. Lancet. 2012 May 5;379(9827):1712-20. PubMed PMID: 22464340. Centola M, Cavet G, Shen Y, et al. Development of a multi-biomarker disease activity test for rheumatoid arthritis. PloS one. 2013;8(4):e60635. PubMed PMID: 23585841. Pubmed Central PMCID: 3621826. Disclosure of Interest K. Hambardzumyan: None declared, R. Bolce Shareholder of: Myriad Genetics Inc., Employee of: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., S. Saevarsdottir: None declared, K. Forslind: None declared, J. Karlsson: None declared, E. Sasso Shareholder of: Myriad Genetics Inc., Employee of: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., D. Chernoff Consultant for: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., C. C. Hwang Shareholder of: Myriad Genetics Inc., Employee of: Crescendo Bioscience, a wholly owned subsidiary of Myriad Genetics, Inc., R. van Vollenhoven Grant/research support from: Abb Vie, BMS, GSK, Pfizer, Roch, UCB, Consultant for: Abb Vie, Biotest, BMS, Crescendo Bioscience, GSK, Janssen, Lilly, Merck, Pfizer, Roch, UCB, Vertex

Background The prediction of radiographic progression in early rheumatoid arthritis (eRA) patients is important for optimal treatment. We previously demonstrated that a multi-biomarker disease activity (MBDA) score at baseline (BL) was predictive for radiographic progression over the first year of treatment. Objectives To evaluate the MBDA score at different time-points and its change during treatment as a predictor of radiographic progression over the first two years of treatment in eRA. Methods The analyses were performed on radiographic progression of patients from the SWEFOT trial, assessed by van der Heijde modified Sharp scores (SHS) from BL to years 1 and 2 (n=220) and from year 1 to year 2 (n=133); and on the MBDA & disease activity scores (DAS28) and C-reactive protein (CRP) at BL (n=220), month 3 (n=220 & n=205) and year 1 (n=133). Radiographic progression was defined as ΔSHS>5. Mann-Whitney U and Chi-square tests were used for comparisons of disease activity measures between progressors and non-progressors, and for determining significance of proportion of radiographic progressors. Results The median values of MBDA score, CRP (mg/L) and DAS28 at BL for progressors (n=41) and non-progressors (n=179) were 70 and 58 (p=0.001), 28 and 18 (p=0.049), and 6.1 and 5.7 (p=0.136), respectively. After 3 months of MTX therapy the corresponding values were 48 and 40 (p=0.001), 9 and 9 (p=0.213), and 4.8 and 4.0 (p=0.009), respectively. At each time-point patients with low MBDA score had a lower mean ΔSHS and a smaller proportion of subsequent radiographic progressors than those with low CRP or low DAS28 (table). The highest risk for progression from BL to year 1 or 2 (25% and 42% respectively), or from year 1 to year 2 (36%), was observed among patients with high MBDA score at BL which remained high at 3 or 12 months. In contrast, patients with high MBDA score at BL and low MBDA score at months 3 or 12 had much lower risk for progression (6%, 18% and 4% respectively). All patients with persistent low MBDA score throughout 1 year did not progress radiographically over 2 years. Those who had a moderate MBDA score at BL and achieved low MBDA at months 3 or 12 did not progress either. Conclusions MBDA scores at BL and at 3 & 12 months of treatment, as well as change in MBDA category were predictive of subsequent radiographic progression during up to 2 years. At all measured time points a low MBDA score or achievement of the latter was associated with low risk for subsequent x-ray progression. Disclosure of Interest K. Hambardzumyan: None declared, R. Bolce Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, S. Saevarsdottir: None declared, K. Forslind: None declared, I. Petersson Speakers bureau: UCB Pharma, Pfizer, AbbVie, P. Geborek: None declared, S. Ernestam: None declared, E. Sasso Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, D. Chernoff Shareholder of: Crescendo Bioscience, Consultant for: Crescendo Bioscience, S. Cruickshank Consultant for: Crescendo Bioscience, R. Van Vollenhoven Grant/research support: Abb Vie, BMS, GSK, Pfizer, Roche, UCB, Consultant for: AbbVie, Biotest, BMS, Crescendo Bioscience, GSK, Janssen, Lilly, Merck, Pfizer, Roche, UCB, Vertex DOI 10.1136/annrheumdis-2014-eular.3719

Background For patients with early RA (eRA), methotrexate (MTX) is recommended as first-line treatment and in non-responders both the addition of conventional non-biological disease modifying anti-rheumatic drug therapy (triple DMARD therapy) and of biological (anti-TNF) therapy are supported by data. Identification of patients with a higher likelihood of responding to one or the other of these options would lead to more personalised medicine and an increased effectiveness of therapy. Objectives To evaluate the change in the multi-biomarker disease activity (MBDA) score during MTX therapy as a predictor of response to subsequent triple versus biological therapy. Methods Patients with eRA and DAS28>3.2 entered the SWEFOT clinical trial and received MTX monotherapy for 3 months, at which time clinical non-responders (DAS28>3.2) were randomised to receive non-biological triple DMARD therapy (arm A) or anti-TNF (infliximab) therapy with MTX (arm B). For this study, 129 non-responders at month 3 (n=62 from arm A and n=67 from arm B) were analysed by MBDA score at baseline (BL) and month 3. The assessment of changes in the MBDA score (ΔMBDA) from BL to month 3 as a predictor for response to triple or anti-TNF therapy at year 1 was done by defining small (≤6), moderate (7-20) and large (>20) decreases by tertiles. Small and moderate decreases were combined together (small/moderate) and compared versus large decreases for arms A and B. The proportion of patients in arm A versus arm B with response at year 1 was evaluated by the odds ratio (OR) for patients with small/moderate versus large decreases. Homogeneity of the odds ratios between the two cohorts was assessed by Breslow-Day test. Results The mean (median) decreases in MBDA score from BL to month 3 for year 1 responders (n=66) and non-responders (n=63) were 12.9 (10) and 10.8 (9), respectively (p=0.431), and month 3 mean (median) MBDA scores were 47.1 (45) and 50.3 (47), respectively (p=0.336). Patients who had small/moderate decreases in MBDA score during MTX monotherapy, 43% responded to subsequent triple therapy and 57% responded to anti-TNF (OR=0.577). In contrast, among patients with a large decreases in MBDA score from BL to month 3, 67% responded to subsequent triple therapy and 37% to anti-TNF treatment (OR=3.33). Thus the relative treatment effect of arm A versus arm B differed according to the degree of change in the MBDA score from BL to 3 months (p=0.032). Conclusions Among patients with eRA who did not achieve low disease activity on MTX monotherapy, those patients with the greatest decreases in MDBA score were more likely to respond to triple therapy whereas patients with lesser decreases of the MBDA score were more likely to respond to anti-TNF therapy. These findings suggest that in MTX non-responders the changes in MBDA score may help guide subsequent therapy. Disclosure of Interest K. Hambardzumyan: None declared, R. Bolce Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, S. Saevarsdottir: None declared, K. Forslind: None declared, I. Petersson Speakers bureau: UCB Pharma, Pfizer, AbbVie, P. Geborek: None declared, S. Ernestam: None declared, E. Sasso Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, D. Chernoff Shareholder of: Crescendo Bioscience, Consultant for: Crescendo Bioscience, S. Cruickshank Consultant for: Crescendo Bioscience, R. Van Vollenhoven Grant/research support: Abb Vie, BMS, GSK, Pfizer, Roche, UCB, Consultant for: AbbVie, Biotest, BMS, Crescendo Bioscience, GSK, Janssen, Lilly, Merck, Pfizer, Roche, UCB, Vertex DOI 10.1136/annrheumdis-2014-eular.4231

Background A multi-biomarker disease activity (MBDA) score has been validated as a measure of disease activity in rheumatoid arthritis (RA). Objectives This study aimed to further validate the MBDA score in relation to “imaging inflammation” (MRI bone marrow edema and synovitis, US power Doppler (PD) score and grey-scale synovitis (GSS)) and “imaging damage” (CT/MRI/US erosions, radiographs) in an investigator-initiated trial (HURRAH trial, NCT00696059). Methods Fifty-two RA patients with active disease despite conventional DMARD therapy were included. All started adalimumab treatment at baseline and continued on methotrexate. MRI, CT, radiographs and US were performed at baseline, week (w)26 and w52. Blood tests were drawn at baseline, w2, 6, 12, 26, 39 and 52 for measurement of 12 serum biomarkers (VCAM-1, EGF, VEGF-A, IL-6, TNF-RI, MMP-1, MMP-3, YKL-40, leptin, resistin, SAA, CRP) used to generate a Vectra® DA algorithm score. Associations between MBDA score, DAS28-CRP and imaging inflammation and damage were evaluated using Spearman9s rank correlations (rho). Smallest detectable difference (SDD) was defined as 2 SD of the intrareader difference. Results Median (IQR) MBDA score was 63 (47-74) at baseline, declined to 48 (35-65) already at w2 and remained at roughly this level at all following visits. Change in MBDA score correlated with change in DAS28 from baseline to w26 (rho=0.71, p<0.001). MBDA score at w26 correlated with MRI synovitis (rho=0.43, p=0.016), MRI bone marrow edema (rho=0.36, p=0.044) and US PD score (rho=0.35, p=0.039), but not US GSS score (rho=0.10, p=0.56). 10 patients (19%) had progressive joint damage above the SDD by one or more imaging modalities from baseline to w52. Two of these 10 patients were in clinical remission (DAS28<2.6), but none were in remission or low disease activity by the MBDA score at w26. The area under the curve (AUC) for the association between MBDA score with change in CT erosion score from baseline to w52 was borderline significant (rho=0.33, p=0.059). No similar trend was observed for AUC of DAS28 (rho=0.15, p=0.42) Baseline MBDA score and change/percent change in MBDA score at w2 or w6 generally did not correlate to change in imaging scores at w26 and w52. However, percent change in MBDA score at w6 correlated to change in MRI synovitis score from baseline to w26 (rho=0.44, p=0.009), and percent change in MBDA score at w2 correlated to change in Sharp/van der Heijde total score from baseline to w26 (rho=0.45, p=0.004). In similar analyses with DAS28 no correlations were found at 0.01 significance level. 15 patients achieved clinical remission (DAS28<2.6) at w26. Inflammation by MRI and/or US was detected in all of these patients; 6 (40%) had a high MBDA score, 7 (47%) had a moderate MBDA score, 0 (0%) had a low MBDA score and only 2 (13%) were categorized as in MBDA remission. Conclusions This study further validates the MBDA score as a measure of disease activity in RA, as it correlates with imaging measures of inflammation and joint damage. Disclosure of Interest S. Krabbe: None declared, R. Bolce Shareholder of: Shareholder of Crescendo Bioscience, Employee of: Employee of Crescendo Bioscience, C. Brahe: None declared, U. Døhn: None declared, G. Wu Employee of: Employee of Crescendo Bioscience, B. Ejbjerg: None declared, M. Hetland: None declared, E. Sasso Shareholder of: Shareholder of Crescendo Bioscience, Employee of: Employee of Crescendo Bioscience, D. Chernoff Shareholder of: Shareholder of Crescendo Bioscience, Consultant for: Consultant for Crescendo Bioscience, M. Hansen: None declared, L. Knudsen: None declared, A. Hansen: None declared, O. Madsen: None declared, M. Hasselquist: None declared, J. Møller: None declared, M. Østergaard: None declared DOI 10.1136/annrheumdis-2014-eular.4460

Background In early rheumatoid arthritis (eRA), predictors of radiographic damage would be very useful for optimal targeting of available therapies. It has been suggested that combining various biomarkers may provide improved prediction. The multi-biomarker disease activity (MBDA) blood test has been validated as a measurement of rheumatoid arthritis (RA) disease activity. MBDA scores range from 1-100 and are based on 12 serum biomarkers: VCAM-1, EGF, VEGF-A, IL-6, TNF-RI, YKL-40, MMP-1, MMP-3, leptin, resistin, SAA, CRP. In the SWEFOT study, patients with eRA (symptom duration < 1 year) were started on methotrexate (MTX); at 3 months responders (DAS28 ≤ 3.2) continued MTX monotherapy and were followed in regular care, whereas non-responders were randomized to receive either triple DMARD therapy or the addition of infliximab. Objectives To study the baseline MBDA score as a predictor of radiographic progression in eRA. Methods Analyses were performed for 296 patients with complete clinical datasets consisting of baseline assessments of DAS28 (based on ESR), DAS28-CRP, CRP and the MBDA score using the validated algorithm; and for a subset of 235 patients that also had radiographs at BL and 1 year (assessed using the van der Heijde modified Sharp score [SvdH]). Spearman’s correlation coefficients (r) were determined for the baseline MBDA score versus the change in SvdH over 1 year. Results For the 296 patients, mean baseline values were: DAS28 = 5.7 and MBDA = 59. For the 235 patient-subset the mean baseline SvdH was 4.4 (median 2.0). Mean change in SvdH from BL to 1 year was 3.1 ± 6.0 (median 1.0). There was a significant correlation between baseline MBDA score and change in SvdH from baseline to 1 year: r = 0.271 (p<0.001); correlations of DAS28, DAS28-CRP, and CRP with change in SvdH were weaker: r = 0.063, 0.014, and 0.178, respectively (p=NS, p=NS, p=0.006). Correlations between BL MBDA and change in SvdH were similar and statistically significant when tested in month-3 responders (n=84) or non-responders (n=181) separately. Conclusions In eRA patients treated initially with MTX, the MBDA score at baseline correlates significantly with radiographic progression during the first year, irrespective of whether the patient responds to MTX or not. Disclosure of Interest K. Hambardzumyan: None Declared, R. Bolce Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, G. Cavet Consultant for: Crescendo Bioscience, D. Chernoff Shareholder of: Crescendo Bioscience, Consultant for: Crescendo Bioscience, D. Haney Shareholder of: Crescendo Bioscience, Employee of: Crescendo Bioscience, K. Forslind: None Declared, I. Petersson: None Declared, P. Geborek: None Declared, S. Ernestam: None Declared, R. van Vollenhoven Grant/research support from: AbbVie, BMS, GSK, MSD, Pfizer, Roche, UCB, Consultant for: AbbVie, BMS, GSK, MSD, Pfizer, Roche, UCB

Background and Objectives The CAMERA II study (Computer Assisted Management in Early RA) demonstrated that the addition of prednisone versus placebo to a MTX-based tight-control strategy increased effectiveness of therapy and reduced need for biological treatment. The present study evaluated changes in biomarker levels over time with MTX+placebo and MTX+prednisone treatment, using the Multi-Biomarker Disease Activity (MBDA) blood test. Materials and Methods Clinical and biomarker assessments were performed at monthly visits to 1 year for 92 patients who had MBDA test results available at baseline and ≥1 subsequent visit. Average number of visits with non-missing disease activity measures was 3.7 per patient. Concentrations of 12 serum biomarkers were combined to produce a score between 1 and 100 using the MBDA algorithm, which generates a validated measure of disease activity. Biomarker responses were also assessed individually but only to 5 months, to avoid subsequent protocol-mandated exposures to non-MTX DMARDs for insufficient responders. Association between DAS28-ESR response and MBDA response was assessed using Spearman’s correlation. Changes from baseline and comparisons of change over time for MTX+placebo versus MTX+prednisone were analysed by t-tests. Biomarker concentrations were analysed as fractions relative to baseline using Mann Whitney U tests. Results Changes from baseline to 1 year in DAS28-ESR and MBDA scores showed a significant correlation in the MTX+placebo arm (r = 0.57, p < 0.001, n = 31) and the MTX+prednisone arm (r = 0.57, p = 0.002, n = 28). Improvements in DAS28-ESR (p < 0.001) and MBDA (p = 0.01) scores were observed as early as 1 month post-BL in the MTX+prednisone arm. Significant reduction in disease activity in the MTX+placebo arm was first observed at 2 months for DAS28-ESR (p = 0.02) and 4 months for MBDA (p = 0.03). Overall, DAS28-ESR and MBDA response profiles were similar, with mean changes at month 5 for MTX+placebo and MTX+prednisone being –2.2/–4.2 for DAS28-ESR, and –13/–21 for MBDA score. Individual biomarker response profiles differed: for some biomarkers, MTX+placebo had little/no effect but MTX+prednisone had significant effect (e.g., MMP-1, TNF-R1, VCAM-1); for others, MTX+placebo had a significant effect that was augmented (e.g., CRP, IL-6, VEGF) or not affected (SAA) by prednisone. Conclusions Responses assessed with the biomarker-based MBDA test and DAS28-ESR were greater and more rapid for therapy with MTX+prednisone than MTX+placebo, even though individual biomarkers differed in their response profiles. The MBDA test may be useful in combination with clinical assessment to evaluate early response to therapy with MTX or MTX+prednisone.