Background: Pentraxin-3 (PTX) levels are elevated in the serum and synovium of rheumatoid arthritis (RA) patients, and high PTX3 levels have been linked to increased disease activity and radiographic damage. Anti-PTX3 circulating antibodies have been found in up to 40% of patients with rheumatic musculoskeletal diseases, and they have been linked to improved outcomes in patients with systemic lupus erythematosus and vasculitis. However, the role of anti-PTX3 antibodies in RA patients has yet to be determined. Objectives: The aim of this study was to determine the levels of anti-PTX3 antibodies in the serum of RA patients and to describe their associations with disease characteristics. Methods: We conducted a single-center, cross-sectional, prospective study on consecutive adult patients diagnosed with RA using the 2010 ACR/EULAR criteria. An expert rheumatologist confirmed the diagnosis. Anti-PTX3 antibodies were measured in serum using a standardised in-house enzyme-linked immunosorbent assay (ELISA) method that had previously been validated in a larger population of patients with connective tissue diseases, including RA. The anti-PTX3 antibody titre was expressed as optical density (OD) at 405 nm. A positivity cut-off was previously determined by receiving operating characteristic (ROC) curve analyses in which sensitivity was calculated in 130 patients with SLE and specificity in 130 healthy subjects at a value of 0.234 (sensitivity 46.2% and specificity 92.8%) [1]. Demographic, clinical, laboratory, and treatment data were collected. Independent examiners evaluated disease activity. At the 6-month follow-up, disease flares were defined as a CDAI increase of >4.5 [2]. Results: The study included 83 patients, 85.5% of whom were females with long disease duration and low disease activity (Table 1). Anti-PTX3 antibodies were found in the blood of all 83 participants, with a median OD titre of 0.111±0.107. A titre above the pre-established cut-off for positivity was found in 7/83 (anti-PTX3+, 8.4%) (median OD titre: 0.341±0.105 in anti-PTX3+ vs. 0.106±0.08 in anti-PTX3-; p<0.001). Anti-PTX3+ and anti-PTX3- patients were similar in terms of age, disease duration, serum CRP and ESR, and erosive disease. All anti-PTX3+ patients tested negative for rheumatoid factor and anti-citrullinated peptide antibody (ACPA) (p=0.013), and none had extra-articular RA (p=0.096; Figure 1A). The absence of ACPA (p=0.008) and extra-articular disease (p=0.013) were significantly associated with anti-PTX3 antibody titres in multivariable linear regression, independent of DAS28. The anti-PTX3 titre was slightly but significantly higher in patients treated with tumor necrosis factor inhibitors (TNFi) (0.149±0.090) than in patients treated with non-TNFi biologics (0.103±0.047; p=0.015; ANOVA; Figure 1B). At six months, we had eight flares, four of which required treatment escalation (bDMARD switch). All flares occurred in anti-PTX3- participants, while none of the anti-PTX3+ RA patients experienced disease flares after 6 months of follow-up. Conclusion: Significant titers of anti-PTX3+ antibodies were detected exclusively in seronegative RA patients and were associated with the absence of extra-articular features and disease flares at 6 months and the use of TNFi. Although preliminary and requiring validation in larger cohorts, these data suggest that anti-PTX3 antibodies may characterise a subset of seronegative RA with a good response to therapy. REFERENCES: [1] Bassi N, Zampieri S, Ghirardello A, Tonon M, Zen M, Cozzi F, Doria A. Pentraxins, anti-pentraxin antibodies, and atherosclerosis. Clin Rev Allergy Immunol. 2009 Aug;37(1):36-43. doi: 10.1007/s12016-008-8098-6. PMID: 19016000. [2] Konzett V, Kerschbaumer A, Smolen JS, et alDefinition of rheumatoid arthritis flare based on SDAI and CDAIAnnals of the Rheumatic Diseases 2024;83:169-176. Acknowledgements: NIL. Disclosure of Interests: None declared.

Background : Mortality rates in patients with rheumatoid arthritis (RA) are 1.5-1.6 fold higher than in the general population 1,2 . No recent data on mortality in large cohortsof RApatientsinItalyare available. Objectives : The aim of this study was to assess standardized mortality ratios (SMRs) and multiple causes of death in RA subjects living in the Veneto Region between

Background::In recent years several biosimilars (BS) of tumour necrosis factor inhibitors (TNF-i) were introduced. At the Padova University Hospital the first BS of etanercept (bsETN) was available in October 2016 and the BS of adalimumab (bsADA) was available in November 2018.Objectives:The objectives of the study were to evaluate the rate of bioriginator-biosimilar (BO-BS) switch in all patients with rheumatoid arthritis (RA), psoriatic arthritis (PSA) and axial spondiloarthritis (axSpA) in the cohort of the Padova University Hospital and to examine factors favouring BO-BS switch. Secondly, we investigated survival of BO-BS switch and BO treatment and factors associated with longer treatment survival.Methods:We considered all patients on ETN originator (boETN) treatment when the first bsETN was available (1st October 2016) and all patients on ADA originator (boADA) when bsADA was available (1st November 2018). Patients were followed until 30 August 2019 and were classified as BO-BS switchers if they underwent a switch from either boETN or boADA to BS during the follow-up, otherwise they were considered as continuing BO treatment. Factors associated with BO-BS switch were tested with a multivariable regression analysis. To test the survival of the BO-BS switch and of the BO treatment, Cox regression analysis was used including all variables achiving a p<0.10 in univariate analysis tested with Log-rank test and Kaplan-Meier curves.Results:Among 1208 patients (553 RA, 433 PSA, 215 axSpA), 560 (46.3%) patients switched to bsETN (391) or bsADA (169). Mean disease duration was 16 (14.2) years and mean duration of the bDMARD treatment was 96.3 (56.8) months. After adjustment for potential confounders, factors associated with BO-BS switch were a longer disease duration, a shorter duration of previous bDMARD treatments and diagnosis (Tab.1) RA patients had almost a 3 fold increased likelihood of being switched to BS compared to PSA and axSPA, while difference between PSA and axSPA was not significant.Following Cox regression analysis we observed a longer drug survival in BO-BS switchers compared to those continuing with BO (HR 1.38; 95% C.I. 1.2-1.58; p<0.001) (Fig. 1). A longer drug survival was also associated with a longer disease duration (.15years: HR 1.75; 95% C.I. 1.5-2; p<0.001), longer mean duration of previous bDMARDs (.5years: HR 4.1; 95% C.I. 3.5-4.7; p<0.001), and diagnosis (RA vs PSA: HR 1.22; 95% C.I. 1.02-1.47; p=0.030; RA vs axSpA: HR 0.89 95% C.I. 0.067-0.97; p=0.023; PSA vs axSpA: HR 0.66; 95% C.I. 0.57-0.77; p<0.001) (Fig 2).Figure 1.Kaplan-Meier curves for treatment survival, Log-rank test.Figure 2.Kaplan-Meier curves for treatment survival in all patients, Log-rank tesConclusion:BO-BS switch was undertaken in almost half of the patients. Patients with longer disease duration and longer bDMARD duration, were the most likely to be switched successfully to BS. BO-BS switching does not affect the survival of the treatment, indeed, it provides sustained effectiveness particularly if undertaken in patients with stable disease activity.Table 1.Factors associated with BO-BS switch, multivariate regression analysis.Disclosure of Interests:DAVIDE ASTORRI: None declared, Francesca Ometto: None declared, LARA FRISO: None declared, BERND RAFFEINER: None declared, Costantino Botsios: None declared, Andrea Doria Consultant of: GSK, Pfizer, Abbvie, Novartis, Ely Lilly, Speakers bureau: UCB pharma, GSK, Pfizer, Janssen, Abbvie, Novartis, Ely Lilly, BMS

Background: Hyperbaric Oxygen Therapy (HOT) proved effective in improving of symptoms of patients affected by fibromyalgia syndrome (FMS) [1]. Objectives: The objective of the present study was to evaluate the effectiveness of HOT compared to hyperbaric treatment with no oxygen therapy (PBO) in the symptoms and working ability in FMS. Methods: We conducted a prospective trial in employed patients with FMS, randomly assigned to HOT or PBO. Patients and evaluating clinicians were blinded to the treatment. HOT arm comprised 40 sessions, 5 days/week, 120 minutes, 100% oxygen at 2ATA; PBO comprised the same sessions without oxygen. Evaluations were at baseline, after 4 (T1) and 8 weeks (T2). Parameters considered were: socio-demographics, biochemistry, clinical evaluation and patient-reported outcomes (PROs). Baseline assessment included questions BELIEF (how much do you hope to improve with this treatment) and HOPE (how much do you expect to improve with this treatment), with VAS response. Spearman’s, Mann-Whitney’s, Kruskal-Wallis and Fisher’s Exact test were used. Results: 12 patients were included and completed the study, 6 in each arm (Tab. 1). No significant difference was observed in clinical measures or PROs at T1 and T2 between HOT and PBO arms, except for Working Productivity and Activity Impairment Questionnaire (WPAI) (result III) (Tab. 2). In both arms, disease duration was associated with worse PROs (Widespread pain index r=0.59,p=0.037, Severity Score r=0.81,p=0.025); higher BMI with improvement in function at T2 (r=0.63,p=0.027); higher baseline scores in BELIEF with reduction symptoms number (r=-0.67, p=0.021), higher scores in HOPE with reduction in Health Assessment Questionnaire (r=-0.057, p=0.039) Table 1. Patients characteristics All HOT PBO Number 12 6 6 Age* 55,5 (44;59,75) 55,5 (47,75;60) 51 (41;58,75) Females** 6 (100) 6 (100) 6 (100) Disease duration * 10 (8,25;26,75) 9,5 (7,5;20,75) 15 (10;26,75) BMI* 25,5 (22,25;31) 25 (22,75;28) 28,5 (23;31,75) Smoke** 0,5 (0;1) 0 (0;1) 1 (0,25;1) HOPE score* 0 (0;1) 0,5 (0;1) 0 (0;0) >=80** 9 (64,3) 5 (62,5) 4 (66,7) BELIEF score* 70 (60;80) 75 (67,5;83,75) 62,5 (60;68,75) >=70** 8 (57,1) 6 (75) 2 (33,3) *median (IQR); **number (%) Table 2. Change from baseline in clinical measures and PROs at T2. Medians (IQR ) P value ≧ 20 Percentage amelioration No. (% ) P value All HOT PBO All HOT PBO Short Form-36 - Physical 1 (-1;4) 3 (-1;5,5) 0,5 (-0,75;3,25) ns 0 (0) 0 (0) 0 (0) ns SF-36 - Mental 4 (-1;8) 6 (2,5;9,5) 1 (-1,75;5,25) ns 0 (0) 0 (0) 0 (0) ns Severity score - total -13 (-15;-2) -15 (-17;-13,5) -3 (-5,5;-0,5) ns 2 (18,2) 2 (40) 0 (0) ns Number symptoms -3 (-4;0) -4 (-4,5;-1,5) -2 (-3;-0,25) ns 2 (18,2) 2 (40) 0 (0) ns SASP score -1 (-3;0) -2 (-3,5;-0,5) -1 (-1,75;-0,25) ns 1 (9,1) 0 (0) 1 (16,7) ns Widespread pain index 0 (-2;0) 0 (-2;0) -0,5 (-1,75;0) ns 2 (18,2) 2 (40) 0 (0) ns Tender Points Count (0-18 -2 (-3;-0,5) -2 (-3,5;-1) -1,25 (-2;-0,13) ns 3 (27,3) 2 (40) 1 (16,7) Ns Health Assessment Questionnaire 0 (0;0) 0 (-0,5;0,5) 0 (0;0) ns 3 (27,3) 1 (20) 2 (33,3) ns Fibromyalgia Impact Questionnaire 10 (0;15) 10 (7,5;15) 2,5 (-3,75;12,5) ns 1 (9,1) 0 (0) 1 (16,7) ns WPAI result-2 -0,5 (-1,25;-0,13) -1,5 (-2;-0,5) -0,25 (-0,5;0,19) ns 3 (21,4) 0 (0) 3 (37,5) ns WPAI result-3 0 (-2,38;0) -2,75 (-3,88;-2) 0 (0;0) ns 5 (35,7) 0 (0) 5 (62,5) P=0.008 WPAI result-1 -1 (-3,25;-0,38) -3,5 (-4,75;-1,5) -0,5 (-1;0,38) ns 6 (42,9) 2 (33,3) 4 (50) ns Conclusion: 8-week HOT treatment does not substantially improve symptoms in FMS compared to PBO. All patients on hyperbaric treatment may experience amelioration of symptoms: other factors should be considered, including beliefs and expectations on the treatment. References: [1]DOI:10.1371/journal.pone.0127012 Disclosure of Interests: None declared

Background: Three definitions of refractory rheumatoid arthritis (RRA) have been proposed: Buch’s (B-RRA), i.e. failure of ≥1 anti-cytokine and ≥1 cell-targeted bDMARD [1]; Kearsley-Fleet’s (KF-RRA), i.e. exposure to ≥3 bDMARDs classes [2]; De Hair’s (DH-RRA), i.e. signs and/or symptoms of RA activity and failure of ≥1 csDMARD and ≥2 bDMARDs [3]. Objectives: To evaluate the rate of RRA according to the three definitions in a monocentric cohort with two cross-sectional analyses in 2012 and 2019. We investigated also the major determinants of each definition. Secondary objective was to evaluate the most frequent treatments in RRA patients. Methods: Patients affected by RA followed at Padova University Hospital were included at two different time points. In the 2012 cohort patients on bDMARDs on 31 st December 2012 and in the 2019 cohort patients on b/target synthetic DMARDs (tsDMARD) on 1 st March 2019. Factors independently associated with RRA definitions were tested with multivariable regression analysis, including all variables achieving a p<0.10 in the univariate analysis. Results: We included 260 patients in the 2012 cohort and 571 in the 2019 cohort. Rate of RRA in 2012 cohort was: 23 (8.8%) B-RRA, 57 (21.9%) KF-RRA and 12 (4.6%) DH-RRA; rate of RRA in 2019 cohort was: 165 (28.9%) B-RRA, 96 (16.8%) KF-RRA and 57 (10%) DH-RRA. Following multivariate regression analysis, in the 2012 cohort a significant association was found between number of bDMARDs treatment and all RRA definitions [Tab.1]. Also in the 2019 cohort the variable associated with all RRA definitions was the number of bDMARDs treatment [Tab.2]. Both in 2012 and 2019, IL6-inhibitors were more frequently prescribed in RRA patients; instead TNF inhibitors were less frequently prescribed in RRA. Conclusion: Rate of RRA in the 2019 cohort was 10-30% which is higher compared to the 2012 cohort. This might be explained by the fact that RRA definitions are mainly affected by the number of bDMARDs. Thus, an accurate RRA definition should consider not only the number of treatments but also the current disease activity. References: [1]Buch MH. Ann Rheum Dis 2018;77:966–969 [2]Kearsley-Fleet L, et al. Ann Rheum Dis 2018;77:1405–1412 [3]De Hair MJH et al. Rheumatology 2018;57:1135-1144 Table 1. Factors associated with three definitions of RRA in the 2012 cohort, multivariate analysis Characteristics OR (95% C.I.) p value B-RRA CRP per mg/L increase 0,81 (0,68-0,95) 0,011 HAQ per unit increase 3,28 (0,85-12,54) 0,84 Combination with any csDMARD 4,61 (0,65-32,59) 0,124 bDMARD treatment duration per year increase 0,58 (0,52-1,03) 0,114 No bDMARDs 91,0 (7,87-1055,58) <0,001 Model constant 0,009 KF-RRA PDN per mg increase 1,49 (0,87-2,56) 0,144 DAS28 per unit increase 4,22 (1,4-12,71) 0,011 No bDMARDs - (0-0) 0,99 Model constant 0,989 DH-RRA Combination with any csDMARD 6,24 (0,69-56,61) 0,614 Comorbidity 12,82 (0,46-1122,91) 0,065 No bDMARDs per unit increase 6,25 (2,75-121,39) 0,003 Model constant 0,002 B-RRA refractory RA according to Buch, KF-RRA refractory RA according to Kearsley-Fleet, DH-RRA refractory RA according to De Hair Table 2. Factors associated with three definitions of RRA in the 2019 cohort, multivariate analysis Characteristics OR (95% C.I.) p value B-RRA No. bDMARDs 18.77 (11.06;31.85) p<0.001 PDN daily dose per 5 mg increase 2.15 (1.17;2.15) 0.014 Model Constant p<0.001 KF-RRA BMI per 5 unit increase 0.61 (0.35;1.05) 0.072 No. bDMARDs 8.69 (5.16;14.64) p<0.001 bDMARD treatment start year per year increase 0.92 (0.84;1.01) 0.087 Model Constant 0.069 DH-RRA No. bDMARDs 3.9 (2.67;5.71) p<0.001 bDMARD treatment start year per year increase 0.91 (0.83;0.99) 0.026 DAS28 per 0.6 unit increase 5.55 (3.34;9.23) p<0.001 RX progression 2.7 (1.21;6.03) 0.015 Model Constant 0.04 B-RRA refractory rheumatoid arthritis according to Buch, KF-RRA refractory rheumatoid arthritis according to Kearsley-Fleet, DH-RRA refractory rheumatoid arthritis according to De Hair, RX progression (mTSS ≥0,5 over the last 24 months) Disclosure of Interests: LARA FRISO: None declared, Francesca Ometto: None declared, DAVIDE ASTORRI: None declared, Costantino Botsios: None declared, Andrea Doria Consultant of: GSK, Pfizer, Abbvie, Novartis, Ely Lilly, Speakers bureau: UCB pharma, GSK, Pfizer, Janssen, Abbvie, Novartis, Ely Lilly, BMS

OBJECTIVES:We aimed to evaluate trends of rheumatoid arthritis (RA) mortality reported as the underlying cause of death (UCD) and as multiple causes of death (MCD) in Italy between 2003 and 2015.METHODS:Analyses were carried out on the Italian National Cause of Death Register, managed by the Italian National Institute of Statistics (ISTAT). Deaths from January 1, 2003 to December 31, 2015 with any mention of RA were included. Diseases are coded according to the International Classification of Diseases, 10th Edition (ICD- 10, 2009 version). Time trends of age-standardised rates were analysed for RA both as UCD and MCD, and the annual percent change (APC) was estimated.RESULTS:Overall, 26,564 deaths with a mention of RA were retrieved out of 7,595,214 deaths (0.35% of all certificates). The mention of RA as MCD increased throughout the study period, meanwhile the selection as the UCD decreased. RA mortality rates based on the UCD declined (males APC -3.1%, CI -3.9, -2.3; females APC -3.3%, CI -4.1, -2.4); while rates based on the MCD were stable. Specifically, rates were stable or declined among younger subjects and increased in subjects aged ≥80 years.CONCLUSIONS:RA was found to be increasingly reported in death certificates in the last two decades in Italy, although it is less frequently reported as the UCD. Due to the increased survival of patients, we observed a shift of RA-related mortality towards the elderly, making RA a comorbidity contributing to death in these patients.

Objective The 5-item Compliance Questionnaire for Rheumatology (CQR5) proved reliability and validity in respect of identification of patients likely to be high adherers (HAs) to anti-rheumatic treatment, or low adherers (LAs), i.e. taking<80% of their medications correctly. The objective of the study was to validate an Italian version of CQR5 (I-CQR5) in rheumatoid arthritis (RA) patients and to investigate factors associated with high adherence. Methods RA patients, undergoing treatment with >= 1 self-administered conventional synthetic disease-modifying anti-rheumatic drug ( csDMARD) or biological DMARD (bDMARD), were enrolled. The cross-cultural adaptation and validation of I-CQR5 followed standardised guidelines. I-CQR5 was completed by patients on one occasion. Data were subjected to factor analysis and Partial Credit model Parametrisation (PCM) to assess construct validity of I-CQR5. Analysis of factors associated with high adherence included demographic, social, clinical and treatment information. Factors achieving a p<0.10 in univariate analysis were included in multivariable analysis. Results Among 604 RA patients, 274 patients were included in the validation and 328 in the analysis of factors associated with adherence. Factor analysis and PCM confirmed the construct validity and consistency of I-CQR5. HAs were found to be 109 (35.2%) of the patients. bDMARD treatment and employment were found to be independently associated with high adherence: OR 2.88 (1.36-6.1), p=0.006 and OR 2.36 (1.21-4.62), p=0.012, respectively. Conclusion Only one-third of RA patients were HAs according to I-CQR5. bDMARDs and employment status increased by almost 3-fold the likelihood of being highly adherent to the anti-rheumatic treatment.

Background Mortality rates in patients with rheumatoid arthritis (RA) are 1.5-1.6 fold higher than in the general population1,2. No recent data on mortality in large cohorts of RA patients in Italy are available. Objectives The aim of this study was to assess standardized mortality ratios (SMRs) and multiple causes of death in RA subjects living in the Veneto Region between 2010 and 2015. Methods We identified in the electronic archive of the Veneto Region a cohort of patients aged 20-89 years who were exempt from copayment for RA in January 2010, and linked them with the archive of causes of deaths of the period 2010-2015. The record-linkage was performed on previously anonymized records. In the Veneto Region a copy of all death certificates is transmitted to the Regional Epidemiology Service for coding of causes of death according to the International Classification of Diseases, 10th Edition. Each subject was followed from 1st January 2010 either until death, or 90 years of age, or 31st December 2015, whichever came first. In the Veneto Region, the archive of causes of death include all diseases mentioned in the death certificate, and the selection of the underlying cause of death (UCOD) is performed by means of the Automated Classification of Medical Entities, a computer program developed by the US National Center for Health Statistics. SMRs with 95% confidence intervals, were computed as the ratios between deaths observed in the cohort, and those expected according to age- and gender-specific regional mortality rates. Results Overall 16,098 residents diagnosed with RA and aged 20-89 years were enrolled in the cohort. Follow-up was complete for above 99% of study subjects. The overall follow-up amounted to 88,599 person-years, with 2,142 registered decedents. The most common causes of death were circulatory diseases (36.6%), neoplasms (24.2%),and respiratory diseases (8.3%). SMR in RA subjects was 1.42 (1.36-1.48).Mortality was significantly increased from circulatory, respiratory, digestive,infectious, hematological diseases and falls (figure 1). Mortality fromneoplasms was similar to that expected based on rates from the generalpopulation. RA was selected as the underlying cause of death in 6.1% of all deaths in the cohort and was mentioned in 25.4% of death certificates. Conclusions Overall, a 42% excess risk of death could be observed among patients with RA in the Veneto Region.These data confirm results from previous studies in large cohorts of RAsubjects [1,2]. References: [1] Sokka T, Abelson B, Pincus T. Mortality in rheumatoid arthritis: 2008 update. Clin Exp Rheumatol 2008;26(Suppl 51):S35-61. [2] van den Hoek J, Boshuize HC, Roorda LD, et al. Mortality in patients with rheumatoid arthritis: a 15-year prospective cohort study. Rheumatol Int 2017;37:487-493. Disclosure of Interest: None declared

Background The 5-item Compliance Questionnaire for Rheumatology (CQR5) allows the identification of patients likely to be high adherers (HA) to anti-rheumatic treatment (i.e. taking ≥80% of their medications correctly), or “low” adherers (LA).1 We have previously validated an Italian version of the questionnaire (I-CQR5) following standardised guidelines. Objectives The objective was to investigate what factors are associated with high treatment adherence according to I-CQR5 in RA. Methods RA patients (disease duration >1 year, undergoing treatment with ≥1 self-administered biological or conventional synthetic disease-modifying anti-rheumatic drugs (bDMARDs, csDMARDs), capable of completing the questionnaire unaided) completed I-CQR5 on one occasion. I-CQR5 were anonymous and clinical data were retrieved anonymously from the local database. To investigate what factors were associated with high adherence, we included demographic and social characteristics, clinical and treatment information. Factors achieving a p<0.10 in univariate analysis were included in the multivariate regression analysis. Results Among 604 RA patients, 328 patients were enrolled, 18 questionnaires were incomplete. Median age of the patients was 57 years,48–67 females were 232 (82%), disease duration was 12 years;7–19 193 (64.3%) patients were treated with bDMARDs and 107 (54.6%) with csDMARD only; 270 (90.3%) were in low disease activity or remission (figure 1). HA were found to be 35.2% (109/310) of the patients.40.2% (79/193) were on bDMARDs and 22.4% (24/107) on csDMARDs. Older age, lower education level, higher prednisone daily dose, use of a csDMARD (particulary hydroxychloroquine and sulfasalazine) and higher patient-VAS were significantly more frequent in LA compared with HA (figure 1). In the multivariate analysis, bDMARD treatment and employment resulted independently associated with high adherence: OR 2.88 (1.36–6.1), p=0.006 and OR 2.36 (1.21–4.62), p=0.012 respectively (table 1). Conclusions Only one third of Italian RA patients were highly adherent to treatment according to I-CQR5. Treatment with bDMARDs and employment status were the major determinants of treatment adherence, increasing by almost 3-fold the likelihood of being highly adherent. Reference [1] Hughes L, et al. A 5 item version of the Compliance Questionnaire for Rheumatology (CQR5) successfully identifies low adherence to DMARDs. BMC Musculoskelet Disord2013;14:286. Disclosure of Interest None declared

Background The 5-item Compliance Questionnaire for Rheumatology (CQR5) allows the identification of patients likely to be high adherers (HA) to anti-rheumatic treatment (i.e. taking ≥80% of their medications correctly), or “low” adherers (LA).1 Objectives The objective was to validate the construct of an Italian version of CQR5 in rheumatoid arthritis (RA). Methods Cross-cultural adaptation comprised: forward translation, synthesis of the translations, back-translation, expert committee assessment and field-testing. Validation was conducted administering the adapted version (I-CQR5) to RA patients (disease duration >1 year, treated with ≥1 self-administered disease-modifying anti-rheumatic drug, capable of completing the questionnaire unaided) on one occasion. Questionnaires were anonymous but contained self-reported data. Construct validity and reliability were assessed with Rasch analysis (Partial Credit model Parametrisation, PCM). Martin-Loef Likelihood ratio test assessed invariance for gender, age, education, social status and disease duration. Results The adaptation process was closed by the expert committee assessment. I-CQR5 is reported in Figure 1. Among 604 patients, 274 were included in the validation process, 6 questionnaires were incomplete. Median age was 57 years (48–67), females were 201 (77%), disease duration was 13.5 years (8.8–19.3), most patients lived with partner/family (159,75%) and had a middle/secondary school education (184,69%). HA were 93 (67%) and LA 179 (35%). Factor analysis revealed ordered thresholds in most items, 2 factors were sufficient to explain variability (Chi-square=0.46,p=0.5) (Tab.1). Item-fit statistics showed overall agreement of items with parametrisation (Infit statistics=0.6–1.4; excluding item no.5). Chi-square showed agreement with PCM parametrisation by item (excluding the item no.1). Martin-Loef likelihood ratio test confirmed unidimensionality (Chi-square 65.8, df=53,p=0.11) and Separation Reliability Index confirmed internal consistency (Patient Separation Index 0.91) (Tab.1). I-CQR5 was invariant to age (Chi-square=40.6, df=28,p=0.059), education level (Chi-square=49.9, df=42,p=0.187), social status (Chi-square=10.5, df=15,p=0.79), disease duration (Chi-square=13.6, df=36,p=0.220); Martin-Loef test was significant for gender (Chi-square=25.4, df=14,p=0.031). Conclusions I-CQR5 was well understood by patients and construct validity, unidimensionality and internal consistency were confirmed by factor analysis and PCM. Reference [1] Hughes L, et al. A 5 item version of the Compliance Questionnaire for Rheumatology (CQR5) successfully identifies low adherence to DMARDs. BMC Musculoskelet Disord2013;14:286. Disclosure of Interest None declared

Objectives To determine factors associated with remission (DAS28<2.6). We specifically considered the association of biologic treatment duration, number of biologic switches and survival of biologic treatment with remission. Methods We conducted a retrospective analysis on a monocentric cohort of rheumatoid arthritis patients. We included patients who were on biologic drugs at the time of the analysis (31st December 2016). We considered patients starting the first biologic treatment since January 2000 and with a follow-up ≥12 months. We considered the following variables: demographics, positive rheumatoid factor (RF)/ anti-citrullinated peptides (ACPA), disease duration at the start of the first biologic treatment, number of biologic switches, clinical assessment at the last follow-up, concomitant DMARDs, prednisone dose, current biologic treatment. We also considered the mean survival of biologic treatments, defined as the duration of the biologic treatment of each patient divided by the number of biologics undergone by the patient. Mann-Whitney test and chi-square test were used to assess the association of continuous and categorical variables with the outcome. Continuous measures are reported as medians and interquartile range. Multivariate regression analysis included all variables reaching a p value <0.2 in univariate analysis Results We collected data of 330 patients. All patients had complete data. One hundred thirty-five patients (40.9%) were in DAS28 remission. Characteristics of the patients are reported in Table I. We considered 609 biologic treatments (abatacept n=61; anti-TNF alpha n=445; anakinra n=43; tocilizumab n=56; rituximab n=5). Total biologic treatment duration in all patients was 9.62 years (5.68–12.53), in patients in remission 8.95 (4.70–12.53) and in patients not in remission 10.12 (6.28–12.62) (p=0.248). Median number of previous biologic switches was 1.00 (0–1.00) in patients in remission and not in remission (p=0.436). Survival of biologic treatments was 5.33 years (2.89–7.72) in all patients, 4.57 (2.54–7.28) in patients in remission and 5.81 (3.49–7.93) in patients not in remission (p=0.013). All clinical assessments at the last follow-up were significantly associated with DAS28 remission. Mean prednisone dose was significantly lower in patients in DAS28 remission but was considered as a consequence of remission rather than a predictor (Table I). The type of current biologic treatment was not significantly associated with DAS28 remission. Variables included in the multivariate regression analysis were: BMI, age, disease duration, positive RF/ACPA mean survival of each biologic treatment. All variables included in the model were independently associated with DAS28 remission. A higher BMI, older age, longer disease duration and positive RF/ACPA were negatively associated with remission. A longer mean survival of biologic treatments was positively associated with DAS28 remission (OR 1.11 per year increase, 95% C.I. 1.03–1.18, p=0.003). Conclusions The duration of biologic treatment and the number of previous biologic switches were not associated with of DAS28 remission. Indeed, a longer survival of biologic treatments was associated with remission. The mean survival of biologic treatments reflects both a smaller number of biologic failures and a prolonged response to each biologic treatment. Disclosure of Interest None declared

Objectives To investigate etanercept (ETN) serum levels and anti-etanercept antibodies (AEA) and evaluate whether ETN levels are associated with SDAI remission in rheumatoid arthritis patients in DAS28 remission treated with full-dose (25 mg twice weekly) and low-dose (25 mg once weekly) ETN. Methods We selected 70 patients in stable (≥12 months) full-dose (35) or low-dose (35) ETN, in remission (DAS28<2.6 for ≥12 months) and with a treatment duration ≥5 years. AEA and basal ETN levels were tested by ELISA. A cut-off >142AU/ml was adopted for defining positive AEA. A concentration of ETN ≥3.1μg/mL was considered as a high ETN level [1]. Multivariate regression analysis was used to investigate whether high ETN levels were associated with SDAI remission. Results One blood sample in the low-dose group resulted inadequate. AEA were present in 1 patient in low-dose (1/34, 2.9% low-dose ETN vs 0/35, 0% full-dose ETN, p=0.507). Full-dose and a lower weight were associated with high ETN levels: 15/35, 42.9% vs 3/34, 8.8% p=0.001 and 55.0 kg (57.0;80.0) and 65.0 (51.0;69.8), p=0.014 for high and low ETN levels respectively. Lower ETN levels, low-dose and a longer low-dose duration were associated with SDAI remission but were not confirmed with multivariate regression (Table 1). Conclusions AEA are almost absent in patients in stable remission, patient with full-dose ETN and lower weight have higher ETN levels. ETN levels are not predictive of SDAI remission. References Daïen CI, Daïen V, Parussini E, et al. Etanercept concentration in patients with rheumatoid arthritis and its potential influence on treatment decisions: a pilot study. J Rheumatol 2012.39(8):1533–8. Disclosure of Interest None declared

Objectives To investigate survival of low-dose etanercept (ETN) (25 mg weekly) and possible predictors. Methods We collected retrospectively data of RA patients starting ETN between April 2001 and September 2014 who achieved and maintained low-dose ETN for at least 12 months. Patients were evaluated every 3 months. Patients who achieved and maintained remission (DAS28 <2.6) for at least 6 months started low dose ETN. If remission was lost in two consecutive visits the patient returned to full dose. We recorded the time to achieve remission (TTR), the time in which ETN dose was reduced since remission (TRL) and the time in which patients returned to full dose or were switched to other biologics. Variables collected are reported in Table I. Survival was analysed with Cox-regression. Results Among 532 patients, 276 were excluded because of missing data or because they were lost from follow-up leaving 256 patients eligible. After 11.28±4.05 years, 151/256 (59.0%) maintained half dose (Table 1). TTR, and TRL were associated to low-dose failure (Table 1) and correlated with each other (TTR and TRL R 0.45, p<0.001). We run two separate analyses for TTR and TRL, covariates were positive rheumatoid factor or anti-citrullinated peptides and mean prednisone dose. TTR was a significant predictor of low-dose failure (OR 1.50 per year increase, 95% C.I. 1.34–1.68, p<0.001) while TRL was not (Table 2). Conclusions A shorter TTR is a predictor of long-term survival of low-dose ETN. Disclosure of Interest None declared

Background Functional outcome of patients with rheumatoid arthritis (RA) depends from control of disease activity and arrest of radiographic damage. Clinical remission can be achieved in the majority of RA patients by combinations of steroids, DMARDs and TNFα blockers. Recent imaging studies identified residual subclinical inflammation in many patients in treatment with DMARDs: active synovitis was demonstrated by magnetic resonance imaging (MRI) and ultrasound (US). Power Doppler (PD) positivity correlated best with radiographic progression (1). On the other hand, TNFα blockers are much more efficacious in blocking radiographic damage by direct inhibition of osteoclasts, but it is not known if PD positivity predicts on-going damage also in RA patients in remission by TNFα blockers. Objectives To determine the usefulness of PD study to predict structural damage in patients considered in clinical remission by TNFα blocker therapy. Methods A prospective observational study was performed on 121 consecutive RA patients in therapy with TNFα blockers and in DAS28 remission since 6 months. At baseline US examination was performed in all patients on the metacarpophalangeal, proximal interphalangeal, wrist and metatarsophalangeal joints. Active synovitis was identified by PD positivity. Semi-quantitative PD scoring 0–3 and number of positive joints were registered. The location of PD was scored as follows: capsular and within synovia without bone contact (location 1) or with bone contact and penetrating bone cortex (location 2). At baseline and after one year X-rays of hands and feet were executed. Radiographic progression was expressed as difference in total Sharp Score modified by van der Heijde (ΔTSS) >0. Risk ratios for presence, grade and location of PD relative to radiological progression were calculated. Results Among the 121 patients, 62 (51,3%) showed no PD signal at the time of remission, whereas 59 (48,7%) had PD signal: in 57,6% of cases grade was 1, in 23,7% grade 2 and in 18,6% grade 3. 38 patients showed PD positivity in one joint, 21 in more joints. Among all PD positive joints 92,4% were at the level of the hands. 74,6% of PD signals were in contact with articular bone surface (location 2). All patients without PD signal did not show radiographic progression. 28,8% of patients showing PD positivity (corresponding to 14% of all patients in clinical remission) underwent radiographic progression. Radiographic progression occurred in 20,6% of patients with PD grade 1, 42,8% with grade 2 and 36,3% with grade 3. Whereas only 6,6% of patients with PD in location 1 progressed, 36,3% of patients with PD in location 2 showed deterioration of radiographic damage. Despite presence of PD, higher PD grades (2 and 3) and PD with contact to bone surface (location 2) determined a significantly increased risk for radiographic progression; PD in location 1 resulted protective. Conclusions PD demonstrated to be very useful in evaluating patients considered in remission. Absence of PD signal guarantees arrest of radiological progression, whereas patients with PD signal are at risk for progression despite TNFα blockers. This risk increases with higher PD grades and with PD in contact or penetrating bone profile. References Brown AK. Arthritis Rheum 2006,54:3761–73 Disclosure of Interest None declared

Background Abatacept (ABT) short-term efficacy and safety profile in the treatment of rheumatoid arthritis (RA) is well established and supported by randomised controlled trials. However, data on ABT long-term effects are still lacking and reported drug survival analyses are often limited to no more than 2-year follow-up. Objectives The aim of the study is to evaluate in a real-life scenario the overall 6-year retention rate of ABT in the treatment of RA and to compare its survival rate as first or second line biotherapy. Methods Data were retrospectively extracted from six local registries including all RA patients treated with ABT between January 2008 and April 2015. The analysis was limited to patients treated with ABT as first or second line biologic agent. The 6-year retention rate was calculated by Kaplan-Meier method and the comparative risk for discontinuation between the two lines of treatment was compared by a stratified log-rank test. Results 236 RA patients (78.5% female, mean age [± standard deviation, SD] 57.6 [±12.7] years, mean disease duration 14.3 [±13.4] years, positive rheumatoid factor 82.2%, positive anti-citrullinated peptide antibodies 71.1%, mean DAS28 5.19 [±1.2]) were retrospectively enrolled. 83 patients were treated with ABT as first-line biotherapy, whereas 153 were second-line users. The overall retention rate was 63.7% at 3 years and 61.8% at 6 years. The 6-year retention rate was significantly higher in the first-line compared with the second-line therapy group (68% and 61.8%, respectively; p=0.04). Overall, ABT was discontinued in 69 (29.2%) patients. Twenty-one (8.9%) patients (4 first-line and 17 second-line users, p=0.15) withdrew ABT because of adverse events, whereas 41 (17.4%) patients (10 first-line and 31 second-line users, p=0.1508) stopped the therapy because of primary or secondary inefficacy. Seven (2.7%) patients discontinued ABT because of other reasons (i.e. lost at follow-up, pregnancy and remission) and were excluded from the retention analysis. Conclusions In our multicentric experience the 6-year retention rate of both first and second line ABT was about 60%. Used as first line ABT showed a significantly higher survival on treatment compared with the second line. Moreover, ABT demonstrated an overall favourable safety profile, with a very low occurrence of adverse events leading to discontinuation. Disclosure of Interest None declared

Inflammatory rheumatic diseases are the leading causes of disability and constitute a frequent medical disorder, leading to inability to work, high comorbidity, and increased mortality. The standard for diagnosing and differentiating arthritis is based on clinical examination, laboratory exams, and imaging findings, such as synovitis, bone edema, or joint erosions. Contrast-enhanced ultrasound (CEUS) examination of the small joints is emerging as a sensitive tool for assessing vascularization and disease activity. Quantitative assessment is mostly performed at the region of interest level, where the mean intensity curve is fitted with an exponential function. We showed that using a more physiologically motivated perfusion curve, and by estimating the kinetic parameters separately pixel by pixel, the quantitative information gathered is able to more effectively characterize the different perfusion patterns. In particular, we demonstrated that a random forest classifier based on pixelwise quantification of the kinetic contrast agent perfusion features can discriminate rheumatoid arthritis from different arthritis forms (psoriatic arthritis, spondyloarthritis, and arthritis in connective tissue disease) with an average accuracy of 97%. On the contrary, clinical evaluation (DAS28), semiquantitative CEUS assessment, serological markers, or region-based parameters do not allow such a high diagnostic accuracy.

Use of real world data is spreading at increasing pace in response to the growing need of ensuring effectiveness and appropriateness in the delivery of health and medical care. Data from clinical trials are essential to define efficacy and safety of medical therapies in experimental conditions, while real world data assume greater relevance when one considers the differences between study populations of clinical trials and people who take the same drug in real life conditions, and the changes that happen over time with the acquisition of new evidence on drugs and treatments. Rheumatoid arthritis is a chronic disease with significant social (direct and indirect) costs that today can be medically managed ensuring both effectiveness and appropriateness. Real world data can contribute significantly to the efficient management of this disease as demonstrated by outstanding examples from Italian registries and databases.

OBJECTIVES This prospective long-term follow-up study evaluated the effects of half-dose etanercept (25 mg weekly) on clinical remission and radiographic progression in a large cohort of patients with rheumatoid arthritis (RA) in clinical remission after etanercept 25 mg bi-weekly. METHODS 524 biologic-naïve RA patients were treated with etanercept 25 mg bi-weekly after failure of conventional drugs. Patients achieving remission (DAS28 <2.6) for ≥12 months were randomised to receive etanercept 25 mg weekly or 25 mg bi-weekly. Patients were assessed at baseline and every 12 weeks. Remission rates, radiographic progression, incidence of infections and costs of the regimens were compared. RESULTS After a mean follow-up of 18±11 months, 347 patients (66.2%) achieved DAS28 remission; 323 were randomised to one of two dose regimens: etanercept 25 weekly (group A, 159 patients) and etanercept 25 mg bi-weekly (group B, 164 patients). At the end of follow-up, 81.8% patients of group A maintained remission for a mean of 3.6±1.5 years. Radiographic progression occurred in a small number of patients of group A and the rate of radiographic progression (TSS >0) was not significantly different in the two groups (18.85% vs. 19.0% after the first year and 16.9% vs. 21.6% after the second year, respectively). The incidence ratio of severe infections was 2.3/1.000 patient-years in group A. Etanercept half-dose regimen resulted in a saving of €3.190.545 with a cost saving up to €827.318 per year. CONCLUSIONS Clinical remission and arrest of radiographic progression persisted in a substantial percentage of patients with RA even after reduction of standard-dose etanercept.

Background Although both diseases are characterized by specific features the differentiation between rheumatoid arthritis (RA) and simil-rheumatoid psoriatic arthritis (srPsA) is extremely difficult except by hard-to-gain biopsy specimens (1). On contrast enhanced magnetic resonance imaging (CE-MRI) some extra-articular manifestations may direct diagnosis, but synovitis could not be discriminated between RA and srPsA in these studies (2-4). Contrast-enhanced ultrasound (CEUS) using “real” intravascular agents is believed to allow more accurate study of synovial vascularization. Objectives To determine the feasibility to discriminate between RA and srPsA using CEUS derived flow parameters by ad hoc developed software program for analysis of synovial vascularization. Methods 64 outclinic patients with polyarthritis of hands, 32 with RA and 32 with srPsA, were recruited. The most active joint was chosen for CEUS examination using a US device (Mylab70, Esaote) equipped with Contrast tuned Imaging (CnTI, Esaote), and as contrast agent sulfur hexafluoride microbubbles (SonoVue; Bracco International). Both the anatomical B-mode image and the CnTI cineloop video were digitally stored for subsequent software analysis. Image analysis was performed firstly applying a semi-automatic detection of synovial boundaries (5). Then, the contrast time-activity curve of all pixels belonging to the synovial and perisynovial region was analyzed fitting a gamma curve f(t) = A(t − t0)a $times$ e(t −t0)/b on the data. The statistics summarizing the distribution of the estimated kinetics parameters in the synovial and in the perisynovial tissue were computed and their difference between the two groups (RA and srPsA) analyzed, so to study the existence of different vascularization patterns. Finally, a supervised classifier (random forest) was trained to classify each patient through its CEUS-derived parameters, validating the classifier diagnostic power using a leave-one-out strategy. To further increase diagnostic power data about DAS28, CRP, ESR and autoantibodies were added. Results Vascularization pattern constituted of 40 flow parameters discriminated effectively RA from srPsA. Accuracy was 0.93 during training and 0.83 during test phase. Adding rheumatoid factor (RF) and anti-CCP increased diagnostic accuracy to 0.99 in training and 0.93 in test phase decreasing needed flow parameters to 28, whereas DAS28, CRP and ESR did not. Conclusions The Dynamic Automated Synovial Imaging (DASI) is actually the only imaging method that accurately discriminates RA from srPsA, especially in the presence of RF and anti-CCP data. References Kruithof E. Arthritis Res Ther 2005;7:569-80. Jevtic V. Handchir Mikrochir Plast Chir 2012;44:163-170. Cimmino MA. J Rheum 2012;39(89):43-8. Schoellnast H. AJR 2006;187:351-7. Veronese E. Med Eng Phys 2013; 35, 188–194. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.3779