Table 1 (abstract O01). The top 8 variables in the final prediction model (full model uses 18 variables) Variable Adjusted Hazard Ratio for time to remission on medications (95% CI) P-value Physician global (0-10) 0.933 (0.888-0.981) 0.005 Weeks from onset to diagnosis 0.997 (0.995-0.999) 0.003 RF+ Polyarthritis 0.383 (0.150-0.978) 0.041 Systemic arthritis 0.550 (0.308-0.983) 0.039 Subtalar joint involved 0.670 (0.482-0.931) 0.015 Symmetric joints involved 0.777 (0.616-0.979) 0.029 Pain intensity (0-10) 0.927 (0.874-0.983) 0.010 No. enthesitis sites 0.960 (0.921-1.001) 0.050 [22]Open in a new tab New tools and some omics O02 TREG SPECIFIC CONSTITUTIVE NRF2 ACTIVATION INDUCES AUTOIMMUNITY Klaus Tenbrock^1, Patricia Klemm^1, Kim Ohl^1 ^1Pediatrics, RWTH AACHEN UNIVERSITY, Aachen, Germany Correspondence: Klaus Tenbrock Introduction: Immune cells are constantly confronted with intracellular and extracellular radical oxygen species (ROS) under steady-state and even more under inflammatory and pathogenic conditions. To investigate the effects of oxidative stress and ROS molecules in regulatory T cells (T[regs]), we deciphered the role of Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) in this context. T[regs] were already found to be more resistant to ROS than effector T cells and activated T[reg] cells show higher expression of genes, which belong to the Nrf2-mediated oxidative stress response compared to activated effector T cells. Objectives: To analyze the specific role of Nrf2 in T[regs.] Methods: To define the role of Nrf2 in Tregs, we generated mice with a knock out of Keap in all hematopoietic cells (VAV^CRE-Keap^fl/fl) with a Treg specific knockout of Keap 1, which is the constitutive binding partner of Nrf2 (Foxp3^CRE-Keap1^fl/fl) and characterized the mice using flow cytometry, T cell suppression assays and metabolomics. Results: While mice bearing a constitutive activation of Nrf2 activation in all immune cells (Vav^creKeap^fl/fl) accumulate high percentages of Foxp3-positive T[regs] in the spleen, lymph nodes and thymus, their suppressive capacity seems to be defective. Interestingly, a T[reg] specific activation of Nrf2 (Foxp3^creKeap^fl/fl) results in an auto-inflammatory phenotype with immune cell infiltrates in the lungs, enhanced effector T cell activation and high percentages of IFN-γ producing effector T cells. Moreover, the constitutive Nrf2 activation in T[regs] increases their in-vitro proliferation, glucose uptake and mTOR activity, while the differentiation and Foxp3 expression in T[regs] declines Conclusion: We demonstrate for the first time that constitutive Nrf2 activation specific to T[regs] affects T[reg] lineage stability and metabolism and might thereby induce an auto-inflammatory phenotype. Our results therefore may have implications for diseases associated with oxidative stress and dysregulated T[reg] responses. Disclosure of Interest None Declared O03 EXTENDED PHENOTYPIC IMMUNOME CHARACTERISATION (EPIC): A WEB-BASED IMMUNE REFERENCE ATLAS Joo Guan Yeo^1,2, Pan Lu^1, Thaschawee Arkachaisri^1,2, Su Li Poh^1, Fauziah Ally^1, Jingyao Leong^1, Kee Thai Yeo^1,2, Loshinidevi D/O Thana Bathi^1, Angela Yun June Tan^2, Liyun Lai^1, Elene Seck Choon Lee^2, Salvatore Albani^1,2 ^1Translational Immunology Institute, SingHealth Duke-NUS Academic Medical Centre; ^2KK Women’s and Children’s Hospital, Singapore, Singapore Correspondence: Joo Guan Yeo Introduction: An atlas of the developing immune system is key to understanding its normal maturation and identifying disease-associated cell subsets. The availability of high dimensional mass cytometry, in comparison to traditional oligo-dimensional approach such as fluorescence-based flow cytometry, provides an opportunity for the creation of this reference. However to date, the power available from these big data has not been fully harnessed due to the absence of clinically relevant and standardised datasets. This results in issues of fragmentation by focusing on individual cell subsets and lack the ability to transverse the whole developmental gradient from neonatal to adult age. There is a critical unmet need for standardised datasets depicting at single cell level and with high dimensionality the entire human immune system. These limitations hamper translational and clinical research. Objectives: To address this need, we aim to construct a mass cytometry based immune atlas from healthy peripheral blood mononuclear cells (PBMC) samples ranging from cord blood to adult age and make this dataset available to the research community via an interactive web portal to enable its mining and comparison with diseased dataset. Methods: The mass cytometry data from 113 healthy individuals (cord blood, newborn to adult) using 63 immune markers encompassing the major immune lineages were obtained. Quality control was performed before dimensional reduction and clustering to identify the cell subsets using our in-house analysis and visualisation pipeline. Their frequencies across the ages were presented as 3-D frequency histograms to create the immune landscape. This database and the analytic pipeline were incorporated into a web-based portal allowing users to interact and upload their own data for comparison. Results: Here, we described EPIC with examples of its potential by exploring the evolution of some representative immune subsets over the full gradient of ages. Three developmental trajectories made up the healthy immune landscape with the most distinct being an increase or decrease in the cell populations with age. The last trajectory constituted an increase in population size in early childhood followed by a region of levelling. There was a distinct segregation of the naive T cell subset enriched in the cord blood/newborn period with the memory T cell subset enriched in adulthood. The naive IL8+ and TNFα+ CD4+ T cells were prominent peaks in the landscape and were found in higher frequency in the cord blood/newborn period. In contrast, the memory IFNɣ+ and TNFα+ CD4+ T cells were enriched in adulthood. For specific cell subsets, transition developmental milestones were observed in the TNFα+ CD4+ T cells where the size of its memory subset would exceed its naive subset at 8 year old. There was a significant reduction and increase with age in the frequency of the naive and memory TNFα+ CD4+ T cells with a Spearman’s correlation coefficient, rho, of -0.4662 and 0.4164 respectively (p<0.0001). A similar intersection was present for the naive and memory regulatory T cell (CD4+, CD25+, Foxp3+, CD152+) subsets at 14 year old with a rho of -0.537 and 0.5034 respectively (p<0.0001). Conclusion: A holistic description of the developing immunome was obtained with key developmental milestones in the T cell compartment identified. This atlas has the translational potential of helping us define the stage of immune maturity and identify the pathological cell subset in both paediatric and adult immune mediated diseases by the direct comparison with this reference atlas using the web-based portal that will be made freely available. Disclosure of Interest None Declared Oral presentation 1 O04 VALIDATION OF THE RECENTLY DEVELOPED EVIDENCE-BASED EUROFEVER CLASSIFICATION CRITERIA FOR HEREDITARY RECURRENT FEVERS (HRF) AND PFAPA Silvia Federici^1,2, Federica Vanoni^3,4, Francesca Bovis^5, Nicola Ruperto^1, Michael Hofer^6, Marco Gattorno^1 and on behalf of the Expert Committee for the Classification Criteria in periodic fever ^1Division of Rheumatology, Istituto Giannina Gaslini, Genova; ^2Division of Rheumatology, IRCCS Ospedale Pediatrico Bambino Gesù, Rome, Italy; ^3Pediatric Rheumatology Unit CHUV, University of Lausanne, Lausanne; ^4Departement of Pediatric of Southern Switzerland, Bellinzona, Switzerland; ^5Biostatistic Unit, Department of Health Sciences, University of Genoa, genoa, Italy; ^6Pediatric Rheumatology Unit, CHUV, Lausanne, Switzerland Correspondence: Silvia Federici Introduction: Provisional Eurofever evidence-based classification criteria for the 4 inherited recurrent fever (FMF, CAPS, TRAPS, MKD) have been published and other diagnostic criteria are available in the literature for FMF, CAPS and PFAPA. Despite a significant increase of accuracy of the recent Eurofever classification criteria in respect to the previous ones, none of them combine the clinical criteria with the results of molecular analysis. This is a major limitation considering the fact that the genetic analysis might be considered per se pathognomonic, and therefore diagnostic, at least in the presence of a confirmatory genotype. Recently new evidence-based classification criteria for hereditary recurrent fevers (HRF) and PFAPA have been identified during a Consensus Conference held in Genoa in March 2017. For each of the monogenic recurrent fever new classification criteria based on genetic and clinical variables were developed. For PFAPA novel classification criteria based on positive and negative clinical variables were also approved by experts (Gattorno et al. PRES 2017, submitted). Objectives: To evaluate the performance of the final set of classification criteria, in discriminating patients with the different HRF and PFAPA in a separate set of real patients coming from the Eurofever Registry and to compare their accuracy with respect to existing criteria. Methods: We selected those patients with recurrent HPF coming from the Eurofever Registry excluding patients belonging to the original dataset of 360 patients used for the development of the criteria themselves. Patients with inherited periodic fever (TRAPS, FMF, MKD and CAPS), PFAPA and undefined periodic fevers were considered and classified according to the indication of each center, without any process of validation by expert. Sensitivity, specificity, accuracy, negative and positive predictive values and AUC-ROC of the new criteria were calculated. Results: A total of 1018 new patients coming from the Eurofever Registry were included. The performance of the criteria coming from the Consensus conference in comparison with the criteria of the literature is listed in Table 1. Overall, their performance was superior (accuracy ranging from 0.81 to 0.98) to the already published literature’s criteria (accuracy 0.56-0.94) with a very high specificity and a variable sensitivity. Most of the patients not classified with the new criteria were negative for genetic analysis or carriers of low-penetrance mutations with an inconsistent clinical phenotype. Conclusion: The validation of the new Eurofever classification criteria in a large group of unselected patients coming from the registry confirms their high specificity and overall better performance in comparison to other criteria available in the literature. It is recommended to use them as classification rather that diagnostic criteria, for clinical trials and pathogenic studies. Disclosure of Interest None Declared Table 1 (abstract O04). See text for description Sensitivity Specificity Accuracy AUC New CAPS criteria (clinical + genetics) 0.72 1 0.96 0.86 Kuemmerle CAPS criteria 0.88 0.82 0.83 0.85 New FMF criteria (clinical + genetics) 0.89 1 0.96 0.94 Livneh criteria 0,9 0,58 0,7 0,74 Tel Hashomer 0,59 0,95 0,64 0,77 Pediatric FMF criteria 0,61 0,94 0,83 0,77 New MKD criteria (clinical + genetics) 0.74 1 0.98 0.87 Clinical guideline 0.77 0.53 0.56 0.65 New TRAPS criteria (clinical + genetics) 0.74 1 0.97 0.87 New PFAPA clinical criteria 0.66 0.97 0.9 0.82 Modified Marshall criteria 0.46 0.91 0.81 0.69 [23]Open in a new tab O05 PERFORMANCE OF THE AUTOINFLAMMATORY DISEASE ACTIVITY INDEX (AIDAI) IN PATIENTS WITH COLCHICINE-RESISTANT FMF, HIDS/MKD AND TRAPS: RESULTS FROM A PIVOTAL, PHASE 3 TRIAL OF CANAKINUMAB Isabelle Kone-Paut^1, Maryam Piram^1, S Benseler^2, J. B Kuemmerle-Deschner^3, A. F Jansson^4, I Rosner^5, A Tomassini^6, S Murias^7, O Karadag^8, J Levy^9, S Smeets^10, F De Benedetti^11 ^1APHP, CHU de Bicêtre, Univ Paris Sud, Le Kremlin Bicêtre, France; ^2Alberta Children's Hospital, Calgary, Canada; ^3University Hospital Tuebingen, Tuebingen; ^4Ludwig Maximilian University, Munich, Germany; ^5Bnai-Zion Medical Center, Rheumatology, Haifa, Israel; ^6Department of Internal Medicine, Università Cattolica Sacro Cuore, Rome, Italy; ^7Hospital La Paz, Madrid, Spain; ^8Hacettepe University Faculty of Medicine, Ankara, Turkey; ^9BIOP, Reinach, Switzerland; ^10Novartis Pharma B.V., Arnhem, Netherlands; ^11IRCCS Ospedale Bambino Gesú, Rome, Italy Correspondence: Isabelle Kone-Paut Introduction: AIDAI is a novel and unique, validated patient (pt)-reported assessment tool to evaluate disease activity in familial Mediterranean fever (FMF), hyper-IgD syndrome/mevalonate kinase deficiency (HIDS/MKD) and TNF receptor-associated periodic syndrome (TRAPS).^1 Objectives: To perform an external validation of AIDAI, we calculated scores over 40 weeks (wks) of canakinumab (CAN) treatment in pts enrolled into the CLUSTER trial and assessed correlation between AIDAI and disease/response characteristics. Methods: CLUSTER consisted of one cohort per disease (crFMF, HIDS/MKD and TRAPS).^2 AIDAI was calculated as the sum of 12 items (Yes=1; No=0)^1 for 30 consecutive days. AIDAI score was calculated if the first score was recorded before ≥29 days. Missing items beyond last evaluable measurement were imputed by last observation carried forward (LOCF). Inactive disease (ID) was defined as AIDAI score <9. Correlation analysis of AIDAI with Sheehan disability score (SDS), child health questionnaire–psychological/physical (CHQ–PsCS/PCS), physician global assessment (PGA), short form 12–physical/mental component summaries (SF12–PCS/MCS), C-reactive protein (CRP) and serum amyloid A (SAA) were performed. Significance was set at p<0.05. Results: Overall, 167 (crFMF: N=59; HIDS/MKD: N=66; TRAPS: N=42) pts were randomized to CAN 150 mg or placebo every 4 wks. Median AIDAI scores in all 3 cohorts decreased from baseline (BL) to Wk 16 (crFMF: 22.5 to 5.0; HIDS/MKD: 41.5 to 12.0; TRAPS: 89.0 to 13.0) and through Wk 40 (crFMF: 1.0; HIDS/MKD: 5.0; TRAPS: 20.5). In all 3 cohorts, the proportion of pts with ID (AIDAI score <9) was higher at Wk 40 versus BL (crFMF: 69.5% vs 5.1%; HIDS/MKD: 56.1% vs 6.1%; TRAPS: 42.9% vs 2.4%). AIDAI at Wk 40 correlated significantly with: SDS in all 3 cohorts; CHQ-PsCS in crFMF and HIDS/MKD; CHQ-PCS in crFMF; PGA in TRAPS; SF12–PCS in crFMF and TRAPS. SF12-MCS, CRP and SAA did not correlate with AIDAI (Table 1). Conclusion: AIDAI scores decreased markedly over 40 weeks of treatment with canakinumab in crFMF, HIDS/MKD and TRAPS, with a relevant percentage of patients having inactive disease score. AIDAI improvements at Week 40 correlated with patient- and physician-driven evaluations. AIDAI is a validated patient-reported tool to assess disease activity and appears to have good sensitivity to change to be used in comparative trials. Patient’s experience on disease activity does not strictly correlate with CRP and SAA, as these reflect more closely biological inflammation than clinical symptoms. ^1Piram M, et al. Ann Rheum Dis. 2014;73:2168-73. ^2De Benedetti F, et al. N Engl J Med. 2018 (in press). Trial registration identifying number: [24]NCT02059291 Disclosure of Interest I. Kone-Paut Grant / Research Support from: Novartis, SOBI, Roche, Consultant for: Novartis, SOBI, Pfizer, AbbVie, Roche, M. Piram Consultant for: Novartis, Pfizer, Abbvie, Speaker Bureau of: Novartis, S. Benseler Consultant for: Novartis, SOBI, AbbVie, J. B. Kuemmerle-Deschner Grant / Research Support from: Novartis, Consultant for: Novartis, SOBI, A. F. Jansson Grant / Research Support from: Novartis, I. Rosner: None Declared, A. Tomassini: None Declared, S. Murias: None Declared, O. Karadag: None Declared, J. Levy Consultant for: Novartis, S. Smeets Employee of: Novartis, F. De Benedetti Grant/Research Support from: Novartis, Roche, Pfizer, SOBI, AbbVie, Novimmune, BMS, Sanofi Table 1 (abstract O05). Correlation between AIDAI and disease activity variables at Week 40 Correlation coefficient (95% CI) SDS CHQ-PsCS CHQ PCS PGA SF12-PCS SF12-MCS SAA CRP crFMF 0.54* (0.20; 0.76) -0.83* (-0.95; -0.49) -0.67* (-0.90; -0.16) 0.24 (-0.08; 0.51) -0.47* (-0.77; 0.0) -0.43 (-0.75; 0.05) -0.07 (-0.38; 0.25) -0.07 (-0.38; 0.24) HIDS/MKD 0.39* (0.05; 0.65) -0.47* (-0.76; -0.03) -0.32 (-0.67; 0.14) 0.24 (-0.03; 0.47) -0.09 (-0.71; 0.61) -0.19 (-0.76; 0.54) 0.06 (-0.22; 0.32) 0.08 (-0.19; 0.34) TRAPS 0.41* (0.02; 0.69) -0.07 (-0.62; 0.53) -0.02 (-0.59; 0.56) 0.37* (0.02; 0.63) -0.60* (-0.86; -0.11) -0.13 (-0.62; 0.43) -0.17 (-0.50; 0.20) -0.18 (-0.51; 0.18) [25]Open in a new tab *p<0.05 O06 INTERRATOR RELIABILITY OF A STANDARDIZED SCORING TOOL TO REPORT MRI FINDINGS IN CHRONIC NONBACTERIAL OSTEOMYELITIS Yongdong (Dan) Zhao^1,2, T. S. Sato^3, Meinrad Beer^4, Mingqian Huang^5, Ramesh S. Iyer^6, Michael McGuire^7, Anh-Vu Ngo^6, Jeffrey P. Otjen^6, Jyoti Panwar^8, Jennifer Stimec^9, Mahesh Thapa^6, Paolo Toma^10, Angela Taneja^11, Nancy E. Gove^2, Polly J. Ferguson^12 ^1Seattle Children’s Hospital, Department of Pediatrics, University of Washington; ^2Center for Clinical and Translational Research, Seattle Children’s Research Institute, Seattle; ^3Radiology, University of Iowa Carver College of Medicine, Iowa City, USA; ^4Department of Diagnostic and Interventional Radiology, University Hospital of Ulm, Ulm, Germany; ^5Department of Radiology, University of Stony Brook, Stony Brook; ^6Department of Radiology, Seattle Children’s Hospital, University of Washington, Seattle; ^7Department of Radiology, Hackensack University Medical Center, Hackensack, USA; ^8Joint Department of Medical Imaging, University Heath Network, University of Toronto; ^9Hospital for Sick Children, Department of Medical Imaging, Toronto, Canada; ^10Department of Imaging, Bambino Gesù Children Hospital, IRCCS, Rome, Italy; ^11Pediatric Rheumatology, Children’s Healthcare of Atlanta, Emory University, Atlanta; ^12Department of Pediatrics, University of Iowa Carver College of Medicine, Iowa City, USA Correspondence: Yongdong (Dan) Zhao Introduction: CNO is an inflammatory bone disease that can result in bone destruction, persistent bone pain and pathological fractures. For clinical and research purposes, serial MRI exams are needed to determine the objective response to treatment. A previously developed MRI scoring tool for CNO in a single center showed sensitivity to detect imaging change after aggressive treatment. Objectives: To use a nominal group technique to develop a practical and consensus-based MRI scoring tool for clinical and research use in CNO. Furthermore, inter-rater reliability were evaluated using whole body (WB) MRI from children with CNO. Methods: Eleven pediatric radiologists, each with at least 5 years of experience reading musculoskeletal MRI from seven different pediatric hospitals in North America and Europe, discussed definitions and grading of signal intensity, extent of signal abnormality within bone marrow and surrounding tissue, physis damage and vertebral fracture on MRI through monthly conference calls and an in-person meeting (Seattle, July 2017). Nine sets of whole body MRIs were used as a pilot reading session at the conference that demonstrated greater than 70% of agreement among the radiologist attendees. Fifty sets of pre-existing WB MRI scans between Jan 2013 and August 2016 from children with CNO at the University of Iowa Children’s Hospital were anonymized and used for an inter-rater reliability study. Inter-rater agreement of presence of abnormal signal and severity were assessed using Fleiss kappa analysis. Results: Signal intensity was rated as absent, < fluid signal or similar to fluid signal within bone marrow. Extent of abnormal signal intensity within bone or surrounding tissues were scored according to the relative affected proportion. Long bones were divided into proximal epiphysis, proximal metaphysis, diaphysis, distal metaphysis and distal epiphysis. Complex bony regions such as the pelvis were divided into easily identifiable anatomical subareas. The agreement among 11 radiologists in readings of femur and tibia were shown in Table 1. There was moderate to substantial agreement (0.44-0.63) in all parameters except the sizes of femoral proximal epiphyseal and femoral proximal metaphyseal lesions (0.28-0.34). There were 38 tibias and 30 femurs identified as active inflammation by at least 9 of 11 radiologists. Data from other bones are being analyzed. Conclusion: A comprehensive standardized scoring tool for MRI in children with CNO was developed. There was moderate to substantial agreements among radiologists in majority of parameters of femur and tibia. If proven reproducible, this tool can be validated in a prospective study and will become a key element of disease activity assessment in CNO. This study was funded by CARRA-AF small grant. Disclosure of Interest Y. D. Zhao Grant / Research Support from: Bristol-Myers Squibbs, CARRA, Clinical Research Scholar Program from Seattle Children's Research Institute, T. Sato: None Declared, M. Beer: None Declared, M. Huang: None Declared, R. Iyer: None Declared, M. McGuire: None Declared, A.-V. Ngo: None Declared, J. Otjen: None Declared, J. Panwar: None Declared, J. Stimec: None Declared, M. Thapa: None Declared, P. Toma: None Declared, A. Taneja: None Declared, N. Gove: None Declared, P. Ferguson Grant/Research Support from: R01AR059703 from NIH/NIAMS Table 1 (abstract O06). See text for description MRI characteristics Kappa values (95% CI) of tibial lesions (n=100) Kappa values (95% CI) of femoral lesions (n=100) Max signal intensity 0.46 (0.44,0.48) 0.48 (0.46,0.5) Size of proximal epiphyseal lesion 0.46 (0.44,0.49) 0.34 (0.32,0.36) Size of proximal metaphyseal lesion 0.54 (0.52,0.55) 0.28 (0.25,0.3) Size of diaphyseal lesion 0.46 (0.44,0.48) 0.44 (0.42,0.46) Size of distal metaphyseal lesion 0.58 (0.57,0.6) 0.47 (0.46,0.49) Size of distal epiphyseal lesion 0.48 (0.45,0.5) 0.49 (0.46,0.51) Soft tissue inflammation 0.6 (0.57,0.63) 0.63 (0.6,0.65) [26]Open in a new tab O07 IMPACT OF PEDIATRIC RHEUMATIC DISEASES ON CAREGIVERS MULTIASSESSMENT QUESTIONNAIRE: CAREGIVERS QUESTIONNAIRE Marcia D. Torres-Made^1, Nadina E. Rubio Pérez^1, Ingris Peláez Ballestas^2, Fernando García Rodriguez^1, Ana V. Villarreal Treviño^1, Brenda D. J. Fortuna Reyna^1, Manuel E. De la O Cavazos^1 ^1Pediatría, Hospital Universitario "Dr. José Eleuterio Gónzalez", Monterrey, Nuevo León; ^2Reumatología, Hospital General Dr. Eduardo Liceaga, Mexico, DF. , Mexico Correspondence: Marcia D. Torres-Made Introduction: Most common pediatric rheumatic disease (PRD) is juvenile idiopathic arthritis (JIA), this influence affects quality of life (QoL) of children and families. Caregiver is a cornerstone to achieve disease control however, there are no information on how this conditions affects them; neither a tool to assess this impact was reported. Objectives: To develop and validate psycho-social impact and financial burden questionnaire in caregivers of JIA patients (CAREGIVERS Questionnaire). Methods: We development and validate the CAREGIVERS Questionnaire in two phases. Phase 1.Design the questionnaire based in non-systematic review and qualitative study: First version (1stQ), exploratory and open questions, was constructed based on literature review by an expert group (pediatric rheumatologists and methodologist). A qualitative study of pragmatic type was conducted on 15 caregivers to identify areas affected after JIA diagnosis. Interviews were reviewed using thematic analysis with Atlas-ti software. Questionnaire was adapted based on qualitative study and second version (2ndQ) was tested on 10 participants to evaluate comprehensibility. Changes in structure were made after pediatric rheumatologists, anthropologist/methodologist, psychologist, and pedagogue analyzed the answers (3rdQ). Phase 2. Validation of questionnaire. 3rdQ was evaluated for face, content, reliability, internal consistency, convergent, and divergent validity. Construct validity was tested for anxiety/depression (Beck), coping strategies (Reyes-Lagunes), QoL (EuroQol), and social and economic factors (Socioeconomic burden). Correlation matrix to assess the construct validity using Cohen’s kappa, and internal consistency tests using Cronbach’s alpha were performed. Results: Qualitative study (1stQ) identified topics like socioeconomic burden, family interactions, alternative medicine, religion, information access, and prognosis. 2ndQ was conformed by 75 items. After pilot study, six items were confuse and modified, 11 were eliminated for being redundant. Items were organized in five domains to develop 3rdQ: disease impact (social, economic, family and relationship), knowledge, future, alternative medicine, and religion (13, 27, 2, 2, and 2 items, respectively). Validity analysis included 32 participants (93% women). Mean age 39 (SD 8.5) years, mean EuroQol VAS 82, minimal depression in 90%. Socioeconomic tool demonstrated family income above 14 minimum wages (65 USD) in 68% and 59% have remunerated job. Reyes-Lagunes describes “life coping” with emotional-negative strategies in 31 caregivers and straight strategies in “children´s health coping” in 100%. Finally, in 3rdQ, 22% felt sadness at diagnosis, 84% change initial feeling and currently 43% are relief. Sixty two percent mention disease has influenced in their financial situation and 72% feel anxiety about their children future. Statistical analysis show adequate inter-item reliability in all domains except for religion, consequently items were adapted to create new dimensions (Table 1). Correlation matrix shown good results in all domains, except for social and economic due to redundant items. After changes, CAREGIVERS Questionnaire was created. Conclusion: We design CAREGIVERS Questionnaire, a specific multidimensional tool to assess burden in caregivers of patients with PRD, and validated in JIA. Further efforts will be performed to validated in other PRD. Disclosure of Interest None Declared Table 1 (abstract O07). See text for description Domains Cronbach’s items Disease impact on:  - Social 0.42 13  - Economic 0.43 13  - Familiar 0.19 4  - Relationship 0.26 2 II. Disease knowledge 0.28 8 III. Discrimination 0.68 4 IV. Use of social networks 0.55 10 V. Future 0.31 2 VI. Religion 0.004 2 [27]Open in a new tab O08 TREAT-TO-TARGET USING FIRST-LINE TREATMENT WITH RECOMBINANT IL-1 RECEPTOR ANTAGONIST IS HIGHLY EFFECTIVE FOR SYSTEMIC JIA: A 5 YEAR FOLLOW-UP STUDY Nienke M. ter Haar^1,2, Evert H. P. van Dijkhuizen^2, Joost F. Swart^2, Wilco de Jager^1, Dirk Holzinger^3,4, Arjen P. Leek^2, Jorg van Loosdregt^1, Annet Van Royen-Kerkhof^2, Nico M. Wulffraat^2, Sytze de Roock^1,2, Sebastiaan J. Vastert^1,2 ^1Laboratory for Translational Immunology; ^2Paediatric Rheumatology and Immunology, University Medical Center Utrecht, Utrecht, Netherlands; ^3Paediatric Rheumatology and Immunology, University of Muenster, Muenster; ^4Paediatric Hematology-Oncology, University of Duisburg-Essen, Essen, Germany Correspondence: Sebastiaan J. Vastert Introduction: Systemic onset JIA (sJIA) is a multifactorial autoinflammatory disease. Historically, the prognosis of sJIA was very poor. The use of interleukin (IL)-1 and IL-6 blockade improved outcome, but still most studies report inactive disease in only 30-50% of patients within one year. It is hypothesized that innate immune activation is most prominent in the early phase of sJIA. To make use of this ‘window of opportunity’, we started a prospective cohort with recombinant IL-1 receptor antagonist (rIL-1RA) as first-line monotherapy in new-onset sJIA patients (Vastert et al. AR 2014). Objectives: To analyse the long-term efficacy of this therapeutic strategy and to separately analyse a subset of sJIA patients without arthritis at disease onset. Methods: In this single centre, prospective cohort study, new-onset sJIA patients with an unsatisfactory response to NSAIDs received rIL-1RA 2mg/kg. Inactive disease (ID) was defined according to the Wallace criteria. In patients reaching ID, rIL-1RA was tapered after 3 months and subsequently stopped. In patients with an incomplete response to rIL-1RA, our protocol directed to increase rIL-1RA dose, add prednisolone or switch to alternative therapy. Clinical data including damage (JADI) and patient reported outcomes (CHAQ and/or JAMAR) were completed in a standardized way. Cytokine profiling was performed using Luminex multiplex immunoassay. Results: Forty-two patients were analysed, with a median follow-up of 5.8 years. At 3 months, 30/42 patients (71%) had ID with rIL-1RA monotherapy and another 6 had ID with rIL-1RA and prednisone. At one year, 32 patients (76%) had ID, of which 22 (52%) were off medication. Twelve patients had no arthritis at onset; 4 of them developed arthritis during flares. The clinical phenotype, inflammatory parameters and cytokine profile of patients without arthritis at onset was similar to patients with arthritis. Hence, after extensive ancillary investigations to exclude other diseases, these non-arthritic patients also received rIL-1RA as first-line therapy. After 3 months, 11/12 non-arthritic patients had ID (10 on rIL-1RA monotherapy), at 1 year all 12 had ID (8/12 off medication). Long-term outcome reflected the high efficacy of first-line rIL-1RA: 95% of all patients had ID and 72% had ID off medication at 3 and 5 years follow-up, only 5% reported articular damage and 5% extra-articular damage, no patient developed growth failure and only one developed obesity during follow-up. Moreover, around 60% of patients reported complete absence of pain and disease, no limitations in daily life and the highest possible score on well-being. One patient died due to MAS. In total, 24 patients (57%) never used other medication besides rIL-1RA and NSAIDs and only 14 patients (33%) used glucocorticoids. Patients who achieved ID at one year were younger and had a significantly shorter disease duration, less active joints and higher neutrophil and leukocyte count before start of rIL-1RA. Multivariate analysis confirmed high neutrophil count as a significant predictor for ID at one year. Furthermore, ID at one year was highly associated with a good response at 1 month. Conclusion: First-line, short course monotherapy with rIL-1RA is a highly efficacious strategy to induce and sustain inactive disease and to prevent disease- and glucocorticoid-related damage. Our data on sJIA patients without arthritis plea for leaving out arthritis as a prerequisite criterion in future disease classification criteria. Disclosure of Interest None Declared O09 CLINICAL OUTCOMES FOLLOWING MYCOPHENOLATE MOFETIL VS. CYCLOPHOSPHAMIDE INDUCTION TREATMENT FOR PROLIFERATIVE JUVENILE-ONSET LUPUS NEPHRITIS Eve M. Smith^1, Michael W. Beresford^1, Christian M. Hedrich^2 ^1Institute of Translational Medicine; ^2Institue of Translational Medicine, University of Liverpool, Liverpool, UK Correspondence: Eve M. Smith Introduction: Treatment of proliferative class III/IV lupus nephritis (LN) in children is largely based upon adult studies. The largest study demonstrated a similar renal response rate with mycophenolate mofetil (MMF) or cyclophosphamide (CYC) induction treatment in adult patients with LN [1]. A further smaller study showed MMF to be more effective than CYC [2]. In both, MMF had a better safety profile [1,2]. Objectives: Within a predominantly Caucasian JSLE cohort, to compare efficacy of MMF vs. CYC in patients with LN, monitoring 1) response to treatment, 2) damage accrual, 3) time to achievement of inactive LN, and 4) time to subsequent LN flare, following MMF or CYC LN induction treatments. Methods: Participants of the UK JSLE Cohort Study (2006-2018), <18 years at the time of diagnosis, with ³4 ACR criteria for SLE, were eligible for inclusion if they had ISN/RPS class III or IV LN. Median values and interquartile ranges quoted. Mann-Whitney U tests for continuous data and Fishers exact test for categorical data. Results: 69/411 UK JSLE Cohort Study patients met the inclusion criteria; 18/61 were excluded due to insufficient follow-up. 34/51 (67%) received MMF (13/34 (38%) class III, 21/34 (62%) class IV LN), and 17/51 (33%) received CYC (8/17 (47%) class III, 9/17 (53%) class IV LN). No statistically significant differences were identified between treatment groups at 4-8, 10-14 months post renal biopsy, and last follow-up, in terms of the renal-BILAG score, urine albumin/creatinine ratio, serum creatinine, ESR, anti-dsDNA antibody, C3 levels and patient/physician global scores (all p>0.05, see Table 1). Standardised Damage Index (SLICC-SDI) scores did not differ between treatment groups at a median of 13 months, or last follow-up (all p>0.05, see Table 1). Inactive LN, following the definition of renal BILAG, was attained at 262 days [141–390] after MMF treatment, and 151 days [117-305] following CYC (p=0.17). Time to subsequent renal flare did not differ; 451 days [157–1266] for MMF, and 343 days [198–635] for CYC (p=0.47)). Conclusion: This is the largest study to date investigating induction treatments for proliferative LN in children. In Caucasian JSLE populations, MMF and CYC may be comparably efficacious in regards to treatment response, damage accrual, and time to next flare. Remission may be reached quicker in patients treated with CYC. Future prospective comparison is warranted to inform LN treatment, given CYC’s poor safety profile. References