Objective The objective of this paper is to identify the prevalence, risk factors, and impact on mortality of neuropsychiatric systemic lupus erythematosus (NPSLE). Methods Patients from the Hanyang BAE lupus cohort were registered and followed from 1998 to 2015. NPSLE was defined using American College of Rheumatology (ACR) case definitions and Ainiala criteria. Demographics, autoantibodies, Systemic Lupus Erythematosus Disease Activity Index (SLEDAI), and Systemic Lupus International Collaborating Clinic (SLICC)/ACR Damage Index were collected at baseline and then annually. Mortality data were derived by linking data from the Korean National Statistics Office. Multivariable logistic regression and Cox regression analysis were conducted in the inception cohort to assess the risk factors and mortality impact of NPSLE. Results Of 1121 registered patients, 429 (38.3%) had NPSLE manifestations according to ACR criteria and 216 (19.3%) by Ainiala criteria. In multivariable logistic regression analysis, higher SLEDAI (OR 1.08, CI 1.01–1.16, p = 0.02) and antiphospholipid antibody positivity (OR 1.72, CI 1.03–2.87, p = 0.04) at SLE diagnosis increased NPSLE risk, while elevated anti-dsDNA antibodies (OR 0.43, CI 0.24–0.78, p < 0.01) and greater education duration (OR 0.92, CI 0.85–1.00, p = 0.04) showed reduced risk of NPSLE. Cox proportional hazard models demonstrated that presence of NPSLE had a three-fold increased risk of mortality (HR 3.09, CI 1.03–9.21, p = 0.04), especially in patients with focal CNS NPSLE (HR = 7.83, CI 2.12–28.96, p < 0.01). Conclusion Higher SLEDAI, antiphospholipid antibody positivity, absence of anti-dsDNA antibody at SLE diagnosis, and fewer years of education are risk factors for development of NPSLE. Presence of NPSLE, especially focal CNS NPSLE, increased the risk of mortality in SLE patients.

Objectives This study aims to identify the factors associated with the development and mortality of pulmonary hypertension (PH) in systemic lupus erythematosus (SLE) patients. Methods We conducted a prospective study of SLE patients in a single tertiary center. PH was defined as a systolic pulmonary arterial pressure ≥30 mmHg on transthoracic echocardiography. We assessed potential associated factors contributing to the development and mortality of PH in SLE patients. Results Of 1110 patients with SLE, 48 patients were identified to have PH. Multivariable analysis indicated that pleuritis or pericarditis (odds ratio (OR) = 4.62), anti-RNP antibody (OR = 2.42), interstitial lung disease (ILD) (OR = 8.34) and cerebro-cardiovascular disease (OR = 13.37) were independently associated with the development of PH in SLE. Subgroup analysis among patients with PH demonstrated that there were no statistically significant factors associated with PH mortality in SLE. Conclusions The prevalence of PH was 4.3% in our cohort. There were significant associations with pleuritis or pericarditis, anti-RNP antibody, ILD, and cerebro-cardiovascular disease in SLE, which may contribute to the development of PH. However, there were no statistically significant factors associated with PH mortality in SLE.

Objective To identify the incidence, risk factors and prognosis for neuropsychiatric lupus (NPSLE) in Korea. Methods 1121 patients with SLE from Hanyang BAE lupus cohort were enrolled and followed from 1998 to 2015. NPSLE was defined using the ACR case definitions and Ainiala Criteria. Demographics and clinical information including ACR Classification criteria for SLE, autoantibodies, SLE Disease Activity Index, the SLICC/ACR damage index (DI) were collected at baseline and then annually. Symptoms of NPSLE were collected from patient interview and medical records. Mortality data were derived by linking with data from the Korean National Statistics Office (KNSO). Multivariable logistic regression and cox regression test were performed to assess the risk factor of NPSLE and predictors of mortality. Results Of 1121 SLE patients, 429 (38.2%) patients had NPSLE events according to ACR definitions and 216 (19.3%) by Ainiala criteria. In multivariable logistic regression analysis, year of education [Odds ratio (OR) 0.92, 95% confidence interval (CI) 0.87 to 0.96, p<0.01] and elevated anti-dsDNA antibodies (OR 0.52, CI: 0.37 to 0.76, p<0.01) decreased the risk of NPSLE. In multivariable cox regression analysis, SLEDAI without NP manifestations at enrollment increased the risk of mortality (OR 1.18, CI: 1.08 to 1.25, p<0.01) in NPSLE patients. Conclusion The 38.2% and 19.3% of SLE patients had NPSLE according to ACR and Ainiala definition of NPSLE. Year of education and elevated anti-dsDNA antibodies decreased the risk of occurrence of NPSLE. SLEDAI without NP manifestations at enrollment increased the risk of mortality in NPSLE patients.

Background and aimsTo identify the prevalence of positive antiphospholipid antibodies (aPL) and related clinical characteristics in Korean SLE patientsMethodsAmong 505 SLE patients from KORean lupus NETwork (KORNET), we selected 469 patients who underwent aPL tests within 2 years of enrollment. They were classified into two groups: 1) aPL (+) group as patients with at least one positive aPL which includes IgG anticardiolipin antibody (aCL >40 GPLU/ml), IgG anti-ß2 glycoprotein I (anti- ß2GPI >40 SGU/ml) and lupus anticoagulant (LAC), and 2) aPL (-) groups as patients with negative aPL. We compared the demographic and clinical characteristics between two groups, and clinical symptoms of thrombosis and obstetric complications were compared according to aPL positivity.ResultsThe 49 (10.4%) SLE patient had at least one positive aPL, and all 3 aPL were positive in 1 patient (0.2%). Although age, sex, autoantibody profile, and the SELENA SLEDAI score were not different between two groups, the SLICC/ACR damage index was higher in aPL (+) group (p=0.03), and 57.1% of aPL (+) patients were using aspirin. More patients in aPL (+) group underwent cerebrovascular disease (CVD) (<0.01), whereas no statistical difference was found in history of ischaemic heart disease or spontaneous abortion. Moreover, renal failure was higher in aPL (+) group (p<0.01), while lupus nephritis was comparable between two groups (p=0.70).ConclusionsTen percent of patients had positive aPL in Korean SLE patients. In addition, CVD and renal failure were more common in patients with aPL (+) SLE patients.

A classic T-cell phenotype in systemic lupus erythematosus (SLE) is the downregulation and replacement of the CD3ζ chain that alters T-cell receptor signaling. However, genetic associations with SLE in the human CD247 locus that encodes CD3ζ are not well established and require replication in independent cohorts. Our aim was therefore to examine, localize and validate CD247-SLE association in a large multiethnic population. We typed 44 contiguous CD247 single-nucleotide polymorphisms (SNPs) in 8922 SLE patients and 8077 controls from four ethnically distinct populations. The strongest associations were found in the Asian population (11 SNPs in intron 1, 4.99 × 10(-4) < P < 4.15 × 10(-2)), where we further identified a five-marker haplotype (rs12141731-rs2949655-rs16859085-rs12144621-rs858554; G-G-A-G-A; P(hap) = 2.12 × 10(-5)) that exceeded the most associated single SNP rs858554 (minor allele frequency in controls = 13%; P = 4.99 × 10(-4), odds ratio = 1.32) in significance. Imputation and subsequent association analysis showed evidence of association (P < 0.05) at 27 additional SNPs within intron 1. Cross-ethnic meta-analysis, assuming an additive genetic model adjusted for population proportions, showed five SNPs with significant P-values (1.40 × 10(-3) < P< 3.97 × 10(-2)), with one (rs704848) remaining significant after Bonferroni correction (P(meta) = 2.66 × 10(-2)). Our study independently confirms and extends the association of SLE with CD247, which is shared by various autoimmune disorders and supports a common T-cell-mediated mechanism.

Background RA treatment has improved significantly with the introduction of TNF inhibitor, but it has little or no effect in about 30% of treated patients. According to EULAR recommendation, patients who failed to TNF inhibitor should switch to a different TNF inhibitor or other biologics including abatacept (ABT), rituximab, or tocilizuamb. However, no consensus has been reached on the strategy of switching. Objectives To compare the effectiveness of second-line biologic therapy in RA patients who failed to their fist TNF inhibitor. Methods We recruited 161 patients who switched biologics after failure of one TNF inhibitor from BIOlogics Pharmacoepidemiologic StudY (BIOPSY), a prospective, national biologics registry in Korea. They were divided into two groups according to the first TNF inhibitor: 101 patients with anti-TNF monoclonal antibodies (MABs: infliximab, adalimumab, golimumab) and 60 patients with etanercept (ETN). The patients who failed with the previous MABs were categorized by MABs→MABs (n=21), MABs→ETN (n=41) and MABs→ABT (n=39). The patients who failed with ETN were divided by two groups of ETN→MABs (n=32) and ETN→ABT (n=28). Drug retention rates were compared across the three or two groups according to the first TNF inhibitor using Kaplan-Meier analysis and log-rank test. The Cox regression model was used to compare risk of drug for discontinuation between groups. Results Among three groups from patients treated with MABs, the demographic and clinical characteristics were comparable with age (p=0.99), gender (p=0.35), their disease duration (p=0.24). Concomitant medication with corticosteroid (p=0.86) and methotrexate (p=0.79) were not different, but duration of previous MABs treatment showed statistical significance (10.2±8.3 in MABs→MABs vs. 6.0±6.9 in MABs→ETN vs. 11.4±12.0 in MABs→ABT, months, p=0.03). At starting time of second biologics, disease activity (DAS28ESR 6.4±1.4 vs. 6.0±1.1 vs. 6.4±1.0, p=0.19) and functional disability (HAQ-DI 1.3±0.7 vs. 1.4±0.7 vs. 1.5±0.7, p=0.51) had no statistical significance between three groups. The drug retention rates during 20 months were comparable (53.2% in MABs→MABs vs. 55.9% in MABs→ETN, 66.0% in MABs→ABT, p=0.39). After adjusting confounding factors, the groups of MABs→ABT (HR 0.32, 95% CI 0.11-0.92) and MABs→ETN (HR 0.34, 95% CI 0.12-0.96) showed lower rate of discontinuation of second biologics compared to MABs→MABs as reference. For patients who failed to ENT, demographic and clinical characteristics were comparable between two groups. At starting time of second biologics, disease activity was higher in ETN→ABT than that of ENT→MABs (6.2±1.2 vs. 5.5±1.2, p=0.02), but the drug retention rates during 20months were not statistical different between two groups (29.8% in ENT→MABs vs. 45.7% in ENT→ABT, p=0.14). Conclusions In the clinical practice, switching to ETN or ABT in patients who failed to MABs showed more favorable drug continuation. For patients with previous ETN, both MABs and ABT had similar drug continuation rate. References Emery P, et al. Ann Rheum Dis. 2014 Jan 29. doi: 10.1136/annrheumdis-2013-203993. Disclosure of Interest None declared

Background Healthcare claims database is one of powerful sources of epidemiologic study for RA patients. It is crucial, however, to assess the validity of using diagnosis codes contained in this healthcare claims database to identify RA cases prior to utilizing them for epidemiological studies. Prior studies have examined the validity of RA diagnostic code and insisted that the addition of disease modifying anti-rheumatic drugs (DMARDs) with RA diagnostic code increase the PPV and accuracy for identifying RA patients in the claims database. However, this algorithm tends to identify only patients in use of DMARDs. Objectives To validate various algorithms using information of medication and health utilization for identifying RA patients in the healthcare claims database in Korea. Methods A total of 134,930 patients≥19 years old with seropositive RA code (M05) covering the period of January 2009 to June 2011 were included in this validation study. We regarded patients registered in the “individual copayment beneficiaries program for rare and intractable diseases (ICBP)” in Korea as gold standard RA. This registration requires an official report by doctors documenting that the patients fulfilled the 1987 ACR criteria for RA. We constructed algorithms using RA diagnostic code and incorporated claims of (1) RA treatment: DMARDs, NSAIDs, or corticosteroid, (2) Laboratory tests: ESR, CRP, or rheumatoid factor, and (3) Health utilization: outpatient visit or hospitalization. We calculated the sensitivity, positive predictive value (PPV), and accuracy of each algorithm and PPV and accuracy were mainly compared between them. Results PPV for identification of RA with RA diagnostic code and any DMARDs or biologics use was 82% and it was higher than those of combination with NSAIDs (52%), corticosteroid (63%), outpatients visits≥3 (74%) or hospitalization (63%). PPV of algorithm for identifying RA with diagnostic code and any DMARDs or biologics or hospitalization was not increased (80%) but the sensitivity (94.9%) was higher than that of RA diagnostic code and any DMARDs or biologics (94.6%). Conclusions We suggest that RA patients can be identified with the algorithm of RA diagnostic code of any DMARDs or biologics with high level of PPV in claims database. Addition of hospitalization increased the sensitivity of algorithm in identifying RA patients. References Cho SK, Sung YK, Choi CB, et al. Development of an algorithm for identifying rheumatoid arthritis in the Korean National Health Insurance claims database. Rheumatol Int. 2013;33:2985-92. Kim SY, Servi A, Polinski JM, et al. Validation of rheumatoid arthritis diagnoses in health care utilization data. Arthritis Res Ther 2011;13:R32. Disclosure of Interest None declared

Background Although subclinical liver disease is common in SLE, strikingly high levels of liver enzymes are rare. Liver enzyme abnormalities in lupus are multifactorial and can be drug induced or caused by disease activity. Objectives Our aim was to determine the cause of high levels of liver enzymes in lupus patients, particularly in patients diagnosed with toxic hepatitis. Methods We performed a retrospective chart review of SLE patients treated at the Inje University Hospital between 2001 and 2013. We defined liver enzyme abnormality as a ≥2 fold increase in 2 or more of the 4 components: bilirubin, AST, ALT and LDH or ALP. Toxic hepatitis was diagnosed by a score 5 in the Roussel Uclaf Causality Assessment Method (RUCAM) and classified according to their pattern of liver enzyme: cholestatic, hepatocellular and mixed. Results Clinical and laboratory findings of liver disease was reviewed in 301 SLE patients who met ≥4 ACR criteria. Of 301 SLE patients, 74 (24.6%) met strict criteria for the liver enzyme abnormality and had the following diagnoses: toxic hepatitis (n=20), viral hepatitis (n=3), autoimmune hepatitis (n=5), liver enzyme elevation associated with infection (n=4) and an indeterminate clinical diagnosis, presumably associated with lupus activity (n=42). In total, 20 patients (6.6%) had presumed toxic hepatitis associated with either herbal medicines (n=8), anti-tuberculosis (Tb) drug (n=4), antibiotics (n=5), NSAIDS (n=1), or anticonvulsant drug (n=2). There were striking abnormalities in patients with toxic hepatitis (mean peak values: AST 610±551 U/L, ALT 369±352 U/L, ALP 598±387 U/L, LDH1,128±761 U/L). 5 had a cholestatic pattern, 9 had a hepatocellular pattern, 3 had a mixed pattern and 3 were undeterminated. Among the 20 patients with toxic hepatitis, 17 were found to have active lupus (SLEDAI≥4). After cessation of the suspected causative medication and subsequent steroids treatment, liver enzymes were improved. Conclusions In our study, herbal medicines was the most common cause of toxic hepatitis in Korean lupus patients. Antibiotics and anti-Tb drugs also could lead to toxic hepatitis in lupus patients. Since most herbal medicines contain a mixture of various products, we could not ascertain what specific ingredient lead to increase in liver enzyme levels. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.2193

Background Recently, it was reported that the antinuclear antibody (ANA) before starting anti-TNF treatment or development of ANA might be related with clinical response of TNF inhibitors or side effect of infusion reactions. Objectives The aims of this study were to examine the prevalence of positive ANA before starting TNF inhibitors in patients with RA and to investigate the impact of ANA positivity on persistency of anti-TNF treatment. Methods Among 355 RA patients who started their first TNF inhibitor between 2000 and 2011, 273 patients who had performed ANA test before starting anti-TNF treatmentwere identified in theretrospective biologics registry in Korea of REtrospective study for Safety and Effectiveness of anti-RA treatment with biologiCs (RESEARCh). We classified all patients into two groups according topositivity of ANA; positive group (ANAtitres at a dilution ≥1:160) and negativegroup (ANAtitres <1:160). The period of observation for evaluating drug survival was defined as the time between starting of TNF inhibitors and their discontinuation or their last study visit if they continued therapy. Kaplan-Meier survival analysis was used to assess persistence, log-rank tests were used to compare drug survival between ANA positive and negative groups, and Cox proportional hazards models were used to assess whether ANA positivity is a potential predictor of treatment discontinuation. Results Among 273 RA patients (mean age of 47.9 years and 85% of female), 131 (48%) patients were ANA positive before starting TNF inhibitors. At the baseline, ANA positive group was more commonly positive with rheumatoid factor (80% vs. 65%, p<0.01) and had higher disease activity with DAS 28 (6.2±0.9 vs. 5.9±0.9, p<0.01) compared to ANA negative group, while other potential predictors were comparable between two groups. In both groups, there was no significant difference in drug discontinuation rates due to all causes (P=0.97), side effects (p=0.83), and non-effectiveness (p=0.67). Cox-proportional hazards models showed that ANA positivity was not a risk factor for discontinuation due to all causes (OR 1.02, CI 0.63-1.65), side effects (OR 1.00, CI 0.98, 1.02), and non-effectiveness (OR 1.00, CI 0.97, 1.02). Conclusions The prevalence of ANA positivity in RA patients who started TNF inhibitors were 48% and ANA positive patients had higher disease activity and RF positivity compared to ANA negative RA patients. However, ANA positivity was not associated with TNF inhibitor discontinuation due to overall causes, side effects, and non-effectiveness. Acknowledgements This study was supported by a grant of the Korea Healthcare technology R&D Project, Ministry of Health and Welfare, Republic of Korea.(HI10C2020). Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.4129

The prevalent vertebral fracture (VF) is a risk factor for future VF, which can be decreased with drug therapy. However, most VFs are not recognized clinically. Vertebral fracture assessment (VFA) by dual-energy x-ray absorptiometry (DXA) and spine x-ray can be performed to detect these prevalent VFs. This study aimed to estimate the costs, effectiveness, and radiation exposure of VF diagnostic strategies. Markov model over a 10-year period was used to calculate the medical costs for diagnostic tests and VF treatment, the reduction of incident VFs of patients who have experienced a VF, and the radiation doses in target population aged over 50. We compared three strategies: ‘VFA followed by confirmatory radiography (VFA screening) ’, ‘only VFA’ and ‘only x-ray’ every 2 years, to ‘no screening before recognition’. We assumed that all patients tested positive for VF received drug therapy. A discount rate of 5% was applied in cost. The results showed the incremental costs for women over age 50 who had VFA screening, only VFA, and only x-ray were $1,112, $1,546, and $1,270 per person, respectively. Future VF incidence was reduced by 29% in both VFA screening and only VFA and 35% in only x-ray as compared with no screening for 10 years. Radiation exposure was highest in the only x-ray strategy. Also, the effectiveness and medical costs were more increased in female and old age people than in male and over age 50. The sensitivity analyses showed that these results are robust to variety assumptions including cycle length, medical costs, and diagnostic accuracy. This study suggests that VFA screening strategy can be relevant option for new VF prevention as considering lower cost and less radiation. This study is expected to provide useful information as establishing the VF diagnostic strategy in clinical practice.

Background Neurological manifestations in systemic lupus erythematosus (SLE) are diverse. Because of its varied manifestations and low prevalence, the ACR has developed nomenclature and case definitions for neuropsychiatric SLE (NPSLE) to facilitate clinical research. Brain MRI has been used for the evaluation of neurologic symptoms. Objectives The purpose of this study was to identify characteristic brain MRI findings in NPSLE and to investigate the association between brain MRI findings and NPSLE manifestations. Methods In total, 145 brain MRIs in 126 patients with NPSLE from 2002 to 2013 from three tertiary university hospitals were retrospectively reviewed. The images were evaluated for the presence of white matter hyperintensity (WMH), gray matter hyperintensity (GMH), parenchymal defects, atrophy, enhancement, and the abnormalities in diffusion-weighted image (DWI). The number, size and location of WMH,GMH and parenchymal defects were evaluated. The NPSLE manifestations of each patient were classified according to the 1999 ACR case definitions for NPSLE syndromes. The associations between MRI findings and manifestations of NPSLE were examined. Results In total, 103 MRIs (71.0%) exhibited abnormalities among the 145 MRIs reviewed. There were 172 NP events that encompassed 16 of 19 NP syndromes. The most common MRI abnormalities were WMHs. One or more WMHs were found in 84 MRIs (57.9%) among the total 145 MRIs. GMHs were observed in 42 MRIs (29.0%). GMHs tended to involve much larger areas than WMHs. Patients with cerebrovascular disease or seizures were more likely to have GMHs than patients with other NP manifestations. 33 MRIs among 42 MRIs which had GMHs also exhibited WMHs. Parenchymal defects were found in 35 MRIs (24.1%). Atrophy was detected in 26 MRIs (17.9%). Brain MRIs were enhanced in 21 of the 126 cases that had undergone enhancement. Patients who had seizures were more likely to demonstrate MRI enhancement than patients with other NP manifestations. DWIs were obtained in 102 MRIs and abnormal DWIs were obtained in 18 MRIs cases. Patients with cerebrovascular disease were more likely to have GMH, parenchymal defects and abnormal DWI than patients with other NP manifestations Conclusions Diverse brain MRI abnormalities were observed in the brain MRI of patients with NPSLE. In addition to WMHs, which were previously known as SLE findings, we also noted the presence of GMHs, parenchymal defects and abnormal DWI in a substantial portion of SLE patients, particularly in those with cerebrovascular disease or seizure. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.2216

Background Patients with rheumatoid arthritis (RA) have an increased risk of fracture caused by loss of joint function, medication and decreased bone mineral density (BMD). Objectives This study aimed to identify the incidence rate of fractures in RA patients and to explore the possible risk factors for fractures in RA patients. Methods A total of 3,557 RA patients in the KORean Observational study Network for Arthritis (KORONA) were included in this study. The information of new fractures was obtained through annual patient self-reported questionnaires. Their demographic profile, disease-specific outcomes, history of prior fracture, medications, and available BMD data were collected. The standardized incidence ratio (SIR) for fractures in RA patients was calculated and possible risk factors for fractures were explored using the logistic regression analysis. To identify the effect of bisphosphonate in RA patients with osteoporosis, we further performed logistic regression analysis. Results A total of 3,557 RA patients were included in this study and the 194 patients with 215 fractures were observed over a mean follow-up of 18 months. Vertebral fractures (n=36) were the most common fractures, followed by wrist (n=33) and rib fractures (n=21). Total fracture crude incidence rates were 41, 34.5 and 42.1 per 100,000 person year for total, male, and female patients, respectively. The SIRs of all types and major osteoporotic fractures were 3.11[95% confidence interval (CI) 2.69-3.53], and 1.41 (CI 1.05-1.86), respectively. Patients who have experienced any fracture had older age (p<0.01), longer disease duration (p=0.02), having osteoporosis (p<0.01), history of prior fracture (p<0.01) and were currently using corticosteroid (p=0.03) and bisphosphonates (p<0.01). Since BMD is most strong predictor for future fracture, logistic regression to identify the risk factors for fractures was performed for patients who had BMD data at baseline (n=866), and only having osteoporosis [Odds ratio (OR) 2.69, CI 1.17-6.18, p=0.02] was independently related to incident fracture in RA patients. In addition, older age (OR 1.03, CI 1.00-1.06, p=0.05) tend to increase the fracture incidence. Subgroup analysis for osteoporosis patients (T score <-2.5, n=210), higher functional disability with HAQ increased fracture risk (OR 2.06, CI 1.07-3.96, p=0.03) and the use of bisphosphonate showed a protective effect for future fracture (OR 0.33, CI 0.13-0.83, p=0.02) in RA patients. Conclusions Rheumatoid arthritis patients had a 3-fold increased risk of fractures as compared to those of general population. Classical risk factors for fracture such as osteoporosis with BMD and old age were associated with fractures in RA patients. In RA patients with osteoporosis, higher functional disability was a risk factor for fracture, while the use of bisphosphonate showed protective effect on their future fracture. Acknowledgements This study was supported by a grant of the Korea Healthcare technology R&D Project, Ministry of Health and Welfare, Republic of Korea.(HI10C2020). Disclosure of Interest : None declared DOI 10.1136/annrheumdis-2014-eular.3982

Background The concerns about development of adverse events (AEs) in elderly RA patients as a result of age-related changes in drug metabolism and the presence of comorbid illnesses are emphasizing due to increasing prevalence of rheumatoid arthritis (RA) in old age. However, they tend to be inadequately represented in RA clinical trials because of the exclusion criteria that are commonly applied. The tolerability and safety of TNF inhibitors in elderly patients have not been also evaluated in clinical practice. Objectives This study aimed to determine the safety of TNF inhibitors and predictors of its discontinuation in elderly RA patients. Methods We recruited 429 RA patients [838 patients-year (PY)] treated with TNF inhibitors from a retrospective biologics registry in Korea of REtrospective study for Safety and Effectiveness of Anti-RA treatment with biologiCs (RESEARCh). We divided patients into two groups of elderly (age ≥65 years) and young RA patients (age <65 years). The period of observation for evaluating safety was defined as the time between starting of TNF inhibitors and termination of treatment. Incidence rate (IR) of serious adverse events (SAEs) in elderly patients was compared to that of young patients. Drug retention rates of both groups were compared using Kaplan-Meier analysis. Potential predictors of discontinuation of TNF inhibitors were also examined using multivariate logistic regression analyses. Results Among total RA patients treated with TNF inhibitors (n=429), the 13.5% of patients (n=58, 110 PY) was classified in elderly group and 371 patients (728 PY) were included in young group. At the baseline, the elderly group had more male than young group (27.6% v.s. 11.9%, p=0.003) and were treated with lower dosage of methotrexate (13.0±3.64 v.s. 14.2±3.55 mg/week, p<0.001). Elderly patients had more comorbid condition such as pulmonary disease, diabetes mellitus and hypertension (p<0.001) and pulmonary tuberculosis history (p=0.017). For the safety, IR of SAE were slightly higher in elderly patients (7.27/100 PY) than young patients (5.22/100 PY). In elderly group, IR of malignancy (1.82/100 PY v.s. 0.96/100 PY) and respiratory disorder (1.82/100 PY v.s. 0.41/100 PY) were higher than young group. In the analysis of retention rates of TNF inhibitors, there was no significant difference between elderly and young patients (p=0.553 by log-rank test). However, there was a difference in the frequent reasons of discontinuation; development of AE (26.32%) and patients9 need (26.32%) in elderly patients, while that was ineffectiveness (34.93%) in young patients. The predictors of discontinuation were also differ in two groups; age (OR 2.01, CI 1.10-1.45) and glucocorticoid use ≥5mg (OR 4.33, CI 1.01-18.60) in elderly group, while short disease duration (OR 2.01, CI 1.10-1.45), first user (v.s. switcher, OR 1.96, CI 1.19-3.13) and infliximab use (v.s. etanercept, OR 1.67, CI 1.05-2.64) in young group. Conclusions The IR of malignancy and respiratory disorder could be increased in elderly patients treated with TNF inhibitors. The retention rate of TNF inhibitors in elderly RA patients was comparable with young patients. Acknowledgements This study was supported by a grant of the Korea Healthcare technology R&D Project, Ministry of Health and Welfare, Republic of Korea.(HI10C2020). Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.3967

Background Systemic lupus erythematosus (SLE) has very high economic burdens on society and healthcare system. Objectives The aim of this study was to estimate the annual direct costs and predictors of cost in Korean patients with SLE. Methods Total 749 patients with SLE were recruited in the Hanyang BAE lupus cohort in Seoul, Korea. We assessed the annual direct costs for the 12 months in 2010. Information was taken directly from the chart review and hospital database. And, the medication costs, which are not available from hospital database, were obtained by micro-costing methodology. Results The estimated total annual direct medical costs amounted to $3,305 (2010 US dollars), of which 60.4% was accounted for by inpatient costs and 39.6% by outpatient costs. Among the cost domains for total direct medical costs, the biggest component was the costs of medication. The mean medication costs were $1,251, which accounted for 38.4% of the total healthcare costs, followed by costs for diagnostic procedures and tests, accounting for 35.6% of the total. The annual direct medical costs increased significantly according to SLICC/American College of Rheumatology Damage Index (SDI) subgroup (total SDI=0, 1 or 2, and ≥3), both at enrollment and in 2010 (p=0.0049 vs. p<0.0001). And, the presence of renal and neuropsychiatric SDI incurred higher costs (p=0.0002 vs. p=0.0007). Total reimbursement rates of patients with SLE were 66.3%, and copayment and non-reimbursement comprised 10.8% and 22.9%, respectively. Reimbursement rates have shown a tendency to increase, whereas, out-of pocket was decreasing gradually each year between 2007 and 2010. In the multivariate regression analyses, the predictors of increased direct costs were higher disease activity (as expressed by the adjusted mean SLEDAI score), higher organ damage (as expressed by the SDI score) and renal and hematologic involvement, whereas longer disease duration predicted lower direct costs. Conclusions We have reported on the first cost-identification study in Korean patients with SLE. This analysis indicates that persons with SLE incurred a mean annual direct cost of $3,305 in 2010 US dollars. Longer disease duration predicted lower costs, whereas higher disease activity, higher organ damage, renal and hematologic involvement predicted higher costs. Disclosure of Interest S.-Y. Park Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., J. Shim Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., D. Kim Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., J. Choi Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., S.-Y. Bang Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., C.-B. Choi Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., Y.-K. Sung Grant/research support from: This study was supported in part by GlaxoSmithKline Korea., S.-C. Bae Grant/research support from: This study was supported in part by GlaxoSmithKline Korea.

Background The importance of tight control is supported by solid clinical evidence. While treating RA with the aim of achieving remission or a low disease activity-as determined by disease activity score employing 28 joints count (DAS28) or the simplified and the clinical disease activity index (SDAI, CDAI) -is reasonable and important, this approach is not widely accepted in the clinical situation. Objectives To identify the effects of tight control of rheumatoid arthritis (RA) on various disease outcomes in a large observational study. Methods We selected 1900 RA patients with a baseline DAS28-ESR of more than 3.2 and who had 1 year of follow-up data. The patients were divided into two groups: (1) disease-modifying antirheumatic drugs (DMARDs)-changed group (patients who changed the types or amounts of their DMARDs) and (2) DMARDs-unchanged group (patients who maintained their DMARDs). We measured various disease outcomes, including the Health Assessment Questionnaire–Disability Index (HAQ-DI), DAS28-ESR, C-reactive protein (CRP), ESR, and global health assessments by both physicians and patients. The t-test was used to identify the effects of tight control of RA on various disease outcomes. Results Patients in the DMARDs-changed group were younger, had a shorter disease duration, used less leflunomide, used more biologic agents. At baseline, they had higher DAS28-ESR (4.62±0.96 in DMARDs-changed group versus 4.42±0.90 in DMARDs-unchanged group, p<0.001), CRP (1.15±1.62 versus 0.81±1.14, p<0.001), global health assessment by both physician (34.05±20.00 versus 26.94±18.16, p<0.001) and patient (48.96±25.34 versus 46.18±24.74, p=0.016). In the comparison between baseline and the 1-year follow-up, the DMARDs-changed group showed greater improvements in HAQ-DI (–0.76±1.43 versus –0.59±1.22, p=0.006), CRP (–0.39±1.71 versus –0.11±1.37 mg/dl, p=0.001), and the global health assessment by a physician (–11.21±24.18 versus –8.10±19.20, p=0.002). Image/graph Conclusions Immediate changes in DMARDs according to disease activity can improve disease outcomes, especially DAS28-ESR, CRP, and the global health assessment by a physician. Disclosure of Interest None Declared

Background Early and aggressive treatment offers better outcomes in early rheumatoid arthritis (RA). The long-term effect of early treatment and effect of early treatment on functional disability have not been studied. Objectives To determine whether early diagnosis and treatment have long-term benefits for disease activity and functional disability in patients with RA enrolled in the KORean Observational study Network for Arthritis (KORONA). Methods A total of 4540 rheumatoid arthritis patients completed questionnaires to establish their demographic profile, medical history, disease-specific outcomes, and RA-related information. We defined early and delayed RA diagnoses as lag-times of shorter than and at least 1 year. Disease activity was measured with the DAS28–ESR, and it was dichotomized using a score of 2.6. Functional disability was measured with the Korean HAQ-DI, and was dichotomized using a score of 1. We used the chi-square test and t-test to identify differences between male and female patients, and crude and multiadjusted models to identify the impacts of early diagnosis on disease activity and functional disability. Results In crude and multiadjusted models, early diagnosis was not associated with a higher remission rate (OR 1.10, CI 0.94-1.28 in crude model, OR 1.05, CI 0.83-1.32 in multiadjusted model). We performed subgroup analysis according to disease duration; there was no association between early diagnosis and remission rate. Early diagnosis was associated with functional disability in crude model (OR 0.79, CI 0.68-0.91) but not in multiadjusted model (OR 0.85, CI 0.72-1.002). In subgroup analysis according to disease duration and disease activity, early diagnosis was independently associated with functional disability for a shorter disease duration (OR 0.78, CI 0.63-0.96 in disease duration <10 years) and the presence of moderate-to-severe disease activity (OR 0.83, CI 0.69-0.99 in DAS28-ESR ≥3.2). Image/graph Conclusions Early diagnosis is associated with functional disability especially with a disease duration of less than 10 years and the presence of moderate-to-severe disease activity. Disclosure of Interest None Declared

To investigate the level of costimulating molecules in systemic lupus erythematosus (SLE), we assessed the plasma concentrations of soluble forms of costimulatory molecules such as programmed death-1 (PD-1), B7-H1 (also called PD-L1 or CD274) and inducible costimulator ligand (ICOSL) in patients with SLE. Plasma concentrations of soluble PD-1, B7-H1 and ICOSL were measured by ELISA using plasma samples from 57 SLE patients with or without the active disease, 21 rheumatoid arthritis (RA) patients and 35 healthy subjects. We also evaluated surface ICOSL expression on B cells using flow cytometry to gain a better understanding of ICOSL expression. To compare the level of ICOSL mRNA expression, reverse transcriptase-polymerase chain reaction (RT-PCR) was performed using total RNA from peripheral blood mononuclear cells (PBMCs) isolated from eight healthy subjects and 11 patients with SLE. The concentration of plasma ICOSL was significantly higher in patients with SLE compared with healthy subjects ( P = 0.005). Plasma ICOSL concentrations in patients with active SLE were also significantly higher than those of either patients with inactive SLE or patients with RA ( P = 0.001, P = 0.015, respectively). Plasma ICOSL concentrations in patients with SLE correlated modestly with the SLE disease activity index score ( r = 0.298, P = 0.024). We also found a significant inverse correlation between the soluble ICOSL expression and the surface ICOSL expression on B cells ( r = −0.690, P = 0.001). However, ICOSL mRNA levels of patients with SLE were comparable with those of the control subjects. There was also no significant difference in plasma B7-H1 concentrations between groups, and plasma PD-1 was not detectable in any of the groups. The plasma concentration of soluble ICOSL might be correlated to the disease severity of lupus. The increased levels of ICOSL in active lupus suggest that this pathway is involved in the pathogenesis of SLE. The mechanism and physiological role of soluble ICOSL in the pathogenesis of SLE, however, remains to be investigated.