I.A. van den Oever

Thromboembolic and cardiovascular risk in rheumatoid arthritis: role of the haemostatic system

Circumstantial evidence suggests that the innate immune system and coagulation system share a common evolutionary origin, which explains the extensive crosstalk between inflammatory cytokines and coagulation factors, with many components being important for both systems. This crosstalk has been extensively studied in sepsis, an acute state of high-grade inflammation. However, rheumatoid arthritis (RA) as well as many other autoimmune diseases can also be considered as a prothrombotic state. More and more studies show that autoimmune diseases, including RA, are a risk factor for cardiovascular disease, and also for venous thromboembolic events, such as pulmonary embolism and deep vein thrombosis. Inflammation and its effect on the haemostatic system is probably the link between these diseases. This viewpoint gives an update of the current literature on thromboembolic risk in RA, but also documents important knowledge gaps. This viewpoint will therefore help to focus on further research topics to improve diagnostic and therapeutic options which may relieve both the proinflammatory and the prothrombotic burden of autoimmune diseases.

Subclinical renal dysfunction is independently associated with cardiovascular events in rheumatoid arthritis: the CARRÉ Study

Background Patients with rheumatoid arthritis (RA) have double the risk of cardiovascular (CV) disease, largely independently of traditional CV risk factors. Renal dysfunction is associated with CV morbidity and mortality in the general population, but data on this association in RA are lacking. Objective To investigate the association between renal function and CV events in RA. Methods The CARRÉ Study is an ongoing prospective cohort study of Dutch patients with RA, which records CV events. Glomerular filtration rate (GFR) was estimated with the abbreviated Modification of Diet in Renal Disease formula. Logistic regression determined the association between estimated GFR and the occurrence of CV events. Results 353 patients were followed for 3 years, and 23 (7%) had a CV event. Patients who had an event had a significantly lower baseline GFR than those who did not (59 vs 79 ml/min, p=0.001). This association remained significant after adjustment for traditional risk factors: in this analysis, a decrease in GFR of 5 ml/min was associated with a 30% (95% CI 7% to 59%) increase in the occurrence of CV events. During follow-up, an unfavourable change in GFR was noted in patients who later had a CV event compared with those who did not. Conclusion These data confirm that, in RA, renal dysfunction is associated with a higher risk of CV disease independently of traditional CV risk factors.

FRI0056 Half of RA Patients Does not Receive Adequate Cardiovascular Risk Management despite A Very High 10-Year Cardiovascular Risk

Background Patients with rheumatoid arthritis (RA) have an increased risk for cardiovascular (CV) events. Since 2011, RA is considered as an independent risk factor for cardiovascular disease (CVD) in the Dutch cardiovascular risk management (CV-RM) guidelines [1]. To date it is unknown whether RA patients receive adequate preventive treatment for CV risk as described in the guidelines. Objectives To investigate if and to what extent the Dutch CV-RM guidelines are implemented in the RA population. Methods Consecutive patients were recruited from two outpatient rheumatology clinics: 1. Reade, Centre of revalidation and rheumatology in Amsterdam and 2. Antonius Hospital in Sneek. CV risk factors, including age, systolic blood pressure, smoking status and total cholesterol/HDL-cholesterol ratio were assessed and subsequently, the 10-years CV risk was calculated. A 10-year risk of developing a cardiovascular event of 20% or higher was classified as high CV risk. A risk below 10% was classified as low and 10 to 20% as intermediate CV risk. Results In total, 390 RA patients were included in this study: 125 from Antonius Hospital and 265 from Reade. Table 1 displays the relevant demographic and cardiovascular risk factors of the total population and the two rheumatology clinics separately. Of the total group 269 RA patients (69%) had a high, 57 (15%) an intermediate and 64 (16%) a low 10-year CV risk (figure 2). Ninety percent of the high CV risk patients should be treated with antihypertensive and/or statin treatment, however only 45% actually used one or both medications (figure 2). Conclusions Two third of the RA patients have a high cardiovascular risk and need cardiovascular risk management. Strikingly, in half of these patients adequate CV-RM is lacking. Therefore, strategies to optimize CV-RM in RA patients are urgently needed. References The Dutch College of General Practitioners, Multidisciplinary guidelines for cardiovascular risk management, revision 2011 Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.3689

AB0456 Good Response on Tocilizumab for Patients with Rheumatoid Arthritis Persists for 3 Years

Background With the advent of biological therapy a range of new drugs have come available for patients with rheumatoid arthritis (RA). Tocilizumab (TCZ) was registered more recently for patients with RA. Obviously, data on long-term efficacy and safety in daily clinical practice is important as results obtained in short-term clinical registration trials might differ from those observed in long-term daily clinical practice. Objectives To ascertain the efficacy of long- term TCZ treatment for RA Methods From May 2009 to August 2014, 40 consecutive patients have been included in the TCZ RA cohort Patients were eligible for inclusion if disease activity was high despite previous treatment with TNF-inhibitor and in the absence of contraindications for the start of TCZ. TCZ was given, intravenously, in a dose of 8 mg/kg once every 4 weeks. Visits were performed regularly based on a fixed protocol. At every visits, disease activity was measured using Disease Activity Score of 28 joints (DAS28). Improvement after baseline as measured using the ΔDAS28. In addition, the Health Assessment Questionnaire was performed at baseline and half yearly thereafter. To investigate the course of DAS28 and improvement over time a t-test was used. Results The mean age was 53 years (SD:18), and 29 (73%) were female. At baseline, the median disease duration was 9 (5-18) years, 29 (74%) were rheumatoid factor positive, 25 (66) had erosive disease, 29 (76%) were anti-CCP positive. At baseline, 22 (56%) used methotrexate (MTX), 24 (62%) used prednisolone, and 11 (28%) a DMARD other than MTX. On average patients had used 3.4 (SD:1.3) DMARDs, and 2.0 (0-3.0) biologicals. The maximum follow-up duration was 4.7 years, during which a total of 23 (58%) patients were stillon drug, and 17 (43%) patients have dropped out 8 patients (47%) due to adverse events, 7 patients (41%) stopped due to inefficacy, and 2 patients (12%) stopped for other reasons. The median follow-up duration was 0.48 (IQR:0.23-1.97) years for patients currently on TCZt, and 1.02 (0.25-1.96) years for drop-outs. The mean DAS-28 dropped significantly from 5.3 (SD: 1.7) at baseline to 3.2 (1.3) at 12 weeks. This improvement sustained and increased further over 3 years, see figure. The mean score on the health assessment questionnaire (HAQ) dropped from 1.6 (0.7) to 1.3 (0.8) at 12 weeks of treatment, and remained stable after that, see figure. Conclusions This cohort of RA patients treated with TCZ intravenously followed long term showed that the initial good response on tocilizumab persists for at least 3 years of follow-up based on DAS28 and HAQ, in patients who remained on TCZ treatment. 43% of patients dropped out during 5 years of follow-up. Disclosure of Interest None declared

SAT0187 Tocilizumab Standard Dose is an Overtreatment with Regard to Systemic IL-6 Receptor Blockade in Rheumatoid Arthritis

Background Tocilizumab (TCZ) inhibits interleukine-6 (IL-6) receptor resulting in inhibition of C-Reactive Protein (CRP) production, a surrogate marker for IL-6 receptor blockade.1 Evolving evidence showed that pharmacokinetics (PK) of TCZ is probably influenced by target-binding.2,3 Objectives To investigate PK of TCZ and the relationship between TCZ concentrations and disease activity in an observational cohort of rheumatoid arthritis (RA) patients. Methods This prospective study consisted of 70 consecutive patients (The Netherlands, n=37; Spain, n=33) treated with TCZ 8 mg/ kg intravenously once per 4 weeks, monitored during 48 weeks. Disease activity was measured with CRP (mg/L) and Disease Activity Score of 28 joints (DAS28), using erythrocyte sedimentation rate (ESR)(mm/hr). TCZ serum trough concentrations and anti-drug antibodies were measured using an enzyme linked immunosorbent assay and an antigen binding test, respectively. Results At baseline, mean DAS28 was 5.4±1.4 and median CRP, 6.9 (2.4-32.5). During 48 weeks, 19 patients discontinued TCZ prematurely due to: inefficacy (n=8), adverse events (n=8) or other reasons (n=3). TCZ concentrations above 1 mg/L were sufficient to normalize CRP production (<10 mg/L)(figure 1A), which was achieved by 69-88% of patients, depending on time point of measurement. Of the 30 patients with increased CRP (>10 mg/L) at baseline, levels were normalized in 17 patients within 4 weeks. In 11 patients normalization of CRP took longer or CRP level was still increased at drop-out. Increased CRP levels in these 11 patients was accompanied by TCZ concentration below 1 mg/L. CRP data of 2 patients was missing. The concentration effect curve at week 24 (last observation carried forward for patients in whom week 12 was the last available visit) showed that the majority was overtreated with TCZ standard dose (figure 1B). At week 24, median TCZ concentration was 11 mg/L (6-19) and mean ΔDAS28 was 2.4±1.6. Mean ΔDAS28 was lower in patients with TCZ concentrations below 1 mg/L vs above, respectively, -0.3±0.8 versus 2.8±1.4 (p<0.001) (independent sample t test). One patient had detectable anti-TCZ antibodies, in combination with low TCZ concentrations, at week 4. Conclusions Increased CRP levels after baseline were accompanied by TCZ concentrations below 1 mg/L. This suggests that PK of TCZ is influenced substantially by target-binding and only marginally by immunogenicity, since, low TCZ concentration with anti-TCZ antibodies was found in only one patient. Due to the direct relationship between TCZ concentration and CRP inhibition, TCZ might be ideally suited for optimizing treatment via a personalised Therapeutic Drug Monitoring approach, aiming for serum concentrations within the optimal range for target blockade. This creates possibilities for dose reduction, since the majority of patients was overtreated with TCZ standard dose. References Swaak. Scand J Rheumatol. 1988. Nishimoto. Blood. 2008. Nishina. Ann Rheum Dis. 2014. Disclosure of Interest E. Kneepkens Speakers bureau: payment for lectures from Pfizer, I. Van Den Oever Speakers bureau: payments for lectures from BMS, C. Plasencia Grant/research support from: Pfizer, paid to institution, Speakers bureau: payment for lectures from Pfizer, D. Pascual-Salcedo Speakers bureau: payment for lectures from Pfizer, D. van der Kleij: None declared, M. Hart: None declared, M. Nurmohamed Consultant for: received consultancy fees from Abbott, Roche, Pfizer, MSD, UCB, SOBI and BMS, Speakers bureau: payment for lectures from Abbott, Roche and Pfizer, A. Balsa Grant/research support from: Pfizer, paid to institution, Consultant for: consultancy fees from Abbvie, Pfizer and MSD, Speakers bureau: payment for lectures from Abbvie, Roche and Pfizer, L. Aarden: None declared, T. Rispens Speakers bureau: received payment for lectures from AbbVie and Pfizer, G. Wolbink Grant/research support from: Pfizer, paid to institution, Speakers bureau: payments for lectures from Pfizer, Amgen, AbbVie, UCB and BMS.

THU0216 Cardiovascular Diseases, Biomechanical Factors and Activity Limitations in Patients with Knee And/Or Hip Osteoarthritis: Results of the AMS-OA Study

Objectives The aim of the study was to investigate differences in biomechanical factors and activity limitations between knee and/or hip osteoarthritis (OA) patients with and without cardiovascular disease (CVD). Methods A total of 458 consecutive patients from the Amsterdam Osteoarthritis (AMS-OA) cohort were included in this cross-sectional study. All patients were diagnosed with OA of the knee and/or hip according to the ACR criteria and referred to the outpatient rehabilitation centre. All provided written informed consent. The Cumulative Illness Rating Scale (CIRS) and a specific CVD questionnaire were used to quantify CVD. Coronary diseases, heart failure and cerebral or peripheral artery diseases were considered as CVD. Biomechanical factors were muscle strength, joint proprioception and joint laxity. Muscle strength was tested with an isokinetic dynamometer, proprioception with a joint motion detection device and varus-valgus laxity was assessed as the movement in the frontal plane in a sitting position. Activity limitations were assessed by Western Ontario and McMaster University Osteoarthritis Index physical function subscale (WOMAC-pf) questionnaire and by the get up and go (GUG) and stair-climb test. Results In total 53 OA patients (11.6%) reported CVD. Sixteen patients (3.5%) had coronary diseases, 18 (3.7%) heart failure, 22 (4.6%) cerebral artery diseases and 3 (0.7%) peripheral artery diseases. The patient characteristics of the total OA population and differences in biomechanical factors and activity limitations between patients with and without CVD are displayed in table 1. OA patients with a history of CVD were significantly older and more often male. They also had a significantly higher BMI compared to OA patients without CVD. OA patients with CVD had less severe OA according to the Kellgren/Lawrence score. Knee and/or hip OA patients with CVD showed lower outcomes for muscle strength, higher scores for proprioception and laxity and also higher scores for the activity limitations tests. However, only the scores on the GUG and stair-climb test were significantly higher in patients with CVD. Flexion and extension muscle strength scores of men with CVD were significantly lower compared to men without CVD (see table 2). In woman with CVD, these scores were also lower, but only significant for the flexion score. The GUG and stairclimb scores were significantly higher in women with CVD compared to woman without CVD. For men this was only true for the stairclimb test up. Conclusions Activity limitations and biomechanical factors differ between knee and/or hip osteoarthritis patients with and without cardiovascular diseases, whereby patients with CVD seem to have worse outcomes, compared to patients without CVD. Furthermore, there are apparent differences between men and woman. Whereas woman with CVD have significantly more activity limitations, the men with CVD have significantly less muscle strength. Future research should focus on the influence of cardiovascular disease and its risk factors on the association between muscle strength and activity limitations. Disclosure of Interest : None declared DOI 10.1136/annrheumdis-2014-eular.2244

FRI0310 Tocilizumab Levels Are Associated with Clinical Response in Patients with Rheumatoid Arthritis

Background Limited data is currently available regarding the pharmacokinetics (PK) and –dynamics (PD) of tocilizumab (TCZ) in daily clinical practice. Objectives To assess the relationship between TCZ levels and clinical response in rheumatoid arthritis (RA) patients. Methods Observational cohort study of 46 consecutive RA patients treated with TCZ 8 mg/kg intravenously once every 4 weeks in the Netherlands (n=25) and Spain (n=21), monitored during 48 weeks. Samples and clinical data were collected at least at baseline, week (w) 24 and 48. TCZ trough levels and titres of anti-drug antibodies (ADA) against TCZ were determined using an ELISA and an Antigen Binding Test (ABT), respectively. Samples which were not trough level were excluded. Disease activity was assessed using the Disease Activity Score of 28 joints (DAS28) and response was defined as a DAS28 remission (DAS28 score <2.6). Values are reported in mean ± SD or median (IQR). Results At baseline patients had a DAS28 score of 5.5±1.4 and 8 patients were biological naive. Duration of follow-up varied between 12 to 48 weeks and 15 patients discontinued TCZ treatment before w48. At w24, 40% of the patients were in remission and at w48 44%. TCZ levels varied widely among patients, at w24 the median TCZ level (mg/L) was 9.6 (3.6-16.7). Figure 1 shows that patients with very low TCZ levels at w24, despite a dosage of 8 mg/kg per 4 weeks, had not improved (ΔDAS28) compared to baseline. Moreover, some patients had high TCZ levels without additional clinical benefit. No ADA against TCZ were detected. Non-parametric testing (not corrected for confounders) showed that TCZ levels at w24 (n=37) and w48 (n=26) were significantly higher in responders compared to non-responders, respectively, 20 (6.3-31.2) vs 6.9 (0.2-11.6) (p=0.004) and 15.5 (7.9-32.2) vs 6.8 (2.9-12.2) (p=0.025). Conclusions Although TCZ trough levels vary greatly, immunogenicity does not seem to be an important factor in the PK/PD of TCZ. However, it might be possible, that ADA were not detected due to drug interference. Another explanation for the variation of TCZ trough levels could be target-binding. Furthermore, this study shows that TCZ trough levels are associated with clinical outcome in RA and that some patients are currently over- or undertreated with TCZ 8 mg/kg per 4 weeks. Therefore, assessing TCZ levels may help to optimize treatment in patient treated with TCZ. Disclosure of Interest E. Kneepkens: None declared, I. Van Den Oever: None declared, C. Plasencia Grant/research support: Pfizer, D. Salcedo Pascual Grant/research support: Pfizer, Speakers bureau: Pfizer, M. Lopez-Casla: None declared, D. Van Der Kleij: None declared, M. Nurmohamed Consultant for: AbbVie, Roche, Pfizer, MSD, UCB, SOBI and BMS, Speakers bureau: AbbVie, Roche, Pfizer, T. Rispens Speakers bureau: AbbVie, A. Balsa Grant/research support: Pfizer, Speakers bureau: Pfizer, Roche, AbbVie, G. Wolbink Grant/research support: Pfizer, Speakers bureau: Pfizer, Amgen DOI 10.1136/annrheumdis-2014-eular.2915

FRI0114 The effect of combination therapy and prednisolone on haemostatic markers in rheumatoid arthritis: The cobra-light study:

Background There is accumulating evidence that rheumatoid arthritis (RA) should be considered as prothrombotic state, explaining the increased risk of thromboembolic events. Suppressing inflammation could reduce this hypercoagulability. COBRA, the combination of step-down prednisolone, methotrexate and sulfasalazine (SSZ) is an effective antirheumatic therapy. Whether glucocorticoids induce a procoagulant state is controversial and depends on the population, thus the effect of prednisolone in RA is unknown. Objectives To evaluate the course of haemostatic markers in RA patients during anti-inflammatory therapy and the dose-dependent effect of prednisolone on coagulation and fibrinolysis. Methods 22 patients diagnosed with early RA, were randomised to either COBRA therapy or an attenuated form (COBRA-light) with halved initial prednisolone dose and without SSZ. At baseline and after 1, 4 and 26 weeks of treatment, 10 ml of citrated blood was collected for measurement of prothrombin time (PT), activated partial thromboplastin time (aPTT) and five haemostatic markers: prothrombin fragment 1+2 (F1+2), factor VIII (FVIII), von Willebrand factor (vWF), plasminogen activator inhibitor (PAI-1) and D-dimer. For statistical analyses t-tests and linear regression were used. Results Baseline characteristics were not significantly different between the 2 groups (each 11 patients). DAS44, CRP, ESR, aPTT, D-dimer and F1+2 decreased during treatment in all patients (table). There was a significant positive association between decrease in CRP and BSE with D-dimer and decrease in CRP, BSE and DAS44 with F1+2 at all time points. There was no difference in the markers between the two groups, except for a stronger decrease in aPTT after 2 weeks (p=0.03) in the COBRA group. This difference was no longer seen at 4 weeks. Table 1. Change in haemostatic factors in all 22 RA patients Baseline 2 weeks 4 weeks 26 weeks DAS44 3.9 (0.6) – – 1.6 (0.9)** ESR (mm/h) 25 (12-43) – – 4 (3-7)** CRP (mg/L) 13.0 (5.5-29.5) 2.5 (2.5-2.7)** 2.5 (2.5-4.5)** 2.5 (2.5-2.7)** aPTT (sec) 29.9 (28.5-34.9) 26.2 (25.1-31.4)** 26.8 (25.5-29.9)** 27.9 (26.6-30.4)* PT (sec) 11.2 (0.4) 11.3 (0.5) 11.1 (0.5) 11.2 (0.5) FVIII (%) 142 (105-179) 151 (112-196)* 152 (123-193) 127 (111-170) D-dimer (FEU/L) 1.18 (0.63-3.34) 0.55 (0.29-1.26)** 0.49 (0.26-0.84)** 0.25 (0.21-0.78)** F1+2 (pmol/L) 342 (253-496) 192 (146-305)* 212 (160-285)** 213 (153-319)* PAI-1 (ng/ml) 35 (25-49) 40 (19-63) 24 (16-53) 42 (17-73) vWF (%) 115 (87-149) 121 (86-157) 138 (94-189)* 101 (86-143) Values are mean (SD) or median (IQR) *p<0,05 and **p<0.001 change compared to baseline. Conclusions Overall, both COBRA and COBRA-light therapy induced an improvement of inflammatory and procoagulant factors in RA. There were no remarkable differences in the haemostatic markers between the two groups, indicating that doses higher than 30 mg of prednisolone have no attributable effect on the procoagulant state in RA. Disclosure of Interest None Declared

FRI0100 Sustained development of cardiovascular disease in rheumatoid arthritis despite cardioprotective treatment: The 7-year prospective carre-study

Background Rheumatoid arthritis (RA) is a chronic inflammatory joint disease which is associated with an increased cardiovascular (CV) risk. Still unknown is whether CV risk factors or the underlying inflammatory process in RA renders these patients more at risk of CV disease (CVD). An increment in these factors over several years might explain this predisposition. Objectives The present study compared changes in these factors over time in RA-patients who did and did not develop CVD during follow-up. Methods 7-year incidence rate of CVD was determined in a prospective cohort of 353 RA-patients. CV risk factors, RA-related factors and medication use were assessed at baseline, at 3-years and at 7-years of follow-up. Associations between the changes in investigated factors and development of CVD were assessed using generalized estimating equation (GEE) analyses. Results After 7 years, there were 62 events over 2361 patientyears of follow-up, incidence rate (IR) of 26.3/1.000 patientyears. This was similar to the IR after 3-years of follow-up. GEE analyses showed that, during follow-up, changes in disease activity score of 28 joints (DAS28) and overall 10-year CV risk estimation (as calculated by SCORE) were associated with incident CVD. Use of biologics increased significantly in patients who did not develop incident CVD vs. those who did (26% vs. 6%), while use of statins increased significantly in patients who did develop incident CVD vs. those who did not (43% vs. 14%). Conclusions An increased risk of incident CVD persists in patients with RA. Changes in both DAS-28 as well as 10-year CV risk estimation (SCORE) were associated with the 7-year CVD incidence. Cardioprotective treatment did not seem to extenuate this association. A more aggressive cardioprotective and anti-inflammatory treatment of RA might mitigate the burden of CVD in RA. Disclosure of Interest None Declared

THU0045 Score and framingham predict cv disease incidence in RA patients better than intima-media thickness

Background Rheumatoid arthritis (RA) is a chronic inflammatory disease with an increased cardiovascular (CV) risk. Carotid intima-media thickness (cIMT) predicts CV events in the general population, and hence, is used as a non-invasive screening tool to identify patients at high CV risk. In RA, it is still unclear whether estimated 10-year CV risk models, such as the Systematic Coronary Risk Evaluation (SCORE) and Framingham accurately predict CV disease incidence or whether cIMT measurement can equally predict CV disease incidence. Objectives To investigate the predictive value of SCORE- and Framingham risk models and cIMT for future CV events in RA. Methods CARRΈ is an ongoing cohort study of CV disease in RA. In a subpopulation (n=141) we measured cIMT at baseline; Of those, 120 had no prior CV disease and 10-year risk of fatal and nonfatal CV disease was calculated. We compared these calculations with actual CV events recorded in a mean follow-up of 9 years. During this time, IMT was measured three times. Univariate logistic regression analyses (with standardized odds ratios) investigated the extent to which individual CV risk factors, SCORE, Framingham and cIMT predicted CV disease incidence. Results Thirteen RA patients (incidence rate: 13.5%, 95% confidence interval (CI): 7.8-23.2) developed a CV event, of which four of these events were fatal (incidence rate: 4.2%, 95%>CI: 1.6-11.1). SCORE- and Framingham predictions were very close to actual CV mortality and morbidity and significantly predicted (fatal) CV events. cIMT also showed a trend in prediction of CV disease. cIMT progressed with an annual rate of 0.005mm. Table 1. Logistic regression of CV disease incidence with CV risk factors, IMT and CV risk models Standardized OR (95% CI) p Age 1.80 (1.00-3.23)* 0.048 Systolic blood pressure 1.64 (0.94-2.86) 0.08 Atherogenic index 1.79 (1.06-3.01)* 0.03 cIMT 1.67 (0.97-2.85) 0.06 3-yearly IMT progression 1.19 (0.71-1.99) 0.52 7-yearly IMT progression 1.13 (0.50-2.55) 0.77 10-year Framingham CV-risk 1.82 (1.12-2.96)* 0.02 10-year SCORE CV-risk 1.83 (1.17-2.88)* 0.009 *Significant association between investigated variables and CV disease incidence (p≤0.05). Conclusions This long term follow up study underscores the value of SCORE and Framingham in prediction of CV disease incidence, also in RA. In this study of RA patients the evidence for cIMT, a surrogate marker of CV disease in the general population, had no predictive value. Disclosure of Interest None Declared