Full Bayesian identification of linear dynamic systems using stable kernels

System identification learns mathematical models of dynamic systems starting from input-output data. Despite its long history, such research area is still extremely active. New challenges are posed by identification of complex physical processes given by the interconnection of dynamic systems. Examples arise in biology and industry, e.g., in the study of brain dynamics or sensor networks. In the last years, regularized kernel-based identification, with inspiration from machine learning, has emerged as an interesting alternative to the classical approach commonly adopted in the literature. In the linear setting, it uses the class of stable kernels to include fundamental features of physical dynamical systems, e.g., smooth exponential decay of impulse responses. Such class includes also unknown continuous parameters, called hyperparameters, which play a similar role as the model discrete order in controlling complexity. In this paper, we develop a linear system identification procedure by casting stable kernels in a full Bayesian framework. Our models incorporate hyperparameters uncertainty and consist of a mixture of dynamic systems over a continuum spectrum of dimensions. They are obtained by overcoming drawbacks related to classical Markov chain Monte Carlo schemes that, when applied to stable kernels, are proved to become nearly reducible (i.e., unable to reconstruct posteriors of interest in reasonable time). Numerical experiments show that full Bayes frequently outperforms the state-of-the-art results on typical benchmark problems. Two real applications related to brain dynamics (neural activity) and sensor networks are also included.

POS1069 COVERAGE OF THE SWEDISH RHEUMATOLOGY QUALITY REGISTER: TO WHAT DEGREE ARE b/tsDMARD TREATMENTS FOR PSORIATIC ARTHRITIS RECORDED?

Background

In a national context, the Swedish Rheumatology Quality register (SRQ) is a major source of information on clinical data for patients treated with biological and targeted synthetic disease-modifying anti-rheumatic drug (b/tsDMARD) for rheumatic diseases. Data from SRQ are fundamental for research on drug effectiveness and safety.

Objectives

To understand to what degree patients with psoriatic arthritis (PsA) and their dispensed b/tsDMARD treatments are recorded in the SRQ.

Methods

We identified all dispensed oral or subcutaneously administered b/tsDMARDs, approved for use in PsA, from the Swedish Prescribed Drug register (PDR, with 100% coverage) in 2018-2019. We required the patients who received the dispensation to have at least one ICD code for PsA (L405, M070, M071, M072, M073) as main diagnosis from a visit to a rheumatology or internal medicine (IM) unit in the National Patient Register (NPR) before dispensation, but no main diagnosis of rheumatoid arthritis (ICD code: M05, M06 (excluding M06.1 and M06.4), M12.3). Furthermore, to limit the assessment to patients with contemporary contact with the specialized rheumatology care, we also required at least one visit with a PsA main diagnosis from rheumatology/IM during 2017-2019. We then checked if the patients and their treatments were registered in SRQ. In a sensitivity analysis, we excluded patients with a visit in dermatology within 6 weeks prior to the first prescription of each b/tsDMARD, in order to exclude patients being prescribed the drug for cutaneous psoriasis.

Results

In 2018-2019, a total of 7922 unique b/tsDMARD prescriptions had been dispensed to 6311 patients with PsA, having contemporary contact with the specialized rheumatology care. Of them, 5687 patients were registered in SRQ (90.1%), of which 94.4% with a PsA diagnosis and 95.5% with at least one registration of a b/tsDMARD. The coverage of the single drugs in SRQ, as compared to dispensations in PDR, ranged between 53.5% to 93.3% (69.3% to 95.8% considering only patients also in SRQ), with the tumor necrosis factor inhibitors (TNFi) having the best coverage (79.4-93.3%) (Table 1). In a sensitivity analysis, among the 5290 patients without a main diagnosis of psoriasis before start of treatment, 4919 (93%) were registered in SRQ, of which 94% with a PsA diagnosis and 96% with at least one registration of a b/tsDMARD.

Table 1.

coverage of the single b/tsDMARD approved for use in PsA patients in the SRQ as compared to the prescribed drug register, 2018-2019

ATCDrug*Total patients in PDRPatients with the same drug in SRQPercentagePercentage only considering patients included in SRQL04AB01Etanercept2753234885.3%92.5%L04AB04Adalimumab2436193379.4%88.8%L04AB05Certolizumab pegol24321387.7%89.9%L04AB06Golimumab41839093.3%95.8%L04AA24Abatacept947377.7%83.9%L04AA32Apremilast54729453.7%69.3%L04AC10Secukinumab83365678.8%84.8%L04AC13Ixekizumab1187160.2%74%L04AC05Ustekinumab21913863%76.2%L04AA29Tofacitinib26119474.3%77%

* In Sweden intravenously administered drugs are not prescribed and therefore not included in the PDR. Also, some hospitals administer ustekinumab in hospital, due to its high cost. However, all b/tsDMARDs, irrespectively of administration route, can be registered in SRQ.

Conclusion

The coverage of b/tsDMARD-treated PsA patients in SRQ as compared to the PDR in Sweden is ranging between 90 and 93%, and SRQ is performing well in the registration of the single drugs, especially for TNFi. These results suggest that research studies based on data from the SRQ are representative of the Swedish b/tsDMARD-treated PsA population.

Disclosure of Interests

Daniela Di Giuseppe: None declared, Ulf Lindström: None declared, Johan K Wallman Consultant of: AbbVie, Amgen, Celgene, Eli Lilly, Novartis, Grant/research support from: AbbVie, Amgen, Eli Lilly, Novartis, Pfizer, Lotta Ljung: None declared, Johan Askling Grant/research support from: Abbvie, Astra-Zeneca, BMS, Eli Lilly, MSD, Pfizer, Roche, Samsung Bioepis, Sanofi, and UCB

OP0307 PULMONARY FIBROSIS IN EARLY RHEUMATOID ARTHRITIS IN RELATION TO GENETIC LOCI AND INDIVIDUAL ACPA SPECIFICITIES

Background

Pulmonary manifestations in rheumatoid arthritis (RA) are common comorbidities but the underlying mechanisms are largely unknown. We found in a previous study 3 SNPs associated with pulmonary fibrosis (PF); rs35705950 (MUC5B), rs111521887 (TOLLIP), and rs2609255 (FAM13A) besides age, rheumatoid factor positivity and methotrexate treatment.

Objectives

To evaluate for the added value of a multiplex of anti-citrullinated peptide antibodies (ACPA) for the development of pulmonary fibrosis (PF) in an inception cohort of RA patients.

Methods

A total of 1184 patients with early RA were consecutively included and followed prospectively from the date of diagnosis (index date) until death or until 31 December 2016. The diagnosis of PF was based on high resolution tomography. The presence of 21 ACPA fine specificities were analysed in plasma sampled at index date, using a custom-made microarray chip (Thermo Fisher Scientific, Uppsala, Sweden). Data on both ACPA and genetic data was available for 841 RA patients, of whom 50 developed PF. Associations were analysed using logistic regression analysis and presented as the odds ratio (OR) with the 95% confidence interval (CI). Models were adjusted for sex, age, DAS28 and presence of RF at RA diagnosis, smoking ever, and HLA-SE and in a second step for the three SNPs (.rs35705950, rs111521887 and rs2609255), respectively.

Results

In unadjusted analyses eight ACPA reactivities were found associated with PF development (p< 0.05-0.001). The number of ACPA reactivities was related to PF development, both in crude and adjusted models (p<0.05 for both). In models concomitantly adjusted for the three SNPs (rs35705950, rs111521887 and rs2609255) respectively, in addition to mentioned adjustments the number of ACPA reactivities (p<0.05 for all three nmodels), Vim60-75 (p<0.05, in all three models), Fibβ62–78 (72) (p<0.001-p<0.05) and F4-CIT-R (p<0.01-p<0.05) were all found significantly associated to PF development irrespective of the SNPs.

Conclusion

The development of PF in an inception cohort of RA patients was associated both with risk genes and, independently of the risk genes, the presence of certain ACPA, and the number of ACPA reactivities.

References

[1]Jönsson E, et al. Pulmonary fibrosis in relation to genetic loci in an inception cohort of patients with early rheumatoid arthritis from northern Sweden. Rheumatology (Oxford). 2021 May 16:keab441. doi: 10.1093/rheumatology/keab441.

Acknowledgements

I have no acknowledgements to declare. The staff and patients at the departments of rheumatology in northern Sweden.

Disclosure of Interests

None declared

POS0568 HOW DOES INSTANTANEOUS RA DISEASE ACTIVITY AFFECT THE SHORT-TERM RISK OF ACUTE CORONARY SYNDROME? – A REGISTER-BASED STUDY

Background

Rheumatoid arthritis (RA) patients have a reduced life expectancy, with cardiovascular disease (CVD) being the most frequent cause of death. The mechanisms behind the increased CVD morbidity and mortality in RA are not fully understood. Systemic inflammation is an important contributor to the accelerated arteriosclerosis in RA, and traditional risk factors for CVD are usually more prevalent in RA patients than in the general population. How much the current state of RA disease control impacts the short-term risk of CVD events remains unclear.

Objectives

To estimate the short-term risks and relative risks of acute coronary syndrome (ACS) in patients with RA as a function of RA disease activity, with particular focus on remission.

Methods

We identified patients with RA from the clinical rheumatology registers (CRR) in Sweden (SE) and Norway (NO), and for these patients we retrieved all registered clinical rheumatology visits from January 1st, 2012 to December 31st, 2020. At each visit, we assessed whether the patient was in remission or not according to multiple definitions including DAS28, ACR criteria, and SDAI. We also categorised the disease activity at each visit into categories (remission, low, moderate and high) using DAS28-ESR. Patients had to be free of any history of ACS in a five-year look back window (assessed at the visit date), and were followed for 6 months from each visit date until ACS event (defined as hospitalization due to ACS or death due to either ACS or sudden death) or censoring (death due to other causes, migration, end of the study). We compared the risk of ACS between patients who were (vs. were not) in remission using Cox regression with robust standard errors (accounting for the correlated data structure), adjusted for covariates (including age at the visit, sex, number of previous treatment courses, use of prednisolone, the expanded risk score in RA (ERS-RA), and defined co-morbidities: diabetes, malignancy, respiratory failure, liver failure and kidney disease) and stratified by country.

Results

We included 43,338 RA patients and their 223,197 visits (211,158 (SE), 12,039 (NO)). 74% of the visits were from women, with a mean age (SD) at visit of 62 (14) years. Several clinical characteristics including treatments and comorbidity history varied with disease activity (Table 1). By contrast, age, number of previous DMARDs, disease duration and smoking habits were relatively similar across categories of disease activity (Table 1).

Table 1.

Median [Q1-Q3] or percentage for clinical characteristics in remission (DAS28-ESR<=2.6) and high disease activity (DAS28-ESR>5.1) categories

VariableRemissionHigh disease activityN (visits)91,49725,364Age, years65[53-72]63[53-72]Disease duration10[4-18]8[2-17]N treatment courses0[0-0]0[0-1]Prednisolon28%58%Tender joint count, 28-joints (TJC)0[0-0]10[6-14]Swollen joint count, 28-joints (SJC)0[0-0]7[4-10]Erythrocyte sedimentation rate, ESR10[6-18]28[15-47]C-reactive protein (CRP)3[1-4]14[5-32]Patient global assessment, PGA13[4-27]70[55-82]Pain12[4-27]70[55-82]ERS-RA8[3-16]12[5-23]Ischemic heart disease6%7%Diabetes9%13%Hypertension39%42%Hyperlipidemia23%25%Ever smoking45%50%

We observed 598 ACS events (in 554 patients) during the 6-month follow-up window. Comparing patients not in remission to patients in remission, adjusting for the covariates described above, indicated that not being in remission increased the risk of ACS occurrence (Figure 1). Similarly, there was an association between DAS28-ESR at the visit and the risk of ACS during the coming six months.

Figure 1.

Hazard ratio (95% confidence interval) comparing patients not in remission to patients in remission, using several remission definitions. Percentage of visits with an ACS event within 6 months (bottom panel).

Conclusion

Being in RA remission at any visit is associated with a noticeably lower risk of ACS during the following months, suggesting that RA disease activity not only affects CVD risk in the longer term but also in the short term.

Acknowledgements

NordForsk and Foreum partially funded this research project.

Disclosure of Interests

Bénédicte Delcoigne: None declared, Sella Aarrestad Provan: None declared, Eirik kristianslund: None declared, Johan Askling Grant/research support from: AbbVie, AstraZeneca, Bristol Myers Squibb, Eli Lilly, Janssen, Merck, Pfizer, Roche, Samsung Bioepis, Sanofi, and UCB, Lotta Ljung: None declared.

Poor long-term prognosis in patients admitted with strong suspicion of acute myocardial infarction but discharged with another diagnosis

BACKGROUND

Characteristics and prognosis of patients admitted with strong suspicion of myocardial infarction (MI) but discharged without an MI diagnosis are not well-described.

OBJECTIVES

To compare background characteristics and cardiovascular outcomes in patients discharged with or without MI diagnosis.

METHODS

The DETermination of the role of Oxygen in suspected Acute Myocardial Infarction (DETO2X-AMI) trial compared 6629 patients with strong suspicion of MI randomized to oxygen or ambient air. The main composite end-point of this subgroup analysis was the incidence of all-cause death, rehospitalization with MI, heart failure (HF) or stroke during a follow-up of 2.1 years (median; range: 1-3.7 years) irrespective of randomized treatment.

RESULTS

1619 (24%) received a non-MI discharge diagnosis, and 5010 patients (76%) were diagnosed with MI. Groups were similar in age, but non-MI patients were more commonly female and had more comorbidities. At thirty days, the incidence of the composite end-point was 2.8% (45 of 1619) in non-MI patients, compared to 5.0% (250 of 5010) in MI patients with lower incidences in all individual end-points. However, for the long-term follow-up, the incidence of the composite end-point increased in the non-MI patients to 17.7% (286 of 1619) as compared to 16.0% (804 of 5010) in MI patients, mainly driven by a higher incidence of all-cause death, stroke and HF.

CONCLUSIONS

Patients admitted with a strong suspicion of MI but discharged with another diagnosis had more favourable outcomes in the short-term perspective, but from one year onwards, cardiovascular outcomes and death deteriorated to a worse long-term prognosis.

POS0484 THE ASSOCIATION BETWEEN SOCIAL STRESSORS AND DISEASE REMISSION AMONG MEN AND WOMEN WITH EARLY RHEUMATOID ARTHRITIS

Background:

The role of psychosocial conditions on the disease course of rheumatoid arthritis (RA) is getting increased attention. In our previous study, low social support and low decision latitude at work were associated with known modifiable risk factors for RA disease development, such as smoking and low educational level (1). Further, smoking and low educational level have previously been shown to be associated with worse RA disease outcome (2-4). Whether psychosocial characteristics are related to RA disease outcome needs further investigation.

Objectives:

To investigate the relationship between two psychosocial characteristics: low social support and low decision latitude at work, and achievement of remission in patients with RA.

Methods:

At inclusion in the Swedish EIRA study, incident RA cases (N=3724) and controls (N=5937), matched for age, sex and residential area, responded to a questionnaire including questions on social support and decision latitude at work. The answers were recoded into separate scores and the distribution of the scores among controls were used to define the exposures. Low social support and low decision latitude at work, respectively, among patients, were set as the level corresponding to the lowest quartile among controls, and were compared with scores corresponding to the remaining three quartiles.

The outcome, disease activity score 28-joint count (DAS28) remission, defined as DAS28<2.6, was captured through linkage with the Swedish Rheumatology Quality Register (SRQ) with data available from diagnosis for 2693 out of 3700 cases for social support and for 847 out of 1248 cases for decision latitude at work.

Logistic regression was used to evaluate the association between low social support or low decision latitude at work, respectively, and the chance of remission at the time-points 3 months, 12 months and 60 months after inclusion. All results were adjusted for age, sex and residential area and the fully adjusted models were also adjusted for smoking, obesity, physical activity and educational level.

Results:

Low social support (n=655) was associated with a reduced chance for remission at all three time points in the model adjusted for age, sex and residential area; OR 3 months 0.77 (95% CI 0.61-0.97), OR 12 months 0.78 (95% CI 0.64-0.95) OR 60 months 0.77 (95% CI 0.59-0.99). This association was diminished after further adjustment. After stratifying for sex, this association was enhanced in women but inverse among men (Figure 1).

No association between low decision latitude at work (n=166) and chance for remission was observed neither in the analyses stratified for matching variables, nor in the full model. This result was only marginally changed after stratifying for sex (Figure 1).

Conclusion:

Low social support was associated with lower chance of remission in early RA, but the association was not independent of other risk factors for worse outcome (smoking, physical activity, obesity and low educational level).

The interrelationship between social stressors and previously known risk factors for worse outcome highlights the importance of supportive actions at many levels to increase the possibility for the individual to make healthy decisions.

References:

[1]Hedenstierna. et al. Scand J Rheumatol. 2021:1-5.

[2]Saevarsdottir, et al. Ann Rheum Dis. 2011;70(3):469-75.

[3]Saevarsdottir, et al. Arthritis Rheum. 2011;63(1):26-36.

[4]Jiang, et al. Arthritis Res Ther. 2015;17:317.

Figure 1.

Odds ratios for assiciation between social stressors and DAS 28 remission

Acknowledgements:

We want to thank all the participants of the EIRA study and the clinical collaborators for their valuable contribution. We also want to thank the staff for their dedicated work with the data collection.

Disclosure of Interests:

None declared

OP0114 SHORT- AND LONGER-TERM RISKS FOR ACUTE CORONARY SYNDROME IN PATIENTS WITH RHEUMATOID ARTHRITIS STARTING TREATMENT WITH DISEASE-MODIFYING ANTI-RHEUMATIC DRUGS. A COLLABORATIVE OBSERVATIONAL HEAD-TO-HEAD STUDY ACROSS FIVE NORDIC RHEUMATOLOGY REGISTERS

Background:

Rheumatoid Arthritis (RA) is associated with increased cardiovascular co-morbidity including acute coronary syndrome (ACS), partly due to effects of systemic inflammation. Disease-modifying anti-rheumatic drugs (DMARDs) may reduce RA disease activity, but act through several pathways and may themselves have an impact on cardiovascular risks. Whether the risks of ACS associated with biologic (b) and targeted synthetic (ts) DMARDs differ is still unknown.

Objectives:

To assess and compare incidences of ACS during treatment of RA with etanercept (ETA), adalimumab (ADA), infliximab (INF), certolizumab pegol (CTZ), golimumab (GOL), rituximab (RIT), abatacept (ABA), tocilizumab (TCZ), baricitinib (BAR) or tofacitinib (TOF).

Methods:

We defined and pooled treatment cohorts of patients starting any of the above treatments between 2008 and 2017 from clinical rheumatology registers in Denmark (DK), Finland (FI), Norway (NO), and Sweden (SE). One patient could contribute several treatment episodes. Age, sex, co-medication (methotrexate, prednisolone), number of previous b/tsDMARDs, CRP, comorbidities (cardiovascular (including ACS (defined as ICD-10: I20.0, I21.0-4, I21.9) and cerebrovascular disease, thromboembolic events, diabetes, hospitalized infection, cancer, kidney failure, COPD) and associated drugs were extracted and used as adjustment in Cox regression analyses comparing the incidence of ACS between treatments. We used several follow-up lengths (1, 2, and up to 5 years) and two different risk windows (ACS on drug [ending follow-up on treatment discontinuation] and ACS ever since treatment start [disregarding any treatment discontinuation]). We also stratified by age and number of previous b/tsDMARDs.

Results:

We included 40850 treatment courses in 24083 patients (DK 7271, FI 3732, NO 1540, and SE 11540; around 75% women). ETA was the most common treatment (27%) whereas BAR and TOF comprised <1%, and the other DMARDs 6-14% each. The proportions with a history of ACS at treatment start ranged from 1.2% (NO) to 1.8% (DK).

We found 780 incident ACS events during 141 326 person-years (pyrs) in the 5-year follow-up time and “ACS ever since treatment start” risk window, resulting in a crude incidence rate of 5.5 events per 1000 pyrs. No event was recorded for BAR nor TOF, which also had the shortest follow-up. Adjusted hazard ratios (HR) increased slightly with longer follow-up times, but the two risk windows provided similar HRs. For the 5-year follow-up, RIT was associated with an increased risk of ACS compared to ETA (Table), while no association was observed for shorter follow-up times. Stratifying on age did not modify the associations. Separate analyses by number of previous b/tsDMARDs suggested that ABA (HR=1.8, 95% CI 1.0-3.3), INF (HR=2.2, 95% CI 1.0-4.6) and RIT (HR=1.9, 95% CI 1.1-3.4) were associated with increased risks of ACS compared to ETA in the subgroup of patients with two or more previous bDMARDs (Figure), whereas no differences were found among patients starting either drug as 1st/2nd bDMARD.

Table 1.

Comparisons of risks for ACS during a 5-year follow-up since start of bDMARD treatment.

DrugN eventspyrsCrude incidence rate/ 1000 pyrsHR (95% CI)1ETA175359174.9ref.ADA115240934.81.0 (0.8-1.3)CTZ54141583.80.9 (0.6-1.2)GOL4090064.41.1 (0.8-1.5)INF106178036.01.2 (0.9-1.5)ABA70107956.51.1 (0.8-1.4)RIT158166229.51.3 (1.0-1.6)TCZ62128664.80.9 (0.5-1.2)BAR036TOF030

Pyrs: person-years; HR: hazards ratio

1 adjustment: see text.

Conclusion:

In this cohort including ≥ 24,000 patients followed for up to 5 years, the ACS incidence rate was 5.5/1000 pyrs, with RIT showing an increased risk compared to ETA. In clinical practice, the choice of bDMARD does not seem to influence ACS risk in the short term. In the longer term, differences in ACS risk between bDMARDs may reflect channeling to these, or truly differential effects in subpopulations of patients.

Acknowledgements:

Partly funded by Nordforsk and Foreum

Disclosure of Interests:

Bénédicte Delcoigne: None declared, Lotta Ljung: None declared, Sella Aa. Provan Speakers bureau: Novartis, Consultant of: Novartis, Grant/research support from: Boehringer- Ingelheim, Bente Glintborg Grant/research support from: Pfizer, BMS, AbbVie, Kathrine Lederballe Gron Grant/research support from: BMS, Merete L. Hetland Consultant of: Abbvie, Biogen, BMS, Celltrion, Eli Lilly, Janssen Biologics B.V, Lundbeck Fonden, MSD, Pfizer, Roche, Samsung Biopies, Sandoz, Novartis, Grant/research support from: Abbvie, Biogen, BMS, Celltrion, Eli Lilly, Janssen Biologics B.V, Lundbeck Fonden, MSD, Pfizer, Roche, Samsung Biopies, Sandoz, Novartis, Niels Steen Krogh: None declared, Nina Trokovic: None declared, Heikki Relas Speakers bureau: Abbvie, Celgene, Pfizer, Grant/research support from: Abbvie, Celgene, Pfizer, Carl Turesson Speakers bureau: Abbvie, Bristol-Myers Squibb, Medac, Pfizer, Roche, Consultant of: Roche, Brigitte Michelsen Consultant of: Novartis (paid to employer), Grant/research support from: Novartis (paid to employer), Johan Askling Consultant of: Abbvie, Astra-Zeneca, BMS, Eli Lilly, MSD, Pfizer, Roche, Samsung Bioepis, Sanofi, and UCB. These entities have entered into agreements with Karolinska Institutet with JA as principal investigator, mainly in the context of safety monitoring of biologics via the ARTIS national safety monitoring system.

OP0222 THE INCIDENCE OF INTERSTITIAL LUNG DISEASE IN PSORIATIC ARTHRITIS COMPARED TO RHEUMATOID ARTHRITIS. DATA FROM OVER 89 000 BDMARD TREATMENT COURSES DERIVED FROM FIVE NORDIC REGISTERS

Background:

Interstitial lung disease (ILD) is an established extra-articular manifestation of rheumatoid arthritis (RA). Few studies have investigated the prevalence of ILD in patients with psoriatic arthritis (PsA). Methotrexate (MTX) is frequently used in the treatment of both RA and PsA and has been suggested to be a risk factor for the development of ILD. It is of interest to understand the interaction between disease and treatment in the development of ILD.

Objectives:

To compare the incidence of ILD between patient with PsA and RA treated with biologic disease modifying antirheumatic drugs (bDMARDS), with or without MTX as a co-medication.

Methods:

Cohorts of patients with RA and PsA starting bDMARD were identified in Nordic registers (Danish nationwide clinical register for patients with RA (DANBIO), Register on antirheumatic and biological therapy in Finland (ROB-FIN), Icelandic nationwide database of biologic therapy (ICEBIO), Norwegian Antirheumatic Drug Register (NOR-DMARD), and the Swedish Rheumatology Quality Register (SRQ)). Linkages to the National Patient Registers and to the Cause of Death Registers were performed in each country to identify cases of ILD. Each individual patient could contribute several treatment courses. ILD was identified as hospital or death certificate ICD10 codes of ILD (J84.1, J84.8, J84.9, J70.2, J70.3, J70.4, J99.0, J99.1, J99.8) given during the follow-up period which was defined as the treatment course duration, plus a 30-day wash-out period added to the end of treatment course period. MTX co-medication was specified as use of MTX at the start of bDMARD. Incidence rates (IR) for any ILD were calculated per 1000 person years at risk (PYR) for each country. The five cohorts were pooled and incidence rate ratios (IRR) for PsA vs. RA were calculated. Hazard ratios (HR) for any ILD in PsA vs. RA were estimated in Cox regression models adjusted for age, gender and repeated observations, and stratified for the use of MTX co-medication.

Results:

Overall 47 987 individual patients representing 89 239 bDMARD treatment courses and contributing 201 279 PYR were included in the study (Table 1). Methotrexate was reported as comedication in 29 916 (33.5 %) of the treatment courses (PsA vs. RA, 30.4 % vs 34.5 %). 970 cases of ILD were identified during the follow-up period. The risk of ILD was consistently lower in patients with PsA compared to patients with RA in all countries. In models stratified for co-medication the HR for ILD in PsA vs. RA was 0.34 (0.21-0.57) in patients treated with MTX and 0.26 (0.18-0.36) in patients not treated with MTX.

Table 1.

Interstitial lung disease in PsA vs. RA in five Nordic biologic registers

DENMARKFINLANDICELANDNORWAYSWEDENRAPsARAPsARAPsARAPsARAPsANumber of individuals78293386494610916754701590999205966393Number of treatment courses17 07266408634184512808592379142738 27910 824Age baseline (SD)57.3 (13.1)49.0 (12.6)53.8 (13.4)48.8 (11.4)53.9 (14.2)50.1 (13.3)53.8 (13.7)48.7 (12.0)57.1 (13.7)50.6 (12.8)Female n (%)12 963 (76)3929 (59)6571 (76)933 (51)969 (76)551 (65)1815 (77)818 (57)29 635 (77)6162 (57)Number of PYR4023513986217984910451727994556265312033427412ILD-events within PYR2182213287232668028IR pr 1000 PYR5.41.66.11.61.50.77.02.35.71.0IRR PsA vs RA crude0.29 (0.18-0.45)0.27 (0.11-0.55)0.46 (0.05-2.42)0.32(0.11-0.78)0.18 (0.12-0.26)HR PsA vs RA0.31 (0.17-0.56)0.46 (0.22-0.96)0.62 (0.12-3.14)0.19 (0.06-0.54)0.25 (0.17-0.37)

PYR: Patient years at risk, IR: Incidence rates, IRR: Incidence rate ratios, HR: Hazard Ratios

Conclusion:

In these preliminary analyses, the incidence of ILD is lower in bDMARD treated PsA vs. RA patients, irrespective of co-medication with MTX. This indicates that the clinician should consider the rheumatological diagnosis when assessing the risk for future ILD in patients treated with bDMARDs and MTX.

Acknowledgements:

Partly funded by NordForsk and FOREUM

Disclosure of Interests:

Sella Aarrestad Provan Consultant of: Novartis, Grant/research support from: Boehringer-Ingelheim, Brigitte Michelsen: None declared, Lotta Ljung: None declared, Thorarinn Jonmundsson: None declared, Björn Gudbjornsson Speakers bureau: Amgen and Novartis, Daniela Di Giuseppe: None declared, Merete Lund Hetland Speakers bureau: Orion Pharma, Biogen, Pfizer, CellTrion, Merck and Samsung Bioepis, Consultant of: Eli Lilly, Grant/research support from: BMS, MSD, AbbVie, Roche, Novartis, Biogen and Pfizer, Guðrún Björk Reynisdóttir: None declared, Bente Glintborg: None declared, Eirik kristianslund: None declared, Heikki Relas: None declared, Kalle Aaltonen: None declared, Dan Nordström Speakers bureau: Abbvie, BMS, Celgene, Eli Lilly, MSD, Novartis, Pfizer, Roche and UCB., Consultant of: Abbvie, BMS, Celgene, Eli Lilly, MSD, Novartis, Pfizer, Roche and UCB., Tore K. Kvien Speakers bureau: Amgen, Celltrion, Egis, Evapharma, Ewopharma, Hikma, Oktal, Sandoz, Sanofi., Consultant of: AbbVie, Amgren, Biogen, Celltrion, Eli Lilly, Gilead, Mylan, Novartis, Pfizer, Roche, Sandoz, Sanofi., Johan Askling Grant/research support from: Abbvie, BMS, Eli Lilly, Merck, Pfizer, Roche, Samsung Bioepis, and Sanofi

Addition of infliximab compared with addition of sulfasalazine and hydroxychloroquine to methotrexate in patients with early rheumatoid arthritis (Swefot trial): 1-year results of a randomised trial

BACKGROUND

New treatment strategies for early rheumatoid arthritis are evolving rapidly. We aimed to compare addition of conventional disease-modifying antirheumatic drugs (sulfasalazine and hydroxychloroquine) with addition of a tumour necrosis factor antagonist (infliximab) to methotrexate in patients with early rheumatoid arthritis.

METHODS

We undertook a randomised trial in 15 rheumatology units in Sweden. We enrolled patients with early rheumatoid arthritis (symptom duration <1 year) and administered methotrexate (up to 20 mg per week). After 3-4 months, those who had not achieved low disease activity but who could tolerate methotrexate were randomly allocated by computer addition of either sulfasalazine and hydroxychloroquine or infliximab. Primary outcome was achievement of a good response according to European League Against Rheumatism (EULAR) criteria at 12 months. Patients were followed up to 24 months; here, we present findings at 12 months. Analysis was by intention to treat and we used non-responder imputation. The Swefot (Swedish Pharmacotherapy) study is registered in the WHO database at the Karolinska University Hospital, number CT20080004.

FINDINGS

487 patients were initially enrolled. Of 258 who had not achieved low disease activity with methotrexate, 130 were allocated sulfasalazine and hydroxychloroquine and 128 were assigned infliximab. 32 of 130 (25%) patients allocated sulfasalazine and hydroxychloroquine achieved the primary outcome compared with 50 of 128 (39%) assigned infliximab (risk ratio 1.59 [95% CI 1.10-2.30], p=0.0160). Adverse events were balanced fairly well between the two groups and accorded with known adverse events of the drugs used. No deaths occurred in either group.

INTERPRETATION

In patients with early rheumatoid arthritis in whom methotrexate treatment failed, addition of a tumour necrosis factor antagonist to methotrexate monotherapy is clinically superior to addition of conventional disease-modifying antirheumatic drugs.

FUNDING

Swedish Rheumatism Association, Schering-Plough.

Interleukin-1 receptor antagonist is associated with both lipid metabolism and inflammation in rheumatoid arthritis

BACKGROUND

There is a relationship between cardiovascular morbidity, inflammatory activity, and changes in the lipid profile in rheumatoid arthritis (RA), although the mechanisms are not fully elaborated. Recent know-ledge that white adipose tissue (WAT) is a producer of immunologically and metabolically active substances gives another perspective to study.

OBJECTIVE

To evaluate the relationship between interleukin-1 receptor antagonist (IL-1Ra) and variables associated with WAT and inflammation in RA.

METHODS

Anthropometric, inflammatory and metabolic variables were assessed in 23 women with RA and 23 matched controls. Spearman, partial correlation and factor analyses were performed.

RESULTS

Inflammatory markers were increased in patients. In both groups, IL-1Ra correlated with leptin independent of age and BMI. IL-1Ra also correlated with haptoglobin and apolipoprotein (Apo) B in patients and with soluble TNF receptor (sTNFR) 1 in controls. In factor analysis, three latent factors were identified among patients. The first loaded on IL-1Ra, leptin, BMI, ApoB and body fat content (BF%), the second loaded on IL1-Ra and sTNF-receptors and the third showed inverse loadings on ApoA-I together with loadings on ESR, haptoglobin, orosomucoid, BF% and BMI.

CONCLUSION

IL-1Ra was associated with markers of inflammation and with fat-related factors in RA patients, suggesting a dualistic relationship of IL-1Ra in RA. IL-1Ra correlated independently with leptin in both patients and controls, indicating a relationship between inflammation and leptin.

Uptake of porcine rubulavirus (LPMV) by PK-15 cells

BACKGROUND

The porcine virus denominated La Piedad Michoacan Virus (LPMV) is a member of the family Paramyxoviridae and is the cause of a disease in pigs present only in Mexico. The disease is characterized by meningoencephalitis and respiratory distress in young pigs, epididymitis and orchitis in boars, and reproductive failure and abortion in sows.

METHODS

The cytopathology, morphology, and distribution of the hemagglutination neuraminidase (HN) and nucleoprotein (NP) proteins of LPMV were investigated following inoculation into PK-15 cells. The cytopathic effect was characterized by cytoplasmic vacuolation and the formation of syncytia and cytoplasmic inclusion bodies.

RESULTS

In immunofluorescence assays using a monoclonal antibody (MAb) against the HN protein at 5-60 min post-infection (early infection), a diffuse immunofluorescence was observed near the cell membrane and adjacent to the nuclear membrane. At 24 h post-infection (late infection), a dust-like immunofluorescence was observed throughout the cytoplasm. LPMV-infected cells incubated with the MAb against the NP protein showed punctate cytoplasmic fluorescence during the early stages of infection. At the late infection stage, these fluorescent particles became larger and were seen predominantly in the cytoplasm of syncytia. This pattern was also apparent by immunohistochemical labeling and immunogold electron microscopy. The latter technique revealed that HN protein was diffusely distributed throughout the cytoplasm. When using the MAb against the NP protein, nucleocapsid organization was the most prominent feature and resulted in the formation of cytoplasmic inclusion bodies visible by light and electron microscopy. Immunogold labeling of purified nucleocapsids was shown by electron microscopy. Virus particles and nucleocapsids were morphologically similar to members of the Paramyxoviridae family.

CONCLUSIONS

The morphologic characteristics of the virions and the distribution patterns of the HN and NP proteins in PK-15 infected cells indicate that the mechanisms of LPMV replication are generally similar to those of the members of the Paramyxoviridae family.

Electron microscopy and X-ray microanalyses of uterine epithelium from lead-injected mice in an experimental delay of implantation

Mice in experimental delay of implantation were injected intravenously with 75 micrograms.g-1 body weight of lead chloride, corresponding to a dose of lead of about 56 micrograms.g-1 body weight. Delay of implantation was obtained by ovariectomy 3 days after mating followed by a depot dose of progesterone every fifth day. Electron microscopy showed that the uterine lumen, which was closed in control mice, was opened in lead-injected mice. This morphology suggested that lead caused an increase in uterine secretion. X-ray microanalysis of pyroantimonate precipitates in the uterine epithelium of injected mice demonstrated lead in the precipitates, suggesting that lead could have a direct effect on the function of the uterine epithelium and that lead also could be secreted into the uterine lumen and affect the blastocysts.

X-ray micro analyses of cations (Na, K, Ca) and anions (S, P, Cl) in uterine secretions during blastocyst implantation in the rat

The mass changes of sodium, potassium, and calcium ions in the rat uterine secretion at blastocyst delay, activation, and attachment have been estimated with X-ray microanalyses of samples of uterine secretions absorbed by small Sephadex beads. A quantification of the ions was attempted by using a standardized coat of gold on the beads as a reference element for normalization of the ion peaks and by fitting the normalized values into corresponding linear regression equations obtained from measurements of step-wise dilutions of a control rat serum. The concentrations of sodium observed at delay, activation, and attachment were 117, 201, and 203 mEq/l, respectively, and those of potassium were 6, 18, and 19 mEq/l, respectively. Calcium values were about 2 mEq/l and decreased at attachment. Among the anions, only the chloride concentration increased at activation and attachment.