Explore the vast range of titles available.

Missing data is very common problem in epidemiological studies, and this problem can affect both the accuracy and the precision of any estimates generated from the study data. The magnitude of the problem will depend to some extent on whether the missing data refer to outcomes, exposures or confounders. There are a number of simple methods commonly used for analysing incomplete data, but which rely on unrealistic assumptions about the mechanism causing the data to be missing. Multiple imputation is a more sophisticated method which can produce precise, reliable estimates in a wide range of circumstances, but still relies on unverifiable assumptions about the missing data. I will outline the problems caused by missing data, and explain when simpler methods of analysing incomplete data would be appropriate. I will also explain the idea behind multiple imputation, and the situations in which it can be used effectively. I will illustrate the differences between conventional analyses and multiple imputations analyses, and show how to interpret the results of multiple imputation. My aim is that you will leave able to confidently decide whether it would be appropriate to use multiple imputation to deal with missing data in a given situation, and to understand the statistical output it generates. Disclosure of Interest None declared DOI 10.1136/annrheumdis-2014-eular.6211

Objective Braces used to treat (PF) osteoarthritis (OA) may reduce contact stress across the PF joint. We hypothesised that in PF OA, braces would decrease knee pain and shrink PF bone marrow lesions (BMLs). Methods Eligible subjects had painful PF OA. Subjects were randomly allocated to brace or no brace for 6 weeks. Knee MRIs were acquired at baseline and 6 weeks. We measured BMLs on post-contrast fat suppressed sagittal and proton density weighted axial images. The primary symptom outcome was change in pain at 6 weeks during a preselected painful activity, and the primary structural outcome was BML volume change in the PF joint. Analyses used multiple linear regression. Results We randomised 126 subjects aged 40–70 years (mean age 55.5  years; 72 females (57.1%)). Mean nominated visual analogue scale (0–10 cm) pain score at baseline was 6.5 cm. 94 knees (75%) had PF BMLs at baseline. Subjects wore the brace for a mean of 7.4 h/day. 6 subjects withdrew during the trial. After accounting for baseline values, the brace group had lower knee pain than the control group at 6 weeks (difference between groups −1.3 cm, 95% CI −2.0 to −0.7; p<0.001) and reduced PF BML volume (difference −490.6 mm3, 95% CI −929.5 to −51.7; p=0.03) but not tibiofemoral volume (difference −53.9 mm3, 95% CI −625.9 to 518.2; p=0.85). Conclusions A PF brace reduces BML volume in the targeted compartment of the knee, and relieves knee pain. Trial registration number UK. ISRCTN50380458.

Background The risk of tuberculosis (TB) in patients with rheumatoid arthritis (RA) is thought to be increased following anti-tumour necrosis factor (anti-TNF) therapy, with a proposed differential risk between the anti-TNF drugs etanercept (ETA), infliximab (INF) and adalimumab (ADA). Objective To compare directly the risk between drugs, to explore time to event, site of infection and the role of ethnicity. Methods Data from the British Society for Rheumatology Biologics Register (BSRBR), a national prospective observational study, were used to compare TB rates in 10 712 anti-TNF treated patients (3913 ETA, 3295 INF, 3504 ADA) and 3232 patients with active RA treated with traditional disease-modifying antirheumatic drugs. Results To April 2008, 40 cases of TB were reported, all in the anti-TNF cohort. The rate of TB was higher for the monoclonal antibodies ADA (144 events/100 000 person-years) and INF (136/100 000 person-years) than for ETA (39/100 000 person-years). After adjustment, the incidence rate ratio compared with ETA-treated patients was 3.1 (95% CI 1.0 to 9.5) for INF and 4.2 (1.4 to 12.4) for ADA. The median time to event was lowest for INF (5.5 months) compared with ETA (13.4 months) and ADA (18.5 months). 13/40 cases occurred after stopping treatment. 25/40 (62%) cases were extrapulmonary, of which 11 were disseminated. Patients of non-white ethnicity had a sixfold increased risk of TB compared with white patients treated with anti-TNF therapy. Conclusion The rate of TB in patients with RA treated with anti-TNF therapy was three- to fourfold higher in patients receiving INF and ADA than in those receiving ETA.

Missing data is very common in observational studies, and the data may be missing for a number of reasons. Missing data will cause estimates to be less precise than they would otherwise have been. They may also result in biased estimates, depending on the mechanism causing them to be missing and the methods used to analyse the observed data. A second common problem in observational studies of the effect of a particular treatment or exposure is confounding by indication: the decision to treat or not treat may depend on patient characteristics which affect the outcome. Therefore, the difference in outcome between treated and untreated subjects is caused partly by the treatment and partly by this difference in patient characteristics, and hence is a biased measure of the effect of treatment. The true effect of treatment is the difference between the outcome if the patient receives treatment and the outcome if they do not receive treatment. Since only one of these outcomes can ever be observed, estimating the effect of treatment can also thought of as a missing data problem. I will present methods which are commonly used to overcome the problems of missing data and confounding by indication, concentrating on the conditions under which they enable the effect of treatment to be estimated without bias.I will also show the parallels between methods that give reliable inference when there is missing data, and methods that give reliable inference when there is confounding by indication. Disclosure of Interest None Declared

Background Muscle atrophy, muscle weakness and localised pain are commonly reported following musculoskeletal injury (MSKI). To mitigate this risk and prepare individuals to return to sport or physically demanding occupations, resistance training (RT) is considered a vital component of rehabilitation. However, to elicit adaptations in muscle strength, exercise guidelines recommend lifting loads ≥ 70% of an individual’s one repetition maximum (1-RM). Unfortunately, individuals with persistent knee pain are often unable to tolerate such high loads and this may negatively impact the duration and extent of their recovery. Low load blood flow restriction (LL-BFR) is an alternative RT technique that has demonstrated improvements in muscle strength, hypertrophy, and pain in the absence of high mechanical loading. However, the effectiveness of high-frequency LL-BFR in a residential rehabilitation environment remains unclear. This study will compare the efficacy of high frequency LL-BFR to ‘conventional’ heavier load resistance training (HL-RT) on measures of physical function and pain in adults with persistent knee pain. Methods This is a multicentre randomised controlled trial (RCT) of 150 UK service personnel (aged 18–55) admitted for a 3-week residential rehabilitation course with persistent knee pain. Participants will be randomised to receive: a) LL-BFR delivered twice daily at 20% 1-RM or b) HL-RT three-times per week at 70% 1-RM. Outcomes will be recorded at baseline (T1), course discharge (T2) and at three-months following course (T3). The primary outcome will be the lower extremity functional scale (LEFS) at T2. Secondary outcomes will include patient reported perceptions of pain, physical and occupational function and objective measures of muscle strength and neuromuscular performance. Additional biomechanical and physiological mechanisms underpinning both RT interventions will also be investigated as part of a nested mechanistic study. Discussion LL-BFR is a rehabilitation modality that has the potential to induce positive clinical adaptations in the absence of high mechanical loads and therefore could be considered a treatment option for patients suffering significant functional deficits who are unable to tolerate heavy load RT. Consequently, results from this study will have a direct clinical application to healthcare service providers and patients involved in the rehabilitation of physically active adults suffering MSKI. Trial registration ClinicalTrials.org reference number, NCT05719922

BackgroundMusculoskeletal injury (MSKI) is the leading cause of medical downgrading and discharge within the UK military, with lower limb MSKI having the greatest incidence, negatively impacting operational readiness. Pain is a primary limiting factor to rehabilitation progress following MSKI. Heavy-load resistance training (RT; ie, loads >70% 1-repetition maximum) is traditionally used but may be contraindicated due to pain, potentially prolonging recovery and leading to failure of essential physical employment standards for UK military personnel. Low-load RT with blood flow restriction (BFR) can promote favourable morphological and physiological adaption, as well as elicit hypoalgesia in healthy and clinical populations (eg, post-operative), and has proven a viable option in military rehabilitation settings. The acceptability and tolerance of higher relative BFR pressures in persistent pain populations are unknown due to the complexity of presentation and the perception of discomfort experienced during BFR exercise. Greater relative pressures (ie, 80% limb occlusion pressure (LOP)) elicit a greater hypoalgesic response in pain-free individuals, but greater perceived discomfort which may not be tolerated in persistent pain populations. However, lower relative pressure (ie, 40% LOP) has elicited hypoalgesia in pain-free individuals, which therefore may be more clinically acceptable and tolerated in persistent pain populations. The primary aim of both randomised controlled trials (RCT) is to investigate the efficacy and acceptability of using high-frequency, low-load BFR-RT in UK military personnel with lower limb MSKI where persistent pain is the primary limiting factor for progression.MethodologyThe presented protocol is a two-phase RCT based within a military rehabilitation setting. Phase One is a 1-week RCT to determine the most efficacious and acceptable BFR-RT protocol (7× BFR-RT sessions over 5 days at 40% or 80% LOP; n=28). Phase Two is a 3-week RCT comparing the most clinically acceptable BFR pressure, determined by Phase One (21× BFR-RT sessions over 15 days; n=26) to usual care within UK Defence Rehabilitation residential rehabilitation practices. Outcomes will be recorded at baseline, daily and following completion of the intervention. The primary outcome will be the brief pain inventory. Secondary outcomes include blood biomarkers for inflammation and pain (Phase Two only), injury-specific outcome measures, lower extremity function scale, objective measures of muscle strength and neuromuscular performance, and pressure pain threshold testing.Ethics and disseminationThe study is approved by the Ministry of Defence Research Ethics Committee (2318/MODREC/24) and Northumbria University. All study findings will be published in scientific peer-reviewed journals and presented at relevant scientific conferences.Trial registration numberRegistered with Clinical Trials. The registration numbers are as follows: NCT06621914 (Phase One) and NCT06621953 (Phase Two).

Factorisation (also known as “factor separation”) is widely used in the analysis of numerical simulations. It allows changes in properties of a system to be attributed to changes in multiple variables associated with that system. There are many possible factorisation methods; here we discuss three previously proposed factorisations that have been applied in the field of climate modelling: the linear factorisation, the factorisation, and the factorisation. We show that, when more than two variables are being considered, none of these three methods possess all four properties of “uniqueness”, “symmetry”, “completeness”, and “purity”. Here, we extend each of these factorisations so that they do possess these properties for any number of variables, resulting in three factorisations – the “linear-sum” factorisation, the “shared-interaction” factorisation, and the “scaled-residual” factorisation. We show that the linear-sum factorisation and the shared-interaction factorisation reduce to be identical in the case of four or fewer variables, and we conjecture that this holds for any number of variables. We present the results of the factorisations in the context of three past studies that used the previously proposed factorisations.

Background:Janus kinase inhibitors (JAKi) are effective therapies for the management of rheumatoid arthritis (RA). However, concerns about their safety in certain “at-risk” populations have resulted in risk minimisation measures being introduced by the European Medicines Agency (EMA) in January 2023, which limits prescription of JAKi to certain patient subgroups unless no suitable alternative is available. The potential impact of these measures on current and future JAKi use in the UK is unknown.Objectives:Using a large national cohort study of patients who have already started their first JAKi, this analysis aimed to describe the proportion of patients who met one or more of these criteria and therefore may be impacted by these changes.Methods:All RA patients who started their first-ever JAK inhibitor (tofacitinib, baricitinib, upadacitinib, or filgotinib) therapy in the BSRBR-RA before 31st May 2022 were included. This cutoff date preceded the EMA warning [1]. Patients with missing data in any EMA criteria related variables were excluded from the analysis. Descriptive statistics were used to present patients’ characteristics at the start of their first-ever JAKi. This analysis defined four at-risk criteria assessed at the start of treatment: 1) aged ≥65 years, 2) increased risk of major cardiovascular problems (history of ischaemic heart disease, stroke, dyslipidaemia, hypertension or diabetes), 3) current or past smokers, 4) increased risk of cancer (defined as prior cancer). Among those who met the criteria, the previous number of distinct biological disease-modifying antirheumatic drugs (bDMARDs) classes of therapy was described (i.e. TNF inhibitor, IL-6 inhibitor).Results:A total of 1362 patients starting their first JAKi were included, and baricitinib was most common (76%), followed by tofacitinib (16%; Table 1). The cohort was 78% female, with an average age of 60 years. Among them, 28% of patients had received ≥3 prior classes of bDMARDs. Eighty percent of patients (1090) met ≥1 EMA risk criterion, with most common individual criterion as current/past smoker (54%) or cardiovascular risk factors (44%). Of those who met ≥1 risk criteria, 645 (59%) met more than one EMA risk criterion. Additionally, 29% had already received ≥3 prior bDMARDs classes, and 52% of the patients previously receiving ≥2 prior bDMARDs classes.Conclusion:We observed that a very high proportion of patients four-in-five commencing JAKi in the UK prior to the EMA advisory would have fallen into the ‘high-risk’ category for whom prescribing would only be advised if no suitable alternatives were available. Whilst almost a third of those had been treated with three or more prior bDMARDs classes, it is likely that for a substantial proportion, suitable alternatives would have existed. These data highlight the enormous impact of the recent licensing updates for the JAKi class.REFERENCES:[1] EMA recommends measures to minimise risk of serious side effects with Janus kinase inhibitors for chronic inflammatory disorders 2022 [updated 28/10/2022. Available from: https://www.ema.europa.eu/en/news/ema-recommends-measures-minimise-risk-serious-side-effects-janus-kinase-inhibitors-chronic accessed 28/11/2023 2023.[2] EMA confirms measures to minimise risk of serious side effects with Janus kinase inhibitors for chronic inflammatory disorders, EMA’s human medicines committee (CHMP) 2023 [updated 23 January 2023. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/janus-kinase-inhibitors-jaki accessed 15 June 2023.Acknowledgements:NIL.Disclosure of Interests:Zixing Tian: None declared, Lianne Kearsley-Fleet: None declared, James Galloway Abbvie, Galapagos, Janssen, Lilly, Novartis, Pfizer, UCB, and Vifor, Abbvie, Galapagos, GSK, Janssen and Pfizer, Kath Watson: None declared, Mark Lunt: None declared, BSRBR-RA Contributors Group: None declared, Kimme Hyrich Abbvie, Pfizer, BMS.

Background: There is increased cardiovascular disease mortality in rheumatoid arthritis. This may reflect an increased prevalence of cardiovascular disease or an increased case fatality in patients with rheumatoid arthritis. Objectives: To examine whether rheumatoid patients with disease onset in the 1980s–1990s have increased mortality, and to compare cardiovascular admission rates in rheumatoid patients with those of the general population. Methods: An inception cohort of 1010 rheumatoid patients attending Stockport rheumatology clinics between 1981 and 1996 was followed up to December 2002 through the Office for National Statistics. Standardised mortality ratios (SMR) were calculated for all-cause and cause specific mortality, using the population of Stockport as reference. Cardiovascular disease admission rates were ascertained for a subgroup of patients using national hospital episode statistics; standardised cardiovascular disease admission rates (SAR) and SMRs were calculated for this subgroup. Results: 470 patients (48%) died during a median follow up of 11.4 years. All-cause mortality was increased in men (SMR = 1.45 (95% confidence interval, 1.22 to 1.71)) and women (SMR = 1.84 (1.64 to 2.05)), as was cardiovascular disease mortality in men (SMR = 1.36 (1.04 to 1.75) and women (SMR = 1.93 (1.65 to 2.26)). No difference in cardiovascular disease admission rates was observed in men (SAR 1.20 (0.89 to 1.58) or women (SAR = 1.10 (0.88 to 1.36)), despite excess cardiovascular disease mortality in this subgroup. Conclusions: Patients with rheumatoid arthritis have reduced life expectancy and excess cardiovascular disease mortality. Nevertheless, standardised admission rates for cardiovascular disease were not raised. This suggests either that cardiovascular disease in rheumatoid arthritis has a higher case fatality than in the general population or that it often goes unrecognised before the fatal event.

BackgroundInterleukin 6 inhibitors (IL-6i) can increase LDL cholesterol levels, which raises concerns about the risk of myocardial infarction (MI) in patients with rheumatoid arthritis (RA) receiving this therapy.ObjectivesThis study aims to compare the risk of MI between people with RA in the UK clinical setting receiving IL-6i or tumour necrosis factor inhibitors (TNFi) overall or by line of therapy (LoT).MethodsPatients with RA registered between 01/10/2001 and 30/05/2022 with BSRBR-RA starting IL-6i or TNFi treatments were included. Occurrence of MI was identified from clinical follow-up forms and through cause of death reported by the national UK death register. Only those MIs occurring whilst patient was actively receiving drug were included. The risk of MI in patients receiving IL6i compared to TNFi was compared using Cox regression, adjusted for baseline co-variates using propensity scores (PS, see Table 1). Follow-up commenced at the start of the drug of interest and patients were censored at occurrence of MI, death, discontinuation of therapy or last follow-up visit, whichever came first. Multiple imputation was used for missing data. To account for known differences in LoT use of TNFi and IL6i (with IL6i more likely as a later line bDMARD), overall analyses adjusted for LoT in PS and secondary analyses by LoT were conducted. Direct switches between originator to biosimilars were considered the same treatment.ResultsA total of 30,022 IL6i or TNFi LoTs in 20,898 patients were included (3,278 IL-6i; 26,744 TNFi), representing 119,797 person-years of exposure. Compared to patients receiving TNFi, patients starting IL-6i treatment were older, had longer disease duration, less likely to use methotrexate and steroids, and had more comorbidities. During follow-up, 409 MIs occurred, 30 on IL-6i and 379 on TNFi. After PS adjustment, the risk of MI was not significantly different between the two treatment overall (HR 0.87, 95% CI 0.56-1.37) or when stratified by LoT (Table 1).ConclusionThis study could not identify any difference in risk of MI between IL-6i and TNFi treatment after patient characteristics and LoT were considered.TableMI in RA patients treated with IL-6i or TNFinPerson-yearsEventsIR per 1000 person year (95% CI)Crude HRPS Adjusted HR (95% CI)(95% CI)All lines of therapy TNFi26,744110,9813793.41 (3.09, 3.78)RefRef IL-6i3,2788,816303.40 (2.38, 4.87)0.93 (0.64, 1.35)0.87 (0.56, 1.37)First line TNFi16,38371,6542443.41 (3.00, 3.86)RefRef IL-6i3521,05743.78 (1.42, 10.08)1.04 (0.39, 2.80)1.26 (0.46, 3.41)Second line TNFi766631,4731023.24 (2.67, 3.94)RefRef IL-6i7491,89942.11 (0.79, 5.61)0.59 (0.22, 1.61)0.64 (0.23, 1.76)Third line TNFi1,7185,652264.60 (3.13, 6.76)RefRef IL-6i1,0592,838113.88 (2.15, 7.00)0.79 (0.38, 1.60)0.80 (0.37, 1.70)Fourth line and above TNFi9772,20373.18 (1.51, 6.67)RefRef IL-6i1,1183,023113.64 (2.92, 6.57)1.16 (0.45, 3.01)0.90 (0.33, 2.47)*Baseline variables used for propensity score adjustment: age, gender, time between BSRBR-RA registration and drug start, DAS28, disease duration, RF status, smoking history, BMI, current methotrexate, total number of prior csDMARDs, line of therapy (only in analyses of all LoTs), prior use of TNFi (not in analyses of first LoT only), prior use of IL6i (not in analyses of first or second LoT only), current use of steroids, ever use of antiplatelet or anticoagulant drugs; history of: hypertension, dyslipidaemia, diabetes, lung disease, renal disease, depression, cancer, venous thromboembolism, ischemic heart disease (myocardial infarction and angina), strokeReference[1]Atzeni, F., Rodríguez-Carrio, J., Popa, C. D., Nurmohamed, M. T., Szűcs, G. & Szekanecz, Z. (2021). ‘Cardiovascular effects of approved drugs for rheumatoid arthritis’, Nat Rev Rheumatol.Acknowledgements:NIL.Disclosure of InterestsZixing Tian: None declared, Lianne Kearsley-Fleet: None declared, Kim Lauper: None declared, Sally Haughton: None declared, Kath Watson: None declared, Mark Lunt: None declared, John Mclaughlin: None declared, Arpana Verma: None declared, Kimme Hyrich Consultant of: Honoraria from Abbvie, Grant/research support from: Grant income from Pfizer and BMS.