Experience With Etanercept, Tocilizumab and Interleukin-1 Inhibitors in Systemic Onset Juvenile Idiopathic Arthritis Patients From the BIKER Registry

Gerd Horneff; Anna Carina Schulz; Jens Klotsche; Anton Hospach; Kirsten Minden; Ivan Foeldvari; Ralf Trauzeddel; Gerd Ganser; Frank Weller-Heinemann; Johannes Perter Haas


Arthritis Res Ther. 2017;19(256) 

In This Article


The German BIKeR register has been approved by the local ethics committee of the Aerztekammer Nordrhein, Düsseldorf, Germany, reference number 2/2015/7441.[13] Written consent was obtained from all patients and their parents and the data were collected pseudonymised. All data were collected prospectively including data on all patients in the German JIA-BIKeR registry, who had sJIA confirmed according to the International League of Associations of Rheumatology (ILAR) criteria, starting treatment with a biologic agent between 2000 and 2015. Patients were only included in analyses if they had assessments at baseline and at least at one follow up visit. Follow up visit reports were collected after 3 and 6 months and 6 monthly thereafter. Reminders were issued about missing and outstanding reports.[14] Patients' characteristics included gender, age, diagnosis, disease duration, previous treatments and initial concomitant treatment and comorbidities. Clinical data included disease activity parameters, number of tender, swollen, active joints and number of joints with limitation of motion, physician's assessment of global disease activity (visual analogue scale, VAS), patient's/parent's assessment of overall wellbeing and pain (both with the VAS), ESR, CRP and functional assessment according to the Childhood Health Assessment Questionnaire (CHAQ) disability index.[15,16]

Effectiveness Assessment

Effectiveness was determined using the JIA-American College of Rheumatology (ACR) response criteria and the Juvenile Disease Activity Score 10 (JADAS-10). The definition used for minimal disease activity was a JADAS-10 ≤ 3.8 and for remission on drug it was a JADAS-10 ≤ 1.0.[17,18] ACR preliminary criteria for remission and inactive disease (Wallace et al.,[19]) were used including: (i) the lowest value of the physician's judgement on global disease activity was 0 on a 100-mm visual analogue scale; (ii) ESR up to 20 mm/h; (iii) CRP up to 6 mg/l; (iv) morning stiffness lasting up to 15 min and (v) the absence of systemic manifestations (fever, rash, pericarditis, hepatomegaly, splenomegaly or lymph node swelling) to fulfil the definition of inactive disease/remission. The JIA-ACR 30/50/70/90 response described by the ACR is defined as 30%/50%/70%/90% or greater improvement in three or more of the six JIA core response variables without greater than 30% worsening in more than one of the remaining core response variables compared with baseline.[20] For this analysis, a 24-month treatment period was chosen as a meaningful time for primary judgement of early effectiveness and ongoing effectiveness and as a compromise with lower patient numbers with prolonged observation due to the nature of the registry analysis. In addition, the last available clinical status was used. The JADAS-10 is an established composite score for assessing disease activity including the four dimensions of physician's global assessment of disease activity, patient's global assessment of overall wellbeing, the active joint count and normalized ESR.

Effectiveness was analysed following the intention-to-treat principle and patients who discontinued due to inefficacy or intolerance were classified as non-responders. This study design is intending to avoid misleading artefacts.

Safety Assessment

Safety assessments included the collection of data on adverse events and serious adverse events. Adverse events were collected throughout the observation and specially requested on each routine follow up visit. In the case of adverse events the investigator assessed and recorded the adverse event in detail on the adverse event form including the date and time of onset, description, severity, time course, duration and outcome, relationship of the adverse event to the biological agent and alternative aetiology for events not considered "probably related" to the drugs that the patient received. Detailed data on adverse events of special interest were collected using special data forms.


The BiKeR data are entered into a Microsoft ACCESS™ (Microsoft Corp., Redmond WA, USA) database. Descriptive statistics were used to report the sociodemographic and clinical characteristics of the patients. This includes the absolute and relative frequencies of categorical data, and either the mean value with standard deviation or median with interquartile range for continuously distributed variables as appropriate. Comparisons of sociodemographic and clinical characteristics at baseline were conducted using the t test and chi square test. The patients cohort treated with etanercept (ETA) had been observed prospectively as well, although ETA had been approved earlier and was the only available biologic agent for treatment of patients with JIA for several years. Since more than 80% of the patients within this study had started treatment before 2008 the ETA cohort was used as a comparator group. In contrast, 74% started therapy after 2008 in the TOC and IL-1i cohort. No adjustments have been carried out for comparison of the ETA and the TOC/IL-1i cohorts, because of the remarkable differences between the cohorts in baseline characteristics, e.g. the concomitant conventional synthetic disease-modifying antirheumatic drug (csDMARD) and steroid use.

The propensity score was estimated for the comparison of effectiveness parameters between the TOC and IL-1 inhibitor cohorts. The likelihood of being treated by either TOC or an IL-1i was modelled by logistic regression including the predictor variables of year of treatment start, number of biologics in the history, concomitant csDMARD and steroid use, disease activity at treatment start, presence of systemic symptoms and disease duration at treatment start. All comparisons between the cohorts were adjusted by the propensity score in effectiveness analyses. Generalized linear mixed models were used to analyse the change in effectiveness parameters. Linear mixed models have the benefit that changes in parameters may be analyzed over time, (i) while using all study visits, (ii) to account for the dependence of measurements over time within the same patient and (iii) to model possible heterogeneity in the response between patients. The means of the linear mixed model were to be used for post-hoc tests to evaluate the change in each effectiveness parameter at a specified follow up. Comparison of adverse events and adverse event rates were calculated using the chi square and Wald test. Statistical analyses were conducted using STATA 12.1.