IL-6 blockade in systemic juvenile idiopathic arthritis – achievement of inactive disease and remission (data from the German AID-registry)

The AID-registry is part of the AID-Net (Network for autoinflammatory diseases), a research initiative funded by the German Federal Ministry of Education and Research. Patient data are documented following a pseudonymisation procedure in an online registry (https://​aid-register.​de); additionally, patient material is collected and stored in a biomaterial bank for serum and plasma [15].

From 2009 to 2014, more than 200 patients with new onset or already established diagnosis of sJIA, which was confirmed by pediatric rheumatologist, were included in the AID-registry. Demographic information, clinical data, and blood samples (serum and EDTA blood) for biomarkers and genetic analysis are collected at study enrollment and follow-up visits. Generally one visit per quarter was scheduled for each patient accounting for an average of 4 visits per year in all sJIA patients. A total of 46/200 (23%) sJIA children were treated with TCZ and longitudinally documented in follow-up. Patients treated with TCZ provided follow-up visits at week 5 and week 12 after treatment start. These patients were available for analysis of the clinical response at week 12 after start of TCZ treatment. The clinical response rate under TCZ at week 12 measured the effectiveness of TCZ and was defined as follows: no symptoms and normal inflammation parameters at that time. Remission after 12 months was evaluated based on Wallace criteria, and therefore we differentiate between active disease (AD), inactive disease (ID) and clinical remission on medication (CRM). Wallace et al. defined ID by meeting the following criteria: no active arthritis, no fever, no exanthema, no serositis, no splenomegaly, no lymphadenopathy, no active uveitis, normal ESR and CRP, no disease activity in physician’s report. CRM was defined as ID for at least 6 months. Active disease (AD) described a visit with symptoms (see above) and increased inflammatory parameters [16]. CRP, ESR (after 1 and 2 h) and leukocytes were measured and the number of patients with normal levels was determined. Different clinical phenotypes were defined: monocyclic (MC) means a persistent flare, after one episode the disease is inactive; polycyclic (PC) are recurrent flares (active disease changed with inactive disease); polyarticular (PA) are flares with arthritis > 4 joints. The concomitant medication before and during TCZ application was analyzed in detail.

We included sJIA patients who fulfilled ILAR classification criteria [17], but also other definitions for sJIA confirming diagnosis by treating pediatric rheumatologists [3, 18‐20]. For study inclusion, a minimum number of 2 visits per year were necessary for each patient. Patients fulfilling these inclusion criteria were included from the following AID-Net centers: Garmisch-Partenkirchen (n = 23), Essen (n = 3), Heidelberg (n = 3), St. Augustin (n = 3), Aachen, Berlin, Bochum, Dresden, Duesseldorf, Hamburg, Krefeld, Landshut and Muenster (each n = 1–2).

Descriptive analyses included mean and range for continuously distributed parameters as well as absolute and relative frequencies for categorical data. The number of adverse events was reported by 100 years under TCZ treatment. The confidence interval for the rate per 100 treatment years was estimated by exact Poisson intervals. The likelihood to achieve a clinical response under TCZ at week 12 was analyzed by logistic regression analyses. Additionally, clinical responses at various visits (response at < 5 weeks, 6–12 weeks, 13–24 weeks, no response at > 24 weeks) were correlated with disease activity at the last available visit in 2014. Logistic regression analyses were conducted in STATA 12.1 (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP).

The AID-registry has been approved by the ethics committees and the data protection responsibles at the University of Duisburg-Essen and Muenster, as well as the Medical Association Nordrhein in Duesseldorf. Parents, children between 9 and 13 years of age and young patients aged ≥14 years provided informed consent.

Forty-six children (19 m, 27 f) with a median age of 11 (range 2–21) years at their last documented visit (up to 04/2014) were included in the analyses. Age of first symptoms was 4 (range 0.6–16) years. Altogether, 680 visits were analyzed. Therapy with TCZ was started at the age of 9 (range 1–18) years. Duration from diagnosis to TCZ start was 24.5 (1–188) months. 28/46 (61%) had a Caucasian origin (Germany, Austria, Poland, Italy, Croatia), 4/46 (9%) were from Turkey and 6/46 (13%) from other East European (Bulgaria, Albania) and Asian countries and 8/46 (17%) of unknown or mixed ethnicity. 34/46 (74%) patients fulfilled ILAR classification criteria, 12/46 (26%) patients did not fulfill ILAR criteria but other definitions for sJIA and diagnosis was confirmed by pediatric rheumatologists [3, 18‐20].

Symptoms at diagnosis of sJIA and at the start of TCZ treatment did not differ significantly. Fever, lymphadenopathy, abdominal pain, hepatomegaly, nausea and vomiting were seen less often at TCZ start, whereas morning stiffness, myalgia, diarrhea, arthralgia and arthritis were observed more frequently. Before TCZ CRP was increased (90 (range 0.1–320) mg/l). Laboratory parameters (CRP, ESR, leukocytes) before diagnosis and under IL-6 inhibition revealed that monocyclic (MC) courses of sJIA, before initiating therapy with TCZ, showed higher CRP and ESR levels than polycyclic (PC) and polyarticular (PA) courses. After TCZ initiation CRP and ESR levels declined irrespective of the disease course. In fact, 27/46 (59%) patients receiving TCZ showed no measurable CRP during therapy or during a clinically diagnosed flare. 10/46 (22%) showed no measurable CRP within the first 4 weeks after starting TCZ.

The median time under TCZ exposure was 14.9 (range 1–48) months for the 46 sJIA patients. In total, 18 (39.1%) patients achieved an inactive disease and 14 (30.4%) a state of remission on medication according to the Wallace criteria. Therefore, a favourable outcome under TCZ treatment was reported for 32/46 (69.6%) of the patients in the registry. Remission on medication was additionally analyzed in a subgroup of patients who were treated with TCZ for at least 12 months. 24/46 sJIA patients (52%) who received TCZ for a median of 23 (range 12–48) months were available for this analysis. The state of remission on medication was achieved for 13/24 (54.2%) sJIA patients, and additional 5/24 (20.8%) were in a state of inactive disease, amounting 18/24 (75%) patients with favourable outcome.

Additionally, clinical response rates within the first 12 weeks of treatment have been estimated in all 46 patients. 16/46 (35%) patients showed an effectiveness of TCZ with no clinical manifestation and normalized inflammation parameters after 12 weeks of treatment. Comparison of initial response (at < 5 weeks, 6–12 weeks, 13–24 weeks, no response at > 24 weeks) and disease activity at last visit suggest that a rapid response to TCZ seems to be related to long term inactivity of sJIA. 7/17 (41%) patients showing inactive disease at the last visit in 2014 had a response to TCZ within 5 weeks. Active disease in the current visit in 2014 was mostly seen in patients who did not show any response within an observation time of > 24 weeks (8/12, 67%).

Forty-six sJIA children were categorized into monocyclic (MC) 12/46 (26%), polycyclic (PC) 16/46 (35%) and polyarticular (PA) 18/46 (39%) disease courses. Comparison of these courses revealed significant differences in the outcome; PC showed the highest clinical response rate (81%, Odds ratio (OR) = 7.0, 95%CI: 1.8–27.2, p = 0.005) followed by the MC courses (59%, OR = 2.9, 95%CI: 1.1–8.6, p = 0.048) compared to the worst outcome for patients with PA courses (29%).

Adverse events (AE) were analyzed for 46 patients contributing to 63 TCZ exposure years. Fourteen adverse events (22.2 AE/100 TCZ exposure years, 95%CI: 12.6–33.7) were reported in 11/46 (24%) patients including leukopenia, infections and elevated transaminases (at least two-fold increased), 2 serious adverse events (SAEs; 3.2 SAE/100 TCZ exposure years, 95%CI: 0.4–11.5) were one Hodgkin’s lymphoma [21] and one gut perforation [22]. No cases of MAS, amyloidosis, or death under TCZ were reported. In 2/46 MAS were diagnosed before TCZ in the medical history. In 5/46 (11%) adverse events were the reason for stopping TCZ administration: neutropenia n = 3, SAEs n = 2.

Only 4/46 (9%) patients (three with PC, one with MC disease course) had to be defined as non-responders over the whole time. In 3/4 patients, symptoms got worse. They subsequently received canakinumab (CANA)/nonsteroidal anti-inflammatory drugs (NSAID), etanercept (ETA)/NSAID and anakinra (ANA)/MTX, respectively. Follow-up data from the fourth non-responder was not available.

Before starting TCZ therapy, 33/46 (72%) patients received disease-modifying antirheumatic drugs (DMARDs) (cyclosporine A (CSA), MTX, azathioprine (AZA), abatacept (ABA), colchicine (COL)), 24/46 (52%) biologicals (ANA, ETA, adalimumab (ADA), CANA, infliximab (IFX)), 23/46 (50%) corticosteroids (CO) and 21/46 (46%) NSAIDs. 6/46 (13%) children received no co-medication before TCZ.

Co-medications at the initiation of TCZ were DMARDs (CSA, MTX, AZA) in 35/46 (76%) patients, NSAID in 16/46 (35%) and corticosteroids (CO) in 18/46 (39%) who received steroids in high- or low-dose and 3/18 (17%) who required additional steroid-pulses. 2/46 (4%) patients received ANA or rituximab (RTX) additionally to TCZ. In our patient population 21/46 (46%) had received TCZ as first biological treatment, 25/46 (54%) as second, third, fourth or fifth line therapy. Among those receiving TCZ as first line treatment, 14/21 (67%) showed ID or CRM after 12 months. The switch of biological agents to TCZ resulted in 14/25 (56%) additional cases in ID or CRM. In 9/46 patients a switch from TCZ to other biological agents was seen because of non-response (n = 4) or adverse effects (n = 5).

Over the whole period of TCZ therapy corticosteroids (CO) were discontinued in 50% (9/18) children. Under observation co-medication were used in 40/46 cases. Co-medications like DMARDs, corticosteroids (CO) and others were very variable and depended on the participating center. Monotherapy with TCZ was rare in responder (n = 2 and n = 1 according to response after 12 weeks and 12 months). After 12 weeks clinical responders (16/46) received TCZ + CO n = 3, TCZ + DMARD etc. n = 11, TCZ + DMARD etc. + CO n = 0 and only TCZ n = 2. After 12 months clinical responders (18/24) received TCZ + CO n = 1, TCZ + DMARD etc. n = 15, TCZ + DMARD etc. + CO n = 1 and only TCZ n = 1 (DMARD = CSA, MTX, AZA, ABA, COL, etc. = ANA, ETA, ADA, CANA, IFX, RTX).