Efficacy and Safety of Tocilizumab in a Real-Life Observational Cohort of Patients With Polyarticular Juvenile Idiopathic Arthritis

Minna-Maija Grönlund; Terhi Remes-Pakarinen; Liisa Kröger; Kati Markula-Patjas; Maria Backström; Anne Putto-Laurila; Kristiina Aalto; Paula Vähäsalo


Rheumatology. 2020;59(4):732-741. 

In This Article


Fifty-six pJIA patients were included in the study. The baseline patient and disease characteristics are shown in Table 1. Tocilizumab was the first-line biologic in six patients (11%) and these treatments were started in the years 2015–17. In patients receiving biologic therapy before tocilizumab, the median number of previous biologics was 2 (range 1–4). The distribution of different biologics used is shown in Figure 1A. The reasons for switching biological agents are shown in Figure 1B. Before starting tocilizumab all patients had used sDMARDs; all had used MTX, and the median number of different sDMARDs used was 2.5 (range 1–5, including prednisolone and sDMARDs in different combinations). The concomitant sDMARD treatment at tocilizumab commencement is shown in Table 1.

Figure 1.

Use of biologics and reasons for their discontinuation before tocilizumab commencement
(A) Distribution of different biological agents used before tocilizumab commencement in patients with polyarticular JIA.*Other biological agents used; 1st biologic: one abatacept, 2nd biologic: one abatacept and two golimumab, 3rd biologic: five abatacept, one golimumab and one certolizumab pegol, 4th biologic: three abatacept and one rituximab. (B) Reasons for discontinuation of biological agents before tocilizumab commencement in patients with polyarticular JIA.

Tocilizumab Modifications, Survival and Discontinuations

The tocilizumab infusion interval was shortened in nine patients due to inefficacy: in one patient to every other week and in eight patients to every 3 weeks. After the shortened infusion interval seven patients had at least one follow-up visit (6–12 months after the change): six patients continued tocilizumab up to the last available clinical visit and one discontinued tocilizumab 1 month after the change due to inefficacy. The median JADAS-10 score decreased from 9.3 (IQR 7–10.8) to 5.8 (IQR 2.7–8.3) (from the clinical visit before the change to the last clinical visit after the change) in those continuing tocilizumab treatment. One patient had infusions every 5 weeks due to frequent neutropenia.

Drug survival at 6 months, 1 and 2 years of treatment was 91, 82 and 64%, respectively (Figure 2). Altogether 16 patients discontinued tocilizumab at the median treatment duration of 6.0 months (range 2.8–20.9 months). The reasons for discontinuation are given in Table 2. For those who discontinued the tocilizumab treatment the median JADAS-10 was 5.7 (IQR 2.3–9.7), the median number of active joints 1.5 (IQR 0–4.0), the median CHAQ-DI 0.625 (IQR 0.1875–1.0) and the median patient's/parent's pain-VAS 23.5 mm (IQR 5–47 mm) at the last visit.

Figure 2.

Tocilizumab survival in patients with polyarticular JIA

Tocilizumab Monotherapy

Nine patients (16%) started tocilizumab treatment as monotherapy and during the follow-up 11 more patients switched to monotherapy. Only one patient with monotherapy restarted sDMARDs at 24 months. The proportion of monotherapy was 28% at 12 months and 38% at 24 months.

The median JADAS-10 scores in the patients with continuous monotherapy from the start of tocilizumab (n = 8) and those with continuous combination therapy (tocilizumab and sDMARD) (n = 31) were comparable at 12 months of follow-up (P = 0.5) (data not shown).

Tocilizumab Efficacy

The proportions of patients with LDA (JADAS-10 ≤3.9), inactive disease (JADAS-10 ≤0.7) and those with clinically inactive disease according to Wallace's preliminary criteria are shown in Figure 3. The median JADAS-10 decreased from 7.6 (IQR 5.4–12.1) at commencement to 1.2 (IQR 0–2.5) at 24 months (P < 0.0001 Friedman's test for difference over time) and the median number of sDMARDs per patient from 2.5 (IQR 2–3) to 1.0 (IQR 0–1) (P < 0.0001 Friedman's test). The proportion of patients receiving intra-articular glucocorticoid injections decreased from 89% at baseline (i.e. injections within 6 months prior to tocilizumab start) to 17% at 24 months (i.e. injections between the 12- and 24-month clinical visits) (P = 0.004 Cochran's Q test for difference over time). The proportion of patients receiving systemic glucocorticoids at the clinical visit decreased from 27% at baseline to 17% at 24 months of follow-up (P < 0.0001 Cochran's Q test). There was a clear decrease in the patient's/parent's pain-VAS from median 28 (IQR 11–51) to 1.5 (IQR 0–40) (P = 0.032 Friedman's test) and CHAQ-DI from 0.625 (IQR 0.125–1.0) to 0.0625 (IQR 0–0.5) from baseline to 24 months (P = 0.0001 Friedman's test), respectively.

Figure 3.

Clinical outcomes of patients with polyarticular JIA remaining on tocilizumab treatment
n1 = The numbers of patients with the data required for calculating JADAS values (the lower value) and assessing the clinically inactive disease by Wallace's criteria (the higher value) at each time point. n2 = The total numbers of patients remaining on tocilizumab treatment. JADAS: 10-joint Juvenile Arthritis DAS.


Eleven patients had a previous history of uveitis at tocilizumab commencement. Five of them had active uveitis when starting tocilizumab; three continued to have active uveitis throughout the tocilizumab treatment (two continued tocilizumab for 24 months and one discontinued at 21 months because of active uveitis and arthritis), one had a single flare and one had no flares during the follow-up. In two patients with no uveitis at the commencement of tocilizumab a uveitis flare was observed at the last visit (18 and 24 months, respectively) and the other four patients had no flares during the follow-up. No new-onset uveitis occurred during the tocilizumab treatment.

Adverse Events

A total of 85.1 patient years (PYs) was captured. Forty patients (71%) experienced a total of 171 AEs (Table 3). The rate of AEs was 200.9/100 PYs. Infections were the most common AEs (54 infections), detected in 24 patients (43%). The most common infections are shown in Table 3; other infections included outpatient pneumonia, sinusitis, conjunctivitis and verrucae (one of each). Gastrointestinal AEs were also frequent [31 AEs in 18 patients (32%)].

Neutropenia was reported 22 times (25.9/100 PYs) in 10 patients (18%) and elevated transaminases 11 times (12.9/100 PYs) in 8 patients (14%). Six patients had recurrent neutropenia and two patients had recurrent elevation of transaminases. One patient had elevated cholesterol values (total cholesterol 5.9 mmol/l, low-density lipoprotein 4.2 mmol/l) and tocilizumab was discontinued because of this and inefficacy.

Three infusion reactions occurred, all during the first tocilizumab infusion and none leading to discontinuation. One patient with shivering, fever, flush and nausea, required overnight hospitalization with i.v. glucocorticoids and hydration. This patient continued tocilizumab treatment for the entire follow-up (26 months) but was given antihistamine medication and acetaminophen before infusions. The other two patients had only skin reactions (undefined rash and urticaria, respectively).

The total number of SAEs was 11 (Table 3), occurring in nine patients (16%). The rate of SAEs was 12.9/100 PYs. Three cases of infectious SAEs were recorded; influenza A infection with concomitant bacterial sepsis, Bartholin's gland abscess and tooth infection, with both of the latter requiring surgical treatment. No cases of mycobacterial infection, death, malignancy, new onset of autoimmunity or pregnancy were reported.