To assess dosing, preliminary safety, and efficacy of canakinumab, a fully human anti–interleukin-1β (anti–IL-1β) antibody, in children with systemic juvenile idiopathic arthritis (JIA) and active systemic features.
To assess dosing, preliminary safety, and efficacy of canakinumab, a fully human anti–interleukin-1β (anti–IL-1β) antibody, in children with systemic juvenile idiopathic arthritis (JIA) and active systemic features.
In this phase II, multicenter, open-label, dosage-escalation study, children with systemic JIA who were ≥4 years of age, had fever, and were receiving ≤0.4 mg/kg/day of corticosteroids were administered a single subcutaneous dose of canakinumab, 0.5–9 mg/kg of body weight, and were redosed upon relapse. Response to treatment was assessed according to an adaptation of the American College of Rheumatology (ACR) pediatric criteria for improvement.
A total of 23 children ages 4–19 years with active disease were enrolled. Of these, 1 patient was excluded from analysis, and 3 of the reenrolled patients were included twice in the efficacy analysis. By day 15 of the first treatment cycle, 15 of 25 patients (60%) had achieved an adapted ACR Pediatric 50 response, with 4 of them achieving inactive disease status. Response was sustained over time, with 11 of 13 patients able to maintain their response throughout the study. In 8 of the 11 responders who had been receiving steroids at baseline, the steroid dosage was decreased from a mean of 0.38 mg/kg/day to 0.13 mg/kg/day over the first 5 months, and 4 of them were able to discontinue steroids. At a dose of 4 mg/kg of canakinumab given subcutaneously every 4 weeks, the median percentage of patients predicted to relapse within 4 weeks was estimated to be 6% (95% confidence interval 1–21). Therapy was generally well tolerated and few patients experienced injection-site reactions.
Canakinumab has a promising preliminary safety and efficacy profile in this limited cohort. Based on the findings of this trial, further studies in a larger population of children with systemic JIA are warranted.
Systemic disease accounts for 10–20% of all patients with juvenile idiopathic arthritis (JIA) (1–4). It is characterized by the presence of chronic arthritis and systemic signs and symptoms, such as daily high spiking fevers, rash, generalized lymphadenopathy, hepatosplenomegaly, and serositis. Systemic JIA is usually associated with laboratory abnormalities that include anemia (5), leukocytosis, thrombocytosis, an elevated C-reactive protein (CRP) level, and an elevated erythrocyte sedimentation rate (ESR). As many as one-third of these patients develop chronic disease associated with destructive arthritis, leading to significant disability and poor quality of life (6, 7). Approximately 10% of children with systemic JIA develop the potentially life-threatening complication of macrophage activation syndrome (8, 9).
Current treatments for systemic JIA have proved largely unsatisfactory (10–14). Management of the disease relies on corticosteroids. Children with systemic JIA do not respond well to disease-modifying agents such as methotrexate, and poor responses have also been reported with newer agents such as anti–tumor necrosis factor α (TNFα) (14, 15). Several reports have suggested a major role for cytokines such as interleukin-6 (IL-6) and, more recently, IL-1 (16–24) in the disease.
IL-1β is a pivotal cytokine that drives inflammation and tissue destruction in arthritis (25, 26) and is overproduced in patients with systemic JIA (18). Emerging evidence suggests that inhibition of IL-1β may provide clinical benefit in systemic JIA (19, 20, 23, 24, 27), as well as in other autoinflammatory conditions (28–31).
Canakinumab (ACZ885 [Ilaris]; Novartis) is a fully human monoclonal antibody that provides potent and selective blockade of IL-1β and has a long half-life. The present phase II, multicenter, open-label, repeated-dose, dose-finding study in children with systemic JIA and active systemic features was designed to establish the optimal dosing regimen for subcutaneous canakinumab and to make a preliminary assessment of its efficacy and safety.
Patients were eligible for inclusion if they were ages 4–20 years, weighed ≥12 kg, and had a confirmed diagnosis of systemic JIA according to the International League of Associations for Rheumatology (ILAR) criteria (32, 33). Patients were included if they had disease of at least 6 months' duration and active disease at the time of enrollment, defined as the presence of at least 2 joints with active arthritis (34), a daily spiking temperature above 38°C for several hours within the week prior to randomization, and CRP levels of >50 mg/liter. Informed consent or parental informed consent was obtained and study protocol approved by the local ethics committees of all participating centers.
Eligible patients had to discontinue anakinra (for lack of efficacy, local side effect, or unwillingness to continue with daily injections) and any disease-modifying or immunosuppressive drugs, with the exception of stable doses of methotrexate, nonsteroidal antiinflammatory drugs, and corticosteroids (>1 week of stable dose of oral prednisone or equivalent at a dose of ≤0.4 mg/kg/day or ≤20 mg/day, whichever was lower).
Patients were excluded from the study if they had taken part in another clinical study within 4 weeks prior to the baseline visit, if they had uncontrolled severe systemic signs or features of macrophage activation syndrome (8), or had received treatment with biologics, azathioprine, cyclophosphamide, chlorambucil, or 6-mercaptopurine within 4–24 weeks prior to the baseline study visit.
The study was a multicenter, open-label, repeat-dose, dose-finding study structured in 2 stages: a dose-finding stage 1 and a fixed-dose regimen stage 2.
This stage consisted of a 15-day screening period (with a 72-hour run-in for any patient previously treated with anakinra). Three cohorts were defined according to the initial dose of canakinumab administered. The initial dose (taken subcutaneously) was 0.5 mg/kg of body weight in cohort 1, 1.5 mg/kg in cohort 2, and 4.5 mg/kg in cohort 3. Patients received their first dose of canakinumab on day 1, and if necessary, a second dose was given on day 3 or day 8 (see below for redosing criteria). In each cohort, the second dose administered was the same as the first dose, giving a total administered dose of 1.0 mg/kg in cohort 1, 3.0 mg/kg in cohort 2, and 9.0 mg/kg in cohort 3. For safety reasons, only children over 12 years of age were included in cohort 1, while for cohorts 2 and 3, no more than 1 patient in the age range of 4–12 years was newly dosed in a given week.
Response to therapy during the first treatment cycle (i.e., period 1 of stage 1) was assessed on days 3, 8, and 15 after dosing, and then every 2 weeks until relapse occurred. Patients in a particular cohort who experienced a response to treatment within 15 days during stage 1 stayed in the study and, upon each relapse (see below for definitions), received an injection of the same total dose to which they responded initially. These patients remained in stage 1 until the fixed-dose regimen was established or were discontinued from the study if they failed to continue to respond.
Patients who showed no response by day 15 during period 1 of stage 1 were considered nonresponders, received no further doses of study drug, and were treated according to standard protocols established at their study center. These patients were followed up for 112 days (±5 days) after the last dose of canakinumab for evaluation of safety, immunogenicity, and pharmacokinetics (PK). Nonresponders in the lower-dose cohorts were allowed to reenter the study in a higher-dose cohort.
Dosing for stage 2 was determined according to available safety data and PK/pharmacodynamic (PD) modeling after all patients had completed at least 1 treatment period 1 during stage 1. All patients who responded during stage 1 were switched to a fixed-dose regimen upon relapse. Responders could remain in the study until they could be recruited into a phase III study.
For the primary outcome, children were defined as responders according to the American College of Rheumatology (ACR) criteria for 30% improvement (Pediatric 30 criteria), with at least 30% improvement in 3 of the 6 core set of variables (physician's global assessment of disease activity on a 0–100-mm visual analog scale [VAS], number of joints with active arthritis, number of joints with limited range of motion, CRP level, functional ability as assessed with the Childhood Health Assessment Questionnaire [C-HAQ]) and with no more than 1 of the remaining variables worsening by >30% (35, 36). In addition, the absence of fever (body temperature ≤37.5°C) during the preceding week was required to qualify a patient as a responder (adapted ACR Pediatric 30 criteria) (37).
For the secondary outcome, children were also evaluated for attainment of the following: (a) adapted ACR Pediatric 50, Pediatric 70, and Pediatric 90 criteria; and (b) inactive disease (modified from the definition of Wallace et al ), which was defined as no joints with active arthritis, absence of fever (body temperature ≤37.5°C), absence of signs of systemic JIA, presence of normal ESR/CRP values, and absence of disease activity according to the physician's global assessment (modified from a score of 0 mm to a score of ≤10 mm on a 0–100-mm VAS to take into account the inherent measurement error of the VAS [39, 40]). In addition, parents rated the patient's pain using a 0–100 mm VAS, where 0 mm = no pain and 100 mm = worst pain.
Redosing occurred in patients who had a persistent/recurrent fever at 48 hours after the first dose (day 3) or in patients who had fever or persistently elevated CRP levels (>10 mg/liter according to values provided by the local laboratory of each participating center) 1 week after the first dose (day 8).
For all subsequent visits during stages 1 and 2, canakinumab was administered upon relapse. Relapse was defined as (a) the reappearance of fever (body temperature >37.5°C) attributable to systemic JIA and not to an infection, as well as a CRP level >30 mg/liter, or (b) attainment of the criteria for polyarticular-course JIA for flare, defined as worsening by at least 30% in 3 of the 6 ACR core set of variables (with ≥30% improvement in not more than 1 of the remaining variables), with the following contingencies: worsening in at least 2 joints with active or limited disease, at least a 20-mm worsening in the physician's or parent's global assessment of disease activity, and a CRP level >30 mg/liter with respect to the previous visit (40, 41). The CRP cutoff value for redosing on day 8 (>10 mg/liter) was chosen pragmatically to facilitate proper dosing during the initial phase of the study. The CRP cutoff value for redosing was raised to >30 mg/liter and concomitant fever was added for any subsequent disease relapse to avoid any minor daily fluctuations in the CRP that could have unnecessarily required a dose of the drug.
Safety and efficacy were assessed at baseline, on day 1 before the dose of study drug was administered, and at each subsequent clinic visit.
Blood samples were collected at each study visit to determine serum concentrations of canakinumab and total IL-1β. Canakinumab concentrations were measured by competitive enzyme-linked immunosorbent assay (ELISA); the lower limit of quantification was 100 ng/ml. The PK parameters for canakinumab were determined by noncompartmental analyses. Noncompartmental PK parameters (from serum subcutaneous PK data) included the following: dose-normalized maximum serum concentration (Cmax), time to maximum observed serum concentration (Tmax), elimination half-life (T1/2), the apparent volume of distribution (Vz/F) during the terminal phase, and apparent total clearance (Cl/F, where F is the absolute bioavailability).
All available PK and median time to relapse data, together with their 95% confidence intervals (95% CIs), as well as the probability of relapse within 30 days were estimated for the different dose levels of stage 1, with data from the 6 dose levels (0.5, 1, 1.5, 3, 4.5, and 9 mg/kg) pooled into 3 dose categories (<3 mg/kg, 3 mg/kg, and >3 mg/kg) using a Weibull gap-time frailty model including covariates (baseline CRP, prednisone equivalent dose, and white blood cell [WBC] count) (45–47) to establish the fixed-dose regimen to be used in stage 2. At the end of the study, the model was again run including data from stage 2, and the estimates were also calculated for canakinumab at a dose of 4 mg/kg of body weight. From the model fit, the median times to relapse for each dose level and the probability of relapse within 30 days was estimated via simulation; the simulation was carried over to define a suitable dosing regimen that would maintain most patients in a flare-free state.
The analysis was mainly descriptive in nature, with data reported as frequencies and medians with ranges or confidence intervals as appropriate and according to the Consolidated Standards of Reporting Trials (CONSORT) group statement (43, 44).
The sample size for the study was chosen in order to minimize the exposure of a pediatric population to the drug and to proceed cautiously, given that this was the first time canakinumab was administered to children. Therefore, 26 patients were recruited, 6 in the first cohort and 10 in each of the other 2 cohorts. Since systemic JIA was considered an orphan indication until anti–IL-1 and anti–IL-6 therapy became available, the expectation was that a new drug should show a substantial response in order to be considered effective by both physicians and parents. Therefore, for the sample size calculation, assuming an underlying true response rate of 80% in a cohort of 10 patients, there is 50% confidence that the true response rate is >70% in ∼68% of the cases (68% power to detect a response).
All patients who received at least 1 dose of study drug in accordance with the protocol were included in the safety and efficacy evaluation. Two subgroups of patients were defined based on their response status (responders and nonresponders) as assessed by PRINTO during the visit on day 15 of period 1 during stage 1. Patients who were reenrolled contributed twice to the efficacy analysis according to their response status from their first and second enrollment in stage 1.
An analysis of the corticosteroid (prednisone equivalent) doses administered over the first 5 months (150 days) after administration of the first dose of canakinumab was performed for responders who were receiving steroids at baseline.
A total of 23 patients were enrolled in the study. Their demographic and baseline characteristics are summarized in Table 1. Three patients reentered the study at a higher dose level; therefore, overall data are available for 26 patient enrollments: 5 into cohort 1, 10 into cohort 2, and 11 into cohort 3 (Figure 1). One patient was excluded from the efficacy analyses because of a protocol violation, having received pulse doses of corticosteroids for acute worsening of disease after the first canakinumab injection. Three patients who did not respond or lost their response to canakinumab in the original cohort were reenrolled in a higher-dose cohort, and these 3 were included twice in the efficacy analysis with data from both enrollments in stage 1.
|Patients included in the safety analysis†|
|No. of patients||23|
|Male, no. (%)||12 (52.2)|
|Caucasian, no. (%)||22 (95.7)|
|Age, median (range) years||10 (4–19)|
|Disease duration, median (range) months||38 (7–204)|
|No. (%) of patients receiving steroid therapy||19 (83)|
|Prednisone-equivalent dose, median (range) mg/kg||0.316 (0.03–0.83)|
|No. (%) of patients who previously received anakinra||17 (69.6)|
|No. (%) of patients receiving MTX‡||6 (26.1)|
|MTX dose, median (range) mg/week||11.3 (7.5–25)|
|Patients included in the efficacy analysis§|
|No. of patients||25|
|Joints with active arthritis, median (range)||20.0 (4.0–62.0)|
|Joints with limited range of motion, median (range)||27.0 (4.0–62.0)|
|CRP, median (range) mg/liter||143.7 (30.5–271.0)|
|ESR, median (range) mm/hour¶||75 (26–121)|
|Physician's global assessment of disease activity, median (range) mm, by VAS||68.0 (5.0–100.0)|
|C-HAQ score, median (range)||2.1 (1.1–3.0)|
|Parent's global assessment of child's overall well-being, median (range) mm, by VAS||67 (12.0–100.0)|
|Parent's assessment of child's pain, median (range) mm, by VAS||76.0 (23.0–100.0)|
Approximately half of the study patients were male. The median disease duration was 38 months, and >70% of them were under treatment with corticosteroids or had been previously treated with anakinra. The 6 ACR core set of variables for juvenile arthritis showed a high degree of articular involvement (median of 20 joints with active disease), inflammation, and disability and a median VAS score of >65 mm for the physician's global assessment of disease activity, the parent's assessment of the child's overall well-being, and the parent's assessment of the child's pain level (Table 1).
Overall, 15 of 25 patients (60%) were classified as responders according to the adapted ACR Pediatric 30 criteria, and 15 (60%) achieved at least a Pediatric 50 response, with 5 of the 25 patients (20%) achieving a Pediatric 90 response, and 4 of them achieving inactive disease status, within 15 days of starting canakinumab treatment (Figure 2). One patient in cohort 1 who did not respond to a total dose of 1 mg/kg responded upon receiving a total dose of 3 mg/kg of canakinumab when reenrolled in cohort 2. Two patients who had initially responded to low and intermediate doses of canakinumab subsequently lost their response and were reenrolled in a higher-dose cohort in stage 1. Responder rates and efficacy evaluations in stage 1 were performed by including data from both first and second enrollment for the reenrolled patients (Figure 2A).
Eleven of the stage 1 responders were redosed upon relapse with canakinumab 4 mg/kg during stage 2. Two of the stage 1 responders did not enter stage 2: 1 patient did not experience a relapse and withdrew from the study after a period of ∼200 days of inactive disease, while another patient received only 1 additional dose of canakinumab during stage 1 and had no disease activity thereafter until the end of study (study day 665). Response was sustained over time, with 11 of 13 patients maintaining their adapted ACR Pediatric 30 response throughout the study and 7 having at least an adapted ACR Pediatric 50 response after the last-available treatment cycle (Figure 2B). The time to relapse ranged from a few days to >240 days. There was large heterogeneity in the relapse pattern among patients, while intrapatient relapse patterns exhibited periodicity, with no evidence of tachyphylaxis (Figure 2C). Overall, 12 patients received canakinumab for more than 1 year, of whom 7 were exposed to the drug for more than 2 years.
Comparisons of clinical variables and PD parameters in responders versus nonresponders at baseline indicated some noteworthy differences between the 2 subgroups at baseline. Most notably, the number of joints with active arthritis decreased from a median of 8.0 at baseline (n = 15) to a median of 2.0 on day 15 (n = 15) in the responders and decreased from a median of 30.5 at baseline (n = 10) to a median of 16.5 on day 15 (n = 6) in the nonresponders.
In addition to decreases in the ESR and the CRP level, patients who responded to canakinumab treatment also showed rapid decreases in the WBC and neutrophil counts, from elevated values at baseline to normal values on day 15. Available data from the responders (n = 13) indicated that the median WBC count decreased from 13.7 × 109/liter at baseline to 9.5 × 109/liter on day 15, and the median neutrophil count decreased from 11.8 × 109/liter to 4.7 × 109/liter at the respective time points. In contrast, in nonresponders (n = 7), the median WBC count was 11.40 × 109/liter at baseline and 11.50 × 109/liter on day 15, and the median neutrophil count was 8.10 × 109/liter at baseline and 7.00 × 109/liter on day 15.
Eleven responders were receiving concomitant steroid therapy for control of systemic JIA symptoms at the time the first dose of canakinumab was administered. Successful tapering of the steroid dose during canakinumab treatment was achieved in 8 of the 11 responders who were receiving corticosteroids at baseline, with 4 of them being able to discontinue steroid treatment by the end of the study. During the study period, the average rate of decrease of the steroid dose was 0.05 mg/kg/month, from a mean of 0.38 mg/kg (95% CI 0.25–0.52) at baseline to 0.13 mg/kg (95% CI 0.01–0.25) at the last available observation.
Seventeen patients had previously received anakinra (with 11 having no response, 3 having a partial response, and 2 having a complete response; response was unknown in 1). Six of 11 patients who had not responded to anakinra achieved at least an adapted ACR Pediatric 50 response on day 15 after a single dose of canakinumab.
The patients who were responders on day 15 of period 1 during stage 1 were eligible to receive repeated treatment upon relapse, and their cases were followed through further cycles of dosing and relapse. Time to relapse ranged from 1 week to 36 weeks; by week 56, 1 patient had not experienced a relapse (Figure 2C). There was a large heterogeneity in the pattern of relapse among the patients and no evidence of tachyphylaxis.
From the results of the time-to-relapse analysis using the Weibull gap-time frailty model, the longest median time to relapse was estimated to be >90 days for the canakinumab 3 mg/kg group, although the 95% CIs were wide and overlapping (Figure 3A). Based on this model, among those receiving subcutaneous canakinumab doses >3 mg/kg, the probability of relapse within 30 days was estimated to be 7% (95% CI 1–23) (Figure 3B). These estimates were further confirmed by the results of the PK flare model that was used to determine the critical systemic threshold concentration of canakinumab at which there would be an acceptably low probability of flare. Doses between 1 mg/kg and 7 mg/kg were simulated via Monte Carlo simulations; Figure 3B shows the resulting percentage of patients relapsing at the end of 4 weeks. As predicted by the simulations, subcutaneous canakinumab 4 mg/kg administered every 4 weeks should ensure that most patients would remain above a critical flare threshold concentration of canakinumab of ∼2 μg/ml. Specifically, the expectation is that a relapse would occur in 6% of patients (95% CI 1–21) after 4 weeks if treated with canakinumab 4 mg/kg. Higher doses may result in slightly greater responses, but 4 mg/kg appears to approach the saturation of response. Thus, canakinumab 4 mg/kg every 4 weeks was deemed the most appropriate dosing regimen for maintaining patients free of relapses. A total of 11 responders entered stage 2 and received canakinumab 4 mg/kg upon relapse until rollover into the phase III trial (Figure 2C).
Following subcutaneous administration, the mean peak serum concentration of canakinumab was reached within ∼2 days in most patients. The half-life of canakinumab was 16.7 ± 5.45 days (mean ± SD). The PK properties of canakinumab following subcutaneous administration were typical of a human IgG1-type antibody, as demonstrated by a low volume of distribution and low mean systemic clearance (Table 2). No relevant differences were noted in the individual PK profiles of canakinumab in responders versus nonresponders.
|Cmax/dose, μg/ml/mg of administered dose (n = 6)†||Tmax, days (n = 6)†||T1/2, days (n = 21)||Cl/F, liters/day (n = 21)||Vz/F, liters (n = 21)|
|Mean ± SD||0.168 ± 0.031||2.62 ± 2.08||16.7 ± 5.45||0.256 ± 0.0994||5.94 ± 2.54|
No deaths or macrophage activation syndrome were reported during the study. A 22-year-old female patient died of pneumococcal sepsis 2.25 years after the last canakinumab injection; she had received 2 injections of 1.5 mg/kg during the study. No patients discontinued the study drug because of adverse events (AEs). With regard to the AEs, the most frequently affected organ systems/classes were gastrointestinal disorders and infections, and the most frequent AEs were cough, abdominal pain, vomiting, diarrhea, and pyrexia (Table 3). The majority of AEs reported were mild to moderate in severity.
|Adverse event||No. (%) of subjects (n = 23)|
|Any adverse event||22 (96)|
|Abdominal pain||8 (35)|
|Abdominal pain upper||5 (22)|
|Pharyngeal erythema||4 (17)|
|Acute tonsillitis||3 (13)|
|Upper respiratory tract infection||3 (13)|
Eleven patients had serious AEs, and in 2 patients (one with Epstein-Barr virus infection during stage 1 and the other with hematoma, prolonged activated partial thromboplastin time, gastroenteritis, and syncope during stage 2), these were suspected to be related to the study drug. All AEs resolved spontaneously or with appropriate medication. There was no obvious relationship between the canakinumab dose and the frequency or type of AE. There were no clinically relevant changes in the laboratory findings, electrocardiographic results, or vital signs.
Canakinumab injections were well tolerated, and there were no reports of severe injection-site reactions. No anticanakinumab antibodies were detected. There was no apparent increase in AEs with higher doses up to 9 mg/kg.
In this dose-escalation study, treatment with canakinumab 4 mg/kg upon relapse was associated with rapid, substantial, and sustained clinical responses in patients with systemic JIA. The clinical efficacy of canakinumab was demonstrated by improvements in ACR criteria for juvenile arthritis (adapted ACR Pediatric 30, Pediatric 50, Pediatric 70, and Pediatric 90 criteria) and reductions in clinical parameters and levels of markers of inflammation. Most patients showing clinical response to treatment were able to taper the dosage of concomitant steroids or to discontinue steroids.
Sixty percent of the patients with systemic JIA treated in this study experienced an improvement in symptoms, meeting at least an adapted ACR Pediatric 50 within 15 days of canakinumab administration, with 4 achieving inactive disease status. These levels of response were substantially maintained in responders for up to 2 years of followup. The clinical response and efficacy results reported herein are likely to underestimate the potential response rate to canakinumab in systemic JIA because patients in the low-dose cohorts are likely to have received subtherapeutic doses of canakinumab.
The study did not show a clear dose-response relationship. This is possibly due to the small number of patients treated in each dose group as well as the nonrandomized design.
Canakinumab provides potent and highly selective blockade of IL-1β (30). The clinical efficacy of canakinumab observed in this study confirms that IL-1β is a pivotal driver in the pathogenesis of a large proportion of systemic JIA. Our findings are consistent with, and extend, evidence from studies with the recombinant form of IL-1 receptor antagonist, anakinra, which provided proof of concept for the blockade of IL-1–associated inflammation as a therapeutic target in systemic JIA and other systemic autoinflammatory conditions, such as cryopyrin-associated periodic syndromes (CAPS) (18, 20, 23, 24, 27, 29, 48). Although anakinra is an effective IL-1 antagonist by blocking interactions between IL-1α as well as IL-1β and their receptors, it has a short half-life of 4–6 hours, thus requiring daily subcutaneous administration (23), which is often associated with injection-site reactions (49). Canakinumab, with its longer half-life and selective targeting of IL-1β, may be a more convenient, better-tolerated agent for the management of systemic JIA.
Treatments targeting other cytokines implicated in the pathogenesis of systemic JIA, such as tocilizumab, which blocks IL-6, have also demonstrated efficacy in clinical studies (16, 17). However, the relative importance of the different cytokines implicated in the pathology of systemic JIA remains to be determined and may partly account for some of the heterogeneity of responses noted in clinical studies in patients with systemic JIA. The exact role of the currently available agents in the management of systemic JIA will be defined by future studies.
As with other studies of IL-1 blockade in systemic JIA (18, 20, 49), heterogeneity of response to canakinumab treatment was observed in this study. Responders to canakinumab had fewer joints with active disease and a higher WBC count at baseline than did nonresponders, which is consistent with the findings of previous studies (20). Identification of predictors of outcome in systemic JIA is important for guiding treatment decisions and for tailoring therapy choices to the patient's risk of disability (7, 20, 50).
Patients with systemic JIA are often heavily dependent on steroid therapy (1, 3, 10, 14). However, long-term steroid treatment is associated with a range of side effects that affect the patient's health and well-being, including an increased risk of vertebral compression fractures, cataracts, growth retardation, and susceptibility to infection. Thus, an important goal for any novel therapeutic used in systemic JIA is to reduce the need for steroids. In this study, most patients (19 of 23 [83%]) were receiving steroid therapy at study entry. Four of the 11 responders who were receiving steroids at baseline were able to discontinue this treatment over the course of the study.
The PK data reported in this study are consistent with previous observations (31), confirming that canakinumab shows typical human IgG1-type antibody clearance. The 17-day half-life of canakinumab determined in this population of children with systemic JIA was found to be shorter than in adult CAPS patients, which was ∼26 days. The results of the Weibull gap-time frailty model and the PK flare model were consistent. The median time to relapse for canakinumab doses of >3 mg/kg was estimated to be 90 days, with a 93% probability of remaining relapse-free during the 30 days following treatment. Results from the PK flare model indicated that a dosing regimen of canakinumab at 4 mg/kg, given every 4 weeks, should maintain patients in a flare-free condition. This regimen has been used in phase III studies in patients with systemic JIA.
The safety data collected in this cohort of patients ages 4–19 years were reassuring, with only 2 patients reporting serious AEs that were suspected to be related to the study drug. There were no severe injection-site reactions, no cases of macrophage activation syndrome, and no evidence of immunogenicity to canakinumab treatment at the doses studied.
In conclusion, this study provides preliminary data showing that treatment with canakinumab 4 mg/kg is associated with rapid and sustained improvement in clinical responses and with the ability to reduce or discontinue steroids in patients with systemic JIA. The results of this phase II trial provide a rationale for further studies in a larger population; a phase III trial is currently under way in children with systemic JIA and features of active systemic disease.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Ruperto had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Drs. Ruperto and Martini, who were involved in the study design, evaluation of the primary outcome, interpretation of the data, and writing, editing, and critical revision of the manuscript, had full access to data from the report for regulatory authorities, and had final responsibility for the decision to submit the manuscript for publication and the journal to which it would be submitted. Drs. Ruperto, Quartier, Wulffraat, Woo, and Martini were members of the steering committee and were involved in the decision to move from dose cohort 1 to cohort 2 and to cohort 3. The remaining coauthors contributed to critical data evaluation and interpretation.
Study conception and design. Ruperto, Woo, Ravelli, Bader-Meunier, Noseda, Chakraborty, Martini, Chioato.
Acquisition of data. Ruperto, Quartier, Wulffraat, Woo, Ravelli, Mouy, Vastert, Noseda, Chioato.
Analysis and interpretation of data. Ruperto, Quartier, Woo, Ravelli, Bader-Meunier, Noseda, D'Ambrosio, Lecot, Chakraborty, Martini, Chioato.
The trial was sponsored by the Novartis Institute for Biomedical Research, Basel, Switzerland, which had full responsibility for the design, planning, and management of the study, the collection, analysis, and interpretation of the data, the writing and submission of the study report to regulatory authorities, the writing of the manuscript, and the right to review the final version before submission. Publication of this article was contingent upon approval by Novartis Institute for Biomedical Research.
The authors would like to thank the patients who participated in the study. The authors and Novartis particularly wish to thank Dr. Christiane Rordorf, study coordinator at Novartis, who recently passed away, for her deep, thoughtful, and invaluable involvement in the design and success of the study.
Laura Carenini, Simona Angioloni, Luca Villa, and Anna Tortorelli (PRINTO Coordinating Center) were involved in the evaluation of response to therapy, Rossella Belleli, Johanna Della Valle, and Beate Kiese (statistics and programming) and Stacey Tannenbaum (modeling and simulation) contributed to the study design and data analysis, and Dr. Agnes Mogenet, Solimda Sotou-Bere, and Yamina Boulahdaj (Centre d'Investigation Clinique, Hôpital Necker–Enfants Malades) assisted with the conduct of the study.