Efficacy of repeated intravenous infusions of an anti–tumor necrosis factor α monoclonal antibody, infliximab, in persistently active, refractory juvenile idiopathic arthritis: Results of an open-label prospective study




To evaluate the efficacy and safety of a chimeric monoclonal anti–tumor necrosis factor α antibody (infliximab) with methotrexate (MTX) in juvenile idiopathic arthritis (JIA) with an active polyarticular course that is not responsive to MTX.


Twenty-four young adults with long-lasting, refractory JIA were enrolled in an open, prospective, 2-year pilot study. Patients received intravenous infliximab at 3 mg/kg of body weight at weeks 0, 2, and 6 and every 8 weeks thereafter, with weekly subcutaneous MTX.


The median duration of therapy was 9.1 months. Significant improvements were observed in the number of joints (28-joint count) with active disease (median 6 at baseline, 2 at 2 weeks, 0 at 6 months, 0 at 1 year; P < 0.05). Pain as well as patient's and physician's global assessments of disease status were assessed on 0–100-mm (0 = best; 100 = worst) visual analog scales (VAS). There were significant improvements in VAS pain scores (45 at baseline, 25 at 2 weeks, 8.5 at 6 months, 10 at 1 year; P < 0.05), patient's global assessment of disease status (50 at baseline, 22 at 2 weeks, 11.5 at 6 months, 18 at 1 year; P < 0.05), and physician's global assessment of disease status (50.5 at baseline, 22.5 at 2 weeks, 6.5 at 6 months, 10 at 1 year; P < 0.01). In addition, there were significant improvements in the erythrocyte sedimentation rate (64 mm/hour at baseline, 36 mm/hour at 2 weeks, 23.5 mm/hour at 6 months, 35 mm/hour at 1 year; P < 0.01) and C-reactive protein level (4.9 mg/dl at baseline, 2.8 mg/dl at 2 weeks, 3.1 mg/dl at 6 months, 3.2 mg/dl at 1 year; P < 0.005). The percentage of patients meeting the American College of Rheumatology 20% improvement criteria at each assessment ranged from 54.2% to 86.7%. Of the responses on the Disease Activity Score in 28 joints, 37.5–63.6% were classified as “good,” 14.3–33.3% were classified as “moderate,” and 18–37.5% were classified as “no response.” Twelve patients (50%) had adverse events, and 5 patients (20.8%) withdrew.


Infliximab plus MTX showed high effectiveness and safety in short- and medium-term treatment of long-lasting refractory JIA. A controlled multicenter clinical trial is needed.

In adults with active rheumatoid arthritis (RA), controlled clinical trials have demonstrated the clinical benefit of tumor necrosis factor α (TNFα)–neutralizing therapy with either a recombinant human TNF receptor (p75)–Fc fusion protein (etanercept) (1) or a chimeric (human and murine) monoclonal antibody against TNFα (infliximab; Centocor, Malvern, PA), administered alone or in combination with methotrexate (MTX) (2, 3). Concurrent treatment with MTX appears to enhance the therapeutic response (4, 5). Treatment with etanercept in 1 controlled clinical trial (6) and in 1 open prospective study (7) led to significant improvement in patients with active polyarticular juvenile idiopathic arthritis (JIA) who did not tolerate or had an inadequate response to MTX. We undertook this study considering the necessity of treating patients affected by long-lasting, persistently active JIA that is refractory to treatment with conventional disease-modifying antirheumatic drugs (DMARDs), nonsteroidal antiinflammatory drugs (NSAIDs), and steroids. We therefore used the guidelines for controlled clinical trials of monoclonal antibodies against TNFα in RA because, as far as we know, there are no reported studies on infliximab therapy of JIA.



We recruited all patients followed up at our center from November 1999 to October 2000 who had active, long-lasting, refractory polyarticular disease not responsive to MTX and 1 or more other DMARDs administered as single or combined therapy (median of previous failed treatments 4.0, mean ± SD 3.8 ± 1.5, range 2–7). In particular, therapy with MTX administered subcutaneously (SC) or intramuscularly (IM) in conventional (10 mg/m2 of body surface area/week) or higher (15–30 mg/m2/week) dosages had previously failed for all patients.

Twenty-four young adult females (only 5 were of pediatric age), fulfilling the European League Against Rheumatism (8) and International League of Associations for Rheumatology (9) criteria for the diagnosis of JIA, were enrolled. The population data are shown in Table 1.

Table 1. Disease subsets and demographic data of the 24 patients with juvenile idiopathic arthritis (JIA)*
  • *

    Except where indicated otherwise, values are the number of patients.

JIA subset
 Systemic onset5
 Seronegative polyarthritis5
 Seropositive polyarthritis1
 Extended oligoarthritis10
 Psoriatic arthritis3
Age at disease onset, mean (range) years6 (1–15.8)
Age at enrollment, mean (range) years22.1 (8.2–32.5)
Disease duration, mean (range) years15.3 (5.2–31.5)

At study entry, all patients still had active disease, with a median Disease Activity Score in 28 joints (DAS28) (10) of 5 (mean ± SD 5.1 ± 1.1, range 3.2–7.9) and a median of 7.0 joints with active disease (all joints counted) (mean ± SD 10.2 ± 7.4, range 3–28). During treatment, 5 patients underwent 1 or more joint replacements (mean of 2 joints replaced per patient, range 1–4).

All enrolled subjects (and/or both parents if the subject was younger than age 18 years) gave their written informed consent for treatment. The local ethics committee approved the trial.

Treatment protocol.

All patients received the first infliximab infusion a mean ± SD of 7 ± 1 days after the baseline examination (week 0). The subsequent infusions were administered at weeks 2 and 6 and every 8 weeks thereafter. The scheduled starting dose was 3 mg/kg of infliximab administered intravenously (IV) over a period of at least 2 hours, and the dosage could be increased according to clinical judgment. NSAIDs and corticosteroids were continued at the preinclusion dose or reduced if possible. Simple analgesics were allowed. MTX was continued or readministered (it had been previously suspended in 6 patients) as IM or SC injections at the previous well-tolerated dosage. All other DMARDs were discontinued at least 1 month prior to trial entry.

Safety monitoring.

A complete physical examination and standard laboratory tests were performed before each infusion. A chest radiograph, an electrocardiogram, and a purified protein derivative test were performed at the screening visit.

Response assessment.

The ongoing study provides for assessment of disease activity indices at baseline (pretreatment; mean ± SD 7 ± 1 days before the first infusion) and just before each subsequent infusion. The following clinical assessments were performed by the same observer (IP), an experienced pediatric rheumatologist: number of swollen joints, number of tender joints, number of joints with limitation of motion, patient's assessment of pain (0–100 mm on a visual analog scale [VAS], with higher scores indicating more severe pain), patient's global assessment of disease status (0–100 mm on a VAS, with higher scores indicating worse status), physician's global assessment of disease status (0–100 mm on a VAS, with higher scores indicating worse status), and an assessment of arthritis-related functional disability (the Health Assessment Questionnaire [HAQ] score; scale ranging from 0 [no difficulty] to 3 [unable to do]) (11). In addition, the observer used the Medical Outcomes Study Short Form 36 (SF-36) health survey to assess the burden of disease in our patients at baseline, after 6 months of treatment (fifth infusion), and after 1 year of treatment (ninth infusion) (12). The SF-36 measures the following 8 aspects of health status on a scale ranging from 0 (worst) to 100 (best): general and mental health, physical activity, social activities, role of physical activity, emotional problems, bodily pain, and vitality.

Routine laboratory indicators of disease activity were further evaluated. These included a complete blood cell count, C-reactive protein (CRP) level by rate nephelometry, and erythrocyte sedimentation rate (ESR; Westergren method).

The improvement in each patient was evaluated at each clinical examination according to the American College of Rheumatology (ACR) criteria (13) and the DAS28 (based on the number of tender joints [28-joint count], the number of swollen joints [28-joint count], and the patient's global assessment of disease status) (10). The DAS28 responses of “good,” “moderate,” or “none” were calculated according to the method of Prevoo et al (10) based on the relationship between baseline absolute values of this index and its percentage reduction from baseline. Furthermore, we calculated the number of joints with active disease (defined as swollen or tender plus limited range joints) at baseline and at each subsequent examination.

Statistical analysis.

Analysis of variance for repeated measures and the Friedman test were used to analyze time variation of the evaluated parameters during the first year of treatment. Pairwise comparisons were based on the Wilcoxon matched pairs signed rank test. The Bonferroni correction was used for multiple comparisons. P values less than 0.05 were considered significant (2-tailed test). The statistical evaluation was based on the last observation carried forward method.


State of treatment.

In this 2-year pilot study, all 24 enrolled patients received at least 2 infusions (range 2–12). Fourteen patients completed 1 year of observation. Nine of these patients completed 1 year of therapy, and 5 (35.7%) dropped out during the first year of therapy, 4 because of infusion reactions and 1 because of a relapse of disease activity. One of the 9 patients who completed the first year of therapy dropped out at month 14 because of an infusion reaction. The median duration of therapy was 9.1 months (mean ± SD 8.8 ± 6.2 months, range 2 weeks–18.8 months). The scheduled starting dose was 3 mg/kg of body weight/infusion. The median dose for maintenance treatment was 4.4 mg/kg/infusion (mean ± SD 4.7 ± 1.7, range 2.9–10). Ten patients had a maintenance dose in the range of 2.9–4 mg/kg/infusion, 10 had a maintenance dose in the range of 4.1–5 mg/kg/infusion, and 4 had a maintenance dose in the range of 5.1–10 mg/kg/infusion.

The median dose of concomitant MTX therapy was 15 mg/week (mean ± SD 12 ± 7, range 5–25). Nineteen patients (79%) were receiving steroid treatment at a median dose of 0.15 mg of prednisone equivalent/kg (mean ± SD 0.17 ± 0.13, range 0.04–0.5). We reduced the dosage of prednisone according to the clinical course, tapering it very slowly. In fact, in 3 patients, the baseline dosage of prednisone was reduced by at least one-third. We never needed to increase the prednisone dosage or to start prednisone therapy during the treatment with infliximab. In 4 of 20 patients, after the first 3 infusions, the scheduled interval between infusions (starting with the fourth and subsequent infusions) had to be shortened from 8 weeks to 6–7 weeks because of a flare of inflammation.

Clinical parameters.

All clinical parameters showed a sharp improvement immediately after the first infusion. The improvement achieved persisted during the subsequent period of therapy (Table 2 and Figure 1). Eight of 24 patients (33.3%) did not have any joints with active disease at the last visit. No significant improvement was observed in the disability index of the HAQ during the first year of treatment (P = 0.733). In the 9 patients who received infliximab for 1 year, the 1-year clinical parameters (assessed just before the ninth infusion) remained substantially unchanged after the first infusion (minimum P = 0.128).

Table 2. Changes in clinical parameters, DAS28 index, and percentages of patients with ACR improvement responses before and during the study (n = 24 patients)*
 Baseline (n = 24)After 2 weeks (n = 24)After 6 months (n = 16)After 1 year (n = 9)P
  • *

    Except where indicated otherwise, values are the median. DAS28 = Disease Activity Score in 28 joints; ACR = American College of Rheumatology; VAS = visual analog scale (0 mm = best condition; 100 mm = worst condition); ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; HAQ = Health Assessment Questionnaire; NS = not significant; ACR20 = 20% improvement in disease activity according to the ACR criteria.

  • By Wilcoxon matched pairs signed rank test.

  • Normal ≤12 mm/hour.

  • §

    Normal 0–1 mg/dl.

  • Value ≥2.1 indicates active disease.

Joints with active disease, 28-joint count6200<0.05
Patient's assessment of pain, 0–100-mm VAS45258.510<0.05
Patient's global assessment, 0–100-mm VAS502211.518<0.05
Physician's global assessment, 0–100-mm VAS50.522.56.510<0.01
ESR, mm/hour643623.535<0.01
CRP level, mg/dl§4.92.833.133.21<0.005
Hemoglobin, gm/liter10.711.312.112.25<0.05
HAQ score1.
ACR20, % of patients54.281.277.8
ACR50, % of patients2563.677.8
ACR70, % of patients12.556.344.4
Figure 1.

Numbers of joints with active disease during the study (all-joint count). Values are the mean. # = P = 0.07; ∗ = P < 0.05, versus baseline.

ACR response and DAS28 index (efficacy of treatment).

The percentage of patients meeting the ACR 20% improvement criteria (the ACR20 responder rate) before each infusion ranged from 54.2% (after 2 weeks of therapy) to 86.7% (in the eighth month of therapy). Six of 9 patients who received infliximab for 1 year (67%) exhibited an ACR20 response for at least 7 of the 8 preinfusion examinations. One patient was an ACR20 responder only 2 times.

Of the DAS28 responses after the first infusion, 37.5% were good (9 patients) and 41.7% were moderate (10 patients), while 5 patients (20.8%) were classified as having no response. In the subsequent period until month 12 of therapy, the good-response mean value was 47% (range 37.5–63.6%), the moderate-response mean value was 23% (range 14.3–33.3%), and the no-response mean value was 23.5% (range 18–37.5%).

SF-36 evaluation.

After 6 months of treatment, the improvement on the SF-36 was weakly significant for bodily pain (P = 0.05), social activities (P = 0.04), emotional problems (P = 0.06), and mental health (P = 0.05). After 1 year of treatment (9 patients), only bodily pain (P = 0.08) and emotional problems (P = 0.06) showed weakly significant improvement.

Adverse events.

Adverse events, which occurred in 12 of 24 patients (50%), included 15 reactions during the infusion (in 7 patients), nausea and vomiting (in 4 patients), tremors and clonus (in 4 patients), hypersomnia (in 2 patients), chest pain with dyspnea (in 2 patients), hypersensitivity (in 2 patients), and fever and chills (in 1 patient). Infusion reactions leading to discontinuation of therapy included the following: symptoms of hypersensitivity reaction, with hypotension, dyspnea, urticaria, nausea, and vomiting occurring in 1 patient during the third infusion (sixth week of therapy) and in 1 patient during the sixth infusion (eighth month of therapy); tremors and clonus in 1 patient at each infusion from the fourth to the tenth (months 4–14 of therapy); and chest pain and dyspnea in 1 patient at the third infusion. The other 3 patients who had adverse events during the drug infusions were treated symptomatically and did not withdraw from the treatment.

Development of antinuclear antibodies (ANAs) and low titers of anti–double-stranded DNA (anti-dsDNA) antibodies was observed in 2 patients (at the second month of therapy) without any clinical features of systemic lupus erythematosus. We did not suspend the treatment. The patients are now at months 12 and 16 of therapy, and the autoantibodies have disappeared. During the whole period of treatment, we recorded only a few mild, easily managed infections (urinary tract infections in 3 patients and upper respiratory tract infections in 2 patients). One patient presented with a recurrent unexplained macrohematuria and treatment was suspended.


Our results show that blockade of TNFα with the chimeric monoclonal antibody infliximab plus MTX is highly effective and safe in the short-term (2 weeks) and medium-term (12 months) treatment of patients with JIA that is long lasting and still active despite previous therapy with MTX at conventional (10 mg/m2/week) or higher (15–30 mg/m2/week) dosages and with 1 or more other conventional DMARDs. Considering the median age of our patients, we evaluated them as adult RA patients, using the ACR response criteria, the DAS28, the HAQ, and the SF-36 outcomes. Treatment with infliximab in this negatively selected population resulted in a clinically impressive and statistically significant reduction in disease activity, as assessed by a number of clinical end points and biochemical markers, and was associated with an improvement in some aspects of quality of life. Just after the first infusion, the majority of the patients in our study reported improvements in pain and morning stiffness and a reduction of fatigue and were in better general condition. After the first infusion, the mean reduction (the mean difference in individual values) in the number of joints with active disease was 36%, the mean reduction in the CRP level was 52.6%, and the mean reduction in the ESR was 42.3%.

Despite the fact that this was a very severely affected group of patients with long-lasting, refractory polyarticular JIA, 8 of 24 of them (33.3%) did not have joints with active disease at the last visit. Our results are similar to those reported in adult RA (5). In contrast with adult RA, we did not observe any significant improvement in the disability index of the HAQ. The majority of our patients were in a very advanced anatomic stage of disease; consequently, their chances of showing any significant functional improvement in a short or medium period of treatment were low. Moreover, during the study, 5 of our young patients underwent joint replacement and were therefore obliged to use some assistive devices to be able to walk, which explains why their HAQ disability index scores remained high.

Only 1 patient in our population, an 8-year-old girl, was in an acute phase of systemic disease, showing a high daily spiking fever. After the first infusion, she showed no improvement in the fever despite the higher dose of infliximab received (10 mg/kg). After the fourth infusion, she had an ACR50 response and maintained it until the seventh infusion, when she withdrew because of an adverse event (unexplained persistent hematuria). This observation contrasts somewhat with the early improvement in fever or rash reported for other patients with systemic JIA who were treated with infliximab (14, 15).

Only 50% of our patients had 1 or more adverse events, and these were usually not serious. This contrasts with results reported from a study of adult RA, in which 95% of patients had at least 1 adverse event (5).

Several blind controlled studies with both etanercept and infliximab have demonstrated a significantly higher rate of infections, including tuberculosis and opportunistic infections, in adults with RA. Varicella complications and streptococcal sepsis with disseminated intravascular coagulation have been reported in children receiving etanercept. We monitored our patients carefully for possible infections, but we recorded only a few mild infections of the upper respiratory tract or the urinary tract, which were easily managed with antimicrobial therapy.

Another concern with monoclonal antibody therapy is the potential formation of antibodies that neutralize the therapeutic agent, thereby limiting its long-term efficacy or causing allergic reactions during retreatment (16). The most common adverse events in our study were the infusion reactions, which were observed in 29% of patients. The response to human antichimeric antibodies was not measured in our patients.

TNFα blockade, either with infliximab or with etanercept, in patients with RA as well as in those with Crohn's disease, is associated with the development of autoantibodies (ANAs and/or anti-dsDNA antibodies) (16) rarely associated with the development of lupus-like syndrome. In our population, only 2 girls developed low titers of anti-dsDNA antibodies without any autoimmune disease. During followup of therapy at months 12 and 16, the autoantibodies had spontaneously disappeared.

Only 1 previous controlled clinical trial (6) and 1 prospective open observational study (7) demonstrated the clinical benefit of TNFα-neutralizing therapy with a recombinant human TNF receptor (p75)–Fc fusion protein (etanercept) in pediatric patients with active polyarticular JIA. Our study is the first to show the clinical benefit of a different TNFα-neutralizing therapy with a chimeric (human and murine) monoclonal antibody against TNFα (infliximab) in patients with JIA. Similar results for the efficacy and safety of infliximab in JIA have been described in preliminary reports by 2 other groups of investigators (17, 18). In our study, therapy with infliximab (repeated IV infusions) in combination with MTX produced rapid, marked, and sustained clinical improvement in JIA with a polyarticular course (regardless of the type of onset). The results of our study confirm that TNFα antagonism is a valid approach to treating JIA with a polyarticular course even in long-lasting, refractory, and still-active disease.

The majority of our patients were young adults with persistently active JIA (only 5 patients were of pediatric age). Since the pathogenetic mechanisms are presumably the same both in children and in young adults with JIA, we can therefore expect similar or better efficacy results with infliximab treatment in children with earlier stages of disease than in young adults with more advanced stages of disease.

In conclusion, if the significant clinical benefit and safety of infliximab observed in our pilot study could be sustained by long-term therapy, this therapeutic approach could be a valid option in patients with active polyarticular JIA who have not had an adequate response to MTX alone. A controlled multicenter clinical trial is needed.