Efficacy of intraarticular infliximab in patients with chronic or recurrent gonarthritis: A clinical randomized trial


  • ISRCTN: 17726268.



To evaluate the efficacy and safety of intraarticular infliximab compared with intraarticular methylprednisolone in patients with gonarthritis.


In 23 patients with recurrent gonarthritis despite previous intraarticular corticosteroid therapy, a total of 41 intraarticular injections (20 infliximab and 21 methylprednisolone) were performed in 28 knees. Initial therapy was randomly assigned, and crossover therapy was eligible within 3 months. The clinical effect was assessed during 6 months of followup. The primary outcome was event-free survival, defined as the time after treatment until local retreatment was performed and/or nonimprovement of the knee joint score. Adverse effects were recorded during followup.


All patients treated with intraarticular infliximab had an insufficient response. In contrast, 8 of the 21 intraarticular methylprednisolone injections were effective (P = 0.004). Between groups, no differences in the patients' age, disease duration, number of disease-modifying antirheumatic drugs, or previous intraarticular methylprednisolone were observed. Reported adverse effects were not related to therapy.


Treatment with intraarticular infliximab injection was not effective in patients with a chronically inflamed knee joint. Intraarticular injection with methylprednisolone was superior despite previous intraarticular corticosteroid therapy. Further investigation is needed to provide these patients with a better alternative.


In chronic and recurrent arthritis, the inflammation and hyperplasia of the synovial tissue results in pain and loss of function, and can cause joint damage. In the case of monarthritis, for instance of the knee, local therapy favors systemic treatment because of fewer side effects and the rapid decrease of symptoms after injection. Currently, intraarticular corticosteroids are the choice of therapy; however, the efficacy of this therapy is rather varying. Relapses of up to 50% within 6 months after injection have been reported (1), and seem dependent on the preparation used, postinjection rest, and the aspiration of joint fluid (2). Furthermore, the application of >3 injections per year in the same joint is advised against because of the risk for procedure- and steroid-associated side effects such as infection or local osteoporosis (2). An alternative local therapy, radiation synovectomy, appears to not be superior to intraarticular corticosteroids in patients with recurrent gonarthritis despite previous local therapy (3). In recent years, several (case) reports were published on the use of intraarticular infliximab (4–10), as well as intraarticular etanercept (11–13).

We performed a randomized, double-blind, optional crossover design trial comparing the efficacy and safety of intraarticular infliximab with intraarticular methylprednisolone in patients with a recurrent or relapsing gonarthritis.


Patients were recruited from the rheumatology outpatient clinic of the Leiden University Medical Centre, were age ≥18 years, and had an active inflammatory gonarthritis despite ≥1 previous intraarticular corticosteroid injection in the preceding year. Patients with arthritis due to infection, gout, or osteoarthritis were excluded. Other exclusion criteria were hemorrhagic disease; participation in any other study that interfered with or could be influenced by this study; use of oral prednisone in excess of 10 mg/day; recent change of disease-modifying antirheumatic drug (DMARD) therapy (≤6 weeks); intraarticular injection with corticosteroids <2 months ago (regardless of which joints); hypersensitivity to methylprednisolone, lidocaine, or infliximab (or other murine proteins); active or latent tuberculosis; acute or chronic infection; multiple sclerosis; decompensatio cordis (New York Heart Association classification III and IV); pregnancy or lactation; and malignancy. According to the guidelines for treatment with systemic anti–tumor necrosis factor α (anti-TNFα) therapy (14), all patients were screened for (latent) tuberculosis, (chronic) infections, and malignancies by the patient's history, physical examination, chest radiograph result, and subcutaneous tuberculin skin test. All patients with an indication for tuberculosis, infection, and/or malignancy (by history, examination, radiograph, and/or tuberculin skin test) were excluded from further participation. The ethics committee approved the study protocol and all participants gave written informed consent.

The affected knees were randomized to receive either intraarticular infliximab 100 mg or intraarticular methylprednisolone 80 mg. In the optional crossover part of the trial, patients were eligible for a second injection (with the other study compound) if the gonarthritis recurred within 3 months. Patients were allowed to use acetaminophen, nonsteroidal antiinflammatory drugs, and other analgesics in a stable dose regimen. DMARD therapy was preferably kept stable, but if a (dose) change of a DMARD occurred, this was not interpreted as nonresponsiveness unless there was a recurrence of arthritis in the injected knee.

In the current prospective, randomized, double-blind, single-center, optional crossover design study, we tested the hypothesis that intraarticular infliximab would be superior to intraarticular methylprednisolone in patients with recurrent or chronic gonarthritis.

Patients were clinically evaluated at 1, 3, and 6 months after the intraarticular injection. The primary outcome measure was event-free survival, defined as the time after treatment until local retreatment (joint aspiration or injection, arthroscopy, or [radio-] synovectomy) was performed because of recurrence or persistence of gonarthritis, and/or nonimprovement of the knee joint score assessed by trained research nurses. The knee joint score (range 0 [minimum] to 7 [maximum] points) encompasses knee tenderness as scored from 0–3 (where 0 = no tenderness, 1 = tenderness when asked, 2 = tenderness on pressure, and 3 = tenderness and wincing) (15), knee swelling as scored from 0–3 (where 0 = no swelling, 1 = little swelling, 2 = moderate swelling, and 3 = abundant swelling) (16), and a patient visual analog scale (VAS; range 0–100, where 0 = no pain and 100 = maximal pain) for knee pain as scored from 0–1 (VAS score divided by 100).The occurrence of side effects was recorded during 6 months of followup.

We have calculated a sample size of 20 affected knees per treatment arm in order to find a statistically significant difference in (non)responders at 6 months. This is based on the assumption of a Type I error of 5% (2-sided), a power of 90%, a difference in response of 25% after 6 months, and a dropout rate of 20%. A difference in response of 25% is considered clinically relevant.

We randomized affected knees according to a computer-generated list, divided in blocks of 8 to prevent a skewed randomization in the case of incomplete inclusion. Randomization was balanced for every 8 injected knees that were included.

The patient and treating rheumatologist, as well as the assessor who determined the clinical parameters, were blinded to the injected medication. The intraarticular injection was given by an independent investigator who was not blinded, because the colors of the 2 medications were different. In all patients, intraarticular fluid was removed by aspiration as much as possible immediately before injection of the study drug.

Nonparametric Mann-Whitney U tests were used to compare baseline characteristics between the infliximab and methylprednisolone groups. Chi-square tests were used to compare the proportions of patients between both treatment arms in which intraarticular injection was effective. Kaplan-Meier survival analyses were used to compare the time to failure between the infliximab and methylprednisolone groups. P values less than 0.05 were considered significant.



Twenty-seven patients were screened during the time period November 2004 to June 2006. Four patients were excluded because of a positive Mantoux test result (n = 2), possible malignancy (n = 1), or withdrawal of informed consent (n = 1), resulting in 23 patients who received ≥1 therapeutic injection. At baseline, these patients (52% male) had a mean ± SD age of 51.4 ± 11.8 years, with a mean ± SD disease duration of 10.6 ± 10.6 years. The underlying diseases were undifferentiated arthritis (n = 8), rheumatoid arthritis (RA; n = 7), psoriatic arthritis (PsA; n = 5), spondylarthropathy (n = 2), and juvenile chronic arthritis (n = 1). Fourteen patients did not receive antirheumatic drugs, and the other patients received methotrexate (MTX; n = 3), sulfasalazine (n = 2), leflunomide (n = 1), combination MTX plus hydroxychloroquine (n = 1), combination MTX plus sulfasalazine and hydroxychloroquine (n = 1), and MTX combined with adalimumab (n = 1).

In the 23 treated patients, a total of 28 knees were injected (Figure 1). In addition to 15 knees injected once, 13 knees received 2 injections (crossover part of the trial). Therefore, a total of 41 intraarticular injections were administered, subdivided into 20 infliximab and 21 methylprednisolone injections.

Figure 1.

Flow of participation: numbers of patients, numbers of therapeutic injections, randomization, and clinical outcome. * Accidental protocol violation.


All patients treated with intraarticular infliximab received retreatment within 6 months because of an insufficient response (either nonimprovement of the knee joint score or recurrence of the gonarthritis) (Figure 1). Seven of the 13 first injections with methylprednisolone were successful, without a relapse before 6 months of followup. The other 6 first methylprednisolone injections, as well as 7 of the 8 crossover (after failure or relapse with intraarticular infliximab) injections, were ineffective: 1 patient had a persistently high knee joint score, and 12 times a relapse occurred that required local therapy. The clinical characteristics of the patients treated with infliximab versus methylprednisolone (compared per injection) showed no difference in the patients' age, disease duration, number of DMARDs, number of previous intraarticular methylprednisolone injections, or clinical disease activity as measured by the knee joint score (Table 1).

Table 1. Clinical characteristics of patients classified per injection*
 Infliximab (n = 20)Methylprednisolone (n = 21)
  • *

    Values are the mean. No significant differences for all parameters. DMARDs = disease-modifying antirheumatic drugs.

Age, years51.253.4
Disease duration, years9.410.4
Number of DMARDs1.71.6
Number of previous intraarticular corticosteroid injections2.22.3
Knee joint score at the time of inclusion4.14.2

Overall, the efficacy of methylprednisolone was superior to infliximab (P = 0.004) (Tables 2 and 3). Figures 2 and 3 show the duration of response, represented as arthritis-free survival after the injections.

Table 2. Outcome of the 41 injections after 6 months of followup
 Infliximab (n = 20)Methylprednisolone (n = 21)
  • *

    P = 0.004.

Sufficient response, no. (%)08 (38)
Insufficient response, no. (%)20 (100)*13 (62)*
Table 3. Outcome of the 28 first injections
 Infliximab (n = 15)Methylprednisolone (n = 13)
Sufficient response, no. (%)07 (54)
Insufficient response, no. (%)15 (100)6 (46)
Figure 2.

Kaplan-Meier curve: time to failure (all 41 injections; intent-to-treat).

Figure 3.

Kaplan-Meier curve: time to failure (28 first injections).


Local adverse reactions after intraarticular administration did not occur with either intervention. In 7 patients, nonlocal side effects were reported. Four patients had a side effect after receiving intraarticular infliximab, including skin disorder (after 180 days), epileptic insult in a patient who had had a cerebrovascular accident in the past (after 79 days), cough (after 38 days), and hypercholesterolemia (after 171 days). Three patients receiving intraarticular methylprednisolone reported a side effect, including ablatio retinae (after 126 days), sinusitis (after 113 days), and blurred vision, mood disorder, and skin disorder (after 149 days). None of the reported side effects were judged to be related to treatment.


Our study demonstrated that intraarticular infliximab is inferior to intraarticular methylprednisolone in patients with a recurrent/chronic gonarthritis despite previous intraarticular corticosteroid injections. No patient treated with intraarticular infliximab as a first injection had a persistently good clinical response. In contrast, 6 of 13 initial injections with intraarticular corticosteroids were effective without relapse during 6 months of followup. In the crossover part of the study, 7 of the 8 patients who had previously failed or relapsed after a first injection with infliximab also failed or relapsed after a second injection with methylprednisolone, and all 4 patients who crossed over from a first injection with corticosteroids to infliximab failed or had an early relapse. There appeared to be no differences in safety.

TNFα plays an important role in chronic inflammation in patients with RA and PsA (17, 18), and in synovial biopsy findings, increased expression of TNFα has been demonstrated (19). Consequently, TNFα-blocking agents reduced disease activity effectively in patients with RA and PsA (20, 21). Infliximab is a chimeric antibody that blocks TNFα and is administered intravenously. In theory, local administration of infliximab into an inflamed joint could be very successful. Several (case) reports of intraarticular infliximab with mostly good clinical outcomes have supported this theory (4–10). In contrast to our study, these reports were not double-blinded and did not directly compare infliximab with corticosteroids. Also, in most of these reports, the followup period was shorter than 6 months. The only reported double-blinded study comparing intraarticular methylprednisolone with intraarticular etanercept showed results similar to ours (11).

It could be argued that the lack of efficacy of intraarticular infliximab was the result of low infliximab concentration in the joint. However, in accordance with the previous successfully treated case reports, we used a dose of 100 mg of infliximab per injection (4, 7, 9, 10). Also, on basic assumption, the administration of infliximab 100 mg in the knee joint should result in a therapeutic range compared with the effective dose of intravenous therapy (22).

Given the high rate of failure, especially after the second injection in the crossover part of the trial (with either infliximab or methylprednisolone), there is a strong suggestion that in many cases the inflammation had already accelerated and would not respond to local treatment. Af Klint et al demonstrated that after intraarticular corticosteroids, a decline of TNFα occurred in synovial biopsy findings, but that macrophage infiltration and proinflammatory endothelial cytokine expression remained unchanged (23). The rate of success of intraarticular methylprednisolone after initial intraarticular infliximab was lower compared with initial intraarticular methylprednisolone (1 [14%] of 7 versus 7 [54%] of 13). The numbers are small, which makes comparison between the groups difficult. If there is a real difference in response, we can only speculate about the reason. Blocking TNFα may result in increased activity of other proinflammatory pathways less responsive to corticosteroids.

We conclude that intraarticular infliximab injection into chronically inflamed knee joints is insufficiently effective and especially not more effective than intraarticular injection with methylprednisolone. Given the fact that intraarticular methylprednisolone also has limited long-term efficacy, further investigations are needed to provide patients with a better alternative.


All authors were involved in contributions to study conception and design, acquisition of data, or analysis and interpretation of data, and drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. van der Bijl 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.


Schering-Plough BV and Centocor, Inc., were not involved in the study design, data collection, data analysis, or writing of the manuscript. Publication of this article was not contingent on the approval of Schering-Plough BV and Centocor, Inc.