Many clinical studies have reported the excellent clinical efficacy of infliximab (IFX), an antitumour necrosis factor α (anti-TNFα) monoclonal antibody, in the treatment of rheumatoid arthritis (RA).1 IFX is also reported to induce a rapid and marked reduction in circulating interleukin 6 (IL-6) levels, suggesting that its efficacy may result from the suppression of IL-6 as well as TNF.2,–,5 In the RISING Study (NCT00691028),6 ,7 we observed patients who showed no response to IFX therapy, despite maintaining a serum IFX level higher than the threshold level for clinical response. Here, we examined data on clinical response to better understand the mechanism of action of IFX.

In this study, patients with methotrexate-refractory RA treated with 3 mg/kg of IFX at weeks 0, 2 and 6 were randomly assigned to receive 3, 6 or 10 mg/kg of IFX every 8 weeks from week 14 to 46 in combination with methotrexate. A total of 271 patients were classified into four groups based on serum IFX, which was used as a surrogate marker of TNF suppression since the TNF level could not be accurately measured in the presence of IFX, and plasma IL-6 levels at week 54 (Group 1, IFX-high/IL-6-low; Group 2, IFX-high/IL-6-high; Group 3, IFX-low/IL-6-low; Group 4, IFX-low/IL-6-high), with cut-off values of IFX and IL-6 of 1.0 μg/ml6 ,8 ,9 and 10 pg/ml (lower 80% confidence limit with one tail at baseline), respectively.

The median IL-6 level at baseline was 28.9 pg/ml, which rapidly decreased at week 10 and remained at 2.4 pg/ml at week 54 with a median suppression rate of 87.2%.

Clinical characteristics such as TNF and C reactive protein (CRP) levels and rheumatoid factor (RF) values showed significant differences among groups at baseline (table 1). At week 54, however, significant differences were observed in each of the disease activity indices, with the lowest disease activity in Group 1, intermediate in Groups 2 and 3 and highest in Group 4. In particular, remission rates were significantly higher in Group 1 (figure 1). A difference was also observed in the improvement of physical function. By contrast, the progression of joint damage was inhibited in the majority of patients, with no significant difference between groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Efficacy of infliximab therapy at week 54 among the four groups classified by serum infliximab and plasma interleukin 6 levels at week 54. Clinical responses were analysed by disease activity score 28 (DAS28)-erythrocyte sedimentation rate (ESR) (A), DAS28-C reactive protein (CRP) (B), simplified disease activity index (SDAI) (C) and clinical disease activity index (CDAI) (D) at week 54. HDA, high disease activity; MDA, moderate disease activity; LDA, low disease activity but no remission. Cut-off values for disease activity in DAS28-CRP were used as reported by Inoue et al (HDA: >4.1; MDA: ≥2.7 to ≤4.1; LDA: ≥2.3 to <2.7; remission: <2.3).10

We speculate that the efficacy of treatment seen in Group 1 resulted from the simultaneous suppression of TNF and IL-6, that in Group 2 from the direct action of TNF neutralisation and that in Group 3 via IL-6 suppression. TNF served as a main modulator of IL-6 in Group 1, as proposed by Feldman et al,2 but exerted only partial modulation in Group 2. RF and anticyclic citrullinated peptide antibody titres were lower in Groups 1 and 2 (IFX-high) than in Groups 3 and 4 (IFX-low). Meanwhile, CRP and matrix metalloproteinase 3 levels were lower in Groups 1 and 3 (IL-6-low) than in Groups 2 and 4 (IL-6-high). However, we found no clear difference in evaluations when using the simplified disease activity index, which includes acute-phase reactants, or the clinical disease activity index, which does not include these reactants. We also analysed the proportion of primary or secondary non-responders in each group, as defined previously,7 and found no obvious difference indicating a primary or secondary failure pattern.

Several limitations of our study warrant mention. First, we set the cut-off values for IFX and IL-6 levels ourselves because there are no standard reference values. Second, serum IFX level was used as a surrogate marker of TNF suppression, which cannot be accurately measured in the presence of IFX. Third, we had no data on other cytokines, so we could not explore their influence on clinical response.

This study demonstrates that the profound suppression of TNF and IL-6 in IFX treatment results in better clinical response, in addition to clinical remission.