Interleukin-6 receptor inhibition with tocilizumab and attainment of disease remission in rheumatoid arthritis: The role of acute-phase reactants


  • Josef S. Smolen,

    Corresponding author
    1. Medical University of Vienna and Hietzing Hospital, Vienna, Austria
    • Division of Rheumatology, Department of Medicine III, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria
    Search for more papers by this author
    • Dr. Smolen has received consulting fees, speaking fees, and/or honoraria from Abbott, Amgen, Bristol-Myers Squibb, Centocor, Schering-Plough, Roche, UCB, and Wyeth (less than $10,000 each).

  • Daniel Aletaha

    1. Medical University of Vienna, Vienna, Austria
    Search for more papers by this author
    • Dr. Aletaha has received consulting fees, speaking fees, and/or honoraria from Bristol-Myers Squibb, Roche, UCB, Abbott, Wyeth, and Schering-Plough (less than $10,000 each).



To determine the effects of tocilizumab on rheumatoid arthritis (RA) disease activity and remission assessment, using measures that do or do not comprise acute-phase reactants.


Simplified Disease Activity Index (SDAI) scores, Clinical Disease Activity Index (CDAI) scores, and the Disease Activity Score in 28 joints (DAS28) were calculated using data from tocilizumab trials in patients with RA in whom disease had remained active despite treatment with disease-modifying antirheumatic drugs. The CDAI does not contain an acute-phase reactant component. Disease activity states, including remission, were defined using established cut points; for the DAS28, an alternative cut point of <2.4 was also used.


Changes in the DAS28, the SDAI score, and the CDAI score among patients receiving tocilizumab were significantly higher than those among patients receiving placebo, and the magnitude of these changes was similar for the SDAI and the CDAI. Among patients who achieved 50% improvement in disease activity according to the American College of Rheumatology criteria, only ∼20% required a reduction in acute-phase reactant values in order to fulfill the criteria. However, DAS28 remission rates were higher (even when using the lower cut point) than the SDAI and CDAI remission rates. Only a minority of tocilizumab-treated patients with DAS28 remission also had disease remission according to the SDAI (26%) or CDAI (∼21%). With infliximab treatment, SDAI and CDAI remission rates were of the same magnitude as those observed with tocilizumab treatment, and DAS28 remission rates were lower. Tocilizumab-treated patients with DAS28 remission but without CDAI remission had significantly higher swollen joint counts but lower erythrocyte sedimentation rates (ESRs) compared with patients with SDAI or CDAI remission.


Disease activity in RA is reduced by tocilizumab treatment, irrespective of the type of composite measure used to evaluate disease activity. Remission rates were much higher using the DAS28 compared with the SDAI and CDAI, due to the high weight of the ESR in the DAS28 and the effect of tocilizumab on the ESR. Using the stringent SDAI and CDAI criteria, however, remission rates in patients treated with tocilizumab were in the same range as those seen in patients treated with tumor necrosis factor inhibitors.

Tumor necrosis factor (TNF), interleukin-1 (IL-1), and IL-6 play important roles in the common final pathways leading to joint inflammation as well as cartilage and bone destruction in rheumatoid arthritis (RA) (1–4). Along with many other cellular and molecular consequences, the acute-phase response is activated by these cytokines (5). Therefore, inhibition of any of these cytokines leads to a rapid reduction in the levels of acute-phase reactants, including the prototypic C-reactive protein (CRP) (6–9). CRP production in the liver is induced primarily, if not exclusively, by IL-6. TNF and IL-1, while also partly activating other acute-phase reactants, appear to mediate their effects on CRP via induction of IL-6 (2, 5, 10–13). The expression of acute-phase reactants is greatly increased in active RA. Both at the initiation of therapy as well as cumulatively over time, the levels of acute-phase reactants are correlated with clinical disease activity (14) and joint destruction (15–17).

Tocilizumab, a humanized monoclonal antibody to the IL-6 receptor (IL-6R), inhibits the binding of IL-6 to its receptor, preventing IL-6–mediated signaling (2), which is efficacious in RA (9, 18, 19). However, although head-to-head studies are not available, the observation that tocilizumab, in contrast to other biologic agents including TNF inhibitors, normalizes mean CRP values in RA clinical trials on the group level (9, 18, 19) suggests that blocking IL-6 conveys the strongest improvement in acute-phase reactant levels currently attainable. The pronounced effects of IL-6R inhibition on acute-phase reactant levels and the fact that acute-phase reactants are part of disease activity measures and response criteria (20) elicit the question of whether the direct biologic effects of tocilizumab on the reduction of acute-phase reactant levels might lead to an overestimation of the response to treatment with tocilizumab. This effect may particularly affect changes in disease activity indices that include acute-phase reactants, such as the Disease Activity Score in 28 joints (DAS28) (21) or the Simplified Disease Activity Index (SDAI) (20), especially in terms of defining remission, which constitutes a major therapeutic target (22).

Given this background, we used data from clinical trials of tocilizumab to evaluate the contribution of reduced acute-phase reactant concentrations to achieving disease remission.


Data sets.

The trial sponsor (Roche) kindly provided an 80% random sample of patient-level data from 3 tocilizumab clinical trials in patients with RA that remained active despite methotrexate (MTX) therapy (the OPTION trial [9] and the LITHE study [23]) or despite treatment with MTX or other disease-modifying antirheumatic drugs (DMARDs) (the TOWARD [Tocilizumab in Combination With Traditional DMARD Therapy] study [18]). In these trials, tocilizumab was used at doses of 4 mg/kg and/or 8 mg/kg and was compared with placebo (all in combination with MTX or other DMARDs). We assessed only those patients for whom all variables needed to calculate composite disease activity scores were available at all relevant time points. Some evaluations were done using the pooled 4 mg/kg tocilizumab arms of the OPTION and LITHE trials (n = 317) and the pooled 8 mg/kg tocilizumab and pooled tocilizumab/placebo arms of all 3 trials (n = 891 and n = 473, respectively) after 6 months of treatment.

For some analyses, we also used data kindly provided by Centocor from the ATTRACT (Anti–Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy) trial (80% random sample; pooled infliximab arms) of infliximab plus MTX compared with placebo plus MTX after 6 months of treatment in patients with active disease despite MTX therapy (24). No statistical comparisons between these biologic therapies were performed, because the ATTRACT trial was performed many years before the tocilizumab trials, and the study populations may not be comparable (25). The comparison with infliximab and the comparison of the different indices each served the purpose of providing respective information on remission attainment on these agents.

Outcome measures.

We determined the DAS28, the SDAI score, and the Clinical Disease Activity Index (CDAI) score, using the following equations: DAS28 = 0.56 × √(TJC28) + 0.28 × √(SJC28) + 0.70 × ln(ESR) + 0.014 × GH; SDAI score = SJC28 + TJC28 + PGA + EGA + CRP; and CDAI score = SJC28 + TJC28 + PGA + EGA, whereby the tender joint count (TJC) and the swollen joint count (SJC) were evaluated using 28 joints, global health (GH) was customarily replaced by the patient's global assessment (PGA) of disease activity (in millimeters on a visual analog scale [VAS]), scores for the PGA and the evaluator's global assessment (EGA) were measured in centimeters on a 0–100-mm VAS, the erythrocyte sedimentation rate (ESR) was measured as mm/hour, and the CRP level was measured as mg/dl.

For all 3 composite measures, cut points for the disease activity states of remission (REM), low disease activity (LDA), moderate disease activity (MDA), and high disease activity (HDA) were determined (21, 26). These established cut points are as follows: for the DAS28, REM <2.6 ≤ LDA ≤3.2 < MDA <5.1 ≤ HDA; for the SDAI, REM ≤3.3 < LDA ≤11 < MDA ≤26 < HDA; for the CDAI, REM ≤2.8 < LDA ≤10 < MDA ≤22 < HDA. For some analyses, we also used a DAS28 remission cut point of <2.4, which had been derived using the same criteria as that for remission defined by the SDAI (3.3) and the CDAI (2.8) (20, 26).

Evaluation of composite disease activity measures.

We assessed and compared the power of the SDAI and the CDAI to discriminate the effects of tocilizumab at doses of 4 mg/kg and 8 mg/kg from the effects of placebo. The SDAI and CDAI differ only with regard to the inclusion or exclusion of the CRP level. In addition, we also calculated DAS28 values for comparative purposes. For each of the 3 tocilizumab trials, we determined the SDAI score, the CDAI score, and the DAS28 at baseline and the mean changes from baseline every month thereafter for 6 months, for the placebo group and the 2 tocilizumab groups.

Discriminative power (comparing active drug with comparator).

We first assessed how the use of different composite disease activity measures, one of which (the CDAI) does not comprise an acute-phase reactant, influenced the power to discriminate treatment-related changes. We therefore calculated the T scores and P values from a t-test, comparing changes in the active-treatment arms (4 mg/kg or 8 mg/kg of tocilizumab) with those in the placebo group, using the DAS28, the SDAI, or the CDAI. We also used the chi-square test to compare the proportions of patients achieving disease remission while receiving active treatment or placebo, according to the different scoring methods.

Impact on remission of reducing acute-phase reactant values.

To test whether the relative therapeutic effect on the clinical components differed from that on the acute-phase reactant components of the 3 composite measures when the state of remission was attained, we determined the values of core set variables in the setting of disease remission as defined by the 3 scoring methods and compared these values using t-tests. Because the established cut point for DAS28 remission (<2.6) was not derived using the same data as those used for deriving the SDAI remission cut point, we also used a DAS28 cut point of <2.4 in some analyses (20, 26, 27). In addition, for each of the composite measures, we calculated the percent change from baseline of values for the respective clinical and acute-phase reactant components.

Composition and weighting of the index components.

Because the ESR has been shown to be relatively highly weighted in the DAS28, while the CRP level has a low weight in the SDAI (28, 29), an exaggerated DAS28 remission rate compared with SDAI and CDAI remission rates could be hypothesized to be indicative of an excessive effect of tocilizumab on the ESR. To evaluate whether discordant remission rates were attributable to effects on acute-phase reactants, we identified all patients with DAS28 remission and compared the differences in core set measure values between those who also had disease remission according to the CDAI and those with low or moderate disease activity according to the CDAI. We used a t-test or analysis of variance (ANOVA) for these comparisons, depending on the number of groups.

To further assess the impact of acute-phase reactants on clinical response assessment, we next looked at the American College of Rheumatology (ACR) definitions of improvement in RA (30), which are traditionally used as the primary and main secondary end points in trials and require a defined percent reduction in the TJC and SJC plus the same reduction in at least 3 of 5 additional variables. We used the ACR definition of 50% improvement (ACR50) as a paradigmatic response that is more meaningful than just the primary end point, the ACR20. We determined the number of patients fulfilling all 5, 4, or 3 of the additional criteria (30) in the pooled tocilizumab 8 mg/kg arms. For informative purposes, we also assessed these data in the ATTRACT trial (24). In patients fulfilling only 3 of the 5 additional criteria, the positive response classification hinges on any 1 of the variables represented among those 3; we identified how frequently the acute-phase reactants (ESR and/or CRP) were represented in the groups fulfilling only 3 of the 5 additional criteria.

Statistical analysis.

All analyses were performed using the SAS package, version 9.1.3.


The baseline characteristics of the pooled populations of study patients are shown in Table 1. When we evaluated the CDAI, a measure that does not include an acute-phase reactant, significant improvement in disease activity was seen as soon as 2 weeks after the start of tocilizumab at 8 mg/kg compared with placebo in the OPTION trial (Figure 1B), with similar changes in the 4 mg/kg group from 8 weeks onward in the other 2 trials (results not shown). This improvement increased consistently and was significantly different from that in the placebo group at almost all evaluated time points (P < 0.05 to P < 0.0001). Importantly, the changes in the SDAI were similar (Figure 1A), although values were slightly higher due to the inclusion of CRP in the formula. Similar to the CDAI, the differences from placebo were significant for both tocilizumab groups (P < 0.05 to P < 0.0001).

Table 1. Baseline demographic characteristics and disease activity measures of patients in the tocilizumab and infliximab trials*
 Tocilizumab/placebo, pooled (n = 473)Tocilizumab 4 mg/kg, pooled (n = 317)Tocilizumab 8 mg/kg, pooled (n = 891)Infliximab/placebo (n = 50)Infliximab 3 mg/kg, pooled (n = 213)
  • *

    Except where indicated otherwise, values are the mean ± SD. RF = rheumatoid factor; RA = rheumatoid arthritis; HAQ = Health Assessment Questionnaire; SDAI = Simplified Disease Activity Index; CDAI = Clinical Disease Activity Index; DAS28 = Disease Activity Score in 28 joints.

Age, years52.2 ± 13.052.5 ± 12.152.7 ± 12.349.7 ± 10.952.7 ± 11.3
Female sex, %84.682.482.480.077.0
RF positive, %75.582.
RA duration, years9.0 ± 8.48.8 ± 7.79.3 ± 8.410.1 ± 8.310.3 ± 8.0
HAQ score1.5 ± 0.61.5 ± 0.61.5 ± 0.61.7 ± 0.61.7 ± 0.6
SDAI score42.6 ± 13.242.7 ± 12.244.7 ± 14.245.0 ± 14.146.6 ± 14.6
CDAI score40.4 ± 12.340.4 ± 11.942.1 ± 13.641.6 ± 13.242.8 ± 13.3
DAS286.6 ± 0.96.6 ± 0.96.7 ± 1.06.6 ± 1.06.7 ± 1.0
Figure 1.

Changes from baseline in the Simplified Disease Activity Index (SDAI) score (A), the Clinical Disease Activity Index (CDAI) score (B), and the Disease Activity Score in 28 joints (DAS28) (C) in the OPTION trial. Baseline values in the placebo, 4 mg/kg tocilizumab (TCZ), and 8 mg/kg tocilizumab groups, respectively, were 43.6, 42.9, and 43.8 for the CDAI; 46.0, 45.6, and 46.6 for the SDAI; and 6.8, 6.8, and 6.9 for the DAS28.

At 6 months, for the placebo group, the 4-mg/kg group, and the 8-mg/kg group, the mean ± SD CDAI scores were 25.1 ± 13.3, 19.0 ± 15.1, and 17.0 ± 13.2, respectively, and the mean ± SD SDAI scores were 27.1 ± 14.3, 20.1 ± 15.6, and 17.2 ± 13.3, respectively; the percent changes from baseline to the end point were 42 ± 29%, 55 ± 38%, and 61 ± 28% for the CDAI and 40 ± 29%, 55 ± 37%, and 63 ± 27% for the SDAI. The observation that the CDAI end point values and changes were quite similar to those for the SDAI suggests that the efficacy of tocilizumab is primarily related to its reduction of values for clinical features such as joint counts and global assessment scores, which are components of the CDAI. Moreover, the data show that the discriminative power of a composite disease activity index is not necessarily inflated by the inclusion of an acute-phase reactant in patients treated with tocilizumab, at least as long as the acute-phase response is not overweighted, such as in the SDAI (29).

For completeness, changes in the DAS28 for the population in the current study are also shown (Figure 1C). The mean ± SD changes in the DAS28 from baseline to 6 months were 17 ± 21%, 21 ± 16%, and 47 ± 22% for the placebo group, the 4-mg/kg tocilizumab group, and the 8-mg/kg tocilizumab group, respectively.

An ACR50 response was observed in 39.8% of the pooled patients treated with tocilizumab 8 mg/kg and 35.2% of the pooled infliximab-treated patients. Among these patients, the majority (∼80% of tocilizumab-treated patients and ∼65% of infliximab-treated patients) fulfilled 4 or all 5 of the additional ACR core set criteria used for determination of the ACR50 response. Thus, proportionally somewhat more patients receiving tocilizumab than those receiving infliximab did not need the critical number of 3 fulfilled variables and, therefore, would have achieved an ACR50 response irrespective of a reduction in acute-phase reactant values. Of those patients meeting just 3 of the 5 criteria (20.3% of ACR50 responders receiving tocilizumab and 34.7% of those receiving infliximab), a reduction in acute-phase reactant values was observed in 97% of those receiving tocilizumab and 58% of those receiving infliximab. Thus, exactly 20% of all ACR50 responders receiving tocilizumab or infliximab required a reduction in acute-phase reactant values. Similar data were obtained for the ACR20 and ACR70 responses (data not shown).

Determination of disease activity states including remission.

For this analysis, we pooled the placebo and tocilizumab 8 mg/kg arms of all 3 trials and the tocilizumab 4 mg/kg arms of the LITHE and OPTION studies and also separately evaluated the infliximab trial. In line with the inclusion criteria for the 3 tocilizumab trials, patients had high or moderate disease activity at baseline (Table 1). Disease activity states as observed after 6 months of therapy are shown in Table 2. Disease remission rates according to the SDAI criteria were quite similar for patients receiving tocilizumab 8 mg/kg and those receiving infliximab (∼7%) (Table 2). Slightly fewer patients in the 8 mg/kg group fulfilled the CDAI remission criteria compared with the SDAI remission criteria (6.4% versus 7.7%). However, compared with SDAI and CDAI remission, DAS28 remission was attained by almost twice as many patients receiving infliximab therapy and 4 times as many receiving tocilizumab at a dose of 8 mg/kg.

Table 2. Disease activity states as defined by the SDAI, CDAI, and DAS28 in pooled patient populations at 6 months*
GroupRemissionLow disease activityModerate disease activityHigh disease activity
  • *

    Values are the percent of patients. SDAI = Simplified Disease Activity Index; CDAI = Clinical Disease Activity Index; DAS28 = Disease Activity Score in 28 joints.

Tocilizumab/placebo (n = 473)    
 DAS28 <
 DAS28 <
Tocilizumab 4 mg/kg (n = 317)    
 DAS28 <2.616.413.941.328.4
 DAS28 <2.412.917.441.328.4
Tocilizumab 8 mg/kg (n = 891)    
 DAS28 <2.629.616.338.515.6
 DAS28 <2.425.920.038.515.6
Infliximab/placebo (n = 50)    
 DAS28 <2.606.030.064.0
 DAS28 <2.406.030.064.0
Infliximab 3 mg/kg (n = 213)    
 DAS28 <2.613.210.342.733.8
 DAS28 <2.410.812.742.733.8

Thus, although tocilizumab and infliximab appeared to convey numerically similar SDAI and CDAI remission rates, the differences between DAS28 remission in these 2 groups suggested that tocilizumab might particularly increase remission rates as assessed by the DAS28. This observation was supported even when a more stringent DAS28 remission cut point (<2.4) corresponding to the SDAI and CDAI remission criteria (26) was used (Table 2). Therefore, we more closely examined the individual core set variables.

Discordance of core set variable values in remission classification by different composite measures.

We analyzed the values of core set variables in patients with disease in remission according to the various composite instruments. Values especially for the mean SJC and the patient's global assessment but also for several other variables were significantly higher in patients with DAS28 remission than in those with SDAI or CDAI remission (Table 3), which is consistent with the low weight that the SJC and the patient's global assessment receive in the DAS28 (21, 28). Among patients with disease in remission according to both the SDAI and the CDAI, the mean and median relative changes from baseline in the various core set variables were >95% for the SJC, the TJC, and the global assessments (Table 3). Relative changes in the CRP level were ∼92% for SDAI remission but were also close to that value (∼89%) for CDAI remission. These data indicate that the relative changes in the CRP level paralleled and were similar to the clinical changes; although SDAI remission rates were slightly higher than CDAI remission rates, use of the SDAI did not appear to increase remission rates, essentially despite including the CRP level in its formula.

Table 3. Mean values and relative changes from baseline for the core set variables among patients achieving remission according to the DAS28, SDAI, or CDAI*
VariablePatients achieving remissionPercent change from baseline
DAS28 (n = 270)SDAI (n = 71)CDAI (n = 59)DAS28SDAICDAI
  • *

    Values are the mean ± SD. SDAI = Simplified Disease Activity Index; CDAI = Clinical Disease Activity Index; PGA = patient's global assessment; VAS = visual analog scale; EGA = evaluator's global assessment; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; HAQ = Health Assessment Questionnaire.

  • P < 0.0001 versus Disease Activity Score in 28 joints (DAS28).

  • P < 0.05 versus DAS28.

Swollen joint count2.0 ± 2.70.2 ± 0.50.2 ± 0.580.9 ± 27.097.1 ± 7.598.1 ± 6.4
Tender joint count1.0 ± 1.50.3 ± 0.50.2 ± 0.489.9 ± 17.195.6 ± 13.896.9 ± 13.4
PGA, cm on VAS1.9 ± 1.80.6 ± 0.60.5 ± 0.696.3 ± 5.1898.6 ± 1.998.6 ± 2.1
EGA, cm on VAS1.3 ± 1.10.5 ± 0.50.5 ± 0.597.6 ± 2.3499.0 ± 1.199.1 ± 1.0
ESR, mm/hour4.6 ± 4.37.1 ± 9.38.5 ± 10.785.4 ± 21.179.9 ± 29.376.2 ± 32.4
CRP, mg/dl0.1 ± 0.50.1 ± 0.30.2 ± 0.688.7 ± 33.791.8 ± 15.789.4 ± 22.5
HAQ0.6 ± 0.60.5 ± 0.40.5 ± 0.553.8 ± 45.066.4 ± 31.866.6 ± 33.0

In accordance with the absolute values, patients with DAS28 remission had lower relative changes in joint counts, especially the SJCs, but somewhat higher changes in the ESR when compared with patients with SDAI and CDAI remission. Because the SJC receives only low weight while the ESR is weighted highly in the DAS28 formula (28), we next addressed the question of whether the effect of tocilizumab on the ESR could lead to high rates of DAS28 remission and simultaneously allow for high residual SJCs in the setting of disease remission.

Analysis of core set measures in patients with discrepancies in remission rates according to different indices.

Among tocilizumab-treated patients with DAS28 remission, ∼75–80% did not have disease remission according to the CDAI or SDAI (Figure 2), while most patients who achieved SDAI or CDAI remission (i.e., ∼95%) also achieved DAS28 remission (data not shown). Among patients treated with infliximab, 50% of those who achieved DAS28 remission did not achieve SDAI or CDAI remission. Interestingly, a considerable proportion of patients who achieved DAS28 remission had moderate disease activity according to the SDAI or CDAI (Figure 2), suggesting the presence of significant residual clinical disease activity.

Figure 2.

Disease activity status as defined by the CDAI and SDAI in patients who achieved DAS28 remission (REM) with tocilizumab or infliximab (INF). LDA = low disease activity; MDA = moderate disease activity (see Figure 1 for other definitions).

For further elucidation, we analyzed the mean values for core set variables in patients with DAS28 remission who also had remission, low disease activity, or moderate disease activity according to the CDAI (Figure 3). The values for all of these variables increased with each CDAI category despite all patients having DAS28 remission (P < 0.0001, by ANOVA), with the exception of the CRP (results not shown). Importantly, the SJC reached high-average levels in patients in DAS28 remission who had moderate disease activity according to the CDAI (mean SJC ∼5); moreover, the patients also considered themselves to have worse disease activity, with a mean score of ∼30 mm for the patient's global assessment among those with CDAI-defined moderate disease activity. This is consistent with the low weight of both the SJC and the patient's global assessment in the DAS28 formula (28).

Figure 3.

Levels of core set variables of disease activity in patients with DAS28 remission who also had disease remission (REM), low disease activity (LDA), or moderate disease activity (MDA) according to the CDAI. As CDAI-defined disease activity worsened, the values for most variables increased despite DAS28 remission. Only the erythrocyte sedimentation rate (ESR) became lower as CDAI-defined disease activity worsened, although the ESR was always in the low-normal range and was far lower than the upper limit of normal (∼20 mm/hour). HAQ = Health Assessment Questionnaire; TJC = tender joint count; SJC = swollen joint count (see Figure 1 for other definitions).

Interestingly, and contrasting all clinical variables, the ESR was the only variable that significantly decreased in value as CDAI-defined disease activity worsened (P < 0.0001, by ANOVA) (Figure 3). Although patients with both DAS28 remission and CDAI remission had a mean ESR of ∼6 mm/hour, the ESR was ∼2 mm/hour in patients with CDAI-defined moderate disease activity. Also, when directly comparing the ESRs in patients with both DAS28 remission and CDAI remission with those in patients with DAS28 remission and CDAI-defined moderate disease activity, the difference was highly significant (P < 0.0001, by t-test). Indeed, the slope of the ESR curve in the DAS28 is steepest in its lowest (and thus normal) range (28, 31). In other words, this very low mean ESR in patients with CDAI-defined moderate disease activity “compensated” for higher values for “alternative” clinical disease activity variables (such as reflected by a higher SJC).

These data support the conclusion that the formula for the DAS28, which underweights the SJC and overweights the ESR, allows for increased proportions of patients with disease remission, and that this may be exaggerated with therapies affecting acute-phase reactants profoundly and even within normal levels.


The results of this study show that tocilizumab, a monoclonal antibody to IL-6R, interferes with clinical disease activity in RA irrespective of its effects on acute-phase reactants. When using the CDAI, a composite index that has been widely validated (21, 28, 31–33) but does not comprise an acute-phase reactant component, tocilizumab treatment at doses of both 4 mg/kg and 8 mg/kg led to a reduction of disease activity that was significantly different from that observed with placebo, with a dose-related effect; this dose-dependent decrease in clinical disease activity is consistent with previous observations on the acute-phase reactant response (9, 18, 19). Indeed, a significant difference from placebo was already seen within the first 2–4 weeks, with continuing improvement thereafter.

The overall reduction in the CDAI was 55% at 6 months, and low disease activity or remission according to established cut points was attained by approximately one-third of the patients receiving the 8 mg/kg dose of tocilizumab; CDAI remission itself was achieved by 6.4% of the patients. This proportion of patients attaining disease remission and low disease activity was almost identical to that seen with infliximab in a similar population of patients who were studied several years previously. Remission, however, was broadly prevalent with the DAS28 criteria and was ∼30% when using the traditional cut point of 2.6 and ∼26% when using a modified cut point of 2.4, which was derived using the same methodology as that used for the SDAI and CDAI remission criteria (20, 26).

Thus, there are several dimensions of trying to address the question of whether the therapeutic results seen with tocilizumab might be overestimated because of specific effects on acute-phase reactant values. Clearly, there seems to be an interaction between the drug used for treatment and the instrument used for assessment of remission. If indices whose components have a high clinical:acute-phase reactant ratio are used, then no important differences in remission rates with tocilizumab are apparent when compared with TNF inhibitors. However, if indices are used in which acute-phase reactants have high weight even within the normal range (e.g., leading to a major difference between an ESR of 2 mm/hour and 6 mm/hour), then such “supra-clinical” effects on the acute-phase reactant values might lead to discordances in remission rates with tocilizumab compared with other drugs, despite the absence of a difference in clinical efficacy. The significance of this finding is not yet obvious.

Both the CRP level and the ESR have been shown to be associated with joint damage (15, 16), and the possibility cannot be excluded that strong interference with acute-phase reactants conveys particularly good radiographic outcomes, although there are indications that SJCs may be more influential in this regard than are acute-phase reactants (34). Thus, although tocilizumab treatment reduces joint damage (23), the impact of its effects on acute-phase reactants in relation to radiographic changes needs to be evaluated in more detail in future analyses.

The evidence provided here expands previous conclusions that the DAS28 overweighs acute-phase reactant measures by revealing that, as a consequence, it allows a higher level of clinical disease activity to be counterbalanced by a very low acute-phase reactant value in patients classified as having disease in remission. Thus, tocilizumab unmasks the potential of the DAS28 to increase remission rates due to minor changes in the ESR, even within its normal range (28, 31). Consequently, agents such as TNF inhibitors and particularly tocilizumab that interfere profoundly with even a normal acute-phase reactant value permit a larger proportion of patients with residual clinical disease activity to fulfill DAS28 remission criteria than is the case with agents that have lesser effects on acute-phase reactants.

In contrast, remission rates according to the CDAI or SDAI, while numerically slightly higher for the SDAI among patients treated with tocilizumab, were in a range similar to that for both IL-6R and TNF inhibitor treatment. This finding is partly attributable to the low weight of the CRP level in the SDAI and partly to the relatively low CRP levels at baseline in tocilizumab trials. The difference among the therapies in terms of DAS28 remission rates, however, supports the contention that interference with IL-6 signaling conveys stronger effects on acute-phase reactants, especially the ESR, than does inhibition of TNF.

The current study has several limitations. In order to attain large numbers of patients with disease in remission according to all measures of disease activity, we pooled patients from different tocilizumab trials for some of the analyses. However, when looking at these trials individually, similar results were attained (data not shown). Furthermore, we assessed a random 80% data set from the clinical trials and evaluated only patients for whom all data needed for our analyses were complete; however, the proportions of responders in this data set were similar to those in published full analyses of the clinical trials. Also, we evaluated only patients in tocilizumab trials in whom disease remained active despite MTX or DMARD therapy; therefore, the data presented are primarily pertinent to this patient population and will have to be confirmed in other patient populations, such as patients who have not received MTX or in whom the response to TNF inhibitors was insufficient. Nevertheless, the majority of patients currently seen in clinical practice belong to the group of patients assessed in this study.

Finally, for pragmatic reasons (the studies were performed several years apart and no head-to-head comparisons of the drugs were performed), we refrained from statistical analyses between tocilizumab and infliximab and simply showed the numerical similarities and disparities; only head-to-head studies will allow for a more detailed comparison of the response to these agents. Nevertheless, as seen when assessing fulfillment of the ACR response components, it appears that the clinical core components of the response criteria are similarly affected by tocilizumab and infliximab, but that the values for acute-phase reactants are more strongly reduced by tocilizumab.

Although our observations are particularly important in terms of IL-6 inhibition, they are not limited to this type of therapy. Indeed, higher (or similar) DAS28 remission rates compared with ACR70 response rates have been observed with TNF inhibitors and even MTX (35–37), supporting the view that the consequences of therapy for acute-phase reactants are of a more general nature and must be taken into consideration when assessing remission using certain instruments. Moreover, remission rates using the DAS28 are generally higher than those using the SDAI and CDAI, which provide more stringent remission criteria because of a stronger relative or exclusive impact of the clinical components (26, 29). Despite the fact that use of the DAS28 generally leads to high or even exaggerated remission rates, it has long been customary to use the DAS28 as the preferred definition of remission in clinical trials, and this approach may need reconsideration. With several classifications of RA remission available, the definition of disease remission is currently being reevaluated by the ACR and the European League Against Rheumatism (38).

Whereas disease remission is a major aim of RA treatment that has only recently become achievable, in patients with longstanding disease who have previously received many types of previous treatments, such as the patients in this study, remission may not be attainable. Therefore, the attainment of low disease activity is an alternative therapeutic goal (22). The results obtained using the pooled tocilizumab clinical trial data show that a significant proportion of patients achieved a state of low disease activity; indeed, the combined proportion of patients with disease in remission and patients with low disease activity appears to be similar to that observed among patients receiving TNF inhibitors. Thus, although DAS28 remission rates are high with tocilizumab therapy, interference with the IL-6 pathways conveys significant clinical efficacy, leading to a large proportion of patients who attain a desirable clinical state.


Both 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. Smolen 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.

Study conception and design. Smolen, Aletaha.

Acquisition of data. Smolen, Aletaha.

Analysis and interpretation of data. Smolen, Aletaha.


The sponsors of the clinical trials, Roche and Centocor, had no role or influence on the design of the present study or the data analysis; they were not involved in writing or submitting the manuscript. Roche provided a research grant, and both Roche and Centocor provided trial data, as stated in Patients and Methods. The companies were informed of the submission of the manuscript and approval. However, publication of this article was not contingent on approval by the study sponsors.


We thank Farideh Alasti for expert assistance with the statistical analyses.