Treatment of rheumatoid arthritis with anakinra, a recombinant human interleukin-1 receptor antagonist, in combination with methotrexate: Results of a twenty-four–week, multicenter, randomized, double-blind, placebo-controlled trial

Authors


Abstract

Objective

To evaluate the efficacy and safety of anakinra in combination with methotrexate (MTX) in patients with active rheumatoid arthritis (RA).

Methods

Patients with moderate-to-severe active RA who were receiving MTX for 6 consecutive months, with stable doses for ≥3 months (those with disease duration of >6 months but <12 years) were randomized into 6 groups: placebo or 0.04, 0.1, 0.4, 1.0, or 2.0 mg/kg of anakinra administered in a single, daily, subcutaneous injection. The primary efficacy end point was the proportion of subjects who met the American College of Rheumatology 20% improvement criteria (attained an ACR20 response) at week 12.

Results

A total of 419 patients were randomized in the study. Patient demographics and disease status were similar in the 6 treatment groups. The ACR20 responses at week 12 in the 5 active treatment plus MTX groups demonstrated a statistically significant (P = 0.001) dose-response relationship compared with the ACR20 response in the placebo plus MTX group. The ACR20 response rate in the anakinra 1.0-mg/kg (46%; P = 0.001) and 2.0-mg/kg (38%; P = 0.007) dose groups was significantly greater than that in the placebo group (19%). The ACR20 responses at 24 weeks were consistent with those at 12 weeks. Similar improvements in anakinra-treated subjects were noted in individual ACR components, erythrocyte sedimentation rate, onset of ACR20 response, sustainability of ACR20 response, and magnitude of ACR response. Anakinra was safe and well tolerated. Injection site reaction was the most frequently noted adverse event, and this led to premature study withdrawal in 7% (1.0-mg/kg group) to 10% (2.0-mg/kg group) of patients receiving higher doses.

Conclusion

In patients with persistently active RA, the combination of anakinra and MTX was safe and well tolerated and provided significantly greater clinical benefit than MTX alone.

Rheumatoid arthritis (RA) is a chronic, progressive, systemic, inflammatory disorder of unknown etiology characterized by symmetric, erosive, and disabling polyarthritis and a wide array of extraarticular complications. As the disease progresses, irreversible joint damage may lead to loss of function and to deformity. Long-term studies have demonstrated that significant disability occurs in 50–70% of patients after 10–15 years of disease (1).

Two of the most important discoveries in the last decade of RA research are 1) joint destruction occurs early in the course of the disease, and 2) cytokines are crucial mediators in the pathogenesis of RA. Patients with active, polyarticular, rheumatoid factor (RF)–positive disease have a >70% probability of developing joint damage or erosions within 2 years of the onset of disease (2–5). Within 3 years, 70% of patients show radiographic damage (2). Although the etiology of RA is poorly understood, the pathogenesis of RA is thought to be mediated by antigen-driven T cells and autoreactive B cells that produce proinflammatory cytokines directly involved in joint destruction (6, 7). Studies of the pathogenesis of RA, both in experimental animal models and in clinical settings, suggest a pivotal role for the cytokine interleukin-1 (IL-1) in synovial inflammation and articular tissue destruction (8–10). IL-1 is present in increased concentrations in synovia of patients with RA (11–13). It promotes inflammation as well as bone and cartilage resorption and appears to play a major role in the promotion of rheumatic inflammation (13). Because of the potent and pathologic effects of IL-1 in RA, regulation of this cytokine is an important component of RA management.

IL-1 receptor antagonist (IL-1Ra) is a naturally occurring, acute-phase, antiinflammatory protein (14). It is a specific inhibitor of IL-1 that acts by blocking the binding of IL-1 to type I receptors (14–16). IL-1Ra exhibits no bioactivity, but antagonizes the effects of IL-1 (15, 16). IL-1Ra is present in the synovial membranes of RA patients (17). Anakinra, a recombinant human form of IL-1Ra, has recently been approved for the treatment of RA.

The objective of IL-1Ra–based therapy is to occupy enough IL-1 receptors with IL-1Ra to block IL-1 cell signaling. It has been shown that >95% receptor occupancy by IL-1Ra is necessary to block IL-1 signaling (14). The availability of recombinant human IL-1Ra, which differs from the native nonglycosylated human IL-1Ra only by the addition of an N-terminal methionine, has made it possible to study this protein as a treatment for RA (18).

Efficacy has previously been shown in a 24-week evaluation of the American College of Rheumatology 20% improvement criteria (ACR20) (19) composite score that indicated a beneficial effect of anakinra monotherapy in patients with early aggressive disease (P = 0.01). In addition, a 41% reduction in disease progression was observed compared with placebo, as assessed by the Larsen score (20). When patients continued to receive anakinra for an additional 24 weeks of treatment, responses were maintained both in the ACR20 composite score and in disease progression (21–23). Treatment was well tolerated, and no major adverse events were observed.

Methotrexate (MTX) is the most commonly used conventional disease-modifying antirheumatic drug (DMARD). Combination regimens that employ MTX plus other biologic agents, such as etanercept or infliximab (both of which inhibit tumor necrosis factor α), have been used successfully to combat disease that remains active and/or progressive despite therapy with MTX alone (24–27). In order to further support the use of combination therapy with MTX and a biologic agent in RA and to evaluate the efficacy and safety of anakinra in combination with MTX, a 24-week, multicenter, randomized, double-blind, placebo-controlled study was performed in patients who continued to have active disease despite receiving maximally tolerated doses of MTX, as assessed by their physicians.

PATIENTS AND METHODS

Patients were eligible for this 24-week, multicenter, randomized study if they met the ACR (formerly, the American Rheumatism Association) 1987 revised criteria for the diagnosis of RA (28), experienced symptoms for >6 months but <12 years, and had active disease (defined as the presence of ≥6 swollen joints and at least 2 of the following: ≥9 tender or painful joints, morning stiffness of ≥45 minutes in duration, or a plasma C-reactive protein [CRP] concentration of >1.5 mg/dl). Eligible patients were required to have been receiving MTX therapy (15–25 mg/week) for at least 6 consecutive months and to have been receiving stable doses for ≥3 months before study entry. Doses of nonsteroidal antiinflammatory drugs and/or oral corticosteroids (≤10 mg/day of prednisone or equivalent) had to have been stable for 4 weeks prior to study entry.

Exclusion criteria included intraarticular or systemic corticosteroid injection within 4 weeks of study enrollment. Patients who received penicillamine, oral or parenteral gold, azathioprine, or cyclosporine within 12 weeks of study initiation and those who received hydroxychloroquine or sulfasalazine within 8 weeks of study initiation were also excluded. The study protocol and informed consent of each subject were approved by the institutional review boards of the study centers.

Treatment. Following assessment of eligibility, subjects were randomized to receive, in addition to their MTX dose (15–25 mg/week), either placebo or anakinra (0.04, 0.1, 0.4, 1.0, or 2.0 mg/kg every day) administered as a single subcutaneous injection by the patient or caregiver. Verbal instructions were given to the patient or caregiver regarding the volume of solution to be injected. Amgen (Thousand Oaks, CA) produced the anakinra with recombinant DNA technology using Escherichia coli fermentation. Study medication was packaged as single-use, unit-dose vials. Each vial contained 1.25 ml of a solution consisting of either placebo or anakinra 10, 40, or 200 mg/ml, sodium chloride, sodium citrate, polysorbate 80, EDTA, citric acid, and sterile water for injection. Subjects were instructed to alternate injection sites among the abdomen, thighs, and backs of the upper arms.

Study design. The study was a multicenter, randomized, double-blind, placebo-controlled trial comparing multiple doses of anakinra versus placebo (MTX alone). Patients were recruited from 36 centers (31 in the US, 3 in Canada, and 2 in Australia). The study was originally designed to evaluate a treatment duration of 12 weeks and included placebo and anakinra dose groups of 0.1, 0.4, and 2.0 mg/kg/day. Clinic visits were scheduled at baseline and at weeks 1, 2, 4, 8, and 12 for a total of 6 visits. A followup visit was scheduled at 2 weeks following the last injection. A total of 105 subjects had been recruited into the 12-week protocol (27 in the placebo group and 28, 23, and 27 in the 0.1-, 0.4-, and 2.0-mg/kg anakinra groups, respectively).

As an amendment to the original study, the duration of double-blind treatment was extended from 12 weeks to 24 weeks, and incremental visits at weeks 16, 20, and 24 were added. The number of anakinra dose groups was broadened to include doses of 0.04 and 1.0 mg/kg/day. Of the 105 patients randomized in the original 12-week protocol, 3 subjects reconsented and remained in their same blinded treatment groups (1 in the placebo group and 1 each in the 0.4-mg/kg and 2.0-mg/kg groups) for the 24-week protocol. A total of 314 additional subjects were randomized into the 24-week protocol, as follows: placebo (n = 47) and anakinra 0.04 (n = 63), 0.1 (n = 46), 0.4 (n = 54), 1.0 (n = 59), and 2.0 (n = 45) mg/kg. For the final analyses, 419 patients were evaluated at the 12-week end point, and 317 patients were evaluated at the 24-week end point.

Efficacy assessment. The primary efficacy end point was the proportion of subjects who attained an ACR20 response at week 12. Secondary efficacy end points included an ACR20 response at week 24, ACR50 and ACR70 responses (29) at weeks 12 and 24, sustained ACR20 response, and change from baseline at weeks 12 and 24 in ACR improvement criteria components, erythrocyte sedimentation rate (ESR), and duration of morning stiffness.

ACR components included the following: number of swollen joints among 66 diarthrodial joints, number of tender/painful joints among 68 diarthrodial joints, patient's assessment of disease activity based on a visual analog scale (VAS) of 0–100, investigator's assessment of disease activity based on a VAS of 0–100, patient's assessment of pain based on a VAS of 0–100, Health Assessment Questionnaire (HAQ) score (a composite score of 8 categories rated on a scale of 0–3) (30), and plasma CRP concentration. All efficacy measurements except the HAQ score were obtained at weeks 0, 1, 2, 4, 8, 12, 16, 20, and 24; the HAQ score was obtained at weeks 0, 4, 8, 12, 16, and 24.

Safety assessment. Safety assessments were obtained at weeks 0, 1, 2, 4, 8, 12, 16, 20, and 24. Safety assessments included treatment-emergent adverse events, clinical laboratory data (complete blood cell count, blood chemistry profile, and urinalysis), and vital signs. Blood samples were obtained at baseline and at weeks 4, 12, and 24 for anti-anakinra antibody screening. Plasma samples were evaluated initially using an enzyme-linked immunosorbent assay method that was modified to include the use of chicken anti-human immunoglobulin reagent to reduce potential interference from RF. Samples identified as either reactive or approaching the assay threshold of being reactive in the initial antibody screen were further characterized using a surface plasma resonance (SPR) biosensor-based confirmatory assay (BIAcore, Stevenage, UK) (31). Samples that were confirmed to be positive by the SPR assay were further evaluated in a biologic cell-based assay for the presence of any generalized antibodies capable of neutralizing the biologic effect of anakinra.

Statistical analysis. The primary analysis of efficacy end points was based on an intent-to-treat population; all subjects were analyzed according to their randomization group. Subjects whose ACR response status could not be ascertained, because of either a missing evaluation or an inability to fully evaluate the response due to missing ACR components, were classified as nonresponders for the relevant time point.

The primary analysis of the ACR20 response rates at weeks 12 and 24 was based on a center-adjusted single omnibus test for a positive dose-response relationship between treatment and ACR20 response rate (32). Given the directional nature of the dose-response alternative, the criterion for statistical significance for this analysis was a 1-tailed alpha level of 0.025. Pairwise differences between each anakinra treatment group and placebo with respect to ACR20 response rate were calculated using a center-adjusted logistic regression model. To control for the Type I error rate due to multiple comparisons, significance of the omnibus test for dose response was a prerequisite for interpreting pairwise differences as being statistically significant.

Analyses similar to those conducted for ACR20 responses were also conducted for the sustained ACR20 response. Subjects achieving a sustained ACR20 response exhibited ACR20 responses for at least 4 (not necessarily consecutive) of the 6 months, with at least 1 response being achieved at week 12 or 24.

The magnitude of the ACR20, ACR50, and ACR70 responses and the assessment of dose-response relationships to treatment were examined by using the Jonckheere-Terpstra test (33, 34) and a 1-tailed alpha level of 0.025. This test assessed whether the magnitude of the ACR response increased with an increasing anakinra dose.

For each ACR component, ESR, and duration of morning stiffness, the mean change from baseline at weeks 12 and 24 was analyzed using least-squares mean values obtained from a repeated measures mixed model analysis of covariance (35) adjusting for center and the corresponding baseline value. An omnibus test for differences among the 6 treatment groups was examined by assessment of a treatment main effect prior to interpretation of any statistical tests for the pairwise comparisons. Except for the 1-tailed tests for dose response, all significance tests were conducted at the 2-tailed 0.05 level.

Analysis of safety end points included all randomized subjects who received at least 1 injection of placebo or anakinra. Adverse events were analyzed using subject incidence rates and summarized by body system for each treatment group and for all anakinra treatment groups combined. (Additional analyses of subject incidence rates examined the exposure-adjusted subject incidence rates for both the cumulative and hazard rate as defined by life-table methods.) Laboratory end points were examined for mean changes over time as well as for shifts from baseline in World Health Organization toxicity grades (36).

RESULTS

Demographic and baseline characteristics. A total of 419 patients were randomized to the 6 treatment groups, as follows: placebo (n = 74), anakinra 0.04 mg/kg (n = 63), anakinra 0.1 mg/kg (n = 74), anakinra 0.4 mg/kg (n = 77), anakinra 1.0 mg/kg (n = 59), and anakinra 2.0 mg/kg (n = 72). The baseline and demographic characteristics of the enrolled subjects are summarized in Table 1. Patients enrolled had moderate-to-severe disease, as demonstrated by seropositivity for RF (range 72.9–83.3%), previous number of DMARD exposures excluding MTX (1.4–2.1), and corticosteroid use (58.4–68.3%). The mean duration of disease at enrollment ranged from 6.3 years to 8.8 years. Disease characteristics of each treatment group at baseline were also similar (Table 2), and also highlighted the moderate-to-severe nature of RA present in the patients who participated in this study.

Table 1. Patient demographics and disease status at baseline*
 Placebo (n = 74)Anakinra dosage, mg/kg/day
0.04(n = 63)0.1(n = 74)0.4(n = 77)1.0(n = 59)2.0(n = 72)
  • *

    RA = rheumatoid arthritis; DMARDs = disease-modifying antirheumatic drugs; NSAID = nonsteroidal antiinflammatory drug; MTX = methotrexate.

  • Excluding MTX.

Mean age, years53.052.653.052.849.054.1
Men, %14.922.220.323.415.337.5
Women, %85.177.879.776.684.762.5
Race, no. (%)
 White67 (90.5)56 (88.9)67 (90.5)64 (83.1)51 (86.4)66 (91.7)
 African American2 (2.7)2 (3.2)1 (1.4)4 (5.2)6 (10.2)4 (5.6)
 Hispanic4 (5.4)3 (4.8)4 (5.4)5 (6.5)1 (1.7)2 (2.8)
 Other1 (1.4)2 (3.2)2 (2.7)4 (5.2)1 (1.7)0 (0.0)
Mean duration of RA, years7.86.38.87.06.58.0
Rheumatoid factor positive, %74.381.078.477.972.983.3
Mean no. of previous DMARDs used2.12.01.91.41.81.9
Corticosteroid use, no. (%)49 (66.2)43 (68.3)48 (64.9)45 (58.4)37 (62.7)47 (65.3)
NSAID use, no. (%)50 (67.6)50 (79.4)52 (70.3)52 (67.5)38 (64.4)47 (65.3)
Mean MTX dose, mg/week16.317.617.017.016.716.8
Table 2. Patient clinical and laboratory characteristics at baseline*
 Placebo (n = 74)Anakinra dosage, mg/kg/day
0.04 (n = 63)0.1 (n = 74)0.4 (n = 77)1.0 (n = 59)2.0 (n = 72)
  • *

    Values are the mean ± SD. HAQ = Health Assessment Questionnaire; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate.

Tender/painful joint count (0–68 joints)28.1 ± 13.923.9 ± 11.425.9 ± 14.827.1 ± 13.022.0 ± 12.924.6 ± 12.8
Swollen joint count (0–66 joints)18.4 ± 9.818.8 ± 8.718.3 ± 9.219.1 ± 9.217.6 ± 8.817.4 ± 8.1
Physician's global assessment (0–100 scale)56.7 ± 18.555.7 ± 19.261.2 ± 17.660.1 ± 18.553.6 ± 17.055.8 ± 18.5
Patient's global assessment (0–100 scale)52.6 ± 21.547.6 ± 21.251.1 ± 21.550.4 ± 19.347.5 ± 21.551.2 ± 21.7
Patient's assessment of pain (0–100 scale)52.5 ± 22.246.4 ± 20.951.6 ± 22.451.2 ± 21.347.5 ± 22.854.6 ± 21.4
HAQ score (0–3 scale)1.4 ± 0.61.4 ± 0.61.5 ± 0.71.5 ± 0.61.3 ± 0.61.3 ± 0.6
CRP level (mg/dl)1.99 ± 2.562.22 ± 3.361.57 ± 1.612.08 ± 2.471.55 ± 2.262.00 ± 2.63
ESR (mm/hour)36.2 ± 27.737.0 ± 23.137.9 ± 24.637.3 ± 26.237.1 ± 25.035.1 ± 20.9
Duration of morning stiffness (minutes/day)139.9 ± 113.2128.5 ± 90.4117.3 ± 90.9120.3 ± 91.0133.6 ± 98.7142.8 ± 98.4

ACR20 response. The results at 12 weeks demonstrated a highly significant dose response across the placebo group and the 5 anakinra dose groups (P = 0.001) (Figure 1). This dose-response relationship provided evidence of an increasing ACR20 response rate with increasing anakinra doses. This response appeared to plateau at the 1–2-mg/kg dose range, indicating that the top of the dose relationship may have been reached. There was a notable difference in response between the groups that received placebo and those that received the two highest doses of anakinra. The ACR20 response rates in the anakinra 0.04-mg/kg (25%; P =0.523) and 0.4-mg/kg (25%; P =0.378) dose groups were not significantly different from that in the placebo group (19%), while the ACR20 response rates in the anakinra 1.0-mg/kg (46%; P =0.001) and 2.0-mg/kg (38%; P =0.007) dose groups were significantly better than that in the placebo group. The ACR20 response rate for the anakinra 0.1-mg/kg dose group (35%) was also significantly better (P =0.014) than that observed in the placebo group. The likelihood of achieving an ACR20 response with anakinra at 1.0 mg/kg was 3.8 times greater than with placebo, and for subjects in the anakinra 2.0-mg/kg group, the likelihood was 2.9 times greater than with placebo.

Figure 1.

Percentages of patients meeting the American College of Rheumatology 20% improvement criteria (achieving an ACR20 response) at weeks 12 and 24. The tests at weeks 12 and 24 are for the dose response adjusted for the study center at the 0.025 significance level, as described in Patients and Methods. Shown are odds ratios (ORs), obtained using logistic regression, of achieving an ACR20 response relative to placebo. P values correspond to pairwise comparisons of each anakinra dose group with placebo, adjusted for the study center.

The ACR20 responses at 24 weeks were consistent with those at 12 weeks. There was still a clear anakinra dose response (P =0.004). At 24 weeks, subjects treated with anakinra 1.0 mg/kg had a significantly better ACR20 response than did those who received placebo (42% versus 23%; P =0.018). Subjects treated with anakinra 0.4 mg/kg or 2.0 mg/kg also had better ACR20 responses (36% and 35%, respectively) than did those who received placebo, but the difference did not achieve statistical significance.

ACR components, ESR, and duration of morning stiffness. The change from baseline in individual components of the ACR improvement criteria and in duration of morning stiffness at weeks 12 and 24 among subjects who received placebo, anakinra 1.0 mg/kg, and anakinra 2.0 mg/kg are shown in Table 3. At week 12, the change from baseline in each parameter among subjects who received either the 1.0-mg/kg or 2.0-mg/kg dose of anakinra was statistically significant (P≤ 0.05) compared with the change from baseline among subjects who received placebo, except for the physician's global assessment (P =0.051 and P =0.004, respectively), patient's assessment of pain (P =0.075 and P =0.003, respectively), number of tender/painful joints (P =0.702 and P =0.344, respectively), and duration of morning stiffness (P =0.031 and P =0.069, respectively).

Table 3. Clinical responses in individual ACR components, ESR, and duration of morning stiffness at weeks 12 and 24 for patients treated with placebo, anakinra 1.0 mg/kg/day, or anakinra 2.0 mg/kg/day*
 Change at 12 weeksChange at 24 weeks
nAdjusted meanP§nAdjusted meanP§
  • *

    Lower doses are omitted for clarity. ACR = American College of Rheumatology (see Table 2 for other definitions).

  • Number of patients in study week.

  • Least-squares mean obtained from repeated measures mixed model adjusted for study center and baseline value.

  • §

    Pairwise comparison of each anakinra dose with placebo, adjusted for study center and baseline value.

Physician's global assessment (0–100 scale)
 Placebo66−12.3138−14.08
 Anakinra 1.0 mg/kg48−19.170.05145−22.340.037
 Anakinra 2.0 mg/kg58−21.790.00432−24.460.013
Patient's global assessment (0–100 scale)
 Placebo66−6.4238−3.61
 Anakinra 1.0 mg/kg48−14.290.03545−13.830.014
 Anakinra 2.0 mg/kg58−17.140.00232−20.440.0001
Patient's assessment of pain (0–100 scale)
 Placebo66−6.0838−2.58
 Anakinra 1.0 mg/kg48−13.000.07545−12.900.060
 Anakinra 2.0 mg/kg58−19.340.00332−22.78<0.0001
No. of swollen joints (0–66 joints)
 Placebo66−2.9538−4.17
 Anakinra 1.0 mg/kg48−7.370.000245−6.300.1063
 Anakinra 2.0 mg/kg58−6.640.000832−7.610.014
No. of tender/painful joints (0–68 joints)
 Placebo66−8.1538−8.28
 Anakinra 1.0 mg/kg48−8.850.70245−8.830.788
 Anakinra 2.0 mg/kg58−9.790.34432−11.180.177
HAQ score (0–3 scale)
 Placebo66−0.1638−0.15
 Anakinra 1.0 mg/kg48−0.350.02545−0.370.036
 Anakinra 2.0 mg/kg58−0.390.00432−0.510.0005
CRP level (mg/dl)
 Placebo640.0438−0.19
 Anakinra 1.0 mg/kg48−0.790.007344−0.770.209
 Anakinra 2.0 mg/kg57−1.040.000230−0.770.250
ESR (mm/hour)
 Placebo630.2336−4.15
 Anakinra 1.0 mg/kg47−9.950.000344−12.360.009
 Anakinra 2.0 mg/kg56−14.85<0.000128−14.450.002
Duration of morning stiffness (minutes/day)
 Placebo66−48.1538−49.56
 Anakinra 1.0 mg/kg48−79.040.03145−74.240.1228
 Anakinra 2.0 mg/kg58−72.710.06932−81.480.059

ESRs at week 12 improved significantly (P< 0.001) from mean baseline values of 37.1 mm/hour in the anakinra 1.0-mg/kg group (−9.95 mm/hour) and 35.1 mm/hour in the anakinra 2.0-mg/kg group (−14.85 mm/hour), while no change from baseline was seen in the placebo group. At week 24, the improvement from baseline in ESRs among subjects who were treated with anakinra 1.0 mg/kg or 2.0 mg/kg was also significantly greater than that in subjects who received placebo (P =0.009 and P =0.002, respectively).

Onset of ACR20 response. The onset of ACR20 response among subjects treated with anakinra 1.0 mg/kg or 2.0 mg/kg was evident as early as 2–4 weeks (Figure 2). At week 2, the odds of achieving an ACR20 response among subjects in the anakinra 1.0-mg/kg and 2.0-mg/kg groups were 1.66 and 2.08 times greater, respectively, than for subjects in the placebo group (P not significant). At week 4, the ACR response rate among subjects treated with anakinra 2.0 mg/kg was significantly (P =0.032) better than that among subjects who received placebo. At week 8, the ACR response rates among subjects who received anakinra 1.0 mg/kg and 2.0 mg/kg were significantly better than that among subjects who received placebo (P =0.027 and P =0.022, respectively).

Figure 2.

Percentages of patients achieving an ACR20 response, by week of visit. For clarity, only responses for the 1.0-mg/kg and 2.0-mg/kg anakinra doses are shown, since they showed the greatest effects versus placebo. P values are indicated for doses and time points that showed significant differences between the placebo-treated and anakinra-treated groups. See Figure 1 for definitions.

Sustainability of ACR20 response. The sustained ACR20 response is an extension of the ACR20 response that includes multiple time points in order to measure the durability of response to therapy. A sustained ACR20 responder was defined as a patient who achieved an ACR20 response for a minimum of 4 of the 6 study months, one of which had to be observed at week 12 or week 24; the 4 positive responses did not need to be consecutive.

The percentages of patients who achieved a sustained ACR20 response are shown in Figure 3. Strong evidence of a dose response continued to be observed, with no evidence of a dose-response plateau as seen with the ACR20 response (Figure 1). Statistically significant evidence of a sustained response over the 24-week treatment period was observed in subjects who received anakinra 1.0 mg/kg and 2.0 mg/kg (P =0.039 and P =0.013, respectively), compared with those who received placebo. The results demonstrated a significant dose response across the placebo and the 5 anakinra dose groups (P =0.002) (Figure 3). This dose-response relationship provided evidence of an increasing ACR20 sustained response rate with increasing anakinra doses.

Figure 3.

Percentages of patients achieving a sustained ACR20 response at week 24. Shown are results from an intent-to-treat population, using nonresponder imputation, as described in Patients and Methods. Also shown are ORs, obtained using logistic regression, of achieving an ACR20 response relative to placebo. P values correspond to pairwise comparisons of each anakinra dose group with placebo, adjusted for the study center. Numbers of evaluable patients at week 24 are reduced from those at week 12. See Figure 1 for definitions.

Magnitude of ACR response. The magnitude of clinical response, as determined by the percentage of patients who achieved ACR20, ACR50, and ACR70 responses at 12 and 24 weeks, is shown in Figure 4. The overall magnitude of clinical response was significantly greater at weeks 12 and 24 among subjects treated with anakinra (P =0.001 and P =0.003, respectively) than that among those who received placebo. The difference was most pronounced in the anakinra 1.0-mg/kg and 2.0-mg/kg dose groups. At week 12, 19% of subjects treated with anakinra 1.0 mg/kg and 24% of those treated with anakinra 2.0 mg/kg achieved an ACR50 response, compared with only 4% of subjects who received placebo. Similarly, at week 12, 5% of subjects treated with anakinra 1.0 mg/kg and 11% of those treated with anakinra 2.0 mg/kg achieved an ACR70 response, compared with 0% of subjects who received placebo. Similar results were seen at week 24.

Figure 4.

Percentages of patients achieving ACR20, ACR50, and ACR70 responses at weeks 12 and 24. Shown are results from an intent-to-treat population, using nonresponder imputation, as described in Patients and Methods. See Figure 1 for definitions.

Safety and tolerability. The percentage of subjects who withdrew prematurely from the study ranged from 16% to 26% across the 6 randomization arms (Table 4). The most common reasons for withdrawal from the study were lack of efficacy (6–14%) and adverse events (1–15%). The percentage of subjects who withdrew prematurely because of lack of efficacy was greatest in the anakinra 0.04-mg/kg group (14%). Most withdrawals in the anakinra 1.0-mg/kg and 2.0-mg/kg groups were due to adverse events (resulting in withdrawal rates of 14% and 15%, respectively), mainly injection-site reactions (Table 5).

Table 4. Distribution of patients who withdrew prematurely from study*
Reason for withdrawalPlacebo (n = 74)Anakinra dosage, mg/kg/day
0.04 (n = 63)0.1 (n = 74)0.4 (n = 77)1.0 (n = 59)2.0 (n = 72)
  • *

    Values are the number (%) of patients.

Any reason14 (18.9)13 (20.6)12 (16.2)17 (22.1)13 (22.0)19 (26.4)
Adverse event3 (4.1)2 (3.2)1 (1.4)5 (6.5)8 (13.6)11 (15.3)
Protocol violation1 (1.4)0 (0.0)0 (0.0)1 (1.3)0 (0.0)0 (0.0)
Noncompliance1 (1.4)1 (1.6)2 (2.7)1 (1.3)0 (0.0)1 (1.4)
Lack of efficacy5 (6.8)9 (14.3)7 (9.5)6 (7.8)4 (6.8)4 (5.6)
Lost to followup0 (0.0)0 (0.0)0 (0.0)3 (3.9)0 (0.0)1 (1.4)
Withdrawal of consent4 (5.4)1 (1.6)2 (2.7)1 (1.3)1 (1.7)2 (2.8)
Table 5. Distribution of withdrawals due to adverse events in ≥2 patients*
Reason for withdrawalPlacebo (n = 74)Anakinra dosage, mg/kg/day
0.04 (n = 63)0.1 (n = 74)0.4 (n = 77)1.0 (n = 59)2.0 (n = 72)
  • *

    Values are the number (%) of patients. Adverse events for which only 1 person withdrew involved the reproductive, gastrointestinal, and musculoskeletal systems as well as the skin and appendages, and included urinary disorders.

  • Includes leukopenia or granulocytopenia.

Adverse event3 (4.1)2 (3.2)2 (2.7)5 (6.5)8 (13.6)11 (15.3)
 Injection site reactions2 (2.7)0 (0.0)0 (0.0)1 (1.3)4 (6.8)7 (9.7)
 Hematologic0 (0.0)1 (1.6)1 (1.4)1 (1.3)1 (1.7)1 (1.4)
 Respiratory1 (1.4)1 (1.6)0 (0.0)1 (1.3)1 (1.7)0 (0.0)
 Body as a whole0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)2 (2.8)
 Chest pain0 (0.0)0 (0.0)0 (0.0)0 (0.0)1 (1.7)1 (1.4)

In general, anakinra was well tolerated. The most frequent adverse events were dose-related injection site reactions, which were experienced by 28% of subjects in the placebo group and by 19%, 38%, 56%, 64%, and 63% of subjects in the anakinra 0.04-, 0.1-, 0.4-, 1.0-, and 2.0-mg/kg groups, respectively. The majority of injection site reactions were mild to moderate in severity and tended to diminish with repeated doses. Injection site reactions included not only pain, erythema, pruritus, and rash, but also injection site trauma categories, such as bruising and bleeding. An injection site reaction was the reason for premature withdrawal from the study in 2 patients (3%) in the placebo group and for 1 patient (1%) in the 0.4-mg/kg anakinra group, 4 patients (7%) in the 1.0-mg/kg anakinra group, and 7 patients (10%) in the 2.0-mg/kg anakinra group (Table 5). The second most frequently reported adverse event was headache, which was experienced by 15% of placebo-treated subjects and by 14–34% of anakinra-treated subjects. Upper respiratory tract infection and sinusitis were reported by 22% and 15%, respectively, of placebo-treated subjects and by 14–24% and 5–14%, respectively, of anakinra-treated subjects. Neither upper respiratory tract infection nor sinusitis was reported as serious or severe or resulted in premature withdrawal from the study. No serious infections were noted during this study.

Abdominal pain was reported more frequently in subjects receiving anakinra (6%) than in those receiving placebo (1%). Consistent with the efficacy profile, musculoskeletal events (arthralgia and worsening of RA) were seen more often in subjects in the placebo-treated group (7% and 11%, respectively) than in the anakinra-treated groups (6% for both).

Five subjects (1.2%) withdrew prematurely from the study because of leukopenia. All but 1 of these 5 subjects developed a white blood cell (WBC) count of ≤3.0 × 109/liter (range 2.5–3.0), the limit specified in the study protocol for withdrawal. The 1 subject who did not develop a WBC count of ≤3.0 × 109/liter had a baseline WBC count of 6.8 × 109/liter, which decreased progressively to 4.2 × 109/liter at week 12, at which time the leukopenia was reported as a serious adverse event. WBC counts returned to normal after discontinuation of anakinra treatment. Fever, infection, or sepsis was not associated with these episodes of transient leukopenia. The occurrence of leukopenia in anakinra-treated subjects was not dose related.

Two subjects (0.5%) were diagnosed as having a new malignancy during the study. One subject (who received placebo), who had a history of adenocarcinoma of the prostate, developed large-cell carcinoma of the lung. The second subject (who received anakinra 2.0 mg/kg), who had a history of ovarian cyst, developed breast cancer. Neither malignancy was considered to be related to the study drug. Both subjects were subsequently withdrawn from the study. No deaths occurred during this study.

Antibodies to anakinra. A total of 57 placebo-treated subjects (13.6%) and 297 anakinra-treated subjects (70.9%) were screened for antibodies to anakinra at weeks 0, 1, 2, 4, 8, 12, 16, 20, and 24. Eight of the 297 anakinra-treated subjects (2.7%) were seroreactive to anakinra at some time during the study, but none were seroreactive at more than one time point. Only 1 of the 57 placebo-treated subjects (1.8%) was seroreactive to anakinra at some time during the study, and this subject was seroreactive to anakinra at baseline only. No evidence of neutralizing antibodies to anakinra was detected.

Injection site reactions occurred in 7 of the 8 anakinra-treated subjects (87.5%) who were seroreactive to anakinra. The single placebo-treated subject who was seroreactive to anakinra did not experience an injection site reaction. No other clinical sequelae or effects on efficacy as a result of seroreactivity to anakinra were observed.

DISCUSSION

Anakinra has been extensively studied as monotherapy for RA and has been demonstrated to be clinically effective. The efficacy of anakinra in RA was studied in a 24-week, multicenter, randomized, double-blind, placebo-controlled study reported by Bresnihan et al in 1998 (21). In that study, 472 patients were randomized to 1 of 4 treatment groups: anakinra 30 mg/day, anakinra 75 mg/day, anakinra 150 mg/day, or placebo. Following 24 weeks of treatment, significantly more subjects (43%) who received anakinra at a dosage of 150 mg/day met the ACR20 response criteria than did subjects who received placebo. Statistically significant improvements compared with placebo were observed at 24 weeks in the anakinra 150-mg/kg treatment group for all 7 individual ACR clinical parameters (37). Additionally, there was a reduction in radiographic progression, as measured by both the Larsen and the Sharp scores, as early as 6 months that persisted after an additional 6 months of followup (23). Patients who received the same dose of anakinra for the entire 12 months showed an acceleration of benefit in the second 6 months compared with the benefit obtained in the first 6 months (38).

Recent clinical trial and epidemiologic data have demonstrated the long-term response to DMARD monotherapy to be unsatisfactory. Pincus et al demonstrated that the majority of patients receiving DMARDs discontinued treatment after 24–36 months due either to lack of efficacy or to adverse events, with more patients continuing MTX long term (39). Evaluation of the clinical trials resulting in the approval of leflunomide for RA treatment demonstrated that  f30–35% of patients receiving MTX or leflunomide achieved an ACR50 response, which is considered by most practitioners to be a clinically important response (40–42). It has also recently been shown that other therapies in combination with MTX are effective in these patients. These include sulfasalazine (43), cyclosporine (44), etanercept (24), infliximab (27), and hydroxychloroquine (43). These results suggest that as many as 50–70% of patients may require additional therapy. MTX has been considered the “gold standard” for RA treatment, and for that reason, anakinra was evaluated in combination with MTX.

Experimental studies have indicated that >95% of IL-1 receptors need to be blocked in order to reach an optimal blockage of IL-1 signaling (14). This study therefore investigated a range of doses of anakinra to determine which doses could fully block the IL-1 signaling pathways. The results of this study demonstrate that the combination of anakinra and MTX was effective in relieving the clinical signs and symptoms of RA in patients who had active disease despite MTX treatment. Evidence of an increasing ACR20 response rate with increasing anakinra dose was demonstrated at weeks 12 and 24. The two highest doses of anakinra—1.0 mg/kg and 2.0 mg/kg—consistently yielded the strongest treatment effects at both time points. The magnitude of clinical efficacy, as measured by ACR50 and ACR70 responses, was found to be dose related and highest in the 1.0-mg/kg and 2.0-mg/kg cohorts.

Compared with placebo-treated subjects, ACR20 response among subjects treated with anakinra 1.0 mg/kg or 2.0 mg/kg occurred as early as 2 weeks of treatment. Examination of the sustained effects of anakinra across the entire 24-week treatment period demonstrates that the ACR20 responses were durable. Additionally, patients receiving anakinra (1.0 mg/kg and 2.0 mg/kg) demonstrated statistically significant improvements in the HAQ disability index at 12 and 24 weeks compared with the placebo-treated patients. The mean improvements in the HAQ score of 0.35 and 0.39 at 12 weeks and 0.37 and 0.51 at 24 weeks for the two groups exceed the reported minimum clinically important difference of 0.22 (45). A reduction of this magnitude in the level of disability is important, since previous studies have reported that both a shortened life span and increased health care costs correlate with higher HAQ scores (46, 47).

Anakinra administered daily for 24 weeks was well tolerated by most patients. None of the adverse events reported in this study were unexpected; all had been reported in earlier studies of anakinra in patients with RA (21, 48). Except for injection site reactions, no dose-related adverse events were evident. Since there were no independent assessors during the patient evaluation visits, there could be concern about unblinding of the trial due to the high level of injection site reactions. This is unlikely, since there was a broad range of injection site reactions across all of the treatment groups, including placebo. The frequency of these reactions ranged from 19% to 64% across the anakinra treatment groups, while 28% of the placebo group experienced injection site reactions. With this distribution, it is unlikely that the study would have been unblinded by the presence of injection site reactions.

The results of this trial demonstrate that the combination of anakinra and MTX offers a new alternative for patients with ongoing active disease despite MTX therapy. Long-term followup and an evaluation of the effect of this combination on retarding radiographic progression will determine the place for this therapy in the treatment armamentarium for patients with RA.

Acknowledgements

The assistance and cooperation of the authors' colleagues is acknowledged. The authors also thank the following principal investigators and their colleagues: Joel Block, Chicago, IL; Jane Box, Charlotte, NC; Peter Brooks, Brisbane Herston, Queensland, Australia; Ken Bulpitt, Los Angeles, CA; Jacques Caldwell, Gainesville, FL; Beverly Carpenter, Cincinnati, OH; Philip Clements, Los Angeles, CA; John Edmonds, Sydney, New South Wales, Australia; Eugene Fung, Waco, TX; Douglas Graham, Hinsdale, IL; Brian Grimmett, Cherry Hill, NJ; John Hanly, Halifax, Nova Scotia, Canada; Jeffrey Kaine, Sarasota, FL; Edward Keystone, Toronto, Ontario, Canada; Elliott Kopp, Raleigh, NC; Albert Lammaertino, Westmont, IL; Joseph Markenson, New York, NY; Richard Martin, Grand Rapids, MI; Robert McKendry, Ottawa, Ontario, Canada; Margaret Michalska, Chicago, IL; Nancy Olsen, Vanderbilt, TN; Jeffrey Poiley, Orlando, FL; Eric Ruderman, Chicago, IL; Marshall Sack, Austin, TX; Yvonne Scherrer, Oakland Park, FL; Thomas Schnitzer, Chicago, IL; William Shergy, Huntsville, AL; John Tesser, Phoenix, AZ; Elizabeth Tindall, Portland, OR; Peter Valen, La Crosse, WI; Ronald van Vollenhoven, Stanford, CA; and Arthur Weaver, Lincoln, NE.

Ancillary