Interleukin-6 receptor inhibition with tocilizumab reduces disease activity in rheumatoid arthritis with inadequate response to disease-modifying antirheumatic drugs: The tocilizumab in combination with traditional disease-modifying antirheumatic drug therapy study

Authors


  • ClinicalTrials.gov identifier: NCT00106574.

Abstract

Objective

To examine the efficacy and safety of the humanized anti–interleukin-6 receptor antibody tocilizumab combined with conventional disease-modifying antirheumatic drugs (DMARDs) in patients with active rheumatoid arthritis (RA).

Methods

A total of 1,220 patients were randomized (2:1 ratio) in the phase III, double-blind, placebo-controlled, multicenter TOWARD (Tocilizumab in Combination With Traditional DMARD Therapy) study. Patients remained on stable doses of DMARDs and received tocilizumab 8 mg/kg or placebo (control group) every 4 weeks for 24 weeks.

Results

At week 24, the proportion of patients achieving a response according to the American College of Rheumatology criteria for 20% improvement (ACR20) was significantly greater in the tocilizumab plus DMARD group than in the control group (61% versus 25%; P < 0.0001). Secondary end points including 50% or 70% improvement (ACR50/70), the Disease Activity Score in 28 joints (DAS28), DAS28 remission responses (DAS28 <2.6), European League Against Rheumatism responses, and systemic markers such as the C-reactive protein and hemoglobin levels showed superiority of tocilizumab plus DMARDs over DMARDs alone. Seventy-three percent of patients in the tocilizumab group had ≥1 adverse event (AE), compared with 61% of patients in the control group. AEs leading to withdrawal from the study were infrequent (4% of patients in the tocilizumab group and 2% of those in the control group). Serious AEs occurred in 6.7% and 4.3% of patients in the tocilizumab and control groups, respectively, and serious infections occurred in 2.7% and 1.9%, respectively. Elevations in the alanine aminotransferase level, from normal at baseline to >3-fold the upper limit of normal, occurred in 4% of patients in the tocilizumab group and 1% of those in the control group, and elevated total cholesterol levels were observed in 23% and 6% of patients, respectively. Sixteen patients started lipid-lowering therapy during the study. Grade 3 neutropenia occurred in 3.7% of patients receiving tocilizumab and none of the patients in the control group, and no grade 4 neutropenia was reported.

Conclusion

Tocilizumab combined with any of the DMARDs evaluated was safe and effective in reducing articular and systemic symptoms in patients with an inadequate response to these agents.

Rheumatoid arthritis (RA) is a chronic systemic inflammatory disease characterized by fatigue, joint pain, and joint swelling, resulting in loss of function, progressive disability, and increased morbidity and mortality (1, 2). RA is often treated with disease-modifying antirheumatic drugs (DMARDs) that relieve inflammatory processes and can slow disease progression (3). Although methotrexate is the most commonly used DMARD, other drugs such as leflunomide, sulfasalazine, and antimalarials, alone or in combination with methotrexate, are used (4). For patients with an inadequate response to conventional DMARDs, biologic agents that inhibit tumor necrosis factor α (TNFα) or B or T lymphocytes are indicated. Nonetheless, a substantial proportion of patients continue to have active synovitis and/or systemic symptoms or fail to maintain clinical benefit from these therapies (3, 5–8).

One potential therapeutic target is interleukin-6 (IL-6), a pleiotropic proinflammatory cytokine produced by a variety of cell types including lymphocytes, monocytes, and fibroblasts. IL-6 is involved in multiple immunologic processes such as T cell activation, B cell proliferation, initiation of acute-phase protein, and stimulation of hematopoietic precursor cell growth, differentiation, and trafficking (9). Of particular relevance to RA, IL-6 induces osteoclast differentiation, contributing to joint destruction and osteoporosis (10).

Tocilizumab is a humanized anti–IL-6 receptor (anti–IL-6R) monoclonal antibody that binds both soluble and membrane-expressed IL-6R, inhibiting IL-6–mediated proinflammatory activity (11). Because IL-6 serum levels correlate with disease activity, targeted inhibition of these effects with tocilizumab represents a novel approach to RA treatment (12–16). Previous phase II/III studies have demonstrated the safety and efficacy of tocilizumab in combination with methotrexate; preliminary evidence of inhibition of radiographic progression with tocilizumab monotherapy compared with DMARDs was recently shown in the SAMURAI study (Study of Active controlled Monotherapy Used for Rheumatoid Arthritis, an IL-6 inhibitor) in Japanese patients with RA (17–20).

The TOWARD (Tocilizumab in cOmbination With traditional DMARD therapy) study examined the efficacy and safety of tocilizumab in combination with a range of DMARDs in patients with moderate-to-severe RA in whom the response to these agents was inadequate.

PATIENTS AND METHODS

Patients.

Patients ages ≥18 years with moderate-to-severe RA of ≥6 months' duration, diagnosed according to the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) 1987 revised criteria for the classification of RA (21), with a swollen joint count (SJC) of ≥6, a tender joint count (TJC) of ≥8, and a C-reactive protein (CRP) level ≥1 mg/dl or an erythrocyte sedimentation rate (ESR) ≥28 mm/hour were enrolled. Eligible patients had received stable doses of permitted DMARDs (methotrexate, chloroquine, hydroxychloroquine, parenteral gold, sulfasalazine, azathioprine, and leflunomide) for ≥8 weeks prior to study entry. Oral glucocorticoids (≤10 mg/day prednisone or equivalent) and nonsteroidal antiinflammatory drugs (NSAIDs)/cyclooxygenase 2 inhibitors were permitted if the doses were stable for ≥6 weeks. Patients who were unsuccessfully treated with an anti-TNF agent or were previously treated with any cell-depleting therapy were excluded. Tuberculosis screening was managed according to local practice.

The study protocol was approved by relevant institutional review boards or ethics committees, and written informed consent was obtained from each patient. The study was conducted in full concordance with the principles of the Declaration of Helsinki and with the laws and regulations of the country in which the research was conducted.

Study drug.

Tocilizumab at a dose of 8 mg/kg and placebo were administered in a blinded manner, intravenously as a 60-minute infusion every 4 weeks. Dose selection was based on results of the phase II CHARISMA (Chugai Humanized Anti-Human Recombinant Interleukin-6 Monoclonal Antibody) study (17).

Study design.

The TOWARD study was a 24-week phase III, randomized, double-blind, placebo-controlled, international, multicenter study conducted in 18 countries (see Appendix A for the investigators and locations). Half of the study sites were US-based, enrolling 41% of patients. Patients were randomized in a 2:1 ratio to receive either tocilizumab or placebo combined with stable DMARD therapy (hereinafter, patients receiving tocilizumab plus DMARDs will be referred to as the tocilizumab group, and those receiving placebo plus DMARDs will be referred to as the control group). All patients received a stable dosage of folate (≥5 mg/week).

Patients were assessed using a dual-assessor approach for efficacy and safety evaluations, to ensure that blinding was not compromised. Tender and swollen joint counts (66 joints assessed and 68 joints assessed, respectively) were performed by a trained joint assessor who had no access to other patient data. No radiologic examinations were performed as part of this study. In patients who experienced an increase in the alanine aminotransferase (ALT) or aspartate aminotransferase (AST) level to ≥3-fold the upper limit of normal, the study treatment was interrupted; if a second ALT or AST elevation to ≥3-fold the upper limit of normal occurred when treatment recommenced, the study treatment was permanently discontinued. The study treatment was stopped in patients with an elevation in the ALT or AST level >5-fold the upper limit of normal, those with indirect bilirubin levels >2-fold the upper limit of normal or a total bilirubin value of >2.5 mg/dl, and in patients whose absolute neutrophil count decreased to <500 cells/mm3.

Patients who failed to achieve ≥20% improvement in both the SJC and the TJC by week 16 could receive rescue therapy consisting of adjustment of the background DMARD dosage and/or a different DMARD, and/or intraarticular/oral glucocorticoids. Such patients were nonresponders for the analysis but remained under study. Nonresponders and patients completing the study with the original treatment could enter an open-label long-term extension study, receiving tocilizumab 8 mg/kg every 4 weeks for up to 5 years.

Study end points.

The primary end point was the proportion of patients who had achieved a response according to the ACR criteria for 20% improvement (ACR20) (22) at week 24. Secondary end points included the proportion of patients with 50% or 70% improvement (ACR50/70) at week 24, as well as the time to onset of ACR20/50/70 responses. The Disease Activity Score in 28 joints (DAS28) (23) based on the erythrocyte sedimentation rate (ESR), the European League Against Rheumatism (EULAR) response (24), changes in hemoglobin levels, and disability index of the Health Assessment Questionnaire (HAQ) (25), Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-F) (26), and Short Form 36 (SF-36) (27) scores were also assessed at week 24.

Safety was assessed by examining adverse events (AEs), serious AEs, infections, withdrawals due to AEs, deaths, and clinically significant changes in vital signs and laboratory test results.

Statistical analysis.

Evaluation of at least 1,200 patients was planned to provide a sufficient number of patients to examine the safety of conventional DMARDs administered in combination with tocilizumab. The sample size with 2:1 randomization provided >90% power to detect an efficacy difference between the tocilizumab and control arms at week 24.

Efficacy analyses were conducted using the intent-to-treat population (ITT), which included all randomized patients who received ≥1 infusion of study treatment. Safety analyses included all randomized patients who received ≥1 infusion of study medication and had ≥1 postrandomization safety assessment. Patients who did not have the required data for a specific time point, who withdrew from the study, or who received rescue therapy were classified as nonresponders.

The primary end point, the proportion of patients with an ACR20 response at week 24, was compared using a Cochran-Mantel-Haenszel chi-square test with adjustment for site; this methodology was also used for the ACR50/70 response, DAS28 remission, and the EULAR response. Changes from baseline in the individual ACR core set parameters (28), the DAS28, the hemoglobin concentration, and FACIT-F and SF-36 scores were summarized by descriptive statistics. The difference between treatment groups for each component at week 24 was compared using an analysis of variance model with adjustment for site.

RESULTS

Patient demographics and baseline characteristics.

A total of 1,220 patients were randomized between March 24, 2005 and August 24, 2006. Of these patients, 1,216 received treatment. Of the 805 patients randomized to receive tocilizumab and the 415 patients randomized to the control group, 803 and 413, respectively, were included in the ITT population (Figure 1). During the study, 96 patients withdrew; 40 of these patients (4% in the tocilizumab group and 2% in the control group) withdrew due to safety reasons. Including patients receiving rescue treatment, 751 patients (93%) in the tocilizumab group and 370 patients (89%) in the control group completed 24 weeks of treatment. Moreover, 91% of patients randomized to receive tocilizumab completed the study on the initial therapy, compared with 78% of patients in the control group.

Figure 1.

Patient disposition. The 415 patients in the disease-modifying antirheumatic drugs (DMARDs) plus placebo group included 2 patients who were randomized to this group but did not receive any study medication and 2 patients who were randomized to this group but received tocilizumab 8 mg/kg plus DMARDs throughout the study. The 805 patients in the tocilizumab plus DMARDs group included 2 patients who were randomized to this group but did not receive any study medication and 3 patients who were randomized to this group but received placebo plus DMARDs throughout the study. Patients who did not achieve 20% improvement from baseline in both the swollen joint count and the tender joint count at week 16 were offered rescue therapy (adjustment of the background DMARD dosage and/or treatment with a different conventional DMARD). Patients receiving rescue therapy were considered nonresponders in the primary efficacy analysis at 24 weeks.

The 2 groups were well balanced with respect to demographics, baseline disease characteristics, and concomitant DMARD, glucocorticoid, and NSAID use at baseline (Table 1). Overall, 76% of patients were treated with 1 DMARD, and 23% were treated with ≥2 DMARDs. The most commonly used DMARD was methotrexate (mean dosage 15 mg/week).

Table 1. Demographics and baseline disease characteristics of the intent-to-treat population of patients with rheumatoid arthritis*
CharacteristicTocilizumab 8 mg/kg plus DMARDs (n = 803)Placebo plus DMARDs (n = 413)
  • *

    Except where indicated otherwise, values are the percent of patients. DAS28 = Disease Activity Score in 28 joints; anti-TNF = anti–tumor necrosis factor; NSAIDs = nonsteroidal antiinflammatory drugs; ESR = erythrocyte sedimentation rate; HAQ = Health Assessment Questionnaire; VAS = visual analog scale.

  • Safety population; for tocilizumab plus disease-modifying antirheumatic drugs (DMARDs), n = 802; for placebo plus DMARDs, n = 414. Patients can be counted within more than one category.

  • Patients contributing to summary statistic; for tocilizumab plus DMARDs, n = 609; for placebo plus DMARDs, n = 304.

Age, mean ± SD years53 ± 1354 ± 13
Weight, mean ± SD kg74 ± 1873 ± 18
Female sex8184
Race  
 White7272
 Asian910
 American Indian or Native Alaskan108
 Black47
 Other33
Disease duration  
 Mean ± SD years9.8 ± 8.89.8 ± 9.1
 Median (minimum–maximum) years7.0 (0.4–46.1)6.8 (0.5–44.4)
Disease duration <2 years1920
DAS28, mean ± SD6.7 ± 1.06.6 ± 1.0
Previous no. of DMARDs/anti-TNF agents taken, mean ± SD1.6 ± 1.61.6 ± 1.6
No. of background DMARDs  
 17775
 2 or more2224
 None11
Medication at baseline  
 Methotrexate75.873.9
 Chloroquine/hydroxychloroquine20.619.8
 Sulfasalazine13.114.3
 Leflunomide12.115.5
 Parenteral gold0.20.7
 Azathioprine2.22.2
 Oral steroids51.254.6
 Folic acid71.870.0
 NSAIDs71.477.1
 Other36.934.5
Methotrexate dosage at baseline, mean mg/week14.715.0
Tender joint count (66 joints assessed), mean ± SD30.1 ± 16.029.1 ± 14.8
Swollen joint count (68 joints assessed), mean ± SD19.7 ± 11.618.7 ± 10.8
C-reactive protein, mean ± SD mg/dl2.6 ± 3.22.6 ± 4.7
ESR, mean ± SD mm/hour48.2 ± 27.549.2 ± 28.3
HAQ disability index score (0–3 scale), mean ± SD1.5 ± 0.61.5 ± 0.6
Patient's assessment of pain, by VAS (0–100-scale), mean ± SD58 ± 2359 ± 23
Patient's assessment of disease activity, by VAS (0–100-mm scale), mean ± SD66 ± 2366 ± 24
Physician's assessment of disease activity, by VAS (0–100-mm scale), mean ± SD64 ± 1663 ± 17

Relevant comorbid conditions included hypertension (31% of patients in the tocilizumab and control groups), osteoporosis (12% in the tocilizumab and control groups), depression (10% and 8%, respectively), gastroesophageal reflux disease (10% and 8%, respectively), diabetes mellitus (9% and 8%, respectively), hyperlipidemia (6% and 7%, respectively), hypercholesterolemia (4% and 6%, respectively), and dyslipidemia (<1% in both groups).

Efficacy.

At week 24, the proportion of ACR20 responders was significantly higher in the tocilizumab group than in the control group (61% versus 25%; P < 0.0001) (Figure 2A). Tocilizumab in combination with any of the study DMARDs resulted in a higher proportion of ACR20 responders than did DMARDs plus placebo, with no obvious differences in response in patients receiving ≥2 DMARDs (Figure 2B). Significantly more patients treated with tocilizumab achieved ACR50 and ACR70 responses at week 24 compared with controls (for ACR50, 38% versus 9%; for ACR70, 21% versus 3% [P < 0.0001 for each]) (Figure 2A), and ACR50 responses were more frequent in patients receiving ≥2 DMARDs (43% of patients). The majority of patients in the tocilizumab group who achieved an ACR20 response also achieved an ACR50 response (62%, versus 37% of patients in the control group). A separation in both the ACR20 and ACR50 responses between the tocilizumab and control groups was apparent by the first scheduled assessment at week 2, and a separation was apparent by week 4 for the ACR70 response (Figure 2C). Moreover, there was a significant decrease from baseline in the number of swollen and tender joints in patients receiving tocilizumab compared with controls (for the SJC, −10.3 versus −4.9; for the TJC, −15.7 versus −8.5 [P < 0.0001 for each]). At week 24, a SJC of zero was recorded for 14% of patients in the tocilizumab group compared with 5% of those in the control group, with a TJC of zero for 12% and 4% of patients, respectively. Furthermore, 7% of patients in the tocilizumab group and 2% of patients in the control group had no swollen joints and no tender joints.

Figure 2.

A, Proportion of patients achieving a response according to the American College of Rheumatology criteria for 20% improvement (ACR20), 50% improvement (ACR50), and 70% improvement (ACR70) at week 24. ∗ = P < 0.0001 versus placebo plus DMARDs. B, Proportion of patients achieving an ACR20 response at week 24, according to background disease-modifying antirheumatic drug (DMARD) therapy. C, Proportion of patients achieving ACR20, ACR50, and ACR70 responses at each visit during the 24-week treatment period.

The mean DAS28 improved incrementally over time. By week 24, mean changes from baseline were greater in the tocilizumab group than in the control group (−3.17 and −1.16, respectively; P < 0.0001) (Figure 3A). Rates of remission responses (DAS28 <2.6) were also higher in the tocilizumab group versus the control group, with 30% of patients and 3% of patients, respectively, achieving clinical remission at week 24 (P < 0.0001) (Figure 3B). Moreover, 45% of patients in the tocilizumab group and 6% of patients in the control group experienced low disease activity (DAS28 ≤3.2). As early as the first scheduled assessment at week 2, 64% of patients in the tocilizumab group had achieved a good or moderate EULAR response, compared with 18% of the control group; by week 24, 80% of patients in the tocilizumab group and 38% of those in the control group had a good or moderate EULAR response (P < 0.0001) (Figure 3B).

Figure 3.

A, Mean change from baseline in the Disease Activity Score in 28 joints (DAS28) at week 24. B, Proportion of patients with a good or moderate European League Against Arthritis (EULAR) response and proportion of patients in clinical remission based on a DAS28 of <2.6. ∗ = P < 0.0001 versus placebo plus disease-modifying antirheumatic drugs (DMARDs).

The levels of inflammation markers decreased significantly in the tocilizumab group versus the control group by week 24 (mean CRP −2.20 versus −0.27 mg/dl; mean ESR −35.6 versus −4.7 mm/hour [P < 0.0001 for each]). In contrast to the small decrease seen in the control group, normalization of mean CRP levels was observed as early as the first visit (week 2) in the tocilizumab group, with maintenance of low levels throughout the study (Figure 4A).

Figure 4.

A, Mean change from baseline in C-reactive protein (CRP) levels over the 24-week study. B, Mean change from baseline in hemoglobin levels over the 24-week study. ∗ = P < 0.0001 versus control group. DMARDs = disease-modifying antirheumatic drugs; ULN = upper limit of normal; LLN = lower limit of normal.

Moreover, mean hemoglobin levels increased as early as week 2, with incremental improvement over time in patients treated with tocilizumab. At week 24, the control group showed essentially no change (mean −0.13 gm/dl), whereas a mean improvement of 0.98 gm/dl was observed in the tocilizumab group (P < 0.0001) (Figure 4B). In patients with baseline hemoglobin levels below the lower limit of normal, the mean increase with tocilizumab was even greater at week 24 (1.7, versus 0.2 gm/dl in the control group).

Furthermore, 60% of patients in the tocilizumab group had a clinically meaningful improvement in physical function at week 24, compared with 34% of patients in the control arm (defined as a change from baseline in the disability index of the HAQ of ≥0.3) (29). At week 24, mean changes from baseline were also significantly higher in the tocilizumab group compared with the control group for the disability index of the HAQ (−0.5 versus −0.2) and FACIT-F scores (8.0 versus 3.6) (P < 0.0001 for each). Mean improvements from baseline in SF-36 scores were higher in the tocilizumab group than in the control group for the both physical (8.9 versus 4.1) and mental (5.3 versus 2.3) components at week 24 (P < 0.0001 for each).

Safety.

The overall incidence of AEs was higher in the tocilizumab group compared with the control group (73% versus 61%) (Table 2). More than 90% of AEs in each group were of mild or moderate intensity. Serious AEs, AEs leading to withdrawal, and AEs leading to dose modification occurred more frequently in the tocilizumab group than in the control group (Table 2). The greatest numbers of patients reporting AEs were in the following background DMARD groups: methotrexate (73% of those receiving tocilizumab and 61% of controls), 2 background DMARDs (76% and 69%, respectively), and leflunomide (67% and 47%, respectively). However, because these groups also included the largest number of patients, no causal relationship can be drawn.

Table 2. Summary of adverse events (safety population)*
 Tocilizumab 8 mg/kg plus DMARDs (n = 802)Placebo plus DMARDs (n = 414)
  • *

    Values are the number (%) of patients. DMARDs = disease-modifying antirheumatic drugs.

Any adverse event584 (72.8)253 (61.1)
Serious adverse event54 (6.7)18 (4.3)
Related serious adverse event23 (2.9)6 (1.4)
Adverse event leading to discontinuation31 (3.9)8 (1.9)
Adverse event leading to dose modification111 (13.8)30 (7.2)
Death2 (<1)2 (<1)
Adverse events occurring in ≥5% of patients  
 Infections and infestations300 (37.4)131 (31.6)
 Gastrointestinal disorders167 (20.8)61 (14.7)
 Musculoskeletal and connective tissue disorders104 (13.0)74 (17.9)
 Skin and subcutaneous tissue disorders133 (16.6)29 (7.0)
 Nervous system disorders93 (11.6)36 (8.7)
 Laboratory investigations94 (11.7)11 (2.7)
 General disorders and administration-site conditions66 (8.2)30 (7.2)
 Injury, poisoning, and procedural complications63 (7.9)29 (7.0)
 Respiratory, thoracic, and mediastinal disorders69 (8.6)21 (5.1)
 Vascular disorders54 (6.7)21 (5.1)
Serious adverse events occurring in ≥3 patients54 (6.7)18 (4.3)
 Total number of events6720
 Infections and infestations22 (2.7)8 (1.9)
 Musculoskeletal disorders2 (0.2)3 (0.7)
 Gastrointestinal disorders9 (1.1)1 (0.2)
 Nervous system disorders6 (0.7)2 (0.5)
 Cardiac disorders3 (0.4)1 (0.2)
 Injuries, poisoning7 (0.9)
 Renal and urinary disorders3 (0.4)
Serious infections occurring in ≥2 patients22 (2.7)8 (1.9)
 Total number of events238
 Cellulitis5 (0.6)
 Pneumonia3 (0.4)2 (0.5)
 Herpes zoster3 (0.4)

The incidence of infections was 37% in the tocilizumab group and 32% in the control group, with rates per 100 patient-years of 120.1 for the tocilizumab group and 95.3 for the control group. Infections with a higher incidence in the tocilizumab group versus the control group were upper respiratory tract infections (9% versus 7%), other respiratory infections (12% versus 10%), and skin and subcutaneous tissue infections (5% versus 3%). Most patients who experienced infections continued treatment without recurrence. The incidence of serious infections was 3% in the tocilizumab group and 2% in the control group (Table 2). The rate of serious infections per 100 patient-years of exposure was 5.9 in the tocilizumab group and 4.7 in the control group. Five serious infections led to study treatment discontinuation: staphylococcal cellulitis, acute pyelonephritis, and sepsis in the tocilizumab group, and 2 cases of pneumonia in the control group. In the 22 patients with serious infections in tocilizumab group, there was no apparent association with background DMARDs, but the majority of patients in this group were also taking glucocorticoids. There were no cases of tuberculosis, although 1 case of asymptomatic Mycobacterium avium-intracellulare was diagnosed in a patient in the tocilizumab group, who had a history of gastroesophageal reflux disease.

The incidence of gastrointestinal disorders was higher in the tocilizumab group than in the control group (Table 2); these disorders included nausea (4% versus 3%), upper abdominal pain (3% versus 1%), mouth ulceration (2% versus 1%), and stomatitis (1% versus 0%). Upper gastrointestinal tract events suggestive of gastric inflammation, gastritis, and ulcer occurred more frequently in the tocilizumab group (3%) than in the control group (1%). These events were readily treated and did not result in withdrawal from the study.

Skin and subcutaneous tissue disorders were experienced more frequently by patients in the tocilizumab group than by patients in the control group (Table 2) and included mild or moderate skin effects (e.g., rash, dermatitis, pruritus, urticaria). With the exception of 1 urticarial event reported during the first infusion in a tocilizumab-treated patient with a strong allergic history, these events were not treatment limiting.

Other events occurring more commonly in the tocilizumab group than in the control group included headache (6% versus 4%) and hypertension (5% versus 3%). A small proportion of these events were reported during and following infusion. Nevertheless, the majority of cases were considered mild in intensity and had little impact on continued study participation. One patient in the tocilizumab group experienced a non–life-threatening anaphylactic reaction and was withdrawn from the study, and 1 patient in this group withdrew from the study due to nonserious hypersensitivity. Two patients in each treatment group died during the 24-week study period. The causes of death were hemorrhagic stroke and postprocedural complications after triple coronary artery bypass graft in the tocilizumab group, and pneumonia and intestinal obstruction in the control group.

A higher proportion of patients in the tocilizumab group had a decrease in the neutrophil count, from high baseline values to normal (>1,500–1,800 cells/mm3) postbaseline values (5%) and from normal baseline values to low postbaseline values (29%), compared with <1% and 4%, respectively, of patients in the control group. The majority of these abnormalities were grade 1 neutropenia (lower limit of normal 1,500 cells/mm3) and grade 2 neutropenia (<1,500–1,000 cells/mm3), with few patients in the tocilizumab group experiencing transient grade 3 neutropenia (<1,000–500 cells/mm3) (Figure 5A). There was no association between low neutrophil counts and the occurrence of infection-related AEs, and none of the patients with grade 3 neutropenia discontinued study treatment because of this event. Three patients had their study medication temporarily interrupted, and 5 patients had their DMARD dosage modified, with subsequent improvement in neutrophil count.

Figure 5.

A, Proportion of patients experiencing reduced neutrophil counts during the 24-week study. B, Proportion of patients with elevations in liver aminotransferase levels and total bilirubin levels (highest elevations at any time point). C, Proportion of patients experiencing elevations in cholesterol levels. ∗ = improvement in the high-density lipoprotein (HDL) cholesterol level. DMARDs = disease-modifying antirheumatic drugs; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ULN = upper limit of normal; LDL = low-density lipoprotein.

Changes in hepatic aminotransferase and bilirubin levels.

More patients in the tocilizumab group than in the control group had an increase in the ALT or ALT/AST level at some point during the study; however, the majority of these elevations were <3-fold the upper limit of normal (Figure 5B). The highest mean increases in hepatic aminotransferase levels were observed 2 weeks after each infusion, during the active portion of the study.

Most elevations in the bilirubin level from normal at baseline were ≤3-fold the upper limit of normal in both groups and were associated with an increase in the indirect bilirubin level in the tocilizumab group, with 1 patient in the control group having a single elevation >3-fold the upper limit of normal (Figure 5B). No patients in the tocilizumab group experienced an increase in the ALT and AST levels to >3-fold the upper limit of normal and a simultaneous increase in the total bilirubin level >2-fold the upper limit of normal during the study. Although no clinical signs or symptoms of hepatic injury were associated with increases in the ALT, AST, or bilirubin level, 3 patients in the tocilizumab group discontinued treatment (1 patient discontinued because of elevated ALT levels, and 2 patients discontinued due to elevated bilirubin levels without concurrent elevations in the hepatic aminotransferase level).

Changes in lipid levels.

Increases in the mean fasting plasma lipid levels (within the normal range) were observed in the tocilizumab group only and were apparent from the time of the first scheduled testing at week 6. Levels remained elevated throughout the study, with mean changes from baseline at week 24 as follows: for total cholesterol, 0.8 mmole/liter; for high-density lipoprotein (HDL) cholesterol, 0.1 mmole/liter; for low-density lipoprotein (LDL) cholesterol, 0.5 mmole/liter. An increase in the total cholesterol level from <240 mg/dl at baseline to ≥240 mg/dl at week 24 was observed in 23% of patients in the tocilizumab group and 5.5% of patients in the control group, and increases in the LDL level from <160 mg/dl at baseline to ≥160 mg/dl at week 24 occurred in 16% and 3% of patients, respectively (Figure 5C). During this study, 16 patients with elevated lipid levels started therapy with statins, which effectively lowered or stabilized cholesterol levels. An improvement in the HDL cholesterol level from <60 mg/dl at baseline to ≥60 mg/dl was also more frequent in the tocilizumab group compared with the control group (15% versus 6%) (Figure 5C). Increases of >30% in the total cholesterol:HDL cholesterol ratio were reported in 12% and 7% of patients in the tocilizumab and control groups, respectively, and increases of >30% in the LDL:HDL cholesterol ratio were observed in 20% and 12% of patients, respectively. Increases in the mean apoliprotein A-I (Apo A-I) and Apo B values (within the normal range) were observed in the tocilizumab group beginning at week 6; however, there was no change in the Apo B:Apo A-I ratio in the majority of patients in the tocilizumab group, with 10% and 8% of patients in the tocilizumab and control groups, respectively, experiencing increases of >30%. Moreover, 7% of patients in the tocilizumab group and 5% of patients in control group experienced a decrease of >30% in the Apo B:Apo A-I ratio.

The mean triglyceride values rose to just above the upper limit of normal at week 6 in the tocilizumab group. Changes in the triglyceride level from <500 mg/dl at baseline to ≥500 mg/dl occurred in 1.4% of patients in the tocilizumab group and <1% of patients in the control group. No clinical symptoms, in particular pancreatitis, were associated with an elevation in the triglyceride level.

DISCUSSION

The use of biologic agents represents an advance in the treatment of RA. However, there are still nonresponders, inadequate responders, patients in whom responsiveness to treatment is lost over time, and patients in whom safety problems develop. Thus, some patients may benefit from treatment with newer biologic agents with a different mechanism of action (30, 31).

TOWARD is the largest study of the novel therapy tocilizumab and the first to show improvement in signs and symptoms with tocilizumab 8 mg/kg in combination with a wide range of DMARDs in patients with moderate-to-severe RA that has responded inadequately to conventional antirheumatic therapy. Moreover, the results show that treatment with the combination of tocilizumab plus DMARDs has a high likelihood of inducing a clinically important benefit in terms of ACR50 and ACR70 response rates and DAS28 remission responses. Tocilizumab was well tolerated in combination with conventional DMARDs, and the safety profile was not affected by the type or number of DMARDs used.

Clinically relevant improvements were seen as early as week 2, with incremental improvement of signs and symptoms over time in patients treated with tocilizumab and DMARDs. The reduction in systemic inflammation was evidenced by the rapid and sustained decrease in acute-phase response markers (the CRP level and the ESR). Importantly, more patients treated with tocilizumab had a clinically meaningful improvement in physical function (the disability index of the HAQ), health-related quality of life (SF-36), and fatigue (FACIT-F). These benefits were associated with a low withdrawal rate due to lack of efficacy in the tocilizumab arm compared with the control arm.

Anemia, a common systemic manifestation of RA, occurs in ∼30% of patients with RA and correlates with levels of disease activity (32). The significant improvement in hemoglobin levels to normal values in this study indicates that tocilizumab treatment has the potential to improve chronic anemia in patients with RA. IL-6 stimulates the production of hepcidin, a liver peptide that modulates hemoglobin production by restricting iron availability (1, 33, 34). Thus, effective blockade of IL-6R signaling is expected to decrease hepcidin levels and increase iron availability for hemoglobin production. There was neither an increase in erythrocyte counts nor an increase in mean hemoglobin levels above normal levels.

The sample size of the study provided an opportunity to identify any major safety issues when tocilizumab is used in combination with commonly prescribed DMARDs. Patients were recruited based largely on practice patterns and not with a goal to assure representation of each DMARD at a similar frequency. Nevertheless, for methotrexate, leflunomide, antimalarials, and sulfasalazine, there was sufficient representation to detect safety interactions that might occur commonly based on pharmacologic effects, i.e., elevations in hepatic aminotransferase levels, neutropenia, or elevations in the lipid level, as well as serious infections. Despite the fact that diverse DMARDs were used in combination with tocilizumab, the incidence and severity of AEs were consistent with those described in previous reports, with the most common AEs being upper respiratory tract infections (17, 19). The overall incidence of serious infections was low, and only a small proportion of patients with serious infections withdrew from the study. Only 1 infection was considered “opportunistic” in the tocilizumab group: an M avium-intracellulare infection diagnosed based on a change in chest radiography, in a patient with no clinical symptoms and with a history of gastroesophageal reflux disease.

Other AEs occurring more frequently in the tocilizumab group were gastrointestinal disorders and skin disorders. At least 1 risk factor, such as concomitant treatment with NSAIDs or glucocorticoids, was present in patients with gastrointestinal tract events. Skin rashes did not suggest hypersensitivity to tocilizumab, and only 1 anaphylactic reaction occurred. Infusion reactions observed within 24 hours of infusion occurred infrequently, and most events defined as infusion reactions were nonspecific and non–treatment-limiting (e.g., transient blood pressure elevations).

A greater reduction in the mean number of neutrophils was seen in the tocilizumab group compared with the control group, consistent with expression of IL-6R on neutrophils and a marked decrease in the acute-phase response. In patients with transient decreases in the neutrophil count to <1,000 mm3, the decreases were not temporally associated with infection. The transient nature of grade 3 neutropenia, the lack of association with infection in this 24-week study, and the lack of need to adjust concomitant treatment suggest that this effect is not a significant issue; however, evaluation of the impact of lower neutrophil counts during long-term treatment will require long-term followup of this and other cohorts.

Approximately 45% of tocilizumab-treated patients had a transient elevation above the upper limit of normal in the level of hepatic ALT, with or without an elevation in the AST level, similar to the incidence observed in other studies (17, 20). In most patients, these were single or infrequent recurrent observations, with normalization spontaneously or after temporary interruption of study treatment and/or lowering of the methotrexate dosage. Although the mean bilirubin levels increased within the normal range in the tocilizumab group, significant elevations in the total bilirubin level did not occur, and there was no association with elevations of the aminotransferase level. There was no evidence that elevations in the hepatic aminotransferase levels were associated with any particular type of DMARD or DMARD combination used in this study or any particular comorbidities. Importantly, there was no clinical evidence of hepatitis or hepatic dysfunction. In 1 patient with type 2 diabetes and recurrent elevations of the aminotransferase level <3-fold the upper limit of normal, a liver biopsy showed fatty liver, and the patient continued with tocilizumab treatment without any adverse consequence.

In this study, mean fasting plasma lipid levels increased upon treatment with tocilizumab, and this increase appeared to coincide with a large decrease in CRP levels. The increase in lipid levels may, in part, be a consequence of effective reduction of inflammation. This effect is consistent with the observations in several studies, however, that elevation of lipids needs further scrutiny (35–38). Although guidelines were included in the protocol, treatment with statins was initiated in very few patients during the study. Patients who did start statin therapy based on the guidelines for lipid management recommended by the Adult Treatment Panel III (National Cholesterol Education Program) and had a followup sample evaluated appeared to have responded to lipid-lowering treatment (mean reduction in total cholesterol levels of 75 mg/dl) (39). It has been suggested that statins may have antiinflammatory properties (40); however, given the limited number of patients starting statin therapy during the study, this should have little impact on the overall response rates.

In the present study of patients with moderate-to-severe RA, treatment with tocilizumab in combination with traditional DMARDs significantly and rapidly reduced disease activity over 24 weeks as compared with treatment with DMARDs plus placebo. Elevations in liver aminotransferase and cholesterol levels were not associated with clinical symptoms. Moreover, although this trial was not powered to examine the correlation between neutropenia and the incidence of infections, there did not appear to be any association between decreases in neutrophil counts and the occurrence of infections. There were no noticeable differences in efficacy and safety responses according to background DMARDs or the number of concomitant DMARDs.

AUTHOR CONTRIBUTIONS

Dr. Genovese 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 design. Genovese, Alecock.

Acquisition of data. Genovese, McKay, Nasonov, da Silva, Alecock, Woodworth, Gomez-Reino.

Analysis and interpretation of data. Genovese, Mysler, Alecock, Woodworth, Gomez-Reino.

Manuscript preparation. Genovese, McKay, Mysler, da Silva, Alecock, Woodworth, Gomez-Reino, and Dr. Marcela Oancea (nonauthor; Phocus Services [supported by F. Hoffman-La Roche Ltd]).

Statistical analysis. Alecock.

Acknowledgements

We would like to acknowledge the collaboration and commitment of all investigators and their staffs, without whom the present study would not have been possible.

APPENDIX A

THE TOWARD INVESTIGATORS

The TOWARD investigators are as follows: J. A. Maldonado-Cocco (Buenos Aires, Argentina), E. Mysler (Buenos Aires, Argentina), D. Siri (Rosario, Argentina); P. Youssef (Sydney, New South Wales, Australia); R. Bonfiglioli (Campinas, Brazil), G. Castelar-Pinheiro (Rio de Janeiro, Brazil), N. da Silva (Goiânia, Brazil); J. Brown (Sainte-Foy, Quebec, Canada), A. Cividino (Hamilton, Ontario, Canada), T. McCarthy (Winnipeg, Manitoba, Canada), R. J. R. Mckendry (Ottawa, Ontario, Canada), B. Nair (Saskatoon, Saskatchewan, Canada); S. Chen (Shanghai, China), X. Han (Shanghai, China), Z. Li (Beijing, China), C. C. Mok (Tuen Mun, China), M. Y. Mok (Hong Kong, China), Y. Zhao (Beijing, China); P. Monge (San Jose, Costa Rica); P. Bradna (Hradec Kralove, Czech Republic), K. Pavelka (Prague, Czech Republic), M. Sedlackova (Prague, Czech Republic); P. Hannonen (Jyvaeskylae, Finland), L. Paimela (Helsinki, Finland); G. Chales (Rennes, France), B. Combe (Montpellier, France), P. Gaudin (La Tronche, France), J. Godde (Marseille, France), J. M. le Parc (Boulogne-Billancourt, France), Y. Maugars (Nantes, France), S. Perrot (Paris, France), J. G. Tebib (Pierre Benite, France), T. Thomas (Saint-Etienne, France); G. Gauler (Osnabrück, Germany), M. Hammer (Sendenhorst, Germany), H. Kellner (Munich, Germany), R. Kurthen (Aachen, Germany), C. Specker (Essen, Germany), U. Von Hinueber (Hildesheim, Germany); G. E. Lugo (Mexico City, Mexico), P. Martinez (Mexico City, Mexico), J. Morales (Leon, Mexico), C. Saldate Alonso (Tijuana, Mexico); G. Guerra (Panama City, Panama); V. I. Mazurov (St. Petersburg, Russia), E. L. Nasonov (Moscow, Russia), T. G. Shemerovskaya (St. Petersburg, Russia), E. I. Shmidt (Moscow, Russia), N. A. Shostak (Moscow, Russia), V. Sorotskaya (Tula, Russia), V. Tyurin (Moscow, Russia), S. Yakushin (Ryazan, Russia); S. Nayiager (Durban, South Africa), B. Traub (Radiokop, South Africa), E. Van Duuren (Pretoria, South Africa); J. Cruz Martinez (Pontevedra, Spain), J. J. Gomez Reino (Santiago de Compostela, Spain); S. R. Dahlqvist (Umea, Sweden), R. van Vollenhoven (Stockholm, Sweden); T. Kitumnuaypong (Bangkok, Thailand), W. Louthrenoo (Chiang Mai, Thailand), M. Osiri (Bangkok, Thailand); B. Bode (Tucson, AZ), D. G. Borenstein (Washington, DC), J. J. Budd III (St. Louis, MO), H. Busch (Jupiter, FL), M. Churchill (Lincoln, NE), C. Codding (Oklahoma City, OK), S. Cooper (Burlington, VT), M. E. Cronin (Milwaukee, WI), J. J. Curran (Chicago, IL), R. Ettlinger (Tacoma, WA), M. Fairfax (Phoenix, AZ), I. F. Fenton (Vernon Hills, IL), J. Fiechtner (Lansing, MI), M. Fisher (Haddon Heights, NJ), M. C. Franklin (Willow Grove, PA), A. Fraser (Bowling Green, KY), E. Fudman (Austin, TX), M. C. Genovese (Palo Alto, CA), M. Goldberg (Passaic, NJ), M. Greenwald (Palm Desert, CA), D. Halter (Houston, TX), R. A. Harrell III (Durham, NC), A. Hou (Upland, CA), E. Huffstutter (Hinson, TN), S. Javed (Newark, DE), H. M. Kenney (Spokane, WA), P. Kent (Minneapolis, MN), E. Kopp (Raleigh, NC), M. Lowenstein (Palm Harbor, FL), C. L. Ludivico (Bethlehem, PA), C. Lue (Little Rock, AR), M. Luggen (Cincinnati, OH), J. Marshall (Baton Rouge, LA), T. E. McAlindon (Boston, MA), J. D. Mckay (Tulsa, OK), J. L. Miller (Tampa, FL), C. Multz (San Jose, CA), E. Peters (Paradise Valley, AZ), A. J. Porges (Roslyn, NY), M. Prupas (Reno, NV), P. Riccardi (Syracuse, NY), D. Ridley (Eagan, MN), W. C. Rizzo (Scottsdale, AZ), J. M. Ross (Allentown, PA), P. Saxe (Delray Beach, FL), J. Schechtman (Peoria, AZ), A. I. Sebba (Palm Harbor, FL), N. Singer (Beachwood, OH), A. K. Singhal (Mesquite, TX), S. B. Tanner (Nashville, TN), M. L. Tiku (New Brunswick, NJ), J. S. Toder (Johnston, RI), R. Trapp (Springfield, IL), J. Udell (Philadelphia, PA), P. A. Valen (La Crosse, WI), M. Chester Wasko (Pittsburgh, PA), N. Wei (Frederick, MD), L. Willis (Oklahoma City, OK), A. E. Winkler (Springfield, MO), S. Wolfe (Dayton, OH).

Ancillary