Efficacy and Safety of Tabalumab in Rheumatoid Arthritis
Efficacy and Safety of Tabalumab in Rheumatoid Arthritis
This multicenter, randomized, placebo-controlled, double-blind phase 3 study (Trial Registration NCT01202760) examined the superiority of tabalumab 120/Q4W or 90/Q2W relative to placebo over 24 weeks in patients with RA on ACR20. Patients were enrolled from 23 countries (Argentina, Australia, Bulgaria, Colombia, Croatia, Hungary, India, Japan, Korea, Lithuania, Malaysia, Mexico, New Zealand, Peru, Poland, Romania, Russia, Slovakia, South Africa, Sri Lanka, Taiwan, Ukraine, and the United States).
The study protocol was approved by the local institutional review boards in accordance with the Declaration of Helsinki and applicable laws and regulations. All patients provided voluntary written informed consent.
Men and women (≥18 years) who had adult-onset RA for 6 months to 15 years, met the ACR (1987 revised) RA criteria, had a Patient's Global Assessment of Disease Activity (PtGA) 20 mm or more of 100 mm on a visual analog scale (VAS), and were in ACR functional class I, II, or III were enrolled. Patients taking conventional DMARDs must have been on a stable dose for 8 weeks or more prior to baseline. Women could not be pregnant or breast-feeding during study participation. Although the study did not require a minimum number of tender or swollen joints for inclusion, in order to obtain a sufficient number of patients with more active disease to assess the primary objective, 50% of enrolled patients or more were required to have 5 or more of 68 tender and 5 or more of 66 swollen joints at baseline.
Key exclusion criteria were use of oral corticosteroids at more than 10 mg/d of prednisone or its equivalent or any parenteral corticosteroid injection within 6 weeks of baseline; current use of biologic DMARDs (must have stopped TNF-α inhibitors or anakinra >28 days, tocilizumab >56 days, abatacept >90 days, rituximab or any other B-cell biotherapy >1 year, or any other biologic DMARD >5 half-lives prior to baseline); use of unstable dose of nonsteroidal anti-inflammatory drugs within 6 weeks of baseline; any serious infection within 3 months of baseline; a bone or joint infection within 6 months of baseline; an active or recent opportunistic infection; a malignancy within 5 years of baseline; a significant uncontrolled medical illness; active hepatitis B or C or HIV; or evidence of active or latent tuberculosis.
Patients were randomized 3:3:2 to receive 120/Q4W, 90/Q2W, or placebo, respectively. Patients were stratified by baseline disease activity (with or without ≥5/68 tender and ≥5/66 swollen joints) and background DMARD therapy to properly evaluate the primary efficacy outcome in patients with more active disease. At baseline (week 0), the 120/Q4W group received a 240-mg loading dose, the 90/Q2W group received a 180-mg loading dose, and the placebo group received placebo injections. Thereafter, patients received study drug every 2 weeks for 24 weeks (last dose administered at week 22); patients in the 120/Q4W group alternated tabalumab and placebo injections during the treatment period. Patients who were nonresponders at week 16 were assigned to receive 90/Q2W for the remainder of the treatment period. For patients who had 5 or more swollen and 5 or more tender joint counts at baseline, nonresponders were defined as patients with less than 20% improvement from baseline in both tender and swollen joint counts. For patients who did not have 5 or more swollen and 5 or more tender joint counts, nonresponders were defined as patients who had 2 or more additional tender and 2 or more additional swollen joints from baseline. After week 24 (or the early discontinuation visit), patients entered a 24- to 48-week posttreatment follow-up. During the 24- to 48-week follow-up, CD20 B cells were followed for each patient until B-cell recovery occurred. B-cell recovery was defined as an absolute B-cell count of 43 cells/μL or more (based on the lower limit of normal among RA patients from 5 previous clinical trials of tabalumab) (data on file) or an absolute B-cell count that was equal to or greater than 50% of the patient's baseline count.
The primary efficacy end point was the proportion of patients achieving ACR20 at week 24. Secondary efficacy end points were the proportion of patients achieving ACR50 and ACR70, mean change in ACR core set, ACR percent improvement (ACR-N), Disease Activity Score in 28 joints c-reactive protein (DAS28-CRP), the Health Assessment Questionnaire–Disability Index (HAQ-DI), and the percentage of good or moderate responders on the European League Against Rheumatism Responder Index in 28 joints (EULAR-28).
Blood samples were collected to evaluate changes in serum immunoglobulin (IgG, IgM, IgA) levels, to assess the development of antitabalumab antibodies (ADAs) and the effect of tabalumab on B cell counts, and to assess rheumatoid factor (RF) and anticyclic citrullinated peptide (anti-CCP) levels. Immunoglobulin concentrations were assayed using standard methods. Antitabalumab antibodies were assayed using validated methods. B cells were assayed by flow cytometry, from which the time course of changes in absolute B-cell counts (CD3-CD20) was assessed. Rheumatoid factor was assessed using standard methods; anti-CCP was assessed using standard enzyme-linked immunosorbent assay methods. Safety was assessed by examining laboratory tests, vital signs, electrocardiograms, adverse events (AEs), and serious AEs (SAEs).
Primary and secondary efficacy analyses were conducted using a prespecified subset of patients who had 5 or more of 68 tender and 5 or more of 66 swollen joints at baseline (efficacy population); these analyses were repeated using the intention-to-treat (ITT) population (all randomized patients). Safety analyses included all randomized patients who received 1 or more dose of study drug (safety population).
A sample size of 1002 randomized patients was anticipated to enroll 501 patients or more who would meet efficacy population criteria. This efficacy population was expected to provide 95% power or greater to detect a difference of 20% in ACR20 response rates at week 24 between each tabalumab group and the placebo group, assuming a placebo response rate of 23%.
For baseline demographics and characteristics, Fisher exact test was used for categorical data, and analysis of variance was used for continuous data. Analyses of categorical efficacy measures (ie, ACR20, ACR50, ACR70, EULAR-28) were conducted using a logistic regression model with treatment, geographic region, baseline DAS28-CRP, TNF–inadequate responders (TNF-IR) treatment history, and background DMARD at baseline as factors. When the logistic regression sample size requirements were not met, Fisher exact test was used. Fisher exact test was also used for categorical safety variables. For continuous efficacy and safety variables (ie, individual components of the ACR, DAS28-CRP, HAQ-DI, CRP, laboratory analytes, vital signs, serum immunoglobulins, and B cells), an analysis-of-covariance model that included treatment, geographic region, TNF-IR treatment history, background DMARD at baseline, and baseline value was used to test the difference of each tabalumab group versus the placebo group.
For ACR20, ACR50, and ACR70 responses, patients who were week 16 nonresponders or who discontinued prior to week 24 were imputed as nonresponders (NRI). For all other measures, the last postbaseline value at or before week 16 for nonresponders and before discontinuation for discontinued patients was used (modified last observation carried forward [mLOCF]) in evaluation. All statistical tests of treatment effects were performed at 2-sided significance levels of 0.05.
Methods
This multicenter, randomized, placebo-controlled, double-blind phase 3 study (Trial Registration NCT01202760) examined the superiority of tabalumab 120/Q4W or 90/Q2W relative to placebo over 24 weeks in patients with RA on ACR20. Patients were enrolled from 23 countries (Argentina, Australia, Bulgaria, Colombia, Croatia, Hungary, India, Japan, Korea, Lithuania, Malaysia, Mexico, New Zealand, Peru, Poland, Romania, Russia, Slovakia, South Africa, Sri Lanka, Taiwan, Ukraine, and the United States).
The study protocol was approved by the local institutional review boards in accordance with the Declaration of Helsinki and applicable laws and regulations. All patients provided voluntary written informed consent.
Patients
Men and women (≥18 years) who had adult-onset RA for 6 months to 15 years, met the ACR (1987 revised) RA criteria, had a Patient's Global Assessment of Disease Activity (PtGA) 20 mm or more of 100 mm on a visual analog scale (VAS), and were in ACR functional class I, II, or III were enrolled. Patients taking conventional DMARDs must have been on a stable dose for 8 weeks or more prior to baseline. Women could not be pregnant or breast-feeding during study participation. Although the study did not require a minimum number of tender or swollen joints for inclusion, in order to obtain a sufficient number of patients with more active disease to assess the primary objective, 50% of enrolled patients or more were required to have 5 or more of 68 tender and 5 or more of 66 swollen joints at baseline.
Key exclusion criteria were use of oral corticosteroids at more than 10 mg/d of prednisone or its equivalent or any parenteral corticosteroid injection within 6 weeks of baseline; current use of biologic DMARDs (must have stopped TNF-α inhibitors or anakinra >28 days, tocilizumab >56 days, abatacept >90 days, rituximab or any other B-cell biotherapy >1 year, or any other biologic DMARD >5 half-lives prior to baseline); use of unstable dose of nonsteroidal anti-inflammatory drugs within 6 weeks of baseline; any serious infection within 3 months of baseline; a bone or joint infection within 6 months of baseline; an active or recent opportunistic infection; a malignancy within 5 years of baseline; a significant uncontrolled medical illness; active hepatitis B or C or HIV; or evidence of active or latent tuberculosis.
Treatments
Patients were randomized 3:3:2 to receive 120/Q4W, 90/Q2W, or placebo, respectively. Patients were stratified by baseline disease activity (with or without ≥5/68 tender and ≥5/66 swollen joints) and background DMARD therapy to properly evaluate the primary efficacy outcome in patients with more active disease. At baseline (week 0), the 120/Q4W group received a 240-mg loading dose, the 90/Q2W group received a 180-mg loading dose, and the placebo group received placebo injections. Thereafter, patients received study drug every 2 weeks for 24 weeks (last dose administered at week 22); patients in the 120/Q4W group alternated tabalumab and placebo injections during the treatment period. Patients who were nonresponders at week 16 were assigned to receive 90/Q2W for the remainder of the treatment period. For patients who had 5 or more swollen and 5 or more tender joint counts at baseline, nonresponders were defined as patients with less than 20% improvement from baseline in both tender and swollen joint counts. For patients who did not have 5 or more swollen and 5 or more tender joint counts, nonresponders were defined as patients who had 2 or more additional tender and 2 or more additional swollen joints from baseline. After week 24 (or the early discontinuation visit), patients entered a 24- to 48-week posttreatment follow-up. During the 24- to 48-week follow-up, CD20 B cells were followed for each patient until B-cell recovery occurred. B-cell recovery was defined as an absolute B-cell count of 43 cells/μL or more (based on the lower limit of normal among RA patients from 5 previous clinical trials of tabalumab) (data on file) or an absolute B-cell count that was equal to or greater than 50% of the patient's baseline count.
End Points
The primary efficacy end point was the proportion of patients achieving ACR20 at week 24. Secondary efficacy end points were the proportion of patients achieving ACR50 and ACR70, mean change in ACR core set, ACR percent improvement (ACR-N), Disease Activity Score in 28 joints c-reactive protein (DAS28-CRP), the Health Assessment Questionnaire–Disability Index (HAQ-DI), and the percentage of good or moderate responders on the European League Against Rheumatism Responder Index in 28 joints (EULAR-28).
Blood samples were collected to evaluate changes in serum immunoglobulin (IgG, IgM, IgA) levels, to assess the development of antitabalumab antibodies (ADAs) and the effect of tabalumab on B cell counts, and to assess rheumatoid factor (RF) and anticyclic citrullinated peptide (anti-CCP) levels. Immunoglobulin concentrations were assayed using standard methods. Antitabalumab antibodies were assayed using validated methods. B cells were assayed by flow cytometry, from which the time course of changes in absolute B-cell counts (CD3-CD20) was assessed. Rheumatoid factor was assessed using standard methods; anti-CCP was assessed using standard enzyme-linked immunosorbent assay methods. Safety was assessed by examining laboratory tests, vital signs, electrocardiograms, adverse events (AEs), and serious AEs (SAEs).
Statistical Analysis
Primary and secondary efficacy analyses were conducted using a prespecified subset of patients who had 5 or more of 68 tender and 5 or more of 66 swollen joints at baseline (efficacy population); these analyses were repeated using the intention-to-treat (ITT) population (all randomized patients). Safety analyses included all randomized patients who received 1 or more dose of study drug (safety population).
A sample size of 1002 randomized patients was anticipated to enroll 501 patients or more who would meet efficacy population criteria. This efficacy population was expected to provide 95% power or greater to detect a difference of 20% in ACR20 response rates at week 24 between each tabalumab group and the placebo group, assuming a placebo response rate of 23%.
For baseline demographics and characteristics, Fisher exact test was used for categorical data, and analysis of variance was used for continuous data. Analyses of categorical efficacy measures (ie, ACR20, ACR50, ACR70, EULAR-28) were conducted using a logistic regression model with treatment, geographic region, baseline DAS28-CRP, TNF–inadequate responders (TNF-IR) treatment history, and background DMARD at baseline as factors. When the logistic regression sample size requirements were not met, Fisher exact test was used. Fisher exact test was also used for categorical safety variables. For continuous efficacy and safety variables (ie, individual components of the ACR, DAS28-CRP, HAQ-DI, CRP, laboratory analytes, vital signs, serum immunoglobulins, and B cells), an analysis-of-covariance model that included treatment, geographic region, TNF-IR treatment history, background DMARD at baseline, and baseline value was used to test the difference of each tabalumab group versus the placebo group.
For ACR20, ACR50, and ACR70 responses, patients who were week 16 nonresponders or who discontinued prior to week 24 were imputed as nonresponders (NRI). For all other measures, the last postbaseline value at or before week 16 for nonresponders and before discontinuation for discontinued patients was used (modified last observation carried forward [mLOCF]) in evaluation. All statistical tests of treatment effects were performed at 2-sided significance levels of 0.05.