Article Text


Extended report
A phase II, double-blind, randomised, placebo-controlled study of BMS945429 (ALD518) in patients with rheumatoid arthritis with an inadequate response to methotrexate
  1. P Mease1,
  2. V Strand2,
  3. L Shalamberidze3,
  4. A Dimic4,
  5. T Raskina5,
  6. Li-An Xu6,
  7. Y Liu7,
  8. J Smith8
  1. 1Division of Rheumatology Research, Swedish Medical Center and University of Washington, Seattle, Washington, USA
  2. 2Division of Immunology and Rheumatology, Stanford University, Palo Alto, California, USA
  3. 3Department of Clinical Pharmacology, V Tsitlanadze Scientifically-Practical Rheumatology Center Tbilisi, Georgia
  4. 4Rheumatology Clinic, Institute of Rehabilitation and Treatment Niska Banja, Serbia
  5. 5Internal Disease Dept, State Institute of Healthcare, Kemerovo, Russian Federation
  6. 6Global Biometric Sciences, Bristol-Myers Squibb, Princeton, New Jersey, USA
  7. 7Global Clinical Research, Bristol-Myers Squibb, Princeton, NJ
  8. 8Alder Biopharmaceuticals Inc, Bothell, Washington, USA
  1. Correspondence to Philip J Mease, Department of Rheumatology, Seattle Rheumatology Associates, 1101 Madison Street Ste 1000, Seattle 98104, Washington, USA; pmease{at}


Background Interleukin 6 (IL-6) plays a key role in the inflammatory cascade in rheumatoid arthritis. BMS945429 is a humanised, monoclonal antibody that potently binds IL-6.

Objective To conduct aphase II study to determine the efficacy and safety of BMS945429 in patients with active rheumatoid arthritis and an inadequate response to methotrexate.

Methods Patients were randomised 1:1:1:1 to BMS945429 (80, 160 or 320 mg; administered intravenously) or placebo plus methotrexate during this 16-week, double-blind trial. The primary efficacy end point was the proportion of patients with a 20% improvement in American College of Rheumatology responses (ACR20) at week 12. Additional end points included ACR50 and ACR70 responses and 28-joint Disease Activity Scores (DAS28).

Results Of 127 randomised and treated patients, 116 completed the trial. ACR20 responders at week 12 were 81% (80 mg; p<0.0001 vs placebo), 71% (160 mg; p=0.0005 vs placebo), 82% (320 mg; p<0.0001 vs placebo) and 27% (placebo), respectively. By week 16, 14% (80 mg), 28% (160 mg) and 44% (320 mg) of BMS945429 patients were in DAS28 remission (DAS28 score <2.6). Statistically significant and clinically meaningful improvements in health-related quality of life (HRQoL) were reported in all active treatment groups. Administration of BMS945429 was associated with increases in liver enzymes and in serum cholesterol. There were no serious infections, infusion reactions or apparent immunogenicity.

Conclusions In this phase II study, BMS945429 was associated with rapid and significant improvements in disease activity and HRQoL in patients with active rheumatoid arthritis and an inadequate response to methotrexate.

Statistics from


Rheumatoid arthritis is a chronic autoimmune disease characterised by inflammation of the synovium and joints, which ultimately leads to the destruction of bone and cartilage.1 The disease affects up to 1% of the global population, and it occurs most often in women.2,,4 Rheumatoid arthritis causes significant disability, adversely impacting health-related quality of life (HRQoL), and is associated with early mortality.5,,7

The pathophysiology of rheumatoid arthritis involves dysregulation of interleukin 6 (IL-6),8 a pleiotropic proinflammatory cytokine that mediates the acute phase response and is associated with other autoimmune conditions.9 10 At the cellular level, IL-6 stimulates B cell differentiation and secretion of antibodies, prevents apoptosis of activated B cells, activates and induces proliferation of T cells and in the presence of IL-2 induces differentiation of mature and immature T cells into cytotoxic T cells.11,,15 IL-6 induces the generation of Th17 cells from naïve T cells, and can inhibit Treg differentiation.16 17 Imbalances in Th17 and Treg activity have been implicated in proinflammatory pathophysiology, autoimmune disease, and in the transition from acute to chronic autoimmune inflammation.18 19 Raised concentrations of IL-6 are found in the synovial fluid of patients with rheumatoid arthritis; the degree of elevation is correlated with disease activity and joint damage.20 21 Initial attempts to develop antibodies to IL-6 as therapeutic agents were not successful,22 23 although tocilizumab, a humanised antibody to the IL-6 receptor, has been approved for the treatment of moderate to severe rheumatoid arthritis in patients who have an inadequate response to tumour necrosis factor antagonists.24

BMS945429 (also known as ALD518) is a humanised monoclonal antibody, produced in Pichia pastoris yeast, directed against IL-6 rather than the IL-6 receptor. The objective of this study was to determine the efficacy and safety of IL-6 blockade by intravenously administered BMS945429 versus placebo in patients with active rheumatoid arthritis and an inadequate response to methotrexate.


This was a phase II, double-blind, randomised, placebo-controlled study of BMS945429 conducted at 24 centres in five countries (see ‘Acknowledgements’) in accordance with ethical principles outlined in the Declaration of Helsinki, and consistent with International Conference on Harmonisation Good Clinical Practice and applicable regulatory requirements. Enrolment began in October 2008 and was completed in June 2009. The protocol and all amendments were reviewed and approved by the institutional review board/independent ethics committee for each site before and during the study, and the study was registered with (id: NCT00867516).


Adults (18–80 years) with rheumatoid arthritis of ≥16 weeks' duration before screening, American College of Rheumatology (ACR) functional status class of 1–3, ≥6 tender and ≥6 swollen joints on a 68/66 joint count at screening, and a serum C-reactive protein (CRP) concentration of ≥10 mg/l were enrolled.

Patients were excluded from the study if their arthritis onset occurred before 16 years of age or if they had a continuing systemic inflammatory condition (eg, systemic lupus erythematosus), nor could they have received any biological treatment or investigational drug in the previous 12 months. A previous history of active tuberculosis or any current or recent serious infection disqualified patients. Patients could not have present or previous malignancies, except a history of cured squamous or basal skin cell carcinoma or cured breast or cervical cancer for ≥10 years without evidence of recurrence. Pregnant or nursing women were excluded, and women of childbearing potential were required to use adequate contraception and have a negative pregnancy test at screening and at baseline. At screening, patients were excluded if they had alanine transaminase (ALT) ≥3×upper limits of normal (ULN); aspartate transaminase (AST) ≥3×ULN; haemoglobin concentration <9 g/dl; absolute neutrophil count <1.5×109/l; platelet count <100×109/l.


After screening, patients were randomised 1:1:1:1 (BMS945429 80 mg, 160 mg, 320 mg or placebo) in blocks of 12.


Patients received intravenous treatment with BMS945429 or placebo at day 1 and at week 8. All patients received methotrexate (≥10 mg/week) throughout the study, at doses that were stable for ≥3 months before the study. Patients were given the drugs intravenously over a period of 1 h on day 1 and at the week 8 visit. Allowed concomitant drugs included folic acid; non-steroidal anti-inflammatory drugs; acetylsalicylic acid (≤100 mg/day); daily doses of glucocorticoids (≤10 mg/day of prednisolone or equivalent, stable for 4 weeks before day 1) and analgesics, except for 24 h before rheumatology evaluations. Narcotics were not permitted.


The primary efficacy end point was the proportion of patients achieving a 20% improvement in ACR response (ACR20) at week 12 in each active treatment group versus placebo.25 Secondary outcomes included ACR 50% and 70% responders and mean changes from baseline to week 12 in each component of the ACR response criteria; 28-joint Disease Activity Scores (DAS28) and EULAR responses26 27; and HRQoL by the Medical Outcomes Survey Short Form-36 (SF-36).28

Adverse events were summarised descriptively, and laboratory results were classified using the common terminology criteria for adverse events.29

Sample size

The target sample size for this study was based on a parallel design for testing individual treatment arms versus placebo (no multiplicity adjustment) using a power of 80% and an individual test type I error of 5%. With an assumed placebo ACR20 response rate of 15% and at least a 50% response with active treatment, the required sample size was 30 patients for each treatment arm or a total of 120 patients.

Statistical methods

All statistical analyses in this report were based on a modified intention-to-treat population, defined as all randomised patients who received one or more dose of the study drug. For analyses of ACR responses, patients with missing values or who discontinued prematurely were classified as non-responders. ACR responses at week 12 were analysed using a Fisher's exact test by comparing each active treatment group with placebo. No multiplicity adjustments were performed for reported p values.

Analyses of mean changes from baseline of secondary efficacy end points were summarised with descriptive statistics by treatment group and visit, based on observed data without imputation. Physical and Mental Component Summary scores of SF-36 were initially assessed and if results of either were statistically significant, mean changes in domains were assessed for statistical significance without p value corrections, as customary, and for improvements meeting or exceeding the minimum clinically important difference of 5–10 points for domain scores and 2.5–5.0 points for Physical and Mental Component Summary scores.


Of 132 randomised patients, 127 received at least one dose of the study drug (modified intention-to-treat population, figure 1) and 116 (91.3%) completed the study. Baseline demographic and disease characteristics are presented in table 1, and were generally balanced between treatment groups.

Table 1

Baseline demographics and disease characteristics

ACR20/50/70 responses at week 12

Responders by ACR20 criteria at week 12 were 81% (BMS945429 80 mg; p<0.0001 vs placebo), 71% (160 mg; p<0.001), 82% (320 mg; p<0.0001), and 27% (placebo). ACR50 responders were 34% (80 mg; p=0.017 vs placebo), 27% (160 mg; p=NS), 50% (320 mg; p<0.001) and 9% (placebo). ACR70 responders were 13% (BMS945429 80 mg; p=NS vs placebo), 12% (160 mg; p=NS), 25% (320 mg; p=0.019) and 3% (placebo). Responders by ACR 20/50/70 over time are depicted in figure 2A–C, with differences in efficacy between active treatment and placebo evident within the first 2 weeks.

Figure 2

Proportion of patients achieving an American College of Rheumatology (ACR)20 (A), ACR50 (B) or ACR70 (C) over time. (D) Mean C reactive protein (CRP) concentration over time. (E) Mean 28-joint Disease Activity Score (DAS28) score over time. MTX, methotrexate.

ACR20/50/70 responses at week 16

At week 16, ACR20 responders were 75% (BMS945429 80 mg; p=0.003 vs placebo), 65% (160 mg; p=0.03), 82% (320 mg; p<0.001), and 36% (placebo). ACR50 responders were 41% (BMS945429 80 mg; p=0.03 vs placebo), 41% (160 mg; p=0.03), 50% (320 mg; p=0.005) and 15% (placebo). ACR70 responders were 22% (BMS945429 80 mg; p=NS vs placebo), 18% (160 mg; p=NS), 43% (320 mg; p=0.002) and 6% (placebo).

Mean changes from baseline in ACR core set components, DAS28 and SF-36

Mean changes from baseline in each component of the ACR response criteria (tender and swollen joint counts, physician global assessment by visual analogue scale (VAS), patient global assessment (VAS), pain (VAS), Health Assessment Questionnaire Disability Index and CRP were numerically greater in each BMS945429 dose group than in the placebo group (table 2). CRP concentrations declined rapidly in the groups receiving an active dose and remained suppressed though week 16 (figure 2D).

Table 2

Mean change from baseline in American College of Rheumatology core set variables and SF-36 Physical/Mental Component Scale scores at week 16

Mean DAS28 scores decreased over time in all treatment groups; changes in the BMS945429 groups were numerically greater by week 12 and at week 16 compared with placebo (figure 2E, table 2). The percentage of patients achieving a good or moderate EULAR response at week 16 was 97%, 94% and 96% in the BMS945429 80 mg, 160 mg and 320 mg groups, respectively, compared with 52% of patients in the placebo group. The proportion of patients in DAS28 remission (DAS28<2.6) was 14%, 28% and 44% at week 16 in the 80 mg, 160 mg and 320 mg groups, respectively, compared with none in the placebo group.

Patients in each active treatment group reported statistically significant and clinically meaningful improvements in HRQoL by Mental Component Summary scores as well as individual domain scores of SF-36 at week 12. Mean changes in Mental Component Summary scores increased by 7.6–13.5, all well exceeding the minimum clinically important difference of 2.5, compared with 1.8 in placebo. Mean changes in Physical Component Summary scores were numerically greater in active treatment groups, and exceeded the minimum clinically important difference: from 6.1 to 8.5 compared with 4.2 in placebo. These are reflected in improvements across all eight domains of SF-36 that exceeded the minimum clinically important difference (5.0 points) in all active groups and were statistically significant (p<0.05) in six domains with 80 mg and 320 mg treatment and one domain with 160 mg (figure 3 and table 3). Reported improvements were maintained through 16 weeks (data not shown).30,,34

Figure 3

Spydergram of Short Form 36 (SF-36) domain scores at baseline and week 12 compared with age/gender matched norms. Legend: SF-36 domains: BP, bodily pain; MH, mental health; PF, physical function; RP, role physical; RE, role emotional; SF, social functioning; VT, vitality. Inner polygon (gray): composite mean baseline scores across all treatment groups; outer polygon (lavender) age/gender-matched normative data based on protocol population; inner polygons: mean changes at week 12 in (orange) placebo; (light blue) 80 mg; (green) 160 mg; (dark blue): 320 mg. Note that values in VT and MH domains at week 12 approach or meet norms.

Table 3

Domain scores of SF-36 at baseline and at week 12


No deaths were reported; two serious adverse events (elevated liver function tests; n=1 each in the 80 mg and 320 mg groups) leading to discontinuation of treatment were reported through week 16. Two additional patients (n=1 each in the 80 mg and 320 mg groups) discontinued treatment owing to raised alanine aminotransferase (ALT)/aspartate aminotransferase (AST) levels. There were no serious infections or infusion reactions. The most common adverse events are listed in table 4.

Table 4

Most common adverse events (occurring in three or more patients overall)

A review of laboratory results noted changes in several parameters, including moderate elevations of liver enzymes in the BMS945429 groups. Fifty-one patients in the BMS945429 groups (n=19 80 mg; n=17 160 mg; n=15 320 mg; n=0 placebo) had at least one ALT≤3× and >1×ULN and 14 (n=4 80 mg; n=2 160 mg; n=8 320 mg; n=0 placebo) had an ALT>3×ULN. Of the 14 patients with ALT levels >3×ULN, seven had other hepatobiliary abnormalities at baseline (n=6 cholelythiasis/cholecystitis; n=1 hepatic steatosis). Raised bilirubin levels were reported in 23 patients (n=8 80 mg; n=6 160 mg; n=6 320 mg; n=3 placebo; all ≤3×ULN), but none met criteria for Hy's Law, that concomitant 3×ULN elevations of ALT and/or AST and 2×ULN of total bilirubin are predictive of liver injury.35

Transient decreases in neutrophil counts were noted in 22 patients (n=11 80 mg; n=9 160 mg; n=2 320 mg; n=0 placebo). Neutropenia was mild (<1500 cells/mm3 and >1000 cells/mm3) in 20/22 patients, and was moderate (500–1000 cells/mm3) in two patients, both in the BMS945429 80 mg group. None were associated with infections. In seven patients treated with BMS945429 decreases in platelets were noted (all counts between 75 000/mm3 and lower limits of normal); none of those episodes were associated with increased bleeding. Mean serum haemoglobin concentrations increased over time in the BMS945429 dose groups.

Baseline total cholesterol (5.2 mmol/l) and triglyceride (1.2 mmol/l) concentrations for the total patient population approached the ULN at baseline. Mean concentrations of total cholesterol and triglycerides increased in the BMS945429 groups compared with placebo. At week 12, mean total cholesterol was 6.6±1.5 mmol/l (80 mg), 6.4±1.4 mmol/l (160 mg), 6.2±1.4 mmol/l (320 mg) and 5.3±1.1 mmol/l (placebo). Serum total cholesterol was ≤7.75 mmol/l in most (72%) of the BMS945429 patients with raised total cholesterol. At week 12, mean serum triglycerides were 1.6±0.9 mmol/l (80 mg), 1.7±1.0 mmol/l (160 mg), 1.7±0.7 mmol/l (320 mg) and 1.3±0.9 mmol/l (placebo). Total cholesterol was not split into component parts.


This is the first study to show that treatment with a humanised antibody to IL-6, BMS945429, was associated with rapid and significant improvement in disease activity compared with placebo in patients with active rheumatoid arthritis and an inadequate response to methotrexate. The highest dose appeared to be associated with better efficacy and significant improvements in HRQoL, although there was a higher incidence of raised liver enzyme. No serious infusion reactions were reported.

The potency and selectivity of BMS945429 for IL-6 and its long half-life probably contribute to the initial efficacy profile seen in this study. Treatment with one dose of BMS945429 was associated with rapid onset of symptom relief and a durable suppression of CRP. Although existing biological agents are associated with improvement in many patients, clinical remission remains difficult to achieve. Forty-three per cent of patients in the BMS945429 320 mg dose group achieved an ACR70 response, and 44% achieved DAS remission by week 16. These responses may be driven in part by the effects of BMS945429 on acute phase reactants. However, the results compare favourably with rheumatoid arthritis trials of biological agents in patients with an inadequate response to methotrexate or other disease-modifying antirheumatic drugs.36,,43 Because the treatment goal for rheumatoid arthritis is remission,44 these results, if confirmed in larger phase III trials, indicate that a significant proportion of patients for whom methotrexate fails could achieve this target with BMS945429.

Improvements in all patient-reported outcomes were statistically significant and exceeded the minimal clinically important difference in the BMS945429 treatment groups. Of interest, SF-36 Mental Component Summary scores improved to a greater extent than the Physical Component Summary scores, in contrast to the pattern typically seen in rheumatoid arthritis trials. In part, this may be due to clinically meaningful improvements in physical function, role physical and bodily pain domains in the placebo group. Nonetheless, improvements in these same physical domains were greater in the BMS945429 treatment groups.

It is important to characterise the safety profile of any new agent with a new mechanism of action. There were cases of mild to moderate neutropenia in the BMS945429 groups, particularly at the lower doses. However, the neutrophil counts did not decrease further after the second dose of BMS945429, and there were no serious infections associated with the reduced neutrophil counts. Moderate elevations in liver enzymes and serum cholesterol were seen, similar to effects seen with blockade of IL-6R signalling by tocilizumab.45 Most individual cases of raised liver enzymes were <3×ULN. Methotrexate may contribute to observed increases in liver enzymes. In a phase II study of BMS945429 as monotherapy in the treatment of non-small cell lung cancer, similar doses to this study were not associated with increases in ALT concentrations.46 Regarding cholesterol changes, it should be noted that inflammation and IL-6 are associated with reductions in circulating lipid concentrations,47,,49 possibly through effects on very low density lipoprotein receptor expression.50 Thus, IL-6 inhibition would be expected to potentially reverse some of these changes.

The study has several limitations. The sample size of the treatment groups is modest, although appropriate for a phase II study. The results cannot be generalised to other populations (eg, inadequate responders to biological agents), who may be more refractory to treatment. No radiographic data were obtained. Conclusions about the overall clinical profile should be tempered by the duration of the study; longer-term data are needed to confirm these data.

In conclusion, BMS945429 is the first direct inhibitor of IL-6 to show efficacy in rheumatoid arthritis, providing a rapid and significant improvement in disease activity in patients with an inadequate response to methotrexate. Its long half-life may allow for sustained response with infrequent dosing. These results warrant additional study of BMS945429 to characterise the efficacy and safety in a wider population of patients with rheumatoid arthritis and for longer periods of time.


The following are additional members of the BMS945429 Study Group. Canada: Rafat Faraawi, Robert McKendry; Georgia: Lali Kilasonia; Poland: Marek Brzosko, Jaroslaw Marcinkiewicz, Stanislaw Sierakowski, Wielslawa Porawska, Cezary Stepniak; Russia: Larisa Eliseeva, Olga Ershova, Vadim Mazurov, Olga Reshetko, Vladimir Simanenkov, Valentina Sorotskaya, Yuriy Uspenskiy, Natalia Vezikova, Sergey Yakushin, Elena Zonova; Serbia: Radmila Petrovic, Roksanda Stojanovic. Dr Dimic would like to acknowledge Aleksandra Stankovic and Jovan Nedovic for their contributions to the conduct of the study. Dr Brian Atkinson of Bristol-Myers Squibb provided professional writing and editorial assistance..


View Abstract


  • Funding Alder Biopharmaceuticals and Bristol-Myers Squibb.

  • Competing interests PM has received research grants from, or served as a consultant to, Abbott, Amgen, BiogenIdec, Bristol-Myers Squibb, Centocor, Genentech, Lilly, Novartis, Pfizer and UCB. PM has served on speaker's bureaus for all of the above except Novartis.

  • Provenance and peer review Not commissioned; externally peer reviewed

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.