Significant improvement in synovitis, osteitis, and bone erosion following golimumab and methotrexate combination therapy as compared with methotrexate alone: A magnetic resonance imaging study of 318 methotrexate-naive rheumatoid arthritis patients

Authors

  • Mikkel Østergaard,

    Corresponding author
    1. Copenhagen University Hospital at Glostrup, Glostrup, Denmark
    • Department of Rheumatology, Copenhagen University Hospital at Glostrup, Nordre Ringvej 57, DK-2600 Glostrup, Denmark
    Search for more papers by this author
    • Dr. Østergaard has received consulting fees, speaking fees, and/or honoraria from Abbott, Bristol-Myers Squibb, Centocor, Roche, GlaxoSmithKline, Pfizer, MSD, UCB, and Wyeth (less than $10,000 each).

  • Paul Emery,

    1. University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
    Search for more papers by this author
    • Dr. Emery has received consulting fees, speaking fees, and/or honoraria from MSD, Pfizer, Roche, Bristol-Myers Squibb, and Abbott (less than $10,000 each).

  • Philip G. Conaghan,

    1. University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
    Search for more papers by this author
    • Dr. Conaghan has received consulting fees, speaking fees, and/or honoraria from Bristol-Myers Squibb, Centocor, Merck, Pfizer, Roche, and Novartis (less than $10,000 each).

  • Roy Fleischmann,

    1. University of Texas Southwest Medical Center, Dallas
    Search for more papers by this author
    • Dr. Fleischmann has received consulting fees, speaking fees, and/or honoraria from Centocor (less than $10,000).

  • Elizabeth C. Hsia,

    1. Centocor Research and Development, Malvern, Pennsylvania, and University of Pennsylvania School of Medicine, Philadelphia
    Search for more papers by this author
    • Drs. Hsia and Rahman own stock or stock options in Johnson & Johnson.

  • Weichun Xu,

    1. Centocor Research and Development, Malvern, Pennsylvania
    Search for more papers by this author
  • Mahboob U. Rahman

    1. Centocor Research and Development, Malvern, Pennsylvania, and University of Pennsylvania School of Medicine, Philadelphia
    Search for more papers by this author
    • Drs. Hsia and Rahman own stock or stock options in Johnson & Johnson.


Abstract

Objective

To evaluate the effects of golimumab on inflammation/structural damage detected by magnetic resonance imaging (MRI) in patients with rheumatoid arthritis (RA).

Methods

Methotrexate (MTX)–naive RA patients (n = 637) were randomized to placebo plus MTX, golimumab 100 mg plus placebo, golimumab 50 mg plus MTX, or golimumab 100 mg plus MTX (subcutaneous golimumab every 4 weeks). Of these, 318 patients participated in an MRI substudy. MRIs (contrast-enhanced; 1.5T) of the wrist and second through fifth metacarpophalangeal joints of the dominant hand were obtained at baseline and weeks 12 and 24. MRIs were scored by 2 independent readers (blinded to image sequence/chronology, patient identity, and treatment group) for synovitis, bone edema/osteitis, and bone erosions using the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) system. Radiographs (hands, wrists, forefeet at baseline and week 28) were scored by 2 other readers (blinded as above) using the modified Sharp/van der Heijde (SvdH) scoring system. Changes from baseline were compared between treatment groups (two-sided analysis of variance on van der Waerden normal scores).

Results

At weeks 12 and 24, combined therapy with golimumab plus MTX versus placebo plus MTX significantly improved RAMRIS scores for synovitis (mean −1.92 versus 0.14 [P < 0.001] at week 12; −2.45 versus −1.04 [P < 0.001] at week 24), osteitis (mean −1.82 versus 0.56 [P < 0.001] at week 12; −2.27 versus −0.32 [P < 0.001] at week 24), and bone erosion (mean −0.40 versus 0.24 [P = 0.016] at week 12; −0.40 versus −0.24 [P = 0.010] at week 24). Results of sensitivity analyses (no missing doses/data and using linear extrapolation) were generally consistent with results of the primary analyses. Changes in SvdH scores among the MRI substudy patients at week 28 showed no significant difference between golimumab plus MTX therapy and placebo plus MTX (mean 0.49 versus 0.92; P = 0.19). Radiographic SvdH scores demonstrated inhibition of structural damage progression by treatment with golimumab plus MTX as compared with placebo plus MTX in the overall study population but required double the number of patients (637 versus 318) and double the length of followup (28 versus 12 weeks) as needed for MRI to demonstrate this.

Conclusion

Improvements in inflammation (synovitis and osteitis) and erosions with golimumab plus MTX therapy exceeded those with placebo plus MTX therapy from week 12 onward, confirming the overall clinical/radiologic findings. MRI was more sensitive than conventional radiography in detecting the progression of erosions.

Rheumatoid arthritis (RA) is a progressive immune-mediated inflammatory disease that can lead to significant physical disability if not treated aggressively during its early stages. While conventional radiographs have been regarded as the gold standard for detecting bone damage and monitoring disease progression, radiographs are only useful in documenting damage to the osseous and cartilaginous components of the joint, a stage of disease that is generally irreversible (1). Conversely, magnetic resonance imaging (MRI) has been shown to be able to detect pre-erosive lesions (synovitis and osteitis) and to be much more sensitive in detecting joint erosions (2–6).

MRI-detected synovitis and bone marrow edema have been shown to foreshadow the development of new erosions over time as detected by either MRI or radiography (7–14). The areas of bone that appear as bone edema or osteitis on MRI have been shown to be heavily infiltrated by inflammatory cells, including osteoclasts (15). Detection and treatment of pre-erosive lesions (synovitis and osteitis) in RA could significantly affect the ensuing disease course.

It has been noted that the severity of overall disease and joint damage in the population of RA patients recruited into clinical studies may be decreasing over time (16). Since radiographic progression rates can be low in patient populations with lower disease activity and less structural damage at baseline, radiography may not be sensitive enough to detect an effect of an experimental compound when there is minimal radiographic progression in the control group (17).

Golimumab is a human monoclonal antibody to tumor necrosis factor α (TNFα). The efficacy and safety of golimumab in patients with RA were established in 3 large, phase 3 clinical studies that included subpopulations of patients with different exposures to prior therapies. One of these studies, the Golimumab Before Employing Methotrexate as the First-Line Option in the Treatment of Rheumatoid Arthritis of Early Onset (GO-BEFORE) study, evaluated patients with active RA who were naive to methotrexate (MTX). The primary efficacy end points related to signs and symptoms, safety, and radiographic data have been reported (17, 18). An MRI substudy was conducted as part of the GO-BEFORE trial to evaluate the effect of golimumab on inflammation and structural damage as detected by MRI in MTX-naive patients with active RA, and the results of the substudy evaluations are presented herein.

PATIENTS AND METHODS

The study design and patient inclusion criteria for the GO-BEFORE study have been published elsewhere (18). The GO-BEFORE study was conducted according to the principles of the Declaration of Helsinki, such that all patients provided written informed consent before participating in the study. The overall GO-BEFORE study population consisted of patients with RA who were MTX-naive (n = 637). A subset of these patients from eligible and willing study sites participated in an MRI substudy (n = 318).

Patients were randomly assigned to receive placebo injections plus MTX capsules (group 1), golimumab 100-mg injections plus placebo capsules (group 2), golimumab 50-mg injections plus MTX capsules (group 3), or golimumab 100-mg injections plus MTX capsules (group 4). Golimumab and placebo injections were administered subcutaneously every 4 weeks. At week 28, patients with <20% improvement in both the tender and swollen joint counts entered a double-blind early escape phase, in which patients in group 1 received golimumab 50 mg plus MTX, patients in group 2 received active MTX plus golimumab 100 mg, and patients in group 3 received golimumab 100 mg plus MTX. Patients in group 4 who met the criteria for early escape had no adjustment of their study medications. Group 1 was an active comparator group, and patients received MTX alone unless they met the early escape criteria. The duration of the active-comparator portion of the GO-BEFORE study was 52 weeks.

Eligible (i.e., MRI at 1.5T could be performed as specified in the study protocol) and willing study sites participated in the MRI substudy, and all patients at these sites participated in the MRI substudy. MRIs of the patient's dominant wrist and metacarpophalangeal (MCP) joints were obtained at baseline and weeks 12, 24, 52, and 104 using a 1.5T MRI with contrast enhancement. The MR sequences were as follows: axial T1 fast spin-echo precontrast, coronal T1 fast spin-echo precontrast, coronal short tau inversion recovery (STIR) or T2 fat-suppressed precontrast, and coronal T1 fat-suppressed postcontrast. Further details of the MRI protocols are summarized in Table 1. Results through week 24 are presented herein.

Table 1. Magnetic resonance imaging protocols for large and small wrist coils*
SequenceTR, msecTE, msecTI, msecSlice thickness, mmGap, mmFOV, cm
  • *

    Depending on availability at the individual centers, a large or a small wrist coil was used. If a large wrist coil was used, it was positioned to include the metacarpophalangeal (MCP) joints and wrist in the field of view (FOV). If a small wrist coil was used, the MCP joints were imaged precontrast, and then the coil was positioned for the wrist joint and the pre- and postcontrast series were imaged. TR = repetition time; TE = echo time; TI = inversion time; FSE = fast spin-echo; STIR = short tau inversion recovery; NA = not applicable.

  • A coronal STIR sequence or a T2-weighted, fat-suppressed sequence (chosen by the center) was obtained with the indicated slice thickness/FOV. The parameters for the STIR sequence are shown.

  • Gadolinium contrast agent (0.1 mmole [0.2 cc]/kg; 20 cc maximum) was administered intravenously via rapid hand injection through prepositioned access, followed by a 20-ml saline flush. Image was to be acquired as soon as possible after completing the saline flush.

Large wrist coil      
 MCP and wrist joints      
  T1 FSE      
   Axial, precontrast42013NA3116
   Coronal, precontrast42013NA2 (inter-leaved)016
  STIR      
   Coronal, precontrast3,760451503116
  T1 fat-suppressed      
   Coronal, postcontrast42013NA4116
Small wrist coil      
 MCP joints      
  T1 FSE      
   Axial, precontrast42013NA3110–12
   Coronal, precontrast42013NA2 (inter-leaved)010–12
  STIR      
   Coronal, precontrast3,760451503110–12
 Wrist joints      
  T1 FSE      
   Axial, precontrast42013NA3110–12
   Coronal, precontrast42013NA2 (inter-leaved)010–12
  STIR      
   Coronal, precontrast3,760451503110–12
  T1 fat-suppressed      
   Coronal, postcontrast42013NA4110–12

Images were scored by 2 independent radiologists who were blinded to image time point and sequence, patient identity, and treatment group. Using the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) system (19), the readers scored synovitis from 0 (normal) to 3 (severe) for each of 3 wrist regions and MCP joints 2, 3, 4, and 5 (yielding a range of synovitis scores of 0–9 for the wrist and 0–12 for the MCP joints); bone edema (osteitis) from 0 (0% bone involvement) to 3 (67–100% bone involvement) for each of 23 wrist and MCP joints (yielding a range of osteitis scores of 0–69); and bone erosions from 0 (0% bone involvement) to 10 (91–100% bone involvement) for each of 23 wrist and MCP joint bones (yielding a range of bone erosion scores of 0–230). Higher RAMRIS scores indicate more severe inflammation/damage.

Randomized patients at MRI substudy sites who had at least 1 MRI image set scored by at least 1 reader were included in the MRI analyses. For each patient at each visit, a RAMRIS score was calculated as the average of each RAMRIS score provided by the 2 independent readers. If a RAMRIS score from one of the readers was missing, the RAMRIS score from the other reader was used. If the score was missing from both readers at any given time point, the score was set to “missing,” and the missing data imputation rules were applied as described below.

The following missing data rules were applied for all patients. First, if the score was missing from baseline through week 24, the change score at week 24 was imputed using the median change from baseline score for all patients in the same screening C-reactive protein (CRP) stratum (<1.5 mg/dl or ≥1.5 mg/dl) at week 24. Second, if the week 24 score was missing, the score was imputed using the last nonmissing observation (including baseline score). Third, if the baseline MRI score was missing, the change score at week 24 was imputed using the median change from baseline score for all patients in the same screening CRP stratum (<1.5 mg/dl or ≥1.5 mg/dl) at week 24.

Sensitivity analyses were conducted, including evaluation of change from baseline to week 24 with no imputation rules (i.e., including only the 238 substudy patients who received subcutaneous study agent through week 24 with no missed doses and no missing baseline or week 24 data for all 3 RAMRIS scores [75% of the 318 substudy patients randomized]) and evaluation of the RAMRIS bone erosion score with linear extrapolation, which was implemented for less than 25% of the substudy patients.

Changes from baseline to week 12 and to week 24 in the RAMRIS scores for synovitis, bone edema (osteitis), and bone erosion scores were compared between treatment groups with the use of a two-sided analysis of variance on van der Waerden normal scores (20) at a 0.05 level of significance. The primary comparison was between the combined golimumab plus MTX group (golimumab 50 mg plus MTX and golimumab 100 mg plus MTX) and the placebo plus MTX group. If this comparison was statistically different, then the individual golimumab dosage groups were also compared with the placebo plus MTX group. The study, however, was not powered for these individual golimumab dosage group comparisons with placebo plus MTX group (details of sample size estimates have been previously reported [18]). Thus, this report will focus on comparisons between the combined golimumab plus MTX group and the placebo plus MTX group. Radiographic data, which are reported as changes in the total modified Sharp/van der Heijde (SvdH) score (21) from baseline (i.e., within 4 weeks of the first dose) to week 28 as determined by 2 independent readers and an adjudicator, were also compared between the combined golimumab treatment groups and the placebo group among the 318 MRI substudy patients.

To assess interreader and intrareader (read–reread) reliability, images from 10% of the patients were randomly selected and reread by each of the 2 readers. Interreader and intrareader reliability coefficients were estimated (22) using the week 24 RAMRIS synovitis (wrist joints), bone erosion, and bone edema reread scores.

RESULTS

A total of 318 randomized patients participated in the GO-BEFORE MRI substudy, including 82 patients in the placebo plus MTX group, 77 in the golimumab 100 mg plus placebo group, 78 in the golimumab 50 mg plus MTX group, and 81 in the golimumab 100 mg plus MTX group (Figure 1). The proportions of MRI substudy patients who met the early escape criteria (12–20%) (Figure 1) were consistent with those for the overall GO-BEFORE population (12–18%) (18).

Figure 1.

Disposition of the study patients in the magnetic resonance imaging (MRI) substudy of the Golimumab Before Employing Methotrexate as the First-Line Option in the Treatment of Rheumatoid Arthritis of Early Onset (GO-BEFORE) trial, by treatment group. MTX = methotrexate.

Baseline demographic and disease characteristics of the MRI substudy patients were generally well balanced across the randomized treatment groups, with the exception of lower RAMRIS scores for bone edema and bone erosion and lower total SvdH scores in the placebo plus MTX group (Table 2). Baseline patient characteristics were also generally consistent with baseline characteristics of the overall GO-BEFORE patient population (18). Mean baseline RAMRIS scores for bone erosion ranged from 17.7 to 25.0 across the randomized treatment groups, and the mean duration of disease ranged from 2.2 years to 4.0 years (Table 2).

Table 2. Baseline clinical characteristics of patients in the GO-BEFORE MRI substudy*
CharacteristicGroup 1, placebo plus MTXGroup 2, golimumab 100 mg plus placeboGolimumab plus MTX
Group 3, 50 mg golimumabGroup 4, 100 mg golimumabGroups 3 and 4 combined
  • *

    GO-BEFORE = Golimumab Before Employing Methotrexate as the First-Line Option in the Treatment of Rheumatoid Arthritis of Early Onset; MRI = magnetic resonance imaging; MTX = methotrexate; IQR = interquartile range; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; SvdH = modified Sharp/van der Heijde score.

  • Due to an inability to obtain postcontrast images of both the wrist and the metacarpophalangeal (MCP) joints at the study site, the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) synovitis scores for the wrist plus MCP joints represent 272 patients, and those for the wrist joints only represent 288 patients.

No. of MRI substudy patients randomized82777881159
No. (%) women69 (84.1)63 (81.8)63 (80.8)62 (76.5)125 (78.6)
Age, years     
 Mean48.647.951.550.350.9
 Median (IQR)50.0 (39.0, 57.0)48.0 (39.0, 56.0)51.5 (45.0, 59.0)50.0 (41.0, 59.0)51.0 (42.0, 59.0)
Disease duration, years     
 Mean2.24.03.53.93.7
 Median (IQR)1.2 (0.5, 3.2)2.1 (0.8, 4.7)1.1 (0.5, 3.1)1.0 (0.6, 3.2)1.0 (0.5, 3.2)
No. of swollen joints (range 0–66)     
 Mean12.612.513.814.314.1
 Median (IQR)9.0 (6.0, 15.0)10.0 (7.0, 15.0)11.0 (8.0, 17.0)10.0 (7.0, 17.0)11.0 (8.0, 17.0)
No. of tender joints (range 0–68)     
 Mean25.525.626.026.126.1
 Median (IQR)23.5 (11.0, 36.0)22.0 (14.0, 33.0)25.0 (14.0, 36.0)25.0 (14.0, 34.0)25.0 (14.0, 35.0)
CRP, mg/dl     
 Mean2.22.12.52.52.3
 Median (IQR)1.2 (0.5, 2.5)1.2 (0.4, 2.9)1.1 (0.4, 2.7)1.5 (0.5, 3.4)1.3 (0.5, 2.8)
ESR, mm/hour     
 Mean42.445.044.043.343.7
 Median (IQR)35.5 (16.0, 61.0)42.0 (25.0, 60.0)39.5 (25.0, 55.0)37.0 (25.0, 58.0)39.0 (25.0, 58.0)
Radiographic total SvdH score (range 0–448)     
 Mean17.323.820.720.520.6
 Median (IQR)4.0 (2.0, 11.5)6.8 (2.5, 29.8)6.0 (2.0, 24.5)6.0 (2.5, 15.8)6.0 (2.0, 20.5)
RAMRIS scores     
 Synovitis     
  Wrist plus MCP joints (range 0–21)     
   Mean8.89.310.39.69.9
   Median (IQR)8.5 (5.0, 12.0)9.0 (4.0, 14.5)10.5 (7.0, 14.0)9.5 (6.0, 13.5)10.0 (6.5, 13.8)
  Wrist joints only (range 0–9)     
   Mean4.84.75.64.95.2
   Median (IQR)5.5 (2.0, 7.0)4.5 (2.5, 7.0)5.5 (3.5, 8.0)5.5 (3.0, 7.0)5.5 (3.0, 7.5)
 Bone edema (osteitis) (range 0–69)     
  Mean8.410.611.59.510.5
  Median (IQR)5.0 (2.5, 11.5)6.5 (1.5, 17.8)9.5 (4.0, 17.0)7.0 (2.0, 13.0)7.5 (3.0, 15.3)
 Bone erosion (range 0–230)     
  Mean17.721.025.021.323.1
  Median (IQR)13.5 (8.8, 19.0)15.8 (10.8, 25.3)15.5 (11.2, 27.0)13.3 (10.0, 19.5)13.9 (11.0, 22.8)

The interreader and intrareader reliability coefficients based on RAMRIS scores at week 24 were 0.90 and 0.95, respectively, for bone erosion; 0.81 and 0.87, respectively, for synovitis; and 0.68 and 0.90, respectively, for bone edema.

Changes in the RAMRIS scores from baseline to weeks 12 and 24 are summarized in Table 3. At week 12, significant improvements in the mean RAMRIS score for synovitis (−1.92 versus 0.14 [P < 0.001] for wrist plus MCP joints and −0.85 versus 0.02 [P < 0.001] for wrist joints only), for bone edema/osteitis (−1.82 versus 0.56; P < 0.001), and for bone erosion (−0.40 versus 0.24; P = 0.016) were observed in the combined golimumab plus MTX group relative to the placebo plus MTX group. Significant improvements in the RAMRIS scores at week 12 were also observed for the golimumab 100 mg plus placebo group versus the placebo plus MTX group (scores for synovitis and for bone edema) (Table 3 and Figure 2).

Table 3. Summary of changes from baseline in the RAMRIS scores and the total SvdH radiographic scores*
CharacteristicGroup 1, placebo plus MTXGroup 2, golimumab 100 mg plus placeboGolimumab plus MTX
Group 3, 50 mg golimumabGroup 4, 100 mg golimumabGroups 3 and 4 combined
  • *

    All P values are versus group 1. IQR = interquartile range; SvdH = modified Sharp/van der Heijde score.

  • Patients in group 1 who did not enter early escape continued taking placebo plus methotrexate (MTX) through week 52.

  • Due to an inability to obtain postcontrast images of both the wrist and the metacarpophalangeal (MCP) joints at the study site, the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) synovitis scores for the wrist plus MCP joints represent 272 patients, and those for the wrist joints only represent 288 patients.

  • §

    The sensitivity analysis using “no imputation” rules included the 238 patients in the magnetic resonance imaging (MRI) substudy who received subcutaneous study agent through week 24 with no missed doses and no missing baseline or week 24 data (75% of the 318 substudy patients randomized).

  • Evaluation of the RAMRIS bone erosion score with linear extrapolation was implemented for <25% of all MRI substudy patients.

No. of MRI substudy patients randomized82777881159
RAMRIS scores for synovitis     
 Wrist plus MCP joints (range 0–21)     
  Change from baseline to week 12     
   No. of patients57575860118
   Mean ± SD0.14 ± 2.98−1.76 ± 2.76−1.66 ± 2.90−2.18 ± 3.26−1.92 ± 3.09
   Median (IQR)0.00 (−1.00, 1.50)−1.00 (−3.50, −0.50)−1.01 (−2.50, −0.17)−2.00 (−4.00, 0.00)−1.50 (−3.50, 0.00)
   P <0.001<0.001<0.001<0.001
  Change from baseline to week 24     
   No. of patients81747780157
   Mean ± SD−1.04 ± 3.04−1.62 ± 3.42−2.21 ± 3.10−2.68 ± 3.61−2.45 ± 3.37
   Median (IQR)−1.00 (−1.63, 0.00)−1.08 (−2.50, 0.00)−1.50 (−3.50, −0.33)−1.50 (−4.50, −0.50)−1.50 (−4.00, −0.50)
   P 0.340.0110.001<0.001
 Wrist joints only (range 0–9)     
  Change from baseline to week 12     
   No. of patients62646064124
   Mean ± SD0.02 ± 1.61−0.61 ± 1.71−0.93 ± 1.56−0.77 ± 1.46−0.85 ± 1.51
   Median (IQR)0.00 (−1.00, 1.00)−0.50 (−1.50, 0.00)−0.50 (−2.25, 0.00)−0.50 (−1.75, 0.00)−0.50 (−2.00, 0.00)
   P 0.0190.0010.003<0.001
  Change from baseline to week 24     
   No. of patients82777881159
   Mean ± SD−0.74 ± 1.86−0.66 ± 1.80−1.29 ± 1.67−1.15 ± 1.89−1.22 ± 1.78
   Median (IQR)−0.50 (−1.00, 1.00)−0.50 (−1.50, 0.00)−1.00 (−2.50, 0.00)−0.50 (−2.00, 0.00)−0.88 (−2.00, 0.00)
   P 0.990.0360.070.024
  Sensitivity analysis for week 24 data, no imputation§     
   No. of patients61525763120
   Mean ± SD−0.91 ± 2.10−0.42 ± 1.87−1.48 ± 1.80−1.19 ± 2.02−1.33 ± 1.92
   Median (IQR)−0.50 (−2.00, 0.00)0.00 (−1.25, 0.25)−1.50 (−6.5, 2.5)−0.50 (−2.25, 0.00)−1.00 (−2.50, 0.00)
   P 0.290.080.340.13
RAMRIS scores for bone edema (osteitis) (range 0–69)     
 Change from baseline to week 12     
  No. of patients62646266128
  Mean ± SD0.56 ± 3.97−2.04 ± 5.69−2.47 ± 4.20−1.22 ± 4.88−1.82 ± 4.59
  Median (IQR)0.25 (−1.00, 2.00)−0.09 (−3.00, 0.50)−1.00 (−3.50, 0.00)−0.50 (−2.50, 0.00)−1.00 (−2.89, 0.00)
  P 0.001<0.0010.003<0.001
 Change from baseline to week 24     
  No. of patients82777881159
  Mean ± SD−0.32 ± 4.66−1.92 ± 6.12−2.47 ± 4.08−2.09 ± 5.57−2.27 ± 4.88
  Median (IQR)0.00 (−1.50, 1.00)−0.50 (−2.00, 0.00)−1.00 (−3.00, 0.00)−0.75 (−2.50, 0.50)−0.75 (−3.00, 0.00)
  P 0.043<0.0010.015<0.001
 Sensitivity analysis for week 24 data, no imputation§     
  No. of patients61545865123
  Mean ± SD−0.39 ± 5.36−1.71 ± 6.00−2.61 ± 4.35−2.40 ± 6.08−2.50 ± 5.32
  Median (IQR)0.00 (−2.00, 1.00)−0.50 (−2.73, 0.50)−1.25 (−3.33, 0.00)−1.00 (−3.00, 0.50)−1.00 (−3.33, 0.50)
  P 0.100.0050.0330.004
RAMRIS scores for bone erosion (range 0–230)     
 Change from baseline to week 12     
  No. of patients62646266128
  Mean ± SD0.24 ± 3.380.51 ± 3.09−0.82 ± 5.200.00 ± 0.83−0.40 ± 3.67
  Median (IQR)0.00 (0.00, 0.52)0.00 (0.00, 0.50)0.00 (−0.50, 0.50)0.00 (−0.07, 0.00)0.00 (−0.50, 0.50)
  P 0.0540.0360.0540.016
 Change from baseline to week 24     
  No. of patients82777881159
  Mean ± SD−0.24 ± 6.390.46 ± 2.71−0.65 ± 5.98−0.15 ± 1.46−0.40 ± 4.31
  Median (IQR)0.00 (0.00, 0.50)0.00 (0.00, 0.50)0.00 (−0.58, 0.00)0.00 (−0.50, 0.00)0.00 (−0.50, 0.00)
  P 0.970.0160.0280.010
 Sensitivity analyses for week 24 data,     
   no imputation§     
  No. of patients61545865123
  Mean ± SD−0.42 ± 7.390.39 ± 2.58−0.89 ± 6.92−0.16 ± 1.61−0.50 ± 4.88
  Median (IQR)0.00 (0.00, 0.52)0.00 (0.00, 0.50)0.00 (−1.00, 0.00)0.00 (−0.50, 0.50)0.00 (−0.62, 0.50)
  P 0.810.0280.070.023
 Linear extrapolation     
  No. of patients82777881159
  Mean ± SD−0.24 ± 6.400.50 ± 2.97−0.70 ± 6.00−0.11 ± 1.57−0.40 ± 4.35
  Median (IQR)0.00 (0.00, 0.50)0.00 (0.00, 1.00)0.00 (−0.82, 0.00)0.00 (−0.50, 0.50)0.00 (−0.50, 0.00)
  P 0.900.0160.0620.017
Radiographic total SvdH scores     
 Change from baseline to week 28     
  No. of patients82777881159
  Mean ± SD0.92 ± 2.960.39 ± 3.180.83 ± 4.380.17 ± 1.410.49 ± 3.23
  Median (IQR)0.00 (0.00, 1.00)0.00 (0.00, 0.50)0.00 (0.00, 0.50)0.00 (0.00, 0.50)0.00 (0.00, 0.50)
  P 0.100.310.230.19
Figure 2.

Changes from baseline to week 12 and week 24 in the Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) system scores for A, synovitis (wrist plus metacarpophalangeal [MCP] joints), B, bone edema/osteitis, and C, bone erosion. Graphs show the mean (solid horizontal line), median (dotted horizontal line), and interquartile range (bars) for changes from baseline in each treatment group. = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001 versus group 1, by analysis of variance on the van der Waerden normal scores. MTX = methotrexate; GLM = golimumab; PBO = placebo.

Similar response patterns were observed at week 24, with significantly greater improvement in the RAMRIS scores for synovitis (−2.45 versus −1.04 [P < 0.001] for wrist plus MCP and −1.22 versus −0.74 [P = 0.024] for wrist only), for bone edema/osteitis (−2.27 versus −0.32; P < 0.001), and for bone erosion (−0.40 versus −0.24; P = 0.010) in the combined golimumab plus MTX group relative to the placebo plus MTX group. Results of sensitivity analyses were generally supportive of the findings of the primary analyses, although several numerical differences between the golimumab plus MTX group and the placebo plus MTX group did not reach statistical significance (Table 3). Significant improvement in bone edema/osteitis from baseline to week 24 was also observed for the golimumab 100 mg plus placebo group (Table 3 and Figure 2).

Representative MRIs obtained at baseline, week 12, and week 24 in a patient who was randomized to receive treatment with golimumab 100 mg plus MTX are shown in Figure 3. Coronal STIR images show bone edema that was extensive at baseline, markedly decreased at week 12, and nearly resolved at week 24 (Figures 3A–C). Corresponding postcontrast T1-weighted images with fat suppression show synovitis that was extensive at baseline, markedly reduced at week 12, and only minimal at week 24 (Figures 3D–F). Precontrast T1-weighted images without fat suppression show no progression of bone erosion during the 24-week followup period (Figures 3G–I).

Figure 3.

Magnetic resonance images of the wrist obtained at baseline, week 12, and week 24 in a patient receiving golimumab 100 mg plus methotrexate. A–C, Coronal short tau inversion recovery (STIR) images show extensive bone edema in many wrist bones, including the lunate, triquetrum, trapezoid, and metacarpal bases 2, 3, and 4 at baseline. The bone edema has markedly decreased at week 12 and has nearly resolved at week 24. D–F, Corresponding postcontrast T1-weighted images with fat suppression show severe synovitis in both the radiocarpal and intercarpal/carpometacarpal joints at baseline and reduced synovitis at week 12 and particularly at week 24. G–I, Precontrast T1-weighted images without fat suppression show no progression of bone erosion during the 24-week followup period. Images shown here are representative of the series of consecutive images that were evaluated.

Comparison of radiographic data, which are reported as the change in SvdH scores from baseline to week 28, among the 318 MRI substudy patients showed no significant difference between the placebo plus MTX group and the combined golimumab plus MTX group (mean of 0.92 versus 0.49 change from baseline to week 28; P = 0.19) (Table 3). The same evaluation using the whole study population (n = 637) at week 28 did show a statistically significant difference between the combined golimumab plus MTX group versus the placebo plus MTX group (17).

DISCUSSION

We evaluated MRI data derived from the GO-BEFORE study, a randomized, controlled study in which the efficacy and safety of golimumab were assessed in MTX-naive patients with active RA. The study results related to signs and symptoms, safety, and radiographic data have been reported (17, 18). These results demonstrated that therapy with golimumab in combination with MTX reduced the signs and symptoms and the radiographic progression of RA in MTX-naive patients, with a safety profile similar to that of other anti-TNF agents.

An MRI substudy was conducted as part of the GO-BEFORE study, in which serial measurements of synovitis, bone edema/osteitis, and bone erosion evaluated according to the RAMRIS system were obtained. This MRI substudy demonstrated very good reliability and a high level of sensitivity to change (23). We also observed good interreader and intrareader reliability, with coefficients of 0.90–0.95 for bone erosion, 0.81–0.87 for synovitis, and 0.68–0.90 for bone edema/osteitis RAMRIS scores at week 24.

Results of the MRI substudy demonstrated that patients who received golimumab plus MTX had improvements in inflammation (as evidenced by the RAMRIS scores for synovitis and bone edema/osteitis) and bone erosion (as evidenced by the RAMRIS scores for bone erosion) that exceeded those observed in patients who received placebo plus MTX. These findings were observed as early as week 12 and continued through week 24 (Figures 2 and 3). These significant improvements in RAMRIS scores were generally observed for both golimumab plus MTX dosage groups, with the exception of the RAMRIS bone erosion score at week 12 (P = 0.054 for 100 mg plus MTX versus placebo plus MTX). This sporadic lack of statistical significance is likely due to the small sample sizes in the individual golimumab dosage groups. Similar trends were observed in the sensitivity analyses conducted for the mean change in RAMRIS scores from baseline to week 24, although several numerical differences between the golimumab plus MTX group versus the placebo plus MTX group were no longer statistically significant.

Treatment with golimumab 100 mg plus placebo also yielded numerically greater improvement in all RAMRIS scores by week 12 than treatment with MTX plus placebo; these differences were statistically significant for wrist synovitis and bone edema/osteitis and approached significance for bone erosion. By week 24, patients treated with MTX monotherapy and with golimumab 100 mg plus placebo had a similar extent of improvement in RAMRIS scores for synovitis and for bone erosion. However, golimumab monotherapy retained an advantage over MTX monotherapy in terms of bone edema/osteitis at week 24. Thus, in these MTX-naive patients with early RA, golimumab demonstrated a faster onset of action in quelling bone inflammation (osteitis) relative to that of MTX alone.

In a comparison of MRI and conventional radiographic assessments (SvdH scores) obtained on 2 occasions 1 year apart in 35 RA patients and 9 healthy control subjects, Ejbjerg and colleagues (24) demonstrated that MRI is significantly superior to radiography in the detection of progressive joint destruction of the hands, wrists, and forefeet. The results from our study are consistent with these findings reported by Ejbjerg and colleagues. As reported previously, golimumab plus MTX demonstrated statistically significant inhibition of radiographic progression compared with placebo plus MTX in the overall study patient population (the GO-BEFORE study) (17). However, achievement of the statistically significant difference required twice as many patients (n = 637) and twice as much time (28 weeks) as was needed for the MRI substudy (n = 318 patients and 12 weeks). While the radiographic evaluation of the MRI substudy population at 6 months did show a numerical difference between the combined golimumab plus MTX group and the placebo plus MTX group, it did not reach statistical significance, documenting that MRI is a more sensitive tool for detecting structural damage. This may be particularly important in evaluating structural damage in current clinical trials recruiting patient populations with milder disease (16) with less potential for radiographic progression.

In this GO-BEFORE MRI substudy of MTX-naive patients, MRI assessment revealed reductions in synovitis, osteitis, and bone erosion at week 12 and onward in patients receiving golimumab plus MTX combination therapy as compared with those receiving MTX monotherapy. Requiring less than half the followup time and half the number of patients, MRI assessment thereby confirmed the clinical and radiographic findings previously reported for the overall study population.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Østergaard had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Østergaard, Emery, Conaghan, Fleischmann, Hsia, Xu, Rahman.

Acquisition of data. Østergaard, Emery, Conaghan, Fleischmann, Hsia, Xu, Rahman.

Analysis and interpretation of data. Østergaard, Emery, Conaghan, Fleischmann, Hsia, Xu, Rahman.

ROLE OF THE STUDY SPONSOR

Centocor Research & Development, a division of Johnson & Johnson Pharmaceutical Research & Development, LLC, and Schering-Plough/Merck Research Institute, Inc. sponsored this study. The following steering committee members who were also authors of this article assisted Centocor scientists with the study design and conduct: Mikkel Østergaard, MD, PhD, DMSc, Paul Emery, MA, MD, FRCP, Philip G. Conaghan, MB, BS, PhD, FRACP, FRCP, and Roy Fleischmann, MD. Data were collected by the investigators and entered into a Centocor database. Centocor statisticians and programmers conducted the analyses, and members of the steering committee, with the assistance of medical writer Michelle Perate, MS (independent consultant for Janssen Biotech Inc., a wholly owned subsidiary of Johnson & Johnson, Inc.), prepared the manuscript. All authors reviewed and approved the manuscript content before submission and jointly agreed to submit the final version of the manuscript. Publication of this study was not contingent upon approval by Centocor or Schering-Plough/Merck.

Acknowledgements

We thank the patients, investigators, and study personnel who made these trials possible. We also thank Michelle Perate, MS (independent consultant for Janssen Biotech Inc., a wholly owned subsidiary of Johnson & Johnson, Inc.) who helped prepare the manuscript.

Ancillary