Article Text

Extended report
Relationship of patient-reported outcomes with MRI measures in rheumatoid arthritis
  1. Joshua F Baker1,2,3,
  2. Philip G Conaghan4,
  3. Paul Emery4,
  4. Daniel G Baker5,
  5. Mikkel Ostergaard6,7
  1. 1Philadelphia Veterans Affairs Medical Center, Philadelphia, Pennsylvania, USA
  2. 2School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA
  3. 3Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania, USA
  4. 4Leeds Institute of Rheumatic and Musculoskeletal Medicine, University of Leeds, and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
  5. 5Janssen Research & Development, LLC, Horsham, Pennsylvania, USA
  6. 6Copenhagen Center for Arthritis Research, Center for Rheumatology and Spine Diseases, Glostrup, Denmark
  7. 7Faculty of Health and Medical Sciences, Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark
  1. Correspondence to Dr Joshua F Baker, Division of Rheumatology, Department of Medicine, 5 White Building, 3600 Spruce Street, Hospital of the University of Pennsylvania, Philadelphia, PA 19104, USA; bakerjo{at}


Purpose We assessed whether MRI measures of synovitis, osteitis and bone erosion were associated with patient-reported outcomes (PROs) in a longitudinal clinical trial setting among patients with rheumatoid arthritis (RA).

Methods This longitudinal cohort of 291 patients with RA was derived from the MRI substudy of the GO-BEFORE randomised controlled trial of golimumab among methotrexate-naïve patients. Correlations between RAMRIS scores (synovitis, osteitis, bone erosion) and physical function (Health Assessment Questionnaire (HAQ)), pain and global patient scores were determined at 0, 12, 24 and 52 weeks. Correlations between interval changes were also assessed. Multivariable regression models using robust generalised estimating equations evaluated associations over all time-points and their relationship to other clinical disease activity measures.

Results Greater synovitis, osteitis and bone erosion scores were positively associated with HAQ at all time-points (all p<0.05) and with pain and patient global scores at 24 and 52 weeks. Over all visits, synovitis was associated with HAQ, pain and patient global scores (p≤0.03) independent of clinical disease activity measures. Improvements in synovitis and bone erosion were also associated with improvements in PROs. Less improvement in synovitis and progression in MRI erosion at 52 weeks were both independently associated with worsening in all PROs at 52 weeks while progression on X-ray was not associated. Similar associations were observed across treatment groups.

Conclusions MRI measures of inflammation and structural damage correlate independently with physical function, pain and patient global assessments. These observations support the validity of MRI biomarkers.

Trial registration number NCT00264537; Post-results.

  • Rheumatoid Arthritis
  • Disease Activity
  • Patient perspective

Statistics from

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.


The use of patient-reported outcomes (PROs) in clinical care of patients with rheumatoid arthritis (RA) enhances communication between patients and their providers. However, PROs are not always reflective of the inflammatory disease burden and may be influenced by comorbid conditions and other factors with disagreement between patients and their providers being common.1–3 In addition, patients who feel well and are in clinical remission can have structural damage progression.4 Thus, objective measures of disease activity are of interest. Because objective outcomes might be expected to correlate more modestly with patient assessments, it is important to determine whether a reduction in objective measures translates into improvement in PROs. MRI measures of synovitis, osteitis and bone erosion correlate with clinical disease activity and predict progression of structural damage.5–7 However, the relationship between MRI measures and PROs has been poorly described.

Studies previously showed that patients who experience structural damage progression as imaged by X-ray have worse physical function as measured by the Health Assessment Questionnaire (HAQ).8 ,9 Averaged over 9 years, an annual increase of 6 units in the van der Heijde-Sharp (vdHS) score was associated with an approximate increase in 0.2 in the HAQ.10 However, while progression of damage on X-ray at 1 year has been an important randomised controlled trial end point for many years, it is not clear that progression on X-ray at 1 year is associated with functional decline.

One small study reported that MRI measures at baseline could predict functional outcomes.11 However, we are aware of no studies that have comprehensively assessed associations between longitudinal MRI measures and changes in PROs. The aims of this study were to determine the extent to which MRI measures are associated with PROs including physical functioning, pain and patient global assessment of disease activity (PtGl) and to determine if changes in MRI measures are associated with changes in these PROs.


This study was a secondary analysis of the GO-BEFORE ( identifier NCT00264537) randomised trial MRI sub-study. Details of the trial design and results have been previously published.12–14 The GO-BEFORE study was performed in 637 methotrexate (MTX) and biologic-naïve subjects; 291 of these were included in this analysis, who had MRIs scored for synovitis, osteitis and/or bone erosion.

Patients 18 years or older who met the American College of Rheumatology 1987 criteria for RA for at least the past 3 months and had active disease were recruited into the MRI substudy at participating sites. Eligible patients were randomly assigned to receive placebo subcutaneous injection with MTX (group 1, N=77), golimumab 100 mg injection with placebo (group 2, N=70), golimumab 50 mg injection with MTX (group 3, N=71) or golimumab 100 mg injection with MTX (group 4, N=73). Subjects remained in their assigned group for 28 weeks and were then allowed predefined increases in treatment intensity (early escape), if they had <20% improvement in both swollen and tender joint counts. Overall, 43/291 (15%) met the early escape criteria between 28 and 52 weeks, including 14/77 (20%) in group 1, 8/70 (11%) in group 2, 10/71 (14%) in group 3 and 13/73 (15%) in group 4. An additional 55 (71%) from group 1 were started on golimumab after 52 weeks. Similarly, an additional 38 (62%) patients in group 2 crossed over from placebo to MTX at 52 weeks.

Data collection at each visit included independent, blinded assessments of disease activity using the Disease Activity Score (DAS)28 with C-reactive protein (CRP) (DAS28(CRP)), completion of the HAQ, as well as assessment of pain and PtGl scores using visual analogue scale.

Magnetic resonance imaging

The details of the acquisition of MRI images has been described elsewhere.7 ,8 ,15 Briefly, MRIs of the patient's dominant wrist and second to fifth metacarpophalangeal (MCP) joints were obtained using 1.5 T MRI with contrast enhancement. Images were scored by two independent readers who were blinded to the image time-point or sequence, patient identity and treatment group. The average score of two readers was determined for synovitis (0–21 for wrist plus MCP joints), osteitis (0–69) and bone erosion (0–230), using the RAMRIS system.7 ,16 Progression in the MRI erosion score was defined as a change >0.5, as previously described.7 ,8 ,17

Radiographs of hands and feet

Radiographs of hands and feet were performed at baseline, week 24 and week 52 and scored by two blinded, centralised readers using the vdHS method as previously described.18 ,19 X-ray progression was defined as a change in vdHS score of >0.5 as in previous studies.7 ,15 ,17 ,20–22

Statistical analysis

Data were analysed using STATA V.14 software (StataCorp, College Station, TX, USA). Correlations between synovitis, osteitis and bone erosion scores on MRI were assessed at individual time-points using Spearman's correlations and overall observations using robust linear models incorporating Generalised Estimating Equations (GEE) in order to cluster on patient. The changes in HAQ, pain and PtGl scores were also determined over the 12-week, 24-week, 52-week and 104-week intervals. Correlation between changes in PROs and changes in MRI measures were determined. Finally, independent associations between changes in synovitis, MRI progression and X-ray progression and changes in PROs at 1year were assessed using multivariable linear regression. Assessment for altered associations across treatment groups was performed by testing the significance of multiplicative interaction terms.


The basic characteristics of the study cohort are shown in table 1 and have been previously described.7 ,17 ,23 The correlations between MRI measures (synovitis, osteitis, bone erosion) with PROs (HAQ, pain and PtGl scores) are presented in table 2. MRI measures were associated with HAQ over all time-points. MRI measures were increasingly associated with pain and PtGl scores at later follow-up times.

Table 1

Basic characteristics of the study population

Table 2

Spearman's correlations at 0, 12, 24 and 52 weeks between synovitis, osteitis, bone erosion and HAQ, pain and patient global

Spearman's correlations between observed changes in MRI measures and changes in the PROs are shown in figure 1. Improvements in synovitis at 12, 24 and 52 weeks were generally associated with greater improvements in HAQ, pain and PtGl scores. Changes in osteitis were not significantly associated with changes in PROs at any follow-up time. Changes in bone erosion were associated positively with changes in pain and patient global at later follow-up times. In contrast, changes in vdHS at 52 weeks were not associated with changes in PROs and the relationship was numerically inverse.

Figure 1

Spearman's correlations between changes in MRI measures of synovitis, osteitis and bone erosion with patient-reported outcomes including HAQ, pain and patient global scores over 12, 24 and 52 weeks of follow-up. The dashed line represents a p value of 0.05. HAQ, Health Assessment Questionnaire; PtGl, patient global assessment of disease activity; vdHS, van der Heijde-Sharp score.

In longitudinal models incorporating all study observations, synovitis was significantly associated with HAQ, and this was independent of the DAS28(CRP) (table 3). Synovitis was also associated with pain and PtGl scores independent of CRP, swollen and tender joint counts. Greater bone erosion was associated with greater HAQ independent of synovitis and DAS28(CRP).

Table 3

Longitudinal univariate and multivariable regression models using robust generalised estimating equations assessing independent associations of synovitis, osteitis and bone erosion with HAQ scores, pain and patient global, over all visits

Univariate and multivariable association between synovitis, bone erosion progression and X-ray progression are presented in table 4. In multivariable models, significant associations were observed between the change in synovitis at 1 year and the change in HAQ (β: 0.053 (0.029 to 0.077) p<0.001), pain (β: 0.16 (0.058 to 0.25) p=0.002) and PtGl score (β: 0.16 (0.066 to 0.25) p=0.001) over the same interval. MRI erosion progression was associated with greater increase in HAQ, pain and PtGl scores (all p<0.01), while X-ray progression was not associated (all p>0.11). In a similar analysis, a unit change of 1 in the MRI erosion score at 1 year was associated with a positive change in HAQ (β: 0.045 (0.014 to 0.076) p=0.005).

Table 4

Multivariable regression to identify independent associations between MRI and X-ray progression with change in PROs at 52 weeks of follow-up

There was substantial overall improvement in HAQ over the first 52 weeks of the study among completers of the 1 year follow-up in the MRI cohort (−0.76 (0.73), p<0.001, N=187), while no significant improvement occurred between 52 and 104 weeks among those who completed the entire 2-year study follow-up (−0.036 (0.35), p=0.15, N=183). There was similar correlation between the change in MRI erosion and the change in HAQ in the first (r=0.16, p=0.03, N=187) and second year (r=0.15, p=0.04, N=183) of follow-up. In contrast, while there was a strong correlation between change in synovitis and change in HAQ over year 1 (r=0.24, p<0.001), poor correlation was seen in year 2 (r=0.032, p=0.67). These relationships were similar across treatment groups.

In these analyses, no attenuation of the coefficients was noted with adjustment for treatment group. In addition, no significant interactions by treatment group were noted.


The current study is the first to systematically assess relationships between MRI measures of synovitis, osteitis and bone erosion with PROs in a large randomised controlled trial setting. MRI measures correlated with physical functioning over the entire study period, and it was observed that (1) less improvement in synovitis and (2) MRI erosion progression were independently associated with unfavourable changes in HAQ, pain and PtGl score at 1 year.

Over all study visits, synovitis was associated with all PROs independent of clinical disease activity. Greater bone erosion scores were also associated independently with higher HAQ. Thus, for two individuals with similar clinical assessments, the individual with greater synovitis on MRI is likely to have worse pain and function. These data indicate that synovitis and bone erosion are complementary to other clinical parameters in terms of relevance to the patient experience.

Improvements in MRI measures of synovitis were generally correlated with greater improvements in all PROs. Specifically, at 52 weeks, the reduction in synovitis noted on MRI was significantly correlated with improvements in HAQ, pain and patient global scores. The correlations between PROs and MRI measures were similar across treatments received. These observations suggest that MRI measures of synovitis may be a reasonable surrogate end point in observational and early interventional studies.

While synovitis showed strong correlations with PROs, osteitis was not significantly associated. One previous study evaluated associations between osteitis and HAQ and found that, while there was correlation at baseline, there was no correlation after 6 years of follow-up.11 The current study, which found poor correlation between the change in HAQ and change in osteitis over time, supports this previous observation.

An important additional observation is that MRI erosion progression was independently associated with greater increases in PROs. We recently showed that MRI erosion has improved discriminative characteristics in the identification of the effective treatment arm.8 These observations from the current study show that short-term progression in bone erosion on MRI is also associated with declines in physical functioning. The current study suggests that progression in the MRI erosion score (>0.5) is associated with a change in HAQ of 0.35 at 1 year. In addition, a 4.4 unit change in MRI erosion score would translate into a change in HAQ of 0.2. The present study did not identify a relation between X-ray progression and functional decline over 1 year. However, as noted, earlier studies have observed such an association averaged over 9 years (6 units of vdHS=0.2 units of HAQ).10 Our results suggest that MRI provides comparable, if not improved, discrimination of functional decline. Of note, we found that changes in synovitis were more strongly correlated with HAQ during the treatment of active inflammation in year 1, while changes in bone erosion were correlated similarly throughout the 2-year period.

The current study is within a randomised controlled trial population and thus may not be entirely generalisable to other populations. Further study in longitudinal observational studies may be helpful to characterise these associations in other populations. This study was also limited in the evaluation of PROs that were available as part of the original randomised controlled trial. Future study should consider incorporating additional PROs in order to better understand their relationships with MRI measures. In addition, there have been advances in MRI techniques since this study was performed that may improve the visualisation of inflammatory and structural changes.

In conclusion, MRI measures of synovitis and bone erosion correlate with PROs. Improvements over time in MRI inflammation and deterioration in MRI damage correlate with changes in function, pain and patient global scores suggesting that these objective measures reflect how patients experience their disease.


Dr Baker would like to acknowledge the support of a Veterans Affairs Clinical Science Research and Development Career Development Award (IK2 CX000955). The contents of this work do not represent the views of the Department of the Veterans Affairs or the US Government.


Supplementary materials

  • Lay summary

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

  • Lay summary

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Handling editor Tore K Kvien

  • Contributors All authors made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data. All authors played a role in drafting the work or revising it critically for important intellectual content. All authors had final approval of the submitted manuscript. All authors agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved.

  • Funding JFB is supported by a Veterans Affairs Clinical Science Research and Development Career Development Award (IK2 CX000955).

  • Competing interests JFB has nothing to disclose. PGC has done speakers bureaus or consultancies for AbbVie, BMS, Janssen, Eli-Lilly, Novartis, Pfizer and Roche. PE has received consulting fees, speaking fees and/or honoraria from Pfizer, Merck, AbbVie, UCB, Roche, BMS, Lilly and Novartis (<US$10 000 each). DGB is an employee of Janssen Biotech. MO has received fees for consultancy or speaker fees and/or research support from Abbott, AbbVie, BMS, Boehringer Ingelheim, Celgene, Centocor, Eli-Lilly, GSK, Janssen, Merck, Mundipharma, Novo Nordisk, Pfizer, Schering-Plough, Roche, UCB, Takeda and Wyeth.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles