Article Text

This article has a correction. Please see:

Download PDFPDF

Concise report
Bone loss in patients with active early rheumatoid arthritis: infliximab and methotrexate compared with methotrexate treatment alone. Explorative analysis from a 12-month randomised, double-blind, placebo-controlled study
  1. G Haugeberg1,2,
  2. P G Conaghan3,
  3. M Quinn3,
  4. P Emery3
  1. 1
    Norwegian University of Science and Technology, Trondheim, Norway
  2. 2
    Departement of Rheumatology, Sørlandet Hospital, Kristiansand, Norway
  3. 3
    Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, University of Leeds, Leeds, UK
  1. Correspondence to Professor P Emery, Section of Musculoskeletal Disease, Chapel Allerton Hospital, Chapeltown Road, Leeds LS7 4SA, UK; p.emery{at}


Objective: To examine the effect of infliximab plus methotrexate (MTX) compared with placebo plus MTX on bone loss in patients with early rheumatoid arthritis (RA) in a double-blind randomised study design. Further, to explore the associations between bone loss and markers of RA disease.

Methods: All 20 patients with RA (10 patients in each treatment group) had active, early RA. Bone mineral density (BMD) was assessed at the hand, lumbar spine (L2–4) and hip by dual energy x-ray absorptiometry at baseline and 12 months’ follow-up. Clinical data were collected at regular visits.

Results: BMD loss was significantly reduced in the infliximab group compared with the placebo group at the femoral neck (−0.35% vs −3.43%, p = 0.01) and total hip (−0.23% vs −2.62%, p = 0.03) but not at the hand (−2.09% vs −2.82%, p = 0.82) and spine (−0.75% vs −1.77%, p = 0.71). Measures of disease process and joint damage were found to be independently associated with bone loss.

Conclusions: This study provides strong evidence of a causal link between inflammation and bone loss in RA. The anti-inflammatory effect of infliximab was potent enough to arrest inflammatory bone loss at the hip but not at the spine and hand.

Statistics from

In patients with rheumatoid arthritis (RA) with active disease the osteoclast plays a pivotal role in the development of generalised osteoporosis, periarticular osteoporosis and erosions.1

Anti-tumour necrosis factor (anti-TNF) treatment has proved to be more successful in retarding radiographic joint destruction than disease-modifying antirheumatic drugs—for example, methotrexate (MTX).2

Information on the effect of anti-inflammatory treatment, including anti-TNF treatment, on both localised and generalised osteoporosis in patients with RA is scarce.3 4 5 6 7

The effect of anti-TNF treatment on both generalised and localised osteoporosis has not previously been studied in a double-blind study design. Thus the aim of this randomised study was to examine the effect of infliximab on both generalised and localised bone loss in patients with early RA with active disease and to explore the associations between bone loss and markers of disease activity, joint damage and treatment.

Materials and methods

Study design, patients and data collection

Study design, clinical data, radiographic and magnetic resonance imaging (MRI) data from this 1-year double-blind randomised clinical trial have previously been described in detail.8

In summary, all examined patients with RA fulfilled the American College of Rheumatology classification criteria, had duration of symptoms of <12 months, had no previous treatment with disease-modifying antirheumatic drugs or oral corticosteroids, had metacarpal joint involvement, were receiving a stable dosage of non-steroidal anti-inflammatory drug and had a poor prognosis.

Ten patients received a standard dosage of infliximab (3 mg/kg of body weight) and 10 patients placebo at baseline, 2 weeks, 6 weeks and then at 8-week intervals. All patients were treated with MTX in escalating doses beginning with a dosage of 7.5 mg once weekly. For the infliximab and placebo group mean total parenteral dose of methylprednisolone in the period between 14 weeks and 54 weeks was 110 mg and 275 mg (no corticosteroids were permitted during the first 14 weeks). During the 54 weeks’ follow-up routine clinical trial data were collected at various time points.

To maintain the original study design of a blinded randomised controlled trial, the treatment code was kept secret for one of us who analysed the data (GH) until the analyses had been conducted.

Bone mineral density

Standardised BMD measurements at left and right hand, spine (L2–4, anterior posterior view) and left and right hip were performed by one technician using the same dual energy x-ray absorptiometry (DXA) equipment (Lunar Expert, Madison, Wisconsin, USA). In vivo short-term precision expressed as a percentage coefficient of variation was 1.07% at the hand, 2.33% at the spine L2–4, 1.43% at the total hip and 2.75% at the femoral neck. Long-term spine phantom percentage coefficient of variation was 0.80%. BMD was expressed as g/cm2. For the hip and whole hand we used the mean values from the left and the right sides.

Analysis and statistical tests

Categorical variables were expressed as numbers and percentages, and continuous variables as means with standard deviation if normally distributed and if not normally distributed as median with range. For group comparison we used a paired and independent two-tailed Student t test for variables with normal distribution and a two-tailed Mann–Whitney U test for skewed variables. For categorical variables we used a χ2 test.

Mean values for C-reactive protein (CRP), 28-joint count Disease Activity Score (DAS28), Health Assessment Questionnaire (HAQ), methylprednisolone and MRI synovitis score calculated from all time points, and progression of MRI erosions and radiographic joint damage from baseline to 54-weeks’ follow-up were used to search for associations between these variables and bone loss. We used Pearson correlation for normally distributed and Spearman correlation coefficients for non-normally distributed variables. Multiple linear regression analysis was used to explore the degree of explanation of the different variables to the bone loss at hand, lumbar spine L2–4 and hip. p Values ⩽0.05 were considered statistically significant. All analyses were performed with SPSS (Statistical Package for Social Sciences) program version 16.0.


Table 1 shows the baseline characteristics. No statistically significant differences between the two treatments groups were found for demographics, clinical disease characteristics, MRI synovitis score, radiographic and MRI joint damage score and BMD.

Table 1

Baseline characteristic for all patients with early rheumatoid arthritis and comparison between infliximab-treated and placebo-treated patients

At 54 weeks’ follow up a significant mean BMD loss was seen for all patients at the hand (−2.44%, p = 0.048), femoral neck (−1.83%, p = 0.007) and total hip (−1.40%, p = 0.017) but not at the lumbar spine L2–4 (−1.26%, p = 0.36). Figure 1 shows that BMD loss was lower in infliximab- than in placebo-treated patients; however, this difference was only statistically significant at femoral neck and total hip.

Figure 1

Percentage change (mean with 95% confidence interval) in bone mineral density (BMD) over 54 weeks in patients with early rheumatoid arthritis with active disease treated with placebo plus methotrexate (MTX) or infliximab plus MTX. Numbers below the graphs show the mean percentage changes with standard error of mean (SEM).

Associations with bone loss

Table 2 shows that no statistically significant correlation was found between change in BMD and age, gender, methylprednisolone and changes in vdHSharp score. A significant correlation between BMD and body mass index was only seen at the spine and between BMD and HAQ only at the femoral neck. CRP was significantly correlated with BMD loss at the hand and femoral neck, whereas DAS28 was only correlated with BMD loss at the femoral neck. A correlation between BMD loss and treatment group was only seen at femoral neck and total hip. For MRI synovitis a significant correlation was only seen for BMD at the spine and femoral neck. For MRI erosions a significant correlation was found with BMD at the femoral neck and total hip.

Table 2

Pearson (Pc) and Spearman (Sc) correlation coefficients (p value) between 54-week bone loss and baseline demographic and mean disease variables and changes in joint damage during 54-week follow-up in patients with early active rheumatoid arthritis

Based on the results from the correlation analysis we tested in multivariate linear regression models (forward procedure) the independent association between age, body mass index, CRP, HAQ, treatment groups, MRI synovitis and erosions with bone loss. Only an increase in CRP was found to be independently associated with hand bone loss (standardised β = −0.658, p = 0.003). At the spine only a high MRI synovitis score was found to be independently associated with bone loss (standardised β = −0.587, p = 0.010). At the femoral neck both an increase in CRP (standardised β = −0.428, p = 0.030) and MRI erosions (standardised β = 0.511, p = 0.012) was found to be independently associated with bone loss, whereas at the total hip only MRI erosions (standardised β = −0.776, p<0.001) was statistically significantly associated with bone loss.


This trial, of a remission-induction approach using a TNFα inhibitor in addition to escalating MTX doses in patients with early poor prognosis RA, showed that inflammatory bone loss was successfully arrested at the hip but not at the hand and lumbar spine. This has not previously been shown in a double-blind randomised study design.

In contrast to our study, previous studies with observational or randomised study design have reported that inflammatory bone loss is arrested by infliximab both at hip and lumbar spine.3 6 This discrepancy may be related to the small number of patients included in our study. As this study was not designed and powered to investigate BMD loss and its determinants, non-associations between bone loss and its determinants may thus be a result of lack of power. Another theoretical explanation of the discrepancy between bone loss at the spine and hip may be that hip joint synovitis was successfully treated in the infliximab group but not in the placebo group. Unfortunately, we do not have data on postmenopausal status or osteoporosis treatment. However, at baseline we found no difference between the two groups in gender distribution, female age or female BMD level. Despite the study limitations we found a strong association between markers of inflammation and generalised bone loss and hand bone loss as previously has also been reported by others.5 7 9 10

In cross-sectional studies an association between deterioration in radiographic hand joint damage score and low bone density at the hand and hip but not at the lumbar spine has been reported.11 12 13 The connection between hand bone loss and radiographic joint damage has also been reported from longitudinal observational studies.4 14 In our study we found no correlation between change in radiographic joint score and bone loss. However, for new MRI erosions a strong correlation was found with hip BMD loss.

In a large double-blind randomised RA study anti-inflammatory treatment with adalimumab was shown to significantly reduce the rate of cortical hand bone mass compared with placebo, but hand bone loss was not arrested and continued during follow-up.7 In our study, which included only a small number of patients, a non-significant lower rate of hand bone loss was seen in the infliximab-treated group (−2.09%) compared with placebo (−2.82%). It is still an open question if higher doses of TNF inhibitors or other treatment strategies would be able to stop inflammatory driven bone loss in RA.

As the numbers of MRI erosions did not increase as reported previously from this study and hand bone loss was not arrested in the infliximab-treated patients, this supports the view that DXA hand bone loss may be a more sensitive tool than even MRI to define remission of bone involvement in early RA.

In conclusion, our study provides strong evidence that there is a causal link between inflammation and bone loss in RA and that bone loss and erosions are features of the same inflammatory disease process. Further studies in RA are needed to elucidate the role of hand bone mass as marker of bone involvement and as response variable to treatment, and to explore the relationship between bone loss and erosions.



  • Funding Grants: PE is an ARC Professor of Rheumatology.

  • Competing interests None.

  • Ethics approval Approval from the ethical committee at NHS Leeds teaching hospitals.

  • 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.

Linked Articles

  • Miscellaneous
    BMJ Publishing Group Ltd and European League Against Rheumatism