Osteoporosis is a well-known complication of rheumatoid arthritis (RA), related to disease activity, steroid use and immobility.1,–,6

In a 1-year observational study it has been suggested that infliximab can have a positive effect on bone mineral density (BMD) in patients with RA.6 In the present study, patients with RA who had been treated with infliximab for 2 years and more in the Slotervaart Hospital or the VU University Medical Center were included if BMD measurements at baseline and at 2-year intervals had been performed. Patients fulfilled the American College of Rheumatology 1987 criteria and had a disease activity score in 28 joints greater than 3.2 at baseline. Infliximab (3 mg/kg) was given at 0, 2, 6 and 14 weeks and thereafter every 8 weeks.

BMD measurements of the spine (L1–L4) and total hip were performed on a Hologic 4500 (Hologic, Waltham, Massachusetts, USA) or a Lunar DPX-IQ (Lunar, Madison, Wisconsin, USA) machine. Repeat measurements were performed on the same machine. In patients with available paired analogue hand x-rays, digital x-ray radiogrammetry (DXR-online; Sectra, Linköping, Sweden) measurement was performed. The study was approved by the local medical ethics committee.

Fifty-two patients were included, baseline characteristics are shown in table 1. BMD of the hip remained stable, BMD of the spine increased, whereas BMD of the hands decreased significantly during the follow-up period of 3.5 years (range 2–5.6 years) (table 2).

The BMD of the spine of prednisone users increased 6.2% (non-users −0.75%, p=0.002). Patients diagnosed with RA 10 years ago or earlier showed an increase of 5.5% (<10 years −0.2%, p=0.009). In a regression model the change in BMD in the spine was significantly associated with prednisone use at baseline and a RA duration of 10 years or more, even after correction for age and bisphosphonate use.

In a study of 102 RA patients treated with infliximab, a comparable discrepancy between generalised and localised changes in BMD was found.7 During 1 year bone loss at the spine and hip was arrested, whereas a significant decrease in BMD in the hand (−0.82%) was observed. Metacarpal cortical bone is possibly more sensitive to and directly influenced by inflammatory cytokines originating from adjacent synovial tissue.

The positive relationship between a change in BMD in the spine and prednisone use could be explained by the anti-inflammatory effect of prednisone, causing a net beneficial effect on bone density. It confirms data from a recent study concerning patients treated with adalimumab for 1 year, in which the concomitant use of prednisone was associated with an increased BMD of the neck of the femur f 2.5%.8 The increase in BMD in the spine in patients with a longer disease duration compared with a stable BMD in patients with a RA duration of less than 10 years could be explained by the fact that the patients with early RA usually have more active disease and are therefore more prone to losing bone. An alternative explanation could be age-related spondylarthrosis in those patients with a longer disease duration; however, this is not in line with the fact that the ages of those with long and short disease durations were overlapping.

A limitation of our study is the relatively small number of patients, but longitudinal data are usually of small sample sizes, and our results are in line with earlier data on the effects of tumour necrosis factor blockade agents on BMD.6,–,8

We found that the long-term use of infliximab in RA patients has beneficial effects on generalised bone loss, which might be associated with an increase in bone strength, and subsequently, a reduction in fractures. The continuing bone loss in the hands could reflect suboptimal suppression of local inflammation.