Generalised osteoporosis is a feature of established rheumatoid arthri
tis but whether this is a consequence of treatment, immobility, or dis
ease activity has been unclear. We estimated bone mineral density by d
ual energy x-ray absorptiometry on 148 patients with early rheumatoid
arthritis before treatment with corticosteroids or disease-modifying d
rugs and 730 normal controls. Scans were done at 12-month intervals in
patients and at 0 and 12 months an 50 of the controls matched for men
opausal status. At presentation, bone mineral density of patients did
not differ from controls. However, patients with disease for less than
6 months had significantly higher spinal bone mineral density than th
ose of longer duration. Over the next 12 months, bone mineral density
loss was greater in patients with rheumatoid arthritis compared with c
ontrols; significantly so for early disease (eg, - 2.4 [0.8] vs - 0.6
[0.4] g/cm(2), p < 0.05 in the spine and - 4.3 [0.8] vs - 0.4 [0.5] g/
cm(2), p < 0.001 in the trochanter). For the lumbar spine, only diseas
e activity was significantly associated with this bone mineral density
loss. For patients with active disease over 2 years, mean bone minera
l density loss at each site was between 5.5 and 10% (p < 0.01 compared
to patients with inactive disease). Suppression of disease activity s
tabilised this bone loss. In patients with rheumatoid arthritis signif
icant amounts of generalised skeletal bone were test early in the dise
ase and the loss was associated with disease activity. These findings
have implications for the management of patients with rheumatoid arthr
itis and possibly other inflammatory diseases.