To date, no studies have been published on incident deformities in patients with rheumatoid arthritis (RA). Morphometric X-ray absorptiometry (MXA) is an alternative to conventional X-rays for identifying vertebral deformities. The aim of the present study was to describe the incidence of vertebral deformities in 255 female RA patients measured by MXA, and the relationship between incident deformities and clinical and demographic variables. MXA is still under evaluation for its ability to identify deformities, so we explored four different cut-off thresholds including fixed percentage reduction and the principle of least significant change (LSC). MXA (T4-L4) and BMD (L2-L4 and total hip; Lunar Expert) were performed on 255 patients (mean age 54.3, range 29.2-70.8 years) at baseline and after a mean period of 2.3 years. MXA scans were analyzed pairwise by the same trained technician, and incident deformities calculated applying LSC with a 99.9% and 99.99% confidence limit, and a fixed reduction of 20% and 25% for anterior, middle or posterior heights. Long term precision (%CV) of height measurements for all vertebrae combined (T4-L4) were 4.8, 4.8 and 4.4, respectively. Frequency and distribution of incident deformities varied from 39 deformities in 33 patients (fixed 20% reduction) to 17 deformities in 15 patients (fixed 25% reduction), and quality control analyses revealed a high number of presumed false deformities. Incidence per 100 patient years varied from 2.9 to 6.7 deformities according to method, and was comparable to those obtained from intervention studies in corticosteroid-induced osteoporosis. Patients with incident deformities were significantly older, had lower BMD, higher disability and more often a previous non-vertebral fractures than those without incident deformities Incident deformities by MXA need further evaluation in secondary osteoporosis. It seems, however, that older patients with previous limb fractures and low BMD are especially prone to this complication.