Background Inflammatory rheumatic diseases such as rheumatoid arthritis (RA) and ankylosing spondylitis (AS) are associated with osteoporosis. Bone mass in RA and AS has been well evaluated by measuring areal bone mineral density (BMD) with DXA. Trabecular bone score (TBS) is a new method evaluating bone microarchitecture by assessing pixel gray- level variations in DXA images from lumbar spine.

Objectives In a case-control study, TBS was evaluated in patients with RA or AS and healthy controls (HC) and the results were compared to lumbar spine (LS) BMD measurements. Changes in LS and hip BMD and TBS score after initiating and during long term anti-TNFa treatment were also examined in a prospective study.

Methods In the case control study, 30 patients with RA (ACR criteria, 19 F; 12 post menopausal women; age [mean ± SD]: 56.9 ± 9.7 yrs; disease duration: 11.7 ± 8.8 yrs; 26 under low dosage corticosteroids [CTC]) and 30 patients with AS (modified NY criteria, 27 M, age 43.8 ± 13.4 yrs; disease duration: 13.0 ± 11.1 yrs; no CTC) were evaluated and compared to 50 HC (29 F, 12 post menopausal women, age: 46.6 ± 11.1 yrs). L2-L4 BMD and hip BMD were measured using DXA (Lunar GE iDXA). TBS was calculated from L2-L4 BMD images (TBS insight^®, Med-Imaps).

In the prospective study, a group of 20 patients requiring TNFa blocking agent (6 F; 12 AS, [age: 40.7 ± 16.1 yrs] and 8 RA [age 60.5 ± 9.7 yrs]; disease duration: 9.6 ± 9.8 yrs; 9 under low dose CTC) were followed for 2 years. LS BMD, hip BMD and TBS score were measured at baseline and after 6, 12 and 24 months of treatment.

Results - Case control study: RA patients had lower BMD and T score at the hip (p < 0.005) compared to HC. Hip T score in patients with AS was also decreased (p = 0.02). LS BMD did not differ between patients and HC. TBS was lower in RA and AS compared to HC: 1.242 ± 0.16 and 1.282 ± 0.13 vs 1.365 ± 0.14, respectively (p= 0.005).

- Prospective study: under anti-TNFa, LS and hip BMD at M24 increased (+ 6.3% and + 2.4% respectively), with significant changes at the spine (p< 0.001). In the whole group, TBS score slightly increased from baseline to M12 (1,304 ± 0.13 to 1.32 ± 0.09) without significance, then returning to initial values at M24 (1.309 ± 0.09). However, in patients with RA, TBS score decreased (baseline to M24: 1.362 ± 0.048 to 1.308 ± 0.07) (p = 0.032) while in patients with AS, TBS progressively increased (1.257 ± 0.15 vs 1.309 ± 0.11; NS).

Conclusions TBS score is decreased in inflammatory rheumatic diseases, especially in RA, suggesting alterations of bone micro architecture. Long term anti TNFa treatment is associated with a positive effect on (LS) bone mass but no major changes in TBS score. Surprisingly and on the contrary of AS, TBS score decreased in RA treated by TNFa blocking agents, suggesting different influences and responses of the bone to this drug’s class.

Disclosure of Interest None Declared