Background The nature of the effect of anti-TNF therapy on new bone formation in ankylosing spondylitis (AS) is still obscure. Clinical trial results suggest that new bone formation is neither inhibited nor augmented by anti-TNF agents. An often discussed hypothesis (“TNF brake”) suggests that syndesmophytes tend to develop in vertebral edges (VE) after resolution of inflammation (INF) due to anti-TNF agents. Our group and others have shown that INF at baseline (BL) as assessed by magnetic resonance imaging (MRI) correlates with syndesmophyte formation as assessed by conventional radiography after 2 years (y). Recently, fatty degeneration (FD) at VE was also identified as a potentially important factor for new bone formation in AS.
Objectives To directly compare the influence of INF and/or FD on the development of syndesmophytes in AS patients after 2 and 5y of anti-TNF therapy.
Methods An experienced reader scored MRIs and x-rays which were blinded for the time point of patients who participated in EASIC. Most patients were treated with infliximab, some with other TNF blockers. Presence or absence of INF, FD and syndesmophytes was documented on the level of vertebral edges (VEs) in the anterior part of the spine. Data were compared using Fisher’s exact test.
Results Complete sets of MRIs and x-rays of 73 AS patients were evaluated at 2 and 5y (1,062 VEs, 258 with and 804 without syndesmophytes or ankylosis at BL). A total of 5 syndesmophytes developed at 5y out of 128 VEs showing FD (3.9%) but no INF at BL, whereas no syndesmophytes developed in 116 VEs with INF but no FD at BL (p<0.05). When FD and INF were both present at BL, 3/61 syndesmophytes developed at 5y (4.9%). In more detail, out of 116 VEs with INF but no FD at BL 32 turned into FD at 2y (27.6%), but none of these gave rise to a syndesmophyte at 5y. The vast majority of those 61 VEs which had both, INF and FD, at BL continued to have FD at 2y (97%), 3 of which gave rise to a syndesmophyte at 5y (4.9%). There were 438 VEs without INF or FD at BL. Out of these 94 VEs with new FD developed at 2y (21.5%) and another 13 VEs at 5y (3%).
Conclusions The hypothesized sequence inflammation - fatty degeneration – new bone formation was not observed in patients with AS treated with anti-TNF agents over 5 years. Although there was a high proportion of VEs with INF at BL that developed FD at 2y, that was not followed by new bone formation at 5y. This finding contrasts with the TNF brake hypothesis. However, it cannot be excluded that the process from FD to syndesmophyte formation may even take longer. On the other hand, the observation that development of FD is decelerated between 2 and 5y of anti-TNF therapy, may suggest that repair processes such as FD which are potentially accelerated by certain medications are not harmful to patients in the vast majority of cases – especially in the first 5 years of treatment.
Disclosure of Interest None Declared