Background/Purpose: It is unclear to what extent low dose glucocorticoids (GCs) impact bone health in patients with rheumatoid arthritis (RA). RA leads to bone loss which might be mitigated with GC via disease control, but GCs themselves can negatively affect the bone. Observational studies are prone to bias and randomized controlled trials (RCTs) possibly underpowered for safety signals. We performed a protocolized (dx.doi.org/10.17504/protocols.io.6qpvr3ombvmk/v1) systematic literature review and individual participant data (IPD) meta-analysis to overcome these limitations.

Methods: MEDLINE (via PubMed), Embase (via Ovid) and the Cochrane Central Register of Controlled Trials (via Cochrane Library) were searched for RCTs that compared low dose GC (≤ 7.5 mg/d prednisone equivalent) with placebo or any other control treatment (e.g., standard of care). Search results were screened separately by two reviewers. There were two co-primary endpoints: Participants with ≥ 1 fractures (clinical/symptomatic) and change in bone mineral density (BMD; in g/cm², as measured by dual x-ray absorptiometry); both over two years. BMD was analysed separately for lumbar spine and femoral (total hip or femoral neck) measuring sites. Analyses were based on one-stage models and the intention-to-treat approach. Missing outcome data were handled via multiple imputations by chained equations. I² was estimated from standard random-effects meta-analysis models. To identify subgroups especially vulnerable to bone loss at the lumbar spine, various baseline characteristics were assessed as effect modifiers. A sensitivity analysis used data ‘as observed’.

Results: Out of 2,336 articles, five RCTs conducted in twelve countries in Europe with 1,112 participants were included and shared IPD: [1-5]. All studies allowed concomitant treatment with DMARDs. Baseline characteristics, including lumbar spine and femoral BMD, were well balanced across groups (Table 1). Regarding femoral BMD, four trials provided femoral neck BMD, one trial had only total hip data available. BMD decreased in both groups over two years (Table 2). There was an increase in bone loss in the GC compared to the control group at the lumbar spine but not at the femur. 24 participants experienced at least one fracture, with similar odds in both groups. Sensitivity analyses yielded consistent results. Subgroup analyses did not reveal groups particularly susceptible to GC-induced bone loss at the lumbar spine (Figure 1). Data on anti-osteoporotic medication was available for one trial where 13% were treated with anti-osteoporotic drugs at baseline (bisphosphonates or denosumab). In those who were randomized to GC and used anti-osteoporotic drugs at baseline (n = 28), there was no bone loss at the lumbar spine.

Conclusion: Low dose GCs, used for two years to treat RA, lead to bone loss at the lumbar spine but not at the femur. Anti-osteoporotic drugs may prevent GC-induced bone loss at the lumbar spine. There was no subgroup identified to be especially vulnerable to GC-induced bone loss. References1. Boers Ann Rheum Dis 20222. Choy Ann Rheum Dis 20053. Wassenberg Arthritis Rheum 20054. Svensson Arthritis Rheum 20055. Kirwan NEJM 1995

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/fractures-and-changes-in-bone-mineral-density-during-two-years-of-low-dose-glucocorticoid-treatment-for-rheumatoid-arthritis-a-systematic-literature-review-and-individual-participant-data-meta-analys/