Background/Purpose: It is widely believed that use of glucocorticoids (GC) increase risk of fracture through reduction in bone quality as well as reduction in bone mineral density (BMD). Data supporting the ‘bone quality hypothesis’ come from a post-hoc analysis of two trials with limited numbers of fractures (van Staa). In a preliminary unadjusted analysis from a large cross-sectional study we surprisingly found that patients on GC had a lower prevalence of fractures.(Bukhari et al)  We now report the effect of GC on fracture in the same population, adjusting for BMD and other risk factors including treatments at presentation.

Methods: We used data from patients referred by their general practitioner or secondary care doctor for first BMD estimation by Dual X-Ray absorptiometry (DEXA) in the North West of England. Patients details recorded at time of scan included age, Body Mass Index (BMI), and the presence of risk factors: smoking, alcohol, family history of osteoporosis, history of clinical fragility fracture, rheumatoid arthritis and drug treatment, specifically GC and antiresorptive agents (mostly bisphosphonates). DEXA assessed BMD in the spine (mean of L1-L4 vertebrae) and femoral neck. Two subpopulations were identified at the extremes: patients on current GC and those with no prior exposure to GC. The prevalence of BMD fractures in each group was compared by chi-square test. A logistic model was fitted to adjust for possible confounding by imbalances in risk factors and BMD.

Results: 20210 subjects were included in the analysis, mean age was 63.2 years (SD 13.0 years), 82% were female. 32% had as history of fracture and 3772 (19%) were on GC at referral. Patients on GC vs unexposed were of similar age: 63.1 vs 62.8 years (p=0.21); As previously reported GC patients had higher bone mass at both sites 1.09 vs 1.04 g/cm2 at lumbar spine (p<0.001) and 0.9 vs 0.90 g/cm2 in the femoral neck (p<0.01). Less GC patients had a history of fracture: (13% vs 22%; OR 0.53 95%CI 0.49,0.58 p<0.001). 19.5% of GC patients were on antiresorptive agents vs 7.2 % of controls (P<0.01) . Modelling steroids as a dependent variable adjusting first for BMD at hip and then BMD at spine and for all the risk factors, gave an odds ratio of fracture 0.56(95%CI 0.52; 0.61) and 0.55(95%CI 0.51,061) repsectively. Restricting the analysis to patients who were not on antiresorptive treatment at the time of scanning (n=18,226 in the spine and 17,490 in the hip) yielded similar results: spine OR  of fracture 0.54 (95%CI 0.49; 0.60); hip OR of fracture 0.54 (95%CI 0.49; 0.60).

Conclusion: The surprising, apparently protective effects of GC on bone quality from this large cross-sectional study cannot be explained by known confounding factors. These results are likely subject to unmeasured confounding and need to be interpreted with caution.  The finding could be due to confounding by indication for DEXA .However, they suggest that current GC use is associated with less prevalent fractures at the same bone mass level, directly contradicting commonly held views about the detrimental effect of GC on bone. It will be worth exploring in a large longitudinal cohort.

References: Van Staa TP et al, Arthritis and Rheumatism 2003:48;3224-9. Bukhari et al, Ann Rheum Dis 2015;74(Suppl2): 527.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/compared-to-non-users-current-glucocorticoids-users-have-less-prevalent-fractures-at-the-same-bone-mass-results-of-a-large-cross-sectional-study/