Background/Purpose:   Patients with rheumatoid arthritis (RA) have an increased fracture risk, however, this is confounded by a coexisting use of steroids. This study aims to quantify the association between fragility fracture risk and traditional risk factors, including bone mineral density (BMD), in an RA population that have never used steroids; further, we aim to differentiate the associations that trabecular (lumbar spine) and cortical bone (femoral neck) have with fracture risk.

Methods: Data was collated from all patients referred for a DXA scan at a North West of England hospital between 2004 and 2015. Patient’s baseline characteristics and BMD were recorded and body mass index (BMI) measured. Only patients with RA were included and those who had previous used steroids were excluded. The population was divided by previous fragility fracture status. Baseline variables were compared using the Student T-test (continuous variables) and chi-squared test (categorical variables). Patients details recorded at time of scan included age, sex, BMI, and the presence of risk factors including smoking, alcohol, and family history of osteoporosis. The predictors of fracture were modelled using traditional risk factors including lumbar spine and femoral neck BMD. This was quantified using logistic regression analysis, adjusted for age, gender and BMI.

Results:   669 patients, 85% female and of median age 65 years (IQR 57.9, 71.9) were included in the analysis. 189 (28%) of patients had previously sustained a fragility fracture. Traditional risk factors were not associated with previous fragility fracture sustainment, as shown in table 1. However, the number of risk factors incrementally increased risk of fracture (OR 7.21, 95% CI 5.22, 9.94). Higher BMD of both the lumbar spine and femoral neck was associated with reduced fragility fracture sustainment, as shown in table 2.  

Conclusion: This study of a large population of RA patients without steroid exposure revealed that low lumbar spine and femoral neck BMD was associated with fragility fracture sustainment, whereas individual traditional risk factors, were not associated with fragility fracture sustainment; rather, only the number of traditional risk factors was associated. Therefore, it may be that BMD is the only important predictor of future fragility fracture risk in RA patients without steroid exposure. Because lumbar spine BMD is not included within the FRAX tool, we would advocate its use when estimating risk.