TITLE: “Ankylosing Spondylitis is associated with an increased risk of osteoporotic fractures: a population-based cohort study.”

Background/Purpose:

Patients with Ankylosing Spondylitis (AS) have reduced bone mass, and altered biomechanics at the spine, but their fracture risk remains unclear. We have used a large population-based public health database to study the association between AS and incident fractures.

Methods:

We screened the SIDIAP Database (www.sidiap.org) to identify those aged 15 years or older with a diagnosis of Ankylosing Spondylitis (ICD-10 M45), and ascertained incident osteoporotic (OP) fractures (any but skull, fingers and toes) in this population in the period 2006-2011. Five controls with no AS or other inflammatory arthritis were matched on gender, age and General Practitioner (GP). SIDIAP contains the anonymised medical records and pharmacy invoice data of a representative >4.9 million people (80% of the total population). Cox regression stratified on matched sets was used to estimate adjusted (body mass index, smoking, alcohol and oral corticosteroids) hazard ratios (HR) according to AS status. We tested for a priori defined interactions with age, gender, inflammatory bowel disease (IBD) and regular NSAID use.

Results:

We identified 6,474 AS patients (0.14% of the population) and 32,346 controls and observed them for a median (inter-quartile range) of 5.98 (2.67-5.99) years. OP and clinical spine fracture rates were 9.64/1,000 person-years (8.52-10.90) and 2.12(1.63-2.76) among AS compared to 8.05 (95%CI 7.57-8.56) and 1.05(0.88-1.24) in controls respectively [Figure]. Adjusted HRs for OP and clinical spine fractures were 1.18 (95% CI 1.02-1.36; p=0.02) and 1.90 (1.37-2.63; p<0.001) respectively for AS patients. Further adjustment for oral corticosteroid use attenuated the association with OP fractures (HR 1.15 (0.99-1.32; p=0.06) but not with clinical spine fractures (HR 1.80 (1.30-2.51); p<0.001). No interactions were present with IBD. By contrast, there were relevant interactions with NSAID use (p=0.001) and gender (p=0.07) on OP fracture: adjusted HRs 1.20 (1.02-1.42; p=0.002) and 0.75 (0.55-1.02; p=0.07) for NSAID non-users VS users respectively [Figure]; and 1.01 (0.80-1.26; p=0.95) for females and 1.28 (1.08-1.53; p=0.006) for males. Similar interactions were found for NSAID use and age on clinical spine fractures (p=0.02 and p=0.03 respectively) [Figure].

Conclusion:

Patients with AS are at 18% higher risk of osteoporotic fracture related to use of oral corticosteroids, and at almost double risk of clinical spine fractures, independently of oral corticosteroids. The excess risk associated with AS is biggest in younger males who do not take oral NSAIDs regularly.

Figure.