Background/Purpose

Systemic glucocorticoid use can increase fracture risk, although dose-specific effects are not well understood, especially in younger adults.  Underlying diseases (e.g., rheumatoid arthritis [RA]) can confound these associations.  This study evaluated the impact of glucocorticoid exposure on risk of fragility fracture in RA patients.

Methods

The 42,034 study patients identified in the MarketScan® claims database were newly treated for RA (index date) between 1/1/2005 and 12/31/2011, aged 18-64 years at index, had benefits coverage for ≥12 months pre-index, and no pre-index cancer.  Patients with pre-index glucocorticoid use were glucocorticoid-free for ≥ 12 months prior to their first glucocorticoid pharmacy claim, and had continuous benefits coverage from first glucocorticoid claim to index date.  Glucocorticoid use (any; cumulative, peak, and average daily dose; cumulative days exposed; days of continuous exposure; days since most recent exposure) was assessed for each patient-day and censored at the earliest of post-index fragility fracture, cancer diagnosis, or end of follow-up.  Exposures were converted to prednisone equivalents and cumulative measures included pre-index use.  Patient demographics at index and pre-index clinical characteristics were assessed.  Incidence rates per 1,000 person-years for post-index fragility fracture were stratified by glucocorticoid exposure.  Cox’s proportional hazards models estimated age- and sex-adjusted fracture risk by exposure level.

Results

Patients averaged 738 (standard deviation [SD] 615) post-index follow-up days.  Most patients (82%) had glucocorticoid exposure during the study (43% with cumulative dose < 675 mgs; 73% had peak dose ≥ 15 mg/day).  Exposed and unexposed patients were demographically similar (74% female; mean age 49.4 [10.7] and 49.8 [10.2] years); 1% had pre-index fracture.  Fragility fracture incidence rates (95% confidence intervals) were 5.1 (4.6, 5.6) and 16.8 (11.4, 24.0) at daily doses of 0 mg/day and ≥ 15 mg/day, respectively, and 4.8 (3.8, 6.0) and 13.7 (10.6, 17.4) at cumulative doses of 0 mg and ≥ 5400 mgs, respectively.  Among patients ever-exposed to glucocorticoids, fracture risk was significantly elevated at cumulative dose ≥ 2700 mg and daily dose ≥ 15 mg/day (Table). Following glucocorticoid discontinuation, fracture risk (adjusted for age, sex, and cumulative dose) was 32% lower at 60-181 days since last exposure versus current exposure. Analyses restricted to patients younger than 50 years also showed highest fracture risk at highest exposures.

Conclusion

In this large, relatively young cohort of RA patients, fracture risk increased by 2- to 3-fold at high levels of daily and cumulative glucocorticoid dose, and began to decline within months of glucocorticoid discontinuation.

Table:  Risk of Post-Index Fragility Fracture among Rheumatoid Arthritis Patients Ever-Exposed to Systemic Glucocorticosteroids