Corticosteroid use in rheumatoid arthritis: prevalence, predictors, correlates, and outcomes.

OBJECTIVE

To determine the rate of current and lifetime use of corticosteroids, the degree of association between corticosteroids and rheumatoid arthritis (RA) activity and outcome, corticosteroid initiation and discontinuation rates, and the predictors associated with initiation and discontinuation.

METHODS

A total of 12,749 patients with RA were evaluated semiannually as to corticosteroid use, RA activity measures, RA outcomes, and predictors of initiation and discontinuation of corticosteroids.

RESULTS

Current corticosteroid use was 35.5% and lifetime use was 65.5%. Rheumatologists varied substantially in their use of corticosteroids. The primary patient-derived determinant of corticosteroid initiation, current use, and discontinuation was symptom severity, although 21-25% of patients in remission or with minimal disease activity continued taking corticosteroids. Within the pool of current users, 24.3% [95% confidence interval (CI) 23.2-25.3%] discontinued corticosteroids yearly, and among patients newly starting corticosteroids this rate was 56.9% (95% CI 53.4-60.7%). Corticosteroid initiation occurred at a rate of 8.9% (95% CI 8.4-9.3%) per year. Among corticosteroid users, persistent use (> 5 years) occurs in about one-third of patients. Corticosteroid use and duration of use is associated with severe outcomes for current and past users. For current users versus non-current users, covariate adjusted outcomes were: mortality 5.7% versus 2.6%, work disability 28.4% versus 17.2%, and total joint replacement 18.5% versus 13.0%.

CONCLUSION

Corticosteroid use is dynamic and is associated with RA severity. Corticosteroid use is also associated with adverse longterm outcomes, but the ability to discern causal associations is severely limited by confounding by indication. The idea of "once on corticosteroids, always on corticosteroids" is incorrect and applies to only a minority of patients.