Background/Purpose: Poor prognostic factors can determine the extent of disease progression, disability and treatment outcomes in patients (pts) with RA. It is currently unknown whether the presence of poor prognostic factors influences treatment decisions in pts with RA. The purpose of this study was to report patterns of medication use and the change in CDAI score between baseline and 12 months by prognostic factors.

Methods: Using the Corrona RA registry, we identified pts with RA who were biologic naïve at enrollment and had a follow-up visit at 12 months (± 3 months). Pts were characterized at enrollment in terms of RA prognosis based on the 2008 ACR treatment recommendations,¹ including functional limitation (based on modified Health Assessment Questionnaire), extra-articular disease (Sjögren’s syndrome, RA lung disease and/or nodules), seropositivity (RF and/or anti-cyclic citrullinated peptide antibodies) and erosions. Pts were categorized as having 0–1, 2 or 3+ poor prognostic factors. Outcomes investigated included the use of a biologic/targeted synthetic DMARD (b/tsDMARD) over the 12-month follow-up period and examination of initiation of any DMARD (e.g. conventional or b/tsDMARD). Logistic regression models (unadjusted and adjusted for sex, age and baseline CDAI) examined the relationship between prognosis and outcomes. In the subset of pts with CDAI score at baseline and 12 months (n=3510), the reduction of disease activity was examined in adjusted models in the three poor prognosis groupings.

Results: There were 3621 pts enrolled on/after January 2005 who met the selection criteria: 1554 (42.9%), 1263 (34.9%) and 804 (22.2%) pts with 0–1, 2 or 3+ poor prognostic factors, respectively. An increased number of poor prognostic factors was associated with older age (median age: 58, 60 and 62 years, p<0.001), greater disease duration (median: 1, 2 and 4 years, p<0.001) and increasing disease activity (median CDAI score: 7, 9.6 and 14, p<0.001, for the 0–1, 2 or 3+ prognosis groupings, respectively). The proportion of pts initiating a b/tsDMARD was greatest in those with 3+ vs 0–1 poor prognostic factors (p=0.024). However, when adjusting for differences in baseline characteristics (including CDAI score), there was no significant relationship between poor prognosis and b/tsDMARD use (Table). There was no significant relationship between poor prognosis category and any DMARD initiation in the unadjusted and adjusted analyses (Table). After adjusting for CDAI score at enrollment, mean reduction in CDAI score over 12 months was significantly less for those with 3+ vs 0–1 poor prognostic factors (Table; p<0.001).

Conclusion: These findings suggest that the presence of poor prognostic factors does not influence treatment decisions. This may warrant reconsideration as there was a diminished reduction in disease activity in those with a greater number of poor prognosis factors. 1. Saag K, et al. Arthritis Rheum 2008;59:762–84.