Background As RCTs used for regulatory approval of RA treatments are not powered to provide serious adverse event rates, there is a need for large-scale observational safety studies. Because of strong channeling bias with new treatments, our study had matched controls to help better understand the long-term safety outcomes, as used in clinical practice, of abatacept (ABA), a selective T-cell co-stimulation modulator. We present preliminary results of our study describing baseline differences to help guide future analysis.
Objectives We present preliminary results of our study describing baseline differences to help guide future analysis.
Methods Patients with RA in the RAlly study were enrolled by their rheumatologist, and either started ABA or changed a DMARD or biologic disease modifiers (BDM) during 2009 through July 2014. Additional patients for both groups were added from the National Data Bank for Rheumatic Diseases (NDB). NDB controls were randomly matched by age, sex, and calendar date to NDB ABA patients in a 3:1 ratio. Control patients were separated into two groups allowing for possible overlap: DMARD (non-biologic) and BDM (biologic disease modifiers).
Results We compared 1274 ABA patients (3830 pt-yrs) with 3656 controls (1346 DMARD, 3094 BDM). The baseline mean (SD) age was 62 (13) yrs, rheumatic disease comorbidity index was 2.1 (1.6), and 15% were male in the ABA group. Significant clinical differences between ABA vs. controls included HAQ 1.30 (0.7) vs. 1.15 (0.7), pain VAS 5.6 (2.7) vs. 4.7 (2.8), patient global 5.0 (2.5) vs. 4.3 (2.5), and prednisone use 47% vs. 38% (see Table). ABA was the first-ever biologic exposure for 16% (n=199), and was taken without concomitant DMARDs in 24% (306). The intravenous form of ABA was used by 91% (1156) and 13% (164) used the sub-cutaneous form available in 2012, with 4% (46) patients treated with both.
Conclusions Preliminary results of our matched-controls safety study indicate that important differences remain in patients initiating ABA vs. other DMARDS and BDM in regular clinical practice. These differences, most notably prior medication exposure, prednisone use, comorbidities, and disease activity, should be accounted for when analyzing safety outcomes.
Acknowledgements The RAlly study is funded by Bristol-Myers Squibb.
Disclosure of Interest None declared