Background/Purpose: Guidelines support early intervention in rheumatoid arthritis (RA).^1,2 The aim was to prospectively test whether outcome assessment at 12 weeks (wks) using a patient-reported tool (RAPID3) was comparable with investigator-based assessment (CDAI) in RA patients (pts) treated with certolizumab pegol (CZP) for predicting attainment of low disease activity (LDA) or better at Wk 52. Both scoring tools are used as clinical measures of disease activity in the absence of inflammatory laboratory markers.

Methods: In this 52-wk study (PREDICT; NCT01255761) pts received CZP standard dosing regimen (400mg at Wks 0, 2, 4 [loading dose] then 200mg Q2W). Pts were ≥18 yrs with adult onset active RA, unsatisfactory response to ≥1 DMARD and ≤2 prior TNF inhibitors. Pts were randomized 1:1 to RAPID3 or CDAI for classification as responder/non-responder at Wk12. Treatment decision at Wk12 was based upon randomization arm (RAPID3 responder, ≤6 or 20% improvement from baseline [BL]; CDAI responder, ≤10 or 20% improvement from BL). Pts who showed no improvement by Wk12 (<1 point improvement in CDAI or no improvement in RAPID3) were classified as failures and withdrawn. All other pts continued CZP 200mg Q2W unless they reached high disease activity (HDA; CDAI>22 or RAPID3>12) at 2 consecutive visits, at which point they escaped. Pts were permitted concomitant DMARDs if their regimen was unmodified between Wks 0 and 12. Long-term treatment success was defined as DAS28(ESR) ≤3.2 (LDA or better) at Wk52. Safety was monitored. The results are presented for the full analysis set (FAS; all pts who had valid BL and post-BL measurements) and both arms were adjusted for baseline DAS28(ESR) scores. Non-responder imputation (NRI) was applied.

Results:

Enrolled pts (FAS; N=733) had longstanding disease (mean±SD 8.9±9.1 years), HDA (mean±SD DAS28[ESR] 6.3±1.1, RAPID3 16.1±5.6, CDAI 40.2±13) and over half (55.5%) were prior TNF inhibitor failures.

At Wk12, 64.7% (238/368) and 76.4% (279/365) of pts randomized to RAPID3 and CDAI, respectively, were classed as responders (Table 1). At Wk52, 31.5% (75/238) of Wk12 RAPID3 responders, and 32.3% (90/279) of Wk12 CDAI responders, achieved DAS28(ESR) LDA or better (Table 1) demonstrating similar positive predictive values of these measures. No new safety signals were observed.

Conclusion:

In pts with chronic, severe RA the physician-based assessment detected greater CZP response at Wk12 compared with the patient-reported tool. Although the two outcome measures were not statistically comparable at Wk12, the positive predictive value for LDA at Wk52 for both assessments was similar. Further analysis is needed to evaluate the sensitivity and specificity of these measures for clinical disease activity assessments and treatment decisions in RA.

References: 1. Saag K.G. Arthritis Rheum 2008;59(6):762-784; 2. Smolen J.S. Curr Med Res Opin 2011;27(2):315-325