Background/Purpose: CDAI (clinical disease activity index) and RAPID3 (routine assessment of patient index data 3) cut points defining responses that best match ACR20/50/70 response in rheumatoid arthritis (RA) patients (pts), are unknown. This study evaluates these cut points in a large study population treated with certolizumab pegol (CZP) plus methotrexate (MTX)1.
Method: ACR responders through Week (Wk) 12 from 393 pts treated with CZP (initial dose of 400 mg at Wks 0, 2 and 4 followed by 200 mg every 2 wks) plus MTX in RAPID 1 trial (NCT00152386) were categorized by proposed response cut points in CDAI (change from baseline CFB ≥6.7, ≥10.0, ≥13.9 and RAPID3 CFB ≥1.8, ≥3.6). ACR20/50/70 responses were compared with these proposed categorizations using cross-tabulations and kappa statistics. CART2 (classification and regression trees) modeling was used to identify CDAI and RAPID3 cut points most closely associated with ACR20/50/70 responses.
Result: At Wk 12, almost all (93–100%) ACR20/50/70 responders achieved CFB in CDAI (≥6.7, ≥10.0, ≥13.9); however fewer pts (73-80%) with CFB in CDAI (≥6.7, ≥10.0, ≥13.9) achieved an ACR20 response (Table 1). CFB in CDAI≥13.9 was most closely associated with ACR20 response (κ=0.57). Association between proposed categorizations and ACR50/70 was weak (κ range: 0.05–0.29). Nearly all (90–100%) ACR20/50/70 responders achieved CFB in RAPID3 (≥1.8, ≥3.6); fewer pts (75–80%) with a CFB in RAPID3 (≥1.8, ≥3.6) also achieved an ACR20 response (Table 2). CFB in RAPID3≥3.6 was most closely associated with ACR20 response (κ=0.55); association between proposed categorizations and ACR50/70 was weak (κ range: 0.08–0.30). CART modeling identified CFB in CDAI ≥13.80, ≥20.15 and CFB in RAPID3 ≥5.46, ≥7.28 as the categorizations most closely associated with ACR50/70 responses. There was better association between ACR50 response and CART-defined CDAI of CFB ≥20.15 (κ=0.43) compared with proposed categorizations (κ range: 0.14–0.29) (Table 1) and better association with ACR70. CART-defined RAPID3 categorizations were more closely associated with ACR20/50/70 responses; with better association between ACR50 and CART-defined RAPID3 categorizations (κ=0.45) than with proposed categorizations (κ range: 0.21–0.30) (Table 2).
Conclusion: Thresholds for CDAI and RAPID3 CFB were identified that defined the closest associations with ACR responses in pts with inadequate response to MTX and high disease activity at baseline (mean DAS28 6.9). Responses based on cut points closely associated with ACR20 were identified, but association with ACR50 and ACR70 was not as strong. Further studies are needed to analyze different pt populations, including those with lower disease activity at baseline.
- Keystone E, et al. Arthritis Rheum 2008; 58(11):3319–29.
- Classification And Regression Trees (CART) software, Salford Systems, CA, USA
Disclosure: M. H. Schiff, UCB Inc, 2, UCB Inc, 5 ; K. Luijtens, UCB Inc, 3 ; O. Davies, UCB Inc, 3, UCB Inc, 1 ; Y. Yazici, UCB Inc, 5 .