Table 1 Performance of “stringent” versus “less stringent” methods of reporting ACR (treatment response) and DAS28 (disease status) criteria according to their discriminatory capacity and their ability to reflect inhibition of structural damage progression or patient satisfaction
TechniqueLess stringentIntermediateStringent
ACR-based techniques at month 6ACR20ACR50ACR70
    Overall success, n/N (%)373/576 (65)205/576 (36)98/576 (17)
    NNS*, n6976125
    Structural damage at year 1†, LR+ (95% CI)1.11 (0.96 to 1.29)1.22 (0.93 to 1.60)1.18 (0.75 to 1.87)
    Patient satisfaction at month 6‡, LR+ (95% CI)2.79 (1.92, 4.05)InfiniteInfinite
DAS28-based techniques at month 6MDASLDASRemission
    Overall success, n/N (%)439/629 (70)147/629 (23)68/629 (11)
    NNS*, n4971103
    Structural damage at year 1§, LR+ (95% CI)1.20 (1.07 to 1.34)1.48 (1.06 to 2.06)1.71 (1.02 to 2.87)
    Patient satisfaction at month 6¶, LR+ (95% CI)2.09 (1.64 to 2.65)16.82 (4.23 to 66.94)Infinite

  • The total number of patients for each analysis was: *n  =  636; †n  =  551; ‡n  =  575; §n  =  602; ¶n  =  623. A lower number of patients needed to study (NNS) value indicates greater discriminatory capacity; higher positive likelihood ratio (LR+) values indicate a greater probability of observing no structural damage or satisfaction with treatment. Abatacept and placebo treatment groups were pooled for radiographic progression and patient satisfaction and unpooled for NNS. ACR, American College of Rheumatology; DAS28, disease activity score in 28 joints; MDAS, moderate disease activity state (DAS28 (C-reactive protein; CRP) <5.1); LDAS, low disease activity state (DAS28 (CRP) ⩽3.2); remission, DAS28 (CRP) <2.6.