Routine Assessment of Patient Index Data 3 and the American College of Rheumatology/European League Against Rheumatism Provisional Remission Definitions as Predictors of Radiographic Outcome in a Rheumatoid Arthritis Clinical Trial With Tocilizumab

Objective: The American College of Rheumatology/European League Against Rheumatism established definitions of remission for rheumatoid arthritis (RA) based on composite scores, including tender (TJC) and swollen joint counts (SJC), patient global visual analog scale (VAS) score, laboratory tests, and, in the Simplified Disease Activity Index (SDAI), the physician global score. Time constraints on a physician's schedule demand an easy yet accurate tool to measure disease activity. We assessed the predictive ability of the Routine Assessment of Patient Index Data 3 (RAPID3) with and without a single swollen joint versus the SDAI and/or Boolean remission criteria for functional and radiographic outcomes. Methods: Data were from the Tocilizumab Safety and the Prevention of Structural Joint Damage phase III trial in RA patients. We assessed the ability at year 1 of a RAPID3 score of ≤3 + 1 SJC, RAPID3 score of ≤3 (remission) without SJC, SDAI score of ≤3.3 (remission), and/or Boolean remission (SJC, TJC, patient global VAS, and C-reactive protein level [mg/dl] all ≤1) to predict year 2 Health Assessment Questionnaire (HAQ) disability index (DI) score of ≤0.5 (normal), no worsening of HAQ DI score from year 1, and no worsening of Genant-modified Total Sharp Score from year 1. Results: Among 690 patients, the mean ± SD baseline Disease Activity Score in 28 joints was 6.5 ± 0.96, RAPID3 score was 14.2 ± 5.51, and the SDAI score was 41.7 ± 13.01. Achieving year 1 measures was associated with good functional and radiographic outcomes at year 2. Sensitivity, specificity, positive predictive values, and negative predictive values were 49.1%, 83.2%, 37.4%, and 88.9% (RAPID3 remission); 26.4%, 91.7%, 36.8%, and 87.1% (RAPID3 + 1 SJC); 26.7%, 90.9%, 37.3%, and 85.9% (SDAI remission); and 17.0%, 96.6%, 47.4%, and 86.4% (Boolean remission), respectively. Conclusion: The predictive ability of RAPID3 (with or without joint count) was similar to that of SDAI and Boolean criteria.