Article Text
Abstract
Background RAPID3 (Routine Assessment of Patient Index Data) indicates differences in efficacy of active versus control treatments at levels similar to DAS28-ESR (Disease Activity Score 28-Erythrocyte Sedimentation Rate) and CDAI (Clinical Disease Activity Index) in clinical trials of adalimumab, abatacept, certolizumab.
Objectives To compare improvement according to RAPID3, DAS 28-ESR, and CDAI in the RA-BEAM trial of baricitinib vs adalimumab and placebo.
Methods Post-hoc analyses were performed of the RA-BEAM trial, in which patients with moderately to severely active rheumatoid arthritis and an inadequate response to methotrexate (MTX) were randomized to baricitinib, adalimumab, or placebo. All patients were to continue stable background MTX and other DMARDs, as well as stable low-dose prednisone and/or NSAIDs, if indicated. A RAPID3-like index was computed from 3 measures: physical function (FN), pain (PN), and patient global assessment (PATGL). FN on a HAQ (Health Assessment Questionnaire) of 20 items [rather than MDHAQ (multidimensional HAQ) of 10 items] was recalculated from 0–3 to 0–10; PN and PATGL (visual analog scales) were recalculated from 0–100 to 0–10, for a 0–30 total score, hence “RAPID3-like”. Mean values at baseline and Week 24 for RAPID3-like, DAS28-ESR, and CDAI, and percent change from baseline were computed in the 3 treatment groups. The proportion of patients with high/moderate activity/severity at Week 24 versus low activity/severity/remission, as well as correlations of the 3 indices at Week 24, were calculated. Statistical significance for percent change was analyzed using Wilcoxon tests, after imputation of missing values using modified last observation carried forward (mLOCF); low activity/severity/remission was compared between groups using a logistic model, adjusting for region and baseline joint erosion status after imputation of missing values using non-responder imputation.
Results Improvement from baseline to Week 24 ranged from 19.2% to 37.0% in placebo patients, 40.0% to 65.9% in baricitinib-treated patients, and 37.6% to 60.9% in adalimumab-treated patients (Table), least in DAS28-ESR, intermediate in RAPID3-like, and highest in CDAI. Changes according to RAPID3-like, DAS28-ESR and CDAI were similar in the 3 treatment groups; baricitinib and adalimumab were superior to placebo according to all indices, and baricitinib was superior to adalimumab according to RAPID3-like and CDAI (Table). Correlations of RAPID3-like with DAS28-ESR and CDAI ranged from r=0.61 to 0.75 and for DAS28-ESR with CDAI from r=0.86 to 0.91 (all p<0.001). The proportion of patients with low activity/severity/remission at Week 24 ranged from 9.6% to 19.7% in placebo patients, 31.6% to 49.9% in baricitinib-treated patients, and 33.6% to 47.6% in adalimumab-treated patients. RAPID3-like results were intermediate between DAS28-ESR and CDAI.
Conclusions RAPID3-like documented greater efficacy of baricitinib versus adalimumab and placebo in the RA-BEAM trial, with results in similar ranges to DAS28-ESR and CDAI. RAPID3 is feasible to provide quantitative, standard medical history data; almost of the time and effort is by the patient rather than a health professional, assuring quantitative data in the infrastructure of usual clinical care.
Disclosure of Interest T. Pincus Shareholder of: Health Report Services, B. Zhu Employee of: Eli Lilly and Company, C. Larmore Employee of: Eli Lilly and Company, J. Bradley Employee of: Eli Lilly and Company, N. Patel Employee of: Eli Lilly and Company, C. Gaich Employee of: Eli Lilly and Company, A. Koch Employee of: Eli Lilly and Company