LAUSR

Background Tocilizumab (TCZ) efficacy and safety in rheumatoid arthritis (RA) have been well established by numerous phase 3 and 4 studies and observational studies. Objectives To explore the extent to which disease duration, inflammation, disease burden, and other baseline factors explain variations in outcomes in studies of RA patients treated with TCZ. Methods This was a pooled analysis of methotrexate-inadequate responding (IR)/conventional synthetic disease-modifying antirheumatic drug (csDMARD)-IR patients with RA allocated to TCZ (intravenous or subcutaneous, monotherapy + combination therapy) in phase 3 and 4 studies. End points were change from baseline to week 24 in Clinical Disease Activity Index (CDAI) and quality of life (Health Assessment Questionnaire–Disability Index [HAQ-DI]) and week 24 ACR50 and CDAI remission (≤2.8). Using a combination of clinically informed and mathematically driven variable selection techniques, models (with study included as a random effect to account for intracorrelation of observations within each study) were built to optimize fit and explain outcome variance. Analysis of covariance and logistic regression were used for CDAI/HAQ-DI change from baseline and for ACR50/CDAI remission, respectively. Results The analyses were performed on 5462 patients from 12 studies. Analysis of baseline characteristics (before TCZ administration) revealed that patients with longer disease duration had been exposed to more csDMARDS and had worse HAQ-DI than patients with shorter disease duration. Statistical modeling of clinical outcomes showed that disease duration accounted for <2% of the variation in HAQ-DI and CDAI change from baseline. Baseline CDAI explained 32% of the variation in CDAI change from baseline. Patients with higher baseline CDAI values tended to have greater improvements, likely due to having more “room” for improvement and to a higher risk for regression to the mean resulting from the fact that inclusion criteria for most trials required defined thresholds of disease activity, including joint counts. Baseline HAQ-DI, neither an inclusion criterion itself nor influenced by other inclusion criteria, explained 15% of the variation in HAQ-DI change from baseline. The odds of achieving ACR50 decreased by 9.2% if disease duration was doubled. The odds of achieving CDAI remission decreased by 15% per 5 additional years of disease duration and decreased by 22% per 10 additional score units of CDAI at baseline. Conclusion In this pooled analysis of TCZ-treated RA patients, disease duration explained statistically significant but practically small variations in clinical outcomes. These findings indicate that TCZ treatment outcomes are not heavily influenced by disease duration or other baseline characteristics. Disclosure of Interests Andrea Rubbert-Roth Consultant for: Chugai, Eli Lilly, Roche, and Sanofi, Speakers bureau: AbbVie, Bristol-Myers Squibb, Chugai, Hexal/Novartis, Janssen, Eli Lilly, Merck Sharp & Dohme, Pfizer, Roche, and Sanofi, Daniel Aletaha Grant/research support from: AbbVie, Bristol-Myers Squibb, and MSD, Consultant for: AbbVie, Bristol-Myers Squibb, Eli Lilly, Janssen, Medac, Merck, MSD, Pfizer Inc, Roche, and UCB, Speakers bureau: AbbVie, Bristol-Myers Squibb, Eli Lilly, Janssen, Medac, Merck, MSD, Pfizer Inc, Roche, and UCB, Jenny Devenport Employee of: F. Hoffmann-La Roche, Paris N. Sidiropoulos Shareholder of: Roche, Employee of: Genentech, Yves Luder Shareholder of: F. Hoffmann-La Roche, Employee of: F. Hoffmann-La Roche, Michael Edwardes Consultant for: Roche, Johannes W. G. Jacobs Grant/research support from: Roche, Consultant for: Roche