Article Text
Abstract
Background RF-CCP- RA patients have inferior response to rituximab (RTX) but a subset respond as well as seropositive RA. RF-CCP- RA may have alternative markers of B cell dysfunction such as ANA or hypergammaglobulinaemia.
Objectives To identify clinical predictors of response to rituximab in RF-CCP- RA as a means to select patients suitable for RTX.
Methods Data from 7 clinical trials were pooled. Binary logistic regression models were used to test whether baseline factors affected odds of ACR20 at 24 weeks in patients negative for RF and CCP patients, controlling for trial. Non-response was imputed for patients who withdrew or were retreated prior to 24 weeks. Interaction terms were added to identify potential treatment effect modifiers (p<0.2 criterion).
Results 396 patients had data available (mean age 52.3, 82% female, mean baseline DAS28CRP 5.5) of whom 155 (39%) achieved ACR20 at 24 weeks. Odds ratios are presented in Table 1. In addition to expected markers of therapy response, we found a relationship between higher IgG and increased odds of response with an interaction with rituximab dose. Overall, 2x1000mg was more effective than placebo, and the level of IgG did not influence the odds of response. In contrast, 2x500mg was increasingly effective compared to placebo with increasing IgG titre. Data on ANA status were more limited. We tested a model including ANA in a set of 301 patients. There was a trend to greater response in ANA positive patients (OR 1.49, CI 0.87 – 2.58). Controlling for ANA the interaction with IgG remained for rituximab 500mg (OR 1.1, CI 0.96 – 1.34)
Conclusions The lower dose RTX appeared less effective in this group of patients overall. A subset of RF-CCP- patients with alternative markers of B cell mediated inflammation have more scope to respond to RTX. However, this effect is less relevant when the licensed dose is used. The efficacy of RTX in RF-CCP-ANA+ arthritis patients supports its efficacy in connective tissue diseases.
Acknowledgement This study was supported by a research grant from Roche.
Disclosure of Interest None declared