We thank Dr. Tekstra and colleagues for their interest in our study demonstrating that the presence of autoantibodies and increased serum IgG levels independently predict the response to rituximab in rheumatoid arthritis. The objective of this work was not to monitor serum B cell biomarkers after rituximab administration, but to identify baseline predictive markers of response to the drug. Since there are now 9 biologic agents available for the management of RA, the identification of predictive factors before treatment institution is particularly challenging with regard to the goal of practicing personalized medicine to appropriately tailor the treatment to the individual patient (1).
We addressed the question of baseline pretreatment serum free light chain levels as a possible predictive marker of response to rituximab, but, consistent with the findings of Tekstra and colleagues' group, neither the absolute value of serum free light chains nor above-normal serum free light chain levels before rituximab treatment was associated with clinical response 24 weeks after rituximab administration (2).
However, we agree that serum free light chain levels in RA have been found by several groups, including ours, to be an accurate and easily assessed surrogate biomarker of disease activity (3, 4). Tekstra et al's demonstration that serum free light chain levels are associated with clinical response to rituximab (3) provides additional data in support of this. But it does not mean that the association is linked to the mechanism of action of rituximab and, as free light chains are correlated with disease activity, exactly the same finding might be observed with any effective treatment of RA.
This discussion clearly illustrates that, as we are reminded by Tekstra et al, it is important to differentiate absolute values of baseline pretreatment biomarkers that are truly predictive of efficacy of a treatment from changes in biomarker levels after treatment, which may or may not be specific to the treatment and, more importantly, may not be more accurate than classic clinical indexes for determining response to a drug.