Objective. Studies comparing 500 mg rituximab and 1,000 mg rituximab doses in rheumatoid arthritis have yielded conflicting data on clinical outcomes, but in all of these studies a subgroup of patients has had excellent responses at the lower dose. Historically, it was considered that rituximab uniformly depleted B cells at both doses. Using highly sensitive assays, we have shown that B cell depletion is variable and predictive of clinical response. Using the same techniques, we undertook the present study to test the hypothesis that the level of B cell depletion, rather than the rituximab dose, determines clinical response.Methods. Nineteen patients were treated with two 500-mg infusions of rituximab, and 61 patients were treated with two 1,000-mg infusions of rituximab. Highly sensitive flow cytometry was performed at 0, 2, 6, 14, and 26 weeks. European League Against Rheumatism (EULAR) response rates at 6 months were compared between patients with and those without complete depletion at each dose.Results. The median B cell count was numerically higher at all time points following therapy in the 500 mg rituximab group. Twenty-five percent of patients in the 500 mg rituximab group had complete depletion at 2 weeks, compared with 49% of those in the 1,000 mg rituximab group. Complete depletion at 2 weeks after treatment with 500 mg rituximab was associated with lower baseline preplasma cell counts (P ؍ 0.047). Most patients responded after either dose, but response was related to B cell depletion. Notably, in the 500 mg rituximab group all patients with complete depletion had a EULAR good response (P ؍ 0.011).Conclusion. This pilot study suggests that the degree of B cell depletion, rather than the dose of rituximab, determines clinical response. It may be possible to predict which patients will respond to lowerdose rituximab, and this may allow more cost-effective treatment.Rituximab in combination with methotrexate is licensed for the treatment of rheumatoid arthritis (RA) after failure of Ն1 anti-tumor necrosis factor (anti-TNF) agents, but data on its optimal use are currently limited. Initial dose selection for rituximab in RA was based on studies in non-Hodgkin's lymphoma (NHL). Major factors influencing serum rituximab concentration and the degree of depletion of tumor cells identified in NHL studies include initial tumor load, organ involvement, and the presence of soluble CD20 shed by lymphoma cells, rather than merely body surface area (1). Equivalent doses of rituximab in combination with cyclophosphamide were used effectively in early studies for the depletion of nonmalignant B cells in the treatment of RA, although the same pharmacokinetic factors do not apply in these patients (2). Subsequent to these small open-label, variable-dose series, treatment with two 500-mg infusions of rituximab was compared with