Pharmacokinetic-pharmacodynamic modelling in a clinical pilot study of rituximab in multiple sclerosis: Towards personalized dosing interval.

Aims: Rituximab (RTX) is used off-label for relapsing-remitting multiple sclerosis, although dosing regimens vary. Observational data suggest the standard 6-month interval may be extended and individualized. We aimed to develop a pharmacokinetic-pharmacodynamic (PKPD) model to describe patient-specific RTX concentrations and CD19+ lymphocyte counts.

Methods: Thirteen treatment-naïve patients initiated RTX treatment with 1000 mg intravenously. Blood samples were collected at 6 time points during the following 6 months. RTX concentrations and CD19+ lymphocyte counts were used to compare different PKPD using the nonlinear mixed-effects software Monolix.

Results: Initial pharmacokinetics of RTX could be described by a 1-compartment model with nonlinear target-mediated elimination of rituximab. Introduction of CD19+ lymphocyte counts as a pharmacodynamic marker resulted in comparable performance of a 1-compartment and a 2-compartment model and can provide descriptions of RTX concentrations and CD19+ lymphocyte counts in individual patients. In both models, nonspecific clearance (CL) contributes approximately 10 times more to the overall elimination than target-mediated clearance (k_deg,1; CL [1/day] = 0.075 ± 0.035 [median 0.061], k_deg,1 [mm^-3/day] = 0.125 ± 0.311 [median 0.007] and CL [1/day] = 0.036 ± 0.027 [median 0.026], k_deg,1 [mm^-3/day] = 0.004 ± 0.002 [median 0.003]).

Conclusions: The PKPD models were able to describe the data both in patients exhibiting enduring CD19+ lymphocyte depletion and in patients exhibiting signs of early CD19+ lymphocyte repopulation. Additional data are required to validate and advance models for prediction of repopulation dynamics and, eventually, individualized RTX dosing.