Purpose: Rituximab (RTX) is a chimeric anti-CD20 monoclonal antibody approved for the treatment of non-Hodgins lymphoma and used off-label to reduce urinary protein excretion in glomerular kidney diseases including membranous nephropathy (MN). The pharmacokinetics of therapeutic proteins, including RTX, is not well described in patients with glomerular kidney diseases. The study aim was to evaluate RTX population pharmacokinetics and pharmacodynamics in patients with MN and nephrotic syndrome.
Methods: Patients with MN (N = 20) received 4 weekly IV infusions of RTX (375 mg/m2) over one month, with a repeat of the identical treatment at 6 months. Patient baseline characteristics were: gender (17M/3F), age (49 ± 13 y), BSA (2.2 ± 0.2 m2), urinary protein excretion (11.3 ± 4.1 g/d), creatinine clearance (72 ± 33 mL/min). The influence of covariates (age, gender, BSA, height, weight) on RTX pharmacokinetics was also assessed. The model selection was based on lowest - 2x log likelihood value using different error models (additive error, multiplicative error), % CVs, and the fit of residual plots. The optimal pharmacokinetic model was selected based on minimizing the objective function (OFV), defined as - 2x log likelihood and provides the overall summary of how closely the model predictions match the data. Population parameters were defined as typical value (θ; V1 = 1.295 L, V2 = 7.1639 L, Cl1 = 0.02030 L/h, Cl2 = 0.916 L/h), random/interindividual variability (Ω; V1 = 0.012 L, V2 = 0.079 L, CL1 = 0.026 L/h, CL2 0.0032 L/h). Boot strapping was performed to verify the PK model. The pharmacodynamic outcome of interest was the absolute urinary protein excretion measured over 24 months. Pharmacodynamic data was described using an indirect inhibitory response model. Population pharmacokinetics/pharmacodynamics was performed using Phoenix V.8.0.
Results: A two-compartment pharmacokinetic model with multiplicative error best described the plasma concentration data. The typical values (θ) and % CV for final model parameters were V1 (1.3 L, 19.28%), V2 (7.2L, 12.04%), CL1 (0.02 L/h, 13.04%) and CL2 (0.92 L/h, 10.18%). Age was statistically significant for affecting clearance (CL1) , with increasing age resulting in CL1 decrease. None of the other covariates significantly impacted the pharmacokinetic parameters. An indirect response pharmacodynamic model best described the absolute decrease in urinary protein excretion.
Conclusion: RTX population pharmacokinetics in MN patients was described by a two-compartment model and pharmacodynamics was represented by an indirect response model. The model will assist in targeting plasma concentrations in RTX treatment for patients with MN and nephrotic syndrome.
Mustafa Ibrahim– Graduate Student, University of Colorado Denver, Aurora
Wisit Cheungapasitpornporn– Physician, Mayo Clinic, Rochester, Minnesota
Fernando Fervenza– Physician, Mayo Clinic
Melanie Joy– Associate Professor, University of Colorado, Colorado