Pharmacokinetic Model-Informed Precision Dosing of Natalizumab in Multiple Sclerosis.

Abstract

Intravenous natalizumab is an effective treatment for relapsing-remitting multiple sclerosis. However, the standard treatment interval of 4 weeks may be excessive for many patients. Personalized interval extension using therapeutic drug monitoring (TDM) can result in adequate drug exposure while reducing hospital visits and healthcare costs. Here, we investigate to which extent TDM-guided personalized dosing can benefit from model-informed precision dosing (MIPD). Individual posterior PK estimates were derived using patient weight and two trough concentrations at the standard dose interval by Bayesian estimation using a newly developed population PK model. MIPD was compared to a previously deployed TDM-guided stratified personalized dosing protocol (SPD) using a decision tree to personalize dosing intervals. Accuracy (mean prediction error) of the predicted dosing intervals was 4.8% versus 24% for model-informed estimates versus decision tree, respectively, when aiming for a 10 μg/mL trough concentration, and 4.8% versus 86% when aiming for 5 μg/mL. Corresponding precision (root mean square error) was 2.3 versus 4.0, and 1.5 versus 5 μg/mL. Finally, we evaluated the feasibility of a MIPD approach to attain a therapeutic trough of 2 μg/mL. Simulating MIPD showed a reduction in the average infusions versus the standard interval by 40%, with an average dose interval of 7 weeks, while maintaining adequate drug exposure. MIPD was concluded to be superior to the conventional TDM-guided personalized dosing approach in terms of enhanced precision in individual dose interval selection, enabling more efficient interval extensions. Simulations supported the clinical deployment of natalizumab MIPD.