Modeling adult COX‐2 cerebrospinal fluid pharmacokinetics to inform pediatric investigation

Abstract

Aim Hysteresis is reported between plasma concentration and analgesic effect from nonsteroidal anti‐inflammatory drugs. It is possible that the temporal delay between plasma and CSF nonsteroidal anti‐inflammatory drugs mirrors this hysteresis. The temporal relationship between plasma and CSF concentrations of COX‐inhibitors (celecoxib, rofecoxib, valdecoxib) has been described. The purpose of this secondary data analysis was to develop a compartmental model for plasma and CSF disposition of these COX‐2 inhibitors. Methods Plasma and CSF concentration‐time profiles and protein binding data in 10 adult volunteers given oral celecoxib 200 mg, valdecoxib 40 mg and rofecoxib 50 mg were available for study. Nonlinear mixed effects models with a single plasma compartment were used to link a single CSF compartment with a transfer factor and an equilibration rate constant (Keq). To enable predictive modeling in pediatrics, celecoxib pharmacokinetics were standardized using allometry. Results Movement of all three unbound plasma COX‐2 drugs into CSF was characterized by a common equilibration half‐time (T1/2keq) of 0.84 h. Influx was faster than efflux and a transfer scaling factor of 2.01 was required to describe conditions at steady‐state. Estimated celecoxib clearance was 49 (95% CI 34–80) L/h/70 kg and the volume of distribution was 346 (95% CI 237–468) L/70 kg. The celecoxib absorption half‐time was 0.35 h with a lag time of 0.62 h. Simulations predicted a 70‐kg adult given oral celecoxib 200 mg with maintenance 100 mg twice daily would have a mean steady‐state total (bound and unbound) plasma concentration of 174 μg L−1 and CSF concentration of 1.1 μg L−1. A child (e.g., 25 kg, typically 7 years) given oral celecoxib 6 mg kg−1 with maintenance of 3 mg kg−1 twice daily would have 282 and 1.7 μg L−1 mean plasma and CSF concentrations, respectively. Conclusions Transfer of unbound COX‐2 inhibitors from plasma to CSF compartment can be described with a delayed effect model using an equilibration rate constant to collapse observed hysteresis. An additional transfer factor was required to account for passage across the blood–brain barrier. Use of a target concentration strategy for dose and consequent plasma (total and unbound) and CSF concentration prediction could be used to inform pediatric clinical studies.