PBPK model shows why plasma EC90 multiples vary for brain-targeted antivirals
- Antiviral dosing often targets a plasma trough above a fixed multiple of in vitro EC90, yet that multiplier is set empirically and ignores poor penetration into sites such as the brain.
- Authors built a rat nirmatrelvir population PK model, fed it into the LeiCNS PBPK platform, and checked it against rat cerebrospinal fluid levels.
- Simulations indicated that plasma unbound fraction and unbound brain-to-plasma partitioning chiefly determine the required multiplier, so a generic cutoff may under-dose CNS targets.
💡 Points
A single EC90 multiple across antivirals is a plasma-centric shortcut, and for brain targets it hides a Kp,uu problem: if unbound brain-to-plasma partitioning is well below 1 (P-gp efflux is the usual suspect for nirmatrelvir-like peptidomimetics), a plasma trough that clears the cutoff can still leave the CNS under-exposed. Validation was rat CSF only, an imperfect proxy for brain parenchyma; human translation remains unproven.