Optimization of antiviral dosing in Herpesviridae encephalitis: a promising approach to improve outcome?

Despite established antiviral therapy for herpes simplex virus, varicella zoster and cytomegalovirus encephalitis, the outcome remains poor.

To assess pharmacokinetic (PK) and pharmacodynamic (PD) data of antiviral drugs in the central nervous system (CNS) to optimize the treatment of Herpesviridae encephalitis.

PUBMED search 1950 to September 2024, terms (1) "encephalitis" and ("HSV" or "VZV" or "CMV") or (2) cerebrospinal and ("(val)acyclovir" or "(val)ganciclovir" or "foscarnet" or "cidofovir").

Antivirals against herpes viruses apparently act in a time-dependent manner. To suppress viral replication, drug concentration in the extracellular space at the site of the infection should be kept above the concentrations active in cell cultures for 24 h/d. Most data reflect delayed drug entry into lumbar cerebrospinal fluid (CSF). Ratios of the areas of the concentration/time curves (AUC) in CSF and serum (AUC_CSF/AUC_S) of acyclovir, ganciclovir and foscarnet are 0.15-0.3 in the absence of meningeal inflammation and increase in severe meningoencephalitis. Elimination half-lives (t_1/2β) are longer in CSF than in plasma. CSF concentrations are rough approximations of drug concentrations in the cerebral extracellular fluid. Lumbar CSF concentrations are usually higher than ventricular or cisternal CSF concentrations tending to overestimate cerebral extracellular fluid concentrations. Provided the availability of adequate measurements in individual CNS compartments, antiviral concentrations and efficacy may be predicted by future PK/PD modelling. Probenecid holds the potential to reduce the efflux of antiviral drugs from the CNS.

The long t_1/2β of antiviral drugs in CSF suggest relatively uniform steady state CSF levels at dosing intervals ≤12h and accumulation after repeated dosing. Probenecid is of unproven utility. To rapidly attain effective concentrations in the infected tissue, physiologically based pharmacokinetic and PK/PD modelling may be helpful. Until reliable PK/PD data are available, doubling the first dose should be considered.