Abstract
Background/Objective: Voriconazole exhibits nonlinear pharmacokinetics and wide interindividual variability driven by CYP2C19 phenotype and clinical covariates, necessitating early therapeutic drug monitoring (TDM). This study aimed to assess how the choice of population pharmacokinetic (PopPK) models influences genotype-stratified voriconazole exposure under a standardized adult regimen, and to delineate model-specific implications for clinical prescribing. Methods: Five CYP2C19-informed PopPK models (Yun, Ling, Wang, Dolton, Friberg) were evaluated under one oral dosing scenario with an identical extensive metabolizers (EM)/intermediate metabolizer (IM)/poor metabolizers (PM) cohort; steady-state exposure metrics were compared across models, with sensitivity checks using model-specific cohorts. Results: Yun predicted the highest exposures with the steepest EM-IM-PM gradient, suggesting a need for caution against upper-tail exceedance when genotype effects are pronounced. Ling yielded intermediate exposures with a modest gradient, consistent with adult central tendencies, thus supporting its use for standard adult initial dosing. Wang primarily distinguished between EM and PM, proving useful for lower-bound checks where underexposure risk or limited genotype information is a concern. Friberg (and Dolton) demonstrated lower exposures with limited genotype separation, offering insights when persistent underexposure is suspected. Conclusions: These model-specific patterns indicate that PopPK model choice can influence initial dose-band selection and the timing of early TDM in routine adult care. Ling can serve as a baseline for standard adult initiation, whereas Yun is appropriate for safety-first scenarios when upper-tail risk from strong genotype effects is anticipated; Wang assists when IM data are lacking or when lower-bound checks are needed. Generalizability beyond standardized adult dosing (e.g., special populations) remains limited.