Population Pharmacokinetic Pharmacodynamic Modeling of Caffeine Using Visual Analogue Scales

Caffeine is a commonly ingested psychoactive substance which affects alertness and cognition. A clinical study was conducted to determine the effect of orally ingested caffeine on visual analogue scale (VAS) responses in healthy, moderate caffeine-consuming volunteers through the use of population pharmacokinetic-pharmacodynamic (PK-PD) modeling. Twelve subjects were recruited for a three-period cross-over study which utilized caffeine containing beverages. Each visit included 8-hour blood plasma and VAS response collection for PK and PD assessment respectively. The VAS used in the study, also called the caffeine analog scale, has been previously validated for caffeine. Population PK-PD modeling was conducted with NONMEM 7.2. Simultaneous and sequential modeling of PK-PD was attempted. Final model selection was based on parameter estimate precision, diagnostic plots, and visual predictive check (VPC) plots. Results showed that a one-compartment open model with first-order absorption and elimination best described the pharmacokinetics of caffeine. Sequential PK-PD modeling was successful and an effect compartment model with linear slope and baseline parameter best described caffeine pharmacodynamics. Diagnostic plots showed no major bias and VPC plots showed agreement between observations and predictions. The model was able to link VAS responses to caffeine concentration in healthy volunteers and may be useful in clinical trial simulations and design.