Comparison of area under the curve for vancomycin from one- and two-compartment models using sparse data

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Therapeutic Drug Monitoring
• Mathematical Modeling in Medicine

Background:

• Vancomycin pharmacokinetics are often described using one- or two-compartment models.
• One-compartment models are frequently used to estimate the area under the curve (AUC) from sparse data during therapeutic drug monitoring.
• The accuracy of AUC estimates from one-compartment models with limited data is uncertain.

Purpose of the Study:

• To compare the accuracy of AUC estimates derived from one- and two-compartment models using sparse vancomycin data.
• To assess the reliability of AUC estimates generated from trough-only data.
• To evaluate the clinical applicability of different modeling approaches for vancomycin AUC prediction.

Main Methods:

• Simulated vancomycin concentration-time data from a robust model for 100 patients.
• Created sparse datasets: trough-only and peak-trough.
• Developed one- and two-compartment models using NONMEM software.
• Calculated 24-hour AUCs (AUC0-24, AUC24-48, AUCavg) using linear trapezoidal formula.
• Assessed AUC deviations from a reference AUC (AUCref).

Main Results:

• One-compartment models (peak-trough and trough-only data) adequately represented the true AUC without statistically significant differences.
• Two-compartment models with peak-trough data also yielded similar AUC estimates to the reference.
• Two-compartment models using only trough-only data showed significant deviations (15.92%-25.16%) from the reference AUC.

Conclusions:

• Both one- and two-compartment models using sparse peak-trough vancomycin data are reliable for AUC prediction.
• One-compartment models utilizing trough-only data provide useful AUC estimates when rich data is unavailable.
• The use of two-compartment models with trough-only vancomycin data is not recommended due to significant inaccuracies.