Impact of Sampling Period on Population Pharmacokinetic Analysis of Antibiotics: Why do You Take Blood Samples Following the Fourth Dose?

Abstract

To date, many population pharmacokinetic models of antibiotics have been developed using blood sampling data after the fourth or fifth dose, which represents steady-state levels. However, if a model developed using blood sampled after the first dose is equivalent to that using blood sampled after the fourth dose, it would be advantageous to utilize the former. The aim of this study was to investigate the effect of blood sampling after the first and/or fourth drug administration on the accuracy and precision of parameter estimates. A previously reported robust, two-compartment model of vancomycin was used for simulation to evaluate the performances of the parameter estimates. The parameter estimation performances were assessed using relative bias and relative root mean square error. Performance was investigated in 72 scenarios consisting of a combination of two blood sampling periods (the first and fourth dose), two total clearances, three infusion times, and four sample sizes. The population pharmacokinetic models from data collected at the first dose and those collected at the fourth dose produced parameter estimates that were similar in accuracy and precision. This study will contribute to increasing the efficiency and simplicity of antibiotic pharmacokinetic studies.

Predicting Antibiotic Effect of Vancomycin Using Pharmacokinetic/Pharmacodynamic Modeling and Simulation: Dense Sampling versus Sparse Sampling

“…However, as shown in Figure 2 and Figure 3 , even if it is generated within the two SDs, when the number of subjects is small, it is highly likely that the parameter estimate is different from the value used for simulation. The accuracy and precision of the population PK/PD parameter estimates are affected by various factors, including the sample size of patients, the number of measurements per patient, sparse or intensive sampling design, the degree of the BSV, and the magnitude of the within-subject variability (WSV) [ 5 , 26 ]. For a one-compartment model with first-order absorption, the sample size required to estimate the absorption rate within a precision level of 20% for a 95% confidence interval and a power of 90% was 30 for a sparse three-sampling scheme and 20 for a dense six-sampling scheme [ 27 ].…”