A population approach to the forecasting of amikacin plasma and urinary levels using a prescribed dosage regimen

We retrospectively analyzed amikacin pharmacokinetics in 19 critically ill patients who received amikacin intravenously. Fourteen subjects (577 serum amikacin concentrations, 167 urine measurements) were studied to obtain dat a for population modeling, while 5 patients (267 serum amikacin concentrati ons, 68 urine measurements) were studied for the assessment of predictive p erformance. The population analysis was performed using serum and urine ami kacin measurements; the renal clearance of amikacin was expressed as a func tion of creatinine clearance. A two-compartment model was fitted to the pop ulation data by using NONMEM. The population characteristics of the pharmac okinetic parameters (fixed and random effects) were estimated using the FOC E method. The population pharmacokinetic parameters with the interindividua l variability (CV%) were as follows: slope (0.254, 126%) and intercept (3 l /h, 59.6%) of the linear model which relate the amikacin renal clearance to the creatinine clearance, initial volume of distribution (17.1 1, 22.2%), intercompartment clearance (5.22 lih, 104%), steady state volume of distrib ution (55.2 1, 64.1%) and urinary elimination (67.5%, 36.3%). The Bayesian approach developed in this study accurately predicts amikacin concentration s in serum and urine and allows for the estimation of amikacin pharmacokine tic parameters, minimizing the risk of bias in the prediction.