• T Martín-Jiménez, M G Papich, and J E Riviere.
• Department of Anatomy, Physiological Sciences and Radiology, College of Veterinary Medicine, North Carolina State University, Raleigh 27606, USA.
• Am. J. Vet. Res. 1998 Dec 1; 59 (12): 1589-98.

ObjectiveTo develop and validate a population pharmacokinetic model for gentamicin in horses, using retrospective clinical data.Animals62 horses that had been treated IV with multiple doses of gentamicin at our veterinary teaching hospital between 1987 and 1996. Procedure-46 horses were assigned to the study group, and 16 to the validation group. Detailed history of dosage, sample collection times, and selected pathophysiologic variables were recorded for each patient. Samples were analyzed by use of a fluorescence polarization immunoassay method. Pharmacostatistical analysis was conducted, using computer software. The predictive model correlates pharmacokinetic parameters to concomitant pathophysiologic variables and estimates the inter- and intraindividual variability in disposition.ResultsA two-compartment model best described the data. Clearance (CI) was linearly correlated to body weight and serum creatinine concentration. Volume of the central compartment (Vd(c)) was linearly related to body weight. Interindividual coefficients of variability for CI and Vd(c) were 24 and 16%, respectively. The residual variability (intraindividual) was 13%; mean prediction error percent (bias) was 2%; and mean absolute prediction error percent (precision) was 29%.ConclusionsPopulation pharmacokinetic analysis allows study of the basic features of gentamicin disposition in horses with sparse data per individual. A considerable proportion of the pharmacokinetic variability of gentamicin in our study population was explained by differences in body weight and serum creatinine concentration.Clinical RelevancePopulation pharmacokinetics can be used to design first-dosage regimens according to the clinical characteristics of individual animals. Population pharmacokinetic models could also be included in Bayesian forecasting strategies to improve plasma concentration predictions in individual patients.

26 keywords...

hide…