Population pharmacokinetics of caffeine in premature neonates

Abstract

In this study, which involved on average only two serum concentrations of caffeine per patient, the use of NONMEM gave us significant and consistent information about the pharmacokinetic profile of caffeine when compared with available bibliographic information. Additionally, parenteral nutrition and low gestational age (< or = 28 weeks) may even come to be considered as risk factors, and their presence may serve as an indicator of the need for periodic monitoring of caffeine concentrations in premature infants.

Clinical Pharmacokinetics of Caffeine in Korean Preterm Infants with Apnea of Prematurity

“…The mean clearance estimate was 1.8 mL/h/kg, which was lower than the average clearance values (7.9-8.9 mL/h/kg) reported by other studies 20,25,26) . The estimate of the population mean volume of distribution of caffeine (748 mL/kg) is similar to values reported by other studies (911 mL/kg) 2,27) .…”

“…The half-life of caffeine in preterm neonates (<33 weeks GA) was 87 hours, compared to 72-96 hours in the term neonate group 20) . The elimination of caffeine was severely suppressed in premature neonates but increased nonlinearly after birth up to 6 weeks of age, and reached adult values at approximately 60 weeks postmenstrual age 2) .…”

“…In a study of neonates with GA of 27.6 weeks, groups that maintained caffeine dosing at 5 and 20 mg/kg had a mean serum caffeine concentration of 14.7 µg/mL (4.8-25.1 µg/mL) and 47.4 µg/mL (18.9-79.8 µg/mL), respectively 2) . In another study, the mean serum caffeine levels of 75 preterm neonates were 11.8 µg/ mL (4.75-26.1 µg/mL), with less than 4% >20 µg/mL 20) . In this and half-life (22.6 h).…”

Sequential Updating of a New Dynamic Pharmacokinetic Model for Caffeine in Premature Neonates

“…The developed PK model allows the predictions of caffeine time-course in a new patient for whom only easily collected covariates (duration of gestation and body mass at birth), and possibly, a few caffeine concentration data would be available. For this, we need to take advantage of all available information: data extracted from medical files, pediatricians' experience and published data on neonate physiology and caffeine population pharmacokinetics [1][2][3] . In this context, a Bayesian approach is efficient, as it allows an easy aggregation of such widely different and uncertain information sources [15][16][17] and naturally allows sequential learning.…”

“…As caffeine elimination is widely variable in infant populations [1][2][3] , a more efficient PK model need to integrate new information on the treated patient. We show here how caffeine concentration predictions of the treated patient can be updated by running a particle algorithm using recorded body mass and/or caffeine concentration data.…”