Objective: The study the population pharmacokinetics of caffeine after intravenous administration to premature infants with apnea.
Methods: A prospective, blinded parallel study in which daily caffeine citrate doses of 30, 15, and 3 mg/kg were administered over 7 days by intermittent intravenous infusion. Arterial blood samples (three to six per patient) were assayed for caffeine content by means of HPLC. Population pharmacokinetic modeling was performed with NONMEM.
Results: Clearance (L/hr) = (0.00000399 . current weight [grams]) + (0.000128 . postnatal age [days]). For gestational age > 28 weeks, volume of distribution (L) = (0.000764 . weight [grams] + (0.0468 . postnatal age [days]); for gestational age < or = 28 weeks, volume of distribution (L) = (0.000755 . weight [grams]) + (0.0224. postnatal age [days]). Interpatient variability (coefficient of variation, in percent) was approximately 25% for clearance and approximately 11% for volume of distribution. Intrapatient error (standard deviation) was 3.9 mg/L. There was insignificant bias between observed and model-predicated serum caffeine concentrations in a separate group of 30 infants.
Conclusions: Caffeine was well tolerated at all doses. Clearance was markedly lower and volume of distribution was higher than the values reported previously for term infants and adults. Both parameters were significantly influenced by postnatal age and current body weight, whereas volume of distribution in infants > 28 weeks' gestational age was higher than that in more premature babies. The predictive performance and the clinical application of the derived population models was satisfactorily shown.