Population Pharmacokinetics of Oral and Intravenous Caffeine in Extremely Premature Neonates

Background: Apnea of prematurity (AP) is a potentially serious condition in >90% of extremely premature neonates. It is characterized by the cessation of airways movements for 15-20 seconds or more and, if left untreated, can have serious sequelae. Caffeine (citrate salt) is now the drug of choice for helping wean premature infants from mechanical ventilation, but little is known of the pharmacokinetic disposition and variability of caffeine in extremely premature neonates.

Methods:  Serum caffeine data (431 samples, median: 4 per subject) were obtained from 110 (52 male) newborn infants having an average birth weight (BW) of 1009 g, gestation age (GA) of 27.5 wk, postnatal age (PNA) of 11.5 d, who were receiving either 5 mg/kg/24 h or 20 mg/kg/24 h of caffeine citrate as part of a dose-ranging clinical study conducted at the Mater Mothers’ Hospital, Brisbane. Of the total 600 doses given, 149 were by the oral (nasogastric) route which permitted estimation of the rate and extent of caffeine absorption.

A population approach using a 1-compartment model with 1st order absorption was conducted using NONMEM (5.1.1) with FOCE, using an exponential interindividual variance (IIV) model and a combined additive-proportional model to estimate the residual unexplained variance (RUV). The influence of patients’ characteristics (e.g. WT, BW, PNA) were screened for significance (P = 0.01) in nested, multiplicative models. Model stability was assessed by the non-parametric bootstrap (N=200).

Results: The following structural model of clearance (CL) and volume (V) best described the data:

CL (L/h) =  0.171 . (WT ÷ 70) 0.75  .  (PNA ÷ 11.6) 0.385  

V (L)      =   59.0 . (WT ÷ 70) 0.75

The estimated absolute bioavailability (F) was 107%. CL (mL/min/kg) was 12-fold lower, V (mL/kg) was 36% greater, and half-life was 20-fold longer than in adults.

Conclusions: This study established that the elimination of caffeine was severely depressed in extremely premature infants, but increased nonlinearly after birth up to 6 weeks of age. While caffeine was absorbed more slowly (absorption half-life: 30 min) than in adults, the absolute bioavailability was excellent, which has favorable implications for switching between intravenous and nasogastric routes. The interoccasion variability about CL was twice the interindividual variability which, among other factors, indicates that routine serum concentration monitoring of caffeine in these patients is not warranted.

Bruce Charles