The population pharmacokinetics and pharmacodynamics of theophylline in neonates with apnea of prematurity

Date of Completion

January 2001


Health Sciences, Pharmacy




The population pharmacokinetics (PK) and pharmacodynamics (PD) of theophylline were investigated in a prospective and retrospective study of 97 premature neonates with apnea (postnatal age [PNA] 0–18.3 weeks, gestational age [GA] 24–33 weeks, daily body weight [DWT] 0.4–3.5 kg). PK and PD data were analyzed with nonlinear mixed effects models. ^ Concentrations of theophylline (n = 451) and its metabolite, caffeine (n = 54), were determined in serum and plasma, respectively, following multiple doses of orally administered theophylline and intravenously administered aminophylline. The disposition of theophylline and caffeine was well described by a parent-metabolite PK model comprising a one-compartment model for each methylxanthine. The model was parameterized as follows: theophylline and caffeine distributive volume (V2 and V3, respectively), theophylline clearance via methylation to caffeine (CL23), theophylline clearance via all other routes (CL20), and caffeine clearance (CL30). The oral absorption of theophylline was modeled as a first order process with rate constant, KA, and bioavailability, F. Parameters characterizing interoccasional and residual variability in the PK of theophylline and caffeine were estimated. Maturational, demographic, and pathophysiologic covariates were incorporated into the model to explain variability and improve predictive performance. Explanatory covariates (p < 0.005) included DWT, PNA, parenteral nutrition (PN, 0 = absent, 1 = present), and loop diuretics (LOOP, 0 = no, 1 = yes). The final PK model was as follows: V2(L) = 1.15 × (DWT/l.221)1.25, CL23 (L/hr) = 0.00211 × (DWT/1.221) 1.29 × (PNA/3.429)−0.322, CL20 (L/hr) = 0.0181 × (DWT/1.221)0.768 × (PNA/3.429) 0.389 × 1.41PN × 0.824LOOP, KA (hr−1) = 5, F = 0.931, V3 (L) = 0.818, and CL30 (L/hr) = (0.146 × DWT−0.00145 × PNA × 7)/24. Interoccasional variability in CL20 was 31% coefficient of variation (CV). Residual variability was 21% and 19% CV at the mean theophylline and caffeine concentration, respectively. ^ The daily frequency of apneic episodes was modeled according to an overdispersed Poisson distribution with a log-normal distribution of interpatient means. The maturational course of daily episode counts was best described by a biexponential function of PNA and was influenced by GA and comorbidity of hyaline membrane disease. An inhibitory maximum effect model was used to describe the reduction of apneic episode frequency by theophylline. The model was parameterized in terms of Emax, the maximum achievable reduction in episode frequency, and IC50, the average concentration resulting in 50% of E max. Emax and IC50 were 58% and 4.3 μg/mL, respectively. ^