Evaluation of theophylline pharmacokinetics in a pediatric population using mixed effects models.

Routine clinical pharmacokinetic data collected prospectively from pediatric patients receiving theophylline were analyzed using the NONMEM (nonlinear mixed effects model) digital computer program. A total of 314 measured serum theophylline concentrations (STCs) were obtained from 84 hospitalized patients ranging in age from 4 months to 15.2 years with the majority of patients between the ages of 1 and 8 years. Fifty-six percent were male. The race/ethnicity distribution was 71.4% Latin, 15.5% black, 11.9% Caucasian, and 1.2% (one subject) Pakistani. Of the total number of observed STCs, 16.2% reflected some degree of outpatient dosing. The pharmacokinetic model used was a one-compartment open model with either zero-order or first-order absorption and first-order elimination. Age was the most important determinant of theophylline clearance (Cl); weight was inferior to age and did not statistically improve the model (p greater than 0.005) when combined with age. Total Cl increased by 10%/year over the age range of 1 to 15 years of age. Black race and male gender were associated with higher Cl values: for a given age, Cl was 34% higher for blacks than the reference population composed of the remaining patients, and Cl for males was 25% higher than that for females. The volume of distribution (Vd) for the population was estimated to be 0.62 L/kg. The interindividual variability in Cl and Vd expressed as coefficients of variation were 19 and 28%, respectively. The residual intraindividual error variance corresponded to a standard deviation of 2.8 micrograms/ml. The STCs that represented some degree of outpatient dosing were 21% lower than those reflecting only inpatient dosing. Alternate models that include weight as a determinant of theophylline clearance are also provided. The NONMEM method of determining population pharmacokinetics is well suited to the pediatric population since it does not require a large number of STCs per patient. In this study a mean of only 3.7 STCs per patient were utilized to provide information which should prove useful in the design and adjustment of theophylline dosage regimens in children.