Nonlinear regression program to individualize oral theophylline dosage
Bauer, L.A.; Black, D.J.; Koup, J.R.
Clinical Pharmacy 3(3): 288-290
1984
ISSN/ISBN: 0278-2677 PMID: 6547379 Document Number: 224753
Theophylline pharmacokinetic values were estimated by two similar computer-assisted methods, and the predictive ability of these methods was evaluated. In each of two study phases, baseline serum theophylline concentrations were obtained in 20 adults receiving oral theophylline for chronic obstructive pulmonary disease; the subjects received no theophylline for 48 hours before start of each phase. In phase 1 all patients received theophylline 400 mg followed by 300 mg every six hours for three doses, and serum theophylline concentrations were measured at 0-6 hours and at 24 hours. In phase 2 patients received individualized theophylline doses calculated from phase 1 data. Predictive method 1 used measured serum theophylline concentrations at 2, 6, and 24 hours to estimate pharmacokinetic values, and method 2 used concentrations at 6 and 24 hours. The computer-predicted values were used to calculate 24-hour serum theophylline concentrations for individualized regimens in phase 2; these were compared with the measured concentration. Pharmacokinetic values predicted by both methods were compared with values based on data obtained after the loading dose in phase 1. Both methods were unbiased when predicting clearance, half-life, and volume of distribution. Estimates of half-life and volume of distribution were significantly more precise with method 1. Both methods underpredicted serum theophylline concentration at 24 hours. The greater number of serum concentrations in method 1 allowed more precise prediction of t1/2 and V; in instances where precision is not critical, method 2 is an acceptable alternative.