Conversion from intravenous aminophylline to sustained-release theophylline: computer simulation versus in vivo...

The accuracy of computer-predicted theophylline serum concentrations compared with actual serum concentrations in patients whose drug therapy was being converted from aminophylline infusions to an oral sustained-release theophylline product (Theo-Dur) was measured. The SIMKIN computer program was used to simulate theophylline serum-concentration curves in nine patients receiving an initial dose of sustained-release theophylline at the time their maintenance aminophylline infusions were discontinued. The sustained-release theophylline doses were calculated using individual patient theophylline clearance values to produce mean theophylline serum concentrations in the normal therapeutic range. Pharmacokinetic variables used in the SIMKIN program were derived from individual patients' serum concentrations and average literature values. Theophylline serum concentrations were measured before the initial dose of the sustained-release product and periodically for 12 hours. Mean SIMKIN-predicted serum theophylline concentrations were within 10% of actual measurements 39% of the time, but data from individual patients varied considerably. Two patients had prolonged absorption lag times that could not be explained. Computer-simulation programs using population-based pharmacokinetic variables to predict theophylline serum concentrations must be tested against in vivo measurements to verify their accuracy.