Ability of three pharmacokinetic equations to predict steady-state serum theophylline concentrations in pediatric...

The pharmacokinetic equations of Chiou, Koup, and Kurland are often used in the pediatric setting to predict steady-state theophylline clearance using non-steady serum theophylline concentrations. However, these equations have not been validated or compared in a pediatric population. We evaluated the ability of these equations to predict steady-state serum theophylline concentrations in 61 children (0.21-14.3 years) who received a continuous intravenous theophylline (0.79 +/- 0.12 mg/kg/h) infusion for a minimum of five half-lives. Theophylline concentrations used in the Kurland equation were obtained 10.8 +/- 4.5 h after initiation of therapy and the time between the two concentrations used in the Chiou and Koup equations was 9.2 +/- 3.9 h. Predicted steady-state theophylline concentration values for the three methods were not different from each other (p = 0.91), nor were they different from the observed steady-state concentration values (p = 0.92). The coefficient of determination for predicted vs. observed steady-state concentrations was statistically significant (p less than 0.001) and was comparable for the three methods. There was no difference in mean bias (p = 0.78), precision (p = 0.82), or % error (p = 0.86) values for the three methods. Regardless of the method used, 75 to 82% of all predicted theophylline concentrations were within 20% of the observed steady-state value. However, on average, all methods underpredicted the clearance and hence overpredicted the serum theophylline concentration. The Kurland method did not predict steady-state concentrations any better in patients who had received theophylline prior to admission.(ABSTRACT TRUNCATED AT 250 WORDS)