Rapid prediction of steady-state serum theophylline concentration in patients treated with intravenous aminophylline.

In 15 acutely ill asthmatics the steady-state serum theophylline concentration was predicted by the method of Chiou et al. using two serum concentration measurements obtained 1 and 5 h after starting a continuous infusion of aminophylline. Two theophylline assays with different precision characteristics were compared. With a precise HPLC-assay the prediction was excellent: prediction error (predicted minus measured concentration) = -0.22 +/- 1.97 mg/l (mean +/- SD); r = 0.922. When the theophylline concentration was determined by a rapid enzyme immunoassay of lower precision, but convenient for clinical use, the prediction was less accurate (prediction error = 0.58 +/- 3.88, r = 0.852). However, it was still clearly superior to dosing recommendations based on the population average of theophylline clearance, even after taking into consideration the effect of smoking, congestive heart failure and cirrhosis (prediction error = 3.62 +/- 13.36, r = 0.560). As employed in this study, the method may be useful in helping the physician to choose the optimal dose in severely ill asthmatics.