To evaluate the dose-effect relationship of a controlled-release theophylline in preschool children, 20 patients with asthma (mean age 4.8 years, range 2 1/2 to 7 years) were given three different dose levels (13.4 +/- 1.4, 18.4 +/- 1.6, and 23.5 +/- 2.0 mg/kg/day, mean +/- SD) at 12-hour intervals for 2 weeks. Subjective variables, peak expiratory flow rate, and co-medications were recorded daily; clinical condition, serum theophylline levels, and lung function measured with a multiple forced oscillation technique were assessed at the end of each period. The morning predose (through) and 4-hour postdose (peak) serum theophylline concentrations increased in an approximately linear fashion with increasing dose. In the majority of patients, dose levels of 20 to 25 mg/kg/day maintained serum concentrations within a clinically effective range, with an acceptable level of fluctuation. However, wide interindividual variations in serum theophylline concentrations were observed, indicating that for optimal treatment individualization of dosage is preferable. Efficacy was related to serum concentration and, less closely, to the dose administered. Symptom scores for night cough, wheeze, and activity showed small improvements between 5 and 10 mg X 1(-1) and marked improvements above 10 mg X 1(-1), whereas lung function values improved in a linear fashion across the serum concentration range. The serum theophylline concentration-response curves varied in an approximately parallel manner between individuals.