Seizures, often with fatal outcome, are a manifestation of pronounced theophylline intoxication. The purpose of this investigation was to characterize the relationship between theophylline concentrations and theophylline-induced convulsions and to develop an animal model suitable for exploring conditions that might predispose theophylline-treated individuals to seizures. Female Lewis rats (approximately 170 g) received an i.v. infusion of theophylline (as aminophylline) at one of three different rates (1.03-5.1 mg/min/rat) until the animals exhibited a maximal seizure (which occurred after 11 +/- 1 to 42 +/- 3 min of infusion). The total dose, the serum concentration (both total and unbound drug) and the brain concentration of theophylline at onset of seizures increased with increasing infusion rate. The theophylline concentration in cerebrospinal fluid at onset of seizures (mean +/- S.D., 232 +/- 17 mg/l, n = 41) was not affected by the infusion rate. The theophylline metabolites 1-methyluric acid and 1,3-dimethyluric acid were found in serum but at very much lower concentrations than those of theophylline. 1-Methylxanthine and caffeine were not detected in any serum sample, 3-methylxanthine was present in low concentrations in only some serum samples and 1-methyluric acid and 3-methylxanthine were found in the brain in low concentrations (less than 10 mg/kg). Theophylline metabolites were not detected in cerebrospinal fluid. Direct i.v. infusion of either 1-methyluric acid, 1,3-dimethyluric acid or 3-methylxanthine did not produce seizures despite high concentrations in serum. Ethylenediamine infusions also did not cause seizures.(ABSTRACT TRUNCATED AT 250 WORDS)