In three prospectively designed studies, first saliva (SA) samples were collected simultaneously with serum (SE) from 86 children who received slow-release theophylline (Th) b.i.d. for treatment of bronchial asthma. SA was collected with a standardized sampling procedure. Regression analysis revealed a close correlation between SE and SA-Th levels (r = 0.959; SE = 1.87 x SA + 0.05) and yielded a "therapeutic SA-Th range" of 5.8 to 10.7 mg/L. In 16 children the diurnal variation was then assessed by measuring 6 SA-Th levels within 24 hr, using the same sampling method. At 0, 9, and 24 hr SE-Th levels were also measured. The levels of SE-Th and SA-Th were again closely correlated (0.914 less than or equal to r less than or equal to 0.949) and both remained constant over the 24-hr period. In the third prospective trial only SA levels were used to monitor Th therapy in 50 asthmatic children. Seven days after starting therapy SA-Th levels were measured and then the Th dose was increased weekly until SA-Th levels above 5.8 mg/L were reached. The following day SE and SA samples were collected simultaneously and their Th levels compared. In 36 of the 47 children who completed the study therapeutic SE-Th levels (greater than or equal to 10 and less than or equal to 20 mg/L) were measured; in one patient it was greater than or equal to 20 mg/L, in 6 it was between 8.0 and 10 mg/L, and only 4 had SE levels less than 8.0 mg/L.(ABSTRACT TRUNCATED AT 250 WORDS)