The relative bioavailability of two slow-release theophylline formulations using bead-filled capsules. Slo-bid and Theo-Dur Sprinkle, was examined in randomized crossover multiple-dose trials using collected saliva. Two groups of seven preschool-age asthmatic children received one of these and Slo-Phyllin Gyrocaps as a reference. Dose-normalized areas under the saliva concentration-time curves were significantly less than the reference for Theo-Dur Sprinkle (p = .02) but not for Slo-bid with the relative bioavailability determinations of 66 +/- 8% (mean +/- SEM) and 109 +/- 5% for Theo-Dur Sprinkle and Slo-bid, respectively. The less complete absorption of Theo-Dur Sprinkle during clinical use in preschool-age children is consistent with the malabsorptive effects of food previously reported for the former but not the latter formulation.