In order to assess whether treatment with oral slow-release theophylline prevents early (EAR) and/or late (LAR) airway responses to allergen, we treated six asthmatic subjects with either placebo or oral theophylline (350 mg b.i.d.) for one week before specific bronchial provocation test (sBPT) with allergen, in a double-blind, cross-over, placebo-controlled study. All subjects had previously shown both EAR and LAR to allergen sBPT. On the day of sBPT, serum theophylline concentration was > 9 micrograms.ml-1 in all subjects treated with theophylline except one (mean 9.8 +/- 2.6 micrograms.ml-1). The total dose (TD) of inhaled allergen, and the provocative dose of allergen causing a 15% fall in forced expiratory volume in one second (PD15FEV1) during sBPT was mildly but significantly lower during theophylline-sBPT than during placebo-sBPT. Theophylline treatment reduced EAR (maximum FEV1 fall from baseline: 26.3 +/- 6.7 vs 32.2 +/- 9.5% after placebo treatment) but not LAR (19.5 +/- 6.5 vs 26.3 +/- 6.1%). A significant protection of LAR (a reduction in FEV1 fall after theophylline treatment 50% more than after placebo treatment) was observed in only 2 out of 6 subjects. Area under the curve was not significantly different after the two treatments. There was no relationship between serum theophylline levels and the degree of protection on EAR or LAR. The decrease in PD15FEV1 methacholine observed at the end of LAR was similar in the two treatments.(ABSTRACT TRUNCATED AT 250 WORDS)