Allergen-induced increase in airway hyperresponsiveness can be used as a model of airway inflammation for assessing antiasthma pharmacologic agents. Steroids and cromolyn, but not beta-agonists, inhibit this increase; theophylline, recently suggested as having anti-inflammatory effects, has not been evaluated in this model. Six atopic subjects with asthma and with late asthmatic responses (N = 5) and postallergen reduction in a provocative concentration of methacholine causing a 20% fall in FEV1 (PC20) (N = 6) were studied. Sustained-release theophylline (Theo-Dur; Astra Pharmaceuticals Canada, Ltd., Mississauga, Canada), 300 mg, and placebo were administered single-blind twice daily for eight doses up to 1 hour before allergen inhalation; cromolyn sodium, 10 mg, was administered in a single dose 10 minutes before allergen inhalation on another day as a "positive control." Mean theophylline levels were in the low therapeutic range, 57 +/- 17 and 58 +/- 13 mumol/L 1 and 8 hours after the last tablet. The FEV1 was 7% and 9% greater after the seventh and eighth doses of theophylline versus placebo (p less than 0.05). Theophylline also produced a significant (p less than 0.05) twofold increase in methacholine PC20. There was a 40% (p = 0.06) reduction in early asthmatic fall in FEV1 and a 25% (not significant) reduction in late FEV1 fall when theophylline was compared to placebo. Theophylline did not influence the geometric mean allergen-induced fall in methacholine PC20 delta log PC20; this was true individually in five of the six subjects. By contrast, cromolyn sodium inhibited all aspects of the allergen response completely.(ABSTRACT TRUNCATED AT 250 WORDS)