To assess whether the withdrawal of salmeterol treatment for 3 days (72 h) can restore its bronchoprotective ability on specific bronchial provocative test (sBPT) with allergen, which was completely lost after 1 week of regular treatment with salmeterol. Single-blind design. We investigated 10 nonsmoking subjects (8 men and 2 women; mean +/- SD age, 24 +/- 8 years) with mild intermittent allergic asthma in the stable phase of the disease, who were never previously treated with regular beta(2)-agonists. Subjects with a previous positive early airway response (EAR) to a screening allergen challenge were considered. They underwent sBPT with allergen after a single dose of inhaled salmeterol, 50 microg (T(1)), and then underwent sBPT after 1 week of regular treatment with inhaled salmeterol, 50 microg bid (T(2)); after that, they continued inhaled salmeterol treatment for 4 days, and then changed to inhaled salmeterol with placebo (two puffs bid) for 3 days (72 h) and underwent sBPT with allergen after a single dose of salmeterol, 50 microg (T(3)). EAR to allergen (DeltaFEV(1) > or = 20% with respect to postdiluent value) was completely abolished by a single dose of salmeterol (T(1); protection index [PI] > or = 50% in all subjects), but it was still present after 1 week of regular treatment with salmeterol (T(2); PI < 50% in all subjects). The maximum FEV(1) percentage fall during sBPT with allergen was significantly lower after withdrawal of regular inhaled salmeterol (T(3)) than after regular treatment with salmeterol (T(2)) (mean, 23% vs 29.5%; range, 4 to 41% vs 18 to 49%, respectively; p < 0.05); a similar result was obtained considering the PI of salmeterol on sBPT with allergen (mean, 44% vs 20%; range, 2 to 86% vs - 11 to 49%, respectively; p < 0.05). However, the maximum FEV(1) percentage fall and PI were significantly different in T(3) than after T(1), and only 4 of 10 patients showed in T(3) a PI > or = 50%. The bronchoprotective effect of salmeterol on allergen-induced EAR, completely lost after 1 week of regular treatment with salmeterol, may be partially restored by the withdrawal of salmeterol therapy for 3 days (72 h). However, this withdrawal time period is not sufficient to recover the baseline bronchoprotective efficacy of the first dose of salmeterol.