The long-acting beta2-agonist salmeterol inhibits in vitro the release of inflammatory mediators up to 20 h. These mediators are involved in ultrasonically nebulized distilled water (UNDW)-induced bronchoconstriction. We investigated whether salmeterol provides prolonged protection against UNDW provocation and whether this effect was paralleled by its bronchodilator effects. Nineteen asthmatic patients (mean forced expiratory volume in one second (FEV1) 84.8% predicted, mean provocative concentration of histamine producing a 20% decrease in FEV1 0.65 mg x mL(-1)) participated in this randomized, double-blind, placebo-controlled crossover trial. After measuring baseline FEV1, patients inhaled 50 microg salmeterol or placebo by metered-dose inhaler. FEV1 was measured after 20 and 40 min, and UNDW provocations and FEV1 measurements were performed after 10, 20 and 34 h. Compared to placebo, salmeterol caused marked bronchodilatation from 20 min up to 20 h after inhalation. Salmeterol also provided more than 20 h of protection against UNDW provocation (still more than one doubling dose). Protection beyond the period of bronchodilatation did not occur. Eleven subjects had a significant reduction in provocative dose of UNDW causing a 20% fall in FEV1 (PD20,UNDW) values between 10 and 20 h, at a time when there was still persistent bronchodilation. No correlation existed between changes in FEV1 and changes in PD20,UNDW. From the equations of regression lines between FEV1 and corresponding PD20,UNDW values, it was calculated that only approximately 25% of the afforded protection was explained by bronchodilatation. In conclusion, a single dose of salmeterol induces both bronchodilatation and protection independently of this bronchodilation against a physiological bronchoconstrictor stimulus for more than 20 h.