Cough variant asthma (CVA) is a common cause of chronic persistent cough, in which allergic airway inflammation may play a role. Although current guidelines recommend bronchodilators and anti-inflammatory drugs for the treatment, comparison of the efficacy of these medications has not been investigated. This study was designed to evaluate the effectiveness of pranlukast, a leukotriene receptor antagonist, and salmeterol, a long-acting beta₂-adrenergic agonist, in the treatment of CVA. The study was a randomized, controlled, parallel-group, multicenter trial. After a 4-week run-in period, 49 patients with newly diagnosed CVA were assigned to receive oral pranlukast (225 mg, b.i.d.) or inhaled salmeterol (100 μg, b.i.d.) for 4 weeks. Primary outcome measure was cough symptom and secondary outcome measures were pulmonary function and eosinophilic airway inflammation. Treatment with pranlukast and salmeterol each decreased cough symptom scores, where the changes from baseline values were significantly greater in the pranlukast group than in the salmeterol group. Forced expiratory volume in 1 second and peak expiratory flow (PEF) increased in the two treatment groups with the same magnitudes, but significant decreases in diurnal variation of PEF and eosinophil counts and eosinophil cationic protein contents in the peripheral blood and induced sputum were observed only in the pranlukast group. In view of antitussive and anti-inflammatory actions, the leukotriene receptor antagonist pranlukast seems to be more effective than the long-acting beta₂-adrenergic agonist salmeterol in the treatment of CVA.