Background: Cough variant asthma (CVA) is a phenotype of asthma presenting solely with coughing, characterized by airway hyperresponsiveness, eosinophilic inflammation and a cough response to bronchodilators. Leukotriene receptor antagonists (LTRAs) are antiasthma medications with anti-inflammatory and bronchodilatory properties. Although LTRAs exert antitussive effects in CVA, the mechanisms involved are unknown. Objectives: This study aimed to clarify the antitussive mechanisms of LTRAs in CVA patients. Methods: We prospectively observed the effect of montelukast (10 mg) daily for 4 weeks in 23 consecutive nonsmoking adults with anti-inflammatory treatment-naive CVA. We evaluated, before and after treatment, the cough visual analogue scale (VAS), pulmonary function (spirometry and impulse oscillation), methacholine airway responsiveness, cough receptor sensitivity, expressed by the concentration of capsaicin inducing 2 or more (C2) and 5 or more (C5) coughs, sputum eosinophil counts and levels of inflammatory mediators, including cysteinyl leukotrienes, leukotriene B4, prostaglandin (PG) D2, PGE2, PGF2α and thromboxane B2. We compared the baseline characteristics of the patients based on the symptomatic response to montelukast, defined as a decrease in the cough VAS of >25% (n = 15) or ≤25% (n = 8). Results: Montelukast significantly decreased the cough VAS (p = 0.0008), sputum eosinophil count (p = 0.013) and cough sensitivity (C2: p = 0.007; C5: p = 0.039), whereas pulmonary function, airway responsiveness and sputum mediator levels remained unchanged. Multivariate analysis showed that a better response to montelukast was associated solely with younger age (p = 0.032). Conclusion: The antitussive effect of montelukast in CVA may be attributed to the attenuation of eosinophilic inflammation rather than its bronchodilatory properties.