Background and objective: Patients presenting with refractory postinfectious cough may respond to glucocorticosteroids but it is unclear whether airway eosinophilic inflammation exists in those patients. We aimed to determine the airway inflammation and causes of subacute cough following acute upper respiratory tract infection (AURTI). Methods: One hundred and sixteen patients with persistent cough lasting 3-8 weeks after upper respiratory tract infection were evaluated with differential cell count in induced sputum, spirometry and methacholine bronchial challenge testing. Results: In patients with subacute cough, sputum eosinophilia (median 8.5%,3.0-73.0%) was identified in 35 (33.6%) patients, 22 (18.5%) without bronchial hyperresponsiveness (BHR) were diagnosed as nonasthmatic eosinophilic bronchitis (NAEB), 13 (14.3%) of whom with BHR were diagnosed as cough variant asthma (CVA). Cough in patients with sputum eosinophilia improved after treatment with corticosteroids. Compared with postinfectious cough (PIC) and NAEB, CVA had significantly higher median eosinophil count in induced sputum (0.5% vs 7.5% vs 20.0%, P < 0.01). MMEF in CVA was significantly lower than PIC and NAEB (P < 0.05). The common causes of subacute cough following acute upper respiratory tract infection (AURTI) were PIC (37.8%), NAEB (18.5%), CVA (14.3%) and upper airway cough syndrome (UACS) (10.1%). Atopic cough (AC) (5.2%) and gastroesophageal refluxrelated cough (GERC) (3.4%) were less common in subacute cough following AURTI, while 9 (7.8%) patients had unexplained cough. Conclusion: Subacute cough following AURTI can be attributed to different entities, eosinophilic airway inflammation is common. Induced sputum should be considered when evaluating patients with subacute cough following acute upper respiratory tract infection.