Objective To study whether the presence of gastric pepsin in the sputum might be used as a reliable criteria of laryngopharyngeal reflux. Methods Fifty-six patients with the symptoms of laryngopharyngitis and fifteen healthy people were recruited. Fifty-six patients were divided into laryngopharyngeal reflux group and chronic laryngitis group by the reflux symptom index (RSI), by the reflux finding score (RFS) and by their treating experiment taking omeprazole 20 mg bid for 2 weeks. Sputum in all three groups was obtained in the morning. Pepsin in the sputum was measured by enzyme-linked immunoadsorbent assay. Results The positive rate of pepsin in sputum among LPR group, chronic laryngopharyngitis group and normal group were 93.8% (30/32), 75.0% (18/24), 20.0% (3/15) respectively, and the median concentration of pepsin were 5. 3 [1.3;53.4] ng/ml, 0.8 [0.1;17.2] ng/ml, 0.0 [0.0;0.0] ng/ml (H=23.29, P=0.000). Compared with co-diagnosis as gold standard, the sensitivity of RSI, RFS treating experiment and the pepsin immunoassay was 59.4%, 84.4%, 81.3% and 93.8%, and the specificity of those was 87.2%, 61.5%, 95. 8% and 46.2% respectively. Conclusions Detection of pepsin in sputum by immunoassay might provide a high sensitive, noninvasive method for laryngopharyngeal reflux.