Purpose of the study Effective screening questionnaires are essential for early detection of obstructive sleep apnea (OSA). The STOP-Bang questionnaire has high sensitivity but low specificity. Dry mouth is a typical clinical sign of OSA. We hypothesised that adding dry mouth in the STOP-Bang questionnaire would improve its specificity.

Study design A survey of the incidence of dry mouth was performed in a general population group and suspected sleep apnea clinical population group. Patients with suspected OSA were assessed by laboratory polysomnography and STOP-Bang questionnaire was performed. Adding the option of dry mouth to the OSA screening questionnaire resulted in a new quesionnaire, where cut-off value, diagnostic efficacy and the predictive parameters (sensitivity, specificity, positive predictive value and negative predictive value) were explored.

Results (In the 912 general population group, the incidence of dry mouth in the snoring group (54.0%) was much higher than that in the non-snoring group (30.5%) (p<0.05). In 207 patients with suspected OSA, the incidence of dry mouth in the OSA group was much higher than that in the non-OSA group (p<0.05). The sensitivity and specificity of the STOP-Bang questionnaire were 88.8% and 23.7% for identifying OSA, and 92.2% and 23.1% for identifying moderate and severe OSA, respectively. Adding the option of dry mouth (dry mouth every morning) to the STOP-Bang questionare resulted in a new questionnaire (STOP-Bang-dry-mouth questionnarie) with 9 items. Its sensitivity and specificity were 81.70% and 42.10% for identifying OSA, and 89.10% and 42.30% for identifying moderate and severe OSA, respectively.

Conclusions The dry mouth symptom correlated with snoring and sleep apnea. The specificity of the STOP-Bang questionnaire can be improved by integrating dry mouth. The diagnostic accuracy of the STOP-Bang-dry mouth questionnaire is yet to be further verified in prospective studies.