Material and Method: Three hundred and three patients (186 males and 117 females) who underwent anthropometric measurement and standard polysomnography were asked to complete the STOP-Bang questionnaire. Subjects were considered high-risk if their scores were ≥3. Patients with significant co-morbidities were excluded.

Results: Screening for OSA involved measurements of STOP-Bang sensitivity, specificity, positive predictive value, and negative predictive value at several apnea-hypopnea index (AHI) cut-off points. At AHI 5, these values were 87.3%, 48.1%, 82.2%, and 52.2%, respectively. At AHI 15, these values were 92.6%, 36.4%, 58.5%, and 83.6%, respectively. The modified STOP-Bang (using a cut-off of BMI >30 kg/m2) showed slightly increased sensitivities at the AHI cut-off points of 5 and 15 with values of 88.7% and 93.2%, respectively, with improved area under the curves. Furthermore, by applying the WHtR of ≥0.55 to those patients who were classified as high-risk by the questionnaires, the specificities for predicting OSA were improved to 85.2% and 76.1% for the aforementioned cut-off points, respectively.

Conclusion: Both STOP-Bang and its modified version were highly sensitive measures for OSA screening in medical or dental clinics. However, the modified version might be more suitable for Thais and Asians, and the additional use of WHtR ≥0.55 might be useful for reducing the unnecessary sleep investigation or management in those who were classified as high-risk patients.

Keywords: Screening, Obstructive sleep apnea (OSA), STOP-Bang, Modified STOP-Bang, waist-to-height ratio (WHtR), Thai patients