Incorporating body-type (apple vs. pear) in STOP-BANG questionnaire improves its validity to detect OSA

The aim of this study is to evaluate whether adding the item of "apple body type" to the STOP-BANG questionnaire enhances diagnostic performance of the questionnaire for detecting obstructive sleep apnea (OSA).

Cross-sectional study.

Sleep center setting.

Two hundred and eight subjects who were referred for an evaluation of possible OSA at Tulane Comprehensive Sleep Center. The exclusion criteria were age<18years old, incomplete or absent questionnaire, incomplete body type identification, polysomnography (PSG) refusal, and pregnant women.

STOP-BANG items and body type data were collected on the initial clinic visit. An overnight PSG was performed on every participant.

Descriptive analyses of the demographic data and PSG variables were performed. The predictive parameters of STOP and STOP-BANG without and with body type score (STOP-Apple and STOPBANG-Apple) were compared.

The STOP questionnaire's sensitivity/specificity/positive likelihood ratio (+LR) (cut-off=2) was 96%/11%/1.1, respectively whereas the STOP-Apple questionnaire (cut-off=3) was 88%/39%/1.5. The STOP-BANG's sensitivity/specificity/+LR (cut-off=3) was 96%/19%/1.2, respectively whereas the STOP-BANG-Apple questionnaire (cut-off=4) was 90%/39%/1.5. The area under the Receiver Operating Characteristic (ROC) curve of STOP-Apple was comparable to the STOP-BANG (P=0.25). The addition of the apple body type item to the STOP-BANG questionnaire in participants with a score≥3 led to increased specificity (67.4%), increased the odds ratio of having OSA of 2.5 (95% CI, 1.2-5.3) and odds ratio of having moderate-severe OSA of 4.7 (95% CI, 2.5-8.7).

In the sleep center setting, adding the body type item to the STOP-BANG questionnaire improves not only clinical prediction for PSG confirmed OSA but also predicts moderate to severe of OSA.