Diagnostic accuracy of the STOP-Bang questionnaire for obstructive sleep apnea in pregnancy: A systematic review and meta-analysis.

BACKGROUND

Obstructive sleep apnea (OSA) in pregnancy is associated with adverse maternal and perinatal outcomes, yet timely diagnostic testing can be limited. Screening questionnaires such as STOP-Bang are therefore used to triage patients for objective testing.

OBJECTIVE

To synthesize evidence on the diagnostic accuracy of the STOP-Bang questionnaire for detecting OSA in pregnancy.

METHODS

We conducted a PRISMA 2020-compliant systematic review (PROSPERO: CRD420261297348). PubMed, Scopus, Web of Science, and EBSCOhost databases were searched from inception through February 2026. Eligible studies evaluated STOP-Bang in pregnant participants against polysomnography (PSG) or home sleep apnea testing (HSAT)/portable monitoring. The risk of bias was assessed using the QUADAS-2. When ≥ 2 studies provided complete (or derivable) 2 × 2 data at STOP-Bang ≥ 3 and AHI/REI ≥ 5, we performed a hierarchical bivariate random-effects diagnostic meta-analysis (Reitsma model; REML) to pool sensitivity and specificity jointly; pathway-based partial verification studies were synthesized narratively.

RESULTS

Five cohort studies (2015-2024) comprising 1,096 pregnant participants were included; 573 underwent objective sleep testing. Reference standards were predominantly HSAT/portable monitoring, with limited PSG in a small subset in one study. Reported diagnostic performance varied across studies, with sensitivity ranging from approximately 0.53 to 0.63 and specificity from 0.36 to 0.90 [1-3], and individual-study AUC values ranging from approximately 0.63 to 0.75. Studies embedded within referral pathways or with partial verification yielded less stable and potentially biased estimates. QUADAS-2 identified flow and timing as the most frequent concern due to partial verification/work-up bias. Hierarchical meta-analysis (STOP-Bang ≥ 3; AHI/REI ≥ 5) yielded pooled sensitivity 0.587 (95% CI 0.464-0.699) and pooled specificity 0.749 (95% CI 0.360-0.941), with AUC 0.604. In sensitivity analysis excluding a class III obesity cohort, pooled specificity increased to 0.860 and AUC to 0.832, while pooled sensitivity remained similar (0.566). Sensitivity was relatively consistent across studies, whereas specificity varied widely.

CONCLUSION

Current evidence does not support STOP-Bang as a reliable stand-alone screening tool for obstructive sleep apnea in pregnancy. Its diagnostic performance appears variable and generally modest, and pregnant patients with suspected OSA should undergo objective sleep evaluation.