Abstract / Summary
Abstract Objective This study aimed to investigate the association between mean nocturnal oxygen saturation (MSpO₂) in pregnant women at high risk of obstructive sleep apnea (OSA) and pregnancy complications. We focused on the MSpO₂during sleep monitoring in the first, second, and third trimesters, which is easy to measure and may have a warning role for pregnancy complications. Given that gestational diabetes mellitus (GDM) is typically diagnosed at 24–28 weeks, we interpreted the third-trimester findings as a concurrent association rather than a prospective prediction. Methods A retrospective analysis was conducted on 569 high-risk OSA pregnant women who underwent portable sleep monitoring. Participants were stratified by trimester (first, n = 90; second, n = 179; third, n = 300) and further categorized by MSpO₂ level (< 95% vs. ≥95%). Maternal and neonatal outcomes were compared, mainly including hypertensive disorders of pregnancy (HDP), GDM, postpartum hemorrhage, delivery gestational age, umbilical artery pH, placental weight, PLGF levels, and the rate of Apgar 1min ≤ 7. Logistic regression and receiver operating characteristic (ROC) analyses were performed to evaluate the discriminative ability of MSpO₂ for GDM. Multivariate logistic regression models were used to assess for effect modification by including an interaction term for obesity status and age. Results Lower MSpO₂(MSpO₂<95%) in the third trimester was significantly associated with a higher incidence of GDM (P = 0.025). After adjusting for age, pre-pregnancy BMI, and PCOS history, MSpO₂ remained an independent protective factor against GDM (adjusted OR = 0.680, 95% CI: 0.533–0.869, P = 0.002). The model combining MSpO₂ with clinical risk factors yielded the highest AUC of 0.731, indicating its utility for concurrent risk stratification. No significant interaction effects were found for age or obesity. In the second trimester, low MSpO₂was associated with lower umbilical artery pH (P = 0.039) and higher placenta-to-newborn weight ratio (P = 0.037). However, no significant association was found between OSA and HDP. The association between oxygen saturation and GDM was consistent across subgroups defined by age and obesity status, with no significant interaction effects observed. Conclusion Lower nocturnal oxygen saturation in the third trimester is independently associated with an increased risk of GDM in high-risk OSA pregnancies. While temporality precludes prediction, MSpO₂ may serve as a useful biomarker for risk stratification, suggesting that the third trimester may represent a critical window during which oxygen metabolism abnormalities impact glucose metabolism. Prospective studies with pre-diagnostic SpO₂ monitoring are warranted.