Correlation between serum HSP70 and PD-L1 levels and vaginal microbiota distribution and pregnancy out-comes in late pregnancy PCOS patients
CHEN Zhi-qin
LIU Li-hua
LIU Xian
LI Yuan
CHEN Wen-ling
Abstract：Objective To identify independent risk factors for adverse pregnancy outcomes in women with Poly-cystic Ovary Syndrome(PCOS)during late pregnancy,to construct and validate a nomogram model,and to explore the correlation between serum Heat Shock Protein 70(HSP70),Programmed death-ligand 1(PD-L1)levels,vaginal mi-crobiota distribution,and pregnancy outcomes in these patients.Methods A total of 180 late-pregnancy PCOS patients admitted from January 2023 to January 2024 were prospectively enrolled as the modeling group,while 180 late-pregnancy women without PCOS during the same period were selected as the control group.Laboratory indicators and clinical data were collected for both groups,and the occurrence of common adverse pregnancy outcomes(fetal distress,macrosomia,polyhydramnios,postpartum hemorrhage,premature rupture of membranes,etc.)was recorded.The incidence of adverse pregnancy outcomes,vaginal microbiota species distribution,serum HSP70 levels,and serum PD-L1 levels were com-pared between the two groups.Multivariate logistic regression analysis was used to identify independent risk factors for ad-verse pregnancy outcomes in late-pregnancy PCOS patients.A nomogram model was constructed based on the modeling group and validated using an external validation group of 80 pregnant women(recruited between February and August 2024 at a 7∶3 split from the modeling group).Results Serum HSP70 and PD-L1 levels were significantly lower in the mod-eling group than in the control group.The main vaginal microorganisms detected in both groups were Lactobacillus,Esche-richia coli,Staphylococcus epidermidis,and Ureaplasma urealyticum.The detection rates of E.coli(48.89％)and Can-dida albicans(24.44％)in the modeling group were higher than those in the control group(36.11％and 11.67％,re-spectively;both P＜0.05).The modeling group showed higher bacterial density and a higher proportion of vaginal micro-ecological imbalance(both P＜0.05).Among the modeling group,those with vaginal microecological imbalance had sig-nificantly lower serum HSP70 and PD-L1 levels than those with normal vaginal microecology(both P＜0.05).Multivari-ate logistic regression showed that age,BMI,vaginal microecological imbalance,Interleukin 6(IL-6),Tumor necrosis factor alpha(TNF-α),HSP70,and PD-L1 were independent risk factors for adverse pregnancy outcomes in PCOS pa-tients(all P＜0.05).The nomogram model demonstrated good predictive performance.The area under the ROC curve(AUC)was 0.900(95％CI:0.855-0.946,P＜0.05)in the modeling group with sensitivity 0.896 and specificity 0.805,and 0.889(95％CI:0.815-0.962,P＜0.05)in the validation group with sensitivity 0.912 and specificity 0.804.Calibration curves indicated good model fit(all P＜0.05).Conclusion Serum HSP70 and PD-L1 levels in late-pregnancy PCOS patients are strongly correlated with vaginal microecology and pregnancy outcomes.The nomogram-based prediction model for adverse pregnancy outcomes in late-pregnancy PCOS patients shows good predictive per-formance and clinical applicability.
Keywords：late pregnancypolycystic ovary syndromeserum heat shock protein 70programmed death receptor ligand 1vaginal microbiotapregnancy outcomes
Publication Date：2026-01-15
Online Publishing Date：2026-03-04(First online date of this platform, not the publication date of the document)
Pages：9( 82-90 )