Background. Gestational diabetes mellitus (GDM) contributes to adverse pregnancy and birth outcomes. In recent decades, extensive research has been devoted to the early prediction of GDM by various methods. Machine learning methods are flexible prediction algorithms with potential advantages over conventional regression. Objective. The purpose of this study was to use machine learning methods to predict GDM and compare their performance with that of logistic regressions. Methods. We performed a retrospective, observational study including women who attended their routine first hospital visits during early pregnancy and had Down’s syndrome screening at 16-20 gestational weeks in a tertiary maternity hospital in China from 2013.1.1 to 2017.12.31. A total of 22,242 singleton pregnancies were included, and 3182 (14.31%) women developed GDM. Candidate predictors included maternal demographic characteristics and medical history (maternal factors) and laboratory values at early pregnancy. The models were derived from the first 70% of the data and then validated with the next 30%. Variables were trained in different machine learning models and traditional logistic regression models. Eight common machine learning methods (GDBT, AdaBoost, LGB, Logistic, Vote, XGB, Decision Tree, and Random Forest) and two common regressions (stepwise logistic regression and logistic regression with RCS) were implemented to predict the occurrence of GDM. Models were compared on discrimination and calibration metrics. Results. In the validation dataset, the machine learning and logistic regression models performed moderately (AUC 0.59-0.74). Overall, the GBDT model performed best (AUC 0.74, 95% CI 0.71-0.76) among the machine learning methods, with negligible differences between them. Fasting blood glucose, HbA1c, triglycerides, and BMI strongly contributed to GDM. A cutoff point for the predictive value at 0.3 in the GBDT model had a negative predictive value of 74.1% (95% CI 69.5%-78.2%) and a sensitivity of 90% (95% CI 88.0%-91.7%), and the cutoff point at 0.7 had a positive predictive value of 93.2% (95% CI 88.2%-96.1%) and a specificity of 99% (95% CI 98.2%-99.4%). Conclusion. In this study, we found that several machine learning methods did not outperform logistic regression in predicting GDM. We developed a model with cutoff points for risk stratification of GDM.

Preeclampsia leads to adverse outcomes for pregnant women. Bisphenol A (BPA) is an environmental endocrine disruptor and has been shown to be positively associated with increased risk of preeclampsia in human studies. We investigated whether BPA exposure causes preeclampsia‐like features in pregnant mice and the associated underlying mechanisms. Experiments were performed in animal models and cell cultures. In pregnant mice, BPA‐exposed mice exhibited preeclampsia‐like features including hypertension, disruption of the circulation, and the placental angiogenesis biomarkers fms‐related tyrosine kinase 1 and placenta growth factor, and glomerular atrophy; urinary protein was not affected. These preeclampsia‐like features correlated with increased retention of smooth muscle cells and reduced vessel areas at the junctional zone of the placenta. In addition, there were disrupted expression of invasion‐related genes including increased tissue inhibitors of metalloproteinases, decreased metalloproteinases, and Wnt family member WJVT2/β‐catenin, which correlated with increased DNA methylation in its promoter region and upregulation of DNA methyltransferase (Dnmt)l. BPA exposure impeded the interaction between the human cytotrophoblast cell line, HTR‐8/SVneo, and endothelium cells. BPA exposure down‐regulated WNT2 expression, and elevated the DNA methylation of WNT2; these results were consistent with in vivo observations. Inhibition of DNMT in HTR‐8/SVneo cells resulted in reduced DNA methylation and increased expression of WNT2 . Taken together, these data demonstrate that BPA exposure alters trophoblast cell invasion and causes abnormal placental vessel remodeling, both of which lead to the development of preeclampsia‐like features in pregnant mice. Our results suggest that this phenomenon involves the epigenetic reprogramming and down‐regulation of WNT2 mediated by DNMT1 .—Ye, Y., Tang, Y., Xiong, Y., Feng, L., Li, X. Bisphenol A exposure alters placentation and causes preeclampsia‐like features in pregnant mice involved in reprogramming of DNA methylation of WNT2. FASEB J. 33,2732–2742 (2019). www.fasebj.org

Long-chain per- and polyfluoroalkyl substances (PFASs) and their short-chain alternatives have been produced and used extensively in China. However, it is unclear whether these compounds contribute to the risk of gestational diabetes mellitus (GDM) in women residing in contaminated areas. The study was performed to explore the association between PFASs varying in chain length and the risk of developing GDM. A nested case-control study was conducted in a prospective cohort of 2,460 pregnant women between July 1, 2017, and January 31, 2019 in Shanghai, China. Twelve PFASs of interest were measured using ultra-performance liquid chromatography/quadrupole time-of-flight mass spectrometry (UPLC-Q/TOF MS) in the sera of pregnant women at 16-20 weeks. GDM was diagnosed by an oral glucose tolerance test administered over 24-28 gestational weeks. The cases and controls were matched by maternal age. The relationship between maternal serum PFAS level and GDM risk was determined by conditional logistic and linear regression analyses. A total of 165 GDM cases and 330 controls were enrolled in the study cohort. The frequencies of detection of PFHpA, PFDS, and PFOSA were all ≤80%. Hence, they were excluded from any further risk analysis. Perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) were detected at relatively high serum levels (medians 6.57 ng/mL and 8.07 ng/mL, respectively). The serum levels of perfluorobutanesulfonic acid (PFBS) and perfluorododecanoic acid (PFDoA) were significantly higher in the GDM group than they were in the control group (P = 0.02 and P < 0.01, respectively) according to a nonparametric Wilcoxon rank sum test. A quartile analysis showed that the odds ratio of GDM would significantly increase at the highest PFBS and PFDoA levels. In the core model, the adjusted ORs were 2.02 (95% CI = 1.04-3.79) and 13.00 (95% CI = 4.74-24.59), respectively, after adjusting for maternal age, sampling time, parity and body mass index [BMI]). Elevated maternal serum PFBS and PFDoA levels in early pregnancy may be associated with a substantially higher GDM risk.

Preeclampsia (PE) is a leading cause of maternal and fetal morbidity and mortality. Although increasing lines of evidence suggest that both the placenta and the decidua likely play roles in the pathogenesis of PE, the molecular mechanism of PE remains elusive partly because of the heterogeneity nature of the maternal-fetal interface. In this study, we perform single-cell RNA-seq on the placenta and the decidual from patients with late-onset PE (LOPE) and women in normal pregnancy. Analyses of single-cell transcriptomes reveal that in LOPE, there are likely a global development deficiency of trophoblasts with impaired invasion of extravillous trophoblasts (EVT) and increased maternal immune rejection and inflammation in the placenta, while there are likely insufficient decidualization of decidual stromal cells (DSC), increased inflammation, and suppressed regulatory functions of decidual immune cells. These findings improve our understanding of the molecular mechanisms of PE.