OBJECTIVE:To investigate whether prophylactic tranexamic acid reduces the incidence of postpartum haemorrhage in women with placenta praevia compared with placebo. DESIGN:Randomised, double blind, placebo controlled, phase 3 study. SETTING:24 maternity units across China between July 2023 and March 2025. PARTICIPANTS:1732 women with placenta praevia undergoing caesarean delivery. INTERVENTIONS:Participants were randomly (1:1) assigned to receive prophylactic oxytocin and either tranexamic acid (1 g in 10 mL) or placebo (10 mL normal saline) diluted in 40 mL normal saline intravenously over 10 minutes, initiated within five minutes of umbilical cord clamping. MAIN OUTCOME MEASURES:The primary outcome was postpartum haemorrhage, defined as calculated estimated blood loss ≥1000 mL or as red cell transfusion within two days after delivery. Serious adverse events included thromboembolic events, seizures, acute kidney or liver injury, and maternal death. RESULTS:Of 1732 women with placenta praevia who were randomised, 38 were excluded because they withdrew consent or were determined to be ineligible after randomisation. Primary outcome data were available for 99.8% (1691/1694) of the remaining women. Placenta accreta spectrum was diagnosed in 303 participants (17.9%). The primary outcome occurred in 29.7% (251/845) of the tranexamic acid group and 35.1% (297/846) of the placebo group (relative risk 0.85, 95.2% confidence interval (CI) 0.75 to 0.96; P=0.01). The rates of serious adverse events were similar between the tranexamic acid group and placebo group (0.5% (4 of 837) v 0.5% (4 of 845); relative risk 1.01, 95% CI 0.25 to 4.00). CONCLUSIONS:In women with placenta praevia who underwent caesarean delivery and received prophylactic oxytocin, treatment with tranexamic acid resulted in a statistically significant yet modest reduction in the incidence of postpartum haemorrhage, with no signal of increased serious adverse events. TRIAL REGISTRATION:ClinicalTrials.gov NCT05811676.

ObjectiveGestational diabetes mellitus (GDM) is associated with gut microbiota dysbiosis and placental dysfunction. This study aimed to investigate whether gut microbiota-derived short-chain fatty acids (SCFAs) ameliorate placental injury and insulin resistance in GDM by suppressing ferroptosis, a novel form of iron-dependent cell death.MethodsA GDM rat model was induced by streptozotocin, and gut microbiota was depleted via a broad-spectrum antibiotic cocktail. Physiological parameters, gut microbiota composition, and SCFA levels were assessed. Placental and metabolic tissues were analyzed for ferroptosis markers (Fe2+, MDA, ROS), key regulatory enzymes (GPX4, ACSL4, LPCAT3), polyunsaturated fatty acid (PUFA) profiles, and inflammation. An in vitro insulin resistance model in human trophoblast cells (HTR-8/SVneo) was used to validate the causal role of SCFAs.ResultsGDM and antibiotic-induced dysbiosis led to a significant depletion of SCFA-producing bacteria and reduced colonic SCFA levels. This was concomitant with severe placental ferroptosis, characterized by iron accumulation, lipid peroxidation, and downregulation of GPX4 alongside upregulation of ACSL4 and LPCAT3. PUFA substrates were significantly consumed, and systemic inflammation was elevated. In vitro, insulin resistance induced trophoblast ferroptosis and dysfunction, which was effectively rescued by SCFA treatment. SCFAs restored GPX4 expression, suppressed ACSL4/LPCAT3, preserved PUFAs, and attenuated inflammation.ConclusionGut microbiota-derived SCFAs play a critical protective role in GDM by inhibiting the ACSL4/LPCAT3-mediated ferroptotic pathway in placental trophoblasts, thereby improving insulin sensitivity and mitigating cellular injury. Our findings highlight the gut-placenta axis as a potential therapeutic target for GDM, warranting further validation in human studies.

Neutrophils are the most abundant leukocytes in human peripheral blood, yet their heterogeneity in pregnancy, especially in gestational diabetes mellitus (GDM), remains incompletely understood. Here, we employed InfinityFlow-based surface marker profiling and single-cell RNA sequencing (scRNA-seq) to delineate neutrophil subsets in healthy and GDM pregnancies. We identified a low-density, immature subgroup (CD10⁻CD49d⁺Ig κ⁺) with distinctive morphology and transcriptomic profiles, contrasted against the CD10⁺ segmented mature neutrophils. In healthy pregnancy, these immature low-density neutrophils (LDNs) expanded by mid-gestation, elevating the immature-to-mature (I–M) neutrophil ratio and circulating myeloid progenitor levels throughout gestation. In GDM, this expansion was markedly blunted, with a consistently lower abundance of immature LDNs and progenitors. Flow cytometry and correlation analyses further linked the lower I–M ratio to impaired insulin sensitivity, underscoring a potential immune-metabolic axis in GDM. Collectively, our study provides a framework for neutrophil phenotyping in pregnancy and links disrupted neutrophil equilibrium and pregnancy complications. Integrated InfinityFlow and scRNA-seq profiling reveal a disrupted neutrophil landscape in GDM: blunted expansion of immature cells and progenitors and a lower I-M ratio linked to metabolic dysfunction, highlighting an immunometabolic axis.

Universal screening for hyperglycaemia in women without pregestational diabetes in China has led to a significant rise in the detection of abnormal glucose metabolism in early pregnancy (EAGM), defined as elevated fasting blood glucose (FPG, 5.1–6.9 mmol/L or 92–124 mg/dL) and/or haemoglobin A1c (HbA1c, 5.7

BACKGROUND:Forceps delivery often leads to increased maternal and neonatal complications, primarily attributed to the rigidity of traditional steel forceps, a major contributor to birth injury. AIMS:This study aims to develop obstetric forceps with reduced rigidity and assess the safety and effectiveness of the newly designed pliant forceps. METHODS:Pliant forceps with varying materials and shapes were produced by three-dimensional printing, and a pilot simulation determined the optimal design. Three fetal mannequins were developed to measure force changes exerted on the fetal head. Simulated births with several resistance intensities, were conducted to compare the real-time force between pliant forceps and Simpson forceps. Perineal distension was assessed by recording maximum perineal distention during a simulated forceps delivery using pliant forceps and Simpson forceps. FINDINGS:The pliant forceps, constructed from polylactic acid with solid blades and angled shanks, featured foam tape on the fetal sides. In simulation studies on term fetal head model, pliant forceps achieved successful assisted vaginal delivery across all resistance levels. Compared to Simpson forceps, pliant forceps consistently exerted lower force on the fetal head during assisted vaginal delivery. The maximum force applied by pliant forceps occurred at RA1 site on fetal head (67.11 ± 4.35 N, Simpson forceps: 99.12 ± 10.53 N, p < 0.001), and Simpson forceps reached its peak at RP2 site (177.37 ± 19.28 N, pliant forceps: 12.87 ± 5.11 N, p < 0.001). Similar results were obtained in simulation experiments on large and small fetal head models. Perineal distension was determined to be smaller in births with pliant forceps compared with that in births with Simpson forceps (lateral perineal distension: 76.6 mm vs. 92.6 mm, p < 0.001). CONCLUSIONS:The three-dimensional-printed pliant forceps demonstrated reduced force on the fetal head and less perineal distension compared to Simpson forceps in simulated births, which holds potential for decreasing birth injuries and maternal birth canal injuries during forceps delivery. Further research is required to ensure the safety and efficacy of pliant forceps before clinical application.

Gestational diabetes mellitus (GDM) is the most common complication during pregnancy and early prediction for high-risk gravidas is crucial to facilitate timely intervention. However, most newly discovered biomarkers require additional blood sampling and high costs, which limits their feasibility in routine clinical practice. This study aimed to assess whether noninvasive retinal parameters could enhance the predictive performance of established models for GDM. This prospective cohort study collected demographic characteristics, glycolipid metabolism indices, and retinal images at 11–13+6 weeks of gestation. The primary outcome was GDM based on oral glucose tolerance test at 24–28 weeks. Variables were selected via least absolute shrinkage and selection operator (LASSO) regression. Six machine learning algorithms (logistic regression, random forest, extreme gradient boosting, categorical boosting [CatBoost], adaptive boosting, and support vector machine) were employed to develop the LIGHT (LIpid+Glucose+opHthalmic+maTernal factors) model incorporating baseline, glycolipid metabolism and retinal features. Model performance was assessed by the area under the receiver-operating-characteristic curve (AUC) and interpreted by the Shapley Additive Explanation method (SHAP). The incremental value of retinal features was evaluated by net reclassification improvement (NRI) and integrated discrimination improvement (IDI). We compared LIGHT model with (1) Baseline model based on demographic characteristics, (2) Glycolipid model including baseline and glycolipid metabolism features, and (3) Eye model using baseline and retinal features. Of the 2114 participants, 1774 pregnancies were included in the final analysis, of which 324 (18.3

Fetal hypoxia is a leading cause of neonatal morbidity and mortality. Cardiotocography (CTG) is widely used to predict fetal hypoxia during labor, but its interpretation remains suboptimal. Artificial intelligence (AI) models have been developed for CTG interpretation, but their clinical utility is limited by two major challenges: demonstrating superiority over human experts and ensuring explainability in real-world settings. A large dataset containing CTG traces from three tertiary hospitals between January 2014 and May 2022 was built for model development. Deep learning architectures, named Cardiotocography Artificial-intelligence Predictors (CAPs), were trained to predict fetal hypoxia from CTG traces based on CNN (CAP-C), Transformer (CAP-T), LSTM (CAP-L), and CfC (CAP-CfC) algorithms. The outcome was fetal hypoxia, determined by either low Apgar score (≤ 7 at 1 or 5 min) or umbilical artery acidemia (grade 1: pH of umbilical artery (pHa) < 7.20; grade 2: pHa < 7.15; grade 3: pHa < 7.10). Model performance was determined by area under the receiver operating characteristic curve (AUROC), evaluated through nationwide AI-human comparison and validated on the CTU-UHB dataset. Gradient-weighted class activation mapping (Grad-CAM) was applied to highlight the CTG regions that contributed most to the model’s predictions. A total of 20,780 CTG traces were obtained for model development, and 467 cases were held out for the nationwide AI-human comparison. Among all models, CAP-L achieved highest AUROC in predicting fetal hypoxia (grade 1: 0.758, 95

ObjectiveGestational diabetes mellitus (GDM) is a frequent pregnancy complication that increases short- and long-term risks for both mother and child. However, its underlying molecular mechanisms remain poorly understood. This study aims to unravel the molecular basis of GDM and explore potential therapeutic targets.MethodsWe integrated genomic, transcriptomic, and single-cell RNA sequencing datasets to delineate cell-type-specific alterations in GDM. Candidate genes were prioritized using Mendelian randomization (MR), followed by quantitative PCR (qPCR) validation in placental samples. Pathway and immune-network analyses were performed to contextualize biological function.ResultsSingle-cell profiling showed marked remodeling of immune compartments in GDM, with prominent changes in monocytes and T-cell subsets. Two-sample MR prioritized 15 genes with putative causal links to GDM, including BNIP3L, COMT, CTSB, LMNA, and SLC7A5. qPCR further demonstrated significant differential expression of CTSB, LMNA, and SLC7A5 between GDM and control placentas (human or mouse). Pathway enrichment implicated CTSB in immune regulation and metabolic processes, whereas LMNA and SLC7A5 mapped to insulin resistance and glucose/amino-acid transport pathways. Immune-network analysis revealed significant correlations between these genes and immune mediators, supporting immune dysregulation as a contributor to GDM pathogenesis.ConclusionThis study provides a comprehensive analysis of the immune-metabolic landscape of GDM. Key genes identified in this study may serve as potential biomarkers and therapeutic targets for early diagnosis and personalized treatment of GDM. Further studies are warranted to elucidate the underlying mechanisms and develop targeted therapies for this disease.

The second trimester of pregnancy is a pivotal stage in human immune system development. Utilizing single-cell RNA sequencing and T cell receptor sequencing, we profiled 2,868,420 immune cells from 321 samples across 23 organs, including adult tissues as comparators. We identify an extrathymic CD4+ T cell subset mediating TOX2+ precursor cells' transition to mature naive CD4+ T cells. Contrary to the prevailing paradigm of fetal immune quiescence, we uncover widespread memory/activated T cells and tissue-resident memory clones shared across organs, indicating systemic immune activity beyond localized barrier defense. Cell-cell communication and functional assays indicate two tolerance mechanisms that suppress fetal T cell activation: ARG1+ neutrophils and a PTGES3/PTGER4 signaling pathway. We also find that hematopoietic stem cells (HSCs) disperse across multiple organs and show that HSCs from non-canonical hematopoietic organs differentiate into diverse immune lineages. These findings provide insights into human immune system maturation and tolerance in fetuses and adults.

INTRODUCTION:This study aimed to evaluate the screening outcomes in women with hyperglycemia in early pregnancy (fasting plasma glucose [FPG] 5.1-6.9 mmol/L and/or HbA1c 39-46 mmol/mol before 20 weeks of gestation). METHODS:This multicenter retrospective cohort study was conducted in China between 2016 and 2022. In our setting, all women without pregestational diabetes performed both FPG and HbA1c screening at the first prenatal visit. Logistic regression models adjusted for confounders were performed to assess the associations of hyperglycemia in early pregnancy with adverse pregnancy outcomes. Subgroup analyses were explored according to the subsequent diagnosis of gestational diabetes (GDM, with or without). RESULTS:Of the 42,999 women in the analysis, 2515 (5.8%) women had hyperglycemia in early pregnancy. Compared with women with normal FPG and HbA1c levels, women with FPG 5.1-6.9 mmol/L and/or HbA1c 39-46 mmol/mol had a 3-fold increased risk of GDM (aOR 3.85; 95% CI 3.52-4.20), and 1-fold higher risk of hypertensive disorders of pregnancy (1.42; 1.20-1.67), shoulder dystocia (1.30; 1.11-1.52), preterm birth (1.30; 1.11-1.52), large-for-gestational-age (1.26; 1.12-1.43), and macrosomia (1.43; 1.19-1.73). Women with hyperglycemia in early pregnancy complicated by GDM were associated with a 50%, 84%, 48% and 24% increase in the odds of developing hypertensive disorders of pregnancy (1.50; 1.21-1.84), preterm premature rupture of membranes (1.84; 1.09-3.10), preterm birth (1.48; 1.22-1.81) and large-for-gestational-age (1.24; 1.05-1.45), respectively, compared with those without hyperglycemia. CONCLUSIONS:Pregnant women with hyperglycemia in early pregnancy have an increased risk of adverse pregnancy outcomes, and women with these conditions complicated by GDM are at higher risk than those without. Further research is needed to explore whether the incidence of GDM can be reduced by early intervention and therefore prevent the relevant adverse pregnancy outcomes.

OBJECTIVES:Adverse social contexts, such as lockdowns and disasters during pregnancy, can significantly impact maternal and neonatal outcomes. However, the specific effects on different populations remain unclear. This study aimed to investigate the variations in pregnancy outcomes resulting from the pandemic lockdown and to identify distinct populations in need of targeted intervention. METHODS:Women who delivered at our institution spanning from 2017 to 2019 (pre-pandemic) and from 2020 to 2022 (during the pandemic lockdown) were included in this study. A comparison was conducted on maternal and neonatal outcomes across a total of 19,382 singleton pregnancies, with a specific focus on those affected by gestational diabetes mellitus (GDM). RESULTS:As a total of 19,382 singleton pregnant women were included, this study found a significant increase in the incidence rates of GDM with an odds ratio of 1.194 (95%CI: 1.109-1.286, p < .001). Additionally, following the pandemic lockdown, there was an increase in rates of premature birth, premature rupture of membranes, and intrahepatic cholestasis of pregnancy. Further analysis of the GDM cohort revealed a notable rise in the risk of Group B Streptococcus infection and neonatal small for gestational age (SGA). Specifically, GDM patients with a body mass index (BMI) less than 18.5 kg/m2 exhibited an increased risk of fetal growth restriction and SGA. CONCLUSIONS:The pandemic lockdown adversely affected pregnancy outcomes. Therefore, targeted screening and clinical management for specific populations should be prioritized.

BACKGROUND:Preeclampsia is a systemic disorder unique to pregnancy that is associated with trophoblast dysfunction. Although the c-Fos (Fos proto-oncogene) is essential for placental development, its mechanistic role in preeclampsia remains unclear. METHODS:We identified c-Fos as a hub gene through an analysis of cell-free RNA from maternal plasma and single-cell RNA sequencing of preeclamptic placentas. We assessed c-Fos expression in the placenta and plasma and examined its correlation with lipids. Functional changes upon the viral modulation of c-Fos in trophoblast cells were evaluated. Untargeted lipidomics and RNA sequencing were conducted to uncover the downstream mechanisms, and the findings were validated in trophoblast cells and a preeclampsia-like murine model. RESULTS:c-Fos expression was decreased in maternal peripheral plasma from early to mid-pregnancy and in the placenta of preeclampsia patients. Furthermore, c-Fos expression levels were negatively correlated with neutral lipid accumulation. c-Fos deficiency impaired trophoblast invasion, proliferation, and syncytialization while promoting apoptosis. Reduced c-Fos expression also led to lipid droplet aggregation and mitochondrial dysfunction. Mechanistically, c-Fos silencing inhibited the p-AMPK (phosphorylated AMP-activated protein kinase)/detyrosinated tubulin pathway, disrupting lipid droplet trafficking and metabolism and promoting lipid droplet accumulation and an altered lipid profile. c-Fos deficiency induced a preeclampsia-like phenotype in mice, whereas c-Fos overexpression partially alleviated preeclampsia symptoms. CONCLUSIONS:c-Fos downregulation inhibits the p-AMPK/detyrosinated tubulin pathway, driving lipid droplet accumulation and consequent trophoblast dysfunction, revealing c-Fos as a potential therapeutic target for preeclampsia.

BACKGROUND:Fetal growth restriction (FGR) is a significant concern due to its potential adverse outcomes for both mothers and infants. Cell-free RNA in maternal plasma has been suggested as a potential biomarker for pregnancy complications, but its effectiveness in predicting FGR remains uncertain. This study aimed to assess the predictive value of cell-free RNA profiling from maternal plasma collected during early to mid-pregnancy for FGR. METHODS:This case-control study included pregnant women diagnosed with FGR who had non-invasive prenatal test data. Differentially expressed genes (DEGs) between FGR and controls groups were identified through the analysis of cell-free RNA and placental microarray dataset which downloaded from the Gene Expression Omnibus database. The intersection of DEGs from cell-free RNA and placenta was explored to explore hub genes. The least absolute shrinkage and selection operator regression was used to select the hub genes from the cell-free RNA DEGs. The prediction model was then constructed using logistic regression with hub genes and clinical characteristics. The predictive accuracy of model was evaluated using receiver operating characteristic analysis, calibration curves, and decision curve analysis. RESULTS:A total of 39 FGR samples and 133 control samples were included in this study. Among them, 405 cell-free RNA DEGs were identified. BIN2 was identified as the intersecting gene that was up-regulated in both cell-free RNA and FGR placental transcripts. Subsequently, RHOA and OAZ1 were selected by least absolute shrinkage and selection operator regression. The hub genes, including BIN2, RHOA and OAZ1, exhibited positive correlations with each other and were up-regulated in the FGR group. A logistic regression model incorporating the hub genes and clinical characteristics was constructed, achieving the highest classification performance with area under the curve of 0.812 (95% CI: 0.719-0.904) in the training cohort, 0.863 (95% CI: 0.736-0.989) in the validation cohort, and 0.786 (95% CI: 0.513-1.000) in the time test cohort. The calibration curve indicated good calibration of the model, and the decision curve analysis demonstrated practical value in clinical application. CONCLUSIONS:An effective prediction model for FGR was developed by integrating maternal plasma cell-free RNA with clinical characteristics, enabling early evaluation of FGR risk.

Maternal high-fat diet (HFD) is known to impair the reproductive function of female offspring, but the underlying epigenetic mechanisms of this developmental programming remain unclear. In this study, female rats were fed either a control diet (CD; 10 % kcal from fat) or an HFD (60 % kcal from fat) prior to and during gestation and lactation. After weaning, female offspring were randomly assigned to continue on either a CD or HFD, resulting in four groups: C/C, C/HF, HF/C, and HF/HF. Maternal HFD significantly reduced primordial follicle numbers, increased follicular atresia and granulosa cell apoptosis, and disrupted sex hormone levels in female offspring at 3 months of age. Female offspring from HFD-fed dams exhibited reduced Kiss1 gene expression and increased promoter methylation in both the ovaries and hypothalami. Additionally, the IL-6/STAT3 pathway was activated, and the expression of Ten-eleven translocation methylcytosine dioxygenase 2 (TET2), a key regulator of DNA demethylation, was downregulated. Post-weaning exposure to a normal diet partially attenuated these effects by 6 months of age. Furthermore, in vitro experiments demonstrated that IL-6/STAT3 signaling downregulated the expression of TET2 and Kisspeptin. Overall, our study demonstrates that maternal HFD consumption reduces TET2 expression in the ovaries and hypothalami of female offspring via the IL-6/STAT3 pathway, leading to Kiss1 promoter hypermethylation and a subsequent decrease in Kisspeptin levels. These findings highlight a potential epigenetic mechanism linking maternal diet to long-term reproductive toxicity in female offspring.

This study explored the ability of quercetin to improve glucose and lipid metabolism disorders induced by a high-fat diet (HFD) in a mouse model of gestational diabetes mellitus (GDM) on the basis of the PCSK9/LDLR axis and its potential molecular mechanisms. For the animal experiments, pregnant C57BL/6J mice were randomly divided into three groups: the control group (CD group), which was fed a standard diet; the model group (HFD group), which was fed a HFD; and the quercetin (CAS: 117–39–5) intervention group (HQ group), which was fed a HFD and given 75 mg·kg− 1·day− 1 quercetin by gavage. Moreover, we created an in vitro GDM hepatocyte model in which BNL CL.2 cells were cultured in high glucose and treated with 16 µM quercetin. The results showed that quercetin significantly improved glucose tolerance in GDM mice, reduced serum total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and proprotein convertase subtilisin/kexin type 9 (PCSK9) levels. It also alleviated hepatic steatosis and pancreatic islet cell hypertrophy. Additionally, it improved the abnormal weights of the placenta and fetus. In in vitro experiments, quercetin regulated the PCSK9/LDLR axis and activated the PI3K/AKT/GSK3β signaling pathway, increasing glucose uptake in liver cells. Molecular docking experiments confirmed that quercetin could directly occupy the PCSK9 catalytic pocket, with binding energies of -9.056 kcal mol− 1 for PCSK9 (human), -9. 193 kcal mol− 1 for PCSK9(mouse) and − 9.1 kcal mol− 1 for EGF-A interface of LDLR, which would sterically hinder LDLR interaction. This study revealed that quercetin improves glucose and lipid metabolism disorders in a GDM model by regulating the PCSK9/LDLR axis and the PI3K/AKT/GSK3β pathway. These findings support the potential use of quercetin as a treatment strategy for GDM.

Preeclampsia (PE), a severe hypertensive disorder during pregnancy, significantly contributes to maternal and neonatal mortality. Existing prediction biomarkers are often invasive and expensive, hindering their widespread application. This study introduces PROMPT (Preeclampsia Risk factor + Ophthalmic data + Mean arterial pressure Prediction Test), an AI-driven model leveraging retinal photography for PE prediction, registered at ChiCTR (ChiCTR2100049850) in August 2021. Analyzing 1812 pregnancies before 14 gestational weeks, we extracted retinal parameters using a deep learning system. The PROMPT achieved an AUC of 0.87 (0.83-0.90) for PE prediction and 0.91 (0.85-0.97) for preterm PE prediction using machine learning, significantly outperforming the baseline model (p < 0.001). It also improved detection of severe adverse pregnancy outcomes from 35% to 41%. Economically, PROMPT was estimated to avert 1809 PE cases and saved over $50 million per 100,000 screenings. These results position PROMPT as a non-invasive and cost-effective tool for prenatal care, especially valuable in low- and middle-income countries.

Background Duration of second stage of labor is crucial for fetal delivery, but the optimal length of this stage remains controversial. While extending the duration of second stage can reduce primary cesarean delivery rates, it may increase maternal and neonatal morbidities as the duration progresses. We aimed to develop a personalized machine learning (ML) model to predict the possible second-stage duration. Methods This multicenter, retrospective study was conducted at four tertiary hospitals in China from September 2013 to October 2022. Data from three hospitals in Guangdong Province was selected as derivation set, and a geographically independent dataset from Fujian Province as the external validation set. Singleton vaginal deliveries with term live birth in a cephalic position were included. The primary outcome was the duration of the second stage of labor. Since durations beyond 3 h were rare, we developed binary classification models with thresholds at 1 h and 2 h. After the optimal features selected by recursive feature elimination (RFE) method, four ML algorithms were employed to build the models. The best model would be selected with the predictive performance and interpreted with Shapley Additive exPlanations method. The study is registered in Clinical Trial (ChiCTR2400085338). Findings Electronic medical records of 79,381 vaginal deliveries were obtained, and 63,401 deliveries meeting the inclusion criteria were included in the fi nal analysis. Eight risk features were selected through the RFE process. Gradient boosting machine implemented by decision tree models achieved the best performance, yielding areas under the curve for 1-h and 2-h models of 0.808 (95% confidence interval [CI] 0.797-0.819) and 0.824 (95% CI 0.804-0.843) in the testing set, and 0.862 (95% CI 0.854-0.870) and 0.859 (95% CI 0.843-0.875) in the external validation set, respectively. Interpretation An explainable and reliable ML model was developed to predict the probable second-stage duration, which could assist in individualized labor management. Factors such as fi rst-stage duration and maternal age are potential predictors for the second stage. Funding National Natural Science Foundation of China (No.82371689, N0.81771602), and National Key Research and Development Program of China (No.2021YFC2700703). Copyright (c) 2025 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction: Sarcopenia may affect the onset of gestational diabetes mellitus (GDM). However, the causal relationship between sarcopenia and GDM remains unclear. In this study, we used a bi-directional Mendelian randomization (MR) approach to explore this intricate relationship. Methods: This study utilized data from FinnGen datasets and genome-wide association studies. A bi-directional MR study was conducted. First, a forward MR analysis evaluated the causality of sarcopenia on GDM risk, with sarcopenia-related traits as exposures and GDM as the outcome. Second, in the reverse MR analysis, we assessed whether GDM influenced sarcopenia-related traits. Finally, sensitivity analysis was conducted to assess the robustness of the MR analysis. Results: Forward MR analysis revealed that appendicular lean mass (odds ratio [OR] = 1.2182, 95% confidence interval [CI]: 1.1397-1.3021, P < 0.0001), right-hand grip strength (OR= 1.4194, 95% CI: 1.0773-1.8701, P= 0.0128), left-hand grip strength (OR= 1.6064, 95% CI: 1.2829-2.0115, P < 0.0001), and usual walking pace (OR= 3.3676, 95% CI: 1.8769-6.0423, P < 0.0001) were associated with an increased risk of GDM. However, according to the reverse MR results, GDM had no causal effect on sarcopenia. No pleiotropy was observed. Conclusion: In summary, sarcopenia had a significant causal influence on GDM, while GDM did not causally affect sarcopenia.