First-trimester metabolomic prediction of stillbirth

Differences in Factors Associated With Preterm and Term Stillbirth: A Secondary Cohort Analysis of the DESiGN Trial

OBJECTIVE

To identify whether maternal and pregnancy characteristics associated with stillbirth differ between preterm and term stillbirth.

DESIGN

Secondary cohort analysis of the DESiGN RCT.

SETTING

Thirteen UK maternity units.

POPULATION

Singleton pregnant women and their babies.

METHODS

Multiple logistic regression was used to assess whether the 12 factors explored were associated with stillbirth. Interaction tests assessed for a difference in these associations between the preterm and term periods.

MAIN OUTCOME MEASURE

Stillbirth stratified by preterm (<37+0 weeks') and term (37+0-42+6 weeks') births.

RESULTS

A total of 195 344 pregnancies were included. Six hundred and sixty-seven were stillborn (3.4 per 1000 births), of which 431 (65%) were preterm. Significant interactions were observed for maternal age, ethnicity, IMD, BMI, parity, smoking, PAPP-A, gestational hypertension, pre-eclampsia and gestational diabetes but not for chronic hypertension and pre-existing diabetes. Stronger associations with term stillbirth were observed in women with obesity compared to BMI 18.5-24.9 kg/m2 (BMI 30.0-34.9 kg/m2 term adjusted OR 2.1 [95% CI 1.4-3.0] vs. preterm aOR 1.1 [0.8-1.7]; BMI ≥ 35.0 kg/m2 term aOR 2.2 [1.4-3.4] vs. preterm aOR 1.5 [1.2-1.8]; p-interaction < 0.01), nulliparity compared to parity 1 (term aOR 1.7 [1.1-2.7] vs. preterm aOR 1.2 [0.9-1.6]; p-interaction < 0.01) and Asian ethnicity compared with White (p-interaction < 0.01). A weaker or lack of association with term, compared to preterm, stillbirth was observed for older maternal age, smoking and pre-eclampsia.

CONCLUSION

Differences in association exist between mothers experiencing preterm and term stillbirth. These differences could contribute to design of timely surveillance and interventions to further mitigate the risk of stillbirth.

Artificial intelligence-driven predictive framework for early detection of still birth

Predictive modeling is becoming increasingly popular in the context of early disease detection. The use of machine learning approaches for predictive modeling can help early detection of diseases thereby enabling medical experts to appropriate medical treatments. Stillbirth prediction is a similar domain where artificial intelligence-based predictive modeling can alleviate this significant global health challenge. Despite advancements in prenatal care, the prevention of stillbirths remains a complex issue requiring further research and interventions. The cardiotocography (CTG) dataset is used in this research work from the UCI machine learning (ML) repository to investigate the efficiency of the proposed approach regarding stillbirth prediction. This research work adopts the Tabular Prior Data Fitted Network (TabPFN) model which was originally designed to solve small tabular classification. TabPFN is used to predict the still or live birth during pregnancy with 97.91% accuracy. To address this life-saving problem with more accurate results and in-depth analysis of ML models, this research work makes use of 13 renowned ML models for performance comparison with the proposed model. The proposed model is evaluated using precision, recall, F-score, Mathews Correlation Coefficient (MCC), and the area under the curve evaluation parameters and the results are 97.87%, 98.26%, 98.05%, 96.42%, and 98.88%, respectively. The results of the proposed model are further evaluated using k-fold cross-validation and its performance comparison with other state-of-the-art studies indicating the superior performance of TabPFN model.

First-Trimester Screening for Miscarriage or Stillbirth-Prediction Model Based on MicroRNA Biomarkers

We evaluated the potential of cardiovascular-disease-associated microRNAs to predict in the early stages of gestation (from 10 to 13 gestational weeks) the occurrence of a miscarriage or stillbirth. The gene expressions of 29 microRNAs were studied retrospectively in peripheral venous blood samples derived from singleton Caucasian pregnancies diagnosed with miscarriage (n = 77 cases; early onset, n = 43 cases; late onset, n = 34 cases) or stillbirth (n = 24 cases; early onset, n = 13 cases; late onset, n = 8 cases; term onset, n = 3 cases) and 80 selected gestational-age-matched controls (normal term pregnancies) using real-time RT-PCR. Altered expressions of nine microRNAs (upregulation of miR-1-3p, miR-16-5p, miR-17-5p, miR-26a-5p, miR-146a-5p, and miR-181a-5p and downregulation of miR-130b-3p, miR-342-3p, and miR-574-3p) were observed in pregnancies with the occurrence of a miscarriage or stillbirth. The screening based on the combination of these nine microRNA biomarkers revealed 99.01% cases at a 10.0% false positive rate (FPR). The predictive model for miscarriage only was based on the altered gene expressions of eight microRNA biomarkers (upregulation of miR-1-3p, miR-16-5p, miR-17-5p, miR-26a-5p, miR-146a-5p, and miR-181a-5p and downregulation of miR-130b-3p and miR-195-5p). It was able to identify 80.52% cases at a 10.0% FPR. Highly efficient early identification of later occurrences of stillbirth was achieved via the combination of eleven microRNA biomarkers (upregulation of miR-1-3p, miR-16-5p, miR-17-5p, miR-20a-5p, miR-146a-5p, and miR-181a-5p and downregulation of miR-130b-3p, miR-145-5p, miR-210-3p, miR-342-3p, and miR-574-3p) or, alternatively, by the combination of just two upregulated microRNA biomarkers (miR-1-3p and miR-181a-5p). The predictive power achieved 95.83% cases at a 10.0% FPR and, alternatively, 91.67% cases at a 10.0% FPR. The models based on the combination of selected cardiovascular-disease-associated microRNAs had very high predictive potential for miscarriages or stillbirths and may be implemented in routine first-trimester screening programs.

Impact of combined consumption of fish oil and probiotics on the serum metabolome in pregnant women with overweight or obesity

BACKGROUND

If a pregnant woman is overweight, this can evoke metabolic alterations that may have health consequences for both mother and child.

METHODS

Pregnant women with overweight/obesity (n = 358) received fish oil+placebo, probiotics+placebo, fish oil+probiotics or placebo+placebo from early pregnancy onwards. The serum metabolome was analysed from fasting samples with a targeted NMR-approach in early and late pregnancy. GDM was diagnosed by OGTT.

FINDINGS

The intervention changed the metabolic profile of the women, but the effect was influenced by their GDM status. In women without GDM, the changes in nine lipids (FDR<0.05) in the fish oil+placebo-group differed when compared to the placebo+placebo-group. The combination of fish oil and probiotics induced changes in more metabolites, 46 of the lipid metabolites differed in women without GDM when compared to placebo+placebo-group; these included reduced increases in the concentrations and lipid constituents of VLDL-particles and less pronounced alterations in the ratios of various lipids in several lipoproteins. In women with GDM, no differences were detected in the changes of any metabolites due to any of the interventions when compared to the placebo+placebo-group (FDR<0.05).

INTERPRETATION

Fish oil and particularly the combination of fish oil and probiotics modified serum lipids in pregnant women with overweight or obesity, while no such effects were seen with probiotics alone. The effects were most evident in the lipid contents of VLDL and LDL only in women without GDM.

FUNDING

State Research Funding for university-level health research in the Turku University Hospital Expert Responsibility Area, Academy of Finland, the Diabetes Research Foundation, the Juho Vainio Foundation, Janssen Research & Development, LLC.

Proposing a machine-learning based method to predict stillbirth before and during delivery and ranking the features: nationwide retrospective cross-sectional study

Background

Stillbirth is defined as fetal loss in pregnancy beyond 28 weeks by WHO. In this study, a machine-learning based method is proposed to predict stillbirth from livebirth and discriminate stillbirth before and during delivery and rank the features.

Method

A two-step stack ensemble classifier is proposed for classifying the instances into stillbirth and livebirth at the first step and then, classifying stillbirth before delivery from stillbirth during the labor at the second step. The proposed SE has two consecutive layers including the same classifiers. The base classifiers in each layer are decision tree, Gradient boosting classifier, logistics regression, random forest and support vector machines which are trained independently and aggregated based on Vote boosting method. Moreover, a new feature ranking method is proposed in this study based on mean decrease accuracy, Gini Index and model coefficients to find high-ranked features.

Results

IMAN registry dataset is used in this study considering all births at or beyond 28th gestational week from 2016/04/01 to 2017/01/01 including 1,415,623 live birth and 5502 stillbirth cases. A combination of maternal demographic features, clinical history, fetal properties, delivery descriptors, environmental features, healthcare service provider descriptors and socio-demographic features are considered. The experimental results show that our proposed SE outperforms the compared classifiers with the average accuracy of 90%, sensitivity of 91%, specificity of 88%. The discrimination of the proposed SE is assessed and the average AUC of ±95%, CI of 90.51% ±1.08 and 90% ±1.12 is obtained on training dataset for model development and test dataset for external validation, respectively. The proposed SE is calibrated using isotopic nonparametric calibration method with the score of 0.07. The process is repeated 10,000 times and AUC of SE classifiers using random different training datasets as null distribution. The obtained p-value to assess the specificity of the proposed SE is 0.0126 which shows the significance of the proposed SE.

Conclusions

Gestational age and fetal height are two most important features for discriminating livebirth from stillbirth. Moreover, hospital, province, delivery main cause, perinatal abnormality, miscarriage number and maternal age are the most important features for classifying stillbirth before and during delivery.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-021-03658-z.

Metabolic infrastructure of pregnant women with methylenetetrahydrofolate reductase polymorphisms: A metabolomic analysis

The aim of this study was to demonstrate the altered metabolic infrastructure of pregnant women with methylenetetrahydrofolate reductase (MTHFR) polymorphisms at first trimester and during delivery. Eight singleton pregnant women with MTHFR polymorphisms were compared with 10 normal pregnant women. Maternal blood samples were obtained twice during their pregnancy period (between the 11th and 14th gestational weeks and during delivery). Metabolomic analysis was performed using GC-MS. The GC-MS based metabolomic profile helped identify 95 metabolites in the plasma samples. In the MTHFR group, the levels of 1-monohexadecanoylglycerol, pyrophosphate, benzoin, and linoleic acid significantly decreased (P ˂ 0.05 for all), whereas the levels of glyceric acid, l-tryptophan, l-alanine, l-proline, norvaline, l-threonine, and myo-inositol significantly increased (P ˂ 0.01 for the first two metabolites, P ˂ 0.05 for the others) at 11-14 gestational weeks. Conversely, the levels of benzoin, 1-monohexadecanoylglycerol, pyruvic acid, l-proline, phosphoric acid, epsilon-caprolactam, and pipecolic acid significantly decreased in the MTHFR group, whereas metabolites such as hexadecanoic acid and 2-hydroxybutyric acid increased significantly in the study group during delivery. An impaired energy metabolism pathway, vitamin B complex disorders, tendency for metabolic acidosis (oxidative stress), and the need for cell/tissue support seem prevalent in pregnancies with MTHFR polymorphisms.

Fungal and plant metabolites in industrially-processed fruit juices in Nigeria

There is scarce data on the mycotoxin profile in retailed fruit juices in Nigeria. Thirty-five industrially-processed fruit juice samples randomly purchased from retailers in Ogun state, Nigeria, were analysed for the presence of > 650 toxic fungal and plant metabolites using a liquid chromatography tandem mass spectrometric method. Only 18 metabolites, including 3-nitropropionic acid, alternariol methylether and emodin, but excluding citrinin, fumonisin B₂, ochratoxin A and patulin, were detected in trace levels in at least one juice sample. Amygdalin, a plant cyanogen, was quantified (2.05-359 µg/L) in 40% of the samples. Although the levels of mycotoxins and toxic plant metabolites found in the juice may be relatively low, daily consumption of juices containing such low levels may contribute to dietary exposures to these natural chemical contaminants in consumers. Fruit juice processors should be encouraged to adhere strictly to good manufacturing practices in order to keep mycotoxins away from the final products.