Prediction of preeclampsia risk in first time pregnant women: Metabolite biomarkers for a clinical test

Presenting a prediction model for HELLP syndrome through data mining

BACKGROUND

The HELLP syndrome represents three complications: hemolysis, elevated liver enzymes, and low platelet count. Since the causes and pathogenesis of HELLP syndrome are not yet fully known and well understood, distinguishing it from other pregnancy-related disorders is complicated. Furthermore, late diagnosis leads to a delay in treatment, which challenges disease management. The present study aimed to present a machine learning (ML) attitude for diagnosing HELLP syndrome based on non-invasive parameters.

METHOD

This cross-sectional study was conducted on 384 patients in Tajrish Hospital, Tehran, Iran, during 2010-2021 in four stages. In the first stage, data elements were identified using a literature review and Delphi method. Then, patient records were gathered, and in the third stage, the dataset was preprocessed and prepared for modeling. Finally, ML models were implemented, and their evaluation metrics were compared.

RESULTS

A total of 21 variables were included in this study after the first stage. Among all the ML algorithms, multi-layer perceptron and deep learning performed the best, with an F1 score of more than 99%.In all three evaluation scenarios of 5fold and 10fold cross-validation, the K-nearest neighbors (KNN), random forest (RF), AdaBoost, XGBoost, and logistic regression (LR) had an F1 score of over 0.95, while this value was around 0.90 for support vector machine (SVM), and the lowest values were below 0.90 for decision tree (DT). According to the modeling output, some variables, such as platelet, gestational age, and alanine aminotransferase (ALT), were the most important in diagnosing HELLP syndrome.

CONCLUSION

The present work indicated that ML algorithms can be used successfully in the development of HELLP syndrome diagnosis models. Other algorithms besides DTs have an F1 score above 0.90. In addition, this study demonstrated that biomarker features (among all features) have the most significant impact on the diagnosis of HELLP syndrome.

Preeclampsia: A comprehensive review

Preeclampsia (PE) is a life-threatening disease of pregnancy and a prominent cause of neonatal and maternal mortality and morbidity. PE affects approximately 5-10% of pregnancies worldwide, posing significant risks to perinatal and maternal health. It is characterized by a variety of interconnected pathological cascades contributing to the stimulation of intravascular inflammation, oxidative stress (OS), endothelial cell activation, and syncytiotrophoblast stress that converge on a common pathway, ultimately resulting in disease progression. The present study was designed and executed to review the existing scientific literature, specifically focusing on the etiology (gestational diabetes mellitus and maternal obesity, insulin resistance, metabolic syndrome, maternal infection, periodontal disease, altered microbiome, and genetics), clinical presentations (hypertension, blood disorders, proteinuria, hepatic dysfunction, renal dysfunction, pulmonary edema, cardiac dysfunction, fetal growth restrictions, and eclampsia), therapeutic clinical biomarkers (creatinine, albuminuria, and cystatin C) along with their associations and mechanisms in PE. In addition, this study provides insights into the potential of nanomedicines for targeting these mechanisms for PE management and treatment. Inflammation, OS, proteinuria, and an altered microbiome are prominent biomarkers associated with progression and PE-related pathogenesis. Understanding the molecular mechanisms, exploring suitable markers, targeted interventions, comprehensive screening, and holistic strategies are critical to decreasing the incidence of PE and promoting maternal-fetal well-being. The present study comprehensively reviewed the etiology, clinical presentations, therapeutic biomarkers, and preventive potential of nanomedicines in the treatment and management of PE.

Cohort profile: Improved Pregnancy Outcomes via Early Detection (IMPROvED), an International Multicentre Prospective Cohort

Background

Improved Pregnancy Outcomes via Early Detection (IMPROvED) is a multi-centre, European phase IIa clinical study. The primary aim of IMPROvED is to enable the assessment and refinement of innovative prototype preeclampsia risk assessment tests based on emerging biomarker technologies. Here we describe IMPROvED's profile and invite researchers to collaborate.

Methods

A total of 4,038 low-risk nulliparous singleton pregnancies were recruited from maternity units in Ireland (N=1,501), United Kingdom (N=1,108), The Netherlands (N=810), and Sweden (N=619) between November 2013 to August 2017. Participants were interviewed by a research midwife at ~11 weeks (optional visit), ~15 weeks, ~20 weeks, ~34 weeks' gestation (optional visit), and postpartum (within 72-hours following delivery).

Findings to date

Clinical data included information on maternal sociodemographic, medical history, and lifestyle factors collected at ~15 weeks' gestation, and maternal measurements, collected at each study visit. Biobank samples included blood, urine, and hair collected at each study visit throughout pregnancy in all units plus umbilical cord/blood samples collected at birth in Ireland and Sweden. A total of 74.0% (N=2,922) had an uncomplicated pregnancy, 3.1% (N=122) developed preeclampsia, 3.6% (N=143) had a spontaneous preterm birth, and 10.5% (N=416) had a small for gestational age baby. We evaluated a panel of metabolite biomarkers and a panel of protein biomarkers at 15 weeks and 20 weeks' gestation for preeclampsia risk assessment. Their translation into tests with clinical application, as conducted by commercial entities, was hampered by technical issues and changes in test requirements. Work on the panel of proteins was abandoned, while work on the use of metabolite biomarkers for preeclampsia risk assessment is ongoing.

Future plans

In accordance with the original goals of the IMPROvED study, the data and biobank are now available for international collaboration to conduct high quality research into the cause and prevention of adverse pregnancy outcomes.

Relationship between the concentration of ergothioneine in plasma and the likelihood of developing pre-eclampsia

Ergothioneine, an antioxidant nutraceutical mainly at present derived from the dietary intake of mushrooms, has been suggested as a preventive for pre-eclampsia (PE). We analysed early pregnancy samples from a cohort of 432 first time mothers as part of the Screening for Endpoints in Pregnancy (SCOPE, European branch) project to determine the concentration of ergothioneine in their plasma. There was a weak association between the ergothioneine levels and maternal age but none for BMI. Of these 432 women, 97 went on to develop pre-term (23) or term (74) PE. If a threshold was set at the 90th percentile of the reference range in the control population (≥462 ng/ml), only one of these 97 women (1%) developed PE, versus 96/397 (24.2%) whose ergothioneine level was below this threshold. One possible interpretation of these findings, consistent with previous experiments in a reduced uterine perfusion model in rats, is that ergothioneine may indeed prove protective against PE in humans. An intervention study of some kind now seems warranted.

First-trimester preterm preeclampsia prediction with metabolite biomarkers: differential prediction according to maternal body mass index

BACKGROUND

Prediction of preeclampsia risk is key to informing effective maternal care. Current screening for preeclampsia at 11 to 13 weeks of gestation using maternal demographic characteristics and medical history with measurements of mean arterial pressure, uterine artery pulsatility index, and serum placental growth factor can identify approximately 75% of women who develop preterm preeclampsia with delivery at <37 weeks of gestation. Further improvements to preeclampsia screening tests will likely require integrating additional biomarkers. Recent research suggests the existence of distinct maternal risk profiles. Therefore, biomarker evaluation should account for the possibility that a biomarker only predicts preeclampsia in a specific maternal phenotype.

OBJECTIVE

This study aimed to verify metabolite biomarkers as preterm preeclampsia predictors early in pregnancy in all women and across body mass index groups.

STUDY DESIGN

Observational case-control study drawn from a large prospective study on the early prediction of pregnancy complications in women attending their routine first hospital visit at King's College Hospital, London, United Kingdom, in 2010 to 2015. Pregnant women underwent a complete first-trimester assessment, including the collection of blood samples for biobanking. In 11- to 13-week plasma samples of 2501 singleton pregnancies, the levels of preselected metabolites implicated in the prediction of pregnancy complications were analyzed using a targeted liquid chromatography-mass spectrometry method, yielding high-quality quantification data on 50 metabolites. The ratios of amino acid levels involved in arginine biosynthesis and nitric oxide synthase pathways were added to the list of biomarkers. Placental growth factor and pregnancy-associated plasma protein A were also available for all study subjects, serving as comparator risk predictors. Data on 1635 control and 106 pregnancies complicated by preterm preeclampsia were considered for this analysis, normalized using multiples of medians. Prediction analyses were performed across the following patient strata: all subjects and the body mass index classes of <25, 25 to <30, and ≥30 kg/m². Adjusted median levels were compared between cases and controls and between each body mass index class group. Odds ratios and 95% confidence intervals were calculated at the mean ±1 standard deviation to gauge clinical prediction merits.

RESULTS

The levels of 13 metabolites were associated with preterm preeclampsia in the entire study population (P<.05) with particularly significant (P<.01) associations found for 6 of them, namely, 2-hydroxy-(2/3)-methylbutyric acid, 25-hydroxyvitamin D3, 2-hydroxybutyric acid, alanine, dodecanoylcarnitine, and 1-(1Z-octadecenyl)-2-oleoyl-sn-glycero-3-phosphocholine. Fold changes in 7 amino acid ratios, all involving glutamine or ornithine, were also significantly different between cases and controls (P<.01). The predictive performance of some metabolites and ratios differed according to body mass index classification; for example, ornithine (P<.001) and several ornithine-related ratios (P<.0001 to P<.01) were only strongly associated with preterm preeclampsia in the body mass index of <25 kg/m² group, whereas dodecanoylcarnitine and 3 glutamine ratios were particularly predictive in the body mass index of ≥30 kg/m² group (P<.01).

CONCLUSION

Single metabolites and ratios of amino acids related to arginine bioavailability and nitric oxide synthase pathways were associated with preterm preeclampsia risk at 11 to 13 weeks of gestation. Differential prediction was observed according to body mass index classes, supporting the existence of distinct maternal risk profiles. Future studies in preeclampsia prediction should account for the possibility of different maternal risk profiles to improve etiologic and prognostic understanding and, ultimately, clinical utility of screening tests.

Early cost-effectiveness analysis of screening for preeclampsia in nulliparous women: A modelling approach in European high-income settings

Background

Preeclampsia causes substantial maternal and perinatal morbidity and mortality and significant societal economic impact. Effective screening would facilitate timely and appropriate prevention and management of preeclampsia.

Objectives

To develop an early cost-effectiveness analysis to assess both costs and health outcomes of a new screening test for preeclampsia from a healthcare payer perspective, in the United Kingdom (UK), Ireland, the Netherlands and Sweden.

Methods

A decision tree over a 9-month time horizon was developed to explore the cost-effectiveness of the new screening test for preeclampsia compared to the current screening strategy. The new test strategy is being developed so that it can stratify healthy low risk nulliparous women early in pregnancy to either a high-risk group with a risk of 1 in 6 or more of developing preeclampsia, or a low-risk group with a risk of 1 in 100 or less. The model simulated 25 plausible scenarios in a hypothetical cohort of 100,000 pregnant women, in which the sensitivity and specificity of the new test were varied to set a benchmark for the minimum test performance that is needed for the test to become cost-effective. The input parameters and costs were mainly derived from published literature. The main outcome was incremental costs per preeclampsia case averted, expressed as an incremental cost-effectiveness ratio (ICER). Deterministic and probabilistic sensitivity analyses were conducted to assess uncertainty.

Results

Base case results showed that the new test strategy would be more effective and less costly compared to the current situation in the UK. In the Netherlands, the majority of scenarios would be cost-effective from a threshold of €50,000 per preeclampsia case averted, while in Ireland and Sweden, the vast majority of scenarios would be considered cost-effective only when a threshold of €100,000 was used. In the best case analyses, ICERs were more favourable in all four participating countries. Aspirin effectiveness, prevalence of preeclampsia, accuracy of the new screening test and cost of regular antenatal care were identified as driving factors for the cost-effectiveness of screening for preeclampsia.

Conclusion

The results indicate that the new screening test for preeclampsia has potential to be cost-effective. Further studies based on proven accuracy of the test will confirm whether the new screening test is a cost-effective additional option to the current situation.

Identification of Biomarkers for Preeclampsia Based on Metabolomics

Background

Preeclampsia (PE) is a significant cause of maternal and neonatal morbidity and mortality worldwide. However, the pathogenesis of PE is unclear and reliable early diagnostic methods are still lacking. The purpose of this review is to summarize potential metabolic biomarkers and pathways of PE, which might facilitate risk prediction and clinical diagnosis, and obtain a better understanding of specific metabolic mechanisms of PE.

Methods

This review included human metabolomics studies related to PE in the PubMed, Google Scholar, and Web of Science databases from January 2000 to November 2021. The reported metabolic biomarkers were systematically examined and compared. Pathway analysis was conducted through the online software MetaboAnalyst 5.0.

Results

Forty-one human studies were included in this systematic review. Several metabolites, such as creatinine, glycine, L-isoleucine, and glucose and biomarkers with consistent trends (decanoylcarnitine, 3-hydroxyisovaleric acid, and octenoylcarnitine), were frequently reported. In addition, eight amino acid metabolism-related, three carbohydrate metabolism-related, one translation-related and one lipid metabolism-related pathways were identified. These biomarkers and pathways, closely related to renal dysfunction, insulin resistance, lipid metabolism disorder, activated inflammation, and impaired nitric oxide production, were very likely to contribute to the progression of PE.

Conclusion

This study summarized several metabolites and metabolic pathways, which may be associated with PE. These high-frequency differential metabolites are promising to be biomarkers of PE for early diagnosis, and the prominent metabolic pathway may provide new insights for the understanding of the pathogenesis of PE.