Lipidomic signatures in maternal blood and placenta: Systematic evidence linking lipid profiles to pregnancy outcomes and fetal growth

Abstract

Introduction

Pregnancy involves profound metabolic adaptations, particularly in lipid metabolism. Disruptions in lipid profiles have been implicated in obstetric complications, yet no systematic synthesis of lipidomic evidence exists across diverse pregnancy outcomes. The primary objective of this review was to identify and synthesize consistent lipid species and lipid classes across studies of maternal and fetal biospecimens, in relation to adverse pregnancy outcomes.

The secondary objective was to evaluate the potential predictive and prognostic utility of lipidomic signatures for risk stratification, including their capacity to provide early biomarkers and to inform postpartum metabolic surveillance.

Methods

A systematic search of five databases (PubMed, Scopus, EMBASE, Web of Science, Cochrane Library) was conducted up to 2024. Eligible studies involved pregnant or postpartum women undergoing lipidomic analysis by mass spectrometry or nuclear magnetic resonance (NMR), with outcomes including preeclampsia, gestational diabetes (GDM), preterm birth, fetal growth restriction (FGR), and postpartum metabolic risk. Data extraction and quality assessment were performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and the Newcastle–Ottawa Scale quality assessment tool.

Results

Sixteen studies with over 5000 participants were included, spanning multiple countries and employing targeted or untargeted lipidomic methods. Maternal plasma, serum, placenta, and cord blood were commonly analyzed. Lipidomic alterations were consistently observed in relation to preeclampsia, GDM, FGR, and congenital anomalies. Key lipid classes included phosphatidylcholines, phosphatidylethanolamines, triglycerides, ceramides, cholesteryl esters, and diacylglycerols. These lipids mapped to metabolic pathways involving insulin resistance, inflammation, and endothelial dysfunction. Some profiles were detectable in early pregnancy, indicating potential for early risk prediction.

Conclusions

Lipidomic profile analysis can reveal consistent metabolic disruptions associated with certain perinatal outcomes (such as preeclampsia, gestational diabetes, preterm birth, fetal intrauterine growth restriction, postpartum cardiometabolic risk and congenital anomalies). Lipidomic profiling reveals consistent metabolic disruptions across pregnancy complications. Lipid signatures hold promise as early biomarkers for obstetric risk stratification and may inform postpartum metabolic surveillance. Standardized approaches and mechanistic validation are needed to support clinical translation.

Abbreviations

CHD

Congenital Heart Disease

CE

Cholesteryl Ester

DAG

Diacylglycerol

FGR

Fetal Growth Restriction

GDM

Gestational Diabetes Mellitus

LDL

Low-Density Lipoprotein

LPC

Lysophosphatidylcholine

NMR

Nuclear Magnetic Resonance

NOS

Newcastle–Ottawa Scale

PC

Phosphatidylcholine

PE

Phosphatidylethanolamine

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

SM

Sphingomyelin

TG

Triglyceride

LC–MS/MS

Liquid Chromatography Tandem Mass Spectrometry

1. Introduction

Pregnancy represents a period of profound physiological transformation, characterized by extensive metabolic adaptations that ensure optimal fetal development and maternal well-being. Among these adaptations, lipid metabolism undergoes dynamic remodeling to meet the increasing energy and biosynthetic demands of both the mother and the fetus. Beyond their structural role in cell membranes, lipids act as crucial mediators of energy homeostasis, inflammation, and intracellular signaling, making them central to pregnancy physiology [1,2].

Traditional lipid measures—such as total cholesterol, triglycerides, and low-density lipoprotein (LDL)—have been linked to pregnancy complications. However, these conventional parameters offer only a limited view of the complex lipid networks involved in gestational processes. Advances in mass spectrometry-based lipidomics now enable high-throughput, high-resolution profiling of hundreds of lipid species spanning diverse classes, including glycerophospholipids, sphingolipids, and neutral lipids [3,4]. These tools provide a deeper understanding of lipid alterations across gestation and their association with adverse maternal and fetal outcomes.

Emerging lipidomic studies have uncovered distinct molecular signatures associated with pregnancy complications. For instance, dysregulation of ceramides and oxidized phospholipids has been implicated in endothelial dysfunction and systemic inflammation characteristic of preeclampsia [2]. In cases of late-onset preeclampsia, elevated triglycerides and disrupted glycerophospholipid metabolism have been consistently reported [5,6]. Similarly, early gestational lipidomic profiles marked by increased diglycerides and reduced cholesteryl esters (CEs) have been predictive of later gestational diabetes mellitus (GDM) onset [3]. These findings not only offer mechanistic insight but also reveal promising early biomarkers for risk stratification.

Moreover, lipidomic alterations are not confined to the gestational period; studies suggest that lipid disturbances may extend into the postpartum phase and influence long-term cardiometabolic risk for both the mother and the child [7,8]. Additionally, maternal lipid profiles have been linked with fetal lipid exposure and outcomes such as intrauterine growth restriction and congenital heart defects, emphasizing the transgenerational implications of maternal lipid metabolism [4,9].

Despite a growing body of evidence, a comprehensive synthesis of lipidomic research across diverse pregnancy outcomes is lacking. This systematic review was designed with dual aims. The primary objective was to evaluate the translational potential of lipidomic signatures as early biomarkers for adverse pregnancy outcomes and postpartum metabolic risk, with a view to informing clinical risk stratification and personalized care. The secondary objective was to synthesize consistent lipid species and lipid classes across studies, thereby offering mechanistic insights into the metabolic pathways underlying obstetric complications.

2. Materials and methods

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [10]. The review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO), registration number: CRD420251064480.

2.1. Search strategy

A systematic and comprehensive search strategy was developed to identify studies investigating lipidomic profiles in the context of pregnancy outcomes. To maximize retrieval sensitivity and specificity, the strategy incorporated both controlled vocabulary (e.g., MeSH terms) and free-text keywords. The search framework encompassed three primary conceptual domains: (1) lipidomics, including terms such as “lipidomics”, “lipid profiling”, “lipid signature”, and “lipid metabolism”; (2) pregnancy-related concepts, including “pregnancy”, “maternal”, “gestation” and “pregnant women”; and (3) adverse pregnancy outcomes, such as “preeclampsia,” “gestational diabetes”, “fetal growth restriction”, “preterm birth” and “birth outcome’’. This strategy was adapted for use across five electronic databases—PubMed/MEDLINE, Scopus, EMBASE, Web of Science, and the Cochrane Library—without time restrictions. Reference lists of included articles and relevant reviews were also manually screened to identify additional eligible studies. The electronic search was last updated in June 2025. Full database-specific search strings, including Boolean operators and applied filters, are provided in Supplementary File 1 to ensure reproducibility.

2.2. Inclusion and exclusion criteria

Studies were included based on the following predefined criteria: a) Population: Pregnant women at any gestational stage or postpartum women, with or without pregnancy-related complications; b) Exposure: Lipidomic profiling performed using mass spectrometry (MS) or nuclear magnetic resonance (NMR) techniques, applied to maternal or fetal biospecimens (e.g., plasma, serum, placenta, or cord blood); c) Outcomes: Any reported association between lipidomic profiles and pregnancy-related outcomes, including—but not limited to—preeclampsia, GDM, fetal growth restriction (FGR), preterm birth, congenital anomalies, or postpartum cardiometabolic risk; d) Study Design: Original research articles reporting observational (cohort, case-control, cross-sectional) or interventional studies; e) Language: Articles published in English and f) Publication Type: Peer-reviewed journal articles.

Reviews, editorials, conference abstracts, case reports, and animal studies were excluded.

2.3. PRISMA process

The study selection process followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Two reviewers independently screened the titles and abstracts of all records retrieved through the database search against the predefined eligibility criteria. Full-text articles were then obtained for studies deemed potentially relevant, and these were assessed independently by the same reviewers for final inclusion. Any discrepancies or disagreements regarding study eligibility were resolved through discussion. When consensus could not be reached, a third reviewer was consulted to achieve resolution.

A total of 650 records were identified through systematic searches of five electronic databases (PubMed/MEDLINE, Scopus, EMBASE, Web of Science, and the Cochrane Library). After removal of 123 duplicates, 527 unique records remained for screening.

Titles and abstracts of these 527 records were screened independently by two reviewers. Of these, 456 records were excluded based on irrelevance to the topic, lack of lipidomic analysis, or non-human/non-original research articles. The full texts of the remaining 71 articles were retrieved for detailed assessment.

Following full-text review, 56 articles were excluded for the following reasons: a) Ineligible population (n = 14); b) Inappropriate or unclear lipidomic methodology (n = 12); c) No relevant pregnancy outcome reported (n = 18) and d) publication type not meeting inclusion criteria (e.g., conference abstracts, reviews) (n = 11).

Ultimately, 16 studies met all the inclusion criteria and were included in the final systematic review. The full study selection process is illustrated in the PRISMA flow diagram (Fig. 1).

Fig. 1.

PRISMA flow diagram of study selection.

Flow diagram depicting the process of identification, screening, eligibility assessment, and inclusion of studies in accordance with PRISMA 2020 guidelines. Out of 650 records identified, 527 were screened after removal of duplicates. Following exclusion of 456 records and full-text assessment of 71 reports, 16 studies met the inclusion criteria and were included in the final review.

2.4. Quality assessment

The quality and risk of bias of included studies were assessed using the Newcastle–Ottawa Scale (NOS) for observational studies [11]. Each study was independently rated by two reviewers, with disagreements resolved by consensus. Quality assessment using the NOS indicated that all included studies were of moderate to high methodological quality. Of the 16 studies, 5 achieved the maximum score of 9/9 stars, 8 scored 8/9 stars, and 3 scored 7/9 stars. No study scored below 7, reflecting overall strong methodological rigor across the evidence base. Full details are provided in Supplementary File 2.

2.5. Data extraction

Data extraction was performed independently by two reviewers using a standardized form. Extracted data included: a) Author(s), year of publication, country; b) Study design and sample size; c) Characteristics of the study population; d) Biospecimen type and timing of collection; e) Analytical platform and lipidomic approach (targeted/untargeted) and f) Main outcomes and findings (including lipid classes/species, associated pathways, statistical models).

2.6. Data synthesis

Due to heterogeneity in study design, population characteristics, biospecimen types, and analytical platforms, a qualitative synthesis approach was employed. Studies were grouped thematically based on the type of pregnancy outcome investigated. Key lipid species and lipid classes were identified across studies, and recurring findings were highlighted to evaluate consistency and potential translational value.

3. Results

A total of 16 studies [[1], [2], [3], [4], [5], [6],9,[12], [13], [14], [15], [16], [17], [18], [19], [20]] were included in this systematic review, encompassing research conducted across diverse geographical regions including the United States, China, the United Kingdom, Australia, Singapore, Serbia, Romania, Finland, and Spain. Collectively, these studies involved over 5000 participants, with individual sample sizes ranging from 40 to 1595 women. The included studies employed various designs, predominantly prospective cohort and nested case-control studies, alongside longitudinal, observational, and cross-sectional analyses. Biospecimens analyzed comprised maternal plasma or serum, placental tissue, fetal cord blood, and postpartum samples, collected at different time points across pregnancy and postpartum. Analytical methods were primarily based on liquid chromatography-mass spectrometry (LC-MS/MS), using either untargeted or targeted lipidomic approaches; some studies incorporated additional platforms such as ˆ1H NMR spectroscopy. Lipidomic profiling enabled the identification of significant alterations in lipid species and networks associated with a wide range of adverse pregnancy and postpartum outcomes, including early- and late-onset preeclampsia, GDM, preterm birth, FGR, small for gestational age (SGA), congenital heart disease (CHD), and future risk of type 2 diabetes. Across studies, lipidomic signatures revealed disruptions in phospholipid metabolism, ceramide pathways, and triglyceride profiles, offering novel biomarkers and mechanistic insights for early risk stratification and understanding the pathophysiology of pregnancy-related complications. A summary of the included studies is provided in Table 1.

Table 1.

Summary of the included studies.

Author(s), Year, Country	Study Design and Sample Size	Study Population Characteristics	Biospecimen Type and Timing	Analytical Platform and Lipidomic Approach	Main Outcomes and Findings
He et al. (2021), USA [2]	Nested case-control study; N = 64 (44 severe preeclampsia, 20 controls)	Multiethnic pregnant women in Hawaii; significant differences in gestational age, gestational diabetes, and parity; only cases had smokers or gestational diabetes	Maternal plasma; collected during pregnancy (biobank-derived samples)	Untargeted lipidomics using LC-MS/MS (TripleTOF 5600); LipidBlast database; WGCNA; machine learning (e.g., random forest)	11 lipid species associated with severe preeclampsia (e.g., LPC 15:0, PC 35:1e, PE 37:2); enriched pathways include insulin signaling, immune response, PL metabolism; RF model: F1 = 0.94, accuracy = 0.88
Yang et al. (2023), China [5]	Case-control study; N = 40 (20 LOPE, 20 controls)	Singleton pregnancies; exclusion of pre-existing conditions; significant differences in BMI, BP, neonatal weight; matched for age, parity, gestational age	Placental tissue; collected immediately post-delivery (within 5 min), snap frozen	Untargeted LC-MS/MS (UHPLC-Q-Exactive Orbitrap MS); XCMS + LipidBlast; WGCNA; OPLS-DA; MetaboAnalyst	226 differentially expressed lipids (94 ↑, 132 ↓); LOPE linked to increased TGs (esp. Unsaturated FAs) and decreased PC, PE, PS; key pathway: glycerophospholipid metabolism; co-expression modules correlated with SBP and BMI
Mires et al. (2024), UK [9]	Observational study; maternal plasma from 98 CHD and 62 control pregnancies	Mothers of children with congenital heart disease (CHD); samples collected postnatally; matched on age and BMI; exclusion of pre-existing metabolic disorders	Maternal plasma; collected 6 weeks postpartum	Untargeted metabolomics and lipidomics via LC-MS (positive and negative ion modes); analyzed using multivariate statistics (PCA, OPLS-DA), machine learning classifiers	Distinct metabolic/lipidomic signatures differentiate CHD vs. control mothers; key lipid species include phosphatidylcholines and acylcarnitines; achieved high classification accuracy using integrated profiles
Huang et al. (2024), China [13]	Case-control; N = 58 (20 controls, 19 EOPE, 19 LOPE)	Pregnant women; EOPE and LOPE classified by gestational age; matched controls; exclusion criteria applied	Maternal plasma samples collected during pregnancy	Untargeted lipidomics using UPLC-MS/MS; multivariate analysis; WGCNA; ROC curve analysis	Strong lipidomic separation between EOPE, LOPE, and controls; AUC = 1.000 (EOPE vs. control), AUC = 0.992 (LOPE vs. control); lipids linked to glycerophospholipid metabolism and correlated with fetal birth weight and urine protein
Alahakoon et al. (2020), Australia	Prospective cross-sectional case-control study	Pregnant women with PE, FGR, PE + FGR, and normal controls; FGR defined by elevated umbilical artery resistance and EFW <10 %	Maternal and fetal serum; collected at delivery	Enzymatic and immunoturbidimetric assays for TC, HDL, LDL, TG, ApoA1, ApoB	Elevated maternal TG in PE; increased fetal TG in PE + FGR; significantly higher fetal ApoB in PE, FGR, PE + FGR; no group differences in TC, HDL, LDL or ApoA1
Chen et al. (2023), Singapore [1]	Prospective longitudinal cohort (S-PRESTO); N = 1595 plasma samples from 976 women	Healthy women from preconception to postpartum; followed through pregnancy and 3 months postpartum	Plasma samples at 3 timepoints: preconception, 26–28 weeks gestation, and 3 months postpartum	LC-MS/MS; quantification of 689 lipid species; analysis includes clustering, regression models, and pathway enrichment	Identified dynamic lipidomic changes during pregnancy; 56 % of lipids changed over time; key lipid clusters associated with BMI, weight gain, glycemic traits; potential early markers of cardiometabolic risk
Antonic et al. (2025), Serbia [6]	Prospective cohort study; N = 90 (70 high-risk controls, 20 developed late-onset preeclampsia)	Pregnant women at high risk for preeclampsia; followed across 4 gestational time points	Maternal serum; collected during 1st trimester, 2nd trimester, 3rd trimester, and delivery	Targeted sphingolipid profiling using LC-MS/MS	S1P, ceramides (C16:0, C24:0), and sphingomyelin C16:0 tracked over time; S1P significantly lower in PE group vs. controls from 2nd trimester onward; ceramides increased in high-risk group over time; sphingomyelin rose in both groups with no group difference
Rahman et al. (2021), USA	Prospective nested case-control; N = 321 (107 GDM, 214 controls)	Multiethnic pregnant women (NICHD Fetal Growth Studies); no preexisting diabetes; matched on site, age, race/ethnicity	Maternal plasma at 10–14 and 15–26 weeks	Untargeted lipidomics (420 metabolites, 328 annotated) using LC-MS; WGCNA; linear mixed models; FDR correction	Lipid networks enriched in diglycerides and saturated/low-unsaturated TGs were associated with increased GDM risk; 40 lipids significantly differed at both timepoints; lipids correlated with glucose, insulin, HbA1c, C-peptide
Wang et al. (2023), USA [20]	Prospective cohort study; N = 1409; 219 developed type 2 diabetes	Postpartum women (within 72 h after singleton delivery) from Boston Medical Center; followed for median 11.8 years	Plasma; collected at 24–72 h postpartum	Untargeted lipidomics via LC-MS/MS; lipidome-wide association; adjusted regression models	33 lipids associated with GDM (16 ↓ including CE, PC plasmalogens; 17 ↑ including DAGs, TAGs); 4 lipids also predicted T2D and mediated 12 % of GDM-to-T2D progression; improved T2D prediction beyond classical risk factors
Williams et al. (2023), USA [19]	Prospective cohort; N = 63 women with obesity (10 developed PE)	Pregnant women with obesity; followed longitudinally by trimester; evaluated by PE status, race, fetal sex	Maternal plasma; collected in each trimester	Targeted lipidomics and standard lipid panels; LC-MS/MS; stratified by trimester, race, and fetal sex	Elevated plasmalogens, PE, and FFA species in 3rd trimester in women with PE; race and pregnancy stage significantly influenced lipidomic profiles; standard lipids showed few differences
Song et al. (2023), USA/Singapore	Prospective longitudinal cohort; N = 321 pregnant women from NICHD Fetal Growth Study	Multiethnic cohort; healthy singleton pregnancies; followed through 4 visits during pregnancy	Maternal plasma; collected at 10–14, 15–26, 23–31, and 33–39 weeks	Untargeted lipidomics via LC-MS/MS; WGCNA and consensus network analysis; FDR-controlled linear mixed models	TGs positively associated with birthweight, head circumference; CEs, PCs, SMs, PEs, and LPCs inversely associated with multiple neonatal size measures; distinct lipid modules linked to fetal growth parameters across gestation
Enthoven et al. (2023), USA [15]	Within-subject longitudinal study; N = 47 pregnant women	Healthy women aged 18–50; paired samples during pregnancy and ∼3 months postpartum	Maternal plasma; 25–28 weeks gestation and ∼3 months postpartum	UPLC-MS/MS metabolomics for 43 sphingolipids; followed by targeted quantitative LC-MS/MS analysis	35 of 43 sphingolipids differed significantly between pregnancy and postpartum; 32 higher during pregnancy, especially sphingomyelins and ceramides; consistent with adaptations in maternal lipid metabolism during gestation
Traila et al. (2025), Romania	Mixed longitudinal and cross-sectional study; N = 107 (65 pregnant women, 42 postpartum controls)	Healthy women; pregnant group sampled across three trimesters; postpartum group used as reference	Plasma; collected at 6–14 weeks, 14–22 weeks, >24 weeks gestation and postpartum	UHPLC-QTOF-ESI+-MS lipidomics; multivariate analysis (PLS-DA), VIP scores, and clustering	Significant lipidomic shifts across pregnancy; 16 lipids showed consistent changes, including PCs, SMs, ceramides, and glycerolipids; lipids discriminated pregnant vs. postpartum states; potential biomarkers for gestational progression
Aung et al. (2021), USA [12]	Nested case-control from a birth cohort; N = 100 (50 preterm, 50 term)	Diverse pregnant women from San Francisco; samples matched by gestational age at collection	Maternal plasma at 15–20 weeks gestation	Untargeted lipidomics via LC-MS/MS; 387 lipids quantified	38 lipid species significantly associated with preterm birth; lower levels of PCs, PEs, and SMs observed in preterm group; lipidome may aid early risk stratification for preterm delivery
Mustaniemi et al. (2023), Finland [14]	Nested case–control study within FinnGeDi; N = 264 (132 GDM, 132 controls)	Pregnant women from the general population in Finland; matched for age and gestational age	Maternal serum; collected at ∼13 weeks gestation (early pregnancy)	Targeted lipidomics via LC-MS/MS; ceramides and traditional lipids; 4 ceramide species and Cer(d18:1/18:0)/Cer(d18:1/16:0) ratio analyzed	Higher levels of Cer(d18:1/18:0), Cer(d18:1/24:1), and the Cer(18:0/16:0) ratio in women who developed GDM; predictive potential of ceramides independent of BMI; supports use of ceramides as early biomarkers for GDM risk
Miranda et al. (2018), Spain [17]	Prospective cohort study; N = 80 (28 AGA, 25 SGA, 27 FGR)	Pregnant women at term; classified as AGA (controls), SGA, or FGR (cases); recruited in Barcelona	Maternal plasma and umbilical cord plasma; collected at delivery (non-fasting)	1H NMR spectroscopy; Liposcale for lipoprotein profiling and phosphatidylcholines; Dolphin for low-molecular-weight metabolites	Lower maternal IDL, VLDL, HDL cholesterol/triglycerides in SGA/FGR vs. AGA; higher fetal VLDL/IDL cholesterol and triglycerides in FGR; altered phosphatidylcholines and glycoproteins; supports disrupted lipid metabolism in fetal growth restriction and potential long-term metabolic risks

AGA – Appropriate for Gestational Age; ApoA1 – Apolipoprotein A1; ApoB – Apolipoprotein B; AUC – Area Under the Curve; BMI – Body Mass Index; BP – Blood Pressure; CE – Cholesteryl Ester; CHD – Congenital Heart Disease; DAG – Diacylglycerol; EOPE – Early-Onset Preeclampsia; EFW – Estimated Fetal Weight; FA – Fatty Acid; FDR – False Discovery Rate; FFA – Free Fatty Acid; FGR – Fetal Growth Restriction; GDM – Gestational Diabetes Mellitus; HDL – High-Density Lipoprotein; IDL – Intermediate-Density Lipoprotein; LC-MS – Liquid Chromatography-Mass Spectrometry; LC-MS/MS – Liquid Chromatography-Tandem Mass Spectrometry; LDL – Low-Density Lipoprotein; LOPE – Late-Onset Preeclampsia; LPC – Lysophosphatidylcholine; MetaboAnalyst – Metabolomics Analysis Platform; NICHD – National Institute of Child Health and Human Development; NMR – Nuclear Magnetic Resonance; OPLS-DA – Orthogonal Partial Least Squares Discriminant Analysis; PCA – Principal Component Analysis; PC – Phosphatidylcholine; PE – Preeclampsia/Phosphatidylethanolamine (context-dependent); PL – Phospholipid; PLS-DA – Partial Least Squares Discriminant Analysis; PS – Phosphatidylserine; RF – Random Forest; ROC – Receiver Operating Characteristic; S1P – Sphingosine-1-Phosphate; SBP – Systolic Blood Pressure; SGA – Small for Gestational Age; SM – Sphingomyelin; S-PRESTO – Singapore Preconception Study of Long-Term Maternal and Child Outcomes; TAG – Triacylglycerol; TC – Total Cholesterol; TG – Triglyceride; T2D – Type 2 Diabetes; UHPLC – Ultra High Performance Liquid Chromatography; UHPLC-Q-Exactive Orbitrap MS – Ultra High Performance Liquid Chromatography Quadrupole Exactive Orbitrap Mass Spectrometry; UHPLC-QTOF-ESI⁺-MS – Ultra High Performance Liquid Chromatography Quadrupole Time of Flight Electrospray Ionization Mass Spectrometry; VIP – Variable Importance in Projection; VLDL – Very-Low-Density Lipoprotein; WGCNA – Weighted Gene Co-expression Network Analysis.

3.1. Preeclampsia

Lipidomic alterations were consistently reported in women who developed preeclampsia, with distinct profiles observed across severe, early-onset preeclampsia (EOPE), and late-onset preeclampsia (LOPE) subtypes. In a nested case-control study by He et al. [2], eleven lipid species—including LPC 15:0, PC 35:1e, and PE 37:2—were significantly associated with severe preeclampsia. These lipids were enriched in pathways related to insulin signaling, immune modulation, and phospholipid metabolism. Yang et al. [5] demonstrated that LOPE was linked to elevated triglycerides (particularly unsaturated fatty acids) and reduced levels of phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylserine (PS). The glycerophospholipid metabolic pathway was notably implicated in both studies.

Further supporting this, Huang et al. [13] reported strong lipidomic separation between EOPE, LOPE, and control pregnancies, with lipid profiles correlating with fetal birth weight and proteinuria. Antonic et al. [6] added a longitudinal perspective, showing that ceramide species (C16:0, C24:0) and sphingosine-1-phosphate (S1P) displayed distinct temporal trajectories in women who later developed preeclampsia, with predictive differences emerging from the second trimester.

Williams et al. [19] conducted a prospective cohort study in the USA involving 63 pregnant women with obesity, 10 of whom developed preeclampsia. Maternal plasma was analyzed across all three trimesters using targeted lipidomics via LC-MS/MS, with stratification by race and fetal sex. They found that women with preeclampsia exhibited significant elevations in plasmalogens, PEs, and free fatty acids (FFAs) in the third trimester. Importantly, both race and gestational stage significantly modulated the lipidomic profiles, whereas conventional lipid panels revealed minimal differences.

The observation that lipidomic signatures were modified by race and fetal sex [19] highlights the importance of considering demographic and biological diversity in biomarker development. These differences carry significant clinical implications, as predictive lipid panels may need to be tailored to specific populations to ensure equitable risk stratification.

3.2. Gestational diabetes mellitus (GDM)

Several studies investigated the role of maternal lipid profiles in predicting GDM. Rahman et al. [3] found that lipid networks enriched in diglycerides and saturated or low-unsaturated triglycerides were positively associated with GDM risk. These findings were consistent across two gestational time points and showed strong correlations with insulin, glucose, and hemoglobin A1c (HbA1c) levels. Similarly, Mustaniemi et al. [14] identified ceramide species—particularly Cer(d18:1/18:0) and Cer(d18:1/24:1)—as predictive markers in early pregnancy, independent of body mass index (BMI), supporting their role as early biomarkers for GDM. Ceramides showed consistent predictive value in early gestation, being associated with subsequent risk of both GDM [14] and preeclampsia [6]. This supports their role as early biomarkers rather than late-stage disease correlates.

3.3. Preterm birth

In a case-control study, Aung et al. [12] reported that 38 lipid species were significantly associated with preterm birth. The preterm group showed lower levels of PCs, PEs, and SMs, suggesting impaired membrane lipid homeostasis. These findings highlight the potential for early lipidomic risk stratification in spontaneous preterm delivery.

3.4. Fetal growth and congenital conditions

Lipidomic profiles were also associated with fetal outcomes beyond maternal complications. Mires et al. [9] demonstrated that maternal plasma lipidomics could distinguish mothers of infants with CHD from controls, particularly through differences in PC species and acylcarnitines. Song et al. [16] reported that certain lipid classes—such as triglycerides (TGs) positively, and CEs, PCs, sphingomyelins (SMs), PEs, and lysophosphatidylcholines (LPCs) negatively—were significantly associated with birthweight and head circumference, suggesting lipidomic influence on fetal growth trajectories. Additionally, Miranda et al. [17] found that mothers of infants with fetal growth restriction (FGR) or SGA had lower levels of maternal intermediate-, very-low-, and high-density lipoprotein cholesterol and triglycerides compared to appropriate for gestational age (AGA) controls, while fetal very-low-density lipoprotein (VLDL) and intermediate-density lipoprotein (IDL), cholesterol and triglycerides were elevated in FGR cases. These findings support the notion that disrupted maternal and fetal lipid metabolism is intricately linked to impaired fetal growth and may have long-term metabolic implications.

3.5. Postpartum cardiometabolic risk

Chen et al. [1] and Wang et al. [20] extended lipidomic profiling into the postpartum period. Wang et al. identified specific lipid species [e.g., diacylglycerols (DAGs), CEs, TGs) that predicted the development of type 2 diabetes among women with prior GDM [20]. These markers improved prediction beyond traditional clinical factors and mediated part of the risk transition from GDM to long-term diabetes.

3.6. Consistently implicated lipid classes

Across the included studies, several lipid classes were recurrently associated with adverse pregnancy outcomes, indicating potential shared metabolic pathways. Alterations in PCs and PEs were observed across multiple biospecimen types—including maternal plasma [5,13], placenta [4], and cord blood [9,16]. This cross-compartmental consistency suggests placental transport and fetal exposure to dysregulated maternal lipid metabolism. PCs were differentially expressed in cases of preeclampsia [2,5,13], GDM [20], FGR and CHD [9,16], and preterm birth [12]. PEs were similarly altered in preeclampsia [2,5,13], preterm birth [12], and in association with fetal growth in the context of GDM [16]. LPCs were notably elevated in severe preeclampsia [2] and inversely associated with neonatal size parameters [16]. TGs were significantly increased in late-onset preeclampsia [5] and GDM [3] and were also linked to fetal birthweight and head circumference [16]. Ceramides and other sphingolipids were implicated in the early prediction of preeclampsia [6], GDM [14], and exhibited dynamic changes between pregnancy and the postpartum period [15]. CEs were found to be decreased in GDM and were associated with the progression to postpartum type 2 diabetes [20], as well as with reduced fetal growth parameters [16]. Finally, DAGs were elevated in GDM and were predictive of maternal cardiometabolic risk postpartum [20]. The convergence of these lipid classes across multiple gestational complications underscores the likelihood of shared pathophysiological mechanisms and supports their potential utility as early biomarkers for risk stratification and clinical decision-making in maternal-fetal health.

To better illustrate the overlap of lipidomic alterations across outcomes, we summarized the lipid classes most consistently implicated in different pregnancy complications (Table 2). This highlights both shared metabolic perturbations (e.g., PCs, TGs, ceramides) and outcome-specific patterns (e.g., acylcarnitines in congenital anomalies).

Table 2.

Summary of lipid classes associated with adverse pregnancy outcomes.

Lipid class	Preeclampsia	Gestational diabetes (GDM)	Preterm birth	FGR/SGA	Congenital anomalies (CHD)	Postpartum cardiometabolic risk
Phosphatidylcholines (PCs)	↑/↓ [2,5,13]	Altered [20]	↓ [12]	↓ [16]	Altered [9]	–
Phosphatidylethanolamines (PEs)	Altered [2,5,13,19]	–	↓ [12]	↓ [16]	–	–
Lysophosphatidylcholines (LPCs)	↑ [2]	–	–	↓ [16]	–	–
Triglycerides (TGs)	↑ [5]	↑ [3]	–	↑/↓ [16,18]	–	↑ [20]
Ceramides/sphingolipids	Altered [6,13]	↑ [14]	–	–	–	Dynamic [15]
Cholesteryl esters (CEs)	↓ [3,20]	↓ [3,20]	–	↓ [16]	–	↓ [20]
Diacylglycerols (DAGs)	–	↑ [3]	–	–	–	↑ [20]
Acylcarnitines	–	–	–	–	Altered [9]	–

4. Discussion

The present systematic review synthesized evidence from 16 studies investigating lipidomic signatures in maternal and fetal biospecimens associated with adverse pregnancy outcomes. The studies included in this review were conducted across many countries and encompassed over 5000 participants. Using both targeted and untargeted lipidomic platforms, the studies analyzed maternal plasma, serum, and placental tissues collected at various gestational stages and postpartum. The key outcomes explored included preeclampsia, GDM, preterm birth, fetal, FGR, congenital anomalies, and postpartum cardiometabolic risk. Despite methodological heterogeneity, a consistent pattern emerged, with specific lipid classes—PCs, PEs, LPCs, TGs, ceramides, CEs, and DAGs—repeatedly implicated across different pregnancy complications. These dysregulated lipid species were mapped onto relevant biological pathways such as glycerophospholipid metabolism, insulin signaling, and inflammation [2,3,5,16,20].

While comprehensive systematic reviews focusing exclusively on lipidomics in pregnancy are lacking, several recent studies have explored the relationship between maternal lipid profiles and adverse pregnancy outcomes. A retrospective cohort study involving over 2000 women with GDM found that elevated mid-pregnancy TG levels were independent risk factors for preeclampsia, preterm birth, macrosomia, postpartum hemorrhage, and intrauterine fetal distress—even under conditions of adequate glycemic control [21].

In a broader context, Preda et al. conducted a systematic review analyzing molecular alterations in lipid metabolism during pregnancy and their contribution to complications such as preeclampsia and intrauterine growth restriction. The review highlighted how dysregulated lipid pathways may contribute to both maternal and fetal adverse outcomes [22]. Similarly, a systematic review by Wang et al. found that maternal lipid levels are significantly associated with the delivery of SGA infants. Specifically, lower maternal concentrations of TGs and LDL cholesterol were associated with an increased risk of SGA births [23]. Moreover, a study by Yamamoto et al. demonstrated that women with GDM had significantly higher plasma TG concentrations compared to normoglycemic pregnant women, emphasizing the potential of lipid parameters as early indicators of metabolic dysfunction during pregnancy [24]. While conventional lipid measures such as total cholesterol, triglycerides, and LDL have been associated with pregnancy complications in systematic reviews and meta-analyses [[21], [22], [23], [24]], these parameters provide only a coarse view of lipid metabolism. Lipidomics advances this field by resolving hundreds of individual lipid species across diverse classes, thereby uncovering molecular signatures that are not captured by routine lipid panels. For instance, whereas elevated triglycerides on standard tests have been linked to GDM and preeclampsia, lipidomic studies differentiate between saturated, unsaturated, and chain-length–specific triglycerides, many of which were detectable in early gestation [3,5,14]. Similarly, phosphatidylcholines, ceramides, and sphingolipids identified through lipidomics provide mechanistic insights into endothelial dysfunction and insulin resistance that conventional lipid markers cannot capture. Thus, positioning lipidomics as an extension and refinement of standard lipid profiling underscores its novelty and translational potential in pregnancy research.

Ceramides and sphingolipids have been implicated in mediating placental dysfunction and inflammatory responses associated with preeclampsia. For instance, Ermini et al. [25] demonstrated that in EOPE, ceramide accumulation in trophoblast cells induces lysosomal biogenesis and exocytosis, leading to the release of ceramide-enriched exosomes containing active lysosomal sphingomyelin phosphodiesterase 1 (L-SMPD1) into the maternal circulation. This process contributes to endothelial dysfunction through paracrine signaling mechanisms. Similarly, Charkiewicz et al. [26] reported altered sphingolipid metabolism in human preeclamptic placentas, implicating disruptions in this lipid class in abnormal trophoblast behavior.

Experimental studies have also shown that disruptions in glycerophospholipid homeostasis and elevated triglyceride synthesis are closely linked to insulin resistance and gestational metabolic disturbances. Tanaka et al. [27] observed impaired lipid metabolism in the placenta of insulin-resistant animal models, providing direct support for the lipid alterations observed in GDM. In broader terms, Kell and Oliver [28] emphasized that lipid metabolic dysregulation is a recurrent feature across obstetric syndromes, although their review did not explore lipid class-specific details. Thus, the present synthesis adds mechanistic specificity and translational relevance by delineating consistent lipidomic signatures across multiple pregnancy outcomes.

The repeated identification of ceramides, TGs, and PCs across the included studies points to shared pathophysiological processes. Ceramides, as bioactive sphingolipids, are known to influence endothelial barrier function, promote oxidative damage, and initiate apoptosis—all mechanisms that contribute to the pathogenesis of preeclampsia and intrauterine growth restriction. Elevated TGs and DAGs may reflect underlying insulin resistance and disrupted energy partitioning, both of which are prominent features of GDM. Furthermore, disturbances in PC and PE metabolism may impair cell membrane integrity and intracellular signaling, thereby interfering with placental nutrient transport and vascular development. These converging mechanistic pathways provide a biologically plausible framework for understanding the lipidomic alterations associated with adverse pregnancy outcomes [[29], [30], [31], [32]].

From a clinical perspective, these findings underscore the promise of lipidomics as a tool for early risk assessment and personalized pregnancy care. Several studies identified lipid alterations in the first or early second trimester—well before clinical onset—suggesting their utility as predictive biomarkers for complications such as preeclampsia and GDM [2,6,14]. Moreover, postpartum lipidomic profiles, particularly those involving DAGs and CEs, were associated with long-term risk for type 2 diabetes, highlighting their relevance for postnatal surveillance [1,20]. These findings may inform the development of lipid-based screening panels or monitoring strategies in high-risk pregnancies and the postpartum setting.

This review has several strengths. It is, to our knowledge, the first systematic synthesis of lipidomic studies across multiple pregnancy outcomes, integrating both targeted and untargeted methods. A comprehensive search strategy, rigorous inclusion criteria, and standardized quality assessment enhance the reliability of findings.

However, limitations must be acknowledged. Considerable heterogeneity in analytical platforms, biospecimen types, and lipid nomenclature limits the comparability of results and precludes meta-analysis. Most included studies were observational, introducing potential bias and limiting causal inference. Another important limitation relates to potential confounding factors. Maternal diet, pre-pregnancy BMI, ethnicity, and the use of medications (e.g., antihypertensives, insulin, or lipid-lowering agents) can all substantially influence lipidomic profiles. These variables were not uniformly measured or adjusted for across the included studies, which may have contributed to heterogeneity in findings. For example, BMI and diet are strong modifiers of circulating lipid species, and population-specific differences in lipid metabolism may affect the generalizability of biomarker candidates. The lack of consistent adjustment for such confounders reduces certainty regarding whether the observed lipid alterations reflect true disease-related signatures or underlying lifestyle and demographic differences.

Although all included studies scored well on the NOS, there are inherent limitations to this assessment. First, the NOS is designed for observational studies and may not fully capture biases specific to lipidomic research, such as analytical variability, batch effects, or differences in lipid annotation. Second, while most studies were rated as high quality, the lack of uniform adjustment for confounders (e.g., maternal BMI, diet, ethnicity) may have introduced residual bias. Third, heterogeneity in biospecimen types, timing of collection, and lipidomic platforms limits the comparability of results despite good overall methodological ratings. Finally, the small number of studies per outcome reduces the power of quality assessment to detect systematic weaknesses.

Translating lipidomic signatures into clinical care will require careful consideration of feasibility, cost-effectiveness, and standardization. Current lipidomic platforms, while increasingly accessible, remain more expensive and technically demanding than conventional lipid assays, limiting their immediate scalability. Moreover, substantial methodological heterogeneity exists in sample processing, analytical platforms, and lipid annotation, underscoring the need for standardized workflows before widespread adoption. Integration into existing obstetric risk models will also be essential. For example, combining lipidomic markers with established predictors such as maternal age, BMI, and blood pressure could enhance the accuracy of early risk stratification tools. Ultimately, prospective validation in large, multiethnic cohorts, coupled with health-economic evaluations, will determine whether lipidomics can transition from a research tool to a clinically viable approach in pregnancy care.

Future studies should prioritize methodological standardization in lipidomic workflows, reporting formats, and lipid annotation. Large-scale, longitudinal cohort studies with multiethnic representation and detailed confounder adjustment are needed to validate candidate lipid biomarkers. Furthermore, functional experiments and multi-omics integration (e.g., transcriptomics, proteomics) could provide mechanistic depth and enhance biomarker discovery.

5. Conclusions

In summary, this systematic review provides the first comprehensive synthesis of lipidomic signatures associated with adverse pregnancy outcomes, integrating data from maternal blood, placenta, and fetal biospecimens. Across a diverse cohort of studies, specific lipid classes—including PCs, PEs, ceramides, TGs, CEs, and DAGs—were repeatedly associated with preeclampsia, GDM, preterm birth, FGR, and postpartum metabolic risk. These findings underscore the biological and clinical significance of lipid metabolism in pregnancy and highlight the promise of lipidomics for early biomarker development. Despite methodological variability and observational study designs, the consistency of lipid pathway disruptions suggests shared pathophysiological mechanisms underlying obstetric syndromes. Future research should focus on validating these biomarkers in large, multiethnic cohorts and exploring their integration into predictive models for personalized obstetric care.

CRediT authorship contribution statement

Eleni Kalapouti: Writing – original draft, Visualization, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Anastasia Bothou: Writing – review & editing, Visualization, Resources, Data curation. Vikentia Harizopoulou: Writing – review & editing, Resources, Methodology, Data curation. Maria Vlachou: Writing – review & editing. Maria Vasiliki Zampeli: Writing – review & editing. Athina Diamanti: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Appendix A. Supplementary data

The following are the Supplementary data to this article.