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.