Non-Targeted Plasma Lipidomic Profiling in Late Pregnancy and Early Postpartum Stages: An Observational Comparative...

Background/Objectives: Pregnancy represents a unique physiological state marked by extensive metabolic adaptations, particularly in lipid pathways essential for maternal adjustments, fetal development, and postpartum recovery. This study aimed to explore these changes through untargeted lipidomic profiling. Methods: This observational, comparative, non-interventional clinical study included 107 women, of which 65 were in the third trimester of pregnancy (mean age 27.9 ± 5 years) and 42 were in the early postpartum period (≤7 days, mean age 28.9 ± 5.9 years). Inclusion criteria were singleton, term pregnancies (37-41 weeks) with neonates weighing > 2500 g and no associated pregnancy-related pathologies; exclusion criteria included multiple gestation, use of lipid-altering medications, maternal age > 40 years, or diagnosed pregnancy complications. Plasma samples were analyzed using High-Performance Liquid Chromatography-Quadrupole Time-Of-Flight-Electrospray Ionization (positive mode)-Mass Spectrometry, data were processed with MetaboAnalyst 6.0 using multivariate and univariate analyses (Partial Least Squares-Discriminant Analysis, Volcano Plot, Random Forest, Receiver Operating Characteristic analysis), with statistical significance set at p < 0.05. Results: Multivariate analysis demonstrated a clear separation between groups with high predictive accuracy as reflected by strong classification metrics (Accuracy = 0.90, R² = 0.75, Q² = 0.68). Several discriminative lipids were consistently identified across statistical models, including 2-Methoxyestrone (AUC = 0.861), Eicosanedioic acid (AUC = 0.854), and Pregnenolone sulfate (AUC = 0.843). These biomarkers were further categorized into five major lipid classes: steroid hormones, long-chain fatty acids, lysophospholipids, ceramides/sphingolipids, and glycerolipids. Conclusions: Untargeted lipidomic profiling revealed distinct metabolic signatures that differentiate late pregnancy from early post-partum states. The identification of robust lipid biomarkers with high discriminative performance highlights their potential utility in maternal health monitoring, obstetric risk assessment, and postpartum recovery surveillance.