Abstract
The etiological heterogeneity of spontaneous preterm birth (sPTB) complicates PTB risk prediction and management. Accurate risk stratification at clinical suspicion is critical to administer life-saving interventions for fetuses at genuine risk, while avoiding unnecessary treatments in pregnancies progressing to term to mitigate long-term complications. Current tools lack predictive accuracy, with 40% of suspected PTB delivering at term. This pilot study investigated signaling lipids as potential biomarkers for sPTB. Prospectively collected midstream urine of 30 women with imminent PTB were retrospectively classified into preterm without chorioamnionitis, preterm with chorioamnionitis, and term, and analyzed using liquid-chromatography mass-spectrometry. Overall-preterm vs. term analysis revealed reduced lipoxygenase- and cytochrome P450-derived oxylipins in the preterm group, suggesting impaired inflammation resolution. Pairwise comparisons showed that these differences were primarily driven by the non-infectious preterm group. Notably, 8,9-DiHETrE and 9-HODE emerged as promising etiology-independent biomarkers for sPTB. Furthermore, oxidative stress dominated non-infectious sPTB cases, while inflammatory cascades characterized the infectious cases. A predictive model incorporating 9-HODE and vaginal discharge demonstrated the ability (AUC 84.2%, sensitivity 73.7%, specificity 88.9%) to differentiate between women delivering preterm and at term. These findings highlight the potential of urinary metabolomics as a non-invasive tool to unravel sPTB pathophysiology and enhance risk prediction.