#AI reads Urine# Integrating urine metabolomic biomarkers and machine learning algorithms to predict preeclampsia

This study aimed to develop a non-invasive method for the early prediction of preeclampsia (a major hypertensive disorder during pregnancy that causes maternal and fetal morbidity and mortality). It recruited 104 healthy pregnant women, 80 patients with gestational hypertension, and 102 patients with preeclampsia, conducted untargeted metabolomic analysis on the participants' urine samples collected before delivery using ultra-high-performance liquid chromatography coupled with mass spectrometry, screened for differential metabolites via multivariate statistical methods, and constructed a predictive model with machine learning algorithms. The results showed that there were 55 significantly abnormal metabolites in preeclampsia patients compared with healthy pregnant women, and 22 significantly abnormal metabolites when preeclampsia patients were compared with gestational hypertension patients; four metabolites—estriol-17-glucuronide, diethylphosphate, 4-deoxythreonic acid, and taurine—were identified as core predictive indicators (with estriol-17-glucuronide having the best predictive performance and a positive correlation with blood pressure), and the abnormal changes of these metabolites were associated with disorders in vitamin B6 metabolism, steroid hormone biosynthesis, and histidine metabolism pathways. The predictive model constructed based on these four metabolites achieved high accuracy in predicting preeclampsia, providing a non-invasive approach for the early clinical screening of preeclampsia and new insights into understanding the metabolic pathogenesis of the disease. The study also noted limitations such as single-center samples and lack of dynamic metabolic data during pregnancy, indicating that future multi-center validation and further mechanistic research are needed to promote clinical translation.

Eur J Med Res. 2025 Nov 10;30(1):1103. doi: 10.1186/s40001-025-03337-1.

Youhe Gao

Statement: During the preparation of this work the author(s) used Doubao / AI reading for summarizing the content. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.

For earlier AI Reads Urine articles：

https://www.keaipublishing.com/en/journals/advances-in-biomarker-sciences-and-technology/ai-reads-urine/