Serum Element-Driven Gut Microbial Disturbances and Their Predictive Implications for Gestational Dyslipidemia: A...

Pregnancy-triggered shifts in maternal serum elements and gut microbiota may disrupt metabolic balance; yet, their interplay in lipid regulation remains elusive. This study aimed to decode how metal elements and gut microbiota across mid-to-late pregnancy contribute to dyslipidemia and enable early risk prediction. In 689 pregnant women, we measured serum metals and profiled gut microbiota using 16S rRNA sequencing, applying interpretable machine learning and causal inference to unravel element-microbiota-dyslipidemia interactions. Higher serum levels of copper and zinc were linked to adverse lipid profiles, while cobalt showed inverse associations. Metals were also associated with broad alterations in microbial composition and diversity, including shifts in alpha diversity and over 30 significant taxa-level associations, primarily within Firmicutes and Bacteroidota. Microbiota features across pregnancy exhibited extensive associations with lipid metabolism, with reduced alpha diversity linked to elevated levels of triglycerides. Finally, a CatBoost model integrating midpregnancy metals and microbiota robustly predicted late gestational dyslipidemia (AUC = 0.91; AP = 0.94), with SHAP analysis revealing risk-defining microbial thresholds and taxa specific nonlinear effects. These findings demonstrate that maternal elements and microbiota jointly contribute to the development of dyslipidemia, supporting potential strategies for the early identification and prevention of gestational dyslipidemia.