Real-world safety profile of mitotane in adrenocortical carcinoma: a retrospective cross-sectional study integrating pharmacovigilance and interpretable machine learning.

Maternal–infant health is increasingly recognized as an ecosystem phenomenon, in which maternal and paternal microbial communities interact with endocrine and immune circuits to shape fertility, placental function, and neonatal development. Across the reproductive timeline, microbial states are not static “snapshots” but dynamic programs that coadapt with host physiology: during pregnancy, for example, the gut microbiome remodels in a gestation‐stage‐specific manner, and growing evidence supports a gut–placenta signaling axis in which microbial metabolites (e.g., short‐chain fatty acids (SCFAs), bile acids (BAs), and indoles) are decoded by placental receptors to tune immune tolerance, vascular remodeling, and metabolic reprogramming. After birth, infant gut assembly follows a nonrandom ecological succession that is highly sensitive to delivery mode, antibiotics, and nutrition. Breast milk serves as a sustained route for both microbial transmission and a selection of substrates that steer early community structure and function. Disruption of these time‐locked microbial trajectories is associated with major perinatal and early‐life morbidities, motivating a life‐course framework that integrates gut microbiomes, metabolite‐mediated crosstalk, and critical developmental windows. However, current microbiome research is often confined to a descriptive synthesis of taxonomic shifts, leaving several critical gaps unresolved or potentially misinterpreted. First, the field frequently overlooks the multikingdom nature of the early‐life ecosystem (including the virome, mycobiome, and resistome) and the emerging paradigm of biparental, particularly paternal microbial contributions. Second, there is a pervasive tendency in clinical studies to equate taxonomic “catch‐up” with true functional restoration. This misinterpretation is especially evident in preterm infants, where compositional recovery often masks persistent functional deficits. Another prominent example is vaginal microbiota transplantation (VMT) for cesarean section‐born infants; while VMT has been shown to partially restore the taxonomic structure of the infant microbiota, whether this intervention can sustainably achieve functional recovery or reduce the long‐term risk of diseases such as asthma and obesity remains unknown. Finally, studies predominantly characterize the “critical window” chronologically and lack a clear understanding of the obligate temporal coupling between microbial metabolic succession and host physiological milestones. To bridge these gaps, this review moves beyond descriptive observations to establish a “developmental synchronization” conceptual framework. Here, we synthesize current evidence from preconception to early childhood, explicitly mapping the “microbiota–metabolite–host target–physiological phenotype” regulatory network. By identifying areas that are poorly understood, such as the persistence of functional dysbiosis and the scarcity of intergenerational longitudinal cohorts, we aim to guide the field away from generic taxonomic profiling toward a mechanistic understanding of early‐life dysbiosis and the development of precision‐targeted, function‐restoring interventions (Figure ).