A notable “upstreaming” trend this week: using endocrine physiology signals earlier in life to forecast later metabolic outcomes. An International Journal of Obesity mother–newborn–offspring study uses machine learning to predict childhood overweight at age 2 leveraging maternal thyroid function/iodine-related parameters plus anthropometrics [1]. In the same week, the obesity multimodal biomarker scoping review explicitly positions endocrine markers (alongside multi-omics, behavior, microbiome, etc.) as part of the combined feature space where ML/AI becomes necessary to integrate and interpret risk [2]. Together, this supports a growing life-course framing: risk prediction starts in pregnancy (or earlier), and “endocrine context” is part of the model, not just a covariate [1] [2].
Why it matters / what changes in tactics
-
Clinicians (especially prenatal/postnatal care): Expect more proposals for “prenatal risk flags.” The practical move is to keep these as adjunctive, not determinative—use them to target education/follow-up intensity rather than to label a child as destined for obesity (risk ≠ fate) [1].
-
Researchers: The key next steps are (a) prospective validation across iodine-sufficient vs iodine-deficient settings, (b) careful handling of measurement timing (when thyroid/iodine markers are drawn), and (c) evaluation of whether risk prediction actually changes outcomes when paired with an intervention pathway [1] [2].
References
-
Ovadia YS, Bilenko N, Mazza O, et al. A prediction model for childhood obesity risk based on maternal thyroid status and related parameters using machine learning: a mother-newborn-offspring study in a mild-to-moderate iodine deficiency area. Int J Obes (Lond). Published online December 26, 2025. doi: https://doi.org/10.1038/s41366-025-01988-y.
PubMed: https://pubmed.ncbi.nlm.nih.gov/41454183/ -
Vahid F, Loyola-Leyva A, Tur J, et al. Multimodal (Bio)Markers and Risk of Obesity - A Comprehensive Scoping Review. Adv Nutr. Published online December 24, 2025. doi: https://doi.org/10.1016/j.advnut.2025.100579.
PubMed: https://pubmed.ncbi.nlm.nih.gov/41453658/