The longitudinal study quantified 147 metabolites using nuclear magnetic resonance spectroscopy in fasting maternal blood samples. The team then modelled children’s growth in weight, length or height, and body mass index from birth to 36 months using latent class growth mixture models. This approach identified three distinct trajectories for each anthropometric measure: a normative pattern, an intermediate pattern, and a non-uniform pattern marked either by accelerated weight and BMI gain or by faltering length and height.

To pinpoint which maternal metabolites set the trajectories apart, the researchers applied partial least squares discriminant analysis. Between 30 and 41 metabolites were needed to differentiate trajectories within each anthropometric measure. Fourteen lipoprotein subclasses were required across weight, length or height, and BMI, with an additional 16 to 31 metabolites contributing in one or two of the trajectories. The prominent role of lipoproteins led the authors to highlight these lipid-carrying particles as especially characteristic of the observed offspring growth patterns.

Beyond classifying trajectories, the study examined how individual metabolite levels related to changes in World Health Organization growth z-scores, using linear mixed-effects regression adjusted for a minimal set of covariates: ethnicity and gestational diabetes mellitus. Numerous associations were observed and validated within each growth trajectory, with effect estimates per unit change in metabolite ranging from −0.47 to 1.5 for weight, −0.87 to 2.2 for length or height, and −0.85 to 1.8 for BMI, all at a statistical significance threshold of P < 0.05. The wide ranges indicate that the direction and magnitude of these associations varied considerably across metabolites and trajectory classes.

The analysis also quantified the contributions of the two covariates. South Asian ethnicity and gestational diabetes were each associated with lower z-scores across the anthropometric measures. For weight, the associations ranged from −0.214 to −0.212 for South Asian ethnicity and −0.044 to −0.041 for gestational diabetes; for length or height, from −0.061 to −0.045 and −0.104 to −0.101 respectively; and for BMI, from −0.157 to −0.182 and −0.083 to −0.071 respectively. These estimates describe associations within this multi-UK cohort rather than causal effects.

The study’s scope carries important limitations. It is an observational analysis of a single birth cohort, and the reported associations between maternal metabolites and child growth do not establish that the metabolites cause differences in growth. The covariate adjustment was deliberately minimal, covering only ethnicity and gestational diabetes, so other maternal and environmental factors could contribute to the observed patterns. The abstract reports associations validated within trajectory classes, but the findings describe characterisation of growth patterns rather than prediction or clinical utility.

By showing that the same metabolite may relate to growth differently depending on which trajectory a child follows, the findings suggest that a single average association across a whole population may mask meaningful heterogeneity. The authors conclude that mid-pregnancy metabolomic profiles, especially lipoproteins, characterise observed offspring growth in different directions and strengths across growth trajectories from birth to 36 months of age.

Subject of Research: Maternal mid-pregnancy metabolomic profiles and offspring growth trajectories from birth to 36 months in a UK multi-ethnic birth cohort

Article Title: Maternal mid-pregnancy metabolic profiles and growth trajectories across the first 36 months of life in a UK multi-ethnic birth cohort

Article References: Zielinska-Pukos, M. A., & Zulyniak, M. A. (2026). Maternal mid-pregnancy metabolic profiles and growth trajectories across the first 36 months of life in a UK multi-ethnic birth cohort. International Journal of Obesity. https://doi.org/10.1038/s41366-026-02233-w

Image Credits: AI Generated

DOI: 10.1038/s41366-026-02233-w

Keywords: metabolomics, pregnancy, child growth, Born in Bradford, lipoproteins, BMI, growth trajectories, gestational diabetes, NMR spectroscopy, birth cohort, Maternal, mid-pregnancy