Fig. 3

Associations between first (a) and third (b) trimester maternal urinary metabolites and fetal growth scores in the Gipuzkoa and Sabadell subcohorts. Percent of change in the standard deviation scores of the fetal growth measure (customized growth curve for each fetus) are presented as a heat map where the colour scale corresponds to the strength of the change for an increase of 50% in the metabolite spectral integral. Only metabolites which were significantly associated with fetal growth in both subcohorts and with a P value < 5% are presented. All the models were adjusted for the time of sampling during gestation and models for birth weight and placental weight were further adjusted for maternal weight and height, father’s height, parity, child’s sex, and gestational age at sampling and at birth. (1) Number of individuals included in the adjusted models, that is, number without missing values for any variable included in each model. *Tentative assignment. 2Py N-methyl- 2-pyridone-5-carboxamide, 3-HBA/3-ABA 3-hydroxybutyrate/3-aminoisobutyrate, UA unassigned, d doublet, Pregnanediol-3-G 5β-pregnane-3α20α-diol-3α-glucuronide, TMAO Trimethylamine oxide. Weeks 12, 20, and 34 stand for the gestational week at which the fetal growth outcome was calculated or the conditional growth between these weeks